diff options
248 files changed, 23908 insertions, 14580 deletions
diff --git a/cmake/Findclang.cmake b/cmake/Findclang.cmake index b4bd80773d..3ce52df893 100644 --- a/cmake/Findclang.cmake +++ b/cmake/Findclang.cmake @@ -9,27 +9,27 @@ find_path(CLANG_INCLUDE_DIRS NAMES clang/Frontend/ASTUnit.h PATHS - /usr/lib/llvm/11/include - /usr/lib/llvm-11/include - /usr/lib/llvm-11.0/include - /usr/local/llvm110/include - /usr/local/llvm11/include + /usr/lib/llvm/12/include + /usr/lib/llvm-12/include + /usr/lib/llvm-12.0/include + /usr/local/llvm120/include + /usr/local/llvm12/include /mingw64/include ) if(ZIG_PREFER_CLANG_CPP_DYLIB) find_library(CLANG_LIBRARIES NAMES - clang-cpp-11.0 - clang-cpp110 + clang-cpp-12.0 + clang-cpp120 clang-cpp PATHS ${CLANG_LIBDIRS} - /usr/lib/llvm/11/lib - /usr/lib/llvm/11/lib64 - /usr/lib/llvm-11/lib - /usr/local/llvm110/lib - /usr/local/llvm11/lib + /usr/lib/llvm/12/lib + /usr/lib/llvm/12/lib64 + /usr/lib/llvm-12/lib + /usr/local/llvm120/lib + /usr/local/llvm12/lib ) endif() @@ -39,11 +39,11 @@ if(NOT CLANG_LIBRARIES) find_library(CLANG_${_prettylibname_}_LIB NAMES ${_libname_} PATHS ${CLANG_LIBDIRS} - /usr/lib/llvm/11/lib - /usr/lib/llvm-11/lib - /usr/lib/llvm-11.0/lib - /usr/local/llvm110/lib - /usr/local/llvm11/lib + /usr/lib/llvm/12/lib + /usr/lib/llvm-12/lib + /usr/lib/llvm-12.0/lib + /usr/local/llvm120/lib + /usr/local/llvm12/lib /mingw64/lib /c/msys64/mingw64/lib c:\\msys64\\mingw64\\lib diff --git a/cmake/Findlld.cmake b/cmake/Findlld.cmake index 3103601ff8..72724ecd1e 100644 --- a/cmake/Findlld.cmake +++ b/cmake/Findlld.cmake @@ -8,16 +8,16 @@ find_path(LLD_INCLUDE_DIRS NAMES lld/Common/Driver.h PATHS - /usr/lib/llvm-11/include - /usr/local/llvm110/include - /usr/local/llvm11/include + /usr/lib/llvm-12/include + /usr/local/llvm120/include + /usr/local/llvm12/include /mingw64/include) -find_library(LLD_LIBRARY NAMES lld-11.0 lld110 lld +find_library(LLD_LIBRARY NAMES lld-12.0 lld120 lld PATHS - /usr/lib/llvm-11/lib - /usr/local/llvm110/lib - /usr/local/llvm11/lib + /usr/lib/llvm-12/lib + /usr/local/llvm120/lib + /usr/local/llvm12/lib ) if(EXISTS ${LLD_LIBRARY}) set(LLD_LIBRARIES ${LLD_LIBRARY}) @@ -27,9 +27,9 @@ else() find_library(LLD_${_prettylibname_}_LIB NAMES ${_libname_} PATHS ${LLD_LIBDIRS} - /usr/lib/llvm-11/lib - /usr/local/llvm110/lib - /usr/local/llvm11/lib + /usr/lib/llvm-12/lib + /usr/local/llvm120/lib + /usr/local/llvm12/lib /mingw64/lib /c/msys64/mingw64/lib c:/msys64/mingw64/lib) diff --git a/cmake/Findllvm.cmake b/cmake/Findllvm.cmake index 02f22037b8..2ab2c23ad0 100644 --- a/cmake/Findllvm.cmake +++ b/cmake/Findllvm.cmake @@ -9,37 +9,37 @@ find_path(LLVM_INCLUDE_DIRS NAMES llvm/IR/IRBuilder.h PATHS - /usr/lib/llvm/11/include - /usr/lib/llvm-11/include - /usr/lib/llvm-11.0/include - /usr/local/llvm11/include - /usr/local/llvm110/include + /usr/lib/llvm/12/include + /usr/lib/llvm-12/include + /usr/lib/llvm-12.0/include + /usr/local/llvm12/include + /usr/local/llvm120/include /mingw64/include ) if(ZIG_PREFER_CLANG_CPP_DYLIB) find_library(LLVM_LIBRARIES NAMES - LLVM-11.0 - LLVM-11 - LLVM-110 + LLVM-12.0 + LLVM-12 + LLVM-120 LLVM PATHS ${LLVM_LIBDIRS} - /usr/lib/llvm/11/lib - /usr/lib/llvm/11/lib64 - /usr/lib/llvm-11/lib - /usr/local/llvm11/lib - /usr/local/llvm110/lib + /usr/lib/llvm/12/lib + /usr/lib/llvm/12/lib64 + /usr/lib/llvm-12/lib + /usr/local/llvm12/lib + /usr/local/llvm120/lib ) find_program(LLVM_CONFIG_EXE - NAMES llvm-config-11 llvm-config-11.0 llvm-config110 llvm-config11 llvm-config + NAMES llvm-config-12 llvm-config-12.0 llvm-config120 llvm-config12 llvm-config PATHS "/mingw64/bin" "/c/msys64/mingw64/bin" "c:/msys64/mingw64/bin" - "C:/Libraries/llvm-11.0.0/bin") + "C:/Libraries/llvm-12.0.0/bin") if ("${LLVM_CONFIG_EXE}" STREQUAL "LLVM_CONFIG_EXE-NOTFOUND") message(FATAL_ERROR "unable to find llvm-config") @@ -54,23 +54,23 @@ if(ZIG_PREFER_CLANG_CPP_DYLIB) OUTPUT_VARIABLE LLVM_CONFIG_VERSION OUTPUT_STRIP_TRAILING_WHITESPACE) - if("${LLVM_CONFIG_VERSION}" VERSION_LESS 11) - message(FATAL_ERROR "expected LLVM 11.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") + if("${LLVM_CONFIG_VERSION}" VERSION_LESS 12) + message(FATAL_ERROR "expected LLVM 12.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") endif() - if("${LLVM_CONFIG_VERSION}" VERSION_EQUAL 12) - message(FATAL_ERROR "expected LLVM 11.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") + if("${LLVM_CONFIG_VERSION}" VERSION_EQUAL 13) + message(FATAL_ERROR "expected LLVM 12.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") endif() - if("${LLVM_CONFIG_VERSION}" VERSION_GREATER 12) - message(FATAL_ERROR "expected LLVM 11.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") + if("${LLVM_CONFIG_VERSION}" VERSION_GREATER 13) + message(FATAL_ERROR "expected LLVM 12.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") endif() elseif(("${ZIG_TARGET_TRIPLE}" STREQUAL "native") OR ZIG_PREFER_LLVM_CONFIG) find_program(LLVM_CONFIG_EXE - NAMES llvm-config-11 llvm-config-11.0 llvm-config110 llvm-config11 llvm-config + NAMES llvm-config-12 llvm-config-12.0 llvm-config120 llvm-config12 llvm-config PATHS "/mingw64/bin" "/c/msys64/mingw64/bin" "c:/msys64/mingw64/bin" - "C:/Libraries/llvm-11.0.0/bin") + "C:/Libraries/llvm-12.0.0/bin") if ("${LLVM_CONFIG_EXE}" STREQUAL "LLVM_CONFIG_EXE-NOTFOUND") message(FATAL_ERROR "unable to find llvm-config") @@ -85,14 +85,14 @@ elseif(("${ZIG_TARGET_TRIPLE}" STREQUAL "native") OR ZIG_PREFER_LLVM_CONFIG) OUTPUT_VARIABLE LLVM_CONFIG_VERSION OUTPUT_STRIP_TRAILING_WHITESPACE) - if("${LLVM_CONFIG_VERSION}" VERSION_LESS 11) - message(FATAL_ERROR "expected LLVM 11.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") + if("${LLVM_CONFIG_VERSION}" VERSION_LESS 12) + message(FATAL_ERROR "expected LLVM 12.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") endif() - if("${LLVM_CONFIG_VERSION}" VERSION_EQUAL 12) - message(FATAL_ERROR "expected LLVM 11.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") + if("${LLVM_CONFIG_VERSION}" VERSION_EQUAL 13) + message(FATAL_ERROR "expected LLVM 12.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") endif() - if("${LLVM_CONFIG_VERSION}" VERSION_GREATER 12) - message(FATAL_ERROR "expected LLVM 11.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") + if("${LLVM_CONFIG_VERSION}" VERSION_GREATER 13) + message(FATAL_ERROR "expected LLVM 12.x but found ${LLVM_CONFIG_VERSION} using ${LLVM_CONFIG_EXE}") endif() execute_process( @@ -166,7 +166,7 @@ elseif(("${ZIG_TARGET_TRIPLE}" STREQUAL "native") OR ZIG_PREFER_LLVM_CONFIG) set(LLVM_LIBRARIES ${LLVM_LIBRARIES} ${LLVM_SYSTEM_LIBS}) if(NOT LLVM_LIBRARIES) - find_library(LLVM_LIBRARIES NAMES LLVM LLVM-11 LLVM-11.0) + find_library(LLVM_LIBRARIES NAMES LLVM LLVM-12 LLVM-12.0) endif() link_directories("${CMAKE_PREFIX_PATH}/lib") @@ -180,11 +180,11 @@ else() find_library(LLVM_${_prettylibname_}_LIB NAMES ${_libname_} PATHS ${LLVM_LIBDIRS} - /usr/lib/llvm/11/lib - /usr/lib/llvm-11/lib - /usr/lib/llvm-11.0/lib - /usr/local/llvm110/lib - /usr/local/llvm11/lib + /usr/lib/llvm/12/lib + /usr/lib/llvm-12/lib + /usr/lib/llvm-12.0/lib + /usr/local/llvm120/lib + /usr/local/llvm12/lib /mingw64/lib /c/msys64/mingw64/lib c:\\msys64\\mingw64\\lib) @@ -194,78 +194,57 @@ else() # This list can be re-generated with `llvm-config --libfiles` and then # reformatting using your favorite text editor. Note we do not execute # `llvm-config` here because we are cross compiling. - FIND_AND_ADD_LLVM_LIB(LLVMXRay) FIND_AND_ADD_LLVM_LIB(LLVMWindowsManifest) - FIND_AND_ADD_LLVM_LIB(LLVMSymbolize) - FIND_AND_ADD_LLVM_LIB(LLVMDebugInfoPDB) - FIND_AND_ADD_LLVM_LIB(LLVMOrcJIT) - FIND_AND_ADD_LLVM_LIB(LLVMOrcError) - FIND_AND_ADD_LLVM_LIB(LLVMJITLink) - FIND_AND_ADD_LLVM_LIB(LLVMObjectYAML) - FIND_AND_ADD_LLVM_LIB(LLVMMCA) - FIND_AND_ADD_LLVM_LIB(LLVMLTO) - FIND_AND_ADD_LLVM_LIB(LLVMPasses) - FIND_AND_ADD_LLVM_LIB(LLVMCoroutines) - FIND_AND_ADD_LLVM_LIB(LLVMObjCARCOpts) - FIND_AND_ADD_LLVM_LIB(LLVMExtensions) - FIND_AND_ADD_LLVM_LIB(LLVMLineEditor) + FIND_AND_ADD_LLVM_LIB(LLVMXRay) FIND_AND_ADD_LLVM_LIB(LLVMLibDriver) - FIND_AND_ADD_LLVM_LIB(LLVMInterpreter) - FIND_AND_ADD_LLVM_LIB(LLVMFuzzMutate) - FIND_AND_ADD_LLVM_LIB(LLVMMCJIT) - FIND_AND_ADD_LLVM_LIB(LLVMExecutionEngine) - FIND_AND_ADD_LLVM_LIB(LLVMRuntimeDyld) - FIND_AND_ADD_LLVM_LIB(LLVMDWARFLinker) FIND_AND_ADD_LLVM_LIB(LLVMDlltoolDriver) - FIND_AND_ADD_LLVM_LIB(LLVMOption) - FIND_AND_ADD_LLVM_LIB(LLVMDebugInfoGSYM) FIND_AND_ADD_LLVM_LIB(LLVMCoverage) + FIND_AND_ADD_LLVM_LIB(LLVMLineEditor) FIND_AND_ADD_LLVM_LIB(LLVMXCoreDisassembler) FIND_AND_ADD_LLVM_LIB(LLVMXCoreCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMXCoreDesc) FIND_AND_ADD_LLVM_LIB(LLVMXCoreInfo) FIND_AND_ADD_LLVM_LIB(LLVMX86Disassembler) - FIND_AND_ADD_LLVM_LIB(LLVMX86CodeGen) FIND_AND_ADD_LLVM_LIB(LLVMX86AsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMX86CodeGen) FIND_AND_ADD_LLVM_LIB(LLVMX86Desc) FIND_AND_ADD_LLVM_LIB(LLVMX86Info) FIND_AND_ADD_LLVM_LIB(LLVMWebAssemblyDisassembler) + FIND_AND_ADD_LLVM_LIB(LLVMWebAssemblyAsmParser) FIND_AND_ADD_LLVM_LIB(LLVMWebAssemblyCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMWebAssemblyDesc) - FIND_AND_ADD_LLVM_LIB(LLVMWebAssemblyAsmParser) FIND_AND_ADD_LLVM_LIB(LLVMWebAssemblyInfo) FIND_AND_ADD_LLVM_LIB(LLVMSystemZDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMSystemZCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMSystemZAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMSystemZCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMSystemZDesc) FIND_AND_ADD_LLVM_LIB(LLVMSystemZInfo) FIND_AND_ADD_LLVM_LIB(LLVMSparcDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMSparcCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMSparcAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMSparcCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMSparcDesc) FIND_AND_ADD_LLVM_LIB(LLVMSparcInfo) FIND_AND_ADD_LLVM_LIB(LLVMRISCVDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMRISCVCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMRISCVAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMRISCVCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMRISCVDesc) - FIND_AND_ADD_LLVM_LIB(LLVMRISCVUtils) FIND_AND_ADD_LLVM_LIB(LLVMRISCVInfo) FIND_AND_ADD_LLVM_LIB(LLVMPowerPCDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMPowerPCCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMPowerPCAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMPowerPCCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMPowerPCDesc) FIND_AND_ADD_LLVM_LIB(LLVMPowerPCInfo) FIND_AND_ADD_LLVM_LIB(LLVMNVPTXCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMNVPTXDesc) FIND_AND_ADD_LLVM_LIB(LLVMNVPTXInfo) FIND_AND_ADD_LLVM_LIB(LLVMMSP430Disassembler) - FIND_AND_ADD_LLVM_LIB(LLVMMSP430CodeGen) FIND_AND_ADD_LLVM_LIB(LLVMMSP430AsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMMSP430CodeGen) FIND_AND_ADD_LLVM_LIB(LLVMMSP430Desc) FIND_AND_ADD_LLVM_LIB(LLVMMSP430Info) FIND_AND_ADD_LLVM_LIB(LLVMMipsDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMMipsCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMMipsAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMMipsCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMMipsDesc) FIND_AND_ADD_LLVM_LIB(LLVMMipsInfo) FIND_AND_ADD_LLVM_LIB(LLVMLanaiDisassembler) @@ -279,44 +258,73 @@ else() FIND_AND_ADD_LLVM_LIB(LLVMHexagonDesc) FIND_AND_ADD_LLVM_LIB(LLVMHexagonInfo) FIND_AND_ADD_LLVM_LIB(LLVMBPFDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMBPFCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMBPFAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMBPFCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMBPFDesc) FIND_AND_ADD_LLVM_LIB(LLVMBPFInfo) FIND_AND_ADD_LLVM_LIB(LLVMAVRDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMAVRCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMAVRAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMAVRCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMAVRDesc) FIND_AND_ADD_LLVM_LIB(LLVMAVRInfo) FIND_AND_ADD_LLVM_LIB(LLVMARMDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMARMCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMARMAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMARMCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMARMDesc) FIND_AND_ADD_LLVM_LIB(LLVMARMUtils) FIND_AND_ADD_LLVM_LIB(LLVMARMInfo) FIND_AND_ADD_LLVM_LIB(LLVMAMDGPUDisassembler) - FIND_AND_ADD_LLVM_LIB(LLVMAMDGPUCodeGen) - FIND_AND_ADD_LLVM_LIB(LLVMMIRParser) - FIND_AND_ADD_LLVM_LIB(LLVMipo) - FIND_AND_ADD_LLVM_LIB(LLVMInstrumentation) - FIND_AND_ADD_LLVM_LIB(LLVMVectorize) - FIND_AND_ADD_LLVM_LIB(LLVMLinker) - FIND_AND_ADD_LLVM_LIB(LLVMIRReader) - FIND_AND_ADD_LLVM_LIB(LLVMAsmParser) - FIND_AND_ADD_LLVM_LIB(LLVMFrontendOpenMP) FIND_AND_ADD_LLVM_LIB(LLVMAMDGPUAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMAMDGPUCodeGen) FIND_AND_ADD_LLVM_LIB(LLVMAMDGPUDesc) FIND_AND_ADD_LLVM_LIB(LLVMAMDGPUUtils) FIND_AND_ADD_LLVM_LIB(LLVMAMDGPUInfo) FIND_AND_ADD_LLVM_LIB(LLVMAArch64Disassembler) - FIND_AND_ADD_LLVM_LIB(LLVMMCDisassembler) + FIND_AND_ADD_LLVM_LIB(LLVMAArch64AsmParser) FIND_AND_ADD_LLVM_LIB(LLVMAArch64CodeGen) + FIND_AND_ADD_LLVM_LIB(LLVMAArch64Desc) + FIND_AND_ADD_LLVM_LIB(LLVMAArch64Utils) + FIND_AND_ADD_LLVM_LIB(LLVMAArch64Info) + FIND_AND_ADD_LLVM_LIB(LLVMOrcJIT) + FIND_AND_ADD_LLVM_LIB(LLVMMCJIT) + FIND_AND_ADD_LLVM_LIB(LLVMJITLink) + FIND_AND_ADD_LLVM_LIB(LLVMOrcTargetProcess) + FIND_AND_ADD_LLVM_LIB(LLVMOrcShared) + FIND_AND_ADD_LLVM_LIB(LLVMInterpreter) + FIND_AND_ADD_LLVM_LIB(LLVMExecutionEngine) + FIND_AND_ADD_LLVM_LIB(LLVMRuntimeDyld) + FIND_AND_ADD_LLVM_LIB(LLVMSymbolize) + FIND_AND_ADD_LLVM_LIB(LLVMDebugInfoPDB) + FIND_AND_ADD_LLVM_LIB(LLVMDebugInfoGSYM) + FIND_AND_ADD_LLVM_LIB(LLVMOption) + FIND_AND_ADD_LLVM_LIB(LLVMObjectYAML) + FIND_AND_ADD_LLVM_LIB(LLVMMCA) + FIND_AND_ADD_LLVM_LIB(LLVMMCDisassembler) + FIND_AND_ADD_LLVM_LIB(LLVMLTO) + FIND_AND_ADD_LLVM_LIB(LLVMPasses) FIND_AND_ADD_LLVM_LIB(LLVMCFGuard) + FIND_AND_ADD_LLVM_LIB(LLVMCoroutines) + FIND_AND_ADD_LLVM_LIB(LLVMObjCARCOpts) + FIND_AND_ADD_LLVM_LIB(LLVMHelloNew) + FIND_AND_ADD_LLVM_LIB(LLVMipo) + FIND_AND_ADD_LLVM_LIB(LLVMVectorize) + FIND_AND_ADD_LLVM_LIB(LLVMLinker) + FIND_AND_ADD_LLVM_LIB(LLVMInstrumentation) + FIND_AND_ADD_LLVM_LIB(LLVMFrontendOpenMP) + FIND_AND_ADD_LLVM_LIB(LLVMFrontendOpenACC) + FIND_AND_ADD_LLVM_LIB(LLVMExtensions) + FIND_AND_ADD_LLVM_LIB(LLVMDWARFLinker) FIND_AND_ADD_LLVM_LIB(LLVMGlobalISel) - FIND_AND_ADD_LLVM_LIB(LLVMSelectionDAG) + FIND_AND_ADD_LLVM_LIB(LLVMMIRParser) FIND_AND_ADD_LLVM_LIB(LLVMAsmPrinter) FIND_AND_ADD_LLVM_LIB(LLVMDebugInfoDWARF) + FIND_AND_ADD_LLVM_LIB(LLVMSelectionDAG) FIND_AND_ADD_LLVM_LIB(LLVMCodeGen) + FIND_AND_ADD_LLVM_LIB(LLVMIRReader) + FIND_AND_ADD_LLVM_LIB(LLVMAsmParser) + FIND_AND_ADD_LLVM_LIB(LLVMInterfaceStub) + FIND_AND_ADD_LLVM_LIB(LLVMFileCheck) + FIND_AND_ADD_LLVM_LIB(LLVMFuzzMutate) FIND_AND_ADD_LLVM_LIB(LLVMTarget) FIND_AND_ADD_LLVM_LIB(LLVMScalarOpts) FIND_AND_ADD_LLVM_LIB(LLVMInstCombine) @@ -327,19 +335,15 @@ else() FIND_AND_ADD_LLVM_LIB(LLVMProfileData) FIND_AND_ADD_LLVM_LIB(LLVMObject) FIND_AND_ADD_LLVM_LIB(LLVMTextAPI) - FIND_AND_ADD_LLVM_LIB(LLVMBitReader) - FIND_AND_ADD_LLVM_LIB(LLVMCore) - FIND_AND_ADD_LLVM_LIB(LLVMRemarks) - FIND_AND_ADD_LLVM_LIB(LLVMBitstreamReader) - FIND_AND_ADD_LLVM_LIB(LLVMAArch64AsmParser) FIND_AND_ADD_LLVM_LIB(LLVMMCParser) - FIND_AND_ADD_LLVM_LIB(LLVMAArch64Desc) FIND_AND_ADD_LLVM_LIB(LLVMMC) FIND_AND_ADD_LLVM_LIB(LLVMDebugInfoCodeView) FIND_AND_ADD_LLVM_LIB(LLVMDebugInfoMSF) + FIND_AND_ADD_LLVM_LIB(LLVMBitReader) + FIND_AND_ADD_LLVM_LIB(LLVMCore) + FIND_AND_ADD_LLVM_LIB(LLVMRemarks) + FIND_AND_ADD_LLVM_LIB(LLVMBitstreamReader) FIND_AND_ADD_LLVM_LIB(LLVMBinaryFormat) - FIND_AND_ADD_LLVM_LIB(LLVMAArch64Utils) - FIND_AND_ADD_LLVM_LIB(LLVMAArch64Info) FIND_AND_ADD_LLVM_LIB(LLVMSupport) FIND_AND_ADD_LLVM_LIB(LLVMDemangle) endif() diff --git a/lib/include/__clang_cuda_builtin_vars.h b/lib/include/__clang_cuda_builtin_vars.h index 2ba1521f25..412e823a82 100644 --- a/lib/include/__clang_cuda_builtin_vars.h +++ b/lib/include/__clang_cuda_builtin_vars.h @@ -55,7 +55,9 @@ struct __cuda_builtin_threadIdx_t { __CUDA_DEVICE_BUILTIN(z,__nvvm_read_ptx_sreg_tid_z()); // threadIdx should be convertible to uint3 (in fact in nvcc, it *is* a // uint3). This function is defined after we pull in vector_types.h. + __attribute__((device)) operator dim3() const; __attribute__((device)) operator uint3() const; + private: __CUDA_DISALLOW_BUILTINVAR_ACCESS(__cuda_builtin_threadIdx_t); }; @@ -66,7 +68,9 @@ struct __cuda_builtin_blockIdx_t { __CUDA_DEVICE_BUILTIN(z,__nvvm_read_ptx_sreg_ctaid_z()); // blockIdx should be convertible to uint3 (in fact in nvcc, it *is* a // uint3). This function is defined after we pull in vector_types.h. + __attribute__((device)) operator dim3() const; __attribute__((device)) operator uint3() const; + private: __CUDA_DISALLOW_BUILTINVAR_ACCESS(__cuda_builtin_blockIdx_t); }; @@ -78,6 +82,8 @@ struct __cuda_builtin_blockDim_t { // blockDim should be convertible to dim3 (in fact in nvcc, it *is* a // dim3). This function is defined after we pull in vector_types.h. __attribute__((device)) operator dim3() const; + __attribute__((device)) operator uint3() const; + private: __CUDA_DISALLOW_BUILTINVAR_ACCESS(__cuda_builtin_blockDim_t); }; @@ -89,6 +95,8 @@ struct __cuda_builtin_gridDim_t { // gridDim should be convertible to dim3 (in fact in nvcc, it *is* a // dim3). This function is defined after we pull in vector_types.h. __attribute__((device)) operator dim3() const; + __attribute__((device)) operator uint3() const; + private: __CUDA_DISALLOW_BUILTINVAR_ACCESS(__cuda_builtin_gridDim_t); }; @@ -108,5 +116,6 @@ __attribute__((device)) const int warpSize = 32; #undef __CUDA_DEVICE_BUILTIN #undef __CUDA_BUILTIN_VAR #undef __CUDA_DISALLOW_BUILTINVAR_ACCESS +#undef __DELETE #endif /* __CUDA_BUILTIN_VARS_H */ diff --git a/lib/include/__clang_cuda_cmath.h b/lib/include/__clang_cuda_cmath.h index 8ba182689a..5bbb59a93b 100644 --- a/lib/include/__clang_cuda_cmath.h +++ b/lib/include/__clang_cuda_cmath.h @@ -66,10 +66,38 @@ __DEVICE__ float frexp(float __arg, int *__exp) { } // For inscrutable reasons, the CUDA headers define these functions for us on -// Windows. For OpenMP we omit these as some old system headers have -// non-conforming `isinf(float)` and `isnan(float)` implementations that return -// an `int`. The system versions of these functions should be fine anyway. -#if !defined(_MSC_VER) && !defined(__OPENMP_NVPTX__) +// Windows. +#if !defined(_MSC_VER) || defined(__OPENMP_NVPTX__) + +// For OpenMP we work around some old system headers that have non-conforming +// `isinf(float)` and `isnan(float)` implementations that return an `int`. We do +// this by providing two versions of these functions, differing only in the +// return type. To avoid conflicting definitions we disable implicit base +// function generation. That means we will end up with two specializations, one +// per type, but only one has a base function defined by the system header. +#if defined(__OPENMP_NVPTX__) +#pragma omp begin declare variant match( \ + implementation = {extension(disable_implicit_base)}) + +// FIXME: We lack an extension to customize the mangling of the variants, e.g., +// add a suffix. This means we would clash with the names of the variants +// (note that we do not create implicit base functions here). To avoid +// this clash we add a new trait to some of them that is always true +// (this is LLVM after all ;)). It will only influence the mangled name +// of the variants inside the inner region and avoid the clash. +#pragma omp begin declare variant match(implementation = {vendor(llvm)}) + +__DEVICE__ int isinf(float __x) { return ::__isinff(__x); } +__DEVICE__ int isinf(double __x) { return ::__isinf(__x); } +__DEVICE__ int isfinite(float __x) { return ::__finitef(__x); } +__DEVICE__ int isfinite(double __x) { return ::__isfinited(__x); } +__DEVICE__ int isnan(float __x) { return ::__isnanf(__x); } +__DEVICE__ int isnan(double __x) { return ::__isnan(__x); } + +#pragma omp end declare variant + +#endif + __DEVICE__ bool isinf(float __x) { return ::__isinff(__x); } __DEVICE__ bool isinf(double __x) { return ::__isinf(__x); } __DEVICE__ bool isfinite(float __x) { return ::__finitef(__x); } @@ -79,6 +107,11 @@ __DEVICE__ bool isfinite(float __x) { return ::__finitef(__x); } __DEVICE__ bool isfinite(double __x) { return ::__isfinited(__x); } __DEVICE__ bool isnan(float __x) { return ::__isnanf(__x); } __DEVICE__ bool isnan(double __x) { return ::__isnan(__x); } + +#if defined(__OPENMP_NVPTX__) +#pragma omp end declare variant +#endif + #endif __DEVICE__ bool isgreater(float __x, float __y) { @@ -142,6 +175,15 @@ __DEVICE__ float sqrt(float __x) { return ::sqrtf(__x); } __DEVICE__ float tan(float __x) { return ::tanf(__x); } __DEVICE__ float tanh(float __x) { return ::tanhf(__x); } +// There was a redefinition error for this this overload in CUDA mode. +// We restrict it to OpenMP mode for now, that is where it is actually needed +// anyway. +#ifdef __OPENMP_NVPTX__ +__DEVICE__ float remquo(float __n, float __d, int *__q) { + return ::remquof(__n, __d, __q); +} +#endif + // Notably missing above is nexttoward. We omit it because // libdevice doesn't provide an implementation, and we don't want to be in the // business of implementing tricky libm functions in this header. diff --git a/lib/include/__clang_cuda_complex_builtins.h b/lib/include/__clang_cuda_complex_builtins.h index d924487ab2..2b701fef0e 100644 --- a/lib/include/__clang_cuda_complex_builtins.h +++ b/lib/include/__clang_cuda_complex_builtins.h @@ -41,6 +41,27 @@ #define _ABSf std::abs #define _LOGBd std::logb #define _LOGBf std::logb +// Rather than pulling in std::max from algorithm everytime, use available ::max. +#define _fmaxd max +#define _fmaxf max +#else +#ifdef __AMDGCN__ +#define _ISNANd __ocml_isnan_f64 +#define _ISNANf __ocml_isnan_f32 +#define _ISINFd __ocml_isinf_f64 +#define _ISINFf __ocml_isinf_f32 +#define _ISFINITEd __ocml_isfinite_f64 +#define _ISFINITEf __ocml_isfinite_f32 +#define _COPYSIGNd __ocml_copysign_f64 +#define _COPYSIGNf __ocml_copysign_f32 +#define _SCALBNd __ocml_scalbn_f64 +#define _SCALBNf __ocml_scalbn_f32 +#define _ABSd __ocml_fabs_f64 +#define _ABSf __ocml_fabs_f32 +#define _LOGBd __ocml_logb_f64 +#define _LOGBf __ocml_logb_f32 +#define _fmaxd __ocml_fmax_f64 +#define _fmaxf __ocml_fmax_f32 #else #define _ISNANd __nv_isnand #define _ISNANf __nv_isnanf @@ -56,6 +77,9 @@ #define _ABSf __nv_fabsf #define _LOGBd __nv_logb #define _LOGBf __nv_logbf +#define _fmaxd __nv_fmax +#define _fmaxf __nv_fmaxf +#endif #endif #if defined(__cplusplus) @@ -167,7 +191,7 @@ __DEVICE__ double _Complex __divdc3(double __a, double __b, double __c, // Can't use std::max, because that's defined in <algorithm>, and we don't // want to pull that in for every compile. The CUDA headers define // ::max(float, float) and ::max(double, double), which is sufficient for us. - double __logbw = _LOGBd(max(_ABSd(__c), _ABSd(__d))); + double __logbw = _LOGBd(_fmaxd(_ABSd(__c), _ABSd(__d))); if (_ISFINITEd(__logbw)) { __ilogbw = (int)__logbw; __c = _SCALBNd(__c, -__ilogbw); @@ -200,7 +224,7 @@ __DEVICE__ double _Complex __divdc3(double __a, double __b, double __c, __DEVICE__ float _Complex __divsc3(float __a, float __b, float __c, float __d) { int __ilogbw = 0; - float __logbw = _LOGBf(max(_ABSf(__c), _ABSf(__d))); + float __logbw = _LOGBf(_fmaxf(_ABSf(__c), _ABSf(__d))); if (_ISFINITEf(__logbw)) { __ilogbw = (int)__logbw; __c = _SCALBNf(__c, -__ilogbw); @@ -249,6 +273,8 @@ __DEVICE__ float _Complex __divsc3(float __a, float __b, float __c, float __d) { #undef _ABSf #undef _LOGBd #undef _LOGBf +#undef _fmaxd +#undef _fmaxf #ifdef __OPENMP_NVPTX__ #pragma omp end declare target diff --git a/lib/include/__clang_cuda_math.h b/lib/include/__clang_cuda_math.h index 332e616702..acb26ad345 100644 --- a/lib/include/__clang_cuda_math.h +++ b/lib/include/__clang_cuda_math.h @@ -195,8 +195,8 @@ __DEVICE__ int max(int __a, int __b) { return __nv_max(__a, __b); } __DEVICE__ int min(int __a, int __b) { return __nv_min(__a, __b); } __DEVICE__ double modf(double __a, double *__b) { return __nv_modf(__a, __b); } __DEVICE__ float modff(float __a, float *__b) { return __nv_modff(__a, __b); } -__DEVICE__ double nearbyint(double __a) { return __nv_nearbyint(__a); } -__DEVICE__ float nearbyintf(float __a) { return __nv_nearbyintf(__a); } +__DEVICE__ double nearbyint(double __a) { return __builtin_nearbyint(__a); } +__DEVICE__ float nearbyintf(float __a) { return __builtin_nearbyintf(__a); } __DEVICE__ double nextafter(double __a, double __b) { return __nv_nextafter(__a, __b); } @@ -249,8 +249,9 @@ __DEVICE__ double rhypot(double __a, double __b) { __DEVICE__ float rhypotf(float __a, float __b) { return __nv_rhypotf(__a, __b); } -__DEVICE__ double rint(double __a) { return __nv_rint(__a); } -__DEVICE__ float rintf(float __a) { return __nv_rintf(__a); } +// __nv_rint* in libdevice is buggy and produces incorrect results. +__DEVICE__ double rint(double __a) { return __builtin_rint(__a); } +__DEVICE__ float rintf(float __a) { return __builtin_rintf(__a); } __DEVICE__ double rnorm(int __a, const double *__b) { return __nv_rnorm(__a, __b); } diff --git a/lib/include/__clang_cuda_math_forward_declares.h b/lib/include/__clang_cuda_math_forward_declares.h index 8a270859e4..c0f1f47cc9 100644 --- a/lib/include/__clang_cuda_math_forward_declares.h +++ b/lib/include/__clang_cuda_math_forward_declares.h @@ -160,6 +160,9 @@ __DEVICE__ double scalbln(double, long); __DEVICE__ float scalbln(float, long); __DEVICE__ double scalbn(double, int); __DEVICE__ float scalbn(float, int); +#ifdef _MSC_VER +__DEVICE__ bool signbit(long double); +#endif __DEVICE__ bool signbit(double); __DEVICE__ bool signbit(float); __DEVICE__ double sin(double); diff --git a/lib/include/__clang_cuda_runtime_wrapper.h b/lib/include/__clang_cuda_runtime_wrapper.h index f43ed55de4..f88c39a9b6 100644 --- a/lib/include/__clang_cuda_runtime_wrapper.h +++ b/lib/include/__clang_cuda_runtime_wrapper.h @@ -377,30 +377,38 @@ __device__ static inline void *malloc(size_t __size) { // Out-of-line implementations from __clang_cuda_builtin_vars.h. These need to // come after we've pulled in the definition of uint3 and dim3. +__device__ inline __cuda_builtin_threadIdx_t::operator dim3() const { + return dim3(x, y, z); +} + __device__ inline __cuda_builtin_threadIdx_t::operator uint3() const { - uint3 ret; - ret.x = x; - ret.y = y; - ret.z = z; - return ret; + return {x, y, z}; +} + +__device__ inline __cuda_builtin_blockIdx_t::operator dim3() const { + return dim3(x, y, z); } __device__ inline __cuda_builtin_blockIdx_t::operator uint3() const { - uint3 ret; - ret.x = x; - ret.y = y; - ret.z = z; - return ret; + return {x, y, z}; } __device__ inline __cuda_builtin_blockDim_t::operator dim3() const { return dim3(x, y, z); } +__device__ inline __cuda_builtin_blockDim_t::operator uint3() const { + return {x, y, z}; +} + __device__ inline __cuda_builtin_gridDim_t::operator dim3() const { return dim3(x, y, z); } +__device__ inline __cuda_builtin_gridDim_t::operator uint3() const { + return {x, y, z}; +} + #include <__clang_cuda_cmath.h> #include <__clang_cuda_intrinsics.h> #include <__clang_cuda_complex_builtins.h> diff --git a/lib/include/__clang_hip_cmath.h b/lib/include/__clang_hip_cmath.h new file mode 100644 index 0000000000..cd22a2df95 --- /dev/null +++ b/lib/include/__clang_hip_cmath.h @@ -0,0 +1,664 @@ +/*===---- __clang_hip_cmath.h - HIP cmath decls -----------------------------=== + * + * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. + * See https://llvm.org/LICENSE.txt for license information. + * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception + * + *===-----------------------------------------------------------------------=== + */ + +#ifndef __CLANG_HIP_CMATH_H__ +#define __CLANG_HIP_CMATH_H__ + +#if !defined(__HIP__) +#error "This file is for HIP and OpenMP AMDGCN device compilation only." +#endif + +#if defined(__cplusplus) +#include <limits> +#include <type_traits> +#include <utility> +#endif +#include <limits.h> +#include <stdint.h> + +#pragma push_macro("__DEVICE__") +#define __DEVICE__ static __device__ inline __attribute__((always_inline)) + +// Start with functions that cannot be defined by DEF macros below. +#if defined(__cplusplus) +__DEVICE__ double abs(double __x) { return ::fabs(__x); } +__DEVICE__ float abs(float __x) { return ::fabsf(__x); } +__DEVICE__ long long abs(long long __n) { return ::llabs(__n); } +__DEVICE__ long abs(long __n) { return ::labs(__n); } +__DEVICE__ float fma(float __x, float __y, float __z) { + return ::fmaf(__x, __y, __z); +} +__DEVICE__ int fpclassify(float __x) { + return __builtin_fpclassify(FP_NAN, FP_INFINITE, FP_NORMAL, FP_SUBNORMAL, + FP_ZERO, __x); +} +__DEVICE__ int fpclassify(double __x) { + return __builtin_fpclassify(FP_NAN, FP_INFINITE, FP_NORMAL, FP_SUBNORMAL, + FP_ZERO, __x); +} +__DEVICE__ float frexp(float __arg, int *__exp) { + return ::frexpf(__arg, __exp); +} +__DEVICE__ bool isfinite(float __x) { return ::__finitef(__x); } +__DEVICE__ bool isfinite(double __x) { return ::__finite(__x); } +__DEVICE__ bool isgreater(float __x, float __y) { + return __builtin_isgreater(__x, __y); +} +__DEVICE__ bool isgreater(double __x, double __y) { + return __builtin_isgreater(__x, __y); +} +__DEVICE__ bool isgreaterequal(float __x, float __y) { + return __builtin_isgreaterequal(__x, __y); +} +__DEVICE__ bool isgreaterequal(double __x, double __y) { + return __builtin_isgreaterequal(__x, __y); +} +__DEVICE__ bool isinf(float __x) { return ::__isinff(__x); } +__DEVICE__ bool isinf(double __x) { return ::__isinf(__x); } +__DEVICE__ bool isless(float __x, float __y) { + return __builtin_isless(__x, __y); +} +__DEVICE__ bool isless(double __x, double __y) { + return __builtin_isless(__x, __y); +} +__DEVICE__ bool islessequal(float __x, float __y) { + return __builtin_islessequal(__x, __y); +} +__DEVICE__ bool islessequal(double __x, double __y) { + return __builtin_islessequal(__x, __y); +} +__DEVICE__ bool islessgreater(float __x, float __y) { + return __builtin_islessgreater(__x, __y); +} +__DEVICE__ bool islessgreater(double __x, double __y) { + return __builtin_islessgreater(__x, __y); +} +__DEVICE__ bool isnan(float __x) { return ::__isnanf(__x); } +__DEVICE__ bool isnan(double __x) { return ::__isnan(__x); } +__DEVICE__ bool isnormal(float __x) { return __builtin_isnormal(__x); } +__DEVICE__ bool isnormal(double __x) { return __builtin_isnormal(__x); } +__DEVICE__ bool isunordered(float __x, float __y) { + return __builtin_isunordered(__x, __y); +} +__DEVICE__ bool isunordered(double __x, double __y) { + return __builtin_isunordered(__x, __y); +} +__DEVICE__ float modf(float __x, float *__iptr) { return ::modff(__x, __iptr); } +__DEVICE__ float pow(float __base, int __iexp) { + return ::powif(__base, __iexp); +} +__DEVICE__ double pow(double __base, int __iexp) { + return ::powi(__base, __iexp); +} +__DEVICE__ float remquo(float __x, float __y, int *__quo) { + return ::remquof(__x, __y, __quo); +} +__DEVICE__ float scalbln(float __x, long int __n) { + return ::scalblnf(__x, __n); +} +__DEVICE__ bool signbit(float __x) { return ::__signbitf(__x); } +__DEVICE__ bool signbit(double __x) { return ::__signbit(__x); } + +// Notably missing above is nexttoward. We omit it because +// ocml doesn't provide an implementation, and we don't want to be in the +// business of implementing tricky libm functions in this header. + +// Other functions. +__DEVICE__ _Float16 fma(_Float16 __x, _Float16 __y, _Float16 __z) { + return __ocml_fma_f16(__x, __y, __z); +} +__DEVICE__ _Float16 pow(_Float16 __base, int __iexp) { + return __ocml_pown_f16(__base, __iexp); +} + +// BEGIN DEF_FUN and HIP_OVERLOAD + +// BEGIN DEF_FUN + +#pragma push_macro("__DEF_FUN1") +#pragma push_macro("__DEF_FUN2") +#pragma push_macro("__DEF_FUN2_FI") + +// Define cmath functions with float argument and returns __retty. +#define __DEF_FUN1(__retty, __func) \ + __DEVICE__ \ + __retty __func(float __x) { return __func##f(__x); } + +// Define cmath functions with two float arguments and returns __retty. +#define __DEF_FUN2(__retty, __func) \ + __DEVICE__ \ + __retty __func(float __x, float __y) { return __func##f(__x, __y); } + +// Define cmath functions with a float and an int argument and returns __retty. +#define __DEF_FUN2_FI(__retty, __func) \ + __DEVICE__ \ + __retty __func(float __x, int __y) { return __func##f(__x, __y); } + +__DEF_FUN1(float, acos) +__DEF_FUN1(float, acosh) +__DEF_FUN1(float, asin) +__DEF_FUN1(float, asinh) +__DEF_FUN1(float, atan) +__DEF_FUN2(float, atan2) +__DEF_FUN1(float, atanh) +__DEF_FUN1(float, cbrt) +__DEF_FUN1(float, ceil) +__DEF_FUN2(float, copysign) +__DEF_FUN1(float, cos) +__DEF_FUN1(float, cosh) +__DEF_FUN1(float, erf) +__DEF_FUN1(float, erfc) +__DEF_FUN1(float, exp) +__DEF_FUN1(float, exp2) +__DEF_FUN1(float, expm1) +__DEF_FUN1(float, fabs) +__DEF_FUN2(float, fdim) +__DEF_FUN1(float, floor) +__DEF_FUN2(float, fmax) +__DEF_FUN2(float, fmin) +__DEF_FUN2(float, fmod) +__DEF_FUN2(float, hypot) +__DEF_FUN1(int, ilogb) +__DEF_FUN2_FI(float, ldexp) +__DEF_FUN1(float, lgamma) +__DEF_FUN1(float, log) +__DEF_FUN1(float, log10) +__DEF_FUN1(float, log1p) +__DEF_FUN1(float, log2) +__DEF_FUN1(float, logb) +__DEF_FUN1(long long, llrint) +__DEF_FUN1(long long, llround) +__DEF_FUN1(long, lrint) +__DEF_FUN1(long, lround) +__DEF_FUN1(float, nearbyint) +__DEF_FUN2(float, nextafter) +__DEF_FUN2(float, pow) +__DEF_FUN2(float, remainder) +__DEF_FUN1(float, rint) +__DEF_FUN1(float, round) +__DEF_FUN2_FI(float, scalbn) +__DEF_FUN1(float, sin) +__DEF_FUN1(float, sinh) +__DEF_FUN1(float, sqrt) +__DEF_FUN1(float, tan) +__DEF_FUN1(float, tanh) +__DEF_FUN1(float, tgamma) +__DEF_FUN1(float, trunc) + +#pragma pop_macro("__DEF_FUN1") +#pragma pop_macro("__DEF_FUN2") +#pragma pop_macro("__DEF_FUN2_FI") + +// END DEF_FUN + +// BEGIN HIP_OVERLOAD + +#pragma push_macro("__HIP_OVERLOAD1") +#pragma push_macro("__HIP_OVERLOAD2") + +// __hip_enable_if::type is a type function which returns __T if __B is true. +template <bool __B, class __T = void> struct __hip_enable_if {}; + +template <class __T> struct __hip_enable_if<true, __T> { typedef __T type; }; + +// decltype is only available in C++11 and above. +#if __cplusplus >= 201103L +// __hip_promote +namespace __hip { + +template <class _Tp> struct __numeric_type { + static void __test(...); + static _Float16 __test(_Float16); + static float __test(float); + static double __test(char); + static double __test(int); + static double __test(unsigned); + static double __test(long); + static double __test(unsigned long); + static double __test(long long); + static double __test(unsigned long long); + static double __test(double); + // No support for long double, use double instead. + static double __test(long double); + + typedef decltype(__test(std::declval<_Tp>())) type; + static const bool value = !std::is_same<type, void>::value; +}; + +template <> struct __numeric_type<void> { static const bool value = true; }; + +template <class _A1, class _A2 = void, class _A3 = void, + bool = __numeric_type<_A1>::value &&__numeric_type<_A2>::value + &&__numeric_type<_A3>::value> +class __promote_imp { +public: + static const bool value = false; +}; + +template <class _A1, class _A2, class _A3> +class __promote_imp<_A1, _A2, _A3, true> { +private: + typedef typename __promote_imp<_A1>::type __type1; + typedef typename __promote_imp<_A2>::type __type2; + typedef typename __promote_imp<_A3>::type __type3; + +public: + typedef decltype(__type1() + __type2() + __type3()) type; + static const bool value = true; +}; + +template <class _A1, class _A2> class __promote_imp<_A1, _A2, void, true> { +private: + typedef typename __promote_imp<_A1>::type __type1; + typedef typename __promote_imp<_A2>::type __type2; + +public: + typedef decltype(__type1() + __type2()) type; + static const bool value = true; +}; + +template <class _A1> class __promote_imp<_A1, void, void, true> { +public: + typedef typename __numeric_type<_A1>::type type; + static const bool value = true; +}; + +template <class _A1, class _A2 = void, class _A3 = void> +class __promote : public __promote_imp<_A1, _A2, _A3> {}; + +} // namespace __hip +#endif //__cplusplus >= 201103L + +// __HIP_OVERLOAD1 is used to resolve function calls with integer argument to +// avoid compilation error due to ambibuity. e.g. floor(5) is resolved with +// floor(double). +#define __HIP_OVERLOAD1(__retty, __fn) \ + template <typename __T> \ + __DEVICE__ typename __hip_enable_if<std::numeric_limits<__T>::is_integer, \ + __retty>::type \ + __fn(__T __x) { \ + return ::__fn((double)__x); \ + } + +// __HIP_OVERLOAD2 is used to resolve function calls with mixed float/double +// or integer argument to avoid compilation error due to ambibuity. e.g. +// max(5.0f, 6.0) is resolved with max(double, double). +#if __cplusplus >= 201103L +#define __HIP_OVERLOAD2(__retty, __fn) \ + template <typename __T1, typename __T2> \ + __DEVICE__ typename __hip_enable_if< \ + std::numeric_limits<__T1>::is_specialized && \ + std::numeric_limits<__T2>::is_specialized, \ + typename __hip::__promote<__T1, __T2>::type>::type \ + __fn(__T1 __x, __T2 __y) { \ + typedef typename __hip::__promote<__T1, __T2>::type __result_type; \ + return __fn((__result_type)__x, (__result_type)__y); \ + } +#else +#define __HIP_OVERLOAD2(__retty, __fn) \ + template <typename __T1, typename __T2> \ + __DEVICE__ \ + typename __hip_enable_if<std::numeric_limits<__T1>::is_specialized && \ + std::numeric_limits<__T2>::is_specialized, \ + __retty>::type \ + __fn(__T1 __x, __T2 __y) { \ + return __fn((double)__x, (double)__y); \ + } +#endif + +__HIP_OVERLOAD1(double, abs) +__HIP_OVERLOAD1(double, acos) +__HIP_OVERLOAD1(double, acosh) +__HIP_OVERLOAD1(double, asin) +__HIP_OVERLOAD1(double, asinh) +__HIP_OVERLOAD1(double, atan) +__HIP_OVERLOAD2(double, atan2) +__HIP_OVERLOAD1(double, atanh) +__HIP_OVERLOAD1(double, cbrt) +__HIP_OVERLOAD1(double, ceil) +__HIP_OVERLOAD2(double, copysign) +__HIP_OVERLOAD1(double, cos) +__HIP_OVERLOAD1(double, cosh) +__HIP_OVERLOAD1(double, erf) +__HIP_OVERLOAD1(double, erfc) +__HIP_OVERLOAD1(double, exp) +__HIP_OVERLOAD1(double, exp2) +__HIP_OVERLOAD1(double, expm1) +__HIP_OVERLOAD1(double, fabs) +__HIP_OVERLOAD2(double, fdim) +__HIP_OVERLOAD1(double, floor) +__HIP_OVERLOAD2(double, fmax) +__HIP_OVERLOAD2(double, fmin) +__HIP_OVERLOAD2(double, fmod) +__HIP_OVERLOAD1(int, fpclassify) +__HIP_OVERLOAD2(double, hypot) +__HIP_OVERLOAD1(int, ilogb) +__HIP_OVERLOAD1(bool, isfinite) +__HIP_OVERLOAD2(bool, isgreater) +__HIP_OVERLOAD2(bool, isgreaterequal) +__HIP_OVERLOAD1(bool, isinf) +__HIP_OVERLOAD2(bool, isless) +__HIP_OVERLOAD2(bool, islessequal) +__HIP_OVERLOAD2(bool, islessgreater) +__HIP_OVERLOAD1(bool, isnan) +__HIP_OVERLOAD1(bool, isnormal) +__HIP_OVERLOAD2(bool, isunordered) +__HIP_OVERLOAD1(double, lgamma) +__HIP_OVERLOAD1(double, log) +__HIP_OVERLOAD1(double, log10) +__HIP_OVERLOAD1(double, log1p) +__HIP_OVERLOAD1(double, log2) +__HIP_OVERLOAD1(double, logb) +__HIP_OVERLOAD1(long long, llrint) +__HIP_OVERLOAD1(long long, llround) +__HIP_OVERLOAD1(long, lrint) +__HIP_OVERLOAD1(long, lround) +__HIP_OVERLOAD1(double, nearbyint) +__HIP_OVERLOAD2(double, nextafter) +__HIP_OVERLOAD2(double, pow) +__HIP_OVERLOAD2(double, remainder) +__HIP_OVERLOAD1(double, rint) +__HIP_OVERLOAD1(double, round) +__HIP_OVERLOAD1(bool, signbit) +__HIP_OVERLOAD1(double, sin) +__HIP_OVERLOAD1(double, sinh) +__HIP_OVERLOAD1(double, sqrt) +__HIP_OVERLOAD1(double, tan) +__HIP_OVERLOAD1(double, tanh) +__HIP_OVERLOAD1(double, tgamma) +__HIP_OVERLOAD1(double, trunc) + +// Overload these but don't add them to std, they are not part of cmath. +__HIP_OVERLOAD2(double, max) +__HIP_OVERLOAD2(double, min) + +// Additional Overloads that don't quite match HIP_OVERLOAD. +#if __cplusplus >= 201103L +template <typename __T1, typename __T2, typename __T3> +__DEVICE__ typename __hip_enable_if< + std::numeric_limits<__T1>::is_specialized && + std::numeric_limits<__T2>::is_specialized && + std::numeric_limits<__T3>::is_specialized, + typename __hip::__promote<__T1, __T2, __T3>::type>::type +fma(__T1 __x, __T2 __y, __T3 __z) { + typedef typename __hip::__promote<__T1, __T2, __T3>::type __result_type; + return ::fma((__result_type)__x, (__result_type)__y, (__result_type)__z); +} +#else +template <typename __T1, typename __T2, typename __T3> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T1>::is_specialized && + std::numeric_limits<__T2>::is_specialized && + std::numeric_limits<__T3>::is_specialized, + double>::type + fma(__T1 __x, __T2 __y, __T3 __z) { + return ::fma((double)__x, (double)__y, (double)__z); +} +#endif + +template <typename __T> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T>::is_integer, double>::type + frexp(__T __x, int *__exp) { + return ::frexp((double)__x, __exp); +} + +template <typename __T> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T>::is_integer, double>::type + ldexp(__T __x, int __exp) { + return ::ldexp((double)__x, __exp); +} + +template <typename __T> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T>::is_integer, double>::type + modf(__T __x, double *__exp) { + return ::modf((double)__x, __exp); +} + +#if __cplusplus >= 201103L +template <typename __T1, typename __T2> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T1>::is_specialized && + std::numeric_limits<__T2>::is_specialized, + typename __hip::__promote<__T1, __T2>::type>::type + remquo(__T1 __x, __T2 __y, int *__quo) { + typedef typename __hip::__promote<__T1, __T2>::type __result_type; + return ::remquo((__result_type)__x, (__result_type)__y, __quo); +} +#else +template <typename __T1, typename __T2> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T1>::is_specialized && + std::numeric_limits<__T2>::is_specialized, + double>::type + remquo(__T1 __x, __T2 __y, int *__quo) { + return ::remquo((double)__x, (double)__y, __quo); +} +#endif + +template <typename __T> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T>::is_integer, double>::type + scalbln(__T __x, long int __exp) { + return ::scalbln((double)__x, __exp); +} + +template <typename __T> +__DEVICE__ + typename __hip_enable_if<std::numeric_limits<__T>::is_integer, double>::type + scalbn(__T __x, int __exp) { + return ::scalbn((double)__x, __exp); +} + +#pragma pop_macro("__HIP_OVERLOAD1") +#pragma pop_macro("__HIP_OVERLOAD2") + +// END HIP_OVERLOAD + +// END DEF_FUN and HIP_OVERLOAD + +#endif // defined(__cplusplus) + +// Define these overloads inside the namespace our standard library uses. +#ifdef _LIBCPP_BEGIN_NAMESPACE_STD +_LIBCPP_BEGIN_NAMESPACE_STD +#else +namespace std { +#ifdef _GLIBCXX_BEGIN_NAMESPACE_VERSION +_GLIBCXX_BEGIN_NAMESPACE_VERSION +#endif +#endif + +// Pull the new overloads we defined above into namespace std. +// using ::abs; - This may be considered for C++. +using ::acos; +using ::acosh; +using ::asin; +using ::asinh; +using ::atan; +using ::atan2; +using ::atanh; +using ::cbrt; +using ::ceil; +using ::copysign; +using ::cos; +using ::cosh; +using ::erf; +using ::erfc; +using ::exp; +using ::exp2; +using ::expm1; +using ::fabs; +using ::fdim; +using ::floor; +using ::fma; +using ::fmax; +using ::fmin; +using ::fmod; +using ::fpclassify; +using ::frexp; +using ::hypot; +using ::ilogb; +using ::isfinite; +using ::isgreater; +using ::isgreaterequal; +using ::isless; +using ::islessequal; +using ::islessgreater; +using ::isnormal; +using ::isunordered; +using ::ldexp; +using ::lgamma; +using ::llrint; +using ::llround; +using ::log; +using ::log10; +using ::log1p; +using ::log2; +using ::logb; +using ::lrint; +using ::lround; +using ::modf; +// using ::nan; - This may be considered for C++. +// using ::nanf; - This may be considered for C++. +// using ::nanl; - This is not yet defined. +using ::nearbyint; +using ::nextafter; +// using ::nexttoward; - Omit this since we do not have a definition. +using ::pow; +using ::remainder; +using ::remquo; +using ::rint; +using ::round; +using ::scalbln; +using ::scalbn; +using ::signbit; +using ::sin; +using ::sinh; +using ::sqrt; +using ::tan; +using ::tanh; +using ::tgamma; +using ::trunc; + +// Well this is fun: We need to pull these symbols in for libc++, but we can't +// pull them in with libstdc++, because its ::isinf and ::isnan are different +// than its std::isinf and std::isnan. +#ifndef __GLIBCXX__ +using ::isinf; +using ::isnan; +#endif + +// Finally, pull the "foobarf" functions that HIP defines into std. +using ::acosf; +using ::acoshf; +using ::asinf; +using ::asinhf; +using ::atan2f; +using ::atanf; +using ::atanhf; +using ::cbrtf; +using ::ceilf; +using ::copysignf; +using ::cosf; +using ::coshf; +using ::erfcf; +using ::erff; +using ::exp2f; +using ::expf; +using ::expm1f; +using ::fabsf; +using ::fdimf; +using ::floorf; +using ::fmaf; +using ::fmaxf; +using ::fminf; +using ::fmodf; +using ::frexpf; +using ::hypotf; +using ::ilogbf; +using ::ldexpf; +using ::lgammaf; +using ::llrintf; +using ::llroundf; +using ::log10f; +using ::log1pf; +using ::log2f; +using ::logbf; +using ::logf; +using ::lrintf; +using ::lroundf; +using ::modff; +using ::nearbyintf; +using ::nextafterf; +// using ::nexttowardf; - Omit this since we do not have a definition. +using ::powf; +using ::remainderf; +using ::remquof; +using ::rintf; +using ::roundf; +using ::scalblnf; +using ::scalbnf; +using ::sinf; +using ::sinhf; +using ::sqrtf; +using ::tanf; +using ::tanhf; +using ::tgammaf; +using ::truncf; + +#ifdef _LIBCPP_END_NAMESPACE_STD +_LIBCPP_END_NAMESPACE_STD +#else +#ifdef _GLIBCXX_BEGIN_NAMESPACE_VERSION +_GLIBCXX_END_NAMESPACE_VERSION +#endif +} // namespace std +#endif + +// Define device-side math functions from <ymath.h> on MSVC. +#if defined(_MSC_VER) + +// Before VS2019, `<ymath.h>` is also included in `<limits>` and other headers. +// But, from VS2019, it's only included in `<complex>`. Need to include +// `<ymath.h>` here to ensure C functions declared there won't be markded as +// `__host__` and `__device__` through `<complex>` wrapper. +#include <ymath.h> + +#if defined(__cplusplus) +extern "C" { +#endif // defined(__cplusplus) +__DEVICE__ __attribute__((overloadable)) double _Cosh(double x, double y) { + return cosh(x) * y; +} +__DEVICE__ __attribute__((overloadable)) float _FCosh(float x, float y) { + return coshf(x) * y; +} +__DEVICE__ __attribute__((overloadable)) short _Dtest(double *p) { + return fpclassify(*p); +} +__DEVICE__ __attribute__((overloadable)) short _FDtest(float *p) { + return fpclassify(*p); +} +__DEVICE__ __attribute__((overloadable)) double _Sinh(double x, double y) { + return sinh(x) * y; +} +__DEVICE__ __attribute__((overloadable)) float _FSinh(float x, float y) { + return sinhf(x) * y; +} +#if defined(__cplusplus) +} +#endif // defined(__cplusplus) +#endif // defined(_MSC_VER) + +#pragma pop_macro("__DEVICE__") + +#endif // __CLANG_HIP_CMATH_H__ diff --git a/lib/include/__clang_hip_libdevice_declares.h b/lib/include/__clang_hip_libdevice_declares.h index e1cd49a39c..ac98907ad5 100644 --- a/lib/include/__clang_hip_libdevice_declares.h +++ b/lib/include/__clang_hip_libdevice_declares.h @@ -10,7 +10,9 @@ #ifndef __CLANG_HIP_LIBDEVICE_DECLARES_H__ #define __CLANG_HIP_LIBDEVICE_DECLARES_H__ +#ifdef __cplusplus extern "C" { +#endif // BEGIN FLOAT __device__ __attribute__((const)) float __ocml_acos_f32(float); @@ -78,6 +80,7 @@ __device__ __attribute__((const)) float __ocml_len4_f32(float, float, float, __device__ __attribute__((pure)) float __ocml_ncdf_f32(float); __device__ __attribute__((pure)) float __ocml_ncdfinv_f32(float); __device__ __attribute__((pure)) float __ocml_pow_f32(float, float); +__device__ __attribute__((pure)) float __ocml_pown_f32(float, int); __device__ __attribute__((pure)) float __ocml_rcbrt_f32(float); __device__ __attribute__((const)) float __ocml_remainder_f32(float, float); __device__ float __ocml_remquo_f32(float, float, @@ -126,10 +129,10 @@ __device__ __attribute__((const)) float __ocml_div_rte_f32(float, float); __device__ __attribute__((const)) float __ocml_div_rtn_f32(float, float); __device__ __attribute__((const)) float __ocml_div_rtp_f32(float, float); __device__ __attribute__((const)) float __ocml_div_rtz_f32(float, float); -__device__ __attribute__((const)) float __ocml_sqrt_rte_f32(float, float); -__device__ __attribute__((const)) float __ocml_sqrt_rtn_f32(float, float); -__device__ __attribute__((const)) float __ocml_sqrt_rtp_f32(float, float); -__device__ __attribute__((const)) float __ocml_sqrt_rtz_f32(float, float); +__device__ __attribute__((const)) float __ocml_sqrt_rte_f32(float); +__device__ __attribute__((const)) float __ocml_sqrt_rtn_f32(float); +__device__ __attribute__((const)) float __ocml_sqrt_rtp_f32(float); +__device__ __attribute__((const)) float __ocml_sqrt_rtz_f32(float); __device__ __attribute__((const)) float __ocml_fma_rte_f32(float, float, float); __device__ __attribute__((const)) float __ocml_fma_rtn_f32(float, float, float); __device__ __attribute__((const)) float __ocml_fma_rtp_f32(float, float, float); @@ -205,6 +208,7 @@ __device__ __attribute__((const)) double __ocml_len4_f64(double, double, double, __device__ __attribute__((pure)) double __ocml_ncdf_f64(double); __device__ __attribute__((pure)) double __ocml_ncdfinv_f64(double); __device__ __attribute__((pure)) double __ocml_pow_f64(double, double); +__device__ __attribute__((pure)) double __ocml_pown_f64(double, int); __device__ __attribute__((pure)) double __ocml_rcbrt_f64(double); __device__ __attribute__((const)) double __ocml_remainder_f64(double, double); __device__ double __ocml_remquo_f64(double, double, @@ -252,10 +256,10 @@ __device__ __attribute__((const)) double __ocml_div_rte_f64(double, double); __device__ __attribute__((const)) double __ocml_div_rtn_f64(double, double); __device__ __attribute__((const)) double __ocml_div_rtp_f64(double, double); __device__ __attribute__((const)) double __ocml_div_rtz_f64(double, double); -__device__ __attribute__((const)) double __ocml_sqrt_rte_f64(double, double); -__device__ __attribute__((const)) double __ocml_sqrt_rtn_f64(double, double); -__device__ __attribute__((const)) double __ocml_sqrt_rtp_f64(double, double); -__device__ __attribute__((const)) double __ocml_sqrt_rtz_f64(double, double); +__device__ __attribute__((const)) double __ocml_sqrt_rte_f64(double); +__device__ __attribute__((const)) double __ocml_sqrt_rtn_f64(double); +__device__ __attribute__((const)) double __ocml_sqrt_rtp_f64(double); +__device__ __attribute__((const)) double __ocml_sqrt_rtz_f64(double); __device__ __attribute__((const)) double __ocml_fma_rte_f64(double, double, double); __device__ __attribute__((const)) double __ocml_fma_rtn_f64(double, double, @@ -290,6 +294,7 @@ __device__ __attribute__((const)) _Float16 __ocml_rsqrt_f16(_Float16); __device__ _Float16 __ocml_sin_f16(_Float16); __device__ __attribute__((const)) _Float16 __ocml_sqrt_f16(_Float16); __device__ __attribute__((const)) _Float16 __ocml_trunc_f16(_Float16); +__device__ __attribute__((pure)) _Float16 __ocml_pown_f16(_Float16, int); typedef _Float16 __2f16 __attribute__((ext_vector_type(2))); typedef short __2i16 __attribute__((ext_vector_type(2))); @@ -313,14 +318,17 @@ __device__ __attribute__((pure)) __2f16 __ocml_log2_2f16(__2f16); __device__ inline __2f16 __llvm_amdgcn_rcp_2f16(__2f16 __x) // Not currently exposed by ROCDL. { - return __2f16{__llvm_amdgcn_rcp_f16(__x.x), __llvm_amdgcn_rcp_f16(__x.y)}; + return (__2f16)(__llvm_amdgcn_rcp_f16(__x.x), __llvm_amdgcn_rcp_f16(__x.y)); } __device__ __attribute__((const)) __2f16 __ocml_rint_2f16(__2f16); __device__ __attribute__((const)) __2f16 __ocml_rsqrt_2f16(__2f16); __device__ __2f16 __ocml_sin_2f16(__2f16); __device__ __attribute__((const)) __2f16 __ocml_sqrt_2f16(__2f16); __device__ __attribute__((const)) __2f16 __ocml_trunc_2f16(__2f16); +__device__ __attribute__((const)) __2f16 __ocml_pown_2f16(__2f16, __2i16); +#ifdef __cplusplus } // extern "C" +#endif #endif // __CLANG_HIP_LIBDEVICE_DECLARES_H__ diff --git a/lib/include/__clang_hip_math.h b/lib/include/__clang_hip_math.h index cf7014b9ae..14d91c66b3 100644 --- a/lib/include/__clang_hip_math.h +++ b/lib/include/__clang_hip_math.h @@ -1,4 +1,4 @@ -/*===---- __clang_hip_math.h - HIP math decls -------------------------------=== +/*===---- __clang_hip_math.h - Device-side HIP math support ----------------=== * * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. * See https://llvm.org/LICENSE.txt for license information. @@ -6,24 +6,57 @@ * *===-----------------------------------------------------------------------=== */ - #ifndef __CLANG_HIP_MATH_H__ #define __CLANG_HIP_MATH_H__ +#if !defined(__HIP__) +#error "This file is for HIP and OpenMP AMDGCN device compilation only." +#endif + +#if defined(__cplusplus) #include <algorithm> +#endif #include <limits.h> -#include <limits> #include <stdint.h> #pragma push_macro("__DEVICE__") -#pragma push_macro("__RETURN_TYPE") +#define __DEVICE__ static __device__ inline __attribute__((always_inline)) -// to be consistent with __clang_cuda_math_forward_declares -#define __DEVICE__ static __device__ +// A few functions return bool type starting only in C++11. +#pragma push_macro("__RETURN_TYPE") +#if defined(__cplusplus) #define __RETURN_TYPE bool +#else +#define __RETURN_TYPE int +#endif + +#if defined (__cplusplus) && __cplusplus < 201103L +// emulate static_assert on type sizes +template<bool> +struct __compare_result{}; +template<> +struct __compare_result<true> { + static const bool valid; +}; + +__DEVICE__ +void __suppress_unused_warning(bool b){}; +template <unsigned int S, unsigned int T> +__DEVICE__ void __static_assert_equal_size() { + __suppress_unused_warning(__compare_result<S == T>::valid); +} + +#define __static_assert_type_size_equal(A, B) \ + __static_assert_equal_size<A,B>() + +#else +#define __static_assert_type_size_equal(A,B) \ + static_assert((A) == (B), "") + +#endif __DEVICE__ -inline uint64_t __make_mantissa_base8(const char *__tagp) { +uint64_t __make_mantissa_base8(const char *__tagp) { uint64_t __r = 0; while (__tagp) { char __tmp = *__tagp; @@ -40,7 +73,7 @@ inline uint64_t __make_mantissa_base8(const char *__tagp) { } __DEVICE__ -inline uint64_t __make_mantissa_base10(const char *__tagp) { +uint64_t __make_mantissa_base10(const char *__tagp) { uint64_t __r = 0; while (__tagp) { char __tmp = *__tagp; @@ -57,7 +90,7 @@ inline uint64_t __make_mantissa_base10(const char *__tagp) { } __DEVICE__ -inline uint64_t __make_mantissa_base16(const char *__tagp) { +uint64_t __make_mantissa_base16(const char *__tagp) { uint64_t __r = 0; while (__tagp) { char __tmp = *__tagp; @@ -78,7 +111,7 @@ inline uint64_t __make_mantissa_base16(const char *__tagp) { } __DEVICE__ -inline uint64_t __make_mantissa(const char *__tagp) { +uint64_t __make_mantissa(const char *__tagp) { if (!__tagp) return 0u; @@ -95,78 +128,124 @@ inline uint64_t __make_mantissa(const char *__tagp) { } // BEGIN FLOAT +#if defined(__cplusplus) __DEVICE__ -inline float abs(float __x) { return __ocml_fabs_f32(__x); } +int abs(int __x) { + int __sgn = __x >> (sizeof(int) * CHAR_BIT - 1); + return (__x ^ __sgn) - __sgn; +} __DEVICE__ -inline float acosf(float __x) { return __ocml_acos_f32(__x); } +long labs(long __x) { + long __sgn = __x >> (sizeof(long) * CHAR_BIT - 1); + return (__x ^ __sgn) - __sgn; +} __DEVICE__ -inline float acoshf(float __x) { return __ocml_acosh_f32(__x); } +long long llabs(long long __x) { + long long __sgn = __x >> (sizeof(long long) * CHAR_BIT - 1); + return (__x ^ __sgn) - __sgn; +} +#endif + __DEVICE__ -inline float asinf(float __x) { return __ocml_asin_f32(__x); } +float acosf(float __x) { return __ocml_acos_f32(__x); } + __DEVICE__ -inline float asinhf(float __x) { return __ocml_asinh_f32(__x); } +float acoshf(float __x) { return __ocml_acosh_f32(__x); } + __DEVICE__ -inline float atan2f(float __x, float __y) { return __ocml_atan2_f32(__x, __y); } +float asinf(float __x) { return __ocml_asin_f32(__x); } + __DEVICE__ -inline float atanf(float __x) { return __ocml_atan_f32(__x); } +float asinhf(float __x) { return __ocml_asinh_f32(__x); } + __DEVICE__ -inline float atanhf(float __x) { return __ocml_atanh_f32(__x); } +float atan2f(float __x, float __y) { return __ocml_atan2_f32(__x, __y); } + __DEVICE__ -inline float cbrtf(float __x) { return __ocml_cbrt_f32(__x); } +float atanf(float __x) { return __ocml_atan_f32(__x); } + __DEVICE__ -inline float ceilf(float __x) { return __ocml_ceil_f32(__x); } +float atanhf(float __x) { return __ocml_atanh_f32(__x); } + __DEVICE__ -inline float copysignf(float __x, float __y) { - return __ocml_copysign_f32(__x, __y); -} +float cbrtf(float __x) { return __ocml_cbrt_f32(__x); } + __DEVICE__ -inline float cosf(float __x) { return __ocml_cos_f32(__x); } +float ceilf(float __x) { return __ocml_ceil_f32(__x); } + __DEVICE__ -inline float coshf(float __x) { return __ocml_cosh_f32(__x); } +float copysignf(float __x, float __y) { return __ocml_copysign_f32(__x, __y); } + __DEVICE__ -inline float cospif(float __x) { return __ocml_cospi_f32(__x); } +float cosf(float __x) { return __ocml_cos_f32(__x); } + __DEVICE__ -inline float cyl_bessel_i0f(float __x) { return __ocml_i0_f32(__x); } +float coshf(float __x) { return __ocml_cosh_f32(__x); } + __DEVICE__ -inline float cyl_bessel_i1f(float __x) { return __ocml_i1_f32(__x); } +float cospif(float __x) { return __ocml_cospi_f32(__x); } + __DEVICE__ -inline float erfcf(float __x) { return __ocml_erfc_f32(__x); } +float cyl_bessel_i0f(float __x) { return __ocml_i0_f32(__x); } + __DEVICE__ -inline float erfcinvf(float __x) { return __ocml_erfcinv_f32(__x); } +float cyl_bessel_i1f(float __x) { return __ocml_i1_f32(__x); } + __DEVICE__ -inline float erfcxf(float __x) { return __ocml_erfcx_f32(__x); } +float erfcf(float __x) { return __ocml_erfc_f32(__x); } + __DEVICE__ -inline float erff(float __x) { return __ocml_erf_f32(__x); } +float erfcinvf(float __x) { return __ocml_erfcinv_f32(__x); } + __DEVICE__ -inline float erfinvf(float __x) { return __ocml_erfinv_f32(__x); } +float erfcxf(float __x) { return __ocml_erfcx_f32(__x); } + __DEVICE__ -inline float exp10f(float __x) { return __ocml_exp10_f32(__x); } +float erff(float __x) { return __ocml_erf_f32(__x); } + __DEVICE__ -inline float exp2f(float __x) { return __ocml_exp2_f32(__x); } +float erfinvf(float __x) { return __ocml_erfinv_f32(__x); } + __DEVICE__ -inline float expf(float __x) { return __ocml_exp_f32(__x); } +float exp10f(float __x) { return __ocml_exp10_f32(__x); } + __DEVICE__ -inline float expm1f(float __x) { return __ocml_expm1_f32(__x); } +float exp2f(float __x) { return __ocml_exp2_f32(__x); } + __DEVICE__ -inline float fabsf(float __x) { return __ocml_fabs_f32(__x); } +float expf(float __x) { return __ocml_exp_f32(__x); } + __DEVICE__ -inline float fdimf(float __x, float __y) { return __ocml_fdim_f32(__x, __y); } +float expm1f(float __x) { return __ocml_expm1_f32(__x); } + __DEVICE__ -inline float fdividef(float __x, float __y) { return __x / __y; } +float fabsf(float __x) { return __ocml_fabs_f32(__x); } + __DEVICE__ -inline float floorf(float __x) { return __ocml_floor_f32(__x); } +float fdimf(float __x, float __y) { return __ocml_fdim_f32(__x, __y); } + __DEVICE__ -inline float fmaf(float __x, float __y, float __z) { +float fdividef(float __x, float __y) { return __x / __y; } + +__DEVICE__ +float floorf(float __x) { return __ocml_floor_f32(__x); } + +__DEVICE__ +float fmaf(float __x, float __y, float __z) { return __ocml_fma_f32(__x, __y, __z); } + __DEVICE__ -inline float fmaxf(float __x, float __y) { return __ocml_fmax_f32(__x, __y); } +float fmaxf(float __x, float __y) { return __ocml_fmax_f32(__x, __y); } + __DEVICE__ -inline float fminf(float __x, float __y) { return __ocml_fmin_f32(__x, __y); } +float fminf(float __x, float __y) { return __ocml_fmin_f32(__x, __y); } + __DEVICE__ -inline float fmodf(float __x, float __y) { return __ocml_fmod_f32(__x, __y); } +float fmodf(float __x, float __y) { return __ocml_fmod_f32(__x, __y); } + __DEVICE__ -inline float frexpf(float __x, int *__nptr) { +float frexpf(float __x, int *__nptr) { int __tmp; float __r = __ocml_frexp_f32(__x, (__attribute__((address_space(5))) int *)&__tmp); @@ -174,24 +253,31 @@ inline float frexpf(float __x, int *__nptr) { return __r; } + __DEVICE__ -inline float hypotf(float __x, float __y) { return __ocml_hypot_f32(__x, __y); } +float hypotf(float __x, float __y) { return __ocml_hypot_f32(__x, __y); } + __DEVICE__ -inline int ilogbf(float __x) { return __ocml_ilogb_f32(__x); } +int ilogbf(float __x) { return __ocml_ilogb_f32(__x); } + __DEVICE__ -inline __RETURN_TYPE isfinite(float __x) { return __ocml_isfinite_f32(__x); } +__RETURN_TYPE __finitef(float __x) { return __ocml_isfinite_f32(__x); } + __DEVICE__ -inline __RETURN_TYPE isinf(float __x) { return __ocml_isinf_f32(__x); } +__RETURN_TYPE __isinff(float __x) { return __ocml_isinf_f32(__x); } + __DEVICE__ -inline __RETURN_TYPE isnan(float __x) { return __ocml_isnan_f32(__x); } +__RETURN_TYPE __isnanf(float __x) { return __ocml_isnan_f32(__x); } + __DEVICE__ -inline float j0f(float __x) { return __ocml_j0_f32(__x); } +float j0f(float __x) { return __ocml_j0_f32(__x); } + __DEVICE__ -inline float j1f(float __x) { return __ocml_j1_f32(__x); } +float j1f(float __x) { return __ocml_j1_f32(__x); } + __DEVICE__ -inline float jnf(int __n, - float __x) { // TODO: we could use Ahmes multiplication - // and the Miller & Brown algorithm +float jnf(int __n, float __x) { // TODO: we could use Ahmes multiplication + // and the Miller & Brown algorithm // for linear recurrences to get O(log n) steps, but it's unclear if // it'd be beneficial in this case. if (__n == 0) @@ -209,50 +295,61 @@ inline float jnf(int __n, return __x1; } + __DEVICE__ -inline float ldexpf(float __x, int __e) { return __ocml_ldexp_f32(__x, __e); } +float ldexpf(float __x, int __e) { return __ocml_ldexp_f32(__x, __e); } + __DEVICE__ -inline float lgammaf(float __x) { return __ocml_lgamma_f32(__x); } +float lgammaf(float __x) { return __ocml_lgamma_f32(__x); } + __DEVICE__ -inline long long int llrintf(float __x) { return __ocml_rint_f32(__x); } +long long int llrintf(float __x) { return __ocml_rint_f32(__x); } + __DEVICE__ -inline long long int llroundf(float __x) { return __ocml_round_f32(__x); } +long long int llroundf(float __x) { return __ocml_round_f32(__x); } + __DEVICE__ -inline float log10f(float __x) { return __ocml_log10_f32(__x); } +float log10f(float __x) { return __ocml_log10_f32(__x); } + __DEVICE__ -inline float log1pf(float __x) { return __ocml_log1p_f32(__x); } +float log1pf(float __x) { return __ocml_log1p_f32(__x); } + __DEVICE__ -inline float log2f(float __x) { return __ocml_log2_f32(__x); } +float log2f(float __x) { return __ocml_log2_f32(__x); } + __DEVICE__ -inline float logbf(float __x) { return __ocml_logb_f32(__x); } +float logbf(float __x) { return __ocml_logb_f32(__x); } + __DEVICE__ -inline float logf(float __x) { return __ocml_log_f32(__x); } +float logf(float __x) { return __ocml_log_f32(__x); } + __DEVICE__ -inline long int lrintf(float __x) { return __ocml_rint_f32(__x); } +long int lrintf(float __x) { return __ocml_rint_f32(__x); } + __DEVICE__ -inline long int lroundf(float __x) { return __ocml_round_f32(__x); } +long int lroundf(float __x) { return __ocml_round_f32(__x); } + __DEVICE__ -inline float modff(float __x, float *__iptr) { +float modff(float __x, float *__iptr) { float __tmp; float __r = __ocml_modf_f32(__x, (__attribute__((address_space(5))) float *)&__tmp); *__iptr = __tmp; - return __r; } + __DEVICE__ -inline float nanf(const char *__tagp) { +float nanf(const char *__tagp) { union { float val; struct ieee_float { - uint32_t mantissa : 22; - uint32_t quiet : 1; - uint32_t exponent : 8; - uint32_t sign : 1; + unsigned int mantissa : 22; + unsigned int quiet : 1; + unsigned int exponent : 8; + unsigned int sign : 1; } bits; - - static_assert(sizeof(float) == sizeof(ieee_float), ""); } __tmp; + __static_assert_type_size_equal(sizeof(__tmp.val), sizeof(__tmp.bits)); __tmp.bits.sign = 0u; __tmp.bits.exponent = ~0u; @@ -261,28 +358,34 @@ inline float nanf(const char *__tagp) { return __tmp.val; } + __DEVICE__ -inline float nearbyintf(float __x) { return __ocml_nearbyint_f32(__x); } +float nearbyintf(float __x) { return __ocml_nearbyint_f32(__x); } + __DEVICE__ -inline float nextafterf(float __x, float __y) { +float nextafterf(float __x, float __y) { return __ocml_nextafter_f32(__x, __y); } + __DEVICE__ -inline float norm3df(float __x, float __y, float __z) { +float norm3df(float __x, float __y, float __z) { return __ocml_len3_f32(__x, __y, __z); } + __DEVICE__ -inline float norm4df(float __x, float __y, float __z, float __w) { +float norm4df(float __x, float __y, float __z, float __w) { return __ocml_len4_f32(__x, __y, __z, __w); } + __DEVICE__ -inline float normcdff(float __x) { return __ocml_ncdf_f32(__x); } +float normcdff(float __x) { return __ocml_ncdf_f32(__x); } + __DEVICE__ -inline float normcdfinvf(float __x) { return __ocml_ncdfinv_f32(__x); } +float normcdfinvf(float __x) { return __ocml_ncdfinv_f32(__x); } + __DEVICE__ -inline float -normf(int __dim, - const float *__a) { // TODO: placeholder until OCML adds support. +float normf(int __dim, + const float *__a) { // TODO: placeholder until OCML adds support. float __r = 0; while (__dim--) { __r += __a[0] * __a[0]; @@ -291,16 +394,23 @@ normf(int __dim, return __ocml_sqrt_f32(__r); } + __DEVICE__ -inline float powf(float __x, float __y) { return __ocml_pow_f32(__x, __y); } +float powf(float __x, float __y) { return __ocml_pow_f32(__x, __y); } + +__DEVICE__ +float powif(float __x, int __y) { return __ocml_pown_f32(__x, __y); } + __DEVICE__ -inline float rcbrtf(float __x) { return __ocml_rcbrt_f32(__x); } +float rcbrtf(float __x) { return __ocml_rcbrt_f32(__x); } + __DEVICE__ -inline float remainderf(float __x, float __y) { +float remainderf(float __x, float __y) { return __ocml_remainder_f32(__x, __y); } + __DEVICE__ -inline float remquof(float __x, float __y, int *__quo) { +float remquof(float __x, float __y, int *__quo) { int __tmp; float __r = __ocml_remquo_f32( __x, __y, (__attribute__((address_space(5))) int *)&__tmp); @@ -308,25 +418,26 @@ inline float remquof(float __x, float __y, int *__quo) { return __r; } + __DEVICE__ -inline float rhypotf(float __x, float __y) { - return __ocml_rhypot_f32(__x, __y); -} +float rhypotf(float __x, float __y) { return __ocml_rhypot_f32(__x, __y); } + __DEVICE__ -inline float rintf(float __x) { return __ocml_rint_f32(__x); } +float rintf(float __x) { return __ocml_rint_f32(__x); } + __DEVICE__ -inline float rnorm3df(float __x, float __y, float __z) { +float rnorm3df(float __x, float __y, float __z) { return __ocml_rlen3_f32(__x, __y, __z); } __DEVICE__ -inline float rnorm4df(float __x, float __y, float __z, float __w) { +float rnorm4df(float __x, float __y, float __z, float __w) { return __ocml_rlen4_f32(__x, __y, __z, __w); } + __DEVICE__ -inline float -rnormf(int __dim, - const float *__a) { // TODO: placeholder until OCML adds support. +float rnormf(int __dim, + const float *__a) { // TODO: placeholder until OCML adds support. float __r = 0; while (__dim--) { __r += __a[0] * __a[0]; @@ -335,59 +446,74 @@ rnormf(int __dim, return __ocml_rsqrt_f32(__r); } + __DEVICE__ -inline float roundf(float __x) { return __ocml_round_f32(__x); } +float roundf(float __x) { return __ocml_round_f32(__x); } + __DEVICE__ -inline float rsqrtf(float __x) { return __ocml_rsqrt_f32(__x); } +float rsqrtf(float __x) { return __ocml_rsqrt_f32(__x); } + __DEVICE__ -inline float scalblnf(float __x, long int __n) { +float scalblnf(float __x, long int __n) { return (__n < INT_MAX) ? __ocml_scalbn_f32(__x, __n) : __ocml_scalb_f32(__x, __n); } + __DEVICE__ -inline float scalbnf(float __x, int __n) { return __ocml_scalbn_f32(__x, __n); } +float scalbnf(float __x, int __n) { return __ocml_scalbn_f32(__x, __n); } + __DEVICE__ -inline __RETURN_TYPE signbit(float __x) { return __ocml_signbit_f32(__x); } +__RETURN_TYPE __signbitf(float __x) { return __ocml_signbit_f32(__x); } + __DEVICE__ -inline void sincosf(float __x, float *__sinptr, float *__cosptr) { +void sincosf(float __x, float *__sinptr, float *__cosptr) { float __tmp; - *__sinptr = __ocml_sincos_f32(__x, (__attribute__((address_space(5))) float *)&__tmp); *__cosptr = __tmp; } + __DEVICE__ -inline void sincospif(float __x, float *__sinptr, float *__cosptr) { +void sincospif(float __x, float *__sinptr, float *__cosptr) { float __tmp; - *__sinptr = __ocml_sincospi_f32( __x, (__attribute__((address_space(5))) float *)&__tmp); *__cosptr = __tmp; } + __DEVICE__ -inline float sinf(float __x) { return __ocml_sin_f32(__x); } +float sinf(float __x) { return __ocml_sin_f32(__x); } + __DEVICE__ -inline float sinhf(float __x) { return __ocml_sinh_f32(__x); } +float sinhf(float __x) { return __ocml_sinh_f32(__x); } + __DEVICE__ -inline float sinpif(float __x) { return __ocml_sinpi_f32(__x); } +float sinpif(float __x) { return __ocml_sinpi_f32(__x); } + __DEVICE__ -inline float sqrtf(float __x) { return __ocml_sqrt_f32(__x); } +float sqrtf(float __x) { return __ocml_sqrt_f32(__x); } + __DEVICE__ -inline float tanf(float __x) { return __ocml_tan_f32(__x); } +float tanf(float __x) { return __ocml_tan_f32(__x); } + __DEVICE__ -inline float tanhf(float __x) { return __ocml_tanh_f32(__x); } +float tanhf(float __x) { return __ocml_tanh_f32(__x); } + __DEVICE__ -inline float tgammaf(float __x) { return __ocml_tgamma_f32(__x); } +float tgammaf(float __x) { return __ocml_tgamma_f32(__x); } + __DEVICE__ -inline float truncf(float __x) { return __ocml_trunc_f32(__x); } +float truncf(float __x) { return __ocml_trunc_f32(__x); } + __DEVICE__ -inline float y0f(float __x) { return __ocml_y0_f32(__x); } +float y0f(float __x) { return __ocml_y0_f32(__x); } + __DEVICE__ -inline float y1f(float __x) { return __ocml_y1_f32(__x); } +float y1f(float __x) { return __ocml_y1_f32(__x); } + __DEVICE__ -inline float ynf(int __n, - float __x) { // TODO: we could use Ahmes multiplication - // and the Miller & Brown algorithm +float ynf(int __n, float __x) { // TODO: we could use Ahmes multiplication + // and the Miller & Brown algorithm // for linear recurrences to get O(log n) steps, but it's unclear if // it'd be beneficial in this case. Placeholder until OCML adds // support. @@ -408,290 +534,343 @@ inline float ynf(int __n, } // BEGIN INTRINSICS + __DEVICE__ -inline float __cosf(float __x) { return __ocml_native_cos_f32(__x); } +float __cosf(float __x) { return __ocml_native_cos_f32(__x); } + __DEVICE__ -inline float __exp10f(float __x) { return __ocml_native_exp10_f32(__x); } +float __exp10f(float __x) { return __ocml_native_exp10_f32(__x); } + __DEVICE__ -inline float __expf(float __x) { return __ocml_native_exp_f32(__x); } +float __expf(float __x) { return __ocml_native_exp_f32(__x); } + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __fadd_rd(float __x, float __y) { - return __ocml_add_rtn_f32(__x, __y); -} -#endif +float __fadd_rd(float __x, float __y) { return __ocml_add_rtn_f32(__x, __y); } __DEVICE__ -inline float __fadd_rn(float __x, float __y) { return __x + __y; } -#if defined OCML_BASIC_ROUNDED_OPERATIONS +float __fadd_rn(float __x, float __y) { return __ocml_add_rte_f32(__x, __y); } __DEVICE__ -inline float __fadd_ru(float __x, float __y) { - return __ocml_add_rtp_f32(__x, __y); -} +float __fadd_ru(float __x, float __y) { return __ocml_add_rtp_f32(__x, __y); } __DEVICE__ -inline float __fadd_rz(float __x, float __y) { - return __ocml_add_rtz_f32(__x, __y); -} +float __fadd_rz(float __x, float __y) { return __ocml_add_rtz_f32(__x, __y); } +#else __DEVICE__ -inline float __fdiv_rd(float __x, float __y) { - return __ocml_div_rtn_f32(__x, __y); -} +float __fadd_rn(float __x, float __y) { return __x + __y; } #endif -__DEVICE__ -inline float __fdiv_rn(float __x, float __y) { return __x / __y; } + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __fdiv_ru(float __x, float __y) { - return __ocml_div_rtp_f32(__x, __y); -} +float __fdiv_rd(float __x, float __y) { return __ocml_div_rtn_f32(__x, __y); } __DEVICE__ -inline float __fdiv_rz(float __x, float __y) { - return __ocml_div_rtz_f32(__x, __y); -} +float __fdiv_rn(float __x, float __y) { return __ocml_div_rte_f32(__x, __y); } +__DEVICE__ +float __fdiv_ru(float __x, float __y) { return __ocml_div_rtp_f32(__x, __y); } +__DEVICE__ +float __fdiv_rz(float __x, float __y) { return __ocml_div_rtz_f32(__x, __y); } +#else +__DEVICE__ +float __fdiv_rn(float __x, float __y) { return __x / __y; } #endif + __DEVICE__ -inline float __fdividef(float __x, float __y) { return __x / __y; } +float __fdividef(float __x, float __y) { return __x / __y; } + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __fmaf_rd(float __x, float __y, float __z) { +float __fmaf_rd(float __x, float __y, float __z) { return __ocml_fma_rtn_f32(__x, __y, __z); } -#endif __DEVICE__ -inline float __fmaf_rn(float __x, float __y, float __z) { - return __ocml_fma_f32(__x, __y, __z); +float __fmaf_rn(float __x, float __y, float __z) { + return __ocml_fma_rte_f32(__x, __y, __z); } -#if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __fmaf_ru(float __x, float __y, float __z) { +float __fmaf_ru(float __x, float __y, float __z) { return __ocml_fma_rtp_f32(__x, __y, __z); } __DEVICE__ -inline float __fmaf_rz(float __x, float __y, float __z) { +float __fmaf_rz(float __x, float __y, float __z) { return __ocml_fma_rtz_f32(__x, __y, __z); } +#else __DEVICE__ -inline float __fmul_rd(float __x, float __y) { - return __ocml_mul_rtn_f32(__x, __y); +float __fmaf_rn(float __x, float __y, float __z) { + return __ocml_fma_f32(__x, __y, __z); } #endif -__DEVICE__ -inline float __fmul_rn(float __x, float __y) { return __x * __y; } + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __fmul_ru(float __x, float __y) { - return __ocml_mul_rtp_f32(__x, __y); -} +float __fmul_rd(float __x, float __y) { return __ocml_mul_rtn_f32(__x, __y); } __DEVICE__ -inline float __fmul_rz(float __x, float __y) { - return __ocml_mul_rtz_f32(__x, __y); -} +float __fmul_rn(float __x, float __y) { return __ocml_mul_rte_f32(__x, __y); } __DEVICE__ -inline float __frcp_rd(float __x) { return __llvm_amdgcn_rcp_f32(__x); } -#endif +float __fmul_ru(float __x, float __y) { return __ocml_mul_rtp_f32(__x, __y); } +__DEVICE__ +float __fmul_rz(float __x, float __y) { return __ocml_mul_rtz_f32(__x, __y); } +#else __DEVICE__ -inline float __frcp_rn(float __x) { return __llvm_amdgcn_rcp_f32(__x); } +float __fmul_rn(float __x, float __y) { return __x * __y; } +#endif + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __frcp_ru(float __x) { return __llvm_amdgcn_rcp_f32(__x); } +float __frcp_rd(float __x) { return __ocml_div_rtn_f32(1.0f, __x); } __DEVICE__ -inline float __frcp_rz(float __x) { return __llvm_amdgcn_rcp_f32(__x); } -#endif +float __frcp_rn(float __x) { return __ocml_div_rte_f32(1.0f, __x); } __DEVICE__ -inline float __frsqrt_rn(float __x) { return __llvm_amdgcn_rsq_f32(__x); } -#if defined OCML_BASIC_ROUNDED_OPERATIONS +float __frcp_ru(float __x) { return __ocml_div_rtp_f32(1.0f, __x); } +__DEVICE__ +float __frcp_rz(float __x) { return __ocml_div_rtz_f32(1.0f, __x); } +#else __DEVICE__ -inline float __fsqrt_rd(float __x) { return __ocml_sqrt_rtn_f32(__x); } +float __frcp_rn(float __x) { return 1.0f / __x; } #endif + __DEVICE__ -inline float __fsqrt_rn(float __x) { return __ocml_native_sqrt_f32(__x); } +float __frsqrt_rn(float __x) { return __llvm_amdgcn_rsq_f32(__x); } + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __fsqrt_ru(float __x) { return __ocml_sqrt_rtp_f32(__x); } +float __fsqrt_rd(float __x) { return __ocml_sqrt_rtn_f32(__x); } __DEVICE__ -inline float __fsqrt_rz(float __x) { return __ocml_sqrt_rtz_f32(__x); } +float __fsqrt_rn(float __x) { return __ocml_sqrt_rte_f32(__x); } __DEVICE__ -inline float __fsub_rd(float __x, float __y) { - return __ocml_sub_rtn_f32(__x, __y); -} -#endif +float __fsqrt_ru(float __x) { return __ocml_sqrt_rtp_f32(__x); } +__DEVICE__ +float __fsqrt_rz(float __x) { return __ocml_sqrt_rtz_f32(__x); } +#else __DEVICE__ -inline float __fsub_rn(float __x, float __y) { return __x - __y; } +float __fsqrt_rn(float __x) { return __ocml_native_sqrt_f32(__x); } +#endif + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline float __fsub_ru(float __x, float __y) { - return __ocml_sub_rtp_f32(__x, __y); -} +float __fsub_rd(float __x, float __y) { return __ocml_sub_rtn_f32(__x, __y); } __DEVICE__ -inline float __fsub_rz(float __x, float __y) { - return __ocml_sub_rtz_f32(__x, __y); -} +float __fsub_rn(float __x, float __y) { return __ocml_sub_rte_f32(__x, __y); } +__DEVICE__ +float __fsub_ru(float __x, float __y) { return __ocml_sub_rtp_f32(__x, __y); } +__DEVICE__ +float __fsub_rz(float __x, float __y) { return __ocml_sub_rtz_f32(__x, __y); } +#else +__DEVICE__ +float __fsub_rn(float __x, float __y) { return __x - __y; } #endif + __DEVICE__ -inline float __log10f(float __x) { return __ocml_native_log10_f32(__x); } +float __log10f(float __x) { return __ocml_native_log10_f32(__x); } + __DEVICE__ -inline float __log2f(float __x) { return __ocml_native_log2_f32(__x); } +float __log2f(float __x) { return __ocml_native_log2_f32(__x); } + __DEVICE__ -inline float __logf(float __x) { return __ocml_native_log_f32(__x); } +float __logf(float __x) { return __ocml_native_log_f32(__x); } + __DEVICE__ -inline float __powf(float __x, float __y) { return __ocml_pow_f32(__x, __y); } +float __powf(float __x, float __y) { return __ocml_pow_f32(__x, __y); } + __DEVICE__ -inline float __saturatef(float __x) { - return (__x < 0) ? 0 : ((__x > 1) ? 1 : __x); -} +float __saturatef(float __x) { return (__x < 0) ? 0 : ((__x > 1) ? 1 : __x); } + __DEVICE__ -inline void __sincosf(float __x, float *__sinptr, float *__cosptr) { +void __sincosf(float __x, float *__sinptr, float *__cosptr) { *__sinptr = __ocml_native_sin_f32(__x); *__cosptr = __ocml_native_cos_f32(__x); } + __DEVICE__ -inline float __sinf(float __x) { return __ocml_native_sin_f32(__x); } +float __sinf(float __x) { return __ocml_native_sin_f32(__x); } + __DEVICE__ -inline float __tanf(float __x) { return __ocml_tan_f32(__x); } +float __tanf(float __x) { return __ocml_tan_f32(__x); } // END INTRINSICS // END FLOAT // BEGIN DOUBLE __DEVICE__ -inline double abs(double __x) { return __ocml_fabs_f64(__x); } -__DEVICE__ -inline double acos(double __x) { return __ocml_acos_f64(__x); } +double acos(double __x) { return __ocml_acos_f64(__x); } + __DEVICE__ -inline double acosh(double __x) { return __ocml_acosh_f64(__x); } +double acosh(double __x) { return __ocml_acosh_f64(__x); } + __DEVICE__ -inline double asin(double __x) { return __ocml_asin_f64(__x); } +double asin(double __x) { return __ocml_asin_f64(__x); } + __DEVICE__ -inline double asinh(double __x) { return __ocml_asinh_f64(__x); } +double asinh(double __x) { return __ocml_asinh_f64(__x); } + __DEVICE__ -inline double atan(double __x) { return __ocml_atan_f64(__x); } +double atan(double __x) { return __ocml_atan_f64(__x); } + __DEVICE__ -inline double atan2(double __x, double __y) { - return __ocml_atan2_f64(__x, __y); -} +double atan2(double __x, double __y) { return __ocml_atan2_f64(__x, __y); } + __DEVICE__ -inline double atanh(double __x) { return __ocml_atanh_f64(__x); } +double atanh(double __x) { return __ocml_atanh_f64(__x); } + __DEVICE__ -inline double cbrt(double __x) { return __ocml_cbrt_f64(__x); } +double cbrt(double __x) { return __ocml_cbrt_f64(__x); } + __DEVICE__ -inline double ceil(double __x) { return __ocml_ceil_f64(__x); } +double ceil(double __x) { return __ocml_ceil_f64(__x); } + __DEVICE__ -inline double copysign(double __x, double __y) { +double copysign(double __x, double __y) { return __ocml_copysign_f64(__x, __y); } + __DEVICE__ -inline double cos(double __x) { return __ocml_cos_f64(__x); } +double cos(double __x) { return __ocml_cos_f64(__x); } + __DEVICE__ -inline double cosh(double __x) { return __ocml_cosh_f64(__x); } +double cosh(double __x) { return __ocml_cosh_f64(__x); } + __DEVICE__ -inline double cospi(double __x) { return __ocml_cospi_f64(__x); } +double cospi(double __x) { return __ocml_cospi_f64(__x); } + __DEVICE__ -inline double cyl_bessel_i0(double __x) { return __ocml_i0_f64(__x); } +double cyl_bessel_i0(double __x) { return __ocml_i0_f64(__x); } + __DEVICE__ -inline double cyl_bessel_i1(double __x) { return __ocml_i1_f64(__x); } +double cyl_bessel_i1(double __x) { return __ocml_i1_f64(__x); } + __DEVICE__ -inline double erf(double __x) { return __ocml_erf_f64(__x); } +double erf(double __x) { return __ocml_erf_f64(__x); } + __DEVICE__ -inline double erfc(double __x) { return __ocml_erfc_f64(__x); } +double erfc(double __x) { return __ocml_erfc_f64(__x); } + __DEVICE__ -inline double erfcinv(double __x) { return __ocml_erfcinv_f64(__x); } +double erfcinv(double __x) { return __ocml_erfcinv_f64(__x); } + __DEVICE__ -inline double erfcx(double __x) { return __ocml_erfcx_f64(__x); } +double erfcx(double __x) { return __ocml_erfcx_f64(__x); } + __DEVICE__ -inline double erfinv(double __x) { return __ocml_erfinv_f64(__x); } +double erfinv(double __x) { return __ocml_erfinv_f64(__x); } + __DEVICE__ -inline double exp(double __x) { return __ocml_exp_f64(__x); } +double exp(double __x) { return __ocml_exp_f64(__x); } + __DEVICE__ -inline double exp10(double __x) { return __ocml_exp10_f64(__x); } +double exp10(double __x) { return __ocml_exp10_f64(__x); } + __DEVICE__ -inline double exp2(double __x) { return __ocml_exp2_f64(__x); } +double exp2(double __x) { return __ocml_exp2_f64(__x); } + __DEVICE__ -inline double expm1(double __x) { return __ocml_expm1_f64(__x); } +double expm1(double __x) { return __ocml_expm1_f64(__x); } + __DEVICE__ -inline double fabs(double __x) { return __ocml_fabs_f64(__x); } +double fabs(double __x) { return __ocml_fabs_f64(__x); } + __DEVICE__ -inline double fdim(double __x, double __y) { return __ocml_fdim_f64(__x, __y); } +double fdim(double __x, double __y) { return __ocml_fdim_f64(__x, __y); } + __DEVICE__ -inline double floor(double __x) { return __ocml_floor_f64(__x); } +double floor(double __x) { return __ocml_floor_f64(__x); } + __DEVICE__ -inline double fma(double __x, double __y, double __z) { +double fma(double __x, double __y, double __z) { return __ocml_fma_f64(__x, __y, __z); } + __DEVICE__ -inline double fmax(double __x, double __y) { return __ocml_fmax_f64(__x, __y); } +double fmax(double __x, double __y) { return __ocml_fmax_f64(__x, __y); } + __DEVICE__ -inline double fmin(double __x, double __y) { return __ocml_fmin_f64(__x, __y); } +double fmin(double __x, double __y) { return __ocml_fmin_f64(__x, __y); } + __DEVICE__ -inline double fmod(double __x, double __y) { return __ocml_fmod_f64(__x, __y); } +double fmod(double __x, double __y) { return __ocml_fmod_f64(__x, __y); } + __DEVICE__ -inline double frexp(double __x, int *__nptr) { +double frexp(double __x, int *__nptr) { int __tmp; double __r = __ocml_frexp_f64(__x, (__attribute__((address_space(5))) int *)&__tmp); *__nptr = __tmp; - return __r; } + __DEVICE__ -inline double hypot(double __x, double __y) { - return __ocml_hypot_f64(__x, __y); -} +double hypot(double __x, double __y) { return __ocml_hypot_f64(__x, __y); } + __DEVICE__ -inline int ilogb(double __x) { return __ocml_ilogb_f64(__x); } +int ilogb(double __x) { return __ocml_ilogb_f64(__x); } + __DEVICE__ -inline __RETURN_TYPE isfinite(double __x) { return __ocml_isfinite_f64(__x); } +__RETURN_TYPE __finite(double __x) { return __ocml_isfinite_f64(__x); } + __DEVICE__ -inline __RETURN_TYPE isinf(double __x) { return __ocml_isinf_f64(__x); } +__RETURN_TYPE __isinf(double __x) { return __ocml_isinf_f64(__x); } + __DEVICE__ -inline __RETURN_TYPE isnan(double __x) { return __ocml_isnan_f64(__x); } +__RETURN_TYPE __isnan(double __x) { return __ocml_isnan_f64(__x); } + __DEVICE__ -inline double j0(double __x) { return __ocml_j0_f64(__x); } +double j0(double __x) { return __ocml_j0_f64(__x); } + __DEVICE__ -inline double j1(double __x) { return __ocml_j1_f64(__x); } +double j1(double __x) { return __ocml_j1_f64(__x); } + __DEVICE__ -inline double jn(int __n, - double __x) { // TODO: we could use Ahmes multiplication - // and the Miller & Brown algorithm +double jn(int __n, double __x) { // TODO: we could use Ahmes multiplication + // and the Miller & Brown algorithm // for linear recurrences to get O(log n) steps, but it's unclear if // it'd be beneficial in this case. Placeholder until OCML adds // support. if (__n == 0) - return j0f(__x); + return j0(__x); if (__n == 1) - return j1f(__x); + return j1(__x); - double __x0 = j0f(__x); - double __x1 = j1f(__x); + double __x0 = j0(__x); + double __x1 = j1(__x); for (int __i = 1; __i < __n; ++__i) { double __x2 = (2 * __i) / __x * __x1 - __x0; __x0 = __x1; __x1 = __x2; } - return __x1; } + __DEVICE__ -inline double ldexp(double __x, int __e) { return __ocml_ldexp_f64(__x, __e); } +double ldexp(double __x, int __e) { return __ocml_ldexp_f64(__x, __e); } + __DEVICE__ -inline double lgamma(double __x) { return __ocml_lgamma_f64(__x); } +double lgamma(double __x) { return __ocml_lgamma_f64(__x); } + __DEVICE__ -inline long long int llrint(double __x) { return __ocml_rint_f64(__x); } +long long int llrint(double __x) { return __ocml_rint_f64(__x); } + __DEVICE__ -inline long long int llround(double __x) { return __ocml_round_f64(__x); } +long long int llround(double __x) { return __ocml_round_f64(__x); } + __DEVICE__ -inline double log(double __x) { return __ocml_log_f64(__x); } +double log(double __x) { return __ocml_log_f64(__x); } + __DEVICE__ -inline double log10(double __x) { return __ocml_log10_f64(__x); } +double log10(double __x) { return __ocml_log10_f64(__x); } + __DEVICE__ -inline double log1p(double __x) { return __ocml_log1p_f64(__x); } +double log1p(double __x) { return __ocml_log1p_f64(__x); } + __DEVICE__ -inline double log2(double __x) { return __ocml_log2_f64(__x); } +double log2(double __x) { return __ocml_log2_f64(__x); } + __DEVICE__ -inline double logb(double __x) { return __ocml_logb_f64(__x); } +double logb(double __x) { return __ocml_logb_f64(__x); } + __DEVICE__ -inline long int lrint(double __x) { return __ocml_rint_f64(__x); } +long int lrint(double __x) { return __ocml_rint_f64(__x); } + __DEVICE__ -inline long int lround(double __x) { return __ocml_round_f64(__x); } +long int lround(double __x) { return __ocml_round_f64(__x); } + __DEVICE__ -inline double modf(double __x, double *__iptr) { +double modf(double __x, double *__iptr) { double __tmp; double __r = __ocml_modf_f64(__x, (__attribute__((address_space(5))) double *)&__tmp); @@ -699,8 +878,9 @@ inline double modf(double __x, double *__iptr) { return __r; } + __DEVICE__ -inline double nan(const char *__tagp) { +double nan(const char *__tagp) { #if !_WIN32 union { double val; @@ -710,8 +890,8 @@ inline double nan(const char *__tagp) { uint32_t exponent : 11; uint32_t sign : 1; } bits; - static_assert(sizeof(double) == sizeof(ieee_double), ""); } __tmp; + __static_assert_type_size_equal(sizeof(__tmp.val), sizeof(__tmp.bits)); __tmp.bits.sign = 0u; __tmp.bits.exponent = ~0u; @@ -720,22 +900,24 @@ inline double nan(const char *__tagp) { return __tmp.val; #else - static_assert(sizeof(uint64_t) == sizeof(double)); - uint64_t val = __make_mantissa(__tagp); - val |= 0xFFF << 51; - return *reinterpret_cast<double *>(&val); + __static_assert_type_size_equal(sizeof(uint64_t), sizeof(double)); + uint64_t __val = __make_mantissa(__tagp); + __val |= 0xFFF << 51; + return *reinterpret_cast<double *>(&__val); #endif } + __DEVICE__ -inline double nearbyint(double __x) { return __ocml_nearbyint_f64(__x); } +double nearbyint(double __x) { return __ocml_nearbyint_f64(__x); } + __DEVICE__ -inline double nextafter(double __x, double __y) { +double nextafter(double __x, double __y) { return __ocml_nextafter_f64(__x, __y); } + __DEVICE__ -inline double -norm(int __dim, - const double *__a) { // TODO: placeholder until OCML adds support. +double norm(int __dim, + const double *__a) { // TODO: placeholder until OCML adds support. double __r = 0; while (__dim--) { __r += __a[0] * __a[0]; @@ -744,28 +926,39 @@ norm(int __dim, return __ocml_sqrt_f64(__r); } + __DEVICE__ -inline double norm3d(double __x, double __y, double __z) { +double norm3d(double __x, double __y, double __z) { return __ocml_len3_f64(__x, __y, __z); } + __DEVICE__ -inline double norm4d(double __x, double __y, double __z, double __w) { +double norm4d(double __x, double __y, double __z, double __w) { return __ocml_len4_f64(__x, __y, __z, __w); } + +__DEVICE__ +double normcdf(double __x) { return __ocml_ncdf_f64(__x); } + __DEVICE__ -inline double normcdf(double __x) { return __ocml_ncdf_f64(__x); } +double normcdfinv(double __x) { return __ocml_ncdfinv_f64(__x); } + __DEVICE__ -inline double normcdfinv(double __x) { return __ocml_ncdfinv_f64(__x); } +double pow(double __x, double __y) { return __ocml_pow_f64(__x, __y); } + __DEVICE__ -inline double pow(double __x, double __y) { return __ocml_pow_f64(__x, __y); } +double powi(double __x, int __y) { return __ocml_pown_f64(__x, __y); } + __DEVICE__ -inline double rcbrt(double __x) { return __ocml_rcbrt_f64(__x); } +double rcbrt(double __x) { return __ocml_rcbrt_f64(__x); } + __DEVICE__ -inline double remainder(double __x, double __y) { +double remainder(double __x, double __y) { return __ocml_remainder_f64(__x, __y); } + __DEVICE__ -inline double remquo(double __x, double __y, int *__quo) { +double remquo(double __x, double __y, int *__quo) { int __tmp; double __r = __ocml_remquo_f64( __x, __y, (__attribute__((address_space(5))) int *)&__tmp); @@ -773,16 +966,16 @@ inline double remquo(double __x, double __y, int *__quo) { return __r; } + __DEVICE__ -inline double rhypot(double __x, double __y) { - return __ocml_rhypot_f64(__x, __y); -} +double rhypot(double __x, double __y) { return __ocml_rhypot_f64(__x, __y); } + __DEVICE__ -inline double rint(double __x) { return __ocml_rint_f64(__x); } +double rint(double __x) { return __ocml_rint_f64(__x); } + __DEVICE__ -inline double -rnorm(int __dim, - const double *__a) { // TODO: placeholder until OCML adds support. +double rnorm(int __dim, + const double *__a) { // TODO: placeholder until OCML adds support. double __r = 0; while (__dim--) { __r += __a[0] * __a[0]; @@ -791,77 +984,93 @@ rnorm(int __dim, return __ocml_rsqrt_f64(__r); } + __DEVICE__ -inline double rnorm3d(double __x, double __y, double __z) { +double rnorm3d(double __x, double __y, double __z) { return __ocml_rlen3_f64(__x, __y, __z); } + __DEVICE__ -inline double rnorm4d(double __x, double __y, double __z, double __w) { +double rnorm4d(double __x, double __y, double __z, double __w) { return __ocml_rlen4_f64(__x, __y, __z, __w); } + __DEVICE__ -inline double round(double __x) { return __ocml_round_f64(__x); } +double round(double __x) { return __ocml_round_f64(__x); } + __DEVICE__ -inline double rsqrt(double __x) { return __ocml_rsqrt_f64(__x); } +double rsqrt(double __x) { return __ocml_rsqrt_f64(__x); } + __DEVICE__ -inline double scalbln(double __x, long int __n) { +double scalbln(double __x, long int __n) { return (__n < INT_MAX) ? __ocml_scalbn_f64(__x, __n) : __ocml_scalb_f64(__x, __n); } __DEVICE__ -inline double scalbn(double __x, int __n) { - return __ocml_scalbn_f64(__x, __n); -} +double scalbn(double __x, int __n) { return __ocml_scalbn_f64(__x, __n); } + __DEVICE__ -inline __RETURN_TYPE signbit(double __x) { return __ocml_signbit_f64(__x); } +__RETURN_TYPE __signbit(double __x) { return __ocml_signbit_f64(__x); } + __DEVICE__ -inline double sin(double __x) { return __ocml_sin_f64(__x); } +double sin(double __x) { return __ocml_sin_f64(__x); } + __DEVICE__ -inline void sincos(double __x, double *__sinptr, double *__cosptr) { +void sincos(double __x, double *__sinptr, double *__cosptr) { double __tmp; *__sinptr = __ocml_sincos_f64( __x, (__attribute__((address_space(5))) double *)&__tmp); *__cosptr = __tmp; } + __DEVICE__ -inline void sincospi(double __x, double *__sinptr, double *__cosptr) { +void sincospi(double __x, double *__sinptr, double *__cosptr) { double __tmp; *__sinptr = __ocml_sincospi_f64( __x, (__attribute__((address_space(5))) double *)&__tmp); *__cosptr = __tmp; } + __DEVICE__ -inline double sinh(double __x) { return __ocml_sinh_f64(__x); } +double sinh(double __x) { return __ocml_sinh_f64(__x); } + __DEVICE__ -inline double sinpi(double __x) { return __ocml_sinpi_f64(__x); } +double sinpi(double __x) { return __ocml_sinpi_f64(__x); } + __DEVICE__ -inline double sqrt(double __x) { return __ocml_sqrt_f64(__x); } +double sqrt(double __x) { return __ocml_sqrt_f64(__x); } + __DEVICE__ -inline double tan(double __x) { return __ocml_tan_f64(__x); } +double tan(double __x) { return __ocml_tan_f64(__x); } + __DEVICE__ -inline double tanh(double __x) { return __ocml_tanh_f64(__x); } +double tanh(double __x) { return __ocml_tanh_f64(__x); } + __DEVICE__ -inline double tgamma(double __x) { return __ocml_tgamma_f64(__x); } +double tgamma(double __x) { return __ocml_tgamma_f64(__x); } + __DEVICE__ -inline double trunc(double __x) { return __ocml_trunc_f64(__x); } +double trunc(double __x) { return __ocml_trunc_f64(__x); } + __DEVICE__ -inline double y0(double __x) { return __ocml_y0_f64(__x); } +double y0(double __x) { return __ocml_y0_f64(__x); } + __DEVICE__ -inline double y1(double __x) { return __ocml_y1_f64(__x); } +double y1(double __x) { return __ocml_y1_f64(__x); } + __DEVICE__ -inline double yn(int __n, - double __x) { // TODO: we could use Ahmes multiplication - // and the Miller & Brown algorithm +double yn(int __n, double __x) { // TODO: we could use Ahmes multiplication + // and the Miller & Brown algorithm // for linear recurrences to get O(log n) steps, but it's unclear if // it'd be beneficial in this case. Placeholder until OCML adds // support. if (__n == 0) - return j0f(__x); + return y0(__x); if (__n == 1) - return j1f(__x); + return y1(__x); - double __x0 = j0f(__x); - double __x1 = j1f(__x); + double __x0 = y0(__x); + double __x1 = y1(__x); for (int __i = 1; __i < __n; ++__i) { double __x2 = (2 * __i) / __x * __x1 - __x0; __x0 = __x1; @@ -874,296 +1083,182 @@ inline double yn(int __n, // BEGIN INTRINSICS #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline double __dadd_rd(double __x, double __y) { +double __dadd_rd(double __x, double __y) { return __ocml_add_rtn_f64(__x, __y); } -#endif __DEVICE__ -inline double __dadd_rn(double __x, double __y) { return __x + __y; } -#if defined OCML_BASIC_ROUNDED_OPERATIONS +double __dadd_rn(double __x, double __y) { + return __ocml_add_rte_f64(__x, __y); +} __DEVICE__ -inline double __dadd_ru(double __x, double __y) { +double __dadd_ru(double __x, double __y) { return __ocml_add_rtp_f64(__x, __y); } __DEVICE__ -inline double __dadd_rz(double __x, double __y) { +double __dadd_rz(double __x, double __y) { return __ocml_add_rtz_f64(__x, __y); } +#else __DEVICE__ -inline double __ddiv_rd(double __x, double __y) { +double __dadd_rn(double __x, double __y) { return __x + __y; } +#endif + +#if defined OCML_BASIC_ROUNDED_OPERATIONS +__DEVICE__ +double __ddiv_rd(double __x, double __y) { return __ocml_div_rtn_f64(__x, __y); } -#endif __DEVICE__ -inline double __ddiv_rn(double __x, double __y) { return __x / __y; } -#if defined OCML_BASIC_ROUNDED_OPERATIONS +double __ddiv_rn(double __x, double __y) { + return __ocml_div_rte_f64(__x, __y); +} __DEVICE__ -inline double __ddiv_ru(double __x, double __y) { +double __ddiv_ru(double __x, double __y) { return __ocml_div_rtp_f64(__x, __y); } __DEVICE__ -inline double __ddiv_rz(double __x, double __y) { +double __ddiv_rz(double __x, double __y) { return __ocml_div_rtz_f64(__x, __y); } +#else +__DEVICE__ +double __ddiv_rn(double __x, double __y) { return __x / __y; } +#endif + +#if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline double __dmul_rd(double __x, double __y) { +double __dmul_rd(double __x, double __y) { return __ocml_mul_rtn_f64(__x, __y); } -#endif __DEVICE__ -inline double __dmul_rn(double __x, double __y) { return __x * __y; } -#if defined OCML_BASIC_ROUNDED_OPERATIONS +double __dmul_rn(double __x, double __y) { + return __ocml_mul_rte_f64(__x, __y); +} __DEVICE__ -inline double __dmul_ru(double __x, double __y) { +double __dmul_ru(double __x, double __y) { return __ocml_mul_rtp_f64(__x, __y); } __DEVICE__ -inline double __dmul_rz(double __x, double __y) { +double __dmul_rz(double __x, double __y) { return __ocml_mul_rtz_f64(__x, __y); } +#else __DEVICE__ -inline double __drcp_rd(double __x) { return __llvm_amdgcn_rcp_f64(__x); } +double __dmul_rn(double __x, double __y) { return __x * __y; } #endif -__DEVICE__ -inline double __drcp_rn(double __x) { return __llvm_amdgcn_rcp_f64(__x); } + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline double __drcp_ru(double __x) { return __llvm_amdgcn_rcp_f64(__x); } +double __drcp_rd(double __x) { return __ocml_div_rtn_f64(1.0, __x); } __DEVICE__ -inline double __drcp_rz(double __x) { return __llvm_amdgcn_rcp_f64(__x); } +double __drcp_rn(double __x) { return __ocml_div_rte_f64(1.0, __x); } __DEVICE__ -inline double __dsqrt_rd(double __x) { return __ocml_sqrt_rtn_f64(__x); } -#endif +double __drcp_ru(double __x) { return __ocml_div_rtp_f64(1.0, __x); } __DEVICE__ -inline double __dsqrt_rn(double __x) { return __ocml_sqrt_f64(__x); } +double __drcp_rz(double __x) { return __ocml_div_rtz_f64(1.0, __x); } +#else +__DEVICE__ +double __drcp_rn(double __x) { return 1.0 / __x; } +#endif + #if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline double __dsqrt_ru(double __x) { return __ocml_sqrt_rtp_f64(__x); } +double __dsqrt_rd(double __x) { return __ocml_sqrt_rtn_f64(__x); } +__DEVICE__ +double __dsqrt_rn(double __x) { return __ocml_sqrt_rte_f64(__x); } +__DEVICE__ +double __dsqrt_ru(double __x) { return __ocml_sqrt_rtp_f64(__x); } __DEVICE__ -inline double __dsqrt_rz(double __x) { return __ocml_sqrt_rtz_f64(__x); } +double __dsqrt_rz(double __x) { return __ocml_sqrt_rtz_f64(__x); } +#else +__DEVICE__ +double __dsqrt_rn(double __x) { return __ocml_sqrt_f64(__x); } +#endif + +#if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline double __dsub_rd(double __x, double __y) { +double __dsub_rd(double __x, double __y) { return __ocml_sub_rtn_f64(__x, __y); } -#endif __DEVICE__ -inline double __dsub_rn(double __x, double __y) { return __x - __y; } -#if defined OCML_BASIC_ROUNDED_OPERATIONS +double __dsub_rn(double __x, double __y) { + return __ocml_sub_rte_f64(__x, __y); +} __DEVICE__ -inline double __dsub_ru(double __x, double __y) { +double __dsub_ru(double __x, double __y) { return __ocml_sub_rtp_f64(__x, __y); } __DEVICE__ -inline double __dsub_rz(double __x, double __y) { +double __dsub_rz(double __x, double __y) { return __ocml_sub_rtz_f64(__x, __y); } +#else +__DEVICE__ +double __dsub_rn(double __x, double __y) { return __x - __y; } +#endif + +#if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline double __fma_rd(double __x, double __y, double __z) { +double __fma_rd(double __x, double __y, double __z) { return __ocml_fma_rtn_f64(__x, __y, __z); } -#endif __DEVICE__ -inline double __fma_rn(double __x, double __y, double __z) { - return __ocml_fma_f64(__x, __y, __z); +double __fma_rn(double __x, double __y, double __z) { + return __ocml_fma_rte_f64(__x, __y, __z); } -#if defined OCML_BASIC_ROUNDED_OPERATIONS __DEVICE__ -inline double __fma_ru(double __x, double __y, double __z) { +double __fma_ru(double __x, double __y, double __z) { return __ocml_fma_rtp_f64(__x, __y, __z); } __DEVICE__ -inline double __fma_rz(double __x, double __y, double __z) { +double __fma_rz(double __x, double __y, double __z) { return __ocml_fma_rtz_f64(__x, __y, __z); } +#else +__DEVICE__ +double __fma_rn(double __x, double __y, double __z) { + return __ocml_fma_f64(__x, __y, __z); +} #endif // END INTRINSICS // END DOUBLE -// BEGIN INTEGER -__DEVICE__ -inline int abs(int __x) { - int __sgn = __x >> (sizeof(int) * CHAR_BIT - 1); - return (__x ^ __sgn) - __sgn; -} -__DEVICE__ -inline long labs(long __x) { - long __sgn = __x >> (sizeof(long) * CHAR_BIT - 1); - return (__x ^ __sgn) - __sgn; -} -__DEVICE__ -inline long long llabs(long long __x) { - long long __sgn = __x >> (sizeof(long long) * CHAR_BIT - 1); - return (__x ^ __sgn) - __sgn; -} +// C only macros +#if !defined(__cplusplus) && __STDC_VERSION__ >= 201112L +#define isfinite(__x) _Generic((__x), float : __finitef, double : __finite)(__x) +#define isinf(__x) _Generic((__x), float : __isinff, double : __isinf)(__x) +#define isnan(__x) _Generic((__x), float : __isnanf, double : __isnan)(__x) +#define signbit(__x) \ + _Generic((__x), float : __signbitf, double : __signbit)(__x) +#endif // !defined(__cplusplus) && __STDC_VERSION__ >= 201112L #if defined(__cplusplus) -__DEVICE__ -inline long abs(long __x) { return labs(__x); } -__DEVICE__ -inline long long abs(long long __x) { return llabs(__x); } -#endif -// END INTEGER - -__DEVICE__ -inline _Float16 fma(_Float16 __x, _Float16 __y, _Float16 __z) { - return __ocml_fma_f16(__x, __y, __z); -} - -__DEVICE__ -inline float fma(float __x, float __y, float __z) { - return fmaf(__x, __y, __z); -} - -#pragma push_macro("__DEF_FUN1") -#pragma push_macro("__DEF_FUN2") -#pragma push_macro("__DEF_FUNI") -#pragma push_macro("__DEF_FLOAT_FUN2I") -#pragma push_macro("__HIP_OVERLOAD1") -#pragma push_macro("__HIP_OVERLOAD2") - -// __hip_enable_if::type is a type function which returns __T if __B is true. -template <bool __B, class __T = void> struct __hip_enable_if {}; - -template <class __T> struct __hip_enable_if<true, __T> { typedef __T type; }; - -// __HIP_OVERLOAD1 is used to resolve function calls with integer argument to -// avoid compilation error due to ambibuity. e.g. floor(5) is resolved with -// floor(double). -#define __HIP_OVERLOAD1(__retty, __fn) \ - template <typename __T> \ - __DEVICE__ typename __hip_enable_if<std::numeric_limits<__T>::is_integer, \ - __retty>::type \ - __fn(__T __x) { \ - return ::__fn((double)__x); \ - } - -// __HIP_OVERLOAD2 is used to resolve function calls with mixed float/double -// or integer argument to avoid compilation error due to ambibuity. e.g. -// max(5.0f, 6.0) is resolved with max(double, double). -#define __HIP_OVERLOAD2(__retty, __fn) \ - template <typename __T1, typename __T2> \ - __DEVICE__ \ - typename __hip_enable_if<std::numeric_limits<__T1>::is_specialized && \ - std::numeric_limits<__T2>::is_specialized, \ - __retty>::type \ - __fn(__T1 __x, __T2 __y) { \ - return __fn((double)__x, (double)__y); \ - } - -// Define cmath functions with float argument and returns float. -#define __DEF_FUN1(__retty, __func) \ - __DEVICE__ \ - inline float __func(float __x) { return __func##f(__x); } \ - __HIP_OVERLOAD1(__retty, __func) - -// Define cmath functions with float argument and returns __retty. -#define __DEF_FUNI(__retty, __func) \ - __DEVICE__ \ - inline __retty __func(float __x) { return __func##f(__x); } \ - __HIP_OVERLOAD1(__retty, __func) - -// define cmath functions with two float arguments. -#define __DEF_FUN2(__retty, __func) \ - __DEVICE__ \ - inline float __func(float __x, float __y) { return __func##f(__x, __y); } \ - __HIP_OVERLOAD2(__retty, __func) - -__DEF_FUN1(double, acos) -__DEF_FUN1(double, acosh) -__DEF_FUN1(double, asin) -__DEF_FUN1(double, asinh) -__DEF_FUN1(double, atan) -__DEF_FUN2(double, atan2); -__DEF_FUN1(double, atanh) -__DEF_FUN1(double, cbrt) -__DEF_FUN1(double, ceil) -__DEF_FUN2(double, copysign); -__DEF_FUN1(double, cos) -__DEF_FUN1(double, cosh) -__DEF_FUN1(double, erf) -__DEF_FUN1(double, erfc) -__DEF_FUN1(double, exp) -__DEF_FUN1(double, exp2) -__DEF_FUN1(double, expm1) -__DEF_FUN1(double, fabs) -__DEF_FUN2(double, fdim); -__DEF_FUN1(double, floor) -__DEF_FUN2(double, fmax); -__DEF_FUN2(double, fmin); -__DEF_FUN2(double, fmod); -//__HIP_OVERLOAD1(int, fpclassify) -__DEF_FUN2(double, hypot); -__DEF_FUNI(int, ilogb) -__HIP_OVERLOAD1(bool, isfinite) -__HIP_OVERLOAD2(bool, isgreater); -__HIP_OVERLOAD2(bool, isgreaterequal); -__HIP_OVERLOAD1(bool, isinf); -__HIP_OVERLOAD2(bool, isless); -__HIP_OVERLOAD2(bool, islessequal); -__HIP_OVERLOAD2(bool, islessgreater); -__HIP_OVERLOAD1(bool, isnan); -//__HIP_OVERLOAD1(bool, isnormal) -__HIP_OVERLOAD2(bool, isunordered); -__DEF_FUN1(double, lgamma) -__DEF_FUN1(double, log) -__DEF_FUN1(double, log10) -__DEF_FUN1(double, log1p) -__DEF_FUN1(double, log2) -__DEF_FUN1(double, logb) -__DEF_FUNI(long long, llrint) -__DEF_FUNI(long long, llround) -__DEF_FUNI(long, lrint) -__DEF_FUNI(long, lround) -__DEF_FUN1(double, nearbyint); -__DEF_FUN2(double, nextafter); -__DEF_FUN2(double, pow); -__DEF_FUN2(double, remainder); -__DEF_FUN1(double, rint); -__DEF_FUN1(double, round); -__HIP_OVERLOAD1(bool, signbit) -__DEF_FUN1(double, sin) -__DEF_FUN1(double, sinh) -__DEF_FUN1(double, sqrt) -__DEF_FUN1(double, tan) -__DEF_FUN1(double, tanh) -__DEF_FUN1(double, tgamma) -__DEF_FUN1(double, trunc); - -// define cmath functions with a float and an integer argument. -#define __DEF_FLOAT_FUN2I(__func) \ - __DEVICE__ \ - inline float __func(float __x, int __y) { return __func##f(__x, __y); } -__DEF_FLOAT_FUN2I(scalbn) - -template <class T> __DEVICE__ inline T min(T __arg1, T __arg2) { +template <class T> __DEVICE__ T min(T __arg1, T __arg2) { return (__arg1 < __arg2) ? __arg1 : __arg2; } -template <class T> __DEVICE__ inline T max(T __arg1, T __arg2) { +template <class T> __DEVICE__ T max(T __arg1, T __arg2) { return (__arg1 > __arg2) ? __arg1 : __arg2; } -__DEVICE__ inline int min(int __arg1, int __arg2) { +__DEVICE__ int min(int __arg1, int __arg2) { return (__arg1 < __arg2) ? __arg1 : __arg2; } -__DEVICE__ inline int max(int __arg1, int __arg2) { +__DEVICE__ int max(int __arg1, int __arg2) { return (__arg1 > __arg2) ? __arg1 : __arg2; } __DEVICE__ -inline float max(float __x, float __y) { return fmaxf(__x, __y); } +float max(float __x, float __y) { return fmaxf(__x, __y); } __DEVICE__ -inline double max(double __x, double __y) { return fmax(__x, __y); } +double max(double __x, double __y) { return fmax(__x, __y); } __DEVICE__ -inline float min(float __x, float __y) { return fminf(__x, __y); } +float min(float __x, float __y) { return fminf(__x, __y); } __DEVICE__ -inline double min(double __x, double __y) { return fmin(__x, __y); } - -__HIP_OVERLOAD2(double, max) -__HIP_OVERLOAD2(double, min) +double min(double __x, double __y) { return fmin(__x, __y); } __host__ inline static int min(int __arg1, int __arg2) { return std::min(__arg1, __arg2); @@ -1172,13 +1267,8 @@ __host__ inline static int min(int __arg1, int __arg2) { __host__ inline static int max(int __arg1, int __arg2) { return std::max(__arg1, __arg2); } +#endif -#pragma pop_macro("__DEF_FUN1") -#pragma pop_macro("__DEF_FUN2") -#pragma pop_macro("__DEF_FUNI") -#pragma pop_macro("__DEF_FLOAT_FUN2I") -#pragma pop_macro("__HIP_OVERLOAD1") -#pragma pop_macro("__HIP_OVERLOAD2") #pragma pop_macro("__DEVICE__") #pragma pop_macro("__RETURN_TYPE") diff --git a/lib/include/__clang_hip_runtime_wrapper.h b/lib/include/__clang_hip_runtime_wrapper.h index addae5605a..81a16a265a 100644 --- a/lib/include/__clang_hip_runtime_wrapper.h +++ b/lib/include/__clang_hip_runtime_wrapper.h @@ -28,6 +28,10 @@ #define __shared__ __attribute__((shared)) #define __constant__ __attribute__((constant)) +#if !defined(__cplusplus) || __cplusplus < 201103L + #define nullptr NULL; +#endif + #if __HIP_ENABLE_DEVICE_MALLOC__ extern "C" __device__ void *__hip_malloc(size_t __size); extern "C" __device__ void *__hip_free(void *__ptr); @@ -51,6 +55,7 @@ static inline __device__ void *free(void *__ptr) { #if !_OPENMP || __HIP_ENABLE_CUDA_WRAPPER_FOR_OPENMP__ #include <__clang_cuda_math_forward_declares.h> +#include <__clang_hip_cmath.h> #include <__clang_cuda_complex_builtins.h> #include <algorithm> diff --git a/lib/include/altivec.h b/lib/include/altivec.h index ac5f438363..4d50d47d51 100644 --- a/lib/include/altivec.h +++ b/lib/include/altivec.h @@ -1709,6 +1709,20 @@ vec_cmpeq(vector double __a, vector double __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpeq(vector signed __int128 __a, vector signed __int128 __b) { + return (vector bool __int128)__builtin_altivec_vcmpequq( + (vector bool __int128)__a, (vector bool __int128)__b); +} + +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpeq(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return (vector bool __int128)__builtin_altivec_vcmpequq( + (vector bool __int128)__a, (vector bool __int128)__b); +} +#endif + #ifdef __POWER9_VECTOR__ /* vec_cmpne */ @@ -1766,35 +1780,25 @@ vec_cmpne(vector unsigned int __a, vector unsigned int __b) { (vector int)__b); } -static __inline__ vector bool long long __ATTRS_o_ai -vec_cmpne(vector bool long long __a, vector bool long long __b) { - return (vector bool long long) - ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); -} - -static __inline__ vector bool long long __ATTRS_o_ai -vec_cmpne(vector signed long long __a, vector signed long long __b) { - return (vector bool long long) - ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); -} - -static __inline__ vector bool long long __ATTRS_o_ai -vec_cmpne(vector unsigned long long __a, vector unsigned long long __b) { - return (vector bool long long) - ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); -} - static __inline__ vector bool int __ATTRS_o_ai vec_cmpne(vector float __a, vector float __b) { return (vector bool int)__builtin_altivec_vcmpnew((vector int)__a, (vector int)__b); } -static __inline__ vector bool long long __ATTRS_o_ai -vec_cmpne(vector double __a, vector double __b) { - return (vector bool long long) - ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); +#ifdef __POWER10_VECTOR__ +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpne(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return (vector bool __int128) ~(__builtin_altivec_vcmpequq( + (vector bool __int128)__a, (vector bool __int128)__b)); +} + +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpne(vector signed __int128 __a, vector signed __int128 __b) { + return (vector bool __int128) ~(__builtin_altivec_vcmpequq( + (vector bool __int128)__a, (vector bool __int128)__b)); } +#endif /* vec_cmpnez */ @@ -1900,6 +1904,86 @@ vec_parity_lsbb(vector signed long long __a) { return __builtin_altivec_vprtybd(__a); } +#else +/* vec_cmpne */ + +static __inline__ vector bool char __ATTRS_o_ai +vec_cmpne(vector bool char __a, vector bool char __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool char __ATTRS_o_ai +vec_cmpne(vector signed char __a, vector signed char __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool char __ATTRS_o_ai +vec_cmpne(vector unsigned char __a, vector unsigned char __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool short __ATTRS_o_ai +vec_cmpne(vector bool short __a, vector bool short __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool short __ATTRS_o_ai +vec_cmpne(vector signed short __a, vector signed short __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool short __ATTRS_o_ai +vec_cmpne(vector unsigned short __a, vector unsigned short __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool int __ATTRS_o_ai +vec_cmpne(vector bool int __a, vector bool int __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool int __ATTRS_o_ai +vec_cmpne(vector signed int __a, vector signed int __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool int __ATTRS_o_ai +vec_cmpne(vector unsigned int __a, vector unsigned int __b) { + return ~(vec_cmpeq(__a, __b)); +} + +static __inline__ vector bool int __ATTRS_o_ai +vec_cmpne(vector float __a, vector float __b) { + return ~(vec_cmpeq(__a, __b)); +} +#endif + +#ifdef __POWER8_VECTOR__ +static __inline__ vector bool long long __ATTRS_o_ai +vec_cmpne(vector bool long long __a, vector bool long long __b) { + return (vector bool long long) + ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); +} + +static __inline__ vector bool long long __ATTRS_o_ai +vec_cmpne(vector signed long long __a, vector signed long long __b) { + return (vector bool long long) + ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); +} + +static __inline__ vector bool long long __ATTRS_o_ai +vec_cmpne(vector unsigned long long __a, vector unsigned long long __b) { + return (vector bool long long) + ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); +} +#endif + +#ifdef __VSX__ +static __inline__ vector bool long long __ATTRS_o_ai +vec_cmpne(vector double __a, vector double __b) { + return (vector bool long long) + ~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b)); +} #endif /* vec_cmpgt */ @@ -1962,6 +2046,20 @@ vec_cmpgt(vector double __a, vector double __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpgt(vector signed __int128 __a, vector signed __int128 __b) { + return (vector bool __int128)__builtin_altivec_vcmpgtsq( + (vector bool __int128)__a, (vector bool __int128)__b); +} + +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpgt(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return (vector bool __int128)__builtin_altivec_vcmpgtuq( + (vector bool __int128)__a, (vector bool __int128)__b); +} +#endif + /* vec_cmpge */ static __inline__ vector bool char __ATTRS_o_ai @@ -2022,6 +2120,18 @@ vec_cmpge(vector unsigned long long __a, vector unsigned long long __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpge(vector signed __int128 __a, vector signed __int128 __b) { + return ~(vec_cmpgt(__b, __a)); +} + +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmpge(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return ~(vec_cmpgt(__b, __a)); +} +#endif + /* vec_vcmpgefp */ static __inline__ vector bool int __attribute__((__always_inline__)) @@ -2134,6 +2244,18 @@ vec_cmple(vector unsigned long long __a, vector unsigned long long __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmple(vector signed __int128 __a, vector signed __int128 __b) { + return vec_cmpge(__b, __a); +} + +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmple(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return vec_cmpge(__b, __a); +} +#endif + /* vec_cmplt */ static __inline__ vector bool char __ATTRS_o_ai @@ -2178,6 +2300,18 @@ vec_cmplt(vector double __a, vector double __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmplt(vector signed __int128 __a, vector signed __int128 __b) { + return vec_cmpgt(__b, __a); +} + +static __inline__ vector bool __int128 __ATTRS_o_ai +vec_cmplt(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return vec_cmpgt(__b, __a); +} +#endif + #ifdef __POWER8_VECTOR__ static __inline__ vector bool long long __ATTRS_o_ai vec_cmplt(vector signed long long __a, vector signed long long __b) { @@ -2702,67 +2836,67 @@ vec_insert_exp(vector unsigned int __a, vector unsigned int __b) { } #if defined(__powerpc64__) -static __inline__ vector signed char __ATTRS_o_ai vec_xl_len(signed char *__a, +static __inline__ vector signed char __ATTRS_o_ai vec_xl_len(const signed char *__a, size_t __b) { return (vector signed char)__builtin_vsx_lxvl(__a, (__b << 56)); } static __inline__ vector unsigned char __ATTRS_o_ai -vec_xl_len(unsigned char *__a, size_t __b) { +vec_xl_len(const unsigned char *__a, size_t __b) { return (vector unsigned char)__builtin_vsx_lxvl(__a, (__b << 56)); } -static __inline__ vector signed short __ATTRS_o_ai vec_xl_len(signed short *__a, +static __inline__ vector signed short __ATTRS_o_ai vec_xl_len(const signed short *__a, size_t __b) { return (vector signed short)__builtin_vsx_lxvl(__a, (__b << 56)); } static __inline__ vector unsigned short __ATTRS_o_ai -vec_xl_len(unsigned short *__a, size_t __b) { +vec_xl_len(const unsigned short *__a, size_t __b) { return (vector unsigned short)__builtin_vsx_lxvl(__a, (__b << 56)); } -static __inline__ vector signed int __ATTRS_o_ai vec_xl_len(signed int *__a, +static __inline__ vector signed int __ATTRS_o_ai vec_xl_len(const signed int *__a, size_t __b) { return (vector signed int)__builtin_vsx_lxvl(__a, (__b << 56)); } -static __inline__ vector unsigned int __ATTRS_o_ai vec_xl_len(unsigned int *__a, +static __inline__ vector unsigned int __ATTRS_o_ai vec_xl_len(const unsigned int *__a, size_t __b) { return (vector unsigned int)__builtin_vsx_lxvl(__a, (__b << 56)); } -static __inline__ vector float __ATTRS_o_ai vec_xl_len(float *__a, size_t __b) { +static __inline__ vector float __ATTRS_o_ai vec_xl_len(const float *__a, size_t __b) { return (vector float)__builtin_vsx_lxvl(__a, (__b << 56)); } static __inline__ vector signed __int128 __ATTRS_o_ai -vec_xl_len(signed __int128 *__a, size_t __b) { +vec_xl_len(const signed __int128 *__a, size_t __b) { return (vector signed __int128)__builtin_vsx_lxvl(__a, (__b << 56)); } static __inline__ vector unsigned __int128 __ATTRS_o_ai -vec_xl_len(unsigned __int128 *__a, size_t __b) { +vec_xl_len(const unsigned __int128 *__a, size_t __b) { return (vector unsigned __int128)__builtin_vsx_lxvl(__a, (__b << 56)); } static __inline__ vector signed long long __ATTRS_o_ai -vec_xl_len(signed long long *__a, size_t __b) { +vec_xl_len(const signed long long *__a, size_t __b) { return (vector signed long long)__builtin_vsx_lxvl(__a, (__b << 56)); } static __inline__ vector unsigned long long __ATTRS_o_ai -vec_xl_len(unsigned long long *__a, size_t __b) { +vec_xl_len(const unsigned long long *__a, size_t __b) { return (vector unsigned long long)__builtin_vsx_lxvl(__a, (__b << 56)); } -static __inline__ vector double __ATTRS_o_ai vec_xl_len(double *__a, +static __inline__ vector double __ATTRS_o_ai vec_xl_len(const double *__a, size_t __b) { return (vector double)__builtin_vsx_lxvl(__a, (__b << 56)); } static __inline__ vector unsigned char __ATTRS_o_ai -vec_xl_len_r(unsigned char *__a, size_t __b) { +vec_xl_len_r(const unsigned char *__a, size_t __b) { vector unsigned char __res = (vector unsigned char)__builtin_vsx_lxvll(__a, (__b << 56)); #ifdef __LITTLE_ENDIAN__ @@ -2862,12 +2996,12 @@ static __inline__ void __ATTRS_o_ai vec_xst_len_r(vector unsigned char __a, #ifdef __VSX__ static __inline__ vector float __ATTRS_o_ai vec_cpsgn(vector float __a, vector float __b) { - return __builtin_vsx_xvcpsgnsp(__a, __b); + return __builtin_vsx_xvcpsgnsp(__b, __a); } static __inline__ vector double __ATTRS_o_ai vec_cpsgn(vector double __a, vector double __b) { - return __builtin_vsx_xvcpsgndp(__a, __b); + return __builtin_vsx_xvcpsgndp(__b, __a); } #endif @@ -2951,6 +3085,42 @@ static __inline__ vector double __ATTRS_o_ai vec_cpsgn(vector double __a, #define vec_vctuxs __builtin_altivec_vctuxs +/* vec_signext */ + +#ifdef __POWER9_VECTOR__ +static __inline__ vector signed int __ATTRS_o_ai +vec_signexti(vector signed char __a) { + return __builtin_altivec_vextsb2w(__a); +} + +static __inline__ vector signed int __ATTRS_o_ai +vec_signexti(vector signed short __a) { + return __builtin_altivec_vextsh2w(__a); +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_signextll(vector signed char __a) { + return __builtin_altivec_vextsb2d(__a); +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_signextll(vector signed short __a) { + return __builtin_altivec_vextsh2d(__a); +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_signextll(vector signed int __a) { + return __builtin_altivec_vextsw2d(__a); +} +#endif + +#ifdef __POWER10_VECTOR__ +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_signextq(vector signed long long __a) { + return __builtin_altivec_vextsd2q(__a); +} +#endif + /* vec_signed */ static __inline__ vector signed int __ATTRS_o_ai @@ -3288,6 +3458,66 @@ static __inline__ vector double __ATTRS_o_ai vec_div(vector double __a, } #endif +/* vec_dive */ + +#ifdef __POWER10_VECTOR__ +static __inline__ vector signed int __ATTRS_o_ai +vec_dive(vector signed int __a, vector signed int __b) { + return __builtin_altivec_vdivesw(__a, __b); +} + +static __inline__ vector unsigned int __ATTRS_o_ai +vec_dive(vector unsigned int __a, vector unsigned int __b) { + return __builtin_altivec_vdiveuw(__a, __b); +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_dive(vector signed long long __a, vector signed long long __b) { + return __builtin_altivec_vdivesd(__a, __b); +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_dive(vector unsigned long long __a, vector unsigned long long __b) { + return __builtin_altivec_vdiveud(__a, __b); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_dive(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return __builtin_altivec_vdiveuq(__a, __b); +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_dive(vector signed __int128 __a, vector signed __int128 __b) { + return __builtin_altivec_vdivesq(__a, __b); +} +#endif + +#ifdef __POWER10_VECTOR__ +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_div(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return __a / __b; +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_div(vector signed __int128 __a, vector signed __int128 __b) { + return __a / __b; +} +#endif /* __POWER10_VECTOR__ */ + +/* vec_xvtdiv */ + +#ifdef __VSX__ +static __inline__ int __ATTRS_o_ai vec_test_swdiv(vector double __a, + vector double __b) { + return __builtin_vsx_xvtdivdp(__a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_test_swdivs(vector float __a, + vector float __b) { + return __builtin_vsx_xvtdivsp(__a, __b); +} +#endif + /* vec_dss */ #define vec_dss __builtin_altivec_dss @@ -3300,23 +3530,19 @@ static __inline__ void __attribute__((__always_inline__)) vec_dssall(void) { /* vec_dst */ #define vec_dst(__PTR, __CW, __STR) \ - __extension__( \ - { __builtin_altivec_dst((const void *)(__PTR), (__CW), (__STR)); }) + __builtin_altivec_dst((const void *)(__PTR), (__CW), (__STR)) /* vec_dstst */ #define vec_dstst(__PTR, __CW, __STR) \ - __extension__( \ - { __builtin_altivec_dstst((const void *)(__PTR), (__CW), (__STR)); }) + __builtin_altivec_dstst((const void *)(__PTR), (__CW), (__STR)) /* vec_dststt */ #define vec_dststt(__PTR, __CW, __STR) \ - __extension__( \ - { __builtin_altivec_dststt((const void *)(__PTR), (__CW), (__STR)); }) + __builtin_altivec_dststt((const void *)(__PTR), (__CW), (__STR)) /* vec_dstt */ #define vec_dstt(__PTR, __CW, __STR) \ - __extension__( \ - { __builtin_altivec_dstt((const void *)(__PTR), (__CW), (__STR)); }) + __builtin_altivec_dstt((const void *)(__PTR), (__CW), (__STR)) /* vec_eqv */ @@ -5467,6 +5693,16 @@ vec_msum(vector unsigned short __a, vector unsigned short __b, return __builtin_altivec_vmsumuhm(__a, __b, __c); } +/* vec_msumc */ + +#ifdef __POWER10_VECTOR__ +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_msumc(vector unsigned long long __a, vector unsigned long long __b, + vector unsigned __int128 __c) { + return __builtin_altivec_vmsumcud(__a, __b, __c); +} +#endif + /* vec_vmsummbm */ static __inline__ vector int __attribute__((__always_inline__)) @@ -5693,6 +5929,26 @@ vec_mule(vector unsigned int __a, vector unsigned int __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_mule(vector signed long long __a, vector signed long long __b) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vmulosd(__a, __b); +#else + return __builtin_altivec_vmulesd(__a, __b); +#endif +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_mule(vector unsigned long long __a, vector unsigned long long __b) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vmuloud(__a, __b); +#else + return __builtin_altivec_vmuleud(__a, __b); +#endif +} +#endif + /* vec_vmulesb */ static __inline__ vector short __attribute__((__always_inline__)) @@ -5737,6 +5993,30 @@ vec_vmuleuh(vector unsigned short __a, vector unsigned short __b) { #endif } +/* vec_mulh */ + +#ifdef __POWER10_VECTOR__ +static __inline__ vector signed int __ATTRS_o_ai +vec_mulh(vector signed int __a, vector signed int __b) { + return __builtin_altivec_vmulhsw(__a, __b); +} + +static __inline__ vector unsigned int __ATTRS_o_ai +vec_mulh(vector unsigned int __a, vector unsigned int __b) { + return __builtin_altivec_vmulhuw(__a, __b); +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_mulh(vector signed long long __a, vector signed long long __b) { + return __builtin_altivec_vmulhsd(__a, __b); +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_mulh(vector unsigned long long __a, vector unsigned long long __b) { + return __builtin_altivec_vmulhud(__a, __b); +} +#endif + /* vec_mulo */ static __inline__ vector short __ATTRS_o_ai vec_mulo(vector signed char __a, @@ -5795,6 +6075,26 @@ vec_mulo(vector unsigned int __a, vector unsigned int __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_mulo(vector signed long long __a, vector signed long long __b) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vmulesd(__a, __b); +#else + return __builtin_altivec_vmulosd(__a, __b); +#endif +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_mulo(vector unsigned long long __a, vector unsigned long long __b) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vmuleud(__a, __b); +#else + return __builtin_altivec_vmuloud(__a, __b); +#endif +} +#endif + /* vec_vmulosb */ static __inline__ vector short __attribute__((__always_inline__)) @@ -7627,6 +7927,18 @@ vec_rl(vector unsigned long long __a, vector unsigned long long __b) { } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_rl(vector signed __int128 __a, vector unsigned __int128 __b) { + return (__b << __a)|(__b >> ((__CHAR_BIT__ * sizeof(vector signed __int128)) - __a)); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_rl(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return (__b << __a)|(__b >> ((__CHAR_BIT__ * sizeof(vector unsigned __int128)) - __a)); +} +#endif + /* vec_rlmi */ #ifdef __POWER9_VECTOR__ static __inline__ vector unsigned int __ATTRS_o_ai @@ -7640,8 +7952,24 @@ vec_rlmi(vector unsigned long long __a, vector unsigned long long __b, vector unsigned long long __c) { return __builtin_altivec_vrldmi(__a, __c, __b); } +#endif + +#ifdef __POWER10_VECTOR__ +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_rlmi(vector unsigned __int128 __a, vector unsigned __int128 __b, + vector unsigned __int128 __c) { + return __builtin_altivec_vrlqmi(__a, __c, __b); +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_rlmi(vector signed __int128 __a, vector signed __int128 __b, + vector signed __int128 __c) { + return __builtin_altivec_vrlqmi(__a, __c, __b); +} +#endif /* vec_rlnm */ +#ifdef __POWER9_VECTOR__ static __inline__ vector unsigned int __ATTRS_o_ai vec_rlnm(vector unsigned int __a, vector unsigned int __b, vector unsigned int __c) { @@ -7657,6 +7985,42 @@ vec_rlnm(vector unsigned long long __a, vector unsigned long long __b, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_rlnm(vector unsigned __int128 __a, vector unsigned __int128 __b, + vector unsigned __int128 __c) { + // Merge __b and __c using an appropriate shuffle. + vector unsigned char TmpB = (vector unsigned char)__b; + vector unsigned char TmpC = (vector unsigned char)__c; + vector unsigned char MaskAndShift = +#ifdef __LITTLE_ENDIAN__ + __builtin_shufflevector(TmpB, TmpC, -1, -1, -1, -1, -1, -1, -1, -1, 16, 0, + 1, -1, -1, -1, -1, -1); +#else + __builtin_shufflevector(TmpB, TmpC, -1, -1, -1, -1, -1, 31, 30, 15, -1, + -1, -1, -1, -1, -1, -1, -1); +#endif + return __builtin_altivec_vrlqnm(__a, (vector unsigned __int128) MaskAndShift); +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_rlnm(vector signed __int128 __a, vector signed __int128 __b, + vector signed __int128 __c) { + // Merge __b and __c using an appropriate shuffle. + vector unsigned char TmpB = (vector unsigned char)__b; + vector unsigned char TmpC = (vector unsigned char)__c; + vector unsigned char MaskAndShift = +#ifdef __LITTLE_ENDIAN__ + __builtin_shufflevector(TmpB, TmpC, -1, -1, -1, -1, -1, -1, -1, -1, 16, 0, + 1, -1, -1, -1, -1, -1); +#else + __builtin_shufflevector(TmpB, TmpC, -1, -1, -1, -1, -1, 31, 30, 15, -1, + -1, -1, -1, -1, -1, -1, -1); +#endif + return __builtin_altivec_vrlqnm(__a, (vector unsigned __int128) MaskAndShift); +} +#endif + /* vec_vrlb */ static __inline__ vector signed char __ATTRS_o_ai @@ -7771,6 +8135,18 @@ vec_vrsqrtefp(vector float __a) { return __builtin_altivec_vrsqrtefp(__a); } +/* vec_xvtsqrt */ + +#ifdef __VSX__ +static __inline__ int __ATTRS_o_ai vec_test_swsqrt(vector double __a) { + return __builtin_vsx_xvtsqrtdp(__a); +} + +static __inline__ int __ATTRS_o_ai vec_test_swsqrts(vector float __a) { + return __builtin_vsx_xvtsqrtsp(__a); +} +#endif + /* vec_sel */ #define __builtin_altivec_vsel_4si vec_sel @@ -7905,6 +8281,46 @@ vec_sel(vector double __a, vector double __b, vector unsigned long long __c) { ((vector long long)__b & (vector long long)__c); return (vector double)__res; } + +static __inline__ vector bool long long __ATTRS_o_ai +vec_sel(vector bool long long __a, vector bool long long __b, + vector bool long long __c) { + return (__a & ~__c) | (__b & __c); +} + +static __inline__ vector bool long long __ATTRS_o_ai +vec_sel(vector bool long long __a, vector bool long long __b, + vector unsigned long long __c) { + return (__a & ~(vector bool long long)__c) | + (__b & (vector bool long long)__c); +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_sel(vector signed long long __a, vector signed long long __b, + vector bool long long __c) { + return (__a & ~(vector signed long long)__c) | + (__b & (vector signed long long)__c); +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_sel(vector signed long long __a, vector signed long long __b, + vector unsigned long long __c) { + return (__a & ~(vector signed long long)__c) | + (__b & (vector signed long long)__c); +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_sel(vector unsigned long long __a, vector unsigned long long __b, + vector bool long long __c) { + return (__a & ~(vector unsigned long long)__c) | + (__b & (vector unsigned long long)__c); +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_sel(vector unsigned long long __a, vector unsigned long long __b, + vector unsigned long long __c) { + return (__a & ~__c) | (__b & __c); +} #endif /* vec_vsel */ @@ -13900,6 +14316,18 @@ static __inline__ int __ATTRS_o_ai vec_all_eq(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_all_eq(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_LT, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_all_eq(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_LT, __a, __b); +} +#endif + /* vec_all_ge */ static __inline__ int __ATTRS_o_ai vec_all_ge(vector signed char __a, @@ -14071,6 +14499,18 @@ static __inline__ int __ATTRS_o_ai vec_all_ge(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_all_ge(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_EQ, __b, __a); +} + +static __inline__ int __ATTRS_o_ai vec_all_ge(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_EQ, __b, __a); +} +#endif + /* vec_all_gt */ static __inline__ int __ATTRS_o_ai vec_all_gt(vector signed char __a, @@ -14242,6 +14682,18 @@ static __inline__ int __ATTRS_o_ai vec_all_gt(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_all_gt(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_LT, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_all_gt(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_LT, __a, __b); +} +#endif + /* vec_all_in */ static __inline__ int __attribute__((__always_inline__)) @@ -14421,6 +14873,18 @@ static __inline__ int __ATTRS_o_ai vec_all_le(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_all_le(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_EQ, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_all_le(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_EQ, __a, __b); +} +#endif + /* vec_all_lt */ static __inline__ int __ATTRS_o_ai vec_all_lt(vector signed char __a, @@ -14593,6 +15057,18 @@ static __inline__ int __ATTRS_o_ai vec_all_lt(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_all_lt(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_LT, __b, __a); +} + +static __inline__ int __ATTRS_o_ai vec_all_lt(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_LT, __b, __a); +} +#endif + /* vec_all_nan */ static __inline__ int __ATTRS_o_ai vec_all_nan(vector float __a) { @@ -14797,6 +15273,18 @@ static __inline__ int __ATTRS_o_ai vec_all_ne(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_all_ne(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_EQ, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_all_ne(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_EQ, __a, __b); +} +#endif + /* vec_all_nge */ static __inline__ int __ATTRS_o_ai vec_all_nge(vector float __a, @@ -15042,6 +15530,18 @@ static __inline__ int __ATTRS_o_ai vec_any_eq(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_any_eq(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_EQ_REV, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_any_eq(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_EQ_REV, __a, __b); +} +#endif + /* vec_any_ge */ static __inline__ int __ATTRS_o_ai vec_any_ge(vector signed char __a, @@ -15221,6 +15721,18 @@ static __inline__ int __ATTRS_o_ai vec_any_ge(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_any_ge(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_LT_REV, __b, __a); +} + +static __inline__ int __ATTRS_o_ai vec_any_ge(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_LT_REV, __b, __a); +} +#endif + /* vec_any_gt */ static __inline__ int __ATTRS_o_ai vec_any_gt(vector signed char __a, @@ -15400,6 +15912,18 @@ static __inline__ int __ATTRS_o_ai vec_any_gt(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_any_gt(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_EQ_REV, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_any_gt(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_EQ_REV, __a, __b); +} +#endif + /* vec_any_le */ static __inline__ int __ATTRS_o_ai vec_any_le(vector signed char __a, @@ -15579,6 +16103,18 @@ static __inline__ int __ATTRS_o_ai vec_any_le(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_any_le(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_LT_REV, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_any_le(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_LT_REV, __a, __b); +} +#endif + /* vec_any_lt */ static __inline__ int __ATTRS_o_ai vec_any_lt(vector signed char __a, @@ -15758,6 +16294,18 @@ static __inline__ int __ATTRS_o_ai vec_any_lt(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_any_lt(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpgtsq_p(__CR6_EQ_REV, __b, __a); +} + +static __inline__ int __ATTRS_o_ai vec_any_lt(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpgtuq_p(__CR6_EQ_REV, __b, __a); +} +#endif + /* vec_any_nan */ static __inline__ int __attribute__((__always_inline__)) @@ -15953,6 +16501,18 @@ static __inline__ int __ATTRS_o_ai vec_any_ne(vector double __a, } #endif +#ifdef __POWER10_VECTOR__ +static __inline__ int __ATTRS_o_ai vec_any_ne(vector signed __int128 __a, + vector signed __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_LT_REV, __a, __b); +} + +static __inline__ int __ATTRS_o_ai vec_any_ne(vector unsigned __int128 __a, + vector unsigned __int128 __b) { + return __builtin_altivec_vcmpequq_p(__CR6_LT_REV, __a, __b); +} +#endif + /* vec_any_nge */ static __inline__ int __attribute__((__always_inline__)) @@ -16353,41 +16913,41 @@ typedef vector unsigned int unaligned_vec_uint __attribute__((aligned(1))); typedef vector float unaligned_vec_float __attribute__((aligned(1))); static inline __ATTRS_o_ai vector signed char vec_xl(signed long long __offset, - signed char *__ptr) { + const signed char *__ptr) { return *(unaligned_vec_schar *)(__ptr + __offset); } static inline __ATTRS_o_ai vector unsigned char -vec_xl(signed long long __offset, unsigned char *__ptr) { +vec_xl(signed long long __offset, const unsigned char *__ptr) { return *(unaligned_vec_uchar*)(__ptr + __offset); } static inline __ATTRS_o_ai vector signed short vec_xl(signed long long __offset, - signed short *__ptr) { + const signed short *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_sshort *)__addr; } static inline __ATTRS_o_ai vector unsigned short -vec_xl(signed long long __offset, unsigned short *__ptr) { +vec_xl(signed long long __offset, const unsigned short *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_ushort *)__addr; } static inline __ATTRS_o_ai vector signed int vec_xl(signed long long __offset, - signed int *__ptr) { + const signed int *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_sint *)__addr; } static inline __ATTRS_o_ai vector unsigned int vec_xl(signed long long __offset, - unsigned int *__ptr) { + const unsigned int *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_uint *)__addr; } static inline __ATTRS_o_ai vector float vec_xl(signed long long __offset, - float *__ptr) { + const float *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_float *)__addr; } @@ -16398,19 +16958,19 @@ typedef vector unsigned long long unaligned_vec_ull __attribute__((aligned(1))); typedef vector double unaligned_vec_double __attribute__((aligned(1))); static inline __ATTRS_o_ai vector signed long long -vec_xl(signed long long __offset, signed long long *__ptr) { +vec_xl(signed long long __offset, const signed long long *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_sll *)__addr; } static inline __ATTRS_o_ai vector unsigned long long -vec_xl(signed long long __offset, unsigned long long *__ptr) { +vec_xl(signed long long __offset, const unsigned long long *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_ull *)__addr; } static inline __ATTRS_o_ai vector double vec_xl(signed long long __offset, - double *__ptr) { + const double *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_double *)__addr; } @@ -16421,13 +16981,13 @@ typedef vector signed __int128 unaligned_vec_si128 __attribute__((aligned(1))); typedef vector unsigned __int128 unaligned_vec_ui128 __attribute__((aligned(1))); static inline __ATTRS_o_ai vector signed __int128 -vec_xl(signed long long __offset, signed __int128 *__ptr) { +vec_xl(signed long long __offset, const signed __int128 *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_si128 *)__addr; } static inline __ATTRS_o_ai vector unsigned __int128 -vec_xl(signed long long __offset, unsigned __int128 *__ptr) { +vec_xl(signed long long __offset, const unsigned __int128 *__ptr) { signed char *__addr = (signed char *)__ptr + __offset; return *(unaligned_vec_ui128 *)__addr; } @@ -16437,71 +16997,71 @@ vec_xl(signed long long __offset, unsigned __int128 *__ptr) { #ifdef __LITTLE_ENDIAN__ static __inline__ vector signed char __ATTRS_o_ai -vec_xl_be(signed long long __offset, signed char *__ptr) { +vec_xl_be(signed long long __offset, const signed char *__ptr) { vector signed char __vec = (vector signed char)__builtin_vsx_lxvd2x_be(__offset, __ptr); return __builtin_shufflevector(__vec, __vec, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8); } static __inline__ vector unsigned char __ATTRS_o_ai -vec_xl_be(signed long long __offset, unsigned char *__ptr) { +vec_xl_be(signed long long __offset, const unsigned char *__ptr) { vector unsigned char __vec = (vector unsigned char)__builtin_vsx_lxvd2x_be(__offset, __ptr); return __builtin_shufflevector(__vec, __vec, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8); } static __inline__ vector signed short __ATTRS_o_ai -vec_xl_be(signed long long __offset, signed short *__ptr) { +vec_xl_be(signed long long __offset, const signed short *__ptr) { vector signed short __vec = (vector signed short)__builtin_vsx_lxvd2x_be(__offset, __ptr); return __builtin_shufflevector(__vec, __vec, 3, 2, 1, 0, 7, 6, 5, 4); } static __inline__ vector unsigned short __ATTRS_o_ai -vec_xl_be(signed long long __offset, unsigned short *__ptr) { +vec_xl_be(signed long long __offset, const unsigned short *__ptr) { vector unsigned short __vec = (vector unsigned short)__builtin_vsx_lxvd2x_be(__offset, __ptr); return __builtin_shufflevector(__vec, __vec, 3, 2, 1, 0, 7, 6, 5, 4); } static __inline__ vector signed int __ATTRS_o_ai -vec_xl_be(signed long long __offset, signed int *__ptr) { +vec_xl_be(signed long long __offset, const signed int *__ptr) { return (vector signed int)__builtin_vsx_lxvw4x_be(__offset, __ptr); } static __inline__ vector unsigned int __ATTRS_o_ai -vec_xl_be(signed long long __offset, unsigned int *__ptr) { +vec_xl_be(signed long long __offset, const unsigned int *__ptr) { return (vector unsigned int)__builtin_vsx_lxvw4x_be(__offset, __ptr); } static __inline__ vector float __ATTRS_o_ai -vec_xl_be(signed long long __offset, float *__ptr) { +vec_xl_be(signed long long __offset, const float *__ptr) { return (vector float)__builtin_vsx_lxvw4x_be(__offset, __ptr); } #ifdef __VSX__ static __inline__ vector signed long long __ATTRS_o_ai -vec_xl_be(signed long long __offset, signed long long *__ptr) { +vec_xl_be(signed long long __offset, const signed long long *__ptr) { return (vector signed long long)__builtin_vsx_lxvd2x_be(__offset, __ptr); } static __inline__ vector unsigned long long __ATTRS_o_ai -vec_xl_be(signed long long __offset, unsigned long long *__ptr) { +vec_xl_be(signed long long __offset, const unsigned long long *__ptr) { return (vector unsigned long long)__builtin_vsx_lxvd2x_be(__offset, __ptr); } static __inline__ vector double __ATTRS_o_ai -vec_xl_be(signed long long __offset, double *__ptr) { +vec_xl_be(signed long long __offset, const double *__ptr) { return (vector double)__builtin_vsx_lxvd2x_be(__offset, __ptr); } #endif #if defined(__POWER8_VECTOR__) && defined(__powerpc64__) static __inline__ vector signed __int128 __ATTRS_o_ai -vec_xl_be(signed long long __offset, signed __int128 *__ptr) { +vec_xl_be(signed long long __offset, const signed __int128 *__ptr) { return vec_xl(__offset, __ptr); } static __inline__ vector unsigned __int128 __ATTRS_o_ai -vec_xl_be(signed long long __offset, unsigned __int128 *__ptr) { +vec_xl_be(signed long long __offset, const unsigned __int128 *__ptr) { return vec_xl(__offset, __ptr); } #endif @@ -16509,6 +17069,54 @@ vec_xl_be(signed long long __offset, unsigned __int128 *__ptr) { #define vec_xl_be vec_xl #endif +#if defined(__POWER10_VECTOR__) && defined(__VSX__) + +/* vect_xl_sext */ + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_sext(signed long long __offset, const signed char *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_sext(signed long long __offset, const signed short *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_sext(signed long long __offset, const signed int *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_sext(signed long long __offset, const signed long long *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +/* vec_xl_zext */ + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_zext(signed long long __offset, const unsigned char *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_zext(signed long long __offset, const unsigned short *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_zext(signed long long __offset, const unsigned int *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_xl_zext(signed long long __offset, const unsigned long long *__pointer) { + return (vector unsigned __int128)*(__pointer + __offset); +} + +#endif + /* vec_xst */ static inline __ATTRS_o_ai void vec_xst(vector signed char __vec, @@ -16597,6 +17205,58 @@ static inline __ATTRS_o_ai void vec_xst(vector unsigned __int128 __vec, } #endif +/* vec_xst_trunc */ + +#if defined(__POWER10_VECTOR__) && defined(__VSX__) +static inline __ATTRS_o_ai void vec_xst_trunc(vector signed __int128 __vec, + signed long long __offset, + signed char *__ptr) { + *(__ptr + __offset) = (signed char)__vec[0]; +} + +static inline __ATTRS_o_ai void vec_xst_trunc(vector unsigned __int128 __vec, + signed long long __offset, + unsigned char *__ptr) { + *(__ptr + __offset) = (unsigned char)__vec[0]; +} + +static inline __ATTRS_o_ai void vec_xst_trunc(vector signed __int128 __vec, + signed long long __offset, + signed short *__ptr) { + *(__ptr + __offset) = (signed short)__vec[0]; +} + +static inline __ATTRS_o_ai void vec_xst_trunc(vector unsigned __int128 __vec, + signed long long __offset, + unsigned short *__ptr) { + *(__ptr + __offset) = (unsigned short)__vec[0]; +} + +static inline __ATTRS_o_ai void vec_xst_trunc(vector signed __int128 __vec, + signed long long __offset, + signed int *__ptr) { + *(__ptr + __offset) = (signed int)__vec[0]; +} + +static inline __ATTRS_o_ai void vec_xst_trunc(vector unsigned __int128 __vec, + signed long long __offset, + unsigned int *__ptr) { + *(__ptr + __offset) = (unsigned int)__vec[0]; +} + +static inline __ATTRS_o_ai void vec_xst_trunc(vector signed __int128 __vec, + signed long long __offset, + signed long long *__ptr) { + *(__ptr + __offset) = (signed long long)__vec[0]; +} + +static inline __ATTRS_o_ai void vec_xst_trunc(vector unsigned __int128 __vec, + signed long long __offset, + unsigned long long *__ptr) { + *(__ptr + __offset) = (unsigned long long)__vec[0]; +} +#endif + /* vec_xst_be */ #ifdef __LITTLE_ENDIAN__ @@ -16763,6 +17423,100 @@ static vector signed char __ATTRS_o_ai vec_nabs(vector signed char __a) { } #ifdef __POWER10_VECTOR__ + +/* vec_extractm */ + +static __inline__ unsigned int __ATTRS_o_ai +vec_extractm(vector unsigned char __a) { + return __builtin_altivec_vextractbm(__a); +} + +static __inline__ unsigned int __ATTRS_o_ai +vec_extractm(vector unsigned short __a) { + return __builtin_altivec_vextracthm(__a); +} + +static __inline__ unsigned int __ATTRS_o_ai +vec_extractm(vector unsigned int __a) { + return __builtin_altivec_vextractwm(__a); +} + +static __inline__ unsigned int __ATTRS_o_ai +vec_extractm(vector unsigned long long __a) { + return __builtin_altivec_vextractdm(__a); +} + +static __inline__ unsigned int __ATTRS_o_ai +vec_extractm(vector unsigned __int128 __a) { + return __builtin_altivec_vextractqm(__a); +} + +/* vec_expandm */ + +static __inline__ vector unsigned char __ATTRS_o_ai +vec_expandm(vector unsigned char __a) { + return __builtin_altivec_vexpandbm(__a); +} + +static __inline__ vector unsigned short __ATTRS_o_ai +vec_expandm(vector unsigned short __a) { + return __builtin_altivec_vexpandhm(__a); +} + +static __inline__ vector unsigned int __ATTRS_o_ai +vec_expandm(vector unsigned int __a) { + return __builtin_altivec_vexpandwm(__a); +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_expandm(vector unsigned long long __a) { + return __builtin_altivec_vexpanddm(__a); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_expandm(vector unsigned __int128 __a) { + return __builtin_altivec_vexpandqm(__a); +} + +/* vec_cntm */ + +#define vec_cntm(__a, __mp) \ + _Generic((__a), vector unsigned char \ + : __builtin_altivec_vcntmbb((__a), (unsigned int)(__mp)), \ + vector unsigned short \ + : __builtin_altivec_vcntmbh((__a), (unsigned int)(__mp)), \ + vector unsigned int \ + : __builtin_altivec_vcntmbw((__a), (unsigned int)(__mp)), \ + vector unsigned long long \ + : __builtin_altivec_vcntmbd((__a), (unsigned int)(__mp))) + +/* vec_gen[b|h|w|d|q]m */ + +static __inline__ vector unsigned char __ATTRS_o_ai +vec_genbm(unsigned long long __bm) { + return __builtin_altivec_mtvsrbm(__bm); +} + +static __inline__ vector unsigned short __ATTRS_o_ai +vec_genhm(unsigned long long __bm) { + return __builtin_altivec_mtvsrhm(__bm); +} + +static __inline__ vector unsigned int __ATTRS_o_ai +vec_genwm(unsigned long long __bm) { + return __builtin_altivec_mtvsrwm(__bm); +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_gendm(unsigned long long __bm) { + return __builtin_altivec_mtvsrdm(__bm); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_genqm(unsigned long long __bm) { + return __builtin_altivec_mtvsrqm(__bm); +} + /* vec_pdep */ static __inline__ vector unsigned long long __ATTRS_o_ai @@ -16881,6 +17635,38 @@ vec_cnttzm(vector unsigned long long __a, vector unsigned long long __b) { return __builtin_altivec_vctzdm(__a, __b); } +/* vec_mod */ + +static __inline__ vector signed int __ATTRS_o_ai +vec_mod(vector signed int __a, vector signed int __b) { + return __a % __b; +} + +static __inline__ vector unsigned int __ATTRS_o_ai +vec_mod(vector unsigned int __a, vector unsigned int __b) { + return __a % __b; +} + +static __inline__ vector signed long long __ATTRS_o_ai +vec_mod(vector signed long long __a, vector signed long long __b) { + return __a % __b; +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_mod(vector unsigned long long __a, vector unsigned long long __b) { + return __a % __b; +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_mod(vector signed __int128 __a, vector signed __int128 __b) { + return __a % __b; +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_mod(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return __a % __b; +} + /* vec_sldbi */ #define vec_sldb(__a, __b, __c) __builtin_altivec_vsldbi(__a, __b, (__c & 0x7)) @@ -17027,6 +17813,92 @@ vec_inserth(vector unsigned int __a, vector unsigned int __b, #endif } +/* vec_extractl */ + +static __inline__ vector unsigned long long __ATTRS_o_ai vec_extractl( + vector unsigned char __a, vector unsigned char __b, unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextdubvrx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextdubvlx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + +static __inline__ vector unsigned long long __ATTRS_o_ai vec_extractl( + vector unsigned short __a, vector unsigned short __b, unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextduhvrx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextduhvlx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + +static __inline__ vector unsigned long long __ATTRS_o_ai vec_extractl( + vector unsigned int __a, vector unsigned int __b, unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextduwvrx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextduwvlx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_extractl(vector unsigned long long __a, vector unsigned long long __b, + unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextddvrx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextddvlx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + +/* vec_extracth */ + +static __inline__ vector unsigned long long __ATTRS_o_ai vec_extracth( + vector unsigned char __a, vector unsigned char __b, unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextdubvlx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextdubvrx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + +static __inline__ vector unsigned long long __ATTRS_o_ai vec_extracth( + vector unsigned short __a, vector unsigned short __b, unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextduhvlx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextduhvrx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + +static __inline__ vector unsigned long long __ATTRS_o_ai vec_extracth( + vector unsigned int __a, vector unsigned int __b, unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextduwvlx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextduwvrx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + +static __inline__ vector unsigned long long __ATTRS_o_ai +vec_extracth(vector unsigned long long __a, vector unsigned long long __b, + unsigned int __c) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vextddvlx(__a, __b, __c); +#else + vector unsigned long long __ret = __builtin_altivec_vextddvrx(__a, __b, __c); + return vec_sld(__ret, __ret, 8); +#endif +} + #ifdef __VSX__ /* vec_permx */ @@ -17095,6 +17967,14 @@ vec_blendv(vector double __a, vector double __b, return __builtin_vsx_xxblendvd(__a, __b, __c); } +/* vec_replace_elt */ + +#define vec_replace_elt __builtin_altivec_vec_replace_elt + +/* vec_replace_unaligned */ + +#define vec_replace_unaligned __builtin_altivec_vec_replace_unaligned + /* vec_splati */ #define vec_splati(__a) \ @@ -17161,6 +18041,197 @@ vec_test_lsbb_all_zeros(vector unsigned char __a) { return __builtin_vsx_xvtlsbb(__a, 0); } #endif /* __VSX__ */ + +/* vec_stril */ + +static __inline__ vector unsigned char __ATTRS_o_ai +vec_stril(vector unsigned char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribr((vector signed char)__a); +#else + return __builtin_altivec_vstribl((vector signed char)__a); +#endif +} + +static __inline__ vector signed char __ATTRS_o_ai +vec_stril(vector signed char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribr(__a); +#else + return __builtin_altivec_vstribl(__a); +#endif +} + +static __inline__ vector unsigned short __ATTRS_o_ai +vec_stril(vector unsigned short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihr((vector signed short)__a); +#else + return __builtin_altivec_vstrihl((vector signed short)__a); +#endif +} + +static __inline__ vector signed short __ATTRS_o_ai +vec_stril(vector signed short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihr(__a); +#else + return __builtin_altivec_vstrihl(__a); +#endif +} + +/* vec_stril_p */ + +static __inline__ int __ATTRS_o_ai vec_stril_p(vector unsigned char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribr_p(__CR6_EQ, (vector signed char)__a); +#else + return __builtin_altivec_vstribl_p(__CR6_EQ, (vector signed char)__a); +#endif +} + +static __inline__ int __ATTRS_o_ai vec_stril_p(vector signed char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribr_p(__CR6_EQ, __a); +#else + return __builtin_altivec_vstribl_p(__CR6_EQ, __a); +#endif +} + +static __inline__ int __ATTRS_o_ai vec_stril_p(vector unsigned short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihr_p(__CR6_EQ, (vector signed short)__a); +#else + return __builtin_altivec_vstrihl_p(__CR6_EQ, (vector signed short)__a); +#endif +} + +static __inline__ int __ATTRS_o_ai vec_stril_p(vector signed short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihr_p(__CR6_EQ, __a); +#else + return __builtin_altivec_vstrihl_p(__CR6_EQ, __a); +#endif +} + +/* vec_strir */ + +static __inline__ vector unsigned char __ATTRS_o_ai +vec_strir(vector unsigned char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribl((vector signed char)__a); +#else + return __builtin_altivec_vstribr((vector signed char)__a); +#endif +} + +static __inline__ vector signed char __ATTRS_o_ai +vec_strir(vector signed char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribl(__a); +#else + return __builtin_altivec_vstribr(__a); +#endif +} + +static __inline__ vector unsigned short __ATTRS_o_ai +vec_strir(vector unsigned short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihl((vector signed short)__a); +#else + return __builtin_altivec_vstrihr((vector signed short)__a); +#endif +} + +static __inline__ vector signed short __ATTRS_o_ai +vec_strir(vector signed short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihl(__a); +#else + return __builtin_altivec_vstrihr(__a); +#endif +} + +/* vec_strir_p */ + +static __inline__ int __ATTRS_o_ai vec_strir_p(vector unsigned char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribl_p(__CR6_EQ, (vector signed char)__a); +#else + return __builtin_altivec_vstribr_p(__CR6_EQ, (vector signed char)__a); +#endif +} + +static __inline__ int __ATTRS_o_ai vec_strir_p(vector signed char __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstribl_p(__CR6_EQ, __a); +#else + return __builtin_altivec_vstribr_p(__CR6_EQ, __a); +#endif +} + +static __inline__ int __ATTRS_o_ai vec_strir_p(vector unsigned short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihl_p(__CR6_EQ, (vector signed short)__a); +#else + return __builtin_altivec_vstrihr_p(__CR6_EQ, (vector signed short)__a); +#endif +} + +static __inline__ int __ATTRS_o_ai vec_strir_p(vector signed short __a) { +#ifdef __LITTLE_ENDIAN__ + return __builtin_altivec_vstrihl_p(__CR6_EQ, __a); +#else + return __builtin_altivec_vstrihr_p(__CR6_EQ, __a); +#endif +} + +/* vs[l | r | ra] */ + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_sl(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return __a << (__b % (vector unsigned __int128)(sizeof(unsigned __int128) * + __CHAR_BIT__)); +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_sl(vector signed __int128 __a, vector unsigned __int128 __b) { + return __a << (__b % (vector unsigned __int128)(sizeof(unsigned __int128) * + __CHAR_BIT__)); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_sr(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return __a >> (__b % (vector unsigned __int128)(sizeof(unsigned __int128) * + __CHAR_BIT__)); +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_sr(vector signed __int128 __a, vector unsigned __int128 __b) { + return ( + vector signed __int128)(((vector unsigned __int128)__a) >> + (__b % + (vector unsigned __int128)(sizeof( + unsigned __int128) * + __CHAR_BIT__))); +} + +static __inline__ vector unsigned __int128 __ATTRS_o_ai +vec_sra(vector unsigned __int128 __a, vector unsigned __int128 __b) { + return ( + vector unsigned __int128)(((vector signed __int128)__a) >> + (__b % + (vector unsigned __int128)(sizeof( + unsigned __int128) * + __CHAR_BIT__))); +} + +static __inline__ vector signed __int128 __ATTRS_o_ai +vec_sra(vector signed __int128 __a, vector unsigned __int128 __b) { + return __a >> (__b % (vector unsigned __int128)(sizeof(unsigned __int128) * + __CHAR_BIT__)); +} + #endif /* __POWER10_VECTOR__ */ #undef __ATTRS_o_ai diff --git a/lib/include/amxintrin.h b/lib/include/amxintrin.h index 58254e21c8..823c7ca1f0 100644 --- a/lib/include/amxintrin.h +++ b/lib/include/amxintrin.h @@ -15,8 +15,8 @@ #define __AMXINTRIN_H #ifdef __x86_64__ -#define __DEFAULT_FN_ATTRS \ - __attribute__((__always_inline__, __nodebug__, __target__("amx-tile"))) +#define __DEFAULT_FN_ATTRS_TILE \ + __attribute__((__always_inline__, __nodebug__, __target__("amx-tile"))) /// Load tile configuration from a 64-byte memory location specified by /// "mem_addr". The tile configuration includes the tile type palette, the @@ -31,9 +31,8 @@ /// /// \param __config /// A pointer to 512-bits configuration -static __inline__ void __DEFAULT_FN_ATTRS -_tile_loadconfig(const void *__config) -{ +static __inline__ void __DEFAULT_FN_ATTRS_TILE +_tile_loadconfig(const void *__config) { __builtin_ia32_tile_loadconfig(__config); } @@ -48,9 +47,8 @@ _tile_loadconfig(const void *__config) /// /// \param __config /// A pointer to 512-bits configuration -static __inline__ void __DEFAULT_FN_ATTRS -_tile_storeconfig(void *__config) -{ +static __inline__ void __DEFAULT_FN_ATTRS_TILE +_tile_storeconfig(void *__config) { __builtin_ia32_tile_storeconfig(__config); } @@ -60,9 +58,7 @@ _tile_storeconfig(void *__config) /// \headerfile <x86intrin.h> /// /// This intrinsic corresponds to the <c> TILERELEASE </c> instruction. -static __inline__ void __DEFAULT_FN_ATTRS -_tile_release(void) -{ +static __inline__ void __DEFAULT_FN_ATTRS_TILE _tile_release(void) { __builtin_ia32_tilerelease(); } @@ -80,8 +76,9 @@ _tile_release(void) /// A pointer to base address. /// \param stride /// The stride between the rows' data to be loaded in memory. -#define _tile_loadd(dst, base, stride) \ - __builtin_ia32_tileloadd64((dst), ((const void *)(base)), (__SIZE_TYPE__)(stride)) +#define _tile_loadd(dst, base, stride) \ + __builtin_ia32_tileloadd64((dst), ((const void *)(base)), \ + (__SIZE_TYPE__)(stride)) /// Load tile rows from memory specifieid by "base" address and "stride" into /// destination tile "dst" using the tile configuration previously configured @@ -99,8 +96,9 @@ _tile_release(void) /// A pointer to base address. /// \param stride /// The stride between the rows' data to be loaded in memory. -#define _tile_stream_loadd(dst, base, stride) \ - __builtin_ia32_tileloaddt164((dst), ((const void *)(base)), (__SIZE_TYPE__)(stride)) +#define _tile_stream_loadd(dst, base, stride) \ + __builtin_ia32_tileloaddt164((dst), ((const void *)(base)), \ + (__SIZE_TYPE__)(stride)) /// Store the tile specified by "src" to memory specifieid by "base" address and /// "stride" using the tile configuration previously configured via @@ -116,7 +114,7 @@ _tile_release(void) /// A pointer to base address. /// \param stride /// The stride between the rows' data to be stored in memory. -#define _tile_stored(dst, base, stride) \ +#define _tile_stored(dst, base, stride) \ __builtin_ia32_tilestored64((dst), ((void *)(base)), (__SIZE_TYPE__)(stride)) /// Zero the tile specified by "tdest". @@ -145,7 +143,8 @@ _tile_release(void) /// The 1st source tile. Max size is 1024 Bytes. /// \param src1 /// The 2nd source tile. Max size is 1024 Bytes. -#define _tile_dpbssd(dst, src0, src1) __builtin_ia32_tdpbssd((dst), (src0), (src1)) +#define _tile_dpbssd(dst, src0, src1) \ + __builtin_ia32_tdpbssd((dst), (src0), (src1)) /// Compute dot-product of bytes in tiles with a source/destination accumulator. /// Multiply groups of 4 adjacent pairs of signed 8-bit integers in src0 with @@ -163,7 +162,8 @@ _tile_release(void) /// The 1st source tile. Max size is 1024 Bytes. /// \param src1 /// The 2nd source tile. Max size is 1024 Bytes. -#define _tile_dpbsud(dst, src0, src1) __builtin_ia32_tdpbsud((dst), (src0), (src1)) +#define _tile_dpbsud(dst, src0, src1) \ + __builtin_ia32_tdpbsud((dst), (src0), (src1)) /// Compute dot-product of bytes in tiles with a source/destination accumulator. /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in src0 with @@ -181,7 +181,8 @@ _tile_release(void) /// The 1st source tile. Max size is 1024 Bytes. /// \param src1 /// The 2nd source tile. Max size is 1024 Bytes. -#define _tile_dpbusd(dst, src0, src1) __builtin_ia32_tdpbusd((dst), (src0), (src1)) +#define _tile_dpbusd(dst, src0, src1) \ + __builtin_ia32_tdpbusd((dst), (src0), (src1)) /// Compute dot-product of bytes in tiles with a source/destination accumulator. /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in src0 with @@ -199,7 +200,8 @@ _tile_release(void) /// The 1st source tile. Max size is 1024 Bytes. /// \param src1 /// The 2nd source tile. Max size is 1024 Bytes. -#define _tile_dpbuud(dst, src0, src1) __builtin_ia32_tdpbuud((dst), (src0), (src1)) +#define _tile_dpbuud(dst, src0, src1) \ + __builtin_ia32_tdpbuud((dst), (src0), (src1)) /// Compute dot-product of BF16 (16-bit) floating-point pairs in tiles src0 and /// src1, accumulating the intermediate single-precision (32-bit) floating-point @@ -216,10 +218,61 @@ _tile_release(void) /// The 1st source tile. Max size is 1024 Bytes. /// \param src1 /// The 2nd source tile. Max size is 1024 Bytes. -#define _tile_dpbf16ps(dst, src0, src1) \ +#define _tile_dpbf16ps(dst, src0, src1) \ __builtin_ia32_tdpbf16ps((dst), (src0), (src1)) -#undef __DEFAULT_FN_ATTRS +#define __DEFAULT_FN_ATTRS_INT8 \ + __attribute__((__always_inline__, __nodebug__, __target__("amx-int8"))) + +typedef int _tile1024i __attribute__((__vector_size__(1024), __aligned__(64))); +static __inline__ _tile1024i __DEFAULT_FN_ATTRS_INT8 +_tile_loadd_internal(unsigned short m, unsigned short n, const void *base, + __SIZE_TYPE__ stride) { + return __builtin_ia32_tileloadd64_internal(m, n, base, + (__SIZE_TYPE__)(stride)); +} + +static __inline__ _tile1024i __DEFAULT_FN_ATTRS_INT8 +_tile_dpbssd_internal(unsigned short m, unsigned short n, unsigned short k, + _tile1024i dst, _tile1024i src1, _tile1024i src2) { + return __builtin_ia32_tdpbssd_internal(m, n, k, dst, src1, src2); +} + +static __inline__ void __DEFAULT_FN_ATTRS_INT8 +_tile_stored_internal(unsigned short m, unsigned short n, void *base, + __SIZE_TYPE__ stride, _tile1024i tile) { + return __builtin_ia32_tilestored64_internal(m, n, base, + (__SIZE_TYPE__)(stride), tile); +} + +typedef struct __tile1024i_str { + const unsigned short row; + const unsigned short col; + _tile1024i tile; +} __tile1024i; + +__DEFAULT_FN_ATTRS_TILE +static void __tile_loadd(__tile1024i *dst, const void *base, + __SIZE_TYPE__ stride) { + dst->tile = _tile_loadd_internal(dst->row, dst->col, base, stride); +} + +__DEFAULT_FN_ATTRS_INT8 +static void __tile_dpbssd(__tile1024i *dst, __tile1024i src1, + __tile1024i src2) { + dst->tile = _tile_dpbssd_internal(src1.row, src2.col, src1.col, dst->tile, + src1.tile, src2.tile); +} + +__DEFAULT_FN_ATTRS_TILE +static void __tile_stored(void *base, __SIZE_TYPE__ stride, __tile1024i src) { + _tile_stored_internal(src.row, src.col, base, stride, src.tile); +} + +__DEFAULT_FN_ATTRS_TILE +static void __tile_zero(__tile1024i *dst) { + dst->tile = __builtin_ia32_tilezero_internal(dst->row, dst->col); +} #endif /* __x86_64__ */ #endif /* __AMXINTRIN_H */ diff --git a/lib/include/arm_acle.h b/lib/include/arm_acle.h index de568b4ff9..c156d89c1f 100644 --- a/lib/include/arm_acle.h +++ b/lib/include/arm_acle.h @@ -639,6 +639,32 @@ __jcvt(double __a) { } #endif +/* Armv8.7-A load/store 64-byte intrinsics */ +#if __ARM_64BIT_STATE && defined(__ARM_FEATURE_LS64) +typedef struct { + uint64_t val[8]; +} data512_t; + +static __inline__ data512_t __attribute__((__always_inline__, __nodebug__)) +__arm_ld64b(const void *__addr) { + data512_t __value; + __builtin_arm_ld64b(__addr, __value.val); + return __value; +} +static __inline__ void __attribute__((__always_inline__, __nodebug__)) +__arm_st64b(void *__addr, data512_t __value) { + __builtin_arm_st64b(__addr, __value.val); +} +static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__)) +__arm_st64bv(void *__addr, data512_t __value) { + return __builtin_arm_st64bv(__addr, __value.val); +} +static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__)) +__arm_st64bv0(void *__addr, data512_t __value) { + return __builtin_arm_st64bv0(__addr, __value.val); +} +#endif + /* 10.1 Special register intrinsics */ #define __arm_rsr(sysreg) __builtin_arm_rsr(sysreg) #define __arm_rsr64(sysreg) __builtin_arm_rsr64(sysreg) diff --git a/lib/include/arm_neon.h b/lib/include/arm_neon.h index da1e17cc00..dbb65b8a49 100644 --- a/lib/include/arm_neon.h +++ b/lib/include/arm_neon.h @@ -40429,6 +40429,638 @@ __ai float32x4_t vcaddq_rot90_f32(float32x4_t __p0, float32x4_t __p1) { } #endif +#ifdef __LITTLE_ENDIAN__ +__ai float32x4_t vcmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#else +__ai float32x4_t vcmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float32x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float32x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float32x4_t __noswap_vcmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float32x2_t vcmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#else +__ai float32x2_t vcmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float32x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float32x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float32x2_t __noswap_vcmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_lane_f32(__p0_169, __p1_169, __p2_169, __p3_169) __extension__ ({ \ + float32x2_t __s0_169 = __p0_169; \ + float32x2_t __s1_169 = __p1_169; \ + float32x2_t __s2_169 = __p2_169; \ + float32x2_t __ret_169; \ +float32x2_t __reint_169 = __s2_169; \ +uint64x1_t __reint1_169 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_169, __p3_169)}; \ + __ret_169 = vcmla_f32(__s0_169, __s1_169, *(float32x2_t *) &__reint1_169); \ + __ret_169; \ +}) +#else +#define vcmla_lane_f32(__p0_170, __p1_170, __p2_170, __p3_170) __extension__ ({ \ + float32x2_t __s0_170 = __p0_170; \ + float32x2_t __s1_170 = __p1_170; \ + float32x2_t __s2_170 = __p2_170; \ + float32x2_t __rev0_170; __rev0_170 = __builtin_shufflevector(__s0_170, __s0_170, 1, 0); \ + float32x2_t __rev1_170; __rev1_170 = __builtin_shufflevector(__s1_170, __s1_170, 1, 0); \ + float32x2_t __rev2_170; __rev2_170 = __builtin_shufflevector(__s2_170, __s2_170, 1, 0); \ + float32x2_t __ret_170; \ +float32x2_t __reint_170 = __rev2_170; \ +uint64x1_t __reint1_170 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_170, __p3_170)}; \ + __ret_170 = __noswap_vcmla_f32(__rev0_170, __rev1_170, *(float32x2_t *) &__reint1_170); \ + __ret_170 = __builtin_shufflevector(__ret_170, __ret_170, 1, 0); \ + __ret_170; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_lane_f32(__p0_171, __p1_171, __p2_171, __p3_171) __extension__ ({ \ + float32x4_t __s0_171 = __p0_171; \ + float32x4_t __s1_171 = __p1_171; \ + float32x2_t __s2_171 = __p2_171; \ + float32x4_t __ret_171; \ +float32x2_t __reint_171 = __s2_171; \ +uint64x2_t __reint1_171 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_171, __p3_171), vget_lane_u64(*(uint64x1_t *) &__reint_171, __p3_171)}; \ + __ret_171 = vcmlaq_f32(__s0_171, __s1_171, *(float32x4_t *) &__reint1_171); \ + __ret_171; \ +}) +#else +#define vcmlaq_lane_f32(__p0_172, __p1_172, __p2_172, __p3_172) __extension__ ({ \ + float32x4_t __s0_172 = __p0_172; \ + float32x4_t __s1_172 = __p1_172; \ + float32x2_t __s2_172 = __p2_172; \ + float32x4_t __rev0_172; __rev0_172 = __builtin_shufflevector(__s0_172, __s0_172, 3, 2, 1, 0); \ + float32x4_t __rev1_172; __rev1_172 = __builtin_shufflevector(__s1_172, __s1_172, 3, 2, 1, 0); \ + float32x2_t __rev2_172; __rev2_172 = __builtin_shufflevector(__s2_172, __s2_172, 1, 0); \ + float32x4_t __ret_172; \ +float32x2_t __reint_172 = __rev2_172; \ +uint64x2_t __reint1_172 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_172, __p3_172), vget_lane_u64(*(uint64x1_t *) &__reint_172, __p3_172)}; \ + __ret_172 = __noswap_vcmlaq_f32(__rev0_172, __rev1_172, *(float32x4_t *) &__reint1_172); \ + __ret_172 = __builtin_shufflevector(__ret_172, __ret_172, 3, 2, 1, 0); \ + __ret_172; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_laneq_f32(__p0_173, __p1_173, __p2_173, __p3_173) __extension__ ({ \ + float32x2_t __s0_173 = __p0_173; \ + float32x2_t __s1_173 = __p1_173; \ + float32x4_t __s2_173 = __p2_173; \ + float32x2_t __ret_173; \ +float32x4_t __reint_173 = __s2_173; \ +uint64x1_t __reint1_173 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_173, __p3_173)}; \ + __ret_173 = vcmla_f32(__s0_173, __s1_173, *(float32x2_t *) &__reint1_173); \ + __ret_173; \ +}) +#else +#define vcmla_laneq_f32(__p0_174, __p1_174, __p2_174, __p3_174) __extension__ ({ \ + float32x2_t __s0_174 = __p0_174; \ + float32x2_t __s1_174 = __p1_174; \ + float32x4_t __s2_174 = __p2_174; \ + float32x2_t __rev0_174; __rev0_174 = __builtin_shufflevector(__s0_174, __s0_174, 1, 0); \ + float32x2_t __rev1_174; __rev1_174 = __builtin_shufflevector(__s1_174, __s1_174, 1, 0); \ + float32x4_t __rev2_174; __rev2_174 = __builtin_shufflevector(__s2_174, __s2_174, 3, 2, 1, 0); \ + float32x2_t __ret_174; \ +float32x4_t __reint_174 = __rev2_174; \ +uint64x1_t __reint1_174 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_174, __p3_174)}; \ + __ret_174 = __noswap_vcmla_f32(__rev0_174, __rev1_174, *(float32x2_t *) &__reint1_174); \ + __ret_174 = __builtin_shufflevector(__ret_174, __ret_174, 1, 0); \ + __ret_174; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_laneq_f32(__p0_175, __p1_175, __p2_175, __p3_175) __extension__ ({ \ + float32x4_t __s0_175 = __p0_175; \ + float32x4_t __s1_175 = __p1_175; \ + float32x4_t __s2_175 = __p2_175; \ + float32x4_t __ret_175; \ +float32x4_t __reint_175 = __s2_175; \ +uint64x2_t __reint1_175 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_175, __p3_175), vgetq_lane_u64(*(uint64x2_t *) &__reint_175, __p3_175)}; \ + __ret_175 = vcmlaq_f32(__s0_175, __s1_175, *(float32x4_t *) &__reint1_175); \ + __ret_175; \ +}) +#else +#define vcmlaq_laneq_f32(__p0_176, __p1_176, __p2_176, __p3_176) __extension__ ({ \ + float32x4_t __s0_176 = __p0_176; \ + float32x4_t __s1_176 = __p1_176; \ + float32x4_t __s2_176 = __p2_176; \ + float32x4_t __rev0_176; __rev0_176 = __builtin_shufflevector(__s0_176, __s0_176, 3, 2, 1, 0); \ + float32x4_t __rev1_176; __rev1_176 = __builtin_shufflevector(__s1_176, __s1_176, 3, 2, 1, 0); \ + float32x4_t __rev2_176; __rev2_176 = __builtin_shufflevector(__s2_176, __s2_176, 3, 2, 1, 0); \ + float32x4_t __ret_176; \ +float32x4_t __reint_176 = __rev2_176; \ +uint64x2_t __reint1_176 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_176, __p3_176), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_176, __p3_176)}; \ + __ret_176 = __noswap_vcmlaq_f32(__rev0_176, __rev1_176, *(float32x4_t *) &__reint1_176); \ + __ret_176 = __builtin_shufflevector(__ret_176, __ret_176, 3, 2, 1, 0); \ + __ret_176; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float32x4_t vcmlaq_rot180_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#else +__ai float32x4_t vcmlaq_rot180_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float32x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float32x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float32x4_t __noswap_vcmlaq_rot180_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float32x2_t vcmla_rot180_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot180_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#else +__ai float32x2_t vcmla_rot180_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float32x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float32x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot180_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float32x2_t __noswap_vcmla_rot180_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot180_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot180_lane_f32(__p0_177, __p1_177, __p2_177, __p3_177) __extension__ ({ \ + float32x2_t __s0_177 = __p0_177; \ + float32x2_t __s1_177 = __p1_177; \ + float32x2_t __s2_177 = __p2_177; \ + float32x2_t __ret_177; \ +float32x2_t __reint_177 = __s2_177; \ +uint64x1_t __reint1_177 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_177, __p3_177)}; \ + __ret_177 = vcmla_rot180_f32(__s0_177, __s1_177, *(float32x2_t *) &__reint1_177); \ + __ret_177; \ +}) +#else +#define vcmla_rot180_lane_f32(__p0_178, __p1_178, __p2_178, __p3_178) __extension__ ({ \ + float32x2_t __s0_178 = __p0_178; \ + float32x2_t __s1_178 = __p1_178; \ + float32x2_t __s2_178 = __p2_178; \ + float32x2_t __rev0_178; __rev0_178 = __builtin_shufflevector(__s0_178, __s0_178, 1, 0); \ + float32x2_t __rev1_178; __rev1_178 = __builtin_shufflevector(__s1_178, __s1_178, 1, 0); \ + float32x2_t __rev2_178; __rev2_178 = __builtin_shufflevector(__s2_178, __s2_178, 1, 0); \ + float32x2_t __ret_178; \ +float32x2_t __reint_178 = __rev2_178; \ +uint64x1_t __reint1_178 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_178, __p3_178)}; \ + __ret_178 = __noswap_vcmla_rot180_f32(__rev0_178, __rev1_178, *(float32x2_t *) &__reint1_178); \ + __ret_178 = __builtin_shufflevector(__ret_178, __ret_178, 1, 0); \ + __ret_178; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot180_lane_f32(__p0_179, __p1_179, __p2_179, __p3_179) __extension__ ({ \ + float32x4_t __s0_179 = __p0_179; \ + float32x4_t __s1_179 = __p1_179; \ + float32x2_t __s2_179 = __p2_179; \ + float32x4_t __ret_179; \ +float32x2_t __reint_179 = __s2_179; \ +uint64x2_t __reint1_179 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_179, __p3_179), vget_lane_u64(*(uint64x1_t *) &__reint_179, __p3_179)}; \ + __ret_179 = vcmlaq_rot180_f32(__s0_179, __s1_179, *(float32x4_t *) &__reint1_179); \ + __ret_179; \ +}) +#else +#define vcmlaq_rot180_lane_f32(__p0_180, __p1_180, __p2_180, __p3_180) __extension__ ({ \ + float32x4_t __s0_180 = __p0_180; \ + float32x4_t __s1_180 = __p1_180; \ + float32x2_t __s2_180 = __p2_180; \ + float32x4_t __rev0_180; __rev0_180 = __builtin_shufflevector(__s0_180, __s0_180, 3, 2, 1, 0); \ + float32x4_t __rev1_180; __rev1_180 = __builtin_shufflevector(__s1_180, __s1_180, 3, 2, 1, 0); \ + float32x2_t __rev2_180; __rev2_180 = __builtin_shufflevector(__s2_180, __s2_180, 1, 0); \ + float32x4_t __ret_180; \ +float32x2_t __reint_180 = __rev2_180; \ +uint64x2_t __reint1_180 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_180, __p3_180), vget_lane_u64(*(uint64x1_t *) &__reint_180, __p3_180)}; \ + __ret_180 = __noswap_vcmlaq_rot180_f32(__rev0_180, __rev1_180, *(float32x4_t *) &__reint1_180); \ + __ret_180 = __builtin_shufflevector(__ret_180, __ret_180, 3, 2, 1, 0); \ + __ret_180; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot180_laneq_f32(__p0_181, __p1_181, __p2_181, __p3_181) __extension__ ({ \ + float32x2_t __s0_181 = __p0_181; \ + float32x2_t __s1_181 = __p1_181; \ + float32x4_t __s2_181 = __p2_181; \ + float32x2_t __ret_181; \ +float32x4_t __reint_181 = __s2_181; \ +uint64x1_t __reint1_181 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_181, __p3_181)}; \ + __ret_181 = vcmla_rot180_f32(__s0_181, __s1_181, *(float32x2_t *) &__reint1_181); \ + __ret_181; \ +}) +#else +#define vcmla_rot180_laneq_f32(__p0_182, __p1_182, __p2_182, __p3_182) __extension__ ({ \ + float32x2_t __s0_182 = __p0_182; \ + float32x2_t __s1_182 = __p1_182; \ + float32x4_t __s2_182 = __p2_182; \ + float32x2_t __rev0_182; __rev0_182 = __builtin_shufflevector(__s0_182, __s0_182, 1, 0); \ + float32x2_t __rev1_182; __rev1_182 = __builtin_shufflevector(__s1_182, __s1_182, 1, 0); \ + float32x4_t __rev2_182; __rev2_182 = __builtin_shufflevector(__s2_182, __s2_182, 3, 2, 1, 0); \ + float32x2_t __ret_182; \ +float32x4_t __reint_182 = __rev2_182; \ +uint64x1_t __reint1_182 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_182, __p3_182)}; \ + __ret_182 = __noswap_vcmla_rot180_f32(__rev0_182, __rev1_182, *(float32x2_t *) &__reint1_182); \ + __ret_182 = __builtin_shufflevector(__ret_182, __ret_182, 1, 0); \ + __ret_182; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot180_laneq_f32(__p0_183, __p1_183, __p2_183, __p3_183) __extension__ ({ \ + float32x4_t __s0_183 = __p0_183; \ + float32x4_t __s1_183 = __p1_183; \ + float32x4_t __s2_183 = __p2_183; \ + float32x4_t __ret_183; \ +float32x4_t __reint_183 = __s2_183; \ +uint64x2_t __reint1_183 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_183, __p3_183), vgetq_lane_u64(*(uint64x2_t *) &__reint_183, __p3_183)}; \ + __ret_183 = vcmlaq_rot180_f32(__s0_183, __s1_183, *(float32x4_t *) &__reint1_183); \ + __ret_183; \ +}) +#else +#define vcmlaq_rot180_laneq_f32(__p0_184, __p1_184, __p2_184, __p3_184) __extension__ ({ \ + float32x4_t __s0_184 = __p0_184; \ + float32x4_t __s1_184 = __p1_184; \ + float32x4_t __s2_184 = __p2_184; \ + float32x4_t __rev0_184; __rev0_184 = __builtin_shufflevector(__s0_184, __s0_184, 3, 2, 1, 0); \ + float32x4_t __rev1_184; __rev1_184 = __builtin_shufflevector(__s1_184, __s1_184, 3, 2, 1, 0); \ + float32x4_t __rev2_184; __rev2_184 = __builtin_shufflevector(__s2_184, __s2_184, 3, 2, 1, 0); \ + float32x4_t __ret_184; \ +float32x4_t __reint_184 = __rev2_184; \ +uint64x2_t __reint1_184 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_184, __p3_184), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_184, __p3_184)}; \ + __ret_184 = __noswap_vcmlaq_rot180_f32(__rev0_184, __rev1_184, *(float32x4_t *) &__reint1_184); \ + __ret_184 = __builtin_shufflevector(__ret_184, __ret_184, 3, 2, 1, 0); \ + __ret_184; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float32x4_t vcmlaq_rot270_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#else +__ai float32x4_t vcmlaq_rot270_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float32x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float32x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float32x4_t __noswap_vcmlaq_rot270_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float32x2_t vcmla_rot270_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot270_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#else +__ai float32x2_t vcmla_rot270_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float32x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float32x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot270_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float32x2_t __noswap_vcmla_rot270_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot270_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot270_lane_f32(__p0_185, __p1_185, __p2_185, __p3_185) __extension__ ({ \ + float32x2_t __s0_185 = __p0_185; \ + float32x2_t __s1_185 = __p1_185; \ + float32x2_t __s2_185 = __p2_185; \ + float32x2_t __ret_185; \ +float32x2_t __reint_185 = __s2_185; \ +uint64x1_t __reint1_185 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_185, __p3_185)}; \ + __ret_185 = vcmla_rot270_f32(__s0_185, __s1_185, *(float32x2_t *) &__reint1_185); \ + __ret_185; \ +}) +#else +#define vcmla_rot270_lane_f32(__p0_186, __p1_186, __p2_186, __p3_186) __extension__ ({ \ + float32x2_t __s0_186 = __p0_186; \ + float32x2_t __s1_186 = __p1_186; \ + float32x2_t __s2_186 = __p2_186; \ + float32x2_t __rev0_186; __rev0_186 = __builtin_shufflevector(__s0_186, __s0_186, 1, 0); \ + float32x2_t __rev1_186; __rev1_186 = __builtin_shufflevector(__s1_186, __s1_186, 1, 0); \ + float32x2_t __rev2_186; __rev2_186 = __builtin_shufflevector(__s2_186, __s2_186, 1, 0); \ + float32x2_t __ret_186; \ +float32x2_t __reint_186 = __rev2_186; \ +uint64x1_t __reint1_186 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_186, __p3_186)}; \ + __ret_186 = __noswap_vcmla_rot270_f32(__rev0_186, __rev1_186, *(float32x2_t *) &__reint1_186); \ + __ret_186 = __builtin_shufflevector(__ret_186, __ret_186, 1, 0); \ + __ret_186; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot270_lane_f32(__p0_187, __p1_187, __p2_187, __p3_187) __extension__ ({ \ + float32x4_t __s0_187 = __p0_187; \ + float32x4_t __s1_187 = __p1_187; \ + float32x2_t __s2_187 = __p2_187; \ + float32x4_t __ret_187; \ +float32x2_t __reint_187 = __s2_187; \ +uint64x2_t __reint1_187 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_187, __p3_187), vget_lane_u64(*(uint64x1_t *) &__reint_187, __p3_187)}; \ + __ret_187 = vcmlaq_rot270_f32(__s0_187, __s1_187, *(float32x4_t *) &__reint1_187); \ + __ret_187; \ +}) +#else +#define vcmlaq_rot270_lane_f32(__p0_188, __p1_188, __p2_188, __p3_188) __extension__ ({ \ + float32x4_t __s0_188 = __p0_188; \ + float32x4_t __s1_188 = __p1_188; \ + float32x2_t __s2_188 = __p2_188; \ + float32x4_t __rev0_188; __rev0_188 = __builtin_shufflevector(__s0_188, __s0_188, 3, 2, 1, 0); \ + float32x4_t __rev1_188; __rev1_188 = __builtin_shufflevector(__s1_188, __s1_188, 3, 2, 1, 0); \ + float32x2_t __rev2_188; __rev2_188 = __builtin_shufflevector(__s2_188, __s2_188, 1, 0); \ + float32x4_t __ret_188; \ +float32x2_t __reint_188 = __rev2_188; \ +uint64x2_t __reint1_188 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_188, __p3_188), vget_lane_u64(*(uint64x1_t *) &__reint_188, __p3_188)}; \ + __ret_188 = __noswap_vcmlaq_rot270_f32(__rev0_188, __rev1_188, *(float32x4_t *) &__reint1_188); \ + __ret_188 = __builtin_shufflevector(__ret_188, __ret_188, 3, 2, 1, 0); \ + __ret_188; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot270_laneq_f32(__p0_189, __p1_189, __p2_189, __p3_189) __extension__ ({ \ + float32x2_t __s0_189 = __p0_189; \ + float32x2_t __s1_189 = __p1_189; \ + float32x4_t __s2_189 = __p2_189; \ + float32x2_t __ret_189; \ +float32x4_t __reint_189 = __s2_189; \ +uint64x1_t __reint1_189 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_189, __p3_189)}; \ + __ret_189 = vcmla_rot270_f32(__s0_189, __s1_189, *(float32x2_t *) &__reint1_189); \ + __ret_189; \ +}) +#else +#define vcmla_rot270_laneq_f32(__p0_190, __p1_190, __p2_190, __p3_190) __extension__ ({ \ + float32x2_t __s0_190 = __p0_190; \ + float32x2_t __s1_190 = __p1_190; \ + float32x4_t __s2_190 = __p2_190; \ + float32x2_t __rev0_190; __rev0_190 = __builtin_shufflevector(__s0_190, __s0_190, 1, 0); \ + float32x2_t __rev1_190; __rev1_190 = __builtin_shufflevector(__s1_190, __s1_190, 1, 0); \ + float32x4_t __rev2_190; __rev2_190 = __builtin_shufflevector(__s2_190, __s2_190, 3, 2, 1, 0); \ + float32x2_t __ret_190; \ +float32x4_t __reint_190 = __rev2_190; \ +uint64x1_t __reint1_190 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_190, __p3_190)}; \ + __ret_190 = __noswap_vcmla_rot270_f32(__rev0_190, __rev1_190, *(float32x2_t *) &__reint1_190); \ + __ret_190 = __builtin_shufflevector(__ret_190, __ret_190, 1, 0); \ + __ret_190; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot270_laneq_f32(__p0_191, __p1_191, __p2_191, __p3_191) __extension__ ({ \ + float32x4_t __s0_191 = __p0_191; \ + float32x4_t __s1_191 = __p1_191; \ + float32x4_t __s2_191 = __p2_191; \ + float32x4_t __ret_191; \ +float32x4_t __reint_191 = __s2_191; \ +uint64x2_t __reint1_191 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_191, __p3_191), vgetq_lane_u64(*(uint64x2_t *) &__reint_191, __p3_191)}; \ + __ret_191 = vcmlaq_rot270_f32(__s0_191, __s1_191, *(float32x4_t *) &__reint1_191); \ + __ret_191; \ +}) +#else +#define vcmlaq_rot270_laneq_f32(__p0_192, __p1_192, __p2_192, __p3_192) __extension__ ({ \ + float32x4_t __s0_192 = __p0_192; \ + float32x4_t __s1_192 = __p1_192; \ + float32x4_t __s2_192 = __p2_192; \ + float32x4_t __rev0_192; __rev0_192 = __builtin_shufflevector(__s0_192, __s0_192, 3, 2, 1, 0); \ + float32x4_t __rev1_192; __rev1_192 = __builtin_shufflevector(__s1_192, __s1_192, 3, 2, 1, 0); \ + float32x4_t __rev2_192; __rev2_192 = __builtin_shufflevector(__s2_192, __s2_192, 3, 2, 1, 0); \ + float32x4_t __ret_192; \ +float32x4_t __reint_192 = __rev2_192; \ +uint64x2_t __reint1_192 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_192, __p3_192), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_192, __p3_192)}; \ + __ret_192 = __noswap_vcmlaq_rot270_f32(__rev0_192, __rev1_192, *(float32x4_t *) &__reint1_192); \ + __ret_192 = __builtin_shufflevector(__ret_192, __ret_192, 3, 2, 1, 0); \ + __ret_192; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float32x4_t vcmlaq_rot90_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#else +__ai float32x4_t vcmlaq_rot90_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float32x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float32x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float32x4_t __noswap_vcmlaq_rot90_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) { + float32x4_t __ret; + __ret = (float32x4_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float32x2_t vcmla_rot90_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot90_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#else +__ai float32x2_t vcmla_rot90_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float32x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float32x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot90_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float32x2_t __noswap_vcmla_rot90_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) { + float32x2_t __ret; + __ret = (float32x2_t) __builtin_neon_vcmla_rot90_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot90_lane_f32(__p0_193, __p1_193, __p2_193, __p3_193) __extension__ ({ \ + float32x2_t __s0_193 = __p0_193; \ + float32x2_t __s1_193 = __p1_193; \ + float32x2_t __s2_193 = __p2_193; \ + float32x2_t __ret_193; \ +float32x2_t __reint_193 = __s2_193; \ +uint64x1_t __reint1_193 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_193, __p3_193)}; \ + __ret_193 = vcmla_rot90_f32(__s0_193, __s1_193, *(float32x2_t *) &__reint1_193); \ + __ret_193; \ +}) +#else +#define vcmla_rot90_lane_f32(__p0_194, __p1_194, __p2_194, __p3_194) __extension__ ({ \ + float32x2_t __s0_194 = __p0_194; \ + float32x2_t __s1_194 = __p1_194; \ + float32x2_t __s2_194 = __p2_194; \ + float32x2_t __rev0_194; __rev0_194 = __builtin_shufflevector(__s0_194, __s0_194, 1, 0); \ + float32x2_t __rev1_194; __rev1_194 = __builtin_shufflevector(__s1_194, __s1_194, 1, 0); \ + float32x2_t __rev2_194; __rev2_194 = __builtin_shufflevector(__s2_194, __s2_194, 1, 0); \ + float32x2_t __ret_194; \ +float32x2_t __reint_194 = __rev2_194; \ +uint64x1_t __reint1_194 = (uint64x1_t) {vget_lane_u64(*(uint64x1_t *) &__reint_194, __p3_194)}; \ + __ret_194 = __noswap_vcmla_rot90_f32(__rev0_194, __rev1_194, *(float32x2_t *) &__reint1_194); \ + __ret_194 = __builtin_shufflevector(__ret_194, __ret_194, 1, 0); \ + __ret_194; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot90_lane_f32(__p0_195, __p1_195, __p2_195, __p3_195) __extension__ ({ \ + float32x4_t __s0_195 = __p0_195; \ + float32x4_t __s1_195 = __p1_195; \ + float32x2_t __s2_195 = __p2_195; \ + float32x4_t __ret_195; \ +float32x2_t __reint_195 = __s2_195; \ +uint64x2_t __reint1_195 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_195, __p3_195), vget_lane_u64(*(uint64x1_t *) &__reint_195, __p3_195)}; \ + __ret_195 = vcmlaq_rot90_f32(__s0_195, __s1_195, *(float32x4_t *) &__reint1_195); \ + __ret_195; \ +}) +#else +#define vcmlaq_rot90_lane_f32(__p0_196, __p1_196, __p2_196, __p3_196) __extension__ ({ \ + float32x4_t __s0_196 = __p0_196; \ + float32x4_t __s1_196 = __p1_196; \ + float32x2_t __s2_196 = __p2_196; \ + float32x4_t __rev0_196; __rev0_196 = __builtin_shufflevector(__s0_196, __s0_196, 3, 2, 1, 0); \ + float32x4_t __rev1_196; __rev1_196 = __builtin_shufflevector(__s1_196, __s1_196, 3, 2, 1, 0); \ + float32x2_t __rev2_196; __rev2_196 = __builtin_shufflevector(__s2_196, __s2_196, 1, 0); \ + float32x4_t __ret_196; \ +float32x2_t __reint_196 = __rev2_196; \ +uint64x2_t __reint1_196 = (uint64x2_t) {vget_lane_u64(*(uint64x1_t *) &__reint_196, __p3_196), vget_lane_u64(*(uint64x1_t *) &__reint_196, __p3_196)}; \ + __ret_196 = __noswap_vcmlaq_rot90_f32(__rev0_196, __rev1_196, *(float32x4_t *) &__reint1_196); \ + __ret_196 = __builtin_shufflevector(__ret_196, __ret_196, 3, 2, 1, 0); \ + __ret_196; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot90_laneq_f32(__p0_197, __p1_197, __p2_197, __p3_197) __extension__ ({ \ + float32x2_t __s0_197 = __p0_197; \ + float32x2_t __s1_197 = __p1_197; \ + float32x4_t __s2_197 = __p2_197; \ + float32x2_t __ret_197; \ +float32x4_t __reint_197 = __s2_197; \ +uint64x1_t __reint1_197 = (uint64x1_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_197, __p3_197)}; \ + __ret_197 = vcmla_rot90_f32(__s0_197, __s1_197, *(float32x2_t *) &__reint1_197); \ + __ret_197; \ +}) +#else +#define vcmla_rot90_laneq_f32(__p0_198, __p1_198, __p2_198, __p3_198) __extension__ ({ \ + float32x2_t __s0_198 = __p0_198; \ + float32x2_t __s1_198 = __p1_198; \ + float32x4_t __s2_198 = __p2_198; \ + float32x2_t __rev0_198; __rev0_198 = __builtin_shufflevector(__s0_198, __s0_198, 1, 0); \ + float32x2_t __rev1_198; __rev1_198 = __builtin_shufflevector(__s1_198, __s1_198, 1, 0); \ + float32x4_t __rev2_198; __rev2_198 = __builtin_shufflevector(__s2_198, __s2_198, 3, 2, 1, 0); \ + float32x2_t __ret_198; \ +float32x4_t __reint_198 = __rev2_198; \ +uint64x1_t __reint1_198 = (uint64x1_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_198, __p3_198)}; \ + __ret_198 = __noswap_vcmla_rot90_f32(__rev0_198, __rev1_198, *(float32x2_t *) &__reint1_198); \ + __ret_198 = __builtin_shufflevector(__ret_198, __ret_198, 1, 0); \ + __ret_198; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot90_laneq_f32(__p0_199, __p1_199, __p2_199, __p3_199) __extension__ ({ \ + float32x4_t __s0_199 = __p0_199; \ + float32x4_t __s1_199 = __p1_199; \ + float32x4_t __s2_199 = __p2_199; \ + float32x4_t __ret_199; \ +float32x4_t __reint_199 = __s2_199; \ +uint64x2_t __reint1_199 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_199, __p3_199), vgetq_lane_u64(*(uint64x2_t *) &__reint_199, __p3_199)}; \ + __ret_199 = vcmlaq_rot90_f32(__s0_199, __s1_199, *(float32x4_t *) &__reint1_199); \ + __ret_199; \ +}) +#else +#define vcmlaq_rot90_laneq_f32(__p0_200, __p1_200, __p2_200, __p3_200) __extension__ ({ \ + float32x4_t __s0_200 = __p0_200; \ + float32x4_t __s1_200 = __p1_200; \ + float32x4_t __s2_200 = __p2_200; \ + float32x4_t __rev0_200; __rev0_200 = __builtin_shufflevector(__s0_200, __s0_200, 3, 2, 1, 0); \ + float32x4_t __rev1_200; __rev1_200 = __builtin_shufflevector(__s1_200, __s1_200, 3, 2, 1, 0); \ + float32x4_t __rev2_200; __rev2_200 = __builtin_shufflevector(__s2_200, __s2_200, 3, 2, 1, 0); \ + float32x4_t __ret_200; \ +float32x4_t __reint_200 = __rev2_200; \ +uint64x2_t __reint1_200 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_200, __p3_200), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_200, __p3_200)}; \ + __ret_200 = __noswap_vcmlaq_rot90_f32(__rev0_200, __rev1_200, *(float32x4_t *) &__reint1_200); \ + __ret_200 = __builtin_shufflevector(__ret_200, __ret_200, 3, 2, 1, 0); \ + __ret_200; \ +}) +#endif + #endif #if defined(__ARM_FEATURE_COMPLEX) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) #ifdef __LITTLE_ENDIAN__ @@ -40499,6 +41131,638 @@ __ai float16x8_t vcaddq_rot90_f16(float16x8_t __p0, float16x8_t __p1) { } #endif +#ifdef __LITTLE_ENDIAN__ +__ai float16x8_t vcmlaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#else +__ai float16x8_t vcmlaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40); + __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); + return __ret; +} +__ai float16x8_t __noswap_vcmlaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float16x4_t vcmla_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#else +__ai float16x4_t vcmla_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float16x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float16x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float16x4_t __noswap_vcmla_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_lane_f16(__p0_201, __p1_201, __p2_201, __p3_201) __extension__ ({ \ + float16x4_t __s0_201 = __p0_201; \ + float16x4_t __s1_201 = __p1_201; \ + float16x4_t __s2_201 = __p2_201; \ + float16x4_t __ret_201; \ +float16x4_t __reint_201 = __s2_201; \ +uint32x2_t __reint1_201 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_201, __p3_201), vget_lane_u32(*(uint32x2_t *) &__reint_201, __p3_201)}; \ + __ret_201 = vcmla_f16(__s0_201, __s1_201, *(float16x4_t *) &__reint1_201); \ + __ret_201; \ +}) +#else +#define vcmla_lane_f16(__p0_202, __p1_202, __p2_202, __p3_202) __extension__ ({ \ + float16x4_t __s0_202 = __p0_202; \ + float16x4_t __s1_202 = __p1_202; \ + float16x4_t __s2_202 = __p2_202; \ + float16x4_t __rev0_202; __rev0_202 = __builtin_shufflevector(__s0_202, __s0_202, 3, 2, 1, 0); \ + float16x4_t __rev1_202; __rev1_202 = __builtin_shufflevector(__s1_202, __s1_202, 3, 2, 1, 0); \ + float16x4_t __rev2_202; __rev2_202 = __builtin_shufflevector(__s2_202, __s2_202, 3, 2, 1, 0); \ + float16x4_t __ret_202; \ +float16x4_t __reint_202 = __rev2_202; \ +uint32x2_t __reint1_202 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_202, __p3_202), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_202, __p3_202)}; \ + __ret_202 = __noswap_vcmla_f16(__rev0_202, __rev1_202, *(float16x4_t *) &__reint1_202); \ + __ret_202 = __builtin_shufflevector(__ret_202, __ret_202, 3, 2, 1, 0); \ + __ret_202; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_lane_f16(__p0_203, __p1_203, __p2_203, __p3_203) __extension__ ({ \ + float16x8_t __s0_203 = __p0_203; \ + float16x8_t __s1_203 = __p1_203; \ + float16x4_t __s2_203 = __p2_203; \ + float16x8_t __ret_203; \ +float16x4_t __reint_203 = __s2_203; \ +uint32x4_t __reint1_203 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_203, __p3_203), vget_lane_u32(*(uint32x2_t *) &__reint_203, __p3_203), vget_lane_u32(*(uint32x2_t *) &__reint_203, __p3_203), vget_lane_u32(*(uint32x2_t *) &__reint_203, __p3_203)}; \ + __ret_203 = vcmlaq_f16(__s0_203, __s1_203, *(float16x8_t *) &__reint1_203); \ + __ret_203; \ +}) +#else +#define vcmlaq_lane_f16(__p0_204, __p1_204, __p2_204, __p3_204) __extension__ ({ \ + float16x8_t __s0_204 = __p0_204; \ + float16x8_t __s1_204 = __p1_204; \ + float16x4_t __s2_204 = __p2_204; \ + float16x8_t __rev0_204; __rev0_204 = __builtin_shufflevector(__s0_204, __s0_204, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_204; __rev1_204 = __builtin_shufflevector(__s1_204, __s1_204, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_204; __rev2_204 = __builtin_shufflevector(__s2_204, __s2_204, 3, 2, 1, 0); \ + float16x8_t __ret_204; \ +float16x4_t __reint_204 = __rev2_204; \ +uint32x4_t __reint1_204 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_204, __p3_204), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_204, __p3_204), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_204, __p3_204), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_204, __p3_204)}; \ + __ret_204 = __noswap_vcmlaq_f16(__rev0_204, __rev1_204, *(float16x8_t *) &__reint1_204); \ + __ret_204 = __builtin_shufflevector(__ret_204, __ret_204, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_204; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_laneq_f16(__p0_205, __p1_205, __p2_205, __p3_205) __extension__ ({ \ + float16x4_t __s0_205 = __p0_205; \ + float16x4_t __s1_205 = __p1_205; \ + float16x8_t __s2_205 = __p2_205; \ + float16x4_t __ret_205; \ +float16x8_t __reint_205 = __s2_205; \ +uint32x2_t __reint1_205 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_205, __p3_205), vgetq_lane_u32(*(uint32x4_t *) &__reint_205, __p3_205)}; \ + __ret_205 = vcmla_f16(__s0_205, __s1_205, *(float16x4_t *) &__reint1_205); \ + __ret_205; \ +}) +#else +#define vcmla_laneq_f16(__p0_206, __p1_206, __p2_206, __p3_206) __extension__ ({ \ + float16x4_t __s0_206 = __p0_206; \ + float16x4_t __s1_206 = __p1_206; \ + float16x8_t __s2_206 = __p2_206; \ + float16x4_t __rev0_206; __rev0_206 = __builtin_shufflevector(__s0_206, __s0_206, 3, 2, 1, 0); \ + float16x4_t __rev1_206; __rev1_206 = __builtin_shufflevector(__s1_206, __s1_206, 3, 2, 1, 0); \ + float16x8_t __rev2_206; __rev2_206 = __builtin_shufflevector(__s2_206, __s2_206, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_206; \ +float16x8_t __reint_206 = __rev2_206; \ +uint32x2_t __reint1_206 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_206, __p3_206), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_206, __p3_206)}; \ + __ret_206 = __noswap_vcmla_f16(__rev0_206, __rev1_206, *(float16x4_t *) &__reint1_206); \ + __ret_206 = __builtin_shufflevector(__ret_206, __ret_206, 3, 2, 1, 0); \ + __ret_206; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_laneq_f16(__p0_207, __p1_207, __p2_207, __p3_207) __extension__ ({ \ + float16x8_t __s0_207 = __p0_207; \ + float16x8_t __s1_207 = __p1_207; \ + float16x8_t __s2_207 = __p2_207; \ + float16x8_t __ret_207; \ +float16x8_t __reint_207 = __s2_207; \ +uint32x4_t __reint1_207 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_207, __p3_207), vgetq_lane_u32(*(uint32x4_t *) &__reint_207, __p3_207), vgetq_lane_u32(*(uint32x4_t *) &__reint_207, __p3_207), vgetq_lane_u32(*(uint32x4_t *) &__reint_207, __p3_207)}; \ + __ret_207 = vcmlaq_f16(__s0_207, __s1_207, *(float16x8_t *) &__reint1_207); \ + __ret_207; \ +}) +#else +#define vcmlaq_laneq_f16(__p0_208, __p1_208, __p2_208, __p3_208) __extension__ ({ \ + float16x8_t __s0_208 = __p0_208; \ + float16x8_t __s1_208 = __p1_208; \ + float16x8_t __s2_208 = __p2_208; \ + float16x8_t __rev0_208; __rev0_208 = __builtin_shufflevector(__s0_208, __s0_208, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_208; __rev1_208 = __builtin_shufflevector(__s1_208, __s1_208, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_208; __rev2_208 = __builtin_shufflevector(__s2_208, __s2_208, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_208; \ +float16x8_t __reint_208 = __rev2_208; \ +uint32x4_t __reint1_208 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_208, __p3_208), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_208, __p3_208), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_208, __p3_208), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_208, __p3_208)}; \ + __ret_208 = __noswap_vcmlaq_f16(__rev0_208, __rev1_208, *(float16x8_t *) &__reint1_208); \ + __ret_208 = __builtin_shufflevector(__ret_208, __ret_208, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_208; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float16x8_t vcmlaq_rot180_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#else +__ai float16x8_t vcmlaq_rot180_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40); + __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); + return __ret; +} +__ai float16x8_t __noswap_vcmlaq_rot180_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float16x4_t vcmla_rot180_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot180_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#else +__ai float16x4_t vcmla_rot180_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float16x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float16x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot180_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float16x4_t __noswap_vcmla_rot180_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot180_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot180_lane_f16(__p0_209, __p1_209, __p2_209, __p3_209) __extension__ ({ \ + float16x4_t __s0_209 = __p0_209; \ + float16x4_t __s1_209 = __p1_209; \ + float16x4_t __s2_209 = __p2_209; \ + float16x4_t __ret_209; \ +float16x4_t __reint_209 = __s2_209; \ +uint32x2_t __reint1_209 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_209, __p3_209), vget_lane_u32(*(uint32x2_t *) &__reint_209, __p3_209)}; \ + __ret_209 = vcmla_rot180_f16(__s0_209, __s1_209, *(float16x4_t *) &__reint1_209); \ + __ret_209; \ +}) +#else +#define vcmla_rot180_lane_f16(__p0_210, __p1_210, __p2_210, __p3_210) __extension__ ({ \ + float16x4_t __s0_210 = __p0_210; \ + float16x4_t __s1_210 = __p1_210; \ + float16x4_t __s2_210 = __p2_210; \ + float16x4_t __rev0_210; __rev0_210 = __builtin_shufflevector(__s0_210, __s0_210, 3, 2, 1, 0); \ + float16x4_t __rev1_210; __rev1_210 = __builtin_shufflevector(__s1_210, __s1_210, 3, 2, 1, 0); \ + float16x4_t __rev2_210; __rev2_210 = __builtin_shufflevector(__s2_210, __s2_210, 3, 2, 1, 0); \ + float16x4_t __ret_210; \ +float16x4_t __reint_210 = __rev2_210; \ +uint32x2_t __reint1_210 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_210, __p3_210), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_210, __p3_210)}; \ + __ret_210 = __noswap_vcmla_rot180_f16(__rev0_210, __rev1_210, *(float16x4_t *) &__reint1_210); \ + __ret_210 = __builtin_shufflevector(__ret_210, __ret_210, 3, 2, 1, 0); \ + __ret_210; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot180_lane_f16(__p0_211, __p1_211, __p2_211, __p3_211) __extension__ ({ \ + float16x8_t __s0_211 = __p0_211; \ + float16x8_t __s1_211 = __p1_211; \ + float16x4_t __s2_211 = __p2_211; \ + float16x8_t __ret_211; \ +float16x4_t __reint_211 = __s2_211; \ +uint32x4_t __reint1_211 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211), vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211), vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211), vget_lane_u32(*(uint32x2_t *) &__reint_211, __p3_211)}; \ + __ret_211 = vcmlaq_rot180_f16(__s0_211, __s1_211, *(float16x8_t *) &__reint1_211); \ + __ret_211; \ +}) +#else +#define vcmlaq_rot180_lane_f16(__p0_212, __p1_212, __p2_212, __p3_212) __extension__ ({ \ + float16x8_t __s0_212 = __p0_212; \ + float16x8_t __s1_212 = __p1_212; \ + float16x4_t __s2_212 = __p2_212; \ + float16x8_t __rev0_212; __rev0_212 = __builtin_shufflevector(__s0_212, __s0_212, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_212; __rev1_212 = __builtin_shufflevector(__s1_212, __s1_212, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_212; __rev2_212 = __builtin_shufflevector(__s2_212, __s2_212, 3, 2, 1, 0); \ + float16x8_t __ret_212; \ +float16x4_t __reint_212 = __rev2_212; \ +uint32x4_t __reint1_212 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_212, __p3_212)}; \ + __ret_212 = __noswap_vcmlaq_rot180_f16(__rev0_212, __rev1_212, *(float16x8_t *) &__reint1_212); \ + __ret_212 = __builtin_shufflevector(__ret_212, __ret_212, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_212; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot180_laneq_f16(__p0_213, __p1_213, __p2_213, __p3_213) __extension__ ({ \ + float16x4_t __s0_213 = __p0_213; \ + float16x4_t __s1_213 = __p1_213; \ + float16x8_t __s2_213 = __p2_213; \ + float16x4_t __ret_213; \ +float16x8_t __reint_213 = __s2_213; \ +uint32x2_t __reint1_213 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_213, __p3_213), vgetq_lane_u32(*(uint32x4_t *) &__reint_213, __p3_213)}; \ + __ret_213 = vcmla_rot180_f16(__s0_213, __s1_213, *(float16x4_t *) &__reint1_213); \ + __ret_213; \ +}) +#else +#define vcmla_rot180_laneq_f16(__p0_214, __p1_214, __p2_214, __p3_214) __extension__ ({ \ + float16x4_t __s0_214 = __p0_214; \ + float16x4_t __s1_214 = __p1_214; \ + float16x8_t __s2_214 = __p2_214; \ + float16x4_t __rev0_214; __rev0_214 = __builtin_shufflevector(__s0_214, __s0_214, 3, 2, 1, 0); \ + float16x4_t __rev1_214; __rev1_214 = __builtin_shufflevector(__s1_214, __s1_214, 3, 2, 1, 0); \ + float16x8_t __rev2_214; __rev2_214 = __builtin_shufflevector(__s2_214, __s2_214, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_214; \ +float16x8_t __reint_214 = __rev2_214; \ +uint32x2_t __reint1_214 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_214, __p3_214), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_214, __p3_214)}; \ + __ret_214 = __noswap_vcmla_rot180_f16(__rev0_214, __rev1_214, *(float16x4_t *) &__reint1_214); \ + __ret_214 = __builtin_shufflevector(__ret_214, __ret_214, 3, 2, 1, 0); \ + __ret_214; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot180_laneq_f16(__p0_215, __p1_215, __p2_215, __p3_215) __extension__ ({ \ + float16x8_t __s0_215 = __p0_215; \ + float16x8_t __s1_215 = __p1_215; \ + float16x8_t __s2_215 = __p2_215; \ + float16x8_t __ret_215; \ +float16x8_t __reint_215 = __s2_215; \ +uint32x4_t __reint1_215 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215), vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215), vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215), vgetq_lane_u32(*(uint32x4_t *) &__reint_215, __p3_215)}; \ + __ret_215 = vcmlaq_rot180_f16(__s0_215, __s1_215, *(float16x8_t *) &__reint1_215); \ + __ret_215; \ +}) +#else +#define vcmlaq_rot180_laneq_f16(__p0_216, __p1_216, __p2_216, __p3_216) __extension__ ({ \ + float16x8_t __s0_216 = __p0_216; \ + float16x8_t __s1_216 = __p1_216; \ + float16x8_t __s2_216 = __p2_216; \ + float16x8_t __rev0_216; __rev0_216 = __builtin_shufflevector(__s0_216, __s0_216, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_216; __rev1_216 = __builtin_shufflevector(__s1_216, __s1_216, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_216; __rev2_216 = __builtin_shufflevector(__s2_216, __s2_216, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_216; \ +float16x8_t __reint_216 = __rev2_216; \ +uint32x4_t __reint1_216 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_216, __p3_216)}; \ + __ret_216 = __noswap_vcmlaq_rot180_f16(__rev0_216, __rev1_216, *(float16x8_t *) &__reint1_216); \ + __ret_216 = __builtin_shufflevector(__ret_216, __ret_216, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_216; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float16x8_t vcmlaq_rot270_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#else +__ai float16x8_t vcmlaq_rot270_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40); + __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); + return __ret; +} +__ai float16x8_t __noswap_vcmlaq_rot270_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float16x4_t vcmla_rot270_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot270_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#else +__ai float16x4_t vcmla_rot270_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float16x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float16x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot270_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float16x4_t __noswap_vcmla_rot270_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot270_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot270_lane_f16(__p0_217, __p1_217, __p2_217, __p3_217) __extension__ ({ \ + float16x4_t __s0_217 = __p0_217; \ + float16x4_t __s1_217 = __p1_217; \ + float16x4_t __s2_217 = __p2_217; \ + float16x4_t __ret_217; \ +float16x4_t __reint_217 = __s2_217; \ +uint32x2_t __reint1_217 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_217, __p3_217), vget_lane_u32(*(uint32x2_t *) &__reint_217, __p3_217)}; \ + __ret_217 = vcmla_rot270_f16(__s0_217, __s1_217, *(float16x4_t *) &__reint1_217); \ + __ret_217; \ +}) +#else +#define vcmla_rot270_lane_f16(__p0_218, __p1_218, __p2_218, __p3_218) __extension__ ({ \ + float16x4_t __s0_218 = __p0_218; \ + float16x4_t __s1_218 = __p1_218; \ + float16x4_t __s2_218 = __p2_218; \ + float16x4_t __rev0_218; __rev0_218 = __builtin_shufflevector(__s0_218, __s0_218, 3, 2, 1, 0); \ + float16x4_t __rev1_218; __rev1_218 = __builtin_shufflevector(__s1_218, __s1_218, 3, 2, 1, 0); \ + float16x4_t __rev2_218; __rev2_218 = __builtin_shufflevector(__s2_218, __s2_218, 3, 2, 1, 0); \ + float16x4_t __ret_218; \ +float16x4_t __reint_218 = __rev2_218; \ +uint32x2_t __reint1_218 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_218, __p3_218), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_218, __p3_218)}; \ + __ret_218 = __noswap_vcmla_rot270_f16(__rev0_218, __rev1_218, *(float16x4_t *) &__reint1_218); \ + __ret_218 = __builtin_shufflevector(__ret_218, __ret_218, 3, 2, 1, 0); \ + __ret_218; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot270_lane_f16(__p0_219, __p1_219, __p2_219, __p3_219) __extension__ ({ \ + float16x8_t __s0_219 = __p0_219; \ + float16x8_t __s1_219 = __p1_219; \ + float16x4_t __s2_219 = __p2_219; \ + float16x8_t __ret_219; \ +float16x4_t __reint_219 = __s2_219; \ +uint32x4_t __reint1_219 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219), vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219), vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219), vget_lane_u32(*(uint32x2_t *) &__reint_219, __p3_219)}; \ + __ret_219 = vcmlaq_rot270_f16(__s0_219, __s1_219, *(float16x8_t *) &__reint1_219); \ + __ret_219; \ +}) +#else +#define vcmlaq_rot270_lane_f16(__p0_220, __p1_220, __p2_220, __p3_220) __extension__ ({ \ + float16x8_t __s0_220 = __p0_220; \ + float16x8_t __s1_220 = __p1_220; \ + float16x4_t __s2_220 = __p2_220; \ + float16x8_t __rev0_220; __rev0_220 = __builtin_shufflevector(__s0_220, __s0_220, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_220; __rev1_220 = __builtin_shufflevector(__s1_220, __s1_220, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_220; __rev2_220 = __builtin_shufflevector(__s2_220, __s2_220, 3, 2, 1, 0); \ + float16x8_t __ret_220; \ +float16x4_t __reint_220 = __rev2_220; \ +uint32x4_t __reint1_220 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_220, __p3_220)}; \ + __ret_220 = __noswap_vcmlaq_rot270_f16(__rev0_220, __rev1_220, *(float16x8_t *) &__reint1_220); \ + __ret_220 = __builtin_shufflevector(__ret_220, __ret_220, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_220; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot270_laneq_f16(__p0_221, __p1_221, __p2_221, __p3_221) __extension__ ({ \ + float16x4_t __s0_221 = __p0_221; \ + float16x4_t __s1_221 = __p1_221; \ + float16x8_t __s2_221 = __p2_221; \ + float16x4_t __ret_221; \ +float16x8_t __reint_221 = __s2_221; \ +uint32x2_t __reint1_221 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_221, __p3_221), vgetq_lane_u32(*(uint32x4_t *) &__reint_221, __p3_221)}; \ + __ret_221 = vcmla_rot270_f16(__s0_221, __s1_221, *(float16x4_t *) &__reint1_221); \ + __ret_221; \ +}) +#else +#define vcmla_rot270_laneq_f16(__p0_222, __p1_222, __p2_222, __p3_222) __extension__ ({ \ + float16x4_t __s0_222 = __p0_222; \ + float16x4_t __s1_222 = __p1_222; \ + float16x8_t __s2_222 = __p2_222; \ + float16x4_t __rev0_222; __rev0_222 = __builtin_shufflevector(__s0_222, __s0_222, 3, 2, 1, 0); \ + float16x4_t __rev1_222; __rev1_222 = __builtin_shufflevector(__s1_222, __s1_222, 3, 2, 1, 0); \ + float16x8_t __rev2_222; __rev2_222 = __builtin_shufflevector(__s2_222, __s2_222, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_222; \ +float16x8_t __reint_222 = __rev2_222; \ +uint32x2_t __reint1_222 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_222, __p3_222), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_222, __p3_222)}; \ + __ret_222 = __noswap_vcmla_rot270_f16(__rev0_222, __rev1_222, *(float16x4_t *) &__reint1_222); \ + __ret_222 = __builtin_shufflevector(__ret_222, __ret_222, 3, 2, 1, 0); \ + __ret_222; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot270_laneq_f16(__p0_223, __p1_223, __p2_223, __p3_223) __extension__ ({ \ + float16x8_t __s0_223 = __p0_223; \ + float16x8_t __s1_223 = __p1_223; \ + float16x8_t __s2_223 = __p2_223; \ + float16x8_t __ret_223; \ +float16x8_t __reint_223 = __s2_223; \ +uint32x4_t __reint1_223 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223), vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223), vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223), vgetq_lane_u32(*(uint32x4_t *) &__reint_223, __p3_223)}; \ + __ret_223 = vcmlaq_rot270_f16(__s0_223, __s1_223, *(float16x8_t *) &__reint1_223); \ + __ret_223; \ +}) +#else +#define vcmlaq_rot270_laneq_f16(__p0_224, __p1_224, __p2_224, __p3_224) __extension__ ({ \ + float16x8_t __s0_224 = __p0_224; \ + float16x8_t __s1_224 = __p1_224; \ + float16x8_t __s2_224 = __p2_224; \ + float16x8_t __rev0_224; __rev0_224 = __builtin_shufflevector(__s0_224, __s0_224, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_224; __rev1_224 = __builtin_shufflevector(__s1_224, __s1_224, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_224; __rev2_224 = __builtin_shufflevector(__s2_224, __s2_224, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_224; \ +float16x8_t __reint_224 = __rev2_224; \ +uint32x4_t __reint1_224 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_224, __p3_224)}; \ + __ret_224 = __noswap_vcmlaq_rot270_f16(__rev0_224, __rev1_224, *(float16x8_t *) &__reint1_224); \ + __ret_224 = __builtin_shufflevector(__ret_224, __ret_224, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_224; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float16x8_t vcmlaq_rot90_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#else +__ai float16x8_t vcmlaq_rot90_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0); + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40); + __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); + return __ret; +} +__ai float16x8_t __noswap_vcmlaq_rot90_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) { + float16x8_t __ret; + __ret = (float16x8_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float16x4_t vcmla_rot90_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot90_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#else +__ai float16x4_t vcmla_rot90_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); + float16x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); + float16x4_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0); + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot90_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8); + __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); + return __ret; +} +__ai float16x4_t __noswap_vcmla_rot90_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) { + float16x4_t __ret; + __ret = (float16x4_t) __builtin_neon_vcmla_rot90_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8); + return __ret; +} +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot90_lane_f16(__p0_225, __p1_225, __p2_225, __p3_225) __extension__ ({ \ + float16x4_t __s0_225 = __p0_225; \ + float16x4_t __s1_225 = __p1_225; \ + float16x4_t __s2_225 = __p2_225; \ + float16x4_t __ret_225; \ +float16x4_t __reint_225 = __s2_225; \ +uint32x2_t __reint1_225 = (uint32x2_t) {vget_lane_u32(*(uint32x2_t *) &__reint_225, __p3_225), vget_lane_u32(*(uint32x2_t *) &__reint_225, __p3_225)}; \ + __ret_225 = vcmla_rot90_f16(__s0_225, __s1_225, *(float16x4_t *) &__reint1_225); \ + __ret_225; \ +}) +#else +#define vcmla_rot90_lane_f16(__p0_226, __p1_226, __p2_226, __p3_226) __extension__ ({ \ + float16x4_t __s0_226 = __p0_226; \ + float16x4_t __s1_226 = __p1_226; \ + float16x4_t __s2_226 = __p2_226; \ + float16x4_t __rev0_226; __rev0_226 = __builtin_shufflevector(__s0_226, __s0_226, 3, 2, 1, 0); \ + float16x4_t __rev1_226; __rev1_226 = __builtin_shufflevector(__s1_226, __s1_226, 3, 2, 1, 0); \ + float16x4_t __rev2_226; __rev2_226 = __builtin_shufflevector(__s2_226, __s2_226, 3, 2, 1, 0); \ + float16x4_t __ret_226; \ +float16x4_t __reint_226 = __rev2_226; \ +uint32x2_t __reint1_226 = (uint32x2_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_226, __p3_226), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_226, __p3_226)}; \ + __ret_226 = __noswap_vcmla_rot90_f16(__rev0_226, __rev1_226, *(float16x4_t *) &__reint1_226); \ + __ret_226 = __builtin_shufflevector(__ret_226, __ret_226, 3, 2, 1, 0); \ + __ret_226; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot90_lane_f16(__p0_227, __p1_227, __p2_227, __p3_227) __extension__ ({ \ + float16x8_t __s0_227 = __p0_227; \ + float16x8_t __s1_227 = __p1_227; \ + float16x4_t __s2_227 = __p2_227; \ + float16x8_t __ret_227; \ +float16x4_t __reint_227 = __s2_227; \ +uint32x4_t __reint1_227 = (uint32x4_t) {vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227), vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227), vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227), vget_lane_u32(*(uint32x2_t *) &__reint_227, __p3_227)}; \ + __ret_227 = vcmlaq_rot90_f16(__s0_227, __s1_227, *(float16x8_t *) &__reint1_227); \ + __ret_227; \ +}) +#else +#define vcmlaq_rot90_lane_f16(__p0_228, __p1_228, __p2_228, __p3_228) __extension__ ({ \ + float16x8_t __s0_228 = __p0_228; \ + float16x8_t __s1_228 = __p1_228; \ + float16x4_t __s2_228 = __p2_228; \ + float16x8_t __rev0_228; __rev0_228 = __builtin_shufflevector(__s0_228, __s0_228, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_228; __rev1_228 = __builtin_shufflevector(__s1_228, __s1_228, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_228; __rev2_228 = __builtin_shufflevector(__s2_228, __s2_228, 3, 2, 1, 0); \ + float16x8_t __ret_228; \ +float16x4_t __reint_228 = __rev2_228; \ +uint32x4_t __reint1_228 = (uint32x4_t) {__noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228), __noswap_vget_lane_u32(*(uint32x2_t *) &__reint_228, __p3_228)}; \ + __ret_228 = __noswap_vcmlaq_rot90_f16(__rev0_228, __rev1_228, *(float16x8_t *) &__reint1_228); \ + __ret_228 = __builtin_shufflevector(__ret_228, __ret_228, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_228; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot90_laneq_f16(__p0_229, __p1_229, __p2_229, __p3_229) __extension__ ({ \ + float16x4_t __s0_229 = __p0_229; \ + float16x4_t __s1_229 = __p1_229; \ + float16x8_t __s2_229 = __p2_229; \ + float16x4_t __ret_229; \ +float16x8_t __reint_229 = __s2_229; \ +uint32x2_t __reint1_229 = (uint32x2_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_229, __p3_229), vgetq_lane_u32(*(uint32x4_t *) &__reint_229, __p3_229)}; \ + __ret_229 = vcmla_rot90_f16(__s0_229, __s1_229, *(float16x4_t *) &__reint1_229); \ + __ret_229; \ +}) +#else +#define vcmla_rot90_laneq_f16(__p0_230, __p1_230, __p2_230, __p3_230) __extension__ ({ \ + float16x4_t __s0_230 = __p0_230; \ + float16x4_t __s1_230 = __p1_230; \ + float16x8_t __s2_230 = __p2_230; \ + float16x4_t __rev0_230; __rev0_230 = __builtin_shufflevector(__s0_230, __s0_230, 3, 2, 1, 0); \ + float16x4_t __rev1_230; __rev1_230 = __builtin_shufflevector(__s1_230, __s1_230, 3, 2, 1, 0); \ + float16x8_t __rev2_230; __rev2_230 = __builtin_shufflevector(__s2_230, __s2_230, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_230; \ +float16x8_t __reint_230 = __rev2_230; \ +uint32x2_t __reint1_230 = (uint32x2_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_230, __p3_230), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_230, __p3_230)}; \ + __ret_230 = __noswap_vcmla_rot90_f16(__rev0_230, __rev1_230, *(float16x4_t *) &__reint1_230); \ + __ret_230 = __builtin_shufflevector(__ret_230, __ret_230, 3, 2, 1, 0); \ + __ret_230; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot90_laneq_f16(__p0_231, __p1_231, __p2_231, __p3_231) __extension__ ({ \ + float16x8_t __s0_231 = __p0_231; \ + float16x8_t __s1_231 = __p1_231; \ + float16x8_t __s2_231 = __p2_231; \ + float16x8_t __ret_231; \ +float16x8_t __reint_231 = __s2_231; \ +uint32x4_t __reint1_231 = (uint32x4_t) {vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231), vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231), vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231), vgetq_lane_u32(*(uint32x4_t *) &__reint_231, __p3_231)}; \ + __ret_231 = vcmlaq_rot90_f16(__s0_231, __s1_231, *(float16x8_t *) &__reint1_231); \ + __ret_231; \ +}) +#else +#define vcmlaq_rot90_laneq_f16(__p0_232, __p1_232, __p2_232, __p3_232) __extension__ ({ \ + float16x8_t __s0_232 = __p0_232; \ + float16x8_t __s1_232 = __p1_232; \ + float16x8_t __s2_232 = __p2_232; \ + float16x8_t __rev0_232; __rev0_232 = __builtin_shufflevector(__s0_232, __s0_232, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_232; __rev1_232 = __builtin_shufflevector(__s1_232, __s1_232, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_232; __rev2_232 = __builtin_shufflevector(__s2_232, __s2_232, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_232; \ +float16x8_t __reint_232 = __rev2_232; \ +uint32x4_t __reint1_232 = (uint32x4_t) {__noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232), __noswap_vgetq_lane_u32(*(uint32x4_t *) &__reint_232, __p3_232)}; \ + __ret_232 = __noswap_vcmlaq_rot90_f16(__rev0_232, __rev1_232, *(float16x8_t *) &__reint1_232); \ + __ret_232 = __builtin_shufflevector(__ret_232, __ret_232, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_232; \ +}) +#endif + #endif #if defined(__ARM_FEATURE_COMPLEX) && defined(__aarch64__) #ifdef __LITTLE_ENDIAN__ @@ -40535,6 +41799,478 @@ __ai float64x2_t vcaddq_rot90_f64(float64x2_t __p0, float64x2_t __p1) { } #endif +#ifdef __LITTLE_ENDIAN__ +__ai float64x2_t vcmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#else +__ai float64x2_t vcmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float64x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float64x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float64x2_t __noswap_vcmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#endif + +__ai float64x1_t vcmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) { + float64x1_t __ret; + __ret = (float64x1_t) __builtin_neon_vcmla_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10); + return __ret; +} +#define vcmla_lane_f64(__p0_233, __p1_233, __p2_233, __p3_233) __extension__ ({ \ + float64x1_t __s0_233 = __p0_233; \ + float64x1_t __s1_233 = __p1_233; \ + float64x1_t __s2_233 = __p2_233; \ + float64x1_t __ret_233; \ +float64x1_t __reint_233 = __s2_233; \ +uint64x2_t __reint1_233 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_233, __p3_233), vgetq_lane_u64(*(uint64x2_t *) &__reint_233, __p3_233)}; \ + __ret_233 = vcmla_f64(__s0_233, __s1_233, *(float64x1_t *) &__reint1_233); \ + __ret_233; \ +}) +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_lane_f64(__p0_234, __p1_234, __p2_234, __p3_234) __extension__ ({ \ + float64x2_t __s0_234 = __p0_234; \ + float64x2_t __s1_234 = __p1_234; \ + float64x1_t __s2_234 = __p2_234; \ + float64x2_t __ret_234; \ +float64x1_t __reint_234 = __s2_234; \ +uint64x2_t __reint1_234 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_234, __p3_234), vgetq_lane_u64(*(uint64x2_t *) &__reint_234, __p3_234)}; \ + __ret_234 = vcmlaq_f64(__s0_234, __s1_234, *(float64x2_t *) &__reint1_234); \ + __ret_234; \ +}) +#else +#define vcmlaq_lane_f64(__p0_235, __p1_235, __p2_235, __p3_235) __extension__ ({ \ + float64x2_t __s0_235 = __p0_235; \ + float64x2_t __s1_235 = __p1_235; \ + float64x1_t __s2_235 = __p2_235; \ + float64x2_t __rev0_235; __rev0_235 = __builtin_shufflevector(__s0_235, __s0_235, 1, 0); \ + float64x2_t __rev1_235; __rev1_235 = __builtin_shufflevector(__s1_235, __s1_235, 1, 0); \ + float64x2_t __ret_235; \ +float64x1_t __reint_235 = __s2_235; \ +uint64x2_t __reint1_235 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_235, __p3_235), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_235, __p3_235)}; \ + __ret_235 = __noswap_vcmlaq_f64(__rev0_235, __rev1_235, *(float64x2_t *) &__reint1_235); \ + __ret_235 = __builtin_shufflevector(__ret_235, __ret_235, 1, 0); \ + __ret_235; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_laneq_f64(__p0_236, __p1_236, __p2_236, __p3_236) __extension__ ({ \ + float64x1_t __s0_236 = __p0_236; \ + float64x1_t __s1_236 = __p1_236; \ + float64x2_t __s2_236 = __p2_236; \ + float64x1_t __ret_236; \ +float64x2_t __reint_236 = __s2_236; \ +uint64x2_t __reint1_236 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_236, __p3_236), vgetq_lane_u64(*(uint64x2_t *) &__reint_236, __p3_236)}; \ + __ret_236 = vcmla_f64(__s0_236, __s1_236, *(float64x1_t *) &__reint1_236); \ + __ret_236; \ +}) +#else +#define vcmla_laneq_f64(__p0_237, __p1_237, __p2_237, __p3_237) __extension__ ({ \ + float64x1_t __s0_237 = __p0_237; \ + float64x1_t __s1_237 = __p1_237; \ + float64x2_t __s2_237 = __p2_237; \ + float64x2_t __rev2_237; __rev2_237 = __builtin_shufflevector(__s2_237, __s2_237, 1, 0); \ + float64x1_t __ret_237; \ +float64x2_t __reint_237 = __rev2_237; \ +uint64x2_t __reint1_237 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_237, __p3_237), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_237, __p3_237)}; \ + __ret_237 = vcmla_f64(__s0_237, __s1_237, *(float64x1_t *) &__reint1_237); \ + __ret_237; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_laneq_f64(__p0_238, __p1_238, __p2_238, __p3_238) __extension__ ({ \ + float64x2_t __s0_238 = __p0_238; \ + float64x2_t __s1_238 = __p1_238; \ + float64x2_t __s2_238 = __p2_238; \ + float64x2_t __ret_238; \ +float64x2_t __reint_238 = __s2_238; \ +uint64x2_t __reint1_238 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_238, __p3_238), vgetq_lane_u64(*(uint64x2_t *) &__reint_238, __p3_238)}; \ + __ret_238 = vcmlaq_f64(__s0_238, __s1_238, *(float64x2_t *) &__reint1_238); \ + __ret_238; \ +}) +#else +#define vcmlaq_laneq_f64(__p0_239, __p1_239, __p2_239, __p3_239) __extension__ ({ \ + float64x2_t __s0_239 = __p0_239; \ + float64x2_t __s1_239 = __p1_239; \ + float64x2_t __s2_239 = __p2_239; \ + float64x2_t __rev0_239; __rev0_239 = __builtin_shufflevector(__s0_239, __s0_239, 1, 0); \ + float64x2_t __rev1_239; __rev1_239 = __builtin_shufflevector(__s1_239, __s1_239, 1, 0); \ + float64x2_t __rev2_239; __rev2_239 = __builtin_shufflevector(__s2_239, __s2_239, 1, 0); \ + float64x2_t __ret_239; \ +float64x2_t __reint_239 = __rev2_239; \ +uint64x2_t __reint1_239 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_239, __p3_239), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_239, __p3_239)}; \ + __ret_239 = __noswap_vcmlaq_f64(__rev0_239, __rev1_239, *(float64x2_t *) &__reint1_239); \ + __ret_239 = __builtin_shufflevector(__ret_239, __ret_239, 1, 0); \ + __ret_239; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float64x2_t vcmlaq_rot180_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#else +__ai float64x2_t vcmlaq_rot180_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float64x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float64x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float64x2_t __noswap_vcmlaq_rot180_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot180_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#endif + +__ai float64x1_t vcmla_rot180_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) { + float64x1_t __ret; + __ret = (float64x1_t) __builtin_neon_vcmla_rot180_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10); + return __ret; +} +#define vcmla_rot180_lane_f64(__p0_240, __p1_240, __p2_240, __p3_240) __extension__ ({ \ + float64x1_t __s0_240 = __p0_240; \ + float64x1_t __s1_240 = __p1_240; \ + float64x1_t __s2_240 = __p2_240; \ + float64x1_t __ret_240; \ +float64x1_t __reint_240 = __s2_240; \ +uint64x2_t __reint1_240 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_240, __p3_240), vgetq_lane_u64(*(uint64x2_t *) &__reint_240, __p3_240)}; \ + __ret_240 = vcmla_rot180_f64(__s0_240, __s1_240, *(float64x1_t *) &__reint1_240); \ + __ret_240; \ +}) +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot180_lane_f64(__p0_241, __p1_241, __p2_241, __p3_241) __extension__ ({ \ + float64x2_t __s0_241 = __p0_241; \ + float64x2_t __s1_241 = __p1_241; \ + float64x1_t __s2_241 = __p2_241; \ + float64x2_t __ret_241; \ +float64x1_t __reint_241 = __s2_241; \ +uint64x2_t __reint1_241 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_241, __p3_241), vgetq_lane_u64(*(uint64x2_t *) &__reint_241, __p3_241)}; \ + __ret_241 = vcmlaq_rot180_f64(__s0_241, __s1_241, *(float64x2_t *) &__reint1_241); \ + __ret_241; \ +}) +#else +#define vcmlaq_rot180_lane_f64(__p0_242, __p1_242, __p2_242, __p3_242) __extension__ ({ \ + float64x2_t __s0_242 = __p0_242; \ + float64x2_t __s1_242 = __p1_242; \ + float64x1_t __s2_242 = __p2_242; \ + float64x2_t __rev0_242; __rev0_242 = __builtin_shufflevector(__s0_242, __s0_242, 1, 0); \ + float64x2_t __rev1_242; __rev1_242 = __builtin_shufflevector(__s1_242, __s1_242, 1, 0); \ + float64x2_t __ret_242; \ +float64x1_t __reint_242 = __s2_242; \ +uint64x2_t __reint1_242 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_242, __p3_242), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_242, __p3_242)}; \ + __ret_242 = __noswap_vcmlaq_rot180_f64(__rev0_242, __rev1_242, *(float64x2_t *) &__reint1_242); \ + __ret_242 = __builtin_shufflevector(__ret_242, __ret_242, 1, 0); \ + __ret_242; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot180_laneq_f64(__p0_243, __p1_243, __p2_243, __p3_243) __extension__ ({ \ + float64x1_t __s0_243 = __p0_243; \ + float64x1_t __s1_243 = __p1_243; \ + float64x2_t __s2_243 = __p2_243; \ + float64x1_t __ret_243; \ +float64x2_t __reint_243 = __s2_243; \ +uint64x2_t __reint1_243 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_243, __p3_243), vgetq_lane_u64(*(uint64x2_t *) &__reint_243, __p3_243)}; \ + __ret_243 = vcmla_rot180_f64(__s0_243, __s1_243, *(float64x1_t *) &__reint1_243); \ + __ret_243; \ +}) +#else +#define vcmla_rot180_laneq_f64(__p0_244, __p1_244, __p2_244, __p3_244) __extension__ ({ \ + float64x1_t __s0_244 = __p0_244; \ + float64x1_t __s1_244 = __p1_244; \ + float64x2_t __s2_244 = __p2_244; \ + float64x2_t __rev2_244; __rev2_244 = __builtin_shufflevector(__s2_244, __s2_244, 1, 0); \ + float64x1_t __ret_244; \ +float64x2_t __reint_244 = __rev2_244; \ +uint64x2_t __reint1_244 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_244, __p3_244), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_244, __p3_244)}; \ + __ret_244 = vcmla_rot180_f64(__s0_244, __s1_244, *(float64x1_t *) &__reint1_244); \ + __ret_244; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot180_laneq_f64(__p0_245, __p1_245, __p2_245, __p3_245) __extension__ ({ \ + float64x2_t __s0_245 = __p0_245; \ + float64x2_t __s1_245 = __p1_245; \ + float64x2_t __s2_245 = __p2_245; \ + float64x2_t __ret_245; \ +float64x2_t __reint_245 = __s2_245; \ +uint64x2_t __reint1_245 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_245, __p3_245), vgetq_lane_u64(*(uint64x2_t *) &__reint_245, __p3_245)}; \ + __ret_245 = vcmlaq_rot180_f64(__s0_245, __s1_245, *(float64x2_t *) &__reint1_245); \ + __ret_245; \ +}) +#else +#define vcmlaq_rot180_laneq_f64(__p0_246, __p1_246, __p2_246, __p3_246) __extension__ ({ \ + float64x2_t __s0_246 = __p0_246; \ + float64x2_t __s1_246 = __p1_246; \ + float64x2_t __s2_246 = __p2_246; \ + float64x2_t __rev0_246; __rev0_246 = __builtin_shufflevector(__s0_246, __s0_246, 1, 0); \ + float64x2_t __rev1_246; __rev1_246 = __builtin_shufflevector(__s1_246, __s1_246, 1, 0); \ + float64x2_t __rev2_246; __rev2_246 = __builtin_shufflevector(__s2_246, __s2_246, 1, 0); \ + float64x2_t __ret_246; \ +float64x2_t __reint_246 = __rev2_246; \ +uint64x2_t __reint1_246 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_246, __p3_246), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_246, __p3_246)}; \ + __ret_246 = __noswap_vcmlaq_rot180_f64(__rev0_246, __rev1_246, *(float64x2_t *) &__reint1_246); \ + __ret_246 = __builtin_shufflevector(__ret_246, __ret_246, 1, 0); \ + __ret_246; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float64x2_t vcmlaq_rot270_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#else +__ai float64x2_t vcmlaq_rot270_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float64x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float64x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float64x2_t __noswap_vcmlaq_rot270_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot270_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#endif + +__ai float64x1_t vcmla_rot270_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) { + float64x1_t __ret; + __ret = (float64x1_t) __builtin_neon_vcmla_rot270_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10); + return __ret; +} +#define vcmla_rot270_lane_f64(__p0_247, __p1_247, __p2_247, __p3_247) __extension__ ({ \ + float64x1_t __s0_247 = __p0_247; \ + float64x1_t __s1_247 = __p1_247; \ + float64x1_t __s2_247 = __p2_247; \ + float64x1_t __ret_247; \ +float64x1_t __reint_247 = __s2_247; \ +uint64x2_t __reint1_247 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_247, __p3_247), vgetq_lane_u64(*(uint64x2_t *) &__reint_247, __p3_247)}; \ + __ret_247 = vcmla_rot270_f64(__s0_247, __s1_247, *(float64x1_t *) &__reint1_247); \ + __ret_247; \ +}) +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot270_lane_f64(__p0_248, __p1_248, __p2_248, __p3_248) __extension__ ({ \ + float64x2_t __s0_248 = __p0_248; \ + float64x2_t __s1_248 = __p1_248; \ + float64x1_t __s2_248 = __p2_248; \ + float64x2_t __ret_248; \ +float64x1_t __reint_248 = __s2_248; \ +uint64x2_t __reint1_248 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_248, __p3_248), vgetq_lane_u64(*(uint64x2_t *) &__reint_248, __p3_248)}; \ + __ret_248 = vcmlaq_rot270_f64(__s0_248, __s1_248, *(float64x2_t *) &__reint1_248); \ + __ret_248; \ +}) +#else +#define vcmlaq_rot270_lane_f64(__p0_249, __p1_249, __p2_249, __p3_249) __extension__ ({ \ + float64x2_t __s0_249 = __p0_249; \ + float64x2_t __s1_249 = __p1_249; \ + float64x1_t __s2_249 = __p2_249; \ + float64x2_t __rev0_249; __rev0_249 = __builtin_shufflevector(__s0_249, __s0_249, 1, 0); \ + float64x2_t __rev1_249; __rev1_249 = __builtin_shufflevector(__s1_249, __s1_249, 1, 0); \ + float64x2_t __ret_249; \ +float64x1_t __reint_249 = __s2_249; \ +uint64x2_t __reint1_249 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_249, __p3_249), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_249, __p3_249)}; \ + __ret_249 = __noswap_vcmlaq_rot270_f64(__rev0_249, __rev1_249, *(float64x2_t *) &__reint1_249); \ + __ret_249 = __builtin_shufflevector(__ret_249, __ret_249, 1, 0); \ + __ret_249; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot270_laneq_f64(__p0_250, __p1_250, __p2_250, __p3_250) __extension__ ({ \ + float64x1_t __s0_250 = __p0_250; \ + float64x1_t __s1_250 = __p1_250; \ + float64x2_t __s2_250 = __p2_250; \ + float64x1_t __ret_250; \ +float64x2_t __reint_250 = __s2_250; \ +uint64x2_t __reint1_250 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_250, __p3_250), vgetq_lane_u64(*(uint64x2_t *) &__reint_250, __p3_250)}; \ + __ret_250 = vcmla_rot270_f64(__s0_250, __s1_250, *(float64x1_t *) &__reint1_250); \ + __ret_250; \ +}) +#else +#define vcmla_rot270_laneq_f64(__p0_251, __p1_251, __p2_251, __p3_251) __extension__ ({ \ + float64x1_t __s0_251 = __p0_251; \ + float64x1_t __s1_251 = __p1_251; \ + float64x2_t __s2_251 = __p2_251; \ + float64x2_t __rev2_251; __rev2_251 = __builtin_shufflevector(__s2_251, __s2_251, 1, 0); \ + float64x1_t __ret_251; \ +float64x2_t __reint_251 = __rev2_251; \ +uint64x2_t __reint1_251 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_251, __p3_251), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_251, __p3_251)}; \ + __ret_251 = vcmla_rot270_f64(__s0_251, __s1_251, *(float64x1_t *) &__reint1_251); \ + __ret_251; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot270_laneq_f64(__p0_252, __p1_252, __p2_252, __p3_252) __extension__ ({ \ + float64x2_t __s0_252 = __p0_252; \ + float64x2_t __s1_252 = __p1_252; \ + float64x2_t __s2_252 = __p2_252; \ + float64x2_t __ret_252; \ +float64x2_t __reint_252 = __s2_252; \ +uint64x2_t __reint1_252 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_252, __p3_252), vgetq_lane_u64(*(uint64x2_t *) &__reint_252, __p3_252)}; \ + __ret_252 = vcmlaq_rot270_f64(__s0_252, __s1_252, *(float64x2_t *) &__reint1_252); \ + __ret_252; \ +}) +#else +#define vcmlaq_rot270_laneq_f64(__p0_253, __p1_253, __p2_253, __p3_253) __extension__ ({ \ + float64x2_t __s0_253 = __p0_253; \ + float64x2_t __s1_253 = __p1_253; \ + float64x2_t __s2_253 = __p2_253; \ + float64x2_t __rev0_253; __rev0_253 = __builtin_shufflevector(__s0_253, __s0_253, 1, 0); \ + float64x2_t __rev1_253; __rev1_253 = __builtin_shufflevector(__s1_253, __s1_253, 1, 0); \ + float64x2_t __rev2_253; __rev2_253 = __builtin_shufflevector(__s2_253, __s2_253, 1, 0); \ + float64x2_t __ret_253; \ +float64x2_t __reint_253 = __rev2_253; \ +uint64x2_t __reint1_253 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_253, __p3_253), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_253, __p3_253)}; \ + __ret_253 = __noswap_vcmlaq_rot270_f64(__rev0_253, __rev1_253, *(float64x2_t *) &__reint1_253); \ + __ret_253 = __builtin_shufflevector(__ret_253, __ret_253, 1, 0); \ + __ret_253; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +__ai float64x2_t vcmlaq_rot90_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#else +__ai float64x2_t vcmlaq_rot90_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); + float64x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); + float64x2_t __rev2; __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0); + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42); + __ret = __builtin_shufflevector(__ret, __ret, 1, 0); + return __ret; +} +__ai float64x2_t __noswap_vcmlaq_rot90_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) { + float64x2_t __ret; + __ret = (float64x2_t) __builtin_neon_vcmlaq_rot90_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42); + return __ret; +} +#endif + +__ai float64x1_t vcmla_rot90_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) { + float64x1_t __ret; + __ret = (float64x1_t) __builtin_neon_vcmla_rot90_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10); + return __ret; +} +#define vcmla_rot90_lane_f64(__p0_254, __p1_254, __p2_254, __p3_254) __extension__ ({ \ + float64x1_t __s0_254 = __p0_254; \ + float64x1_t __s1_254 = __p1_254; \ + float64x1_t __s2_254 = __p2_254; \ + float64x1_t __ret_254; \ +float64x1_t __reint_254 = __s2_254; \ +uint64x2_t __reint1_254 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_254, __p3_254), vgetq_lane_u64(*(uint64x2_t *) &__reint_254, __p3_254)}; \ + __ret_254 = vcmla_rot90_f64(__s0_254, __s1_254, *(float64x1_t *) &__reint1_254); \ + __ret_254; \ +}) +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot90_lane_f64(__p0_255, __p1_255, __p2_255, __p3_255) __extension__ ({ \ + float64x2_t __s0_255 = __p0_255; \ + float64x2_t __s1_255 = __p1_255; \ + float64x1_t __s2_255 = __p2_255; \ + float64x2_t __ret_255; \ +float64x1_t __reint_255 = __s2_255; \ +uint64x2_t __reint1_255 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_255, __p3_255), vgetq_lane_u64(*(uint64x2_t *) &__reint_255, __p3_255)}; \ + __ret_255 = vcmlaq_rot90_f64(__s0_255, __s1_255, *(float64x2_t *) &__reint1_255); \ + __ret_255; \ +}) +#else +#define vcmlaq_rot90_lane_f64(__p0_256, __p1_256, __p2_256, __p3_256) __extension__ ({ \ + float64x2_t __s0_256 = __p0_256; \ + float64x2_t __s1_256 = __p1_256; \ + float64x1_t __s2_256 = __p2_256; \ + float64x2_t __rev0_256; __rev0_256 = __builtin_shufflevector(__s0_256, __s0_256, 1, 0); \ + float64x2_t __rev1_256; __rev1_256 = __builtin_shufflevector(__s1_256, __s1_256, 1, 0); \ + float64x2_t __ret_256; \ +float64x1_t __reint_256 = __s2_256; \ +uint64x2_t __reint1_256 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_256, __p3_256), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_256, __p3_256)}; \ + __ret_256 = __noswap_vcmlaq_rot90_f64(__rev0_256, __rev1_256, *(float64x2_t *) &__reint1_256); \ + __ret_256 = __builtin_shufflevector(__ret_256, __ret_256, 1, 0); \ + __ret_256; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmla_rot90_laneq_f64(__p0_257, __p1_257, __p2_257, __p3_257) __extension__ ({ \ + float64x1_t __s0_257 = __p0_257; \ + float64x1_t __s1_257 = __p1_257; \ + float64x2_t __s2_257 = __p2_257; \ + float64x1_t __ret_257; \ +float64x2_t __reint_257 = __s2_257; \ +uint64x2_t __reint1_257 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_257, __p3_257), vgetq_lane_u64(*(uint64x2_t *) &__reint_257, __p3_257)}; \ + __ret_257 = vcmla_rot90_f64(__s0_257, __s1_257, *(float64x1_t *) &__reint1_257); \ + __ret_257; \ +}) +#else +#define vcmla_rot90_laneq_f64(__p0_258, __p1_258, __p2_258, __p3_258) __extension__ ({ \ + float64x1_t __s0_258 = __p0_258; \ + float64x1_t __s1_258 = __p1_258; \ + float64x2_t __s2_258 = __p2_258; \ + float64x2_t __rev2_258; __rev2_258 = __builtin_shufflevector(__s2_258, __s2_258, 1, 0); \ + float64x1_t __ret_258; \ +float64x2_t __reint_258 = __rev2_258; \ +uint64x2_t __reint1_258 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_258, __p3_258), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_258, __p3_258)}; \ + __ret_258 = vcmla_rot90_f64(__s0_258, __s1_258, *(float64x1_t *) &__reint1_258); \ + __ret_258; \ +}) +#endif + +#ifdef __LITTLE_ENDIAN__ +#define vcmlaq_rot90_laneq_f64(__p0_259, __p1_259, __p2_259, __p3_259) __extension__ ({ \ + float64x2_t __s0_259 = __p0_259; \ + float64x2_t __s1_259 = __p1_259; \ + float64x2_t __s2_259 = __p2_259; \ + float64x2_t __ret_259; \ +float64x2_t __reint_259 = __s2_259; \ +uint64x2_t __reint1_259 = (uint64x2_t) {vgetq_lane_u64(*(uint64x2_t *) &__reint_259, __p3_259), vgetq_lane_u64(*(uint64x2_t *) &__reint_259, __p3_259)}; \ + __ret_259 = vcmlaq_rot90_f64(__s0_259, __s1_259, *(float64x2_t *) &__reint1_259); \ + __ret_259; \ +}) +#else +#define vcmlaq_rot90_laneq_f64(__p0_260, __p1_260, __p2_260, __p3_260) __extension__ ({ \ + float64x2_t __s0_260 = __p0_260; \ + float64x2_t __s1_260 = __p1_260; \ + float64x2_t __s2_260 = __p2_260; \ + float64x2_t __rev0_260; __rev0_260 = __builtin_shufflevector(__s0_260, __s0_260, 1, 0); \ + float64x2_t __rev1_260; __rev1_260 = __builtin_shufflevector(__s1_260, __s1_260, 1, 0); \ + float64x2_t __rev2_260; __rev2_260 = __builtin_shufflevector(__s2_260, __s2_260, 1, 0); \ + float64x2_t __ret_260; \ +float64x2_t __reint_260 = __rev2_260; \ +uint64x2_t __reint1_260 = (uint64x2_t) {__noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_260, __p3_260), __noswap_vgetq_lane_u64(*(uint64x2_t *) &__reint_260, __p3_260)}; \ + __ret_260 = __noswap_vcmlaq_rot90_f64(__rev0_260, __rev1_260, *(float64x2_t *) &__reint1_260); \ + __ret_260 = __builtin_shufflevector(__ret_260, __ret_260, 1, 0); \ + __ret_260; \ +}) +#endif + #endif #if defined(__ARM_FEATURE_DOTPROD) #ifdef __LITTLE_ENDIAN__ @@ -40630,228 +42366,228 @@ __ai int32x2_t __noswap_vdot_s32(int32x2_t __p0, int8x8_t __p1, int8x8_t __p2) { #endif #ifdef __LITTLE_ENDIAN__ -#define vdotq_lane_u32(__p0_169, __p1_169, __p2_169, __p3_169) __extension__ ({ \ - uint32x4_t __s0_169 = __p0_169; \ - uint8x16_t __s1_169 = __p1_169; \ - uint8x8_t __s2_169 = __p2_169; \ - uint32x4_t __ret_169; \ -uint8x8_t __reint_169 = __s2_169; \ -uint32x4_t __reint1_169 = splatq_lane_u32(*(uint32x2_t *) &__reint_169, __p3_169); \ - __ret_169 = vdotq_u32(__s0_169, __s1_169, *(uint8x16_t *) &__reint1_169); \ - __ret_169; \ +#define vdotq_lane_u32(__p0_261, __p1_261, __p2_261, __p3_261) __extension__ ({ \ + uint32x4_t __s0_261 = __p0_261; \ + uint8x16_t __s1_261 = __p1_261; \ + uint8x8_t __s2_261 = __p2_261; \ + uint32x4_t __ret_261; \ +uint8x8_t __reint_261 = __s2_261; \ +uint32x4_t __reint1_261 = splatq_lane_u32(*(uint32x2_t *) &__reint_261, __p3_261); \ + __ret_261 = vdotq_u32(__s0_261, __s1_261, *(uint8x16_t *) &__reint1_261); \ + __ret_261; \ }) #else -#define vdotq_lane_u32(__p0_170, __p1_170, __p2_170, __p3_170) __extension__ ({ \ - uint32x4_t __s0_170 = __p0_170; \ - uint8x16_t __s1_170 = __p1_170; \ - uint8x8_t __s2_170 = __p2_170; \ - uint32x4_t __rev0_170; __rev0_170 = __builtin_shufflevector(__s0_170, __s0_170, 3, 2, 1, 0); \ - uint8x16_t __rev1_170; __rev1_170 = __builtin_shufflevector(__s1_170, __s1_170, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __rev2_170; __rev2_170 = __builtin_shufflevector(__s2_170, __s2_170, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_170; \ -uint8x8_t __reint_170 = __rev2_170; \ -uint32x4_t __reint1_170 = __noswap_splatq_lane_u32(*(uint32x2_t *) &__reint_170, __p3_170); \ - __ret_170 = __noswap_vdotq_u32(__rev0_170, __rev1_170, *(uint8x16_t *) &__reint1_170); \ - __ret_170 = __builtin_shufflevector(__ret_170, __ret_170, 3, 2, 1, 0); \ - __ret_170; \ +#define vdotq_lane_u32(__p0_262, __p1_262, __p2_262, __p3_262) __extension__ ({ \ + uint32x4_t __s0_262 = __p0_262; \ + uint8x16_t __s1_262 = __p1_262; \ + uint8x8_t __s2_262 = __p2_262; \ + uint32x4_t __rev0_262; __rev0_262 = __builtin_shufflevector(__s0_262, __s0_262, 3, 2, 1, 0); \ + uint8x16_t __rev1_262; __rev1_262 = __builtin_shufflevector(__s1_262, __s1_262, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __rev2_262; __rev2_262 = __builtin_shufflevector(__s2_262, __s2_262, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_262; \ +uint8x8_t __reint_262 = __rev2_262; \ +uint32x4_t __reint1_262 = __noswap_splatq_lane_u32(*(uint32x2_t *) &__reint_262, __p3_262); \ + __ret_262 = __noswap_vdotq_u32(__rev0_262, __rev1_262, *(uint8x16_t *) &__reint1_262); \ + __ret_262 = __builtin_shufflevector(__ret_262, __ret_262, 3, 2, 1, 0); \ + __ret_262; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdotq_lane_s32(__p0_171, __p1_171, __p2_171, __p3_171) __extension__ ({ \ - int32x4_t __s0_171 = __p0_171; \ - int8x16_t __s1_171 = __p1_171; \ - int8x8_t __s2_171 = __p2_171; \ - int32x4_t __ret_171; \ -int8x8_t __reint_171 = __s2_171; \ -int32x4_t __reint1_171 = splatq_lane_s32(*(int32x2_t *) &__reint_171, __p3_171); \ - __ret_171 = vdotq_s32(__s0_171, __s1_171, *(int8x16_t *) &__reint1_171); \ - __ret_171; \ +#define vdotq_lane_s32(__p0_263, __p1_263, __p2_263, __p3_263) __extension__ ({ \ + int32x4_t __s0_263 = __p0_263; \ + int8x16_t __s1_263 = __p1_263; \ + int8x8_t __s2_263 = __p2_263; \ + int32x4_t __ret_263; \ +int8x8_t __reint_263 = __s2_263; \ +int32x4_t __reint1_263 = splatq_lane_s32(*(int32x2_t *) &__reint_263, __p3_263); \ + __ret_263 = vdotq_s32(__s0_263, __s1_263, *(int8x16_t *) &__reint1_263); \ + __ret_263; \ }) #else -#define vdotq_lane_s32(__p0_172, __p1_172, __p2_172, __p3_172) __extension__ ({ \ - int32x4_t __s0_172 = __p0_172; \ - int8x16_t __s1_172 = __p1_172; \ - int8x8_t __s2_172 = __p2_172; \ - int32x4_t __rev0_172; __rev0_172 = __builtin_shufflevector(__s0_172, __s0_172, 3, 2, 1, 0); \ - int8x16_t __rev1_172; __rev1_172 = __builtin_shufflevector(__s1_172, __s1_172, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __rev2_172; __rev2_172 = __builtin_shufflevector(__s2_172, __s2_172, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_172; \ -int8x8_t __reint_172 = __rev2_172; \ -int32x4_t __reint1_172 = __noswap_splatq_lane_s32(*(int32x2_t *) &__reint_172, __p3_172); \ - __ret_172 = __noswap_vdotq_s32(__rev0_172, __rev1_172, *(int8x16_t *) &__reint1_172); \ - __ret_172 = __builtin_shufflevector(__ret_172, __ret_172, 3, 2, 1, 0); \ - __ret_172; \ +#define vdotq_lane_s32(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \ + int32x4_t __s0_264 = __p0_264; \ + int8x16_t __s1_264 = __p1_264; \ + int8x8_t __s2_264 = __p2_264; \ + int32x4_t __rev0_264; __rev0_264 = __builtin_shufflevector(__s0_264, __s0_264, 3, 2, 1, 0); \ + int8x16_t __rev1_264; __rev1_264 = __builtin_shufflevector(__s1_264, __s1_264, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __rev2_264; __rev2_264 = __builtin_shufflevector(__s2_264, __s2_264, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_264; \ +int8x8_t __reint_264 = __rev2_264; \ +int32x4_t __reint1_264 = __noswap_splatq_lane_s32(*(int32x2_t *) &__reint_264, __p3_264); \ + __ret_264 = __noswap_vdotq_s32(__rev0_264, __rev1_264, *(int8x16_t *) &__reint1_264); \ + __ret_264 = __builtin_shufflevector(__ret_264, __ret_264, 3, 2, 1, 0); \ + __ret_264; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdot_lane_u32(__p0_173, __p1_173, __p2_173, __p3_173) __extension__ ({ \ - uint32x2_t __s0_173 = __p0_173; \ - uint8x8_t __s1_173 = __p1_173; \ - uint8x8_t __s2_173 = __p2_173; \ - uint32x2_t __ret_173; \ -uint8x8_t __reint_173 = __s2_173; \ -uint32x2_t __reint1_173 = splat_lane_u32(*(uint32x2_t *) &__reint_173, __p3_173); \ - __ret_173 = vdot_u32(__s0_173, __s1_173, *(uint8x8_t *) &__reint1_173); \ - __ret_173; \ +#define vdot_lane_u32(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \ + uint32x2_t __s0_265 = __p0_265; \ + uint8x8_t __s1_265 = __p1_265; \ + uint8x8_t __s2_265 = __p2_265; \ + uint32x2_t __ret_265; \ +uint8x8_t __reint_265 = __s2_265; \ +uint32x2_t __reint1_265 = splat_lane_u32(*(uint32x2_t *) &__reint_265, __p3_265); \ + __ret_265 = vdot_u32(__s0_265, __s1_265, *(uint8x8_t *) &__reint1_265); \ + __ret_265; \ }) #else -#define vdot_lane_u32(__p0_174, __p1_174, __p2_174, __p3_174) __extension__ ({ \ - uint32x2_t __s0_174 = __p0_174; \ - uint8x8_t __s1_174 = __p1_174; \ - uint8x8_t __s2_174 = __p2_174; \ - uint32x2_t __rev0_174; __rev0_174 = __builtin_shufflevector(__s0_174, __s0_174, 1, 0); \ - uint8x8_t __rev1_174; __rev1_174 = __builtin_shufflevector(__s1_174, __s1_174, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __rev2_174; __rev2_174 = __builtin_shufflevector(__s2_174, __s2_174, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x2_t __ret_174; \ -uint8x8_t __reint_174 = __rev2_174; \ -uint32x2_t __reint1_174 = __noswap_splat_lane_u32(*(uint32x2_t *) &__reint_174, __p3_174); \ - __ret_174 = __noswap_vdot_u32(__rev0_174, __rev1_174, *(uint8x8_t *) &__reint1_174); \ - __ret_174 = __builtin_shufflevector(__ret_174, __ret_174, 1, 0); \ - __ret_174; \ +#define vdot_lane_u32(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \ + uint32x2_t __s0_266 = __p0_266; \ + uint8x8_t __s1_266 = __p1_266; \ + uint8x8_t __s2_266 = __p2_266; \ + uint32x2_t __rev0_266; __rev0_266 = __builtin_shufflevector(__s0_266, __s0_266, 1, 0); \ + uint8x8_t __rev1_266; __rev1_266 = __builtin_shufflevector(__s1_266, __s1_266, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __rev2_266; __rev2_266 = __builtin_shufflevector(__s2_266, __s2_266, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x2_t __ret_266; \ +uint8x8_t __reint_266 = __rev2_266; \ +uint32x2_t __reint1_266 = __noswap_splat_lane_u32(*(uint32x2_t *) &__reint_266, __p3_266); \ + __ret_266 = __noswap_vdot_u32(__rev0_266, __rev1_266, *(uint8x8_t *) &__reint1_266); \ + __ret_266 = __builtin_shufflevector(__ret_266, __ret_266, 1, 0); \ + __ret_266; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdot_lane_s32(__p0_175, __p1_175, __p2_175, __p3_175) __extension__ ({ \ - int32x2_t __s0_175 = __p0_175; \ - int8x8_t __s1_175 = __p1_175; \ - int8x8_t __s2_175 = __p2_175; \ - int32x2_t __ret_175; \ -int8x8_t __reint_175 = __s2_175; \ -int32x2_t __reint1_175 = splat_lane_s32(*(int32x2_t *) &__reint_175, __p3_175); \ - __ret_175 = vdot_s32(__s0_175, __s1_175, *(int8x8_t *) &__reint1_175); \ - __ret_175; \ +#define vdot_lane_s32(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \ + int32x2_t __s0_267 = __p0_267; \ + int8x8_t __s1_267 = __p1_267; \ + int8x8_t __s2_267 = __p2_267; \ + int32x2_t __ret_267; \ +int8x8_t __reint_267 = __s2_267; \ +int32x2_t __reint1_267 = splat_lane_s32(*(int32x2_t *) &__reint_267, __p3_267); \ + __ret_267 = vdot_s32(__s0_267, __s1_267, *(int8x8_t *) &__reint1_267); \ + __ret_267; \ }) #else -#define vdot_lane_s32(__p0_176, __p1_176, __p2_176, __p3_176) __extension__ ({ \ - int32x2_t __s0_176 = __p0_176; \ - int8x8_t __s1_176 = __p1_176; \ - int8x8_t __s2_176 = __p2_176; \ - int32x2_t __rev0_176; __rev0_176 = __builtin_shufflevector(__s0_176, __s0_176, 1, 0); \ - int8x8_t __rev1_176; __rev1_176 = __builtin_shufflevector(__s1_176, __s1_176, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __rev2_176; __rev2_176 = __builtin_shufflevector(__s2_176, __s2_176, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x2_t __ret_176; \ -int8x8_t __reint_176 = __rev2_176; \ -int32x2_t __reint1_176 = __noswap_splat_lane_s32(*(int32x2_t *) &__reint_176, __p3_176); \ - __ret_176 = __noswap_vdot_s32(__rev0_176, __rev1_176, *(int8x8_t *) &__reint1_176); \ - __ret_176 = __builtin_shufflevector(__ret_176, __ret_176, 1, 0); \ - __ret_176; \ +#define vdot_lane_s32(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \ + int32x2_t __s0_268 = __p0_268; \ + int8x8_t __s1_268 = __p1_268; \ + int8x8_t __s2_268 = __p2_268; \ + int32x2_t __rev0_268; __rev0_268 = __builtin_shufflevector(__s0_268, __s0_268, 1, 0); \ + int8x8_t __rev1_268; __rev1_268 = __builtin_shufflevector(__s1_268, __s1_268, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __rev2_268; __rev2_268 = __builtin_shufflevector(__s2_268, __s2_268, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x2_t __ret_268; \ +int8x8_t __reint_268 = __rev2_268; \ +int32x2_t __reint1_268 = __noswap_splat_lane_s32(*(int32x2_t *) &__reint_268, __p3_268); \ + __ret_268 = __noswap_vdot_s32(__rev0_268, __rev1_268, *(int8x8_t *) &__reint1_268); \ + __ret_268 = __builtin_shufflevector(__ret_268, __ret_268, 1, 0); \ + __ret_268; \ }) #endif #endif #if defined(__ARM_FEATURE_DOTPROD) && defined(__aarch64__) #ifdef __LITTLE_ENDIAN__ -#define vdotq_laneq_u32(__p0_177, __p1_177, __p2_177, __p3_177) __extension__ ({ \ - uint32x4_t __s0_177 = __p0_177; \ - uint8x16_t __s1_177 = __p1_177; \ - uint8x16_t __s2_177 = __p2_177; \ - uint32x4_t __ret_177; \ -uint8x16_t __reint_177 = __s2_177; \ -uint32x4_t __reint1_177 = splatq_laneq_u32(*(uint32x4_t *) &__reint_177, __p3_177); \ - __ret_177 = vdotq_u32(__s0_177, __s1_177, *(uint8x16_t *) &__reint1_177); \ - __ret_177; \ +#define vdotq_laneq_u32(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \ + uint32x4_t __s0_269 = __p0_269; \ + uint8x16_t __s1_269 = __p1_269; \ + uint8x16_t __s2_269 = __p2_269; \ + uint32x4_t __ret_269; \ +uint8x16_t __reint_269 = __s2_269; \ +uint32x4_t __reint1_269 = splatq_laneq_u32(*(uint32x4_t *) &__reint_269, __p3_269); \ + __ret_269 = vdotq_u32(__s0_269, __s1_269, *(uint8x16_t *) &__reint1_269); \ + __ret_269; \ }) #else -#define vdotq_laneq_u32(__p0_178, __p1_178, __p2_178, __p3_178) __extension__ ({ \ - uint32x4_t __s0_178 = __p0_178; \ - uint8x16_t __s1_178 = __p1_178; \ - uint8x16_t __s2_178 = __p2_178; \ - uint32x4_t __rev0_178; __rev0_178 = __builtin_shufflevector(__s0_178, __s0_178, 3, 2, 1, 0); \ - uint8x16_t __rev1_178; __rev1_178 = __builtin_shufflevector(__s1_178, __s1_178, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __rev2_178; __rev2_178 = __builtin_shufflevector(__s2_178, __s2_178, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_178; \ -uint8x16_t __reint_178 = __rev2_178; \ -uint32x4_t __reint1_178 = __noswap_splatq_laneq_u32(*(uint32x4_t *) &__reint_178, __p3_178); \ - __ret_178 = __noswap_vdotq_u32(__rev0_178, __rev1_178, *(uint8x16_t *) &__reint1_178); \ - __ret_178 = __builtin_shufflevector(__ret_178, __ret_178, 3, 2, 1, 0); \ - __ret_178; \ +#define vdotq_laneq_u32(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \ + uint32x4_t __s0_270 = __p0_270; \ + uint8x16_t __s1_270 = __p1_270; \ + uint8x16_t __s2_270 = __p2_270; \ + uint32x4_t __rev0_270; __rev0_270 = __builtin_shufflevector(__s0_270, __s0_270, 3, 2, 1, 0); \ + uint8x16_t __rev1_270; __rev1_270 = __builtin_shufflevector(__s1_270, __s1_270, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __rev2_270; __rev2_270 = __builtin_shufflevector(__s2_270, __s2_270, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_270; \ +uint8x16_t __reint_270 = __rev2_270; \ +uint32x4_t __reint1_270 = __noswap_splatq_laneq_u32(*(uint32x4_t *) &__reint_270, __p3_270); \ + __ret_270 = __noswap_vdotq_u32(__rev0_270, __rev1_270, *(uint8x16_t *) &__reint1_270); \ + __ret_270 = __builtin_shufflevector(__ret_270, __ret_270, 3, 2, 1, 0); \ + __ret_270; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdotq_laneq_s32(__p0_179, __p1_179, __p2_179, __p3_179) __extension__ ({ \ - int32x4_t __s0_179 = __p0_179; \ - int8x16_t __s1_179 = __p1_179; \ - int8x16_t __s2_179 = __p2_179; \ - int32x4_t __ret_179; \ -int8x16_t __reint_179 = __s2_179; \ -int32x4_t __reint1_179 = splatq_laneq_s32(*(int32x4_t *) &__reint_179, __p3_179); \ - __ret_179 = vdotq_s32(__s0_179, __s1_179, *(int8x16_t *) &__reint1_179); \ - __ret_179; \ +#define vdotq_laneq_s32(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \ + int32x4_t __s0_271 = __p0_271; \ + int8x16_t __s1_271 = __p1_271; \ + int8x16_t __s2_271 = __p2_271; \ + int32x4_t __ret_271; \ +int8x16_t __reint_271 = __s2_271; \ +int32x4_t __reint1_271 = splatq_laneq_s32(*(int32x4_t *) &__reint_271, __p3_271); \ + __ret_271 = vdotq_s32(__s0_271, __s1_271, *(int8x16_t *) &__reint1_271); \ + __ret_271; \ }) #else -#define vdotq_laneq_s32(__p0_180, __p1_180, __p2_180, __p3_180) __extension__ ({ \ - int32x4_t __s0_180 = __p0_180; \ - int8x16_t __s1_180 = __p1_180; \ - int8x16_t __s2_180 = __p2_180; \ - int32x4_t __rev0_180; __rev0_180 = __builtin_shufflevector(__s0_180, __s0_180, 3, 2, 1, 0); \ - int8x16_t __rev1_180; __rev1_180 = __builtin_shufflevector(__s1_180, __s1_180, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __rev2_180; __rev2_180 = __builtin_shufflevector(__s2_180, __s2_180, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_180; \ -int8x16_t __reint_180 = __rev2_180; \ -int32x4_t __reint1_180 = __noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_180, __p3_180); \ - __ret_180 = __noswap_vdotq_s32(__rev0_180, __rev1_180, *(int8x16_t *) &__reint1_180); \ - __ret_180 = __builtin_shufflevector(__ret_180, __ret_180, 3, 2, 1, 0); \ - __ret_180; \ +#define vdotq_laneq_s32(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \ + int32x4_t __s0_272 = __p0_272; \ + int8x16_t __s1_272 = __p1_272; \ + int8x16_t __s2_272 = __p2_272; \ + int32x4_t __rev0_272; __rev0_272 = __builtin_shufflevector(__s0_272, __s0_272, 3, 2, 1, 0); \ + int8x16_t __rev1_272; __rev1_272 = __builtin_shufflevector(__s1_272, __s1_272, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __rev2_272; __rev2_272 = __builtin_shufflevector(__s2_272, __s2_272, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_272; \ +int8x16_t __reint_272 = __rev2_272; \ +int32x4_t __reint1_272 = __noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_272, __p3_272); \ + __ret_272 = __noswap_vdotq_s32(__rev0_272, __rev1_272, *(int8x16_t *) &__reint1_272); \ + __ret_272 = __builtin_shufflevector(__ret_272, __ret_272, 3, 2, 1, 0); \ + __ret_272; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdot_laneq_u32(__p0_181, __p1_181, __p2_181, __p3_181) __extension__ ({ \ - uint32x2_t __s0_181 = __p0_181; \ - uint8x8_t __s1_181 = __p1_181; \ - uint8x16_t __s2_181 = __p2_181; \ - uint32x2_t __ret_181; \ -uint8x16_t __reint_181 = __s2_181; \ -uint32x2_t __reint1_181 = splat_laneq_u32(*(uint32x4_t *) &__reint_181, __p3_181); \ - __ret_181 = vdot_u32(__s0_181, __s1_181, *(uint8x8_t *) &__reint1_181); \ - __ret_181; \ +#define vdot_laneq_u32(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \ + uint32x2_t __s0_273 = __p0_273; \ + uint8x8_t __s1_273 = __p1_273; \ + uint8x16_t __s2_273 = __p2_273; \ + uint32x2_t __ret_273; \ +uint8x16_t __reint_273 = __s2_273; \ +uint32x2_t __reint1_273 = splat_laneq_u32(*(uint32x4_t *) &__reint_273, __p3_273); \ + __ret_273 = vdot_u32(__s0_273, __s1_273, *(uint8x8_t *) &__reint1_273); \ + __ret_273; \ }) #else -#define vdot_laneq_u32(__p0_182, __p1_182, __p2_182, __p3_182) __extension__ ({ \ - uint32x2_t __s0_182 = __p0_182; \ - uint8x8_t __s1_182 = __p1_182; \ - uint8x16_t __s2_182 = __p2_182; \ - uint32x2_t __rev0_182; __rev0_182 = __builtin_shufflevector(__s0_182, __s0_182, 1, 0); \ - uint8x8_t __rev1_182; __rev1_182 = __builtin_shufflevector(__s1_182, __s1_182, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __rev2_182; __rev2_182 = __builtin_shufflevector(__s2_182, __s2_182, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x2_t __ret_182; \ -uint8x16_t __reint_182 = __rev2_182; \ -uint32x2_t __reint1_182 = __noswap_splat_laneq_u32(*(uint32x4_t *) &__reint_182, __p3_182); \ - __ret_182 = __noswap_vdot_u32(__rev0_182, __rev1_182, *(uint8x8_t *) &__reint1_182); \ - __ret_182 = __builtin_shufflevector(__ret_182, __ret_182, 1, 0); \ - __ret_182; \ +#define vdot_laneq_u32(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \ + uint32x2_t __s0_274 = __p0_274; \ + uint8x8_t __s1_274 = __p1_274; \ + uint8x16_t __s2_274 = __p2_274; \ + uint32x2_t __rev0_274; __rev0_274 = __builtin_shufflevector(__s0_274, __s0_274, 1, 0); \ + uint8x8_t __rev1_274; __rev1_274 = __builtin_shufflevector(__s1_274, __s1_274, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __rev2_274; __rev2_274 = __builtin_shufflevector(__s2_274, __s2_274, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x2_t __ret_274; \ +uint8x16_t __reint_274 = __rev2_274; \ +uint32x2_t __reint1_274 = __noswap_splat_laneq_u32(*(uint32x4_t *) &__reint_274, __p3_274); \ + __ret_274 = __noswap_vdot_u32(__rev0_274, __rev1_274, *(uint8x8_t *) &__reint1_274); \ + __ret_274 = __builtin_shufflevector(__ret_274, __ret_274, 1, 0); \ + __ret_274; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdot_laneq_s32(__p0_183, __p1_183, __p2_183, __p3_183) __extension__ ({ \ - int32x2_t __s0_183 = __p0_183; \ - int8x8_t __s1_183 = __p1_183; \ - int8x16_t __s2_183 = __p2_183; \ - int32x2_t __ret_183; \ -int8x16_t __reint_183 = __s2_183; \ -int32x2_t __reint1_183 = splat_laneq_s32(*(int32x4_t *) &__reint_183, __p3_183); \ - __ret_183 = vdot_s32(__s0_183, __s1_183, *(int8x8_t *) &__reint1_183); \ - __ret_183; \ +#define vdot_laneq_s32(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \ + int32x2_t __s0_275 = __p0_275; \ + int8x8_t __s1_275 = __p1_275; \ + int8x16_t __s2_275 = __p2_275; \ + int32x2_t __ret_275; \ +int8x16_t __reint_275 = __s2_275; \ +int32x2_t __reint1_275 = splat_laneq_s32(*(int32x4_t *) &__reint_275, __p3_275); \ + __ret_275 = vdot_s32(__s0_275, __s1_275, *(int8x8_t *) &__reint1_275); \ + __ret_275; \ }) #else -#define vdot_laneq_s32(__p0_184, __p1_184, __p2_184, __p3_184) __extension__ ({ \ - int32x2_t __s0_184 = __p0_184; \ - int8x8_t __s1_184 = __p1_184; \ - int8x16_t __s2_184 = __p2_184; \ - int32x2_t __rev0_184; __rev0_184 = __builtin_shufflevector(__s0_184, __s0_184, 1, 0); \ - int8x8_t __rev1_184; __rev1_184 = __builtin_shufflevector(__s1_184, __s1_184, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __rev2_184; __rev2_184 = __builtin_shufflevector(__s2_184, __s2_184, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x2_t __ret_184; \ -int8x16_t __reint_184 = __rev2_184; \ -int32x2_t __reint1_184 = __noswap_splat_laneq_s32(*(int32x4_t *) &__reint_184, __p3_184); \ - __ret_184 = __noswap_vdot_s32(__rev0_184, __rev1_184, *(int8x8_t *) &__reint1_184); \ - __ret_184 = __builtin_shufflevector(__ret_184, __ret_184, 1, 0); \ - __ret_184; \ +#define vdot_laneq_s32(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \ + int32x2_t __s0_276 = __p0_276; \ + int8x8_t __s1_276 = __p1_276; \ + int8x16_t __s2_276 = __p2_276; \ + int32x2_t __rev0_276; __rev0_276 = __builtin_shufflevector(__s0_276, __s0_276, 1, 0); \ + int8x8_t __rev1_276; __rev1_276 = __builtin_shufflevector(__s1_276, __s1_276, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __rev2_276; __rev2_276 = __builtin_shufflevector(__s2_276, __s2_276, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x2_t __ret_276; \ +int8x16_t __reint_276 = __rev2_276; \ +int32x2_t __reint1_276 = __noswap_splat_laneq_s32(*(int32x4_t *) &__reint_276, __p3_276); \ + __ret_276 = __noswap_vdot_s32(__rev0_276, __rev1_276, *(int8x8_t *) &__reint1_276); \ + __ret_276 = __builtin_shufflevector(__ret_276, __ret_276, 1, 0); \ + __ret_276; \ }) #endif @@ -42518,44 +44254,44 @@ __ai float16x4_t vmul_f16(float16x4_t __p0, float16x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_lane_f16(__p0_185, __p1_185, __p2_185) __extension__ ({ \ - float16x8_t __s0_185 = __p0_185; \ - float16x4_t __s1_185 = __p1_185; \ - float16x8_t __ret_185; \ - __ret_185 = __s0_185 * splatq_lane_f16(__s1_185, __p2_185); \ - __ret_185; \ +#define vmulq_lane_f16(__p0_277, __p1_277, __p2_277) __extension__ ({ \ + float16x8_t __s0_277 = __p0_277; \ + float16x4_t __s1_277 = __p1_277; \ + float16x8_t __ret_277; \ + __ret_277 = __s0_277 * splatq_lane_f16(__s1_277, __p2_277); \ + __ret_277; \ }) #else -#define vmulq_lane_f16(__p0_186, __p1_186, __p2_186) __extension__ ({ \ - float16x8_t __s0_186 = __p0_186; \ - float16x4_t __s1_186 = __p1_186; \ - float16x8_t __rev0_186; __rev0_186 = __builtin_shufflevector(__s0_186, __s0_186, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __rev1_186; __rev1_186 = __builtin_shufflevector(__s1_186, __s1_186, 3, 2, 1, 0); \ - float16x8_t __ret_186; \ - __ret_186 = __rev0_186 * __noswap_splatq_lane_f16(__rev1_186, __p2_186); \ - __ret_186 = __builtin_shufflevector(__ret_186, __ret_186, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_186; \ +#define vmulq_lane_f16(__p0_278, __p1_278, __p2_278) __extension__ ({ \ + float16x8_t __s0_278 = __p0_278; \ + float16x4_t __s1_278 = __p1_278; \ + float16x8_t __rev0_278; __rev0_278 = __builtin_shufflevector(__s0_278, __s0_278, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev1_278; __rev1_278 = __builtin_shufflevector(__s1_278, __s1_278, 3, 2, 1, 0); \ + float16x8_t __ret_278; \ + __ret_278 = __rev0_278 * __noswap_splatq_lane_f16(__rev1_278, __p2_278); \ + __ret_278 = __builtin_shufflevector(__ret_278, __ret_278, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_278; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmul_lane_f16(__p0_187, __p1_187, __p2_187) __extension__ ({ \ - float16x4_t __s0_187 = __p0_187; \ - float16x4_t __s1_187 = __p1_187; \ - float16x4_t __ret_187; \ - __ret_187 = __s0_187 * splat_lane_f16(__s1_187, __p2_187); \ - __ret_187; \ +#define vmul_lane_f16(__p0_279, __p1_279, __p2_279) __extension__ ({ \ + float16x4_t __s0_279 = __p0_279; \ + float16x4_t __s1_279 = __p1_279; \ + float16x4_t __ret_279; \ + __ret_279 = __s0_279 * splat_lane_f16(__s1_279, __p2_279); \ + __ret_279; \ }) #else -#define vmul_lane_f16(__p0_188, __p1_188, __p2_188) __extension__ ({ \ - float16x4_t __s0_188 = __p0_188; \ - float16x4_t __s1_188 = __p1_188; \ - float16x4_t __rev0_188; __rev0_188 = __builtin_shufflevector(__s0_188, __s0_188, 3, 2, 1, 0); \ - float16x4_t __rev1_188; __rev1_188 = __builtin_shufflevector(__s1_188, __s1_188, 3, 2, 1, 0); \ - float16x4_t __ret_188; \ - __ret_188 = __rev0_188 * __noswap_splat_lane_f16(__rev1_188, __p2_188); \ - __ret_188 = __builtin_shufflevector(__ret_188, __ret_188, 3, 2, 1, 0); \ - __ret_188; \ +#define vmul_lane_f16(__p0_280, __p1_280, __p2_280) __extension__ ({ \ + float16x4_t __s0_280 = __p0_280; \ + float16x4_t __s1_280 = __p1_280; \ + float16x4_t __rev0_280; __rev0_280 = __builtin_shufflevector(__s0_280, __s0_280, 3, 2, 1, 0); \ + float16x4_t __rev1_280; __rev1_280 = __builtin_shufflevector(__s1_280, __s1_280, 3, 2, 1, 0); \ + float16x4_t __ret_280; \ + __ret_280 = __rev0_280 * __noswap_splat_lane_f16(__rev1_280, __p2_280); \ + __ret_280 = __builtin_shufflevector(__ret_280, __ret_280, 3, 2, 1, 0); \ + __ret_280; \ }) #endif @@ -43297,140 +45033,140 @@ __ai float16x4_t vdiv_f16(float16x4_t __p0, float16x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsh_lane_f16(__p0_189, __p1_189, __p2_189, __p3_189) __extension__ ({ \ - float16_t __s0_189 = __p0_189; \ - float16_t __s1_189 = __p1_189; \ - float16x4_t __s2_189 = __p2_189; \ - float16_t __ret_189; \ - __ret_189 = vfmah_lane_f16(__s0_189, -__s1_189, __s2_189, __p3_189); \ - __ret_189; \ +#define vfmsh_lane_f16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \ + float16_t __s0_281 = __p0_281; \ + float16_t __s1_281 = __p1_281; \ + float16x4_t __s2_281 = __p2_281; \ + float16_t __ret_281; \ + __ret_281 = vfmah_lane_f16(__s0_281, -__s1_281, __s2_281, __p3_281); \ + __ret_281; \ }) #else -#define vfmsh_lane_f16(__p0_190, __p1_190, __p2_190, __p3_190) __extension__ ({ \ - float16_t __s0_190 = __p0_190; \ - float16_t __s1_190 = __p1_190; \ - float16x4_t __s2_190 = __p2_190; \ - float16x4_t __rev2_190; __rev2_190 = __builtin_shufflevector(__s2_190, __s2_190, 3, 2, 1, 0); \ - float16_t __ret_190; \ - __ret_190 = __noswap_vfmah_lane_f16(__s0_190, -__s1_190, __rev2_190, __p3_190); \ - __ret_190; \ +#define vfmsh_lane_f16(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \ + float16_t __s0_282 = __p0_282; \ + float16_t __s1_282 = __p1_282; \ + float16x4_t __s2_282 = __p2_282; \ + float16x4_t __rev2_282; __rev2_282 = __builtin_shufflevector(__s2_282, __s2_282, 3, 2, 1, 0); \ + float16_t __ret_282; \ + __ret_282 = __noswap_vfmah_lane_f16(__s0_282, -__s1_282, __rev2_282, __p3_282); \ + __ret_282; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsq_lane_f16(__p0_191, __p1_191, __p2_191, __p3_191) __extension__ ({ \ - float16x8_t __s0_191 = __p0_191; \ - float16x8_t __s1_191 = __p1_191; \ - float16x4_t __s2_191 = __p2_191; \ - float16x8_t __ret_191; \ - __ret_191 = vfmaq_lane_f16(__s0_191, -__s1_191, __s2_191, __p3_191); \ - __ret_191; \ +#define vfmsq_lane_f16(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \ + float16x8_t __s0_283 = __p0_283; \ + float16x8_t __s1_283 = __p1_283; \ + float16x4_t __s2_283 = __p2_283; \ + float16x8_t __ret_283; \ + __ret_283 = vfmaq_lane_f16(__s0_283, -__s1_283, __s2_283, __p3_283); \ + __ret_283; \ }) #else -#define vfmsq_lane_f16(__p0_192, __p1_192, __p2_192, __p3_192) __extension__ ({ \ - float16x8_t __s0_192 = __p0_192; \ - float16x8_t __s1_192 = __p1_192; \ - float16x4_t __s2_192 = __p2_192; \ - float16x8_t __rev0_192; __rev0_192 = __builtin_shufflevector(__s0_192, __s0_192, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev1_192; __rev1_192 = __builtin_shufflevector(__s1_192, __s1_192, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __rev2_192; __rev2_192 = __builtin_shufflevector(__s2_192, __s2_192, 3, 2, 1, 0); \ - float16x8_t __ret_192; \ - __ret_192 = __noswap_vfmaq_lane_f16(__rev0_192, -__rev1_192, __rev2_192, __p3_192); \ - __ret_192 = __builtin_shufflevector(__ret_192, __ret_192, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_192; \ +#define vfmsq_lane_f16(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \ + float16x8_t __s0_284 = __p0_284; \ + float16x8_t __s1_284 = __p1_284; \ + float16x4_t __s2_284 = __p2_284; \ + float16x8_t __rev0_284; __rev0_284 = __builtin_shufflevector(__s0_284, __s0_284, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_284; __rev1_284 = __builtin_shufflevector(__s1_284, __s1_284, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_284; __rev2_284 = __builtin_shufflevector(__s2_284, __s2_284, 3, 2, 1, 0); \ + float16x8_t __ret_284; \ + __ret_284 = __noswap_vfmaq_lane_f16(__rev0_284, -__rev1_284, __rev2_284, __p3_284); \ + __ret_284 = __builtin_shufflevector(__ret_284, __ret_284, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_284; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfms_lane_f16(__p0_193, __p1_193, __p2_193, __p3_193) __extension__ ({ \ - float16x4_t __s0_193 = __p0_193; \ - float16x4_t __s1_193 = __p1_193; \ - float16x4_t __s2_193 = __p2_193; \ - float16x4_t __ret_193; \ - __ret_193 = vfma_lane_f16(__s0_193, -__s1_193, __s2_193, __p3_193); \ - __ret_193; \ +#define vfms_lane_f16(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \ + float16x4_t __s0_285 = __p0_285; \ + float16x4_t __s1_285 = __p1_285; \ + float16x4_t __s2_285 = __p2_285; \ + float16x4_t __ret_285; \ + __ret_285 = vfma_lane_f16(__s0_285, -__s1_285, __s2_285, __p3_285); \ + __ret_285; \ }) #else -#define vfms_lane_f16(__p0_194, __p1_194, __p2_194, __p3_194) __extension__ ({ \ - float16x4_t __s0_194 = __p0_194; \ - float16x4_t __s1_194 = __p1_194; \ - float16x4_t __s2_194 = __p2_194; \ - float16x4_t __rev0_194; __rev0_194 = __builtin_shufflevector(__s0_194, __s0_194, 3, 2, 1, 0); \ - float16x4_t __rev1_194; __rev1_194 = __builtin_shufflevector(__s1_194, __s1_194, 3, 2, 1, 0); \ - float16x4_t __rev2_194; __rev2_194 = __builtin_shufflevector(__s2_194, __s2_194, 3, 2, 1, 0); \ - float16x4_t __ret_194; \ - __ret_194 = __noswap_vfma_lane_f16(__rev0_194, -__rev1_194, __rev2_194, __p3_194); \ - __ret_194 = __builtin_shufflevector(__ret_194, __ret_194, 3, 2, 1, 0); \ - __ret_194; \ +#define vfms_lane_f16(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \ + float16x4_t __s0_286 = __p0_286; \ + float16x4_t __s1_286 = __p1_286; \ + float16x4_t __s2_286 = __p2_286; \ + float16x4_t __rev0_286; __rev0_286 = __builtin_shufflevector(__s0_286, __s0_286, 3, 2, 1, 0); \ + float16x4_t __rev1_286; __rev1_286 = __builtin_shufflevector(__s1_286, __s1_286, 3, 2, 1, 0); \ + float16x4_t __rev2_286; __rev2_286 = __builtin_shufflevector(__s2_286, __s2_286, 3, 2, 1, 0); \ + float16x4_t __ret_286; \ + __ret_286 = __noswap_vfma_lane_f16(__rev0_286, -__rev1_286, __rev2_286, __p3_286); \ + __ret_286 = __builtin_shufflevector(__ret_286, __ret_286, 3, 2, 1, 0); \ + __ret_286; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsh_laneq_f16(__p0_195, __p1_195, __p2_195, __p3_195) __extension__ ({ \ - float16_t __s0_195 = __p0_195; \ - float16_t __s1_195 = __p1_195; \ - float16x8_t __s2_195 = __p2_195; \ - float16_t __ret_195; \ - __ret_195 = vfmah_laneq_f16(__s0_195, -__s1_195, __s2_195, __p3_195); \ - __ret_195; \ +#define vfmsh_laneq_f16(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \ + float16_t __s0_287 = __p0_287; \ + float16_t __s1_287 = __p1_287; \ + float16x8_t __s2_287 = __p2_287; \ + float16_t __ret_287; \ + __ret_287 = vfmah_laneq_f16(__s0_287, -__s1_287, __s2_287, __p3_287); \ + __ret_287; \ }) #else -#define vfmsh_laneq_f16(__p0_196, __p1_196, __p2_196, __p3_196) __extension__ ({ \ - float16_t __s0_196 = __p0_196; \ - float16_t __s1_196 = __p1_196; \ - float16x8_t __s2_196 = __p2_196; \ - float16x8_t __rev2_196; __rev2_196 = __builtin_shufflevector(__s2_196, __s2_196, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16_t __ret_196; \ - __ret_196 = __noswap_vfmah_laneq_f16(__s0_196, -__s1_196, __rev2_196, __p3_196); \ - __ret_196; \ +#define vfmsh_laneq_f16(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \ + float16_t __s0_288 = __p0_288; \ + float16_t __s1_288 = __p1_288; \ + float16x8_t __s2_288 = __p2_288; \ + float16x8_t __rev2_288; __rev2_288 = __builtin_shufflevector(__s2_288, __s2_288, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16_t __ret_288; \ + __ret_288 = __noswap_vfmah_laneq_f16(__s0_288, -__s1_288, __rev2_288, __p3_288); \ + __ret_288; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsq_laneq_f16(__p0_197, __p1_197, __p2_197, __p3_197) __extension__ ({ \ - float16x8_t __s0_197 = __p0_197; \ - float16x8_t __s1_197 = __p1_197; \ - float16x8_t __s2_197 = __p2_197; \ - float16x8_t __ret_197; \ - __ret_197 = vfmaq_laneq_f16(__s0_197, -__s1_197, __s2_197, __p3_197); \ - __ret_197; \ +#define vfmsq_laneq_f16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \ + float16x8_t __s0_289 = __p0_289; \ + float16x8_t __s1_289 = __p1_289; \ + float16x8_t __s2_289 = __p2_289; \ + float16x8_t __ret_289; \ + __ret_289 = vfmaq_laneq_f16(__s0_289, -__s1_289, __s2_289, __p3_289); \ + __ret_289; \ }) #else -#define vfmsq_laneq_f16(__p0_198, __p1_198, __p2_198, __p3_198) __extension__ ({ \ - float16x8_t __s0_198 = __p0_198; \ - float16x8_t __s1_198 = __p1_198; \ - float16x8_t __s2_198 = __p2_198; \ - float16x8_t __rev0_198; __rev0_198 = __builtin_shufflevector(__s0_198, __s0_198, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev1_198; __rev1_198 = __builtin_shufflevector(__s1_198, __s1_198, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev2_198; __rev2_198 = __builtin_shufflevector(__s2_198, __s2_198, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __ret_198; \ - __ret_198 = __noswap_vfmaq_laneq_f16(__rev0_198, -__rev1_198, __rev2_198, __p3_198); \ - __ret_198 = __builtin_shufflevector(__ret_198, __ret_198, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_198; \ +#define vfmsq_laneq_f16(__p0_290, __p1_290, __p2_290, __p3_290) __extension__ ({ \ + float16x8_t __s0_290 = __p0_290; \ + float16x8_t __s1_290 = __p1_290; \ + float16x8_t __s2_290 = __p2_290; \ + float16x8_t __rev0_290; __rev0_290 = __builtin_shufflevector(__s0_290, __s0_290, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_290; __rev1_290 = __builtin_shufflevector(__s1_290, __s1_290, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_290; __rev2_290 = __builtin_shufflevector(__s2_290, __s2_290, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_290; \ + __ret_290 = __noswap_vfmaq_laneq_f16(__rev0_290, -__rev1_290, __rev2_290, __p3_290); \ + __ret_290 = __builtin_shufflevector(__ret_290, __ret_290, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_290; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfms_laneq_f16(__p0_199, __p1_199, __p2_199, __p3_199) __extension__ ({ \ - float16x4_t __s0_199 = __p0_199; \ - float16x4_t __s1_199 = __p1_199; \ - float16x8_t __s2_199 = __p2_199; \ - float16x4_t __ret_199; \ - __ret_199 = vfma_laneq_f16(__s0_199, -__s1_199, __s2_199, __p3_199); \ - __ret_199; \ +#define vfms_laneq_f16(__p0_291, __p1_291, __p2_291, __p3_291) __extension__ ({ \ + float16x4_t __s0_291 = __p0_291; \ + float16x4_t __s1_291 = __p1_291; \ + float16x8_t __s2_291 = __p2_291; \ + float16x4_t __ret_291; \ + __ret_291 = vfma_laneq_f16(__s0_291, -__s1_291, __s2_291, __p3_291); \ + __ret_291; \ }) #else -#define vfms_laneq_f16(__p0_200, __p1_200, __p2_200, __p3_200) __extension__ ({ \ - float16x4_t __s0_200 = __p0_200; \ - float16x4_t __s1_200 = __p1_200; \ - float16x8_t __s2_200 = __p2_200; \ - float16x4_t __rev0_200; __rev0_200 = __builtin_shufflevector(__s0_200, __s0_200, 3, 2, 1, 0); \ - float16x4_t __rev1_200; __rev1_200 = __builtin_shufflevector(__s1_200, __s1_200, 3, 2, 1, 0); \ - float16x8_t __rev2_200; __rev2_200 = __builtin_shufflevector(__s2_200, __s2_200, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __ret_200; \ - __ret_200 = __noswap_vfma_laneq_f16(__rev0_200, -__rev1_200, __rev2_200, __p3_200); \ - __ret_200 = __builtin_shufflevector(__ret_200, __ret_200, 3, 2, 1, 0); \ - __ret_200; \ +#define vfms_laneq_f16(__p0_292, __p1_292, __p2_292, __p3_292) __extension__ ({ \ + float16x4_t __s0_292 = __p0_292; \ + float16x4_t __s1_292 = __p1_292; \ + float16x8_t __s2_292 = __p2_292; \ + float16x4_t __rev0_292; __rev0_292 = __builtin_shufflevector(__s0_292, __s0_292, 3, 2, 1, 0); \ + float16x4_t __rev1_292; __rev1_292 = __builtin_shufflevector(__s1_292, __s1_292, 3, 2, 1, 0); \ + float16x8_t __rev2_292; __rev2_292 = __builtin_shufflevector(__s2_292, __s2_292, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_292; \ + __ret_292 = __noswap_vfma_laneq_f16(__rev0_292, -__rev1_292, __rev2_292, __p3_292); \ + __ret_292 = __builtin_shufflevector(__ret_292, __ret_292, 3, 2, 1, 0); \ + __ret_292; \ }) #endif @@ -43617,44 +45353,44 @@ __ai float16x4_t vdiv_f16(float16x4_t __p0, float16x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_laneq_f16(__p0_201, __p1_201, __p2_201) __extension__ ({ \ - float16x8_t __s0_201 = __p0_201; \ - float16x8_t __s1_201 = __p1_201; \ - float16x8_t __ret_201; \ - __ret_201 = __s0_201 * splatq_laneq_f16(__s1_201, __p2_201); \ - __ret_201; \ +#define vmulq_laneq_f16(__p0_293, __p1_293, __p2_293) __extension__ ({ \ + float16x8_t __s0_293 = __p0_293; \ + float16x8_t __s1_293 = __p1_293; \ + float16x8_t __ret_293; \ + __ret_293 = __s0_293 * splatq_laneq_f16(__s1_293, __p2_293); \ + __ret_293; \ }) #else -#define vmulq_laneq_f16(__p0_202, __p1_202, __p2_202) __extension__ ({ \ - float16x8_t __s0_202 = __p0_202; \ - float16x8_t __s1_202 = __p1_202; \ - float16x8_t __rev0_202; __rev0_202 = __builtin_shufflevector(__s0_202, __s0_202, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev1_202; __rev1_202 = __builtin_shufflevector(__s1_202, __s1_202, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __ret_202; \ - __ret_202 = __rev0_202 * __noswap_splatq_laneq_f16(__rev1_202, __p2_202); \ - __ret_202 = __builtin_shufflevector(__ret_202, __ret_202, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_202; \ +#define vmulq_laneq_f16(__p0_294, __p1_294, __p2_294) __extension__ ({ \ + float16x8_t __s0_294 = __p0_294; \ + float16x8_t __s1_294 = __p1_294; \ + float16x8_t __rev0_294; __rev0_294 = __builtin_shufflevector(__s0_294, __s0_294, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_294; __rev1_294 = __builtin_shufflevector(__s1_294, __s1_294, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_294; \ + __ret_294 = __rev0_294 * __noswap_splatq_laneq_f16(__rev1_294, __p2_294); \ + __ret_294 = __builtin_shufflevector(__ret_294, __ret_294, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_294; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmul_laneq_f16(__p0_203, __p1_203, __p2_203) __extension__ ({ \ - float16x4_t __s0_203 = __p0_203; \ - float16x8_t __s1_203 = __p1_203; \ - float16x4_t __ret_203; \ - __ret_203 = __s0_203 * splat_laneq_f16(__s1_203, __p2_203); \ - __ret_203; \ +#define vmul_laneq_f16(__p0_295, __p1_295, __p2_295) __extension__ ({ \ + float16x4_t __s0_295 = __p0_295; \ + float16x8_t __s1_295 = __p1_295; \ + float16x4_t __ret_295; \ + __ret_295 = __s0_295 * splat_laneq_f16(__s1_295, __p2_295); \ + __ret_295; \ }) #else -#define vmul_laneq_f16(__p0_204, __p1_204, __p2_204) __extension__ ({ \ - float16x4_t __s0_204 = __p0_204; \ - float16x8_t __s1_204 = __p1_204; \ - float16x4_t __rev0_204; __rev0_204 = __builtin_shufflevector(__s0_204, __s0_204, 3, 2, 1, 0); \ - float16x8_t __rev1_204; __rev1_204 = __builtin_shufflevector(__s1_204, __s1_204, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __ret_204; \ - __ret_204 = __rev0_204 * __noswap_splat_laneq_f16(__rev1_204, __p2_204); \ - __ret_204 = __builtin_shufflevector(__ret_204, __ret_204, 3, 2, 1, 0); \ - __ret_204; \ +#define vmul_laneq_f16(__p0_296, __p1_296, __p2_296) __extension__ ({ \ + float16x4_t __s0_296 = __p0_296; \ + float16x8_t __s1_296 = __p1_296; \ + float16x4_t __rev0_296; __rev0_296 = __builtin_shufflevector(__s0_296, __s0_296, 3, 2, 1, 0); \ + float16x8_t __rev1_296; __rev1_296 = __builtin_shufflevector(__s1_296, __s1_296, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_296; \ + __ret_296 = __rev0_296 * __noswap_splat_laneq_f16(__rev1_296, __p2_296); \ + __ret_296 = __builtin_shufflevector(__ret_296, __ret_296, 3, 2, 1, 0); \ + __ret_296; \ }) #endif @@ -43722,44 +45458,44 @@ __ai float16x4_t __noswap_vmulx_f16(float16x4_t __p0, float16x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxq_lane_f16(__p0_205, __p1_205, __p2_205) __extension__ ({ \ - float16x8_t __s0_205 = __p0_205; \ - float16x4_t __s1_205 = __p1_205; \ - float16x8_t __ret_205; \ - __ret_205 = vmulxq_f16(__s0_205, splatq_lane_f16(__s1_205, __p2_205)); \ - __ret_205; \ +#define vmulxq_lane_f16(__p0_297, __p1_297, __p2_297) __extension__ ({ \ + float16x8_t __s0_297 = __p0_297; \ + float16x4_t __s1_297 = __p1_297; \ + float16x8_t __ret_297; \ + __ret_297 = vmulxq_f16(__s0_297, splatq_lane_f16(__s1_297, __p2_297)); \ + __ret_297; \ }) #else -#define vmulxq_lane_f16(__p0_206, __p1_206, __p2_206) __extension__ ({ \ - float16x8_t __s0_206 = __p0_206; \ - float16x4_t __s1_206 = __p1_206; \ - float16x8_t __rev0_206; __rev0_206 = __builtin_shufflevector(__s0_206, __s0_206, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __rev1_206; __rev1_206 = __builtin_shufflevector(__s1_206, __s1_206, 3, 2, 1, 0); \ - float16x8_t __ret_206; \ - __ret_206 = __noswap_vmulxq_f16(__rev0_206, __noswap_splatq_lane_f16(__rev1_206, __p2_206)); \ - __ret_206 = __builtin_shufflevector(__ret_206, __ret_206, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_206; \ +#define vmulxq_lane_f16(__p0_298, __p1_298, __p2_298) __extension__ ({ \ + float16x8_t __s0_298 = __p0_298; \ + float16x4_t __s1_298 = __p1_298; \ + float16x8_t __rev0_298; __rev0_298 = __builtin_shufflevector(__s0_298, __s0_298, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev1_298; __rev1_298 = __builtin_shufflevector(__s1_298, __s1_298, 3, 2, 1, 0); \ + float16x8_t __ret_298; \ + __ret_298 = __noswap_vmulxq_f16(__rev0_298, __noswap_splatq_lane_f16(__rev1_298, __p2_298)); \ + __ret_298 = __builtin_shufflevector(__ret_298, __ret_298, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_298; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulx_lane_f16(__p0_207, __p1_207, __p2_207) __extension__ ({ \ - float16x4_t __s0_207 = __p0_207; \ - float16x4_t __s1_207 = __p1_207; \ - float16x4_t __ret_207; \ - __ret_207 = vmulx_f16(__s0_207, splat_lane_f16(__s1_207, __p2_207)); \ - __ret_207; \ +#define vmulx_lane_f16(__p0_299, __p1_299, __p2_299) __extension__ ({ \ + float16x4_t __s0_299 = __p0_299; \ + float16x4_t __s1_299 = __p1_299; \ + float16x4_t __ret_299; \ + __ret_299 = vmulx_f16(__s0_299, splat_lane_f16(__s1_299, __p2_299)); \ + __ret_299; \ }) #else -#define vmulx_lane_f16(__p0_208, __p1_208, __p2_208) __extension__ ({ \ - float16x4_t __s0_208 = __p0_208; \ - float16x4_t __s1_208 = __p1_208; \ - float16x4_t __rev0_208; __rev0_208 = __builtin_shufflevector(__s0_208, __s0_208, 3, 2, 1, 0); \ - float16x4_t __rev1_208; __rev1_208 = __builtin_shufflevector(__s1_208, __s1_208, 3, 2, 1, 0); \ - float16x4_t __ret_208; \ - __ret_208 = __noswap_vmulx_f16(__rev0_208, __noswap_splat_lane_f16(__rev1_208, __p2_208)); \ - __ret_208 = __builtin_shufflevector(__ret_208, __ret_208, 3, 2, 1, 0); \ - __ret_208; \ +#define vmulx_lane_f16(__p0_300, __p1_300, __p2_300) __extension__ ({ \ + float16x4_t __s0_300 = __p0_300; \ + float16x4_t __s1_300 = __p1_300; \ + float16x4_t __rev0_300; __rev0_300 = __builtin_shufflevector(__s0_300, __s0_300, 3, 2, 1, 0); \ + float16x4_t __rev1_300; __rev1_300 = __builtin_shufflevector(__s1_300, __s1_300, 3, 2, 1, 0); \ + float16x4_t __ret_300; \ + __ret_300 = __noswap_vmulx_f16(__rev0_300, __noswap_splat_lane_f16(__rev1_300, __p2_300)); \ + __ret_300 = __builtin_shufflevector(__ret_300, __ret_300, 3, 2, 1, 0); \ + __ret_300; \ }) #endif @@ -43783,44 +45519,44 @@ __ai float16x4_t __noswap_vmulx_f16(float16x4_t __p0, float16x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxq_laneq_f16(__p0_209, __p1_209, __p2_209) __extension__ ({ \ - float16x8_t __s0_209 = __p0_209; \ - float16x8_t __s1_209 = __p1_209; \ - float16x8_t __ret_209; \ - __ret_209 = vmulxq_f16(__s0_209, splatq_laneq_f16(__s1_209, __p2_209)); \ - __ret_209; \ +#define vmulxq_laneq_f16(__p0_301, __p1_301, __p2_301) __extension__ ({ \ + float16x8_t __s0_301 = __p0_301; \ + float16x8_t __s1_301 = __p1_301; \ + float16x8_t __ret_301; \ + __ret_301 = vmulxq_f16(__s0_301, splatq_laneq_f16(__s1_301, __p2_301)); \ + __ret_301; \ }) #else -#define vmulxq_laneq_f16(__p0_210, __p1_210, __p2_210) __extension__ ({ \ - float16x8_t __s0_210 = __p0_210; \ - float16x8_t __s1_210 = __p1_210; \ - float16x8_t __rev0_210; __rev0_210 = __builtin_shufflevector(__s0_210, __s0_210, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev1_210; __rev1_210 = __builtin_shufflevector(__s1_210, __s1_210, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __ret_210; \ - __ret_210 = __noswap_vmulxq_f16(__rev0_210, __noswap_splatq_laneq_f16(__rev1_210, __p2_210)); \ - __ret_210 = __builtin_shufflevector(__ret_210, __ret_210, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_210; \ +#define vmulxq_laneq_f16(__p0_302, __p1_302, __p2_302) __extension__ ({ \ + float16x8_t __s0_302 = __p0_302; \ + float16x8_t __s1_302 = __p1_302; \ + float16x8_t __rev0_302; __rev0_302 = __builtin_shufflevector(__s0_302, __s0_302, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev1_302; __rev1_302 = __builtin_shufflevector(__s1_302, __s1_302, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_302; \ + __ret_302 = __noswap_vmulxq_f16(__rev0_302, __noswap_splatq_laneq_f16(__rev1_302, __p2_302)); \ + __ret_302 = __builtin_shufflevector(__ret_302, __ret_302, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_302; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulx_laneq_f16(__p0_211, __p1_211, __p2_211) __extension__ ({ \ - float16x4_t __s0_211 = __p0_211; \ - float16x8_t __s1_211 = __p1_211; \ - float16x4_t __ret_211; \ - __ret_211 = vmulx_f16(__s0_211, splat_laneq_f16(__s1_211, __p2_211)); \ - __ret_211; \ +#define vmulx_laneq_f16(__p0_303, __p1_303, __p2_303) __extension__ ({ \ + float16x4_t __s0_303 = __p0_303; \ + float16x8_t __s1_303 = __p1_303; \ + float16x4_t __ret_303; \ + __ret_303 = vmulx_f16(__s0_303, splat_laneq_f16(__s1_303, __p2_303)); \ + __ret_303; \ }) #else -#define vmulx_laneq_f16(__p0_212, __p1_212, __p2_212) __extension__ ({ \ - float16x4_t __s0_212 = __p0_212; \ - float16x8_t __s1_212 = __p1_212; \ - float16x4_t __rev0_212; __rev0_212 = __builtin_shufflevector(__s0_212, __s0_212, 3, 2, 1, 0); \ - float16x8_t __rev1_212; __rev1_212 = __builtin_shufflevector(__s1_212, __s1_212, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __ret_212; \ - __ret_212 = __noswap_vmulx_f16(__rev0_212, __noswap_splat_laneq_f16(__rev1_212, __p2_212)); \ - __ret_212 = __builtin_shufflevector(__ret_212, __ret_212, 3, 2, 1, 0); \ - __ret_212; \ +#define vmulx_laneq_f16(__p0_304, __p1_304, __p2_304) __extension__ ({ \ + float16x4_t __s0_304 = __p0_304; \ + float16x8_t __s1_304 = __p1_304; \ + float16x4_t __rev0_304; __rev0_304 = __builtin_shufflevector(__s0_304, __s0_304, 3, 2, 1, 0); \ + float16x8_t __rev1_304; __rev1_304 = __builtin_shufflevector(__s1_304, __s1_304, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_304; \ + __ret_304 = __noswap_vmulx_f16(__rev0_304, __noswap_splat_laneq_f16(__rev1_304, __p2_304)); \ + __ret_304 = __builtin_shufflevector(__ret_304, __ret_304, 3, 2, 1, 0); \ + __ret_304; \ }) #endif @@ -44336,54 +46072,54 @@ __ai int32x2_t __noswap_vusdot_s32(int32x2_t __p0, uint8x8_t __p1, int8x8_t __p2 #endif #ifdef __LITTLE_ENDIAN__ -#define vusdotq_lane_s32(__p0_213, __p1_213, __p2_213, __p3_213) __extension__ ({ \ - int32x4_t __s0_213 = __p0_213; \ - uint8x16_t __s1_213 = __p1_213; \ - int8x8_t __s2_213 = __p2_213; \ - int32x4_t __ret_213; \ -int8x8_t __reint_213 = __s2_213; \ - __ret_213 = vusdotq_s32(__s0_213, __s1_213, (int8x16_t)(splatq_lane_s32(*(int32x2_t *) &__reint_213, __p3_213))); \ - __ret_213; \ +#define vusdotq_lane_s32(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \ + int32x4_t __s0_305 = __p0_305; \ + uint8x16_t __s1_305 = __p1_305; \ + int8x8_t __s2_305 = __p2_305; \ + int32x4_t __ret_305; \ +int8x8_t __reint_305 = __s2_305; \ + __ret_305 = vusdotq_s32(__s0_305, __s1_305, (int8x16_t)(splatq_lane_s32(*(int32x2_t *) &__reint_305, __p3_305))); \ + __ret_305; \ }) #else -#define vusdotq_lane_s32(__p0_214, __p1_214, __p2_214, __p3_214) __extension__ ({ \ - int32x4_t __s0_214 = __p0_214; \ - uint8x16_t __s1_214 = __p1_214; \ - int8x8_t __s2_214 = __p2_214; \ - int32x4_t __rev0_214; __rev0_214 = __builtin_shufflevector(__s0_214, __s0_214, 3, 2, 1, 0); \ - uint8x16_t __rev1_214; __rev1_214 = __builtin_shufflevector(__s1_214, __s1_214, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __rev2_214; __rev2_214 = __builtin_shufflevector(__s2_214, __s2_214, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_214; \ -int8x8_t __reint_214 = __rev2_214; \ - __ret_214 = __noswap_vusdotq_s32(__rev0_214, __rev1_214, (int8x16_t)(__noswap_splatq_lane_s32(*(int32x2_t *) &__reint_214, __p3_214))); \ - __ret_214 = __builtin_shufflevector(__ret_214, __ret_214, 3, 2, 1, 0); \ - __ret_214; \ +#define vusdotq_lane_s32(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \ + int32x4_t __s0_306 = __p0_306; \ + uint8x16_t __s1_306 = __p1_306; \ + int8x8_t __s2_306 = __p2_306; \ + int32x4_t __rev0_306; __rev0_306 = __builtin_shufflevector(__s0_306, __s0_306, 3, 2, 1, 0); \ + uint8x16_t __rev1_306; __rev1_306 = __builtin_shufflevector(__s1_306, __s1_306, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __rev2_306; __rev2_306 = __builtin_shufflevector(__s2_306, __s2_306, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_306; \ +int8x8_t __reint_306 = __rev2_306; \ + __ret_306 = __noswap_vusdotq_s32(__rev0_306, __rev1_306, (int8x16_t)(__noswap_splatq_lane_s32(*(int32x2_t *) &__reint_306, __p3_306))); \ + __ret_306 = __builtin_shufflevector(__ret_306, __ret_306, 3, 2, 1, 0); \ + __ret_306; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vusdot_lane_s32(__p0_215, __p1_215, __p2_215, __p3_215) __extension__ ({ \ - int32x2_t __s0_215 = __p0_215; \ - uint8x8_t __s1_215 = __p1_215; \ - int8x8_t __s2_215 = __p2_215; \ - int32x2_t __ret_215; \ -int8x8_t __reint_215 = __s2_215; \ - __ret_215 = vusdot_s32(__s0_215, __s1_215, (int8x8_t)(splat_lane_s32(*(int32x2_t *) &__reint_215, __p3_215))); \ - __ret_215; \ +#define vusdot_lane_s32(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \ + int32x2_t __s0_307 = __p0_307; \ + uint8x8_t __s1_307 = __p1_307; \ + int8x8_t __s2_307 = __p2_307; \ + int32x2_t __ret_307; \ +int8x8_t __reint_307 = __s2_307; \ + __ret_307 = vusdot_s32(__s0_307, __s1_307, (int8x8_t)(splat_lane_s32(*(int32x2_t *) &__reint_307, __p3_307))); \ + __ret_307; \ }) #else -#define vusdot_lane_s32(__p0_216, __p1_216, __p2_216, __p3_216) __extension__ ({ \ - int32x2_t __s0_216 = __p0_216; \ - uint8x8_t __s1_216 = __p1_216; \ - int8x8_t __s2_216 = __p2_216; \ - int32x2_t __rev0_216; __rev0_216 = __builtin_shufflevector(__s0_216, __s0_216, 1, 0); \ - uint8x8_t __rev1_216; __rev1_216 = __builtin_shufflevector(__s1_216, __s1_216, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __rev2_216; __rev2_216 = __builtin_shufflevector(__s2_216, __s2_216, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x2_t __ret_216; \ -int8x8_t __reint_216 = __rev2_216; \ - __ret_216 = __noswap_vusdot_s32(__rev0_216, __rev1_216, (int8x8_t)(__noswap_splat_lane_s32(*(int32x2_t *) &__reint_216, __p3_216))); \ - __ret_216 = __builtin_shufflevector(__ret_216, __ret_216, 1, 0); \ - __ret_216; \ +#define vusdot_lane_s32(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \ + int32x2_t __s0_308 = __p0_308; \ + uint8x8_t __s1_308 = __p1_308; \ + int8x8_t __s2_308 = __p2_308; \ + int32x2_t __rev0_308; __rev0_308 = __builtin_shufflevector(__s0_308, __s0_308, 1, 0); \ + uint8x8_t __rev1_308; __rev1_308 = __builtin_shufflevector(__s1_308, __s1_308, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __rev2_308; __rev2_308 = __builtin_shufflevector(__s2_308, __s2_308, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x2_t __ret_308; \ +int8x8_t __reint_308 = __rev2_308; \ + __ret_308 = __noswap_vusdot_s32(__rev0_308, __rev1_308, (int8x8_t)(__noswap_splat_lane_s32(*(int32x2_t *) &__reint_308, __p3_308))); \ + __ret_308 = __builtin_shufflevector(__ret_308, __ret_308, 1, 0); \ + __ret_308; \ }) #endif @@ -44480,98 +46216,98 @@ __ai int16x4_t vqrdmlah_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) { #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahq_lane_s32(__p0_217, __p1_217, __p2_217, __p3_217) __extension__ ({ \ - int32x4_t __s0_217 = __p0_217; \ - int32x4_t __s1_217 = __p1_217; \ - int32x2_t __s2_217 = __p2_217; \ - int32x4_t __ret_217; \ - __ret_217 = vqaddq_s32(__s0_217, vqrdmulhq_s32(__s1_217, splatq_lane_s32(__s2_217, __p3_217))); \ - __ret_217; \ +#define vqrdmlahq_lane_s32(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \ + int32x4_t __s0_309 = __p0_309; \ + int32x4_t __s1_309 = __p1_309; \ + int32x2_t __s2_309 = __p2_309; \ + int32x4_t __ret_309; \ + __ret_309 = vqaddq_s32(__s0_309, vqrdmulhq_s32(__s1_309, splatq_lane_s32(__s2_309, __p3_309))); \ + __ret_309; \ }) #else -#define vqrdmlahq_lane_s32(__p0_218, __p1_218, __p2_218, __p3_218) __extension__ ({ \ - int32x4_t __s0_218 = __p0_218; \ - int32x4_t __s1_218 = __p1_218; \ - int32x2_t __s2_218 = __p2_218; \ - int32x4_t __rev0_218; __rev0_218 = __builtin_shufflevector(__s0_218, __s0_218, 3, 2, 1, 0); \ - int32x4_t __rev1_218; __rev1_218 = __builtin_shufflevector(__s1_218, __s1_218, 3, 2, 1, 0); \ - int32x2_t __rev2_218; __rev2_218 = __builtin_shufflevector(__s2_218, __s2_218, 1, 0); \ - int32x4_t __ret_218; \ - __ret_218 = __noswap_vqaddq_s32(__rev0_218, __noswap_vqrdmulhq_s32(__rev1_218, __noswap_splatq_lane_s32(__rev2_218, __p3_218))); \ - __ret_218 = __builtin_shufflevector(__ret_218, __ret_218, 3, 2, 1, 0); \ - __ret_218; \ +#define vqrdmlahq_lane_s32(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \ + int32x4_t __s0_310 = __p0_310; \ + int32x4_t __s1_310 = __p1_310; \ + int32x2_t __s2_310 = __p2_310; \ + int32x4_t __rev0_310; __rev0_310 = __builtin_shufflevector(__s0_310, __s0_310, 3, 2, 1, 0); \ + int32x4_t __rev1_310; __rev1_310 = __builtin_shufflevector(__s1_310, __s1_310, 3, 2, 1, 0); \ + int32x2_t __rev2_310; __rev2_310 = __builtin_shufflevector(__s2_310, __s2_310, 1, 0); \ + int32x4_t __ret_310; \ + __ret_310 = __noswap_vqaddq_s32(__rev0_310, __noswap_vqrdmulhq_s32(__rev1_310, __noswap_splatq_lane_s32(__rev2_310, __p3_310))); \ + __ret_310 = __builtin_shufflevector(__ret_310, __ret_310, 3, 2, 1, 0); \ + __ret_310; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahq_lane_s16(__p0_219, __p1_219, __p2_219, __p3_219) __extension__ ({ \ - int16x8_t __s0_219 = __p0_219; \ - int16x8_t __s1_219 = __p1_219; \ - int16x4_t __s2_219 = __p2_219; \ - int16x8_t __ret_219; \ - __ret_219 = vqaddq_s16(__s0_219, vqrdmulhq_s16(__s1_219, splatq_lane_s16(__s2_219, __p3_219))); \ - __ret_219; \ +#define vqrdmlahq_lane_s16(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \ + int16x8_t __s0_311 = __p0_311; \ + int16x8_t __s1_311 = __p1_311; \ + int16x4_t __s2_311 = __p2_311; \ + int16x8_t __ret_311; \ + __ret_311 = vqaddq_s16(__s0_311, vqrdmulhq_s16(__s1_311, splatq_lane_s16(__s2_311, __p3_311))); \ + __ret_311; \ }) #else -#define vqrdmlahq_lane_s16(__p0_220, __p1_220, __p2_220, __p3_220) __extension__ ({ \ - int16x8_t __s0_220 = __p0_220; \ - int16x8_t __s1_220 = __p1_220; \ - int16x4_t __s2_220 = __p2_220; \ - int16x8_t __rev0_220; __rev0_220 = __builtin_shufflevector(__s0_220, __s0_220, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_220; __rev1_220 = __builtin_shufflevector(__s1_220, __s1_220, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev2_220; __rev2_220 = __builtin_shufflevector(__s2_220, __s2_220, 3, 2, 1, 0); \ - int16x8_t __ret_220; \ - __ret_220 = __noswap_vqaddq_s16(__rev0_220, __noswap_vqrdmulhq_s16(__rev1_220, __noswap_splatq_lane_s16(__rev2_220, __p3_220))); \ - __ret_220 = __builtin_shufflevector(__ret_220, __ret_220, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_220; \ +#define vqrdmlahq_lane_s16(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \ + int16x8_t __s0_312 = __p0_312; \ + int16x8_t __s1_312 = __p1_312; \ + int16x4_t __s2_312 = __p2_312; \ + int16x8_t __rev0_312; __rev0_312 = __builtin_shufflevector(__s0_312, __s0_312, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_312; __rev1_312 = __builtin_shufflevector(__s1_312, __s1_312, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev2_312; __rev2_312 = __builtin_shufflevector(__s2_312, __s2_312, 3, 2, 1, 0); \ + int16x8_t __ret_312; \ + __ret_312 = __noswap_vqaddq_s16(__rev0_312, __noswap_vqrdmulhq_s16(__rev1_312, __noswap_splatq_lane_s16(__rev2_312, __p3_312))); \ + __ret_312 = __builtin_shufflevector(__ret_312, __ret_312, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_312; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlah_lane_s32(__p0_221, __p1_221, __p2_221, __p3_221) __extension__ ({ \ - int32x2_t __s0_221 = __p0_221; \ - int32x2_t __s1_221 = __p1_221; \ - int32x2_t __s2_221 = __p2_221; \ - int32x2_t __ret_221; \ - __ret_221 = vqadd_s32(__s0_221, vqrdmulh_s32(__s1_221, splat_lane_s32(__s2_221, __p3_221))); \ - __ret_221; \ +#define vqrdmlah_lane_s32(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \ + int32x2_t __s0_313 = __p0_313; \ + int32x2_t __s1_313 = __p1_313; \ + int32x2_t __s2_313 = __p2_313; \ + int32x2_t __ret_313; \ + __ret_313 = vqadd_s32(__s0_313, vqrdmulh_s32(__s1_313, splat_lane_s32(__s2_313, __p3_313))); \ + __ret_313; \ }) #else -#define vqrdmlah_lane_s32(__p0_222, __p1_222, __p2_222, __p3_222) __extension__ ({ \ - int32x2_t __s0_222 = __p0_222; \ - int32x2_t __s1_222 = __p1_222; \ - int32x2_t __s2_222 = __p2_222; \ - int32x2_t __rev0_222; __rev0_222 = __builtin_shufflevector(__s0_222, __s0_222, 1, 0); \ - int32x2_t __rev1_222; __rev1_222 = __builtin_shufflevector(__s1_222, __s1_222, 1, 0); \ - int32x2_t __rev2_222; __rev2_222 = __builtin_shufflevector(__s2_222, __s2_222, 1, 0); \ - int32x2_t __ret_222; \ - __ret_222 = __noswap_vqadd_s32(__rev0_222, __noswap_vqrdmulh_s32(__rev1_222, __noswap_splat_lane_s32(__rev2_222, __p3_222))); \ - __ret_222 = __builtin_shufflevector(__ret_222, __ret_222, 1, 0); \ - __ret_222; \ +#define vqrdmlah_lane_s32(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \ + int32x2_t __s0_314 = __p0_314; \ + int32x2_t __s1_314 = __p1_314; \ + int32x2_t __s2_314 = __p2_314; \ + int32x2_t __rev0_314; __rev0_314 = __builtin_shufflevector(__s0_314, __s0_314, 1, 0); \ + int32x2_t __rev1_314; __rev1_314 = __builtin_shufflevector(__s1_314, __s1_314, 1, 0); \ + int32x2_t __rev2_314; __rev2_314 = __builtin_shufflevector(__s2_314, __s2_314, 1, 0); \ + int32x2_t __ret_314; \ + __ret_314 = __noswap_vqadd_s32(__rev0_314, __noswap_vqrdmulh_s32(__rev1_314, __noswap_splat_lane_s32(__rev2_314, __p3_314))); \ + __ret_314 = __builtin_shufflevector(__ret_314, __ret_314, 1, 0); \ + __ret_314; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlah_lane_s16(__p0_223, __p1_223, __p2_223, __p3_223) __extension__ ({ \ - int16x4_t __s0_223 = __p0_223; \ - int16x4_t __s1_223 = __p1_223; \ - int16x4_t __s2_223 = __p2_223; \ - int16x4_t __ret_223; \ - __ret_223 = vqadd_s16(__s0_223, vqrdmulh_s16(__s1_223, splat_lane_s16(__s2_223, __p3_223))); \ - __ret_223; \ +#define vqrdmlah_lane_s16(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \ + int16x4_t __s0_315 = __p0_315; \ + int16x4_t __s1_315 = __p1_315; \ + int16x4_t __s2_315 = __p2_315; \ + int16x4_t __ret_315; \ + __ret_315 = vqadd_s16(__s0_315, vqrdmulh_s16(__s1_315, splat_lane_s16(__s2_315, __p3_315))); \ + __ret_315; \ }) #else -#define vqrdmlah_lane_s16(__p0_224, __p1_224, __p2_224, __p3_224) __extension__ ({ \ - int16x4_t __s0_224 = __p0_224; \ - int16x4_t __s1_224 = __p1_224; \ - int16x4_t __s2_224 = __p2_224; \ - int16x4_t __rev0_224; __rev0_224 = __builtin_shufflevector(__s0_224, __s0_224, 3, 2, 1, 0); \ - int16x4_t __rev1_224; __rev1_224 = __builtin_shufflevector(__s1_224, __s1_224, 3, 2, 1, 0); \ - int16x4_t __rev2_224; __rev2_224 = __builtin_shufflevector(__s2_224, __s2_224, 3, 2, 1, 0); \ - int16x4_t __ret_224; \ - __ret_224 = __noswap_vqadd_s16(__rev0_224, __noswap_vqrdmulh_s16(__rev1_224, __noswap_splat_lane_s16(__rev2_224, __p3_224))); \ - __ret_224 = __builtin_shufflevector(__ret_224, __ret_224, 3, 2, 1, 0); \ - __ret_224; \ +#define vqrdmlah_lane_s16(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \ + int16x4_t __s0_316 = __p0_316; \ + int16x4_t __s1_316 = __p1_316; \ + int16x4_t __s2_316 = __p2_316; \ + int16x4_t __rev0_316; __rev0_316 = __builtin_shufflevector(__s0_316, __s0_316, 3, 2, 1, 0); \ + int16x4_t __rev1_316; __rev1_316 = __builtin_shufflevector(__s1_316, __s1_316, 3, 2, 1, 0); \ + int16x4_t __rev2_316; __rev2_316 = __builtin_shufflevector(__s2_316, __s2_316, 3, 2, 1, 0); \ + int16x4_t __ret_316; \ + __ret_316 = __noswap_vqadd_s16(__rev0_316, __noswap_vqrdmulh_s16(__rev1_316, __noswap_splat_lane_s16(__rev2_316, __p3_316))); \ + __ret_316 = __builtin_shufflevector(__ret_316, __ret_316, 3, 2, 1, 0); \ + __ret_316; \ }) #endif @@ -44648,292 +46384,292 @@ __ai int16x4_t vqrdmlsh_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) { #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshq_lane_s32(__p0_225, __p1_225, __p2_225, __p3_225) __extension__ ({ \ - int32x4_t __s0_225 = __p0_225; \ - int32x4_t __s1_225 = __p1_225; \ - int32x2_t __s2_225 = __p2_225; \ - int32x4_t __ret_225; \ - __ret_225 = vqsubq_s32(__s0_225, vqrdmulhq_s32(__s1_225, splatq_lane_s32(__s2_225, __p3_225))); \ - __ret_225; \ +#define vqrdmlshq_lane_s32(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \ + int32x4_t __s0_317 = __p0_317; \ + int32x4_t __s1_317 = __p1_317; \ + int32x2_t __s2_317 = __p2_317; \ + int32x4_t __ret_317; \ + __ret_317 = vqsubq_s32(__s0_317, vqrdmulhq_s32(__s1_317, splatq_lane_s32(__s2_317, __p3_317))); \ + __ret_317; \ }) #else -#define vqrdmlshq_lane_s32(__p0_226, __p1_226, __p2_226, __p3_226) __extension__ ({ \ - int32x4_t __s0_226 = __p0_226; \ - int32x4_t __s1_226 = __p1_226; \ - int32x2_t __s2_226 = __p2_226; \ - int32x4_t __rev0_226; __rev0_226 = __builtin_shufflevector(__s0_226, __s0_226, 3, 2, 1, 0); \ - int32x4_t __rev1_226; __rev1_226 = __builtin_shufflevector(__s1_226, __s1_226, 3, 2, 1, 0); \ - int32x2_t __rev2_226; __rev2_226 = __builtin_shufflevector(__s2_226, __s2_226, 1, 0); \ - int32x4_t __ret_226; \ - __ret_226 = __noswap_vqsubq_s32(__rev0_226, __noswap_vqrdmulhq_s32(__rev1_226, __noswap_splatq_lane_s32(__rev2_226, __p3_226))); \ - __ret_226 = __builtin_shufflevector(__ret_226, __ret_226, 3, 2, 1, 0); \ - __ret_226; \ +#define vqrdmlshq_lane_s32(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \ + int32x4_t __s0_318 = __p0_318; \ + int32x4_t __s1_318 = __p1_318; \ + int32x2_t __s2_318 = __p2_318; \ + int32x4_t __rev0_318; __rev0_318 = __builtin_shufflevector(__s0_318, __s0_318, 3, 2, 1, 0); \ + int32x4_t __rev1_318; __rev1_318 = __builtin_shufflevector(__s1_318, __s1_318, 3, 2, 1, 0); \ + int32x2_t __rev2_318; __rev2_318 = __builtin_shufflevector(__s2_318, __s2_318, 1, 0); \ + int32x4_t __ret_318; \ + __ret_318 = __noswap_vqsubq_s32(__rev0_318, __noswap_vqrdmulhq_s32(__rev1_318, __noswap_splatq_lane_s32(__rev2_318, __p3_318))); \ + __ret_318 = __builtin_shufflevector(__ret_318, __ret_318, 3, 2, 1, 0); \ + __ret_318; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshq_lane_s16(__p0_227, __p1_227, __p2_227, __p3_227) __extension__ ({ \ - int16x8_t __s0_227 = __p0_227; \ - int16x8_t __s1_227 = __p1_227; \ - int16x4_t __s2_227 = __p2_227; \ - int16x8_t __ret_227; \ - __ret_227 = vqsubq_s16(__s0_227, vqrdmulhq_s16(__s1_227, splatq_lane_s16(__s2_227, __p3_227))); \ - __ret_227; \ +#define vqrdmlshq_lane_s16(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \ + int16x8_t __s0_319 = __p0_319; \ + int16x8_t __s1_319 = __p1_319; \ + int16x4_t __s2_319 = __p2_319; \ + int16x8_t __ret_319; \ + __ret_319 = vqsubq_s16(__s0_319, vqrdmulhq_s16(__s1_319, splatq_lane_s16(__s2_319, __p3_319))); \ + __ret_319; \ }) #else -#define vqrdmlshq_lane_s16(__p0_228, __p1_228, __p2_228, __p3_228) __extension__ ({ \ - int16x8_t __s0_228 = __p0_228; \ - int16x8_t __s1_228 = __p1_228; \ - int16x4_t __s2_228 = __p2_228; \ - int16x8_t __rev0_228; __rev0_228 = __builtin_shufflevector(__s0_228, __s0_228, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_228; __rev1_228 = __builtin_shufflevector(__s1_228, __s1_228, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev2_228; __rev2_228 = __builtin_shufflevector(__s2_228, __s2_228, 3, 2, 1, 0); \ - int16x8_t __ret_228; \ - __ret_228 = __noswap_vqsubq_s16(__rev0_228, __noswap_vqrdmulhq_s16(__rev1_228, __noswap_splatq_lane_s16(__rev2_228, __p3_228))); \ - __ret_228 = __builtin_shufflevector(__ret_228, __ret_228, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_228; \ +#define vqrdmlshq_lane_s16(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \ + int16x8_t __s0_320 = __p0_320; \ + int16x8_t __s1_320 = __p1_320; \ + int16x4_t __s2_320 = __p2_320; \ + int16x8_t __rev0_320; __rev0_320 = __builtin_shufflevector(__s0_320, __s0_320, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_320; __rev1_320 = __builtin_shufflevector(__s1_320, __s1_320, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev2_320; __rev2_320 = __builtin_shufflevector(__s2_320, __s2_320, 3, 2, 1, 0); \ + int16x8_t __ret_320; \ + __ret_320 = __noswap_vqsubq_s16(__rev0_320, __noswap_vqrdmulhq_s16(__rev1_320, __noswap_splatq_lane_s16(__rev2_320, __p3_320))); \ + __ret_320 = __builtin_shufflevector(__ret_320, __ret_320, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_320; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlsh_lane_s32(__p0_229, __p1_229, __p2_229, __p3_229) __extension__ ({ \ - int32x2_t __s0_229 = __p0_229; \ - int32x2_t __s1_229 = __p1_229; \ - int32x2_t __s2_229 = __p2_229; \ - int32x2_t __ret_229; \ - __ret_229 = vqsub_s32(__s0_229, vqrdmulh_s32(__s1_229, splat_lane_s32(__s2_229, __p3_229))); \ - __ret_229; \ +#define vqrdmlsh_lane_s32(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \ + int32x2_t __s0_321 = __p0_321; \ + int32x2_t __s1_321 = __p1_321; \ + int32x2_t __s2_321 = __p2_321; \ + int32x2_t __ret_321; \ + __ret_321 = vqsub_s32(__s0_321, vqrdmulh_s32(__s1_321, splat_lane_s32(__s2_321, __p3_321))); \ + __ret_321; \ }) #else -#define vqrdmlsh_lane_s32(__p0_230, __p1_230, __p2_230, __p3_230) __extension__ ({ \ - int32x2_t __s0_230 = __p0_230; \ - int32x2_t __s1_230 = __p1_230; \ - int32x2_t __s2_230 = __p2_230; \ - int32x2_t __rev0_230; __rev0_230 = __builtin_shufflevector(__s0_230, __s0_230, 1, 0); \ - int32x2_t __rev1_230; __rev1_230 = __builtin_shufflevector(__s1_230, __s1_230, 1, 0); \ - int32x2_t __rev2_230; __rev2_230 = __builtin_shufflevector(__s2_230, __s2_230, 1, 0); \ - int32x2_t __ret_230; \ - __ret_230 = __noswap_vqsub_s32(__rev0_230, __noswap_vqrdmulh_s32(__rev1_230, __noswap_splat_lane_s32(__rev2_230, __p3_230))); \ - __ret_230 = __builtin_shufflevector(__ret_230, __ret_230, 1, 0); \ - __ret_230; \ +#define vqrdmlsh_lane_s32(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \ + int32x2_t __s0_322 = __p0_322; \ + int32x2_t __s1_322 = __p1_322; \ + int32x2_t __s2_322 = __p2_322; \ + int32x2_t __rev0_322; __rev0_322 = __builtin_shufflevector(__s0_322, __s0_322, 1, 0); \ + int32x2_t __rev1_322; __rev1_322 = __builtin_shufflevector(__s1_322, __s1_322, 1, 0); \ + int32x2_t __rev2_322; __rev2_322 = __builtin_shufflevector(__s2_322, __s2_322, 1, 0); \ + int32x2_t __ret_322; \ + __ret_322 = __noswap_vqsub_s32(__rev0_322, __noswap_vqrdmulh_s32(__rev1_322, __noswap_splat_lane_s32(__rev2_322, __p3_322))); \ + __ret_322 = __builtin_shufflevector(__ret_322, __ret_322, 1, 0); \ + __ret_322; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlsh_lane_s16(__p0_231, __p1_231, __p2_231, __p3_231) __extension__ ({ \ - int16x4_t __s0_231 = __p0_231; \ - int16x4_t __s1_231 = __p1_231; \ - int16x4_t __s2_231 = __p2_231; \ - int16x4_t __ret_231; \ - __ret_231 = vqsub_s16(__s0_231, vqrdmulh_s16(__s1_231, splat_lane_s16(__s2_231, __p3_231))); \ - __ret_231; \ +#define vqrdmlsh_lane_s16(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \ + int16x4_t __s0_323 = __p0_323; \ + int16x4_t __s1_323 = __p1_323; \ + int16x4_t __s2_323 = __p2_323; \ + int16x4_t __ret_323; \ + __ret_323 = vqsub_s16(__s0_323, vqrdmulh_s16(__s1_323, splat_lane_s16(__s2_323, __p3_323))); \ + __ret_323; \ }) #else -#define vqrdmlsh_lane_s16(__p0_232, __p1_232, __p2_232, __p3_232) __extension__ ({ \ - int16x4_t __s0_232 = __p0_232; \ - int16x4_t __s1_232 = __p1_232; \ - int16x4_t __s2_232 = __p2_232; \ - int16x4_t __rev0_232; __rev0_232 = __builtin_shufflevector(__s0_232, __s0_232, 3, 2, 1, 0); \ - int16x4_t __rev1_232; __rev1_232 = __builtin_shufflevector(__s1_232, __s1_232, 3, 2, 1, 0); \ - int16x4_t __rev2_232; __rev2_232 = __builtin_shufflevector(__s2_232, __s2_232, 3, 2, 1, 0); \ - int16x4_t __ret_232; \ - __ret_232 = __noswap_vqsub_s16(__rev0_232, __noswap_vqrdmulh_s16(__rev1_232, __noswap_splat_lane_s16(__rev2_232, __p3_232))); \ - __ret_232 = __builtin_shufflevector(__ret_232, __ret_232, 3, 2, 1, 0); \ - __ret_232; \ +#define vqrdmlsh_lane_s16(__p0_324, __p1_324, __p2_324, __p3_324) __extension__ ({ \ + int16x4_t __s0_324 = __p0_324; \ + int16x4_t __s1_324 = __p1_324; \ + int16x4_t __s2_324 = __p2_324; \ + int16x4_t __rev0_324; __rev0_324 = __builtin_shufflevector(__s0_324, __s0_324, 3, 2, 1, 0); \ + int16x4_t __rev1_324; __rev1_324 = __builtin_shufflevector(__s1_324, __s1_324, 3, 2, 1, 0); \ + int16x4_t __rev2_324; __rev2_324 = __builtin_shufflevector(__s2_324, __s2_324, 3, 2, 1, 0); \ + int16x4_t __ret_324; \ + __ret_324 = __noswap_vqsub_s16(__rev0_324, __noswap_vqrdmulh_s16(__rev1_324, __noswap_splat_lane_s16(__rev2_324, __p3_324))); \ + __ret_324 = __builtin_shufflevector(__ret_324, __ret_324, 3, 2, 1, 0); \ + __ret_324; \ }) #endif #endif #if defined(__ARM_FEATURE_QRDMX) && defined(__aarch64__) #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahq_laneq_s32(__p0_233, __p1_233, __p2_233, __p3_233) __extension__ ({ \ - int32x4_t __s0_233 = __p0_233; \ - int32x4_t __s1_233 = __p1_233; \ - int32x4_t __s2_233 = __p2_233; \ - int32x4_t __ret_233; \ - __ret_233 = vqaddq_s32(__s0_233, vqrdmulhq_s32(__s1_233, splatq_laneq_s32(__s2_233, __p3_233))); \ - __ret_233; \ +#define vqrdmlahq_laneq_s32(__p0_325, __p1_325, __p2_325, __p3_325) __extension__ ({ \ + int32x4_t __s0_325 = __p0_325; \ + int32x4_t __s1_325 = __p1_325; \ + int32x4_t __s2_325 = __p2_325; \ + int32x4_t __ret_325; \ + __ret_325 = vqaddq_s32(__s0_325, vqrdmulhq_s32(__s1_325, splatq_laneq_s32(__s2_325, __p3_325))); \ + __ret_325; \ }) #else -#define vqrdmlahq_laneq_s32(__p0_234, __p1_234, __p2_234, __p3_234) __extension__ ({ \ - int32x4_t __s0_234 = __p0_234; \ - int32x4_t __s1_234 = __p1_234; \ - int32x4_t __s2_234 = __p2_234; \ - int32x4_t __rev0_234; __rev0_234 = __builtin_shufflevector(__s0_234, __s0_234, 3, 2, 1, 0); \ - int32x4_t __rev1_234; __rev1_234 = __builtin_shufflevector(__s1_234, __s1_234, 3, 2, 1, 0); \ - int32x4_t __rev2_234; __rev2_234 = __builtin_shufflevector(__s2_234, __s2_234, 3, 2, 1, 0); \ - int32x4_t __ret_234; \ - __ret_234 = __noswap_vqaddq_s32(__rev0_234, __noswap_vqrdmulhq_s32(__rev1_234, __noswap_splatq_laneq_s32(__rev2_234, __p3_234))); \ - __ret_234 = __builtin_shufflevector(__ret_234, __ret_234, 3, 2, 1, 0); \ - __ret_234; \ +#define vqrdmlahq_laneq_s32(__p0_326, __p1_326, __p2_326, __p3_326) __extension__ ({ \ + int32x4_t __s0_326 = __p0_326; \ + int32x4_t __s1_326 = __p1_326; \ + int32x4_t __s2_326 = __p2_326; \ + int32x4_t __rev0_326; __rev0_326 = __builtin_shufflevector(__s0_326, __s0_326, 3, 2, 1, 0); \ + int32x4_t __rev1_326; __rev1_326 = __builtin_shufflevector(__s1_326, __s1_326, 3, 2, 1, 0); \ + int32x4_t __rev2_326; __rev2_326 = __builtin_shufflevector(__s2_326, __s2_326, 3, 2, 1, 0); \ + int32x4_t __ret_326; \ + __ret_326 = __noswap_vqaddq_s32(__rev0_326, __noswap_vqrdmulhq_s32(__rev1_326, __noswap_splatq_laneq_s32(__rev2_326, __p3_326))); \ + __ret_326 = __builtin_shufflevector(__ret_326, __ret_326, 3, 2, 1, 0); \ + __ret_326; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahq_laneq_s16(__p0_235, __p1_235, __p2_235, __p3_235) __extension__ ({ \ - int16x8_t __s0_235 = __p0_235; \ - int16x8_t __s1_235 = __p1_235; \ - int16x8_t __s2_235 = __p2_235; \ - int16x8_t __ret_235; \ - __ret_235 = vqaddq_s16(__s0_235, vqrdmulhq_s16(__s1_235, splatq_laneq_s16(__s2_235, __p3_235))); \ - __ret_235; \ +#define vqrdmlahq_laneq_s16(__p0_327, __p1_327, __p2_327, __p3_327) __extension__ ({ \ + int16x8_t __s0_327 = __p0_327; \ + int16x8_t __s1_327 = __p1_327; \ + int16x8_t __s2_327 = __p2_327; \ + int16x8_t __ret_327; \ + __ret_327 = vqaddq_s16(__s0_327, vqrdmulhq_s16(__s1_327, splatq_laneq_s16(__s2_327, __p3_327))); \ + __ret_327; \ }) #else -#define vqrdmlahq_laneq_s16(__p0_236, __p1_236, __p2_236, __p3_236) __extension__ ({ \ - int16x8_t __s0_236 = __p0_236; \ - int16x8_t __s1_236 = __p1_236; \ - int16x8_t __s2_236 = __p2_236; \ - int16x8_t __rev0_236; __rev0_236 = __builtin_shufflevector(__s0_236, __s0_236, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_236; __rev1_236 = __builtin_shufflevector(__s1_236, __s1_236, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_236; __rev2_236 = __builtin_shufflevector(__s2_236, __s2_236, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_236; \ - __ret_236 = __noswap_vqaddq_s16(__rev0_236, __noswap_vqrdmulhq_s16(__rev1_236, __noswap_splatq_laneq_s16(__rev2_236, __p3_236))); \ - __ret_236 = __builtin_shufflevector(__ret_236, __ret_236, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_236; \ +#define vqrdmlahq_laneq_s16(__p0_328, __p1_328, __p2_328, __p3_328) __extension__ ({ \ + int16x8_t __s0_328 = __p0_328; \ + int16x8_t __s1_328 = __p1_328; \ + int16x8_t __s2_328 = __p2_328; \ + int16x8_t __rev0_328; __rev0_328 = __builtin_shufflevector(__s0_328, __s0_328, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_328; __rev1_328 = __builtin_shufflevector(__s1_328, __s1_328, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_328; __rev2_328 = __builtin_shufflevector(__s2_328, __s2_328, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_328; \ + __ret_328 = __noswap_vqaddq_s16(__rev0_328, __noswap_vqrdmulhq_s16(__rev1_328, __noswap_splatq_laneq_s16(__rev2_328, __p3_328))); \ + __ret_328 = __builtin_shufflevector(__ret_328, __ret_328, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_328; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlah_laneq_s32(__p0_237, __p1_237, __p2_237, __p3_237) __extension__ ({ \ - int32x2_t __s0_237 = __p0_237; \ - int32x2_t __s1_237 = __p1_237; \ - int32x4_t __s2_237 = __p2_237; \ - int32x2_t __ret_237; \ - __ret_237 = vqadd_s32(__s0_237, vqrdmulh_s32(__s1_237, splat_laneq_s32(__s2_237, __p3_237))); \ - __ret_237; \ +#define vqrdmlah_laneq_s32(__p0_329, __p1_329, __p2_329, __p3_329) __extension__ ({ \ + int32x2_t __s0_329 = __p0_329; \ + int32x2_t __s1_329 = __p1_329; \ + int32x4_t __s2_329 = __p2_329; \ + int32x2_t __ret_329; \ + __ret_329 = vqadd_s32(__s0_329, vqrdmulh_s32(__s1_329, splat_laneq_s32(__s2_329, __p3_329))); \ + __ret_329; \ }) #else -#define vqrdmlah_laneq_s32(__p0_238, __p1_238, __p2_238, __p3_238) __extension__ ({ \ - int32x2_t __s0_238 = __p0_238; \ - int32x2_t __s1_238 = __p1_238; \ - int32x4_t __s2_238 = __p2_238; \ - int32x2_t __rev0_238; __rev0_238 = __builtin_shufflevector(__s0_238, __s0_238, 1, 0); \ - int32x2_t __rev1_238; __rev1_238 = __builtin_shufflevector(__s1_238, __s1_238, 1, 0); \ - int32x4_t __rev2_238; __rev2_238 = __builtin_shufflevector(__s2_238, __s2_238, 3, 2, 1, 0); \ - int32x2_t __ret_238; \ - __ret_238 = __noswap_vqadd_s32(__rev0_238, __noswap_vqrdmulh_s32(__rev1_238, __noswap_splat_laneq_s32(__rev2_238, __p3_238))); \ - __ret_238 = __builtin_shufflevector(__ret_238, __ret_238, 1, 0); \ - __ret_238; \ +#define vqrdmlah_laneq_s32(__p0_330, __p1_330, __p2_330, __p3_330) __extension__ ({ \ + int32x2_t __s0_330 = __p0_330; \ + int32x2_t __s1_330 = __p1_330; \ + int32x4_t __s2_330 = __p2_330; \ + int32x2_t __rev0_330; __rev0_330 = __builtin_shufflevector(__s0_330, __s0_330, 1, 0); \ + int32x2_t __rev1_330; __rev1_330 = __builtin_shufflevector(__s1_330, __s1_330, 1, 0); \ + int32x4_t __rev2_330; __rev2_330 = __builtin_shufflevector(__s2_330, __s2_330, 3, 2, 1, 0); \ + int32x2_t __ret_330; \ + __ret_330 = __noswap_vqadd_s32(__rev0_330, __noswap_vqrdmulh_s32(__rev1_330, __noswap_splat_laneq_s32(__rev2_330, __p3_330))); \ + __ret_330 = __builtin_shufflevector(__ret_330, __ret_330, 1, 0); \ + __ret_330; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlah_laneq_s16(__p0_239, __p1_239, __p2_239, __p3_239) __extension__ ({ \ - int16x4_t __s0_239 = __p0_239; \ - int16x4_t __s1_239 = __p1_239; \ - int16x8_t __s2_239 = __p2_239; \ - int16x4_t __ret_239; \ - __ret_239 = vqadd_s16(__s0_239, vqrdmulh_s16(__s1_239, splat_laneq_s16(__s2_239, __p3_239))); \ - __ret_239; \ +#define vqrdmlah_laneq_s16(__p0_331, __p1_331, __p2_331, __p3_331) __extension__ ({ \ + int16x4_t __s0_331 = __p0_331; \ + int16x4_t __s1_331 = __p1_331; \ + int16x8_t __s2_331 = __p2_331; \ + int16x4_t __ret_331; \ + __ret_331 = vqadd_s16(__s0_331, vqrdmulh_s16(__s1_331, splat_laneq_s16(__s2_331, __p3_331))); \ + __ret_331; \ }) #else -#define vqrdmlah_laneq_s16(__p0_240, __p1_240, __p2_240, __p3_240) __extension__ ({ \ - int16x4_t __s0_240 = __p0_240; \ - int16x4_t __s1_240 = __p1_240; \ - int16x8_t __s2_240 = __p2_240; \ - int16x4_t __rev0_240; __rev0_240 = __builtin_shufflevector(__s0_240, __s0_240, 3, 2, 1, 0); \ - int16x4_t __rev1_240; __rev1_240 = __builtin_shufflevector(__s1_240, __s1_240, 3, 2, 1, 0); \ - int16x8_t __rev2_240; __rev2_240 = __builtin_shufflevector(__s2_240, __s2_240, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __ret_240; \ - __ret_240 = __noswap_vqadd_s16(__rev0_240, __noswap_vqrdmulh_s16(__rev1_240, __noswap_splat_laneq_s16(__rev2_240, __p3_240))); \ - __ret_240 = __builtin_shufflevector(__ret_240, __ret_240, 3, 2, 1, 0); \ - __ret_240; \ +#define vqrdmlah_laneq_s16(__p0_332, __p1_332, __p2_332, __p3_332) __extension__ ({ \ + int16x4_t __s0_332 = __p0_332; \ + int16x4_t __s1_332 = __p1_332; \ + int16x8_t __s2_332 = __p2_332; \ + int16x4_t __rev0_332; __rev0_332 = __builtin_shufflevector(__s0_332, __s0_332, 3, 2, 1, 0); \ + int16x4_t __rev1_332; __rev1_332 = __builtin_shufflevector(__s1_332, __s1_332, 3, 2, 1, 0); \ + int16x8_t __rev2_332; __rev2_332 = __builtin_shufflevector(__s2_332, __s2_332, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __ret_332; \ + __ret_332 = __noswap_vqadd_s16(__rev0_332, __noswap_vqrdmulh_s16(__rev1_332, __noswap_splat_laneq_s16(__rev2_332, __p3_332))); \ + __ret_332 = __builtin_shufflevector(__ret_332, __ret_332, 3, 2, 1, 0); \ + __ret_332; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshq_laneq_s32(__p0_241, __p1_241, __p2_241, __p3_241) __extension__ ({ \ - int32x4_t __s0_241 = __p0_241; \ - int32x4_t __s1_241 = __p1_241; \ - int32x4_t __s2_241 = __p2_241; \ - int32x4_t __ret_241; \ - __ret_241 = vqsubq_s32(__s0_241, vqrdmulhq_s32(__s1_241, splatq_laneq_s32(__s2_241, __p3_241))); \ - __ret_241; \ +#define vqrdmlshq_laneq_s32(__p0_333, __p1_333, __p2_333, __p3_333) __extension__ ({ \ + int32x4_t __s0_333 = __p0_333; \ + int32x4_t __s1_333 = __p1_333; \ + int32x4_t __s2_333 = __p2_333; \ + int32x4_t __ret_333; \ + __ret_333 = vqsubq_s32(__s0_333, vqrdmulhq_s32(__s1_333, splatq_laneq_s32(__s2_333, __p3_333))); \ + __ret_333; \ }) #else -#define vqrdmlshq_laneq_s32(__p0_242, __p1_242, __p2_242, __p3_242) __extension__ ({ \ - int32x4_t __s0_242 = __p0_242; \ - int32x4_t __s1_242 = __p1_242; \ - int32x4_t __s2_242 = __p2_242; \ - int32x4_t __rev0_242; __rev0_242 = __builtin_shufflevector(__s0_242, __s0_242, 3, 2, 1, 0); \ - int32x4_t __rev1_242; __rev1_242 = __builtin_shufflevector(__s1_242, __s1_242, 3, 2, 1, 0); \ - int32x4_t __rev2_242; __rev2_242 = __builtin_shufflevector(__s2_242, __s2_242, 3, 2, 1, 0); \ - int32x4_t __ret_242; \ - __ret_242 = __noswap_vqsubq_s32(__rev0_242, __noswap_vqrdmulhq_s32(__rev1_242, __noswap_splatq_laneq_s32(__rev2_242, __p3_242))); \ - __ret_242 = __builtin_shufflevector(__ret_242, __ret_242, 3, 2, 1, 0); \ - __ret_242; \ +#define vqrdmlshq_laneq_s32(__p0_334, __p1_334, __p2_334, __p3_334) __extension__ ({ \ + int32x4_t __s0_334 = __p0_334; \ + int32x4_t __s1_334 = __p1_334; \ + int32x4_t __s2_334 = __p2_334; \ + int32x4_t __rev0_334; __rev0_334 = __builtin_shufflevector(__s0_334, __s0_334, 3, 2, 1, 0); \ + int32x4_t __rev1_334; __rev1_334 = __builtin_shufflevector(__s1_334, __s1_334, 3, 2, 1, 0); \ + int32x4_t __rev2_334; __rev2_334 = __builtin_shufflevector(__s2_334, __s2_334, 3, 2, 1, 0); \ + int32x4_t __ret_334; \ + __ret_334 = __noswap_vqsubq_s32(__rev0_334, __noswap_vqrdmulhq_s32(__rev1_334, __noswap_splatq_laneq_s32(__rev2_334, __p3_334))); \ + __ret_334 = __builtin_shufflevector(__ret_334, __ret_334, 3, 2, 1, 0); \ + __ret_334; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshq_laneq_s16(__p0_243, __p1_243, __p2_243, __p3_243) __extension__ ({ \ - int16x8_t __s0_243 = __p0_243; \ - int16x8_t __s1_243 = __p1_243; \ - int16x8_t __s2_243 = __p2_243; \ - int16x8_t __ret_243; \ - __ret_243 = vqsubq_s16(__s0_243, vqrdmulhq_s16(__s1_243, splatq_laneq_s16(__s2_243, __p3_243))); \ - __ret_243; \ +#define vqrdmlshq_laneq_s16(__p0_335, __p1_335, __p2_335, __p3_335) __extension__ ({ \ + int16x8_t __s0_335 = __p0_335; \ + int16x8_t __s1_335 = __p1_335; \ + int16x8_t __s2_335 = __p2_335; \ + int16x8_t __ret_335; \ + __ret_335 = vqsubq_s16(__s0_335, vqrdmulhq_s16(__s1_335, splatq_laneq_s16(__s2_335, __p3_335))); \ + __ret_335; \ }) #else -#define vqrdmlshq_laneq_s16(__p0_244, __p1_244, __p2_244, __p3_244) __extension__ ({ \ - int16x8_t __s0_244 = __p0_244; \ - int16x8_t __s1_244 = __p1_244; \ - int16x8_t __s2_244 = __p2_244; \ - int16x8_t __rev0_244; __rev0_244 = __builtin_shufflevector(__s0_244, __s0_244, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_244; __rev1_244 = __builtin_shufflevector(__s1_244, __s1_244, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_244; __rev2_244 = __builtin_shufflevector(__s2_244, __s2_244, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_244; \ - __ret_244 = __noswap_vqsubq_s16(__rev0_244, __noswap_vqrdmulhq_s16(__rev1_244, __noswap_splatq_laneq_s16(__rev2_244, __p3_244))); \ - __ret_244 = __builtin_shufflevector(__ret_244, __ret_244, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_244; \ +#define vqrdmlshq_laneq_s16(__p0_336, __p1_336, __p2_336, __p3_336) __extension__ ({ \ + int16x8_t __s0_336 = __p0_336; \ + int16x8_t __s1_336 = __p1_336; \ + int16x8_t __s2_336 = __p2_336; \ + int16x8_t __rev0_336; __rev0_336 = __builtin_shufflevector(__s0_336, __s0_336, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_336; __rev1_336 = __builtin_shufflevector(__s1_336, __s1_336, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_336; __rev2_336 = __builtin_shufflevector(__s2_336, __s2_336, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_336; \ + __ret_336 = __noswap_vqsubq_s16(__rev0_336, __noswap_vqrdmulhq_s16(__rev1_336, __noswap_splatq_laneq_s16(__rev2_336, __p3_336))); \ + __ret_336 = __builtin_shufflevector(__ret_336, __ret_336, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_336; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlsh_laneq_s32(__p0_245, __p1_245, __p2_245, __p3_245) __extension__ ({ \ - int32x2_t __s0_245 = __p0_245; \ - int32x2_t __s1_245 = __p1_245; \ - int32x4_t __s2_245 = __p2_245; \ - int32x2_t __ret_245; \ - __ret_245 = vqsub_s32(__s0_245, vqrdmulh_s32(__s1_245, splat_laneq_s32(__s2_245, __p3_245))); \ - __ret_245; \ +#define vqrdmlsh_laneq_s32(__p0_337, __p1_337, __p2_337, __p3_337) __extension__ ({ \ + int32x2_t __s0_337 = __p0_337; \ + int32x2_t __s1_337 = __p1_337; \ + int32x4_t __s2_337 = __p2_337; \ + int32x2_t __ret_337; \ + __ret_337 = vqsub_s32(__s0_337, vqrdmulh_s32(__s1_337, splat_laneq_s32(__s2_337, __p3_337))); \ + __ret_337; \ }) #else -#define vqrdmlsh_laneq_s32(__p0_246, __p1_246, __p2_246, __p3_246) __extension__ ({ \ - int32x2_t __s0_246 = __p0_246; \ - int32x2_t __s1_246 = __p1_246; \ - int32x4_t __s2_246 = __p2_246; \ - int32x2_t __rev0_246; __rev0_246 = __builtin_shufflevector(__s0_246, __s0_246, 1, 0); \ - int32x2_t __rev1_246; __rev1_246 = __builtin_shufflevector(__s1_246, __s1_246, 1, 0); \ - int32x4_t __rev2_246; __rev2_246 = __builtin_shufflevector(__s2_246, __s2_246, 3, 2, 1, 0); \ - int32x2_t __ret_246; \ - __ret_246 = __noswap_vqsub_s32(__rev0_246, __noswap_vqrdmulh_s32(__rev1_246, __noswap_splat_laneq_s32(__rev2_246, __p3_246))); \ - __ret_246 = __builtin_shufflevector(__ret_246, __ret_246, 1, 0); \ - __ret_246; \ +#define vqrdmlsh_laneq_s32(__p0_338, __p1_338, __p2_338, __p3_338) __extension__ ({ \ + int32x2_t __s0_338 = __p0_338; \ + int32x2_t __s1_338 = __p1_338; \ + int32x4_t __s2_338 = __p2_338; \ + int32x2_t __rev0_338; __rev0_338 = __builtin_shufflevector(__s0_338, __s0_338, 1, 0); \ + int32x2_t __rev1_338; __rev1_338 = __builtin_shufflevector(__s1_338, __s1_338, 1, 0); \ + int32x4_t __rev2_338; __rev2_338 = __builtin_shufflevector(__s2_338, __s2_338, 3, 2, 1, 0); \ + int32x2_t __ret_338; \ + __ret_338 = __noswap_vqsub_s32(__rev0_338, __noswap_vqrdmulh_s32(__rev1_338, __noswap_splat_laneq_s32(__rev2_338, __p3_338))); \ + __ret_338 = __builtin_shufflevector(__ret_338, __ret_338, 1, 0); \ + __ret_338; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlsh_laneq_s16(__p0_247, __p1_247, __p2_247, __p3_247) __extension__ ({ \ - int16x4_t __s0_247 = __p0_247; \ - int16x4_t __s1_247 = __p1_247; \ - int16x8_t __s2_247 = __p2_247; \ - int16x4_t __ret_247; \ - __ret_247 = vqsub_s16(__s0_247, vqrdmulh_s16(__s1_247, splat_laneq_s16(__s2_247, __p3_247))); \ - __ret_247; \ +#define vqrdmlsh_laneq_s16(__p0_339, __p1_339, __p2_339, __p3_339) __extension__ ({ \ + int16x4_t __s0_339 = __p0_339; \ + int16x4_t __s1_339 = __p1_339; \ + int16x8_t __s2_339 = __p2_339; \ + int16x4_t __ret_339; \ + __ret_339 = vqsub_s16(__s0_339, vqrdmulh_s16(__s1_339, splat_laneq_s16(__s2_339, __p3_339))); \ + __ret_339; \ }) #else -#define vqrdmlsh_laneq_s16(__p0_248, __p1_248, __p2_248, __p3_248) __extension__ ({ \ - int16x4_t __s0_248 = __p0_248; \ - int16x4_t __s1_248 = __p1_248; \ - int16x8_t __s2_248 = __p2_248; \ - int16x4_t __rev0_248; __rev0_248 = __builtin_shufflevector(__s0_248, __s0_248, 3, 2, 1, 0); \ - int16x4_t __rev1_248; __rev1_248 = __builtin_shufflevector(__s1_248, __s1_248, 3, 2, 1, 0); \ - int16x8_t __rev2_248; __rev2_248 = __builtin_shufflevector(__s2_248, __s2_248, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __ret_248; \ - __ret_248 = __noswap_vqsub_s16(__rev0_248, __noswap_vqrdmulh_s16(__rev1_248, __noswap_splat_laneq_s16(__rev2_248, __p3_248))); \ - __ret_248 = __builtin_shufflevector(__ret_248, __ret_248, 3, 2, 1, 0); \ - __ret_248; \ +#define vqrdmlsh_laneq_s16(__p0_340, __p1_340, __p2_340, __p3_340) __extension__ ({ \ + int16x4_t __s0_340 = __p0_340; \ + int16x4_t __s1_340 = __p1_340; \ + int16x8_t __s2_340 = __p2_340; \ + int16x4_t __rev0_340; __rev0_340 = __builtin_shufflevector(__s0_340, __s0_340, 3, 2, 1, 0); \ + int16x4_t __rev1_340; __rev1_340 = __builtin_shufflevector(__s1_340, __s1_340, 3, 2, 1, 0); \ + int16x8_t __rev2_340; __rev2_340 = __builtin_shufflevector(__s2_340, __s2_340, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __ret_340; \ + __ret_340 = __noswap_vqsub_s16(__rev0_340, __noswap_vqrdmulh_s16(__rev1_340, __noswap_splat_laneq_s16(__rev2_340, __p3_340))); \ + __ret_340 = __builtin_shufflevector(__ret_340, __ret_340, 3, 2, 1, 0); \ + __ret_340; \ }) #endif @@ -45860,9 +47596,9 @@ __ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) { __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1); return __ret; } -__ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1); +__ai uint64_t vceqd_s64(int64_t __p0, int64_t __p1) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vceqd_s64(__p0, __p1); return __ret; } __ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) { @@ -45897,22 +47633,6 @@ __ai uint64x1_t vceqz_p64(poly64x1_t __p0) { return __ret; } #ifdef __LITTLE_ENDIAN__ -__ai uint16x4_t vceqz_p16(poly16x4_t __p0) { - uint16x4_t __ret; - __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17); - return __ret; -} -#else -__ai uint16x4_t vceqz_p16(poly16x4_t __p0) { - poly16x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); - uint16x4_t __ret; - __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17); - __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); - return __ret; -} -#endif - -#ifdef __LITTLE_ENDIAN__ __ai uint8x16_t vceqzq_p8(poly8x16_t __p0) { uint8x16_t __ret; __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48); @@ -45945,22 +47665,6 @@ __ai uint64x2_t vceqzq_p64(poly64x2_t __p0) { #endif #ifdef __LITTLE_ENDIAN__ -__ai uint16x8_t vceqzq_p16(poly16x8_t __p0) { - uint16x8_t __ret; - __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49); - return __ret; -} -#else -__ai uint16x8_t vceqzq_p16(poly16x8_t __p0) { - poly16x8_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0); - uint16x8_t __ret; - __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49); - __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); - return __ret; -} -#endif - -#ifdef __LITTLE_ENDIAN__ __ai uint8x16_t vceqzq_u8(uint8x16_t __p0) { uint8x16_t __ret; __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48); @@ -46252,9 +47956,9 @@ __ai uint64_t vceqzd_u64(uint64_t __p0) { __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0); return __ret; } -__ai int64_t vceqzd_s64(int64_t __p0) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0); +__ai uint64_t vceqzd_s64(int64_t __p0) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vceqzd_s64(__p0); return __ret; } __ai uint64_t vceqzd_f64(float64_t __p0) { @@ -46333,9 +48037,9 @@ __ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) { __ret = (uint64x1_t)(__p0 >= __p1); return __ret; } -__ai int64_t vcged_s64(int64_t __p0, int64_t __p1) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1); +__ai uint64_t vcged_s64(int64_t __p0, int64_t __p1) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vcged_s64(__p0, __p1); return __ret; } __ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) { @@ -46523,9 +48227,9 @@ __ai uint16x4_t vcgez_s16(int16x4_t __p0) { } #endif -__ai int64_t vcgezd_s64(int64_t __p0) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0); +__ai uint64_t vcgezd_s64(int64_t __p0) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vcgezd_s64(__p0); return __ret; } __ai uint64_t vcgezd_f64(float64_t __p0) { @@ -46604,9 +48308,9 @@ __ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) { __ret = (uint64x1_t)(__p0 > __p1); return __ret; } -__ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1); +__ai uint64_t vcgtd_s64(int64_t __p0, int64_t __p1) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vcgtd_s64(__p0, __p1); return __ret; } __ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) { @@ -46794,9 +48498,9 @@ __ai uint16x4_t vcgtz_s16(int16x4_t __p0) { } #endif -__ai int64_t vcgtzd_s64(int64_t __p0) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0); +__ai uint64_t vcgtzd_s64(int64_t __p0) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vcgtzd_s64(__p0); return __ret; } __ai uint64_t vcgtzd_f64(float64_t __p0) { @@ -46880,9 +48584,9 @@ __ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) { __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1); return __ret; } -__ai int64_t vcled_s64(int64_t __p0, int64_t __p1) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1); +__ai uint64_t vcled_s64(int64_t __p0, int64_t __p1) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vcled_s64(__p0, __p1); return __ret; } __ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) { @@ -47065,9 +48769,9 @@ __ai uint16x4_t vclez_s16(int16x4_t __p0) { } #endif -__ai int64_t vclezd_s64(int64_t __p0) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vclezd_s64(__p0); +__ai uint64_t vclezd_s64(int64_t __p0) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vclezd_s64(__p0); return __ret; } __ai uint64_t vclezd_f64(float64_t __p0) { @@ -47151,9 +48855,9 @@ __ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) { __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1); return __ret; } -__ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1); +__ai uint64_t vcltd_s64(int64_t __p0, int64_t __p1) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vcltd_s64(__p0, __p1); return __ret; } __ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) { @@ -47336,9 +49040,9 @@ __ai uint16x4_t vcltz_s16(int16x4_t __p0) { } #endif -__ai int64_t vcltzd_s64(int64_t __p0) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0); +__ai uint64_t vcltzd_s64(int64_t __p0) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vcltzd_s64(__p0); return __ret; } __ai uint64_t vcltzd_f64(float64_t __p0) { @@ -47382,892 +49086,892 @@ __ai float64x2_t vcombine_f64(float64x1_t __p0, float64x1_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_p8(__p0_249, __p1_249, __p2_249, __p3_249) __extension__ ({ \ - poly8x16_t __s0_249 = __p0_249; \ - poly8x8_t __s2_249 = __p2_249; \ - poly8x16_t __ret_249; \ - __ret_249 = vsetq_lane_p8(vget_lane_p8(__s2_249, __p3_249), __s0_249, __p1_249); \ - __ret_249; \ +#define vcopyq_lane_p8(__p0_341, __p1_341, __p2_341, __p3_341) __extension__ ({ \ + poly8x16_t __s0_341 = __p0_341; \ + poly8x8_t __s2_341 = __p2_341; \ + poly8x16_t __ret_341; \ + __ret_341 = vsetq_lane_p8(vget_lane_p8(__s2_341, __p3_341), __s0_341, __p1_341); \ + __ret_341; \ }) #else -#define vcopyq_lane_p8(__p0_250, __p1_250, __p2_250, __p3_250) __extension__ ({ \ - poly8x16_t __s0_250 = __p0_250; \ - poly8x8_t __s2_250 = __p2_250; \ - poly8x16_t __rev0_250; __rev0_250 = __builtin_shufflevector(__s0_250, __s0_250, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x8_t __rev2_250; __rev2_250 = __builtin_shufflevector(__s2_250, __s2_250, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x16_t __ret_250; \ - __ret_250 = __noswap_vsetq_lane_p8(__noswap_vget_lane_p8(__rev2_250, __p3_250), __rev0_250, __p1_250); \ - __ret_250 = __builtin_shufflevector(__ret_250, __ret_250, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_250; \ +#define vcopyq_lane_p8(__p0_342, __p1_342, __p2_342, __p3_342) __extension__ ({ \ + poly8x16_t __s0_342 = __p0_342; \ + poly8x8_t __s2_342 = __p2_342; \ + poly8x16_t __rev0_342; __rev0_342 = __builtin_shufflevector(__s0_342, __s0_342, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x8_t __rev2_342; __rev2_342 = __builtin_shufflevector(__s2_342, __s2_342, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x16_t __ret_342; \ + __ret_342 = __noswap_vsetq_lane_p8(__noswap_vget_lane_p8(__rev2_342, __p3_342), __rev0_342, __p1_342); \ + __ret_342 = __builtin_shufflevector(__ret_342, __ret_342, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_342; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_p16(__p0_251, __p1_251, __p2_251, __p3_251) __extension__ ({ \ - poly16x8_t __s0_251 = __p0_251; \ - poly16x4_t __s2_251 = __p2_251; \ - poly16x8_t __ret_251; \ - __ret_251 = vsetq_lane_p16(vget_lane_p16(__s2_251, __p3_251), __s0_251, __p1_251); \ - __ret_251; \ +#define vcopyq_lane_p16(__p0_343, __p1_343, __p2_343, __p3_343) __extension__ ({ \ + poly16x8_t __s0_343 = __p0_343; \ + poly16x4_t __s2_343 = __p2_343; \ + poly16x8_t __ret_343; \ + __ret_343 = vsetq_lane_p16(vget_lane_p16(__s2_343, __p3_343), __s0_343, __p1_343); \ + __ret_343; \ }) #else -#define vcopyq_lane_p16(__p0_252, __p1_252, __p2_252, __p3_252) __extension__ ({ \ - poly16x8_t __s0_252 = __p0_252; \ - poly16x4_t __s2_252 = __p2_252; \ - poly16x8_t __rev0_252; __rev0_252 = __builtin_shufflevector(__s0_252, __s0_252, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly16x4_t __rev2_252; __rev2_252 = __builtin_shufflevector(__s2_252, __s2_252, 3, 2, 1, 0); \ - poly16x8_t __ret_252; \ - __ret_252 = __noswap_vsetq_lane_p16(__noswap_vget_lane_p16(__rev2_252, __p3_252), __rev0_252, __p1_252); \ - __ret_252 = __builtin_shufflevector(__ret_252, __ret_252, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_252; \ +#define vcopyq_lane_p16(__p0_344, __p1_344, __p2_344, __p3_344) __extension__ ({ \ + poly16x8_t __s0_344 = __p0_344; \ + poly16x4_t __s2_344 = __p2_344; \ + poly16x8_t __rev0_344; __rev0_344 = __builtin_shufflevector(__s0_344, __s0_344, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly16x4_t __rev2_344; __rev2_344 = __builtin_shufflevector(__s2_344, __s2_344, 3, 2, 1, 0); \ + poly16x8_t __ret_344; \ + __ret_344 = __noswap_vsetq_lane_p16(__noswap_vget_lane_p16(__rev2_344, __p3_344), __rev0_344, __p1_344); \ + __ret_344 = __builtin_shufflevector(__ret_344, __ret_344, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_344; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_u8(__p0_253, __p1_253, __p2_253, __p3_253) __extension__ ({ \ - uint8x16_t __s0_253 = __p0_253; \ - uint8x8_t __s2_253 = __p2_253; \ - uint8x16_t __ret_253; \ - __ret_253 = vsetq_lane_u8(vget_lane_u8(__s2_253, __p3_253), __s0_253, __p1_253); \ - __ret_253; \ +#define vcopyq_lane_u8(__p0_345, __p1_345, __p2_345, __p3_345) __extension__ ({ \ + uint8x16_t __s0_345 = __p0_345; \ + uint8x8_t __s2_345 = __p2_345; \ + uint8x16_t __ret_345; \ + __ret_345 = vsetq_lane_u8(vget_lane_u8(__s2_345, __p3_345), __s0_345, __p1_345); \ + __ret_345; \ }) #else -#define vcopyq_lane_u8(__p0_254, __p1_254, __p2_254, __p3_254) __extension__ ({ \ - uint8x16_t __s0_254 = __p0_254; \ - uint8x8_t __s2_254 = __p2_254; \ - uint8x16_t __rev0_254; __rev0_254 = __builtin_shufflevector(__s0_254, __s0_254, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __rev2_254; __rev2_254 = __builtin_shufflevector(__s2_254, __s2_254, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __ret_254; \ - __ret_254 = __noswap_vsetq_lane_u8(__noswap_vget_lane_u8(__rev2_254, __p3_254), __rev0_254, __p1_254); \ - __ret_254 = __builtin_shufflevector(__ret_254, __ret_254, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_254; \ +#define vcopyq_lane_u8(__p0_346, __p1_346, __p2_346, __p3_346) __extension__ ({ \ + uint8x16_t __s0_346 = __p0_346; \ + uint8x8_t __s2_346 = __p2_346; \ + uint8x16_t __rev0_346; __rev0_346 = __builtin_shufflevector(__s0_346, __s0_346, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __rev2_346; __rev2_346 = __builtin_shufflevector(__s2_346, __s2_346, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __ret_346; \ + __ret_346 = __noswap_vsetq_lane_u8(__noswap_vget_lane_u8(__rev2_346, __p3_346), __rev0_346, __p1_346); \ + __ret_346 = __builtin_shufflevector(__ret_346, __ret_346, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_346; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_u32(__p0_255, __p1_255, __p2_255, __p3_255) __extension__ ({ \ - uint32x4_t __s0_255 = __p0_255; \ - uint32x2_t __s2_255 = __p2_255; \ - uint32x4_t __ret_255; \ - __ret_255 = vsetq_lane_u32(vget_lane_u32(__s2_255, __p3_255), __s0_255, __p1_255); \ - __ret_255; \ +#define vcopyq_lane_u32(__p0_347, __p1_347, __p2_347, __p3_347) __extension__ ({ \ + uint32x4_t __s0_347 = __p0_347; \ + uint32x2_t __s2_347 = __p2_347; \ + uint32x4_t __ret_347; \ + __ret_347 = vsetq_lane_u32(vget_lane_u32(__s2_347, __p3_347), __s0_347, __p1_347); \ + __ret_347; \ }) #else -#define vcopyq_lane_u32(__p0_256, __p1_256, __p2_256, __p3_256) __extension__ ({ \ - uint32x4_t __s0_256 = __p0_256; \ - uint32x2_t __s2_256 = __p2_256; \ - uint32x4_t __rev0_256; __rev0_256 = __builtin_shufflevector(__s0_256, __s0_256, 3, 2, 1, 0); \ - uint32x2_t __rev2_256; __rev2_256 = __builtin_shufflevector(__s2_256, __s2_256, 1, 0); \ - uint32x4_t __ret_256; \ - __ret_256 = __noswap_vsetq_lane_u32(__noswap_vget_lane_u32(__rev2_256, __p3_256), __rev0_256, __p1_256); \ - __ret_256 = __builtin_shufflevector(__ret_256, __ret_256, 3, 2, 1, 0); \ - __ret_256; \ +#define vcopyq_lane_u32(__p0_348, __p1_348, __p2_348, __p3_348) __extension__ ({ \ + uint32x4_t __s0_348 = __p0_348; \ + uint32x2_t __s2_348 = __p2_348; \ + uint32x4_t __rev0_348; __rev0_348 = __builtin_shufflevector(__s0_348, __s0_348, 3, 2, 1, 0); \ + uint32x2_t __rev2_348; __rev2_348 = __builtin_shufflevector(__s2_348, __s2_348, 1, 0); \ + uint32x4_t __ret_348; \ + __ret_348 = __noswap_vsetq_lane_u32(__noswap_vget_lane_u32(__rev2_348, __p3_348), __rev0_348, __p1_348); \ + __ret_348 = __builtin_shufflevector(__ret_348, __ret_348, 3, 2, 1, 0); \ + __ret_348; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_u64(__p0_257, __p1_257, __p2_257, __p3_257) __extension__ ({ \ - uint64x2_t __s0_257 = __p0_257; \ - uint64x1_t __s2_257 = __p2_257; \ - uint64x2_t __ret_257; \ - __ret_257 = vsetq_lane_u64(vget_lane_u64(__s2_257, __p3_257), __s0_257, __p1_257); \ - __ret_257; \ +#define vcopyq_lane_u64(__p0_349, __p1_349, __p2_349, __p3_349) __extension__ ({ \ + uint64x2_t __s0_349 = __p0_349; \ + uint64x1_t __s2_349 = __p2_349; \ + uint64x2_t __ret_349; \ + __ret_349 = vsetq_lane_u64(vget_lane_u64(__s2_349, __p3_349), __s0_349, __p1_349); \ + __ret_349; \ }) #else -#define vcopyq_lane_u64(__p0_258, __p1_258, __p2_258, __p3_258) __extension__ ({ \ - uint64x2_t __s0_258 = __p0_258; \ - uint64x1_t __s2_258 = __p2_258; \ - uint64x2_t __rev0_258; __rev0_258 = __builtin_shufflevector(__s0_258, __s0_258, 1, 0); \ - uint64x2_t __ret_258; \ - __ret_258 = __noswap_vsetq_lane_u64(vget_lane_u64(__s2_258, __p3_258), __rev0_258, __p1_258); \ - __ret_258 = __builtin_shufflevector(__ret_258, __ret_258, 1, 0); \ - __ret_258; \ +#define vcopyq_lane_u64(__p0_350, __p1_350, __p2_350, __p3_350) __extension__ ({ \ + uint64x2_t __s0_350 = __p0_350; \ + uint64x1_t __s2_350 = __p2_350; \ + uint64x2_t __rev0_350; __rev0_350 = __builtin_shufflevector(__s0_350, __s0_350, 1, 0); \ + uint64x2_t __ret_350; \ + __ret_350 = __noswap_vsetq_lane_u64(vget_lane_u64(__s2_350, __p3_350), __rev0_350, __p1_350); \ + __ret_350 = __builtin_shufflevector(__ret_350, __ret_350, 1, 0); \ + __ret_350; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_u16(__p0_259, __p1_259, __p2_259, __p3_259) __extension__ ({ \ - uint16x8_t __s0_259 = __p0_259; \ - uint16x4_t __s2_259 = __p2_259; \ - uint16x8_t __ret_259; \ - __ret_259 = vsetq_lane_u16(vget_lane_u16(__s2_259, __p3_259), __s0_259, __p1_259); \ - __ret_259; \ +#define vcopyq_lane_u16(__p0_351, __p1_351, __p2_351, __p3_351) __extension__ ({ \ + uint16x8_t __s0_351 = __p0_351; \ + uint16x4_t __s2_351 = __p2_351; \ + uint16x8_t __ret_351; \ + __ret_351 = vsetq_lane_u16(vget_lane_u16(__s2_351, __p3_351), __s0_351, __p1_351); \ + __ret_351; \ }) #else -#define vcopyq_lane_u16(__p0_260, __p1_260, __p2_260, __p3_260) __extension__ ({ \ - uint16x8_t __s0_260 = __p0_260; \ - uint16x4_t __s2_260 = __p2_260; \ - uint16x8_t __rev0_260; __rev0_260 = __builtin_shufflevector(__s0_260, __s0_260, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __rev2_260; __rev2_260 = __builtin_shufflevector(__s2_260, __s2_260, 3, 2, 1, 0); \ - uint16x8_t __ret_260; \ - __ret_260 = __noswap_vsetq_lane_u16(__noswap_vget_lane_u16(__rev2_260, __p3_260), __rev0_260, __p1_260); \ - __ret_260 = __builtin_shufflevector(__ret_260, __ret_260, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_260; \ +#define vcopyq_lane_u16(__p0_352, __p1_352, __p2_352, __p3_352) __extension__ ({ \ + uint16x8_t __s0_352 = __p0_352; \ + uint16x4_t __s2_352 = __p2_352; \ + uint16x8_t __rev0_352; __rev0_352 = __builtin_shufflevector(__s0_352, __s0_352, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __rev2_352; __rev2_352 = __builtin_shufflevector(__s2_352, __s2_352, 3, 2, 1, 0); \ + uint16x8_t __ret_352; \ + __ret_352 = __noswap_vsetq_lane_u16(__noswap_vget_lane_u16(__rev2_352, __p3_352), __rev0_352, __p1_352); \ + __ret_352 = __builtin_shufflevector(__ret_352, __ret_352, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_352; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_s8(__p0_261, __p1_261, __p2_261, __p3_261) __extension__ ({ \ - int8x16_t __s0_261 = __p0_261; \ - int8x8_t __s2_261 = __p2_261; \ - int8x16_t __ret_261; \ - __ret_261 = vsetq_lane_s8(vget_lane_s8(__s2_261, __p3_261), __s0_261, __p1_261); \ - __ret_261; \ +#define vcopyq_lane_s8(__p0_353, __p1_353, __p2_353, __p3_353) __extension__ ({ \ + int8x16_t __s0_353 = __p0_353; \ + int8x8_t __s2_353 = __p2_353; \ + int8x16_t __ret_353; \ + __ret_353 = vsetq_lane_s8(vget_lane_s8(__s2_353, __p3_353), __s0_353, __p1_353); \ + __ret_353; \ }) #else -#define vcopyq_lane_s8(__p0_262, __p1_262, __p2_262, __p3_262) __extension__ ({ \ - int8x16_t __s0_262 = __p0_262; \ - int8x8_t __s2_262 = __p2_262; \ - int8x16_t __rev0_262; __rev0_262 = __builtin_shufflevector(__s0_262, __s0_262, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __rev2_262; __rev2_262 = __builtin_shufflevector(__s2_262, __s2_262, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_262; \ - __ret_262 = __noswap_vsetq_lane_s8(__noswap_vget_lane_s8(__rev2_262, __p3_262), __rev0_262, __p1_262); \ - __ret_262 = __builtin_shufflevector(__ret_262, __ret_262, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_262; \ +#define vcopyq_lane_s8(__p0_354, __p1_354, __p2_354, __p3_354) __extension__ ({ \ + int8x16_t __s0_354 = __p0_354; \ + int8x8_t __s2_354 = __p2_354; \ + int8x16_t __rev0_354; __rev0_354 = __builtin_shufflevector(__s0_354, __s0_354, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __rev2_354; __rev2_354 = __builtin_shufflevector(__s2_354, __s2_354, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_354; \ + __ret_354 = __noswap_vsetq_lane_s8(__noswap_vget_lane_s8(__rev2_354, __p3_354), __rev0_354, __p1_354); \ + __ret_354 = __builtin_shufflevector(__ret_354, __ret_354, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_354; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_f32(__p0_263, __p1_263, __p2_263, __p3_263) __extension__ ({ \ - float32x4_t __s0_263 = __p0_263; \ - float32x2_t __s2_263 = __p2_263; \ - float32x4_t __ret_263; \ - __ret_263 = vsetq_lane_f32(vget_lane_f32(__s2_263, __p3_263), __s0_263, __p1_263); \ - __ret_263; \ +#define vcopyq_lane_f32(__p0_355, __p1_355, __p2_355, __p3_355) __extension__ ({ \ + float32x4_t __s0_355 = __p0_355; \ + float32x2_t __s2_355 = __p2_355; \ + float32x4_t __ret_355; \ + __ret_355 = vsetq_lane_f32(vget_lane_f32(__s2_355, __p3_355), __s0_355, __p1_355); \ + __ret_355; \ }) #else -#define vcopyq_lane_f32(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \ - float32x4_t __s0_264 = __p0_264; \ - float32x2_t __s2_264 = __p2_264; \ - float32x4_t __rev0_264; __rev0_264 = __builtin_shufflevector(__s0_264, __s0_264, 3, 2, 1, 0); \ - float32x2_t __rev2_264; __rev2_264 = __builtin_shufflevector(__s2_264, __s2_264, 1, 0); \ - float32x4_t __ret_264; \ - __ret_264 = __noswap_vsetq_lane_f32(__noswap_vget_lane_f32(__rev2_264, __p3_264), __rev0_264, __p1_264); \ - __ret_264 = __builtin_shufflevector(__ret_264, __ret_264, 3, 2, 1, 0); \ - __ret_264; \ +#define vcopyq_lane_f32(__p0_356, __p1_356, __p2_356, __p3_356) __extension__ ({ \ + float32x4_t __s0_356 = __p0_356; \ + float32x2_t __s2_356 = __p2_356; \ + float32x4_t __rev0_356; __rev0_356 = __builtin_shufflevector(__s0_356, __s0_356, 3, 2, 1, 0); \ + float32x2_t __rev2_356; __rev2_356 = __builtin_shufflevector(__s2_356, __s2_356, 1, 0); \ + float32x4_t __ret_356; \ + __ret_356 = __noswap_vsetq_lane_f32(__noswap_vget_lane_f32(__rev2_356, __p3_356), __rev0_356, __p1_356); \ + __ret_356 = __builtin_shufflevector(__ret_356, __ret_356, 3, 2, 1, 0); \ + __ret_356; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_s32(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \ - int32x4_t __s0_265 = __p0_265; \ - int32x2_t __s2_265 = __p2_265; \ - int32x4_t __ret_265; \ - __ret_265 = vsetq_lane_s32(vget_lane_s32(__s2_265, __p3_265), __s0_265, __p1_265); \ - __ret_265; \ +#define vcopyq_lane_s32(__p0_357, __p1_357, __p2_357, __p3_357) __extension__ ({ \ + int32x4_t __s0_357 = __p0_357; \ + int32x2_t __s2_357 = __p2_357; \ + int32x4_t __ret_357; \ + __ret_357 = vsetq_lane_s32(vget_lane_s32(__s2_357, __p3_357), __s0_357, __p1_357); \ + __ret_357; \ }) #else -#define vcopyq_lane_s32(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \ - int32x4_t __s0_266 = __p0_266; \ - int32x2_t __s2_266 = __p2_266; \ - int32x4_t __rev0_266; __rev0_266 = __builtin_shufflevector(__s0_266, __s0_266, 3, 2, 1, 0); \ - int32x2_t __rev2_266; __rev2_266 = __builtin_shufflevector(__s2_266, __s2_266, 1, 0); \ - int32x4_t __ret_266; \ - __ret_266 = __noswap_vsetq_lane_s32(__noswap_vget_lane_s32(__rev2_266, __p3_266), __rev0_266, __p1_266); \ - __ret_266 = __builtin_shufflevector(__ret_266, __ret_266, 3, 2, 1, 0); \ - __ret_266; \ +#define vcopyq_lane_s32(__p0_358, __p1_358, __p2_358, __p3_358) __extension__ ({ \ + int32x4_t __s0_358 = __p0_358; \ + int32x2_t __s2_358 = __p2_358; \ + int32x4_t __rev0_358; __rev0_358 = __builtin_shufflevector(__s0_358, __s0_358, 3, 2, 1, 0); \ + int32x2_t __rev2_358; __rev2_358 = __builtin_shufflevector(__s2_358, __s2_358, 1, 0); \ + int32x4_t __ret_358; \ + __ret_358 = __noswap_vsetq_lane_s32(__noswap_vget_lane_s32(__rev2_358, __p3_358), __rev0_358, __p1_358); \ + __ret_358 = __builtin_shufflevector(__ret_358, __ret_358, 3, 2, 1, 0); \ + __ret_358; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_s64(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \ - int64x2_t __s0_267 = __p0_267; \ - int64x1_t __s2_267 = __p2_267; \ - int64x2_t __ret_267; \ - __ret_267 = vsetq_lane_s64(vget_lane_s64(__s2_267, __p3_267), __s0_267, __p1_267); \ - __ret_267; \ +#define vcopyq_lane_s64(__p0_359, __p1_359, __p2_359, __p3_359) __extension__ ({ \ + int64x2_t __s0_359 = __p0_359; \ + int64x1_t __s2_359 = __p2_359; \ + int64x2_t __ret_359; \ + __ret_359 = vsetq_lane_s64(vget_lane_s64(__s2_359, __p3_359), __s0_359, __p1_359); \ + __ret_359; \ }) #else -#define vcopyq_lane_s64(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \ - int64x2_t __s0_268 = __p0_268; \ - int64x1_t __s2_268 = __p2_268; \ - int64x2_t __rev0_268; __rev0_268 = __builtin_shufflevector(__s0_268, __s0_268, 1, 0); \ - int64x2_t __ret_268; \ - __ret_268 = __noswap_vsetq_lane_s64(vget_lane_s64(__s2_268, __p3_268), __rev0_268, __p1_268); \ - __ret_268 = __builtin_shufflevector(__ret_268, __ret_268, 1, 0); \ - __ret_268; \ +#define vcopyq_lane_s64(__p0_360, __p1_360, __p2_360, __p3_360) __extension__ ({ \ + int64x2_t __s0_360 = __p0_360; \ + int64x1_t __s2_360 = __p2_360; \ + int64x2_t __rev0_360; __rev0_360 = __builtin_shufflevector(__s0_360, __s0_360, 1, 0); \ + int64x2_t __ret_360; \ + __ret_360 = __noswap_vsetq_lane_s64(vget_lane_s64(__s2_360, __p3_360), __rev0_360, __p1_360); \ + __ret_360 = __builtin_shufflevector(__ret_360, __ret_360, 1, 0); \ + __ret_360; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_s16(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \ - int16x8_t __s0_269 = __p0_269; \ - int16x4_t __s2_269 = __p2_269; \ - int16x8_t __ret_269; \ - __ret_269 = vsetq_lane_s16(vget_lane_s16(__s2_269, __p3_269), __s0_269, __p1_269); \ - __ret_269; \ +#define vcopyq_lane_s16(__p0_361, __p1_361, __p2_361, __p3_361) __extension__ ({ \ + int16x8_t __s0_361 = __p0_361; \ + int16x4_t __s2_361 = __p2_361; \ + int16x8_t __ret_361; \ + __ret_361 = vsetq_lane_s16(vget_lane_s16(__s2_361, __p3_361), __s0_361, __p1_361); \ + __ret_361; \ }) #else -#define vcopyq_lane_s16(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \ - int16x8_t __s0_270 = __p0_270; \ - int16x4_t __s2_270 = __p2_270; \ - int16x8_t __rev0_270; __rev0_270 = __builtin_shufflevector(__s0_270, __s0_270, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev2_270; __rev2_270 = __builtin_shufflevector(__s2_270, __s2_270, 3, 2, 1, 0); \ - int16x8_t __ret_270; \ - __ret_270 = __noswap_vsetq_lane_s16(__noswap_vget_lane_s16(__rev2_270, __p3_270), __rev0_270, __p1_270); \ - __ret_270 = __builtin_shufflevector(__ret_270, __ret_270, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_270; \ +#define vcopyq_lane_s16(__p0_362, __p1_362, __p2_362, __p3_362) __extension__ ({ \ + int16x8_t __s0_362 = __p0_362; \ + int16x4_t __s2_362 = __p2_362; \ + int16x8_t __rev0_362; __rev0_362 = __builtin_shufflevector(__s0_362, __s0_362, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev2_362; __rev2_362 = __builtin_shufflevector(__s2_362, __s2_362, 3, 2, 1, 0); \ + int16x8_t __ret_362; \ + __ret_362 = __noswap_vsetq_lane_s16(__noswap_vget_lane_s16(__rev2_362, __p3_362), __rev0_362, __p1_362); \ + __ret_362 = __builtin_shufflevector(__ret_362, __ret_362, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_362; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_p8(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \ - poly8x8_t __s0_271 = __p0_271; \ - poly8x8_t __s2_271 = __p2_271; \ - poly8x8_t __ret_271; \ - __ret_271 = vset_lane_p8(vget_lane_p8(__s2_271, __p3_271), __s0_271, __p1_271); \ - __ret_271; \ +#define vcopy_lane_p8(__p0_363, __p1_363, __p2_363, __p3_363) __extension__ ({ \ + poly8x8_t __s0_363 = __p0_363; \ + poly8x8_t __s2_363 = __p2_363; \ + poly8x8_t __ret_363; \ + __ret_363 = vset_lane_p8(vget_lane_p8(__s2_363, __p3_363), __s0_363, __p1_363); \ + __ret_363; \ }) #else -#define vcopy_lane_p8(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \ - poly8x8_t __s0_272 = __p0_272; \ - poly8x8_t __s2_272 = __p2_272; \ - poly8x8_t __rev0_272; __rev0_272 = __builtin_shufflevector(__s0_272, __s0_272, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x8_t __rev2_272; __rev2_272 = __builtin_shufflevector(__s2_272, __s2_272, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x8_t __ret_272; \ - __ret_272 = __noswap_vset_lane_p8(__noswap_vget_lane_p8(__rev2_272, __p3_272), __rev0_272, __p1_272); \ - __ret_272 = __builtin_shufflevector(__ret_272, __ret_272, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_272; \ +#define vcopy_lane_p8(__p0_364, __p1_364, __p2_364, __p3_364) __extension__ ({ \ + poly8x8_t __s0_364 = __p0_364; \ + poly8x8_t __s2_364 = __p2_364; \ + poly8x8_t __rev0_364; __rev0_364 = __builtin_shufflevector(__s0_364, __s0_364, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x8_t __rev2_364; __rev2_364 = __builtin_shufflevector(__s2_364, __s2_364, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x8_t __ret_364; \ + __ret_364 = __noswap_vset_lane_p8(__noswap_vget_lane_p8(__rev2_364, __p3_364), __rev0_364, __p1_364); \ + __ret_364 = __builtin_shufflevector(__ret_364, __ret_364, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_364; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_p16(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \ - poly16x4_t __s0_273 = __p0_273; \ - poly16x4_t __s2_273 = __p2_273; \ - poly16x4_t __ret_273; \ - __ret_273 = vset_lane_p16(vget_lane_p16(__s2_273, __p3_273), __s0_273, __p1_273); \ - __ret_273; \ +#define vcopy_lane_p16(__p0_365, __p1_365, __p2_365, __p3_365) __extension__ ({ \ + poly16x4_t __s0_365 = __p0_365; \ + poly16x4_t __s2_365 = __p2_365; \ + poly16x4_t __ret_365; \ + __ret_365 = vset_lane_p16(vget_lane_p16(__s2_365, __p3_365), __s0_365, __p1_365); \ + __ret_365; \ }) #else -#define vcopy_lane_p16(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \ - poly16x4_t __s0_274 = __p0_274; \ - poly16x4_t __s2_274 = __p2_274; \ - poly16x4_t __rev0_274; __rev0_274 = __builtin_shufflevector(__s0_274, __s0_274, 3, 2, 1, 0); \ - poly16x4_t __rev2_274; __rev2_274 = __builtin_shufflevector(__s2_274, __s2_274, 3, 2, 1, 0); \ - poly16x4_t __ret_274; \ - __ret_274 = __noswap_vset_lane_p16(__noswap_vget_lane_p16(__rev2_274, __p3_274), __rev0_274, __p1_274); \ - __ret_274 = __builtin_shufflevector(__ret_274, __ret_274, 3, 2, 1, 0); \ - __ret_274; \ +#define vcopy_lane_p16(__p0_366, __p1_366, __p2_366, __p3_366) __extension__ ({ \ + poly16x4_t __s0_366 = __p0_366; \ + poly16x4_t __s2_366 = __p2_366; \ + poly16x4_t __rev0_366; __rev0_366 = __builtin_shufflevector(__s0_366, __s0_366, 3, 2, 1, 0); \ + poly16x4_t __rev2_366; __rev2_366 = __builtin_shufflevector(__s2_366, __s2_366, 3, 2, 1, 0); \ + poly16x4_t __ret_366; \ + __ret_366 = __noswap_vset_lane_p16(__noswap_vget_lane_p16(__rev2_366, __p3_366), __rev0_366, __p1_366); \ + __ret_366 = __builtin_shufflevector(__ret_366, __ret_366, 3, 2, 1, 0); \ + __ret_366; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_u8(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \ - uint8x8_t __s0_275 = __p0_275; \ - uint8x8_t __s2_275 = __p2_275; \ - uint8x8_t __ret_275; \ - __ret_275 = vset_lane_u8(vget_lane_u8(__s2_275, __p3_275), __s0_275, __p1_275); \ - __ret_275; \ +#define vcopy_lane_u8(__p0_367, __p1_367, __p2_367, __p3_367) __extension__ ({ \ + uint8x8_t __s0_367 = __p0_367; \ + uint8x8_t __s2_367 = __p2_367; \ + uint8x8_t __ret_367; \ + __ret_367 = vset_lane_u8(vget_lane_u8(__s2_367, __p3_367), __s0_367, __p1_367); \ + __ret_367; \ }) #else -#define vcopy_lane_u8(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \ - uint8x8_t __s0_276 = __p0_276; \ - uint8x8_t __s2_276 = __p2_276; \ - uint8x8_t __rev0_276; __rev0_276 = __builtin_shufflevector(__s0_276, __s0_276, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __rev2_276; __rev2_276 = __builtin_shufflevector(__s2_276, __s2_276, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __ret_276; \ - __ret_276 = __noswap_vset_lane_u8(__noswap_vget_lane_u8(__rev2_276, __p3_276), __rev0_276, __p1_276); \ - __ret_276 = __builtin_shufflevector(__ret_276, __ret_276, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_276; \ +#define vcopy_lane_u8(__p0_368, __p1_368, __p2_368, __p3_368) __extension__ ({ \ + uint8x8_t __s0_368 = __p0_368; \ + uint8x8_t __s2_368 = __p2_368; \ + uint8x8_t __rev0_368; __rev0_368 = __builtin_shufflevector(__s0_368, __s0_368, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __rev2_368; __rev2_368 = __builtin_shufflevector(__s2_368, __s2_368, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __ret_368; \ + __ret_368 = __noswap_vset_lane_u8(__noswap_vget_lane_u8(__rev2_368, __p3_368), __rev0_368, __p1_368); \ + __ret_368 = __builtin_shufflevector(__ret_368, __ret_368, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_368; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_u32(__p0_277, __p1_277, __p2_277, __p3_277) __extension__ ({ \ - uint32x2_t __s0_277 = __p0_277; \ - uint32x2_t __s2_277 = __p2_277; \ - uint32x2_t __ret_277; \ - __ret_277 = vset_lane_u32(vget_lane_u32(__s2_277, __p3_277), __s0_277, __p1_277); \ - __ret_277; \ +#define vcopy_lane_u32(__p0_369, __p1_369, __p2_369, __p3_369) __extension__ ({ \ + uint32x2_t __s0_369 = __p0_369; \ + uint32x2_t __s2_369 = __p2_369; \ + uint32x2_t __ret_369; \ + __ret_369 = vset_lane_u32(vget_lane_u32(__s2_369, __p3_369), __s0_369, __p1_369); \ + __ret_369; \ }) #else -#define vcopy_lane_u32(__p0_278, __p1_278, __p2_278, __p3_278) __extension__ ({ \ - uint32x2_t __s0_278 = __p0_278; \ - uint32x2_t __s2_278 = __p2_278; \ - uint32x2_t __rev0_278; __rev0_278 = __builtin_shufflevector(__s0_278, __s0_278, 1, 0); \ - uint32x2_t __rev2_278; __rev2_278 = __builtin_shufflevector(__s2_278, __s2_278, 1, 0); \ - uint32x2_t __ret_278; \ - __ret_278 = __noswap_vset_lane_u32(__noswap_vget_lane_u32(__rev2_278, __p3_278), __rev0_278, __p1_278); \ - __ret_278 = __builtin_shufflevector(__ret_278, __ret_278, 1, 0); \ - __ret_278; \ +#define vcopy_lane_u32(__p0_370, __p1_370, __p2_370, __p3_370) __extension__ ({ \ + uint32x2_t __s0_370 = __p0_370; \ + uint32x2_t __s2_370 = __p2_370; \ + uint32x2_t __rev0_370; __rev0_370 = __builtin_shufflevector(__s0_370, __s0_370, 1, 0); \ + uint32x2_t __rev2_370; __rev2_370 = __builtin_shufflevector(__s2_370, __s2_370, 1, 0); \ + uint32x2_t __ret_370; \ + __ret_370 = __noswap_vset_lane_u32(__noswap_vget_lane_u32(__rev2_370, __p3_370), __rev0_370, __p1_370); \ + __ret_370 = __builtin_shufflevector(__ret_370, __ret_370, 1, 0); \ + __ret_370; \ }) #endif -#define vcopy_lane_u64(__p0_279, __p1_279, __p2_279, __p3_279) __extension__ ({ \ - uint64x1_t __s0_279 = __p0_279; \ - uint64x1_t __s2_279 = __p2_279; \ - uint64x1_t __ret_279; \ - __ret_279 = vset_lane_u64(vget_lane_u64(__s2_279, __p3_279), __s0_279, __p1_279); \ - __ret_279; \ +#define vcopy_lane_u64(__p0_371, __p1_371, __p2_371, __p3_371) __extension__ ({ \ + uint64x1_t __s0_371 = __p0_371; \ + uint64x1_t __s2_371 = __p2_371; \ + uint64x1_t __ret_371; \ + __ret_371 = vset_lane_u64(vget_lane_u64(__s2_371, __p3_371), __s0_371, __p1_371); \ + __ret_371; \ }) #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_u16(__p0_280, __p1_280, __p2_280, __p3_280) __extension__ ({ \ - uint16x4_t __s0_280 = __p0_280; \ - uint16x4_t __s2_280 = __p2_280; \ - uint16x4_t __ret_280; \ - __ret_280 = vset_lane_u16(vget_lane_u16(__s2_280, __p3_280), __s0_280, __p1_280); \ - __ret_280; \ +#define vcopy_lane_u16(__p0_372, __p1_372, __p2_372, __p3_372) __extension__ ({ \ + uint16x4_t __s0_372 = __p0_372; \ + uint16x4_t __s2_372 = __p2_372; \ + uint16x4_t __ret_372; \ + __ret_372 = vset_lane_u16(vget_lane_u16(__s2_372, __p3_372), __s0_372, __p1_372); \ + __ret_372; \ }) #else -#define vcopy_lane_u16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \ - uint16x4_t __s0_281 = __p0_281; \ - uint16x4_t __s2_281 = __p2_281; \ - uint16x4_t __rev0_281; __rev0_281 = __builtin_shufflevector(__s0_281, __s0_281, 3, 2, 1, 0); \ - uint16x4_t __rev2_281; __rev2_281 = __builtin_shufflevector(__s2_281, __s2_281, 3, 2, 1, 0); \ - uint16x4_t __ret_281; \ - __ret_281 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_281, __p3_281), __rev0_281, __p1_281); \ - __ret_281 = __builtin_shufflevector(__ret_281, __ret_281, 3, 2, 1, 0); \ - __ret_281; \ +#define vcopy_lane_u16(__p0_373, __p1_373, __p2_373, __p3_373) __extension__ ({ \ + uint16x4_t __s0_373 = __p0_373; \ + uint16x4_t __s2_373 = __p2_373; \ + uint16x4_t __rev0_373; __rev0_373 = __builtin_shufflevector(__s0_373, __s0_373, 3, 2, 1, 0); \ + uint16x4_t __rev2_373; __rev2_373 = __builtin_shufflevector(__s2_373, __s2_373, 3, 2, 1, 0); \ + uint16x4_t __ret_373; \ + __ret_373 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_373, __p3_373), __rev0_373, __p1_373); \ + __ret_373 = __builtin_shufflevector(__ret_373, __ret_373, 3, 2, 1, 0); \ + __ret_373; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_s8(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \ - int8x8_t __s0_282 = __p0_282; \ - int8x8_t __s2_282 = __p2_282; \ - int8x8_t __ret_282; \ - __ret_282 = vset_lane_s8(vget_lane_s8(__s2_282, __p3_282), __s0_282, __p1_282); \ - __ret_282; \ +#define vcopy_lane_s8(__p0_374, __p1_374, __p2_374, __p3_374) __extension__ ({ \ + int8x8_t __s0_374 = __p0_374; \ + int8x8_t __s2_374 = __p2_374; \ + int8x8_t __ret_374; \ + __ret_374 = vset_lane_s8(vget_lane_s8(__s2_374, __p3_374), __s0_374, __p1_374); \ + __ret_374; \ }) #else -#define vcopy_lane_s8(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \ - int8x8_t __s0_283 = __p0_283; \ - int8x8_t __s2_283 = __p2_283; \ - int8x8_t __rev0_283; __rev0_283 = __builtin_shufflevector(__s0_283, __s0_283, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __rev2_283; __rev2_283 = __builtin_shufflevector(__s2_283, __s2_283, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __ret_283; \ - __ret_283 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_283, __p3_283), __rev0_283, __p1_283); \ - __ret_283 = __builtin_shufflevector(__ret_283, __ret_283, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_283; \ +#define vcopy_lane_s8(__p0_375, __p1_375, __p2_375, __p3_375) __extension__ ({ \ + int8x8_t __s0_375 = __p0_375; \ + int8x8_t __s2_375 = __p2_375; \ + int8x8_t __rev0_375; __rev0_375 = __builtin_shufflevector(__s0_375, __s0_375, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __rev2_375; __rev2_375 = __builtin_shufflevector(__s2_375, __s2_375, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __ret_375; \ + __ret_375 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_375, __p3_375), __rev0_375, __p1_375); \ + __ret_375 = __builtin_shufflevector(__ret_375, __ret_375, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_375; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_f32(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \ - float32x2_t __s0_284 = __p0_284; \ - float32x2_t __s2_284 = __p2_284; \ - float32x2_t __ret_284; \ - __ret_284 = vset_lane_f32(vget_lane_f32(__s2_284, __p3_284), __s0_284, __p1_284); \ - __ret_284; \ +#define vcopy_lane_f32(__p0_376, __p1_376, __p2_376, __p3_376) __extension__ ({ \ + float32x2_t __s0_376 = __p0_376; \ + float32x2_t __s2_376 = __p2_376; \ + float32x2_t __ret_376; \ + __ret_376 = vset_lane_f32(vget_lane_f32(__s2_376, __p3_376), __s0_376, __p1_376); \ + __ret_376; \ }) #else -#define vcopy_lane_f32(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \ - float32x2_t __s0_285 = __p0_285; \ - float32x2_t __s2_285 = __p2_285; \ - float32x2_t __rev0_285; __rev0_285 = __builtin_shufflevector(__s0_285, __s0_285, 1, 0); \ - float32x2_t __rev2_285; __rev2_285 = __builtin_shufflevector(__s2_285, __s2_285, 1, 0); \ - float32x2_t __ret_285; \ - __ret_285 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_285, __p3_285), __rev0_285, __p1_285); \ - __ret_285 = __builtin_shufflevector(__ret_285, __ret_285, 1, 0); \ - __ret_285; \ +#define vcopy_lane_f32(__p0_377, __p1_377, __p2_377, __p3_377) __extension__ ({ \ + float32x2_t __s0_377 = __p0_377; \ + float32x2_t __s2_377 = __p2_377; \ + float32x2_t __rev0_377; __rev0_377 = __builtin_shufflevector(__s0_377, __s0_377, 1, 0); \ + float32x2_t __rev2_377; __rev2_377 = __builtin_shufflevector(__s2_377, __s2_377, 1, 0); \ + float32x2_t __ret_377; \ + __ret_377 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_377, __p3_377), __rev0_377, __p1_377); \ + __ret_377 = __builtin_shufflevector(__ret_377, __ret_377, 1, 0); \ + __ret_377; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_s32(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \ - int32x2_t __s0_286 = __p0_286; \ - int32x2_t __s2_286 = __p2_286; \ - int32x2_t __ret_286; \ - __ret_286 = vset_lane_s32(vget_lane_s32(__s2_286, __p3_286), __s0_286, __p1_286); \ - __ret_286; \ +#define vcopy_lane_s32(__p0_378, __p1_378, __p2_378, __p3_378) __extension__ ({ \ + int32x2_t __s0_378 = __p0_378; \ + int32x2_t __s2_378 = __p2_378; \ + int32x2_t __ret_378; \ + __ret_378 = vset_lane_s32(vget_lane_s32(__s2_378, __p3_378), __s0_378, __p1_378); \ + __ret_378; \ }) #else -#define vcopy_lane_s32(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \ - int32x2_t __s0_287 = __p0_287; \ - int32x2_t __s2_287 = __p2_287; \ - int32x2_t __rev0_287; __rev0_287 = __builtin_shufflevector(__s0_287, __s0_287, 1, 0); \ - int32x2_t __rev2_287; __rev2_287 = __builtin_shufflevector(__s2_287, __s2_287, 1, 0); \ - int32x2_t __ret_287; \ - __ret_287 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_287, __p3_287), __rev0_287, __p1_287); \ - __ret_287 = __builtin_shufflevector(__ret_287, __ret_287, 1, 0); \ - __ret_287; \ +#define vcopy_lane_s32(__p0_379, __p1_379, __p2_379, __p3_379) __extension__ ({ \ + int32x2_t __s0_379 = __p0_379; \ + int32x2_t __s2_379 = __p2_379; \ + int32x2_t __rev0_379; __rev0_379 = __builtin_shufflevector(__s0_379, __s0_379, 1, 0); \ + int32x2_t __rev2_379; __rev2_379 = __builtin_shufflevector(__s2_379, __s2_379, 1, 0); \ + int32x2_t __ret_379; \ + __ret_379 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_379, __p3_379), __rev0_379, __p1_379); \ + __ret_379 = __builtin_shufflevector(__ret_379, __ret_379, 1, 0); \ + __ret_379; \ }) #endif -#define vcopy_lane_s64(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \ - int64x1_t __s0_288 = __p0_288; \ - int64x1_t __s2_288 = __p2_288; \ - int64x1_t __ret_288; \ - __ret_288 = vset_lane_s64(vget_lane_s64(__s2_288, __p3_288), __s0_288, __p1_288); \ - __ret_288; \ +#define vcopy_lane_s64(__p0_380, __p1_380, __p2_380, __p3_380) __extension__ ({ \ + int64x1_t __s0_380 = __p0_380; \ + int64x1_t __s2_380 = __p2_380; \ + int64x1_t __ret_380; \ + __ret_380 = vset_lane_s64(vget_lane_s64(__s2_380, __p3_380), __s0_380, __p1_380); \ + __ret_380; \ }) #ifdef __LITTLE_ENDIAN__ -#define vcopy_lane_s16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \ - int16x4_t __s0_289 = __p0_289; \ - int16x4_t __s2_289 = __p2_289; \ - int16x4_t __ret_289; \ - __ret_289 = vset_lane_s16(vget_lane_s16(__s2_289, __p3_289), __s0_289, __p1_289); \ - __ret_289; \ +#define vcopy_lane_s16(__p0_381, __p1_381, __p2_381, __p3_381) __extension__ ({ \ + int16x4_t __s0_381 = __p0_381; \ + int16x4_t __s2_381 = __p2_381; \ + int16x4_t __ret_381; \ + __ret_381 = vset_lane_s16(vget_lane_s16(__s2_381, __p3_381), __s0_381, __p1_381); \ + __ret_381; \ }) #else -#define vcopy_lane_s16(__p0_290, __p1_290, __p2_290, __p3_290) __extension__ ({ \ - int16x4_t __s0_290 = __p0_290; \ - int16x4_t __s2_290 = __p2_290; \ - int16x4_t __rev0_290; __rev0_290 = __builtin_shufflevector(__s0_290, __s0_290, 3, 2, 1, 0); \ - int16x4_t __rev2_290; __rev2_290 = __builtin_shufflevector(__s2_290, __s2_290, 3, 2, 1, 0); \ - int16x4_t __ret_290; \ - __ret_290 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_290, __p3_290), __rev0_290, __p1_290); \ - __ret_290 = __builtin_shufflevector(__ret_290, __ret_290, 3, 2, 1, 0); \ - __ret_290; \ +#define vcopy_lane_s16(__p0_382, __p1_382, __p2_382, __p3_382) __extension__ ({ \ + int16x4_t __s0_382 = __p0_382; \ + int16x4_t __s2_382 = __p2_382; \ + int16x4_t __rev0_382; __rev0_382 = __builtin_shufflevector(__s0_382, __s0_382, 3, 2, 1, 0); \ + int16x4_t __rev2_382; __rev2_382 = __builtin_shufflevector(__s2_382, __s2_382, 3, 2, 1, 0); \ + int16x4_t __ret_382; \ + __ret_382 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_382, __p3_382), __rev0_382, __p1_382); \ + __ret_382 = __builtin_shufflevector(__ret_382, __ret_382, 3, 2, 1, 0); \ + __ret_382; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_p8(__p0_291, __p1_291, __p2_291, __p3_291) __extension__ ({ \ - poly8x16_t __s0_291 = __p0_291; \ - poly8x16_t __s2_291 = __p2_291; \ - poly8x16_t __ret_291; \ - __ret_291 = vsetq_lane_p8(vgetq_lane_p8(__s2_291, __p3_291), __s0_291, __p1_291); \ - __ret_291; \ +#define vcopyq_laneq_p8(__p0_383, __p1_383, __p2_383, __p3_383) __extension__ ({ \ + poly8x16_t __s0_383 = __p0_383; \ + poly8x16_t __s2_383 = __p2_383; \ + poly8x16_t __ret_383; \ + __ret_383 = vsetq_lane_p8(vgetq_lane_p8(__s2_383, __p3_383), __s0_383, __p1_383); \ + __ret_383; \ }) #else -#define vcopyq_laneq_p8(__p0_292, __p1_292, __p2_292, __p3_292) __extension__ ({ \ - poly8x16_t __s0_292 = __p0_292; \ - poly8x16_t __s2_292 = __p2_292; \ - poly8x16_t __rev0_292; __rev0_292 = __builtin_shufflevector(__s0_292, __s0_292, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x16_t __rev2_292; __rev2_292 = __builtin_shufflevector(__s2_292, __s2_292, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x16_t __ret_292; \ - __ret_292 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_292, __p3_292), __rev0_292, __p1_292); \ - __ret_292 = __builtin_shufflevector(__ret_292, __ret_292, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_292; \ +#define vcopyq_laneq_p8(__p0_384, __p1_384, __p2_384, __p3_384) __extension__ ({ \ + poly8x16_t __s0_384 = __p0_384; \ + poly8x16_t __s2_384 = __p2_384; \ + poly8x16_t __rev0_384; __rev0_384 = __builtin_shufflevector(__s0_384, __s0_384, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x16_t __rev2_384; __rev2_384 = __builtin_shufflevector(__s2_384, __s2_384, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x16_t __ret_384; \ + __ret_384 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_384, __p3_384), __rev0_384, __p1_384); \ + __ret_384 = __builtin_shufflevector(__ret_384, __ret_384, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_384; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_p16(__p0_293, __p1_293, __p2_293, __p3_293) __extension__ ({ \ - poly16x8_t __s0_293 = __p0_293; \ - poly16x8_t __s2_293 = __p2_293; \ - poly16x8_t __ret_293; \ - __ret_293 = vsetq_lane_p16(vgetq_lane_p16(__s2_293, __p3_293), __s0_293, __p1_293); \ - __ret_293; \ +#define vcopyq_laneq_p16(__p0_385, __p1_385, __p2_385, __p3_385) __extension__ ({ \ + poly16x8_t __s0_385 = __p0_385; \ + poly16x8_t __s2_385 = __p2_385; \ + poly16x8_t __ret_385; \ + __ret_385 = vsetq_lane_p16(vgetq_lane_p16(__s2_385, __p3_385), __s0_385, __p1_385); \ + __ret_385; \ }) #else -#define vcopyq_laneq_p16(__p0_294, __p1_294, __p2_294, __p3_294) __extension__ ({ \ - poly16x8_t __s0_294 = __p0_294; \ - poly16x8_t __s2_294 = __p2_294; \ - poly16x8_t __rev0_294; __rev0_294 = __builtin_shufflevector(__s0_294, __s0_294, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly16x8_t __rev2_294; __rev2_294 = __builtin_shufflevector(__s2_294, __s2_294, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly16x8_t __ret_294; \ - __ret_294 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_294, __p3_294), __rev0_294, __p1_294); \ - __ret_294 = __builtin_shufflevector(__ret_294, __ret_294, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_294; \ +#define vcopyq_laneq_p16(__p0_386, __p1_386, __p2_386, __p3_386) __extension__ ({ \ + poly16x8_t __s0_386 = __p0_386; \ + poly16x8_t __s2_386 = __p2_386; \ + poly16x8_t __rev0_386; __rev0_386 = __builtin_shufflevector(__s0_386, __s0_386, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly16x8_t __rev2_386; __rev2_386 = __builtin_shufflevector(__s2_386, __s2_386, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly16x8_t __ret_386; \ + __ret_386 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_386, __p3_386), __rev0_386, __p1_386); \ + __ret_386 = __builtin_shufflevector(__ret_386, __ret_386, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_386; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_u8(__p0_295, __p1_295, __p2_295, __p3_295) __extension__ ({ \ - uint8x16_t __s0_295 = __p0_295; \ - uint8x16_t __s2_295 = __p2_295; \ - uint8x16_t __ret_295; \ - __ret_295 = vsetq_lane_u8(vgetq_lane_u8(__s2_295, __p3_295), __s0_295, __p1_295); \ - __ret_295; \ +#define vcopyq_laneq_u8(__p0_387, __p1_387, __p2_387, __p3_387) __extension__ ({ \ + uint8x16_t __s0_387 = __p0_387; \ + uint8x16_t __s2_387 = __p2_387; \ + uint8x16_t __ret_387; \ + __ret_387 = vsetq_lane_u8(vgetq_lane_u8(__s2_387, __p3_387), __s0_387, __p1_387); \ + __ret_387; \ }) #else -#define vcopyq_laneq_u8(__p0_296, __p1_296, __p2_296, __p3_296) __extension__ ({ \ - uint8x16_t __s0_296 = __p0_296; \ - uint8x16_t __s2_296 = __p2_296; \ - uint8x16_t __rev0_296; __rev0_296 = __builtin_shufflevector(__s0_296, __s0_296, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __rev2_296; __rev2_296 = __builtin_shufflevector(__s2_296, __s2_296, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __ret_296; \ - __ret_296 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_296, __p3_296), __rev0_296, __p1_296); \ - __ret_296 = __builtin_shufflevector(__ret_296, __ret_296, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_296; \ +#define vcopyq_laneq_u8(__p0_388, __p1_388, __p2_388, __p3_388) __extension__ ({ \ + uint8x16_t __s0_388 = __p0_388; \ + uint8x16_t __s2_388 = __p2_388; \ + uint8x16_t __rev0_388; __rev0_388 = __builtin_shufflevector(__s0_388, __s0_388, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __rev2_388; __rev2_388 = __builtin_shufflevector(__s2_388, __s2_388, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __ret_388; \ + __ret_388 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_388, __p3_388), __rev0_388, __p1_388); \ + __ret_388 = __builtin_shufflevector(__ret_388, __ret_388, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_388; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_u32(__p0_297, __p1_297, __p2_297, __p3_297) __extension__ ({ \ - uint32x4_t __s0_297 = __p0_297; \ - uint32x4_t __s2_297 = __p2_297; \ - uint32x4_t __ret_297; \ - __ret_297 = vsetq_lane_u32(vgetq_lane_u32(__s2_297, __p3_297), __s0_297, __p1_297); \ - __ret_297; \ +#define vcopyq_laneq_u32(__p0_389, __p1_389, __p2_389, __p3_389) __extension__ ({ \ + uint32x4_t __s0_389 = __p0_389; \ + uint32x4_t __s2_389 = __p2_389; \ + uint32x4_t __ret_389; \ + __ret_389 = vsetq_lane_u32(vgetq_lane_u32(__s2_389, __p3_389), __s0_389, __p1_389); \ + __ret_389; \ }) #else -#define vcopyq_laneq_u32(__p0_298, __p1_298, __p2_298, __p3_298) __extension__ ({ \ - uint32x4_t __s0_298 = __p0_298; \ - uint32x4_t __s2_298 = __p2_298; \ - uint32x4_t __rev0_298; __rev0_298 = __builtin_shufflevector(__s0_298, __s0_298, 3, 2, 1, 0); \ - uint32x4_t __rev2_298; __rev2_298 = __builtin_shufflevector(__s2_298, __s2_298, 3, 2, 1, 0); \ - uint32x4_t __ret_298; \ - __ret_298 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_298, __p3_298), __rev0_298, __p1_298); \ - __ret_298 = __builtin_shufflevector(__ret_298, __ret_298, 3, 2, 1, 0); \ - __ret_298; \ +#define vcopyq_laneq_u32(__p0_390, __p1_390, __p2_390, __p3_390) __extension__ ({ \ + uint32x4_t __s0_390 = __p0_390; \ + uint32x4_t __s2_390 = __p2_390; \ + uint32x4_t __rev0_390; __rev0_390 = __builtin_shufflevector(__s0_390, __s0_390, 3, 2, 1, 0); \ + uint32x4_t __rev2_390; __rev2_390 = __builtin_shufflevector(__s2_390, __s2_390, 3, 2, 1, 0); \ + uint32x4_t __ret_390; \ + __ret_390 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_390, __p3_390), __rev0_390, __p1_390); \ + __ret_390 = __builtin_shufflevector(__ret_390, __ret_390, 3, 2, 1, 0); \ + __ret_390; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_u64(__p0_299, __p1_299, __p2_299, __p3_299) __extension__ ({ \ - uint64x2_t __s0_299 = __p0_299; \ - uint64x2_t __s2_299 = __p2_299; \ - uint64x2_t __ret_299; \ - __ret_299 = vsetq_lane_u64(vgetq_lane_u64(__s2_299, __p3_299), __s0_299, __p1_299); \ - __ret_299; \ +#define vcopyq_laneq_u64(__p0_391, __p1_391, __p2_391, __p3_391) __extension__ ({ \ + uint64x2_t __s0_391 = __p0_391; \ + uint64x2_t __s2_391 = __p2_391; \ + uint64x2_t __ret_391; \ + __ret_391 = vsetq_lane_u64(vgetq_lane_u64(__s2_391, __p3_391), __s0_391, __p1_391); \ + __ret_391; \ }) #else -#define vcopyq_laneq_u64(__p0_300, __p1_300, __p2_300, __p3_300) __extension__ ({ \ - uint64x2_t __s0_300 = __p0_300; \ - uint64x2_t __s2_300 = __p2_300; \ - uint64x2_t __rev0_300; __rev0_300 = __builtin_shufflevector(__s0_300, __s0_300, 1, 0); \ - uint64x2_t __rev2_300; __rev2_300 = __builtin_shufflevector(__s2_300, __s2_300, 1, 0); \ - uint64x2_t __ret_300; \ - __ret_300 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_300, __p3_300), __rev0_300, __p1_300); \ - __ret_300 = __builtin_shufflevector(__ret_300, __ret_300, 1, 0); \ - __ret_300; \ +#define vcopyq_laneq_u64(__p0_392, __p1_392, __p2_392, __p3_392) __extension__ ({ \ + uint64x2_t __s0_392 = __p0_392; \ + uint64x2_t __s2_392 = __p2_392; \ + uint64x2_t __rev0_392; __rev0_392 = __builtin_shufflevector(__s0_392, __s0_392, 1, 0); \ + uint64x2_t __rev2_392; __rev2_392 = __builtin_shufflevector(__s2_392, __s2_392, 1, 0); \ + uint64x2_t __ret_392; \ + __ret_392 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_392, __p3_392), __rev0_392, __p1_392); \ + __ret_392 = __builtin_shufflevector(__ret_392, __ret_392, 1, 0); \ + __ret_392; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_u16(__p0_301, __p1_301, __p2_301, __p3_301) __extension__ ({ \ - uint16x8_t __s0_301 = __p0_301; \ - uint16x8_t __s2_301 = __p2_301; \ - uint16x8_t __ret_301; \ - __ret_301 = vsetq_lane_u16(vgetq_lane_u16(__s2_301, __p3_301), __s0_301, __p1_301); \ - __ret_301; \ +#define vcopyq_laneq_u16(__p0_393, __p1_393, __p2_393, __p3_393) __extension__ ({ \ + uint16x8_t __s0_393 = __p0_393; \ + uint16x8_t __s2_393 = __p2_393; \ + uint16x8_t __ret_393; \ + __ret_393 = vsetq_lane_u16(vgetq_lane_u16(__s2_393, __p3_393), __s0_393, __p1_393); \ + __ret_393; \ }) #else -#define vcopyq_laneq_u16(__p0_302, __p1_302, __p2_302, __p3_302) __extension__ ({ \ - uint16x8_t __s0_302 = __p0_302; \ - uint16x8_t __s2_302 = __p2_302; \ - uint16x8_t __rev0_302; __rev0_302 = __builtin_shufflevector(__s0_302, __s0_302, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev2_302; __rev2_302 = __builtin_shufflevector(__s2_302, __s2_302, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __ret_302; \ - __ret_302 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_302, __p3_302), __rev0_302, __p1_302); \ - __ret_302 = __builtin_shufflevector(__ret_302, __ret_302, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_302; \ +#define vcopyq_laneq_u16(__p0_394, __p1_394, __p2_394, __p3_394) __extension__ ({ \ + uint16x8_t __s0_394 = __p0_394; \ + uint16x8_t __s2_394 = __p2_394; \ + uint16x8_t __rev0_394; __rev0_394 = __builtin_shufflevector(__s0_394, __s0_394, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev2_394; __rev2_394 = __builtin_shufflevector(__s2_394, __s2_394, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __ret_394; \ + __ret_394 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_394, __p3_394), __rev0_394, __p1_394); \ + __ret_394 = __builtin_shufflevector(__ret_394, __ret_394, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_394; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_s8(__p0_303, __p1_303, __p2_303, __p3_303) __extension__ ({ \ - int8x16_t __s0_303 = __p0_303; \ - int8x16_t __s2_303 = __p2_303; \ - int8x16_t __ret_303; \ - __ret_303 = vsetq_lane_s8(vgetq_lane_s8(__s2_303, __p3_303), __s0_303, __p1_303); \ - __ret_303; \ +#define vcopyq_laneq_s8(__p0_395, __p1_395, __p2_395, __p3_395) __extension__ ({ \ + int8x16_t __s0_395 = __p0_395; \ + int8x16_t __s2_395 = __p2_395; \ + int8x16_t __ret_395; \ + __ret_395 = vsetq_lane_s8(vgetq_lane_s8(__s2_395, __p3_395), __s0_395, __p1_395); \ + __ret_395; \ }) #else -#define vcopyq_laneq_s8(__p0_304, __p1_304, __p2_304, __p3_304) __extension__ ({ \ - int8x16_t __s0_304 = __p0_304; \ - int8x16_t __s2_304 = __p2_304; \ - int8x16_t __rev0_304; __rev0_304 = __builtin_shufflevector(__s0_304, __s0_304, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __rev2_304; __rev2_304 = __builtin_shufflevector(__s2_304, __s2_304, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_304; \ - __ret_304 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_304, __p3_304), __rev0_304, __p1_304); \ - __ret_304 = __builtin_shufflevector(__ret_304, __ret_304, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_304; \ +#define vcopyq_laneq_s8(__p0_396, __p1_396, __p2_396, __p3_396) __extension__ ({ \ + int8x16_t __s0_396 = __p0_396; \ + int8x16_t __s2_396 = __p2_396; \ + int8x16_t __rev0_396; __rev0_396 = __builtin_shufflevector(__s0_396, __s0_396, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __rev2_396; __rev2_396 = __builtin_shufflevector(__s2_396, __s2_396, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_396; \ + __ret_396 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_396, __p3_396), __rev0_396, __p1_396); \ + __ret_396 = __builtin_shufflevector(__ret_396, __ret_396, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_396; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_f32(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \ - float32x4_t __s0_305 = __p0_305; \ - float32x4_t __s2_305 = __p2_305; \ - float32x4_t __ret_305; \ - __ret_305 = vsetq_lane_f32(vgetq_lane_f32(__s2_305, __p3_305), __s0_305, __p1_305); \ - __ret_305; \ +#define vcopyq_laneq_f32(__p0_397, __p1_397, __p2_397, __p3_397) __extension__ ({ \ + float32x4_t __s0_397 = __p0_397; \ + float32x4_t __s2_397 = __p2_397; \ + float32x4_t __ret_397; \ + __ret_397 = vsetq_lane_f32(vgetq_lane_f32(__s2_397, __p3_397), __s0_397, __p1_397); \ + __ret_397; \ }) #else -#define vcopyq_laneq_f32(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \ - float32x4_t __s0_306 = __p0_306; \ - float32x4_t __s2_306 = __p2_306; \ - float32x4_t __rev0_306; __rev0_306 = __builtin_shufflevector(__s0_306, __s0_306, 3, 2, 1, 0); \ - float32x4_t __rev2_306; __rev2_306 = __builtin_shufflevector(__s2_306, __s2_306, 3, 2, 1, 0); \ - float32x4_t __ret_306; \ - __ret_306 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_306, __p3_306), __rev0_306, __p1_306); \ - __ret_306 = __builtin_shufflevector(__ret_306, __ret_306, 3, 2, 1, 0); \ - __ret_306; \ +#define vcopyq_laneq_f32(__p0_398, __p1_398, __p2_398, __p3_398) __extension__ ({ \ + float32x4_t __s0_398 = __p0_398; \ + float32x4_t __s2_398 = __p2_398; \ + float32x4_t __rev0_398; __rev0_398 = __builtin_shufflevector(__s0_398, __s0_398, 3, 2, 1, 0); \ + float32x4_t __rev2_398; __rev2_398 = __builtin_shufflevector(__s2_398, __s2_398, 3, 2, 1, 0); \ + float32x4_t __ret_398; \ + __ret_398 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_398, __p3_398), __rev0_398, __p1_398); \ + __ret_398 = __builtin_shufflevector(__ret_398, __ret_398, 3, 2, 1, 0); \ + __ret_398; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_s32(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \ - int32x4_t __s0_307 = __p0_307; \ - int32x4_t __s2_307 = __p2_307; \ - int32x4_t __ret_307; \ - __ret_307 = vsetq_lane_s32(vgetq_lane_s32(__s2_307, __p3_307), __s0_307, __p1_307); \ - __ret_307; \ +#define vcopyq_laneq_s32(__p0_399, __p1_399, __p2_399, __p3_399) __extension__ ({ \ + int32x4_t __s0_399 = __p0_399; \ + int32x4_t __s2_399 = __p2_399; \ + int32x4_t __ret_399; \ + __ret_399 = vsetq_lane_s32(vgetq_lane_s32(__s2_399, __p3_399), __s0_399, __p1_399); \ + __ret_399; \ }) #else -#define vcopyq_laneq_s32(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \ - int32x4_t __s0_308 = __p0_308; \ - int32x4_t __s2_308 = __p2_308; \ - int32x4_t __rev0_308; __rev0_308 = __builtin_shufflevector(__s0_308, __s0_308, 3, 2, 1, 0); \ - int32x4_t __rev2_308; __rev2_308 = __builtin_shufflevector(__s2_308, __s2_308, 3, 2, 1, 0); \ - int32x4_t __ret_308; \ - __ret_308 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_308, __p3_308), __rev0_308, __p1_308); \ - __ret_308 = __builtin_shufflevector(__ret_308, __ret_308, 3, 2, 1, 0); \ - __ret_308; \ +#define vcopyq_laneq_s32(__p0_400, __p1_400, __p2_400, __p3_400) __extension__ ({ \ + int32x4_t __s0_400 = __p0_400; \ + int32x4_t __s2_400 = __p2_400; \ + int32x4_t __rev0_400; __rev0_400 = __builtin_shufflevector(__s0_400, __s0_400, 3, 2, 1, 0); \ + int32x4_t __rev2_400; __rev2_400 = __builtin_shufflevector(__s2_400, __s2_400, 3, 2, 1, 0); \ + int32x4_t __ret_400; \ + __ret_400 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_400, __p3_400), __rev0_400, __p1_400); \ + __ret_400 = __builtin_shufflevector(__ret_400, __ret_400, 3, 2, 1, 0); \ + __ret_400; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_s64(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \ - int64x2_t __s0_309 = __p0_309; \ - int64x2_t __s2_309 = __p2_309; \ - int64x2_t __ret_309; \ - __ret_309 = vsetq_lane_s64(vgetq_lane_s64(__s2_309, __p3_309), __s0_309, __p1_309); \ - __ret_309; \ +#define vcopyq_laneq_s64(__p0_401, __p1_401, __p2_401, __p3_401) __extension__ ({ \ + int64x2_t __s0_401 = __p0_401; \ + int64x2_t __s2_401 = __p2_401; \ + int64x2_t __ret_401; \ + __ret_401 = vsetq_lane_s64(vgetq_lane_s64(__s2_401, __p3_401), __s0_401, __p1_401); \ + __ret_401; \ }) #else -#define vcopyq_laneq_s64(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \ - int64x2_t __s0_310 = __p0_310; \ - int64x2_t __s2_310 = __p2_310; \ - int64x2_t __rev0_310; __rev0_310 = __builtin_shufflevector(__s0_310, __s0_310, 1, 0); \ - int64x2_t __rev2_310; __rev2_310 = __builtin_shufflevector(__s2_310, __s2_310, 1, 0); \ - int64x2_t __ret_310; \ - __ret_310 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_310, __p3_310), __rev0_310, __p1_310); \ - __ret_310 = __builtin_shufflevector(__ret_310, __ret_310, 1, 0); \ - __ret_310; \ +#define vcopyq_laneq_s64(__p0_402, __p1_402, __p2_402, __p3_402) __extension__ ({ \ + int64x2_t __s0_402 = __p0_402; \ + int64x2_t __s2_402 = __p2_402; \ + int64x2_t __rev0_402; __rev0_402 = __builtin_shufflevector(__s0_402, __s0_402, 1, 0); \ + int64x2_t __rev2_402; __rev2_402 = __builtin_shufflevector(__s2_402, __s2_402, 1, 0); \ + int64x2_t __ret_402; \ + __ret_402 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_402, __p3_402), __rev0_402, __p1_402); \ + __ret_402 = __builtin_shufflevector(__ret_402, __ret_402, 1, 0); \ + __ret_402; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_s16(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \ - int16x8_t __s0_311 = __p0_311; \ - int16x8_t __s2_311 = __p2_311; \ - int16x8_t __ret_311; \ - __ret_311 = vsetq_lane_s16(vgetq_lane_s16(__s2_311, __p3_311), __s0_311, __p1_311); \ - __ret_311; \ +#define vcopyq_laneq_s16(__p0_403, __p1_403, __p2_403, __p3_403) __extension__ ({ \ + int16x8_t __s0_403 = __p0_403; \ + int16x8_t __s2_403 = __p2_403; \ + int16x8_t __ret_403; \ + __ret_403 = vsetq_lane_s16(vgetq_lane_s16(__s2_403, __p3_403), __s0_403, __p1_403); \ + __ret_403; \ }) #else -#define vcopyq_laneq_s16(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \ - int16x8_t __s0_312 = __p0_312; \ - int16x8_t __s2_312 = __p2_312; \ - int16x8_t __rev0_312; __rev0_312 = __builtin_shufflevector(__s0_312, __s0_312, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_312; __rev2_312 = __builtin_shufflevector(__s2_312, __s2_312, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_312; \ - __ret_312 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_312, __p3_312), __rev0_312, __p1_312); \ - __ret_312 = __builtin_shufflevector(__ret_312, __ret_312, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_312; \ +#define vcopyq_laneq_s16(__p0_404, __p1_404, __p2_404, __p3_404) __extension__ ({ \ + int16x8_t __s0_404 = __p0_404; \ + int16x8_t __s2_404 = __p2_404; \ + int16x8_t __rev0_404; __rev0_404 = __builtin_shufflevector(__s0_404, __s0_404, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_404; __rev2_404 = __builtin_shufflevector(__s2_404, __s2_404, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_404; \ + __ret_404 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_404, __p3_404), __rev0_404, __p1_404); \ + __ret_404 = __builtin_shufflevector(__ret_404, __ret_404, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_404; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_p8(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \ - poly8x8_t __s0_313 = __p0_313; \ - poly8x16_t __s2_313 = __p2_313; \ - poly8x8_t __ret_313; \ - __ret_313 = vset_lane_p8(vgetq_lane_p8(__s2_313, __p3_313), __s0_313, __p1_313); \ - __ret_313; \ +#define vcopy_laneq_p8(__p0_405, __p1_405, __p2_405, __p3_405) __extension__ ({ \ + poly8x8_t __s0_405 = __p0_405; \ + poly8x16_t __s2_405 = __p2_405; \ + poly8x8_t __ret_405; \ + __ret_405 = vset_lane_p8(vgetq_lane_p8(__s2_405, __p3_405), __s0_405, __p1_405); \ + __ret_405; \ }) #else -#define vcopy_laneq_p8(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \ - poly8x8_t __s0_314 = __p0_314; \ - poly8x16_t __s2_314 = __p2_314; \ - poly8x8_t __rev0_314; __rev0_314 = __builtin_shufflevector(__s0_314, __s0_314, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x16_t __rev2_314; __rev2_314 = __builtin_shufflevector(__s2_314, __s2_314, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x8_t __ret_314; \ - __ret_314 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_314, __p3_314), __rev0_314, __p1_314); \ - __ret_314 = __builtin_shufflevector(__ret_314, __ret_314, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_314; \ +#define vcopy_laneq_p8(__p0_406, __p1_406, __p2_406, __p3_406) __extension__ ({ \ + poly8x8_t __s0_406 = __p0_406; \ + poly8x16_t __s2_406 = __p2_406; \ + poly8x8_t __rev0_406; __rev0_406 = __builtin_shufflevector(__s0_406, __s0_406, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x16_t __rev2_406; __rev2_406 = __builtin_shufflevector(__s2_406, __s2_406, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x8_t __ret_406; \ + __ret_406 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_406, __p3_406), __rev0_406, __p1_406); \ + __ret_406 = __builtin_shufflevector(__ret_406, __ret_406, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_406; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_p16(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \ - poly16x4_t __s0_315 = __p0_315; \ - poly16x8_t __s2_315 = __p2_315; \ - poly16x4_t __ret_315; \ - __ret_315 = vset_lane_p16(vgetq_lane_p16(__s2_315, __p3_315), __s0_315, __p1_315); \ - __ret_315; \ +#define vcopy_laneq_p16(__p0_407, __p1_407, __p2_407, __p3_407) __extension__ ({ \ + poly16x4_t __s0_407 = __p0_407; \ + poly16x8_t __s2_407 = __p2_407; \ + poly16x4_t __ret_407; \ + __ret_407 = vset_lane_p16(vgetq_lane_p16(__s2_407, __p3_407), __s0_407, __p1_407); \ + __ret_407; \ }) #else -#define vcopy_laneq_p16(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \ - poly16x4_t __s0_316 = __p0_316; \ - poly16x8_t __s2_316 = __p2_316; \ - poly16x4_t __rev0_316; __rev0_316 = __builtin_shufflevector(__s0_316, __s0_316, 3, 2, 1, 0); \ - poly16x8_t __rev2_316; __rev2_316 = __builtin_shufflevector(__s2_316, __s2_316, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly16x4_t __ret_316; \ - __ret_316 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_316, __p3_316), __rev0_316, __p1_316); \ - __ret_316 = __builtin_shufflevector(__ret_316, __ret_316, 3, 2, 1, 0); \ - __ret_316; \ +#define vcopy_laneq_p16(__p0_408, __p1_408, __p2_408, __p3_408) __extension__ ({ \ + poly16x4_t __s0_408 = __p0_408; \ + poly16x8_t __s2_408 = __p2_408; \ + poly16x4_t __rev0_408; __rev0_408 = __builtin_shufflevector(__s0_408, __s0_408, 3, 2, 1, 0); \ + poly16x8_t __rev2_408; __rev2_408 = __builtin_shufflevector(__s2_408, __s2_408, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly16x4_t __ret_408; \ + __ret_408 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_408, __p3_408), __rev0_408, __p1_408); \ + __ret_408 = __builtin_shufflevector(__ret_408, __ret_408, 3, 2, 1, 0); \ + __ret_408; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_u8(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \ - uint8x8_t __s0_317 = __p0_317; \ - uint8x16_t __s2_317 = __p2_317; \ - uint8x8_t __ret_317; \ - __ret_317 = vset_lane_u8(vgetq_lane_u8(__s2_317, __p3_317), __s0_317, __p1_317); \ - __ret_317; \ +#define vcopy_laneq_u8(__p0_409, __p1_409, __p2_409, __p3_409) __extension__ ({ \ + uint8x8_t __s0_409 = __p0_409; \ + uint8x16_t __s2_409 = __p2_409; \ + uint8x8_t __ret_409; \ + __ret_409 = vset_lane_u8(vgetq_lane_u8(__s2_409, __p3_409), __s0_409, __p1_409); \ + __ret_409; \ }) #else -#define vcopy_laneq_u8(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \ - uint8x8_t __s0_318 = __p0_318; \ - uint8x16_t __s2_318 = __p2_318; \ - uint8x8_t __rev0_318; __rev0_318 = __builtin_shufflevector(__s0_318, __s0_318, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __rev2_318; __rev2_318 = __builtin_shufflevector(__s2_318, __s2_318, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __ret_318; \ - __ret_318 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_318, __p3_318), __rev0_318, __p1_318); \ - __ret_318 = __builtin_shufflevector(__ret_318, __ret_318, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_318; \ +#define vcopy_laneq_u8(__p0_410, __p1_410, __p2_410, __p3_410) __extension__ ({ \ + uint8x8_t __s0_410 = __p0_410; \ + uint8x16_t __s2_410 = __p2_410; \ + uint8x8_t __rev0_410; __rev0_410 = __builtin_shufflevector(__s0_410, __s0_410, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __rev2_410; __rev2_410 = __builtin_shufflevector(__s2_410, __s2_410, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __ret_410; \ + __ret_410 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_410, __p3_410), __rev0_410, __p1_410); \ + __ret_410 = __builtin_shufflevector(__ret_410, __ret_410, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_410; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_u32(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \ - uint32x2_t __s0_319 = __p0_319; \ - uint32x4_t __s2_319 = __p2_319; \ - uint32x2_t __ret_319; \ - __ret_319 = vset_lane_u32(vgetq_lane_u32(__s2_319, __p3_319), __s0_319, __p1_319); \ - __ret_319; \ +#define vcopy_laneq_u32(__p0_411, __p1_411, __p2_411, __p3_411) __extension__ ({ \ + uint32x2_t __s0_411 = __p0_411; \ + uint32x4_t __s2_411 = __p2_411; \ + uint32x2_t __ret_411; \ + __ret_411 = vset_lane_u32(vgetq_lane_u32(__s2_411, __p3_411), __s0_411, __p1_411); \ + __ret_411; \ }) #else -#define vcopy_laneq_u32(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \ - uint32x2_t __s0_320 = __p0_320; \ - uint32x4_t __s2_320 = __p2_320; \ - uint32x2_t __rev0_320; __rev0_320 = __builtin_shufflevector(__s0_320, __s0_320, 1, 0); \ - uint32x4_t __rev2_320; __rev2_320 = __builtin_shufflevector(__s2_320, __s2_320, 3, 2, 1, 0); \ - uint32x2_t __ret_320; \ - __ret_320 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_320, __p3_320), __rev0_320, __p1_320); \ - __ret_320 = __builtin_shufflevector(__ret_320, __ret_320, 1, 0); \ - __ret_320; \ +#define vcopy_laneq_u32(__p0_412, __p1_412, __p2_412, __p3_412) __extension__ ({ \ + uint32x2_t __s0_412 = __p0_412; \ + uint32x4_t __s2_412 = __p2_412; \ + uint32x2_t __rev0_412; __rev0_412 = __builtin_shufflevector(__s0_412, __s0_412, 1, 0); \ + uint32x4_t __rev2_412; __rev2_412 = __builtin_shufflevector(__s2_412, __s2_412, 3, 2, 1, 0); \ + uint32x2_t __ret_412; \ + __ret_412 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_412, __p3_412), __rev0_412, __p1_412); \ + __ret_412 = __builtin_shufflevector(__ret_412, __ret_412, 1, 0); \ + __ret_412; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_u64(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \ - uint64x1_t __s0_321 = __p0_321; \ - uint64x2_t __s2_321 = __p2_321; \ - uint64x1_t __ret_321; \ - __ret_321 = vset_lane_u64(vgetq_lane_u64(__s2_321, __p3_321), __s0_321, __p1_321); \ - __ret_321; \ +#define vcopy_laneq_u64(__p0_413, __p1_413, __p2_413, __p3_413) __extension__ ({ \ + uint64x1_t __s0_413 = __p0_413; \ + uint64x2_t __s2_413 = __p2_413; \ + uint64x1_t __ret_413; \ + __ret_413 = vset_lane_u64(vgetq_lane_u64(__s2_413, __p3_413), __s0_413, __p1_413); \ + __ret_413; \ }) #else -#define vcopy_laneq_u64(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \ - uint64x1_t __s0_322 = __p0_322; \ - uint64x2_t __s2_322 = __p2_322; \ - uint64x2_t __rev2_322; __rev2_322 = __builtin_shufflevector(__s2_322, __s2_322, 1, 0); \ - uint64x1_t __ret_322; \ - __ret_322 = vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_322, __p3_322), __s0_322, __p1_322); \ - __ret_322; \ +#define vcopy_laneq_u64(__p0_414, __p1_414, __p2_414, __p3_414) __extension__ ({ \ + uint64x1_t __s0_414 = __p0_414; \ + uint64x2_t __s2_414 = __p2_414; \ + uint64x2_t __rev2_414; __rev2_414 = __builtin_shufflevector(__s2_414, __s2_414, 1, 0); \ + uint64x1_t __ret_414; \ + __ret_414 = vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_414, __p3_414), __s0_414, __p1_414); \ + __ret_414; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_u16(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \ - uint16x4_t __s0_323 = __p0_323; \ - uint16x8_t __s2_323 = __p2_323; \ - uint16x4_t __ret_323; \ - __ret_323 = vset_lane_u16(vgetq_lane_u16(__s2_323, __p3_323), __s0_323, __p1_323); \ - __ret_323; \ +#define vcopy_laneq_u16(__p0_415, __p1_415, __p2_415, __p3_415) __extension__ ({ \ + uint16x4_t __s0_415 = __p0_415; \ + uint16x8_t __s2_415 = __p2_415; \ + uint16x4_t __ret_415; \ + __ret_415 = vset_lane_u16(vgetq_lane_u16(__s2_415, __p3_415), __s0_415, __p1_415); \ + __ret_415; \ }) #else -#define vcopy_laneq_u16(__p0_324, __p1_324, __p2_324, __p3_324) __extension__ ({ \ - uint16x4_t __s0_324 = __p0_324; \ - uint16x8_t __s2_324 = __p2_324; \ - uint16x4_t __rev0_324; __rev0_324 = __builtin_shufflevector(__s0_324, __s0_324, 3, 2, 1, 0); \ - uint16x8_t __rev2_324; __rev2_324 = __builtin_shufflevector(__s2_324, __s2_324, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __ret_324; \ - __ret_324 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_324, __p3_324), __rev0_324, __p1_324); \ - __ret_324 = __builtin_shufflevector(__ret_324, __ret_324, 3, 2, 1, 0); \ - __ret_324; \ +#define vcopy_laneq_u16(__p0_416, __p1_416, __p2_416, __p3_416) __extension__ ({ \ + uint16x4_t __s0_416 = __p0_416; \ + uint16x8_t __s2_416 = __p2_416; \ + uint16x4_t __rev0_416; __rev0_416 = __builtin_shufflevector(__s0_416, __s0_416, 3, 2, 1, 0); \ + uint16x8_t __rev2_416; __rev2_416 = __builtin_shufflevector(__s2_416, __s2_416, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __ret_416; \ + __ret_416 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_416, __p3_416), __rev0_416, __p1_416); \ + __ret_416 = __builtin_shufflevector(__ret_416, __ret_416, 3, 2, 1, 0); \ + __ret_416; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_s8(__p0_325, __p1_325, __p2_325, __p3_325) __extension__ ({ \ - int8x8_t __s0_325 = __p0_325; \ - int8x16_t __s2_325 = __p2_325; \ - int8x8_t __ret_325; \ - __ret_325 = vset_lane_s8(vgetq_lane_s8(__s2_325, __p3_325), __s0_325, __p1_325); \ - __ret_325; \ +#define vcopy_laneq_s8(__p0_417, __p1_417, __p2_417, __p3_417) __extension__ ({ \ + int8x8_t __s0_417 = __p0_417; \ + int8x16_t __s2_417 = __p2_417; \ + int8x8_t __ret_417; \ + __ret_417 = vset_lane_s8(vgetq_lane_s8(__s2_417, __p3_417), __s0_417, __p1_417); \ + __ret_417; \ }) #else -#define vcopy_laneq_s8(__p0_326, __p1_326, __p2_326, __p3_326) __extension__ ({ \ - int8x8_t __s0_326 = __p0_326; \ - int8x16_t __s2_326 = __p2_326; \ - int8x8_t __rev0_326; __rev0_326 = __builtin_shufflevector(__s0_326, __s0_326, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __rev2_326; __rev2_326 = __builtin_shufflevector(__s2_326, __s2_326, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __ret_326; \ - __ret_326 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_326, __p3_326), __rev0_326, __p1_326); \ - __ret_326 = __builtin_shufflevector(__ret_326, __ret_326, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_326; \ +#define vcopy_laneq_s8(__p0_418, __p1_418, __p2_418, __p3_418) __extension__ ({ \ + int8x8_t __s0_418 = __p0_418; \ + int8x16_t __s2_418 = __p2_418; \ + int8x8_t __rev0_418; __rev0_418 = __builtin_shufflevector(__s0_418, __s0_418, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __rev2_418; __rev2_418 = __builtin_shufflevector(__s2_418, __s2_418, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __ret_418; \ + __ret_418 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_418, __p3_418), __rev0_418, __p1_418); \ + __ret_418 = __builtin_shufflevector(__ret_418, __ret_418, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_418; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_f32(__p0_327, __p1_327, __p2_327, __p3_327) __extension__ ({ \ - float32x2_t __s0_327 = __p0_327; \ - float32x4_t __s2_327 = __p2_327; \ - float32x2_t __ret_327; \ - __ret_327 = vset_lane_f32(vgetq_lane_f32(__s2_327, __p3_327), __s0_327, __p1_327); \ - __ret_327; \ +#define vcopy_laneq_f32(__p0_419, __p1_419, __p2_419, __p3_419) __extension__ ({ \ + float32x2_t __s0_419 = __p0_419; \ + float32x4_t __s2_419 = __p2_419; \ + float32x2_t __ret_419; \ + __ret_419 = vset_lane_f32(vgetq_lane_f32(__s2_419, __p3_419), __s0_419, __p1_419); \ + __ret_419; \ }) #else -#define vcopy_laneq_f32(__p0_328, __p1_328, __p2_328, __p3_328) __extension__ ({ \ - float32x2_t __s0_328 = __p0_328; \ - float32x4_t __s2_328 = __p2_328; \ - float32x2_t __rev0_328; __rev0_328 = __builtin_shufflevector(__s0_328, __s0_328, 1, 0); \ - float32x4_t __rev2_328; __rev2_328 = __builtin_shufflevector(__s2_328, __s2_328, 3, 2, 1, 0); \ - float32x2_t __ret_328; \ - __ret_328 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_328, __p3_328), __rev0_328, __p1_328); \ - __ret_328 = __builtin_shufflevector(__ret_328, __ret_328, 1, 0); \ - __ret_328; \ +#define vcopy_laneq_f32(__p0_420, __p1_420, __p2_420, __p3_420) __extension__ ({ \ + float32x2_t __s0_420 = __p0_420; \ + float32x4_t __s2_420 = __p2_420; \ + float32x2_t __rev0_420; __rev0_420 = __builtin_shufflevector(__s0_420, __s0_420, 1, 0); \ + float32x4_t __rev2_420; __rev2_420 = __builtin_shufflevector(__s2_420, __s2_420, 3, 2, 1, 0); \ + float32x2_t __ret_420; \ + __ret_420 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_420, __p3_420), __rev0_420, __p1_420); \ + __ret_420 = __builtin_shufflevector(__ret_420, __ret_420, 1, 0); \ + __ret_420; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_s32(__p0_329, __p1_329, __p2_329, __p3_329) __extension__ ({ \ - int32x2_t __s0_329 = __p0_329; \ - int32x4_t __s2_329 = __p2_329; \ - int32x2_t __ret_329; \ - __ret_329 = vset_lane_s32(vgetq_lane_s32(__s2_329, __p3_329), __s0_329, __p1_329); \ - __ret_329; \ +#define vcopy_laneq_s32(__p0_421, __p1_421, __p2_421, __p3_421) __extension__ ({ \ + int32x2_t __s0_421 = __p0_421; \ + int32x4_t __s2_421 = __p2_421; \ + int32x2_t __ret_421; \ + __ret_421 = vset_lane_s32(vgetq_lane_s32(__s2_421, __p3_421), __s0_421, __p1_421); \ + __ret_421; \ }) #else -#define vcopy_laneq_s32(__p0_330, __p1_330, __p2_330, __p3_330) __extension__ ({ \ - int32x2_t __s0_330 = __p0_330; \ - int32x4_t __s2_330 = __p2_330; \ - int32x2_t __rev0_330; __rev0_330 = __builtin_shufflevector(__s0_330, __s0_330, 1, 0); \ - int32x4_t __rev2_330; __rev2_330 = __builtin_shufflevector(__s2_330, __s2_330, 3, 2, 1, 0); \ - int32x2_t __ret_330; \ - __ret_330 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_330, __p3_330), __rev0_330, __p1_330); \ - __ret_330 = __builtin_shufflevector(__ret_330, __ret_330, 1, 0); \ - __ret_330; \ +#define vcopy_laneq_s32(__p0_422, __p1_422, __p2_422, __p3_422) __extension__ ({ \ + int32x2_t __s0_422 = __p0_422; \ + int32x4_t __s2_422 = __p2_422; \ + int32x2_t __rev0_422; __rev0_422 = __builtin_shufflevector(__s0_422, __s0_422, 1, 0); \ + int32x4_t __rev2_422; __rev2_422 = __builtin_shufflevector(__s2_422, __s2_422, 3, 2, 1, 0); \ + int32x2_t __ret_422; \ + __ret_422 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_422, __p3_422), __rev0_422, __p1_422); \ + __ret_422 = __builtin_shufflevector(__ret_422, __ret_422, 1, 0); \ + __ret_422; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_s64(__p0_331, __p1_331, __p2_331, __p3_331) __extension__ ({ \ - int64x1_t __s0_331 = __p0_331; \ - int64x2_t __s2_331 = __p2_331; \ - int64x1_t __ret_331; \ - __ret_331 = vset_lane_s64(vgetq_lane_s64(__s2_331, __p3_331), __s0_331, __p1_331); \ - __ret_331; \ +#define vcopy_laneq_s64(__p0_423, __p1_423, __p2_423, __p3_423) __extension__ ({ \ + int64x1_t __s0_423 = __p0_423; \ + int64x2_t __s2_423 = __p2_423; \ + int64x1_t __ret_423; \ + __ret_423 = vset_lane_s64(vgetq_lane_s64(__s2_423, __p3_423), __s0_423, __p1_423); \ + __ret_423; \ }) #else -#define vcopy_laneq_s64(__p0_332, __p1_332, __p2_332, __p3_332) __extension__ ({ \ - int64x1_t __s0_332 = __p0_332; \ - int64x2_t __s2_332 = __p2_332; \ - int64x2_t __rev2_332; __rev2_332 = __builtin_shufflevector(__s2_332, __s2_332, 1, 0); \ - int64x1_t __ret_332; \ - __ret_332 = vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_332, __p3_332), __s0_332, __p1_332); \ - __ret_332; \ +#define vcopy_laneq_s64(__p0_424, __p1_424, __p2_424, __p3_424) __extension__ ({ \ + int64x1_t __s0_424 = __p0_424; \ + int64x2_t __s2_424 = __p2_424; \ + int64x2_t __rev2_424; __rev2_424 = __builtin_shufflevector(__s2_424, __s2_424, 1, 0); \ + int64x1_t __ret_424; \ + __ret_424 = vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_424, __p3_424), __s0_424, __p1_424); \ + __ret_424; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_s16(__p0_333, __p1_333, __p2_333, __p3_333) __extension__ ({ \ - int16x4_t __s0_333 = __p0_333; \ - int16x8_t __s2_333 = __p2_333; \ - int16x4_t __ret_333; \ - __ret_333 = vset_lane_s16(vgetq_lane_s16(__s2_333, __p3_333), __s0_333, __p1_333); \ - __ret_333; \ +#define vcopy_laneq_s16(__p0_425, __p1_425, __p2_425, __p3_425) __extension__ ({ \ + int16x4_t __s0_425 = __p0_425; \ + int16x8_t __s2_425 = __p2_425; \ + int16x4_t __ret_425; \ + __ret_425 = vset_lane_s16(vgetq_lane_s16(__s2_425, __p3_425), __s0_425, __p1_425); \ + __ret_425; \ }) #else -#define vcopy_laneq_s16(__p0_334, __p1_334, __p2_334, __p3_334) __extension__ ({ \ - int16x4_t __s0_334 = __p0_334; \ - int16x8_t __s2_334 = __p2_334; \ - int16x4_t __rev0_334; __rev0_334 = __builtin_shufflevector(__s0_334, __s0_334, 3, 2, 1, 0); \ - int16x8_t __rev2_334; __rev2_334 = __builtin_shufflevector(__s2_334, __s2_334, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __ret_334; \ - __ret_334 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_334, __p3_334), __rev0_334, __p1_334); \ - __ret_334 = __builtin_shufflevector(__ret_334, __ret_334, 3, 2, 1, 0); \ - __ret_334; \ +#define vcopy_laneq_s16(__p0_426, __p1_426, __p2_426, __p3_426) __extension__ ({ \ + int16x4_t __s0_426 = __p0_426; \ + int16x8_t __s2_426 = __p2_426; \ + int16x4_t __rev0_426; __rev0_426 = __builtin_shufflevector(__s0_426, __s0_426, 3, 2, 1, 0); \ + int16x8_t __rev2_426; __rev2_426 = __builtin_shufflevector(__s2_426, __s2_426, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __ret_426; \ + __ret_426 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_426, __p3_426), __rev0_426, __p1_426); \ + __ret_426 = __builtin_shufflevector(__ret_426, __ret_426, 3, 2, 1, 0); \ + __ret_426; \ }) #endif @@ -49009,85 +50713,85 @@ __ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) { }) #endif -#define vdup_lane_p64(__p0_335, __p1_335) __extension__ ({ \ - poly64x1_t __s0_335 = __p0_335; \ - poly64x1_t __ret_335; \ - __ret_335 = splat_lane_p64(__s0_335, __p1_335); \ - __ret_335; \ +#define vdup_lane_p64(__p0_427, __p1_427) __extension__ ({ \ + poly64x1_t __s0_427 = __p0_427; \ + poly64x1_t __ret_427; \ + __ret_427 = splat_lane_p64(__s0_427, __p1_427); \ + __ret_427; \ }) #ifdef __LITTLE_ENDIAN__ -#define vdupq_lane_p64(__p0_336, __p1_336) __extension__ ({ \ - poly64x1_t __s0_336 = __p0_336; \ - poly64x2_t __ret_336; \ - __ret_336 = splatq_lane_p64(__s0_336, __p1_336); \ - __ret_336; \ +#define vdupq_lane_p64(__p0_428, __p1_428) __extension__ ({ \ + poly64x1_t __s0_428 = __p0_428; \ + poly64x2_t __ret_428; \ + __ret_428 = splatq_lane_p64(__s0_428, __p1_428); \ + __ret_428; \ }) #else -#define vdupq_lane_p64(__p0_337, __p1_337) __extension__ ({ \ - poly64x1_t __s0_337 = __p0_337; \ - poly64x2_t __ret_337; \ - __ret_337 = __noswap_splatq_lane_p64(__s0_337, __p1_337); \ - __ret_337 = __builtin_shufflevector(__ret_337, __ret_337, 1, 0); \ - __ret_337; \ +#define vdupq_lane_p64(__p0_429, __p1_429) __extension__ ({ \ + poly64x1_t __s0_429 = __p0_429; \ + poly64x2_t __ret_429; \ + __ret_429 = __noswap_splatq_lane_p64(__s0_429, __p1_429); \ + __ret_429 = __builtin_shufflevector(__ret_429, __ret_429, 1, 0); \ + __ret_429; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_lane_f64(__p0_338, __p1_338) __extension__ ({ \ - float64x1_t __s0_338 = __p0_338; \ - float64x2_t __ret_338; \ - __ret_338 = splatq_lane_f64(__s0_338, __p1_338); \ - __ret_338; \ +#define vdupq_lane_f64(__p0_430, __p1_430) __extension__ ({ \ + float64x1_t __s0_430 = __p0_430; \ + float64x2_t __ret_430; \ + __ret_430 = splatq_lane_f64(__s0_430, __p1_430); \ + __ret_430; \ }) #else -#define vdupq_lane_f64(__p0_339, __p1_339) __extension__ ({ \ - float64x1_t __s0_339 = __p0_339; \ - float64x2_t __ret_339; \ - __ret_339 = __noswap_splatq_lane_f64(__s0_339, __p1_339); \ - __ret_339 = __builtin_shufflevector(__ret_339, __ret_339, 1, 0); \ - __ret_339; \ +#define vdupq_lane_f64(__p0_431, __p1_431) __extension__ ({ \ + float64x1_t __s0_431 = __p0_431; \ + float64x2_t __ret_431; \ + __ret_431 = __noswap_splatq_lane_f64(__s0_431, __p1_431); \ + __ret_431 = __builtin_shufflevector(__ret_431, __ret_431, 1, 0); \ + __ret_431; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_lane_f16(__p0_340, __p1_340) __extension__ ({ \ - float16x4_t __s0_340 = __p0_340; \ - float16x8_t __ret_340; \ - __ret_340 = splatq_lane_f16(__s0_340, __p1_340); \ - __ret_340; \ +#define vdupq_lane_f16(__p0_432, __p1_432) __extension__ ({ \ + float16x4_t __s0_432 = __p0_432; \ + float16x8_t __ret_432; \ + __ret_432 = splatq_lane_f16(__s0_432, __p1_432); \ + __ret_432; \ }) #else -#define vdupq_lane_f16(__p0_341, __p1_341) __extension__ ({ \ - float16x4_t __s0_341 = __p0_341; \ - float16x4_t __rev0_341; __rev0_341 = __builtin_shufflevector(__s0_341, __s0_341, 3, 2, 1, 0); \ - float16x8_t __ret_341; \ - __ret_341 = __noswap_splatq_lane_f16(__rev0_341, __p1_341); \ - __ret_341 = __builtin_shufflevector(__ret_341, __ret_341, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_341; \ +#define vdupq_lane_f16(__p0_433, __p1_433) __extension__ ({ \ + float16x4_t __s0_433 = __p0_433; \ + float16x4_t __rev0_433; __rev0_433 = __builtin_shufflevector(__s0_433, __s0_433, 3, 2, 1, 0); \ + float16x8_t __ret_433; \ + __ret_433 = __noswap_splatq_lane_f16(__rev0_433, __p1_433); \ + __ret_433 = __builtin_shufflevector(__ret_433, __ret_433, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_433; \ }) #endif -#define vdup_lane_f64(__p0_342, __p1_342) __extension__ ({ \ - float64x1_t __s0_342 = __p0_342; \ - float64x1_t __ret_342; \ - __ret_342 = splat_lane_f64(__s0_342, __p1_342); \ - __ret_342; \ +#define vdup_lane_f64(__p0_434, __p1_434) __extension__ ({ \ + float64x1_t __s0_434 = __p0_434; \ + float64x1_t __ret_434; \ + __ret_434 = splat_lane_f64(__s0_434, __p1_434); \ + __ret_434; \ }) #ifdef __LITTLE_ENDIAN__ -#define vdup_lane_f16(__p0_343, __p1_343) __extension__ ({ \ - float16x4_t __s0_343 = __p0_343; \ - float16x4_t __ret_343; \ - __ret_343 = splat_lane_f16(__s0_343, __p1_343); \ - __ret_343; \ +#define vdup_lane_f16(__p0_435, __p1_435) __extension__ ({ \ + float16x4_t __s0_435 = __p0_435; \ + float16x4_t __ret_435; \ + __ret_435 = splat_lane_f16(__s0_435, __p1_435); \ + __ret_435; \ }) #else -#define vdup_lane_f16(__p0_344, __p1_344) __extension__ ({ \ - float16x4_t __s0_344 = __p0_344; \ - float16x4_t __rev0_344; __rev0_344 = __builtin_shufflevector(__s0_344, __s0_344, 3, 2, 1, 0); \ - float16x4_t __ret_344; \ - __ret_344 = __noswap_splat_lane_f16(__rev0_344, __p1_344); \ - __ret_344 = __builtin_shufflevector(__ret_344, __ret_344, 3, 2, 1, 0); \ - __ret_344; \ +#define vdup_lane_f16(__p0_436, __p1_436) __extension__ ({ \ + float16x4_t __s0_436 = __p0_436; \ + float16x4_t __rev0_436; __rev0_436 = __builtin_shufflevector(__s0_436, __s0_436, 3, 2, 1, 0); \ + float16x4_t __ret_436; \ + __ret_436 = __noswap_splat_lane_f16(__rev0_436, __p1_436); \ + __ret_436 = __builtin_shufflevector(__ret_436, __ret_436, 3, 2, 1, 0); \ + __ret_436; \ }) #endif @@ -49296,502 +51000,502 @@ __ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_p8(__p0_345, __p1_345) __extension__ ({ \ - poly8x16_t __s0_345 = __p0_345; \ - poly8x8_t __ret_345; \ - __ret_345 = splat_laneq_p8(__s0_345, __p1_345); \ - __ret_345; \ +#define vdup_laneq_p8(__p0_437, __p1_437) __extension__ ({ \ + poly8x16_t __s0_437 = __p0_437; \ + poly8x8_t __ret_437; \ + __ret_437 = splat_laneq_p8(__s0_437, __p1_437); \ + __ret_437; \ }) #else -#define vdup_laneq_p8(__p0_346, __p1_346) __extension__ ({ \ - poly8x16_t __s0_346 = __p0_346; \ - poly8x16_t __rev0_346; __rev0_346 = __builtin_shufflevector(__s0_346, __s0_346, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x8_t __ret_346; \ - __ret_346 = __noswap_splat_laneq_p8(__rev0_346, __p1_346); \ - __ret_346 = __builtin_shufflevector(__ret_346, __ret_346, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_346; \ +#define vdup_laneq_p8(__p0_438, __p1_438) __extension__ ({ \ + poly8x16_t __s0_438 = __p0_438; \ + poly8x16_t __rev0_438; __rev0_438 = __builtin_shufflevector(__s0_438, __s0_438, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x8_t __ret_438; \ + __ret_438 = __noswap_splat_laneq_p8(__rev0_438, __p1_438); \ + __ret_438 = __builtin_shufflevector(__ret_438, __ret_438, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_438; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_p64(__p0_347, __p1_347) __extension__ ({ \ - poly64x2_t __s0_347 = __p0_347; \ - poly64x1_t __ret_347; \ - __ret_347 = splat_laneq_p64(__s0_347, __p1_347); \ - __ret_347; \ +#define vdup_laneq_p64(__p0_439, __p1_439) __extension__ ({ \ + poly64x2_t __s0_439 = __p0_439; \ + poly64x1_t __ret_439; \ + __ret_439 = splat_laneq_p64(__s0_439, __p1_439); \ + __ret_439; \ }) #else -#define vdup_laneq_p64(__p0_348, __p1_348) __extension__ ({ \ - poly64x2_t __s0_348 = __p0_348; \ - poly64x2_t __rev0_348; __rev0_348 = __builtin_shufflevector(__s0_348, __s0_348, 1, 0); \ - poly64x1_t __ret_348; \ - __ret_348 = __noswap_splat_laneq_p64(__rev0_348, __p1_348); \ - __ret_348; \ +#define vdup_laneq_p64(__p0_440, __p1_440) __extension__ ({ \ + poly64x2_t __s0_440 = __p0_440; \ + poly64x2_t __rev0_440; __rev0_440 = __builtin_shufflevector(__s0_440, __s0_440, 1, 0); \ + poly64x1_t __ret_440; \ + __ret_440 = __noswap_splat_laneq_p64(__rev0_440, __p1_440); \ + __ret_440; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_p16(__p0_349, __p1_349) __extension__ ({ \ - poly16x8_t __s0_349 = __p0_349; \ - poly16x4_t __ret_349; \ - __ret_349 = splat_laneq_p16(__s0_349, __p1_349); \ - __ret_349; \ +#define vdup_laneq_p16(__p0_441, __p1_441) __extension__ ({ \ + poly16x8_t __s0_441 = __p0_441; \ + poly16x4_t __ret_441; \ + __ret_441 = splat_laneq_p16(__s0_441, __p1_441); \ + __ret_441; \ }) #else -#define vdup_laneq_p16(__p0_350, __p1_350) __extension__ ({ \ - poly16x8_t __s0_350 = __p0_350; \ - poly16x8_t __rev0_350; __rev0_350 = __builtin_shufflevector(__s0_350, __s0_350, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly16x4_t __ret_350; \ - __ret_350 = __noswap_splat_laneq_p16(__rev0_350, __p1_350); \ - __ret_350 = __builtin_shufflevector(__ret_350, __ret_350, 3, 2, 1, 0); \ - __ret_350; \ +#define vdup_laneq_p16(__p0_442, __p1_442) __extension__ ({ \ + poly16x8_t __s0_442 = __p0_442; \ + poly16x8_t __rev0_442; __rev0_442 = __builtin_shufflevector(__s0_442, __s0_442, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly16x4_t __ret_442; \ + __ret_442 = __noswap_splat_laneq_p16(__rev0_442, __p1_442); \ + __ret_442 = __builtin_shufflevector(__ret_442, __ret_442, 3, 2, 1, 0); \ + __ret_442; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_p8(__p0_351, __p1_351) __extension__ ({ \ - poly8x16_t __s0_351 = __p0_351; \ - poly8x16_t __ret_351; \ - __ret_351 = splatq_laneq_p8(__s0_351, __p1_351); \ - __ret_351; \ +#define vdupq_laneq_p8(__p0_443, __p1_443) __extension__ ({ \ + poly8x16_t __s0_443 = __p0_443; \ + poly8x16_t __ret_443; \ + __ret_443 = splatq_laneq_p8(__s0_443, __p1_443); \ + __ret_443; \ }) #else -#define vdupq_laneq_p8(__p0_352, __p1_352) __extension__ ({ \ - poly8x16_t __s0_352 = __p0_352; \ - poly8x16_t __rev0_352; __rev0_352 = __builtin_shufflevector(__s0_352, __s0_352, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly8x16_t __ret_352; \ - __ret_352 = __noswap_splatq_laneq_p8(__rev0_352, __p1_352); \ - __ret_352 = __builtin_shufflevector(__ret_352, __ret_352, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_352; \ +#define vdupq_laneq_p8(__p0_444, __p1_444) __extension__ ({ \ + poly8x16_t __s0_444 = __p0_444; \ + poly8x16_t __rev0_444; __rev0_444 = __builtin_shufflevector(__s0_444, __s0_444, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly8x16_t __ret_444; \ + __ret_444 = __noswap_splatq_laneq_p8(__rev0_444, __p1_444); \ + __ret_444 = __builtin_shufflevector(__ret_444, __ret_444, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_444; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_p64(__p0_353, __p1_353) __extension__ ({ \ - poly64x2_t __s0_353 = __p0_353; \ - poly64x2_t __ret_353; \ - __ret_353 = splatq_laneq_p64(__s0_353, __p1_353); \ - __ret_353; \ +#define vdupq_laneq_p64(__p0_445, __p1_445) __extension__ ({ \ + poly64x2_t __s0_445 = __p0_445; \ + poly64x2_t __ret_445; \ + __ret_445 = splatq_laneq_p64(__s0_445, __p1_445); \ + __ret_445; \ }) #else -#define vdupq_laneq_p64(__p0_354, __p1_354) __extension__ ({ \ - poly64x2_t __s0_354 = __p0_354; \ - poly64x2_t __rev0_354; __rev0_354 = __builtin_shufflevector(__s0_354, __s0_354, 1, 0); \ - poly64x2_t __ret_354; \ - __ret_354 = __noswap_splatq_laneq_p64(__rev0_354, __p1_354); \ - __ret_354 = __builtin_shufflevector(__ret_354, __ret_354, 1, 0); \ - __ret_354; \ +#define vdupq_laneq_p64(__p0_446, __p1_446) __extension__ ({ \ + poly64x2_t __s0_446 = __p0_446; \ + poly64x2_t __rev0_446; __rev0_446 = __builtin_shufflevector(__s0_446, __s0_446, 1, 0); \ + poly64x2_t __ret_446; \ + __ret_446 = __noswap_splatq_laneq_p64(__rev0_446, __p1_446); \ + __ret_446 = __builtin_shufflevector(__ret_446, __ret_446, 1, 0); \ + __ret_446; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_p16(__p0_355, __p1_355) __extension__ ({ \ - poly16x8_t __s0_355 = __p0_355; \ - poly16x8_t __ret_355; \ - __ret_355 = splatq_laneq_p16(__s0_355, __p1_355); \ - __ret_355; \ +#define vdupq_laneq_p16(__p0_447, __p1_447) __extension__ ({ \ + poly16x8_t __s0_447 = __p0_447; \ + poly16x8_t __ret_447; \ + __ret_447 = splatq_laneq_p16(__s0_447, __p1_447); \ + __ret_447; \ }) #else -#define vdupq_laneq_p16(__p0_356, __p1_356) __extension__ ({ \ - poly16x8_t __s0_356 = __p0_356; \ - poly16x8_t __rev0_356; __rev0_356 = __builtin_shufflevector(__s0_356, __s0_356, 7, 6, 5, 4, 3, 2, 1, 0); \ - poly16x8_t __ret_356; \ - __ret_356 = __noswap_splatq_laneq_p16(__rev0_356, __p1_356); \ - __ret_356 = __builtin_shufflevector(__ret_356, __ret_356, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_356; \ +#define vdupq_laneq_p16(__p0_448, __p1_448) __extension__ ({ \ + poly16x8_t __s0_448 = __p0_448; \ + poly16x8_t __rev0_448; __rev0_448 = __builtin_shufflevector(__s0_448, __s0_448, 7, 6, 5, 4, 3, 2, 1, 0); \ + poly16x8_t __ret_448; \ + __ret_448 = __noswap_splatq_laneq_p16(__rev0_448, __p1_448); \ + __ret_448 = __builtin_shufflevector(__ret_448, __ret_448, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_448; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_u8(__p0_357, __p1_357) __extension__ ({ \ - uint8x16_t __s0_357 = __p0_357; \ - uint8x16_t __ret_357; \ - __ret_357 = splatq_laneq_u8(__s0_357, __p1_357); \ - __ret_357; \ +#define vdupq_laneq_u8(__p0_449, __p1_449) __extension__ ({ \ + uint8x16_t __s0_449 = __p0_449; \ + uint8x16_t __ret_449; \ + __ret_449 = splatq_laneq_u8(__s0_449, __p1_449); \ + __ret_449; \ }) #else -#define vdupq_laneq_u8(__p0_358, __p1_358) __extension__ ({ \ - uint8x16_t __s0_358 = __p0_358; \ - uint8x16_t __rev0_358; __rev0_358 = __builtin_shufflevector(__s0_358, __s0_358, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __ret_358; \ - __ret_358 = __noswap_splatq_laneq_u8(__rev0_358, __p1_358); \ - __ret_358 = __builtin_shufflevector(__ret_358, __ret_358, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_358; \ +#define vdupq_laneq_u8(__p0_450, __p1_450) __extension__ ({ \ + uint8x16_t __s0_450 = __p0_450; \ + uint8x16_t __rev0_450; __rev0_450 = __builtin_shufflevector(__s0_450, __s0_450, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __ret_450; \ + __ret_450 = __noswap_splatq_laneq_u8(__rev0_450, __p1_450); \ + __ret_450 = __builtin_shufflevector(__ret_450, __ret_450, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_450; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_u32(__p0_359, __p1_359) __extension__ ({ \ - uint32x4_t __s0_359 = __p0_359; \ - uint32x4_t __ret_359; \ - __ret_359 = splatq_laneq_u32(__s0_359, __p1_359); \ - __ret_359; \ +#define vdupq_laneq_u32(__p0_451, __p1_451) __extension__ ({ \ + uint32x4_t __s0_451 = __p0_451; \ + uint32x4_t __ret_451; \ + __ret_451 = splatq_laneq_u32(__s0_451, __p1_451); \ + __ret_451; \ }) #else -#define vdupq_laneq_u32(__p0_360, __p1_360) __extension__ ({ \ - uint32x4_t __s0_360 = __p0_360; \ - uint32x4_t __rev0_360; __rev0_360 = __builtin_shufflevector(__s0_360, __s0_360, 3, 2, 1, 0); \ - uint32x4_t __ret_360; \ - __ret_360 = __noswap_splatq_laneq_u32(__rev0_360, __p1_360); \ - __ret_360 = __builtin_shufflevector(__ret_360, __ret_360, 3, 2, 1, 0); \ - __ret_360; \ +#define vdupq_laneq_u32(__p0_452, __p1_452) __extension__ ({ \ + uint32x4_t __s0_452 = __p0_452; \ + uint32x4_t __rev0_452; __rev0_452 = __builtin_shufflevector(__s0_452, __s0_452, 3, 2, 1, 0); \ + uint32x4_t __ret_452; \ + __ret_452 = __noswap_splatq_laneq_u32(__rev0_452, __p1_452); \ + __ret_452 = __builtin_shufflevector(__ret_452, __ret_452, 3, 2, 1, 0); \ + __ret_452; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_u64(__p0_361, __p1_361) __extension__ ({ \ - uint64x2_t __s0_361 = __p0_361; \ - uint64x2_t __ret_361; \ - __ret_361 = splatq_laneq_u64(__s0_361, __p1_361); \ - __ret_361; \ +#define vdupq_laneq_u64(__p0_453, __p1_453) __extension__ ({ \ + uint64x2_t __s0_453 = __p0_453; \ + uint64x2_t __ret_453; \ + __ret_453 = splatq_laneq_u64(__s0_453, __p1_453); \ + __ret_453; \ }) #else -#define vdupq_laneq_u64(__p0_362, __p1_362) __extension__ ({ \ - uint64x2_t __s0_362 = __p0_362; \ - uint64x2_t __rev0_362; __rev0_362 = __builtin_shufflevector(__s0_362, __s0_362, 1, 0); \ - uint64x2_t __ret_362; \ - __ret_362 = __noswap_splatq_laneq_u64(__rev0_362, __p1_362); \ - __ret_362 = __builtin_shufflevector(__ret_362, __ret_362, 1, 0); \ - __ret_362; \ +#define vdupq_laneq_u64(__p0_454, __p1_454) __extension__ ({ \ + uint64x2_t __s0_454 = __p0_454; \ + uint64x2_t __rev0_454; __rev0_454 = __builtin_shufflevector(__s0_454, __s0_454, 1, 0); \ + uint64x2_t __ret_454; \ + __ret_454 = __noswap_splatq_laneq_u64(__rev0_454, __p1_454); \ + __ret_454 = __builtin_shufflevector(__ret_454, __ret_454, 1, 0); \ + __ret_454; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_u16(__p0_363, __p1_363) __extension__ ({ \ - uint16x8_t __s0_363 = __p0_363; \ - uint16x8_t __ret_363; \ - __ret_363 = splatq_laneq_u16(__s0_363, __p1_363); \ - __ret_363; \ +#define vdupq_laneq_u16(__p0_455, __p1_455) __extension__ ({ \ + uint16x8_t __s0_455 = __p0_455; \ + uint16x8_t __ret_455; \ + __ret_455 = splatq_laneq_u16(__s0_455, __p1_455); \ + __ret_455; \ }) #else -#define vdupq_laneq_u16(__p0_364, __p1_364) __extension__ ({ \ - uint16x8_t __s0_364 = __p0_364; \ - uint16x8_t __rev0_364; __rev0_364 = __builtin_shufflevector(__s0_364, __s0_364, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __ret_364; \ - __ret_364 = __noswap_splatq_laneq_u16(__rev0_364, __p1_364); \ - __ret_364 = __builtin_shufflevector(__ret_364, __ret_364, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_364; \ +#define vdupq_laneq_u16(__p0_456, __p1_456) __extension__ ({ \ + uint16x8_t __s0_456 = __p0_456; \ + uint16x8_t __rev0_456; __rev0_456 = __builtin_shufflevector(__s0_456, __s0_456, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __ret_456; \ + __ret_456 = __noswap_splatq_laneq_u16(__rev0_456, __p1_456); \ + __ret_456 = __builtin_shufflevector(__ret_456, __ret_456, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_456; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_s8(__p0_365, __p1_365) __extension__ ({ \ - int8x16_t __s0_365 = __p0_365; \ - int8x16_t __ret_365; \ - __ret_365 = splatq_laneq_s8(__s0_365, __p1_365); \ - __ret_365; \ +#define vdupq_laneq_s8(__p0_457, __p1_457) __extension__ ({ \ + int8x16_t __s0_457 = __p0_457; \ + int8x16_t __ret_457; \ + __ret_457 = splatq_laneq_s8(__s0_457, __p1_457); \ + __ret_457; \ }) #else -#define vdupq_laneq_s8(__p0_366, __p1_366) __extension__ ({ \ - int8x16_t __s0_366 = __p0_366; \ - int8x16_t __rev0_366; __rev0_366 = __builtin_shufflevector(__s0_366, __s0_366, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_366; \ - __ret_366 = __noswap_splatq_laneq_s8(__rev0_366, __p1_366); \ - __ret_366 = __builtin_shufflevector(__ret_366, __ret_366, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_366; \ +#define vdupq_laneq_s8(__p0_458, __p1_458) __extension__ ({ \ + int8x16_t __s0_458 = __p0_458; \ + int8x16_t __rev0_458; __rev0_458 = __builtin_shufflevector(__s0_458, __s0_458, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_458; \ + __ret_458 = __noswap_splatq_laneq_s8(__rev0_458, __p1_458); \ + __ret_458 = __builtin_shufflevector(__ret_458, __ret_458, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_458; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_f64(__p0_367, __p1_367) __extension__ ({ \ - float64x2_t __s0_367 = __p0_367; \ - float64x2_t __ret_367; \ - __ret_367 = splatq_laneq_f64(__s0_367, __p1_367); \ - __ret_367; \ +#define vdupq_laneq_f64(__p0_459, __p1_459) __extension__ ({ \ + float64x2_t __s0_459 = __p0_459; \ + float64x2_t __ret_459; \ + __ret_459 = splatq_laneq_f64(__s0_459, __p1_459); \ + __ret_459; \ }) #else -#define vdupq_laneq_f64(__p0_368, __p1_368) __extension__ ({ \ - float64x2_t __s0_368 = __p0_368; \ - float64x2_t __rev0_368; __rev0_368 = __builtin_shufflevector(__s0_368, __s0_368, 1, 0); \ - float64x2_t __ret_368; \ - __ret_368 = __noswap_splatq_laneq_f64(__rev0_368, __p1_368); \ - __ret_368 = __builtin_shufflevector(__ret_368, __ret_368, 1, 0); \ - __ret_368; \ +#define vdupq_laneq_f64(__p0_460, __p1_460) __extension__ ({ \ + float64x2_t __s0_460 = __p0_460; \ + float64x2_t __rev0_460; __rev0_460 = __builtin_shufflevector(__s0_460, __s0_460, 1, 0); \ + float64x2_t __ret_460; \ + __ret_460 = __noswap_splatq_laneq_f64(__rev0_460, __p1_460); \ + __ret_460 = __builtin_shufflevector(__ret_460, __ret_460, 1, 0); \ + __ret_460; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_f32(__p0_369, __p1_369) __extension__ ({ \ - float32x4_t __s0_369 = __p0_369; \ - float32x4_t __ret_369; \ - __ret_369 = splatq_laneq_f32(__s0_369, __p1_369); \ - __ret_369; \ +#define vdupq_laneq_f32(__p0_461, __p1_461) __extension__ ({ \ + float32x4_t __s0_461 = __p0_461; \ + float32x4_t __ret_461; \ + __ret_461 = splatq_laneq_f32(__s0_461, __p1_461); \ + __ret_461; \ }) #else -#define vdupq_laneq_f32(__p0_370, __p1_370) __extension__ ({ \ - float32x4_t __s0_370 = __p0_370; \ - float32x4_t __rev0_370; __rev0_370 = __builtin_shufflevector(__s0_370, __s0_370, 3, 2, 1, 0); \ - float32x4_t __ret_370; \ - __ret_370 = __noswap_splatq_laneq_f32(__rev0_370, __p1_370); \ - __ret_370 = __builtin_shufflevector(__ret_370, __ret_370, 3, 2, 1, 0); \ - __ret_370; \ +#define vdupq_laneq_f32(__p0_462, __p1_462) __extension__ ({ \ + float32x4_t __s0_462 = __p0_462; \ + float32x4_t __rev0_462; __rev0_462 = __builtin_shufflevector(__s0_462, __s0_462, 3, 2, 1, 0); \ + float32x4_t __ret_462; \ + __ret_462 = __noswap_splatq_laneq_f32(__rev0_462, __p1_462); \ + __ret_462 = __builtin_shufflevector(__ret_462, __ret_462, 3, 2, 1, 0); \ + __ret_462; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_f16(__p0_371, __p1_371) __extension__ ({ \ - float16x8_t __s0_371 = __p0_371; \ - float16x8_t __ret_371; \ - __ret_371 = splatq_laneq_f16(__s0_371, __p1_371); \ - __ret_371; \ +#define vdupq_laneq_f16(__p0_463, __p1_463) __extension__ ({ \ + float16x8_t __s0_463 = __p0_463; \ + float16x8_t __ret_463; \ + __ret_463 = splatq_laneq_f16(__s0_463, __p1_463); \ + __ret_463; \ }) #else -#define vdupq_laneq_f16(__p0_372, __p1_372) __extension__ ({ \ - float16x8_t __s0_372 = __p0_372; \ - float16x8_t __rev0_372; __rev0_372 = __builtin_shufflevector(__s0_372, __s0_372, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __ret_372; \ - __ret_372 = __noswap_splatq_laneq_f16(__rev0_372, __p1_372); \ - __ret_372 = __builtin_shufflevector(__ret_372, __ret_372, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_372; \ +#define vdupq_laneq_f16(__p0_464, __p1_464) __extension__ ({ \ + float16x8_t __s0_464 = __p0_464; \ + float16x8_t __rev0_464; __rev0_464 = __builtin_shufflevector(__s0_464, __s0_464, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_464; \ + __ret_464 = __noswap_splatq_laneq_f16(__rev0_464, __p1_464); \ + __ret_464 = __builtin_shufflevector(__ret_464, __ret_464, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_464; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_s32(__p0_373, __p1_373) __extension__ ({ \ - int32x4_t __s0_373 = __p0_373; \ - int32x4_t __ret_373; \ - __ret_373 = splatq_laneq_s32(__s0_373, __p1_373); \ - __ret_373; \ +#define vdupq_laneq_s32(__p0_465, __p1_465) __extension__ ({ \ + int32x4_t __s0_465 = __p0_465; \ + int32x4_t __ret_465; \ + __ret_465 = splatq_laneq_s32(__s0_465, __p1_465); \ + __ret_465; \ }) #else -#define vdupq_laneq_s32(__p0_374, __p1_374) __extension__ ({ \ - int32x4_t __s0_374 = __p0_374; \ - int32x4_t __rev0_374; __rev0_374 = __builtin_shufflevector(__s0_374, __s0_374, 3, 2, 1, 0); \ - int32x4_t __ret_374; \ - __ret_374 = __noswap_splatq_laneq_s32(__rev0_374, __p1_374); \ - __ret_374 = __builtin_shufflevector(__ret_374, __ret_374, 3, 2, 1, 0); \ - __ret_374; \ +#define vdupq_laneq_s32(__p0_466, __p1_466) __extension__ ({ \ + int32x4_t __s0_466 = __p0_466; \ + int32x4_t __rev0_466; __rev0_466 = __builtin_shufflevector(__s0_466, __s0_466, 3, 2, 1, 0); \ + int32x4_t __ret_466; \ + __ret_466 = __noswap_splatq_laneq_s32(__rev0_466, __p1_466); \ + __ret_466 = __builtin_shufflevector(__ret_466, __ret_466, 3, 2, 1, 0); \ + __ret_466; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_s64(__p0_375, __p1_375) __extension__ ({ \ - int64x2_t __s0_375 = __p0_375; \ - int64x2_t __ret_375; \ - __ret_375 = splatq_laneq_s64(__s0_375, __p1_375); \ - __ret_375; \ +#define vdupq_laneq_s64(__p0_467, __p1_467) __extension__ ({ \ + int64x2_t __s0_467 = __p0_467; \ + int64x2_t __ret_467; \ + __ret_467 = splatq_laneq_s64(__s0_467, __p1_467); \ + __ret_467; \ }) #else -#define vdupq_laneq_s64(__p0_376, __p1_376) __extension__ ({ \ - int64x2_t __s0_376 = __p0_376; \ - int64x2_t __rev0_376; __rev0_376 = __builtin_shufflevector(__s0_376, __s0_376, 1, 0); \ - int64x2_t __ret_376; \ - __ret_376 = __noswap_splatq_laneq_s64(__rev0_376, __p1_376); \ - __ret_376 = __builtin_shufflevector(__ret_376, __ret_376, 1, 0); \ - __ret_376; \ +#define vdupq_laneq_s64(__p0_468, __p1_468) __extension__ ({ \ + int64x2_t __s0_468 = __p0_468; \ + int64x2_t __rev0_468; __rev0_468 = __builtin_shufflevector(__s0_468, __s0_468, 1, 0); \ + int64x2_t __ret_468; \ + __ret_468 = __noswap_splatq_laneq_s64(__rev0_468, __p1_468); \ + __ret_468 = __builtin_shufflevector(__ret_468, __ret_468, 1, 0); \ + __ret_468; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdupq_laneq_s16(__p0_377, __p1_377) __extension__ ({ \ - int16x8_t __s0_377 = __p0_377; \ - int16x8_t __ret_377; \ - __ret_377 = splatq_laneq_s16(__s0_377, __p1_377); \ - __ret_377; \ +#define vdupq_laneq_s16(__p0_469, __p1_469) __extension__ ({ \ + int16x8_t __s0_469 = __p0_469; \ + int16x8_t __ret_469; \ + __ret_469 = splatq_laneq_s16(__s0_469, __p1_469); \ + __ret_469; \ }) #else -#define vdupq_laneq_s16(__p0_378, __p1_378) __extension__ ({ \ - int16x8_t __s0_378 = __p0_378; \ - int16x8_t __rev0_378; __rev0_378 = __builtin_shufflevector(__s0_378, __s0_378, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_378; \ - __ret_378 = __noswap_splatq_laneq_s16(__rev0_378, __p1_378); \ - __ret_378 = __builtin_shufflevector(__ret_378, __ret_378, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_378; \ +#define vdupq_laneq_s16(__p0_470, __p1_470) __extension__ ({ \ + int16x8_t __s0_470 = __p0_470; \ + int16x8_t __rev0_470; __rev0_470 = __builtin_shufflevector(__s0_470, __s0_470, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_470; \ + __ret_470 = __noswap_splatq_laneq_s16(__rev0_470, __p1_470); \ + __ret_470 = __builtin_shufflevector(__ret_470, __ret_470, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_470; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_u8(__p0_379, __p1_379) __extension__ ({ \ - uint8x16_t __s0_379 = __p0_379; \ - uint8x8_t __ret_379; \ - __ret_379 = splat_laneq_u8(__s0_379, __p1_379); \ - __ret_379; \ +#define vdup_laneq_u8(__p0_471, __p1_471) __extension__ ({ \ + uint8x16_t __s0_471 = __p0_471; \ + uint8x8_t __ret_471; \ + __ret_471 = splat_laneq_u8(__s0_471, __p1_471); \ + __ret_471; \ }) #else -#define vdup_laneq_u8(__p0_380, __p1_380) __extension__ ({ \ - uint8x16_t __s0_380 = __p0_380; \ - uint8x16_t __rev0_380; __rev0_380 = __builtin_shufflevector(__s0_380, __s0_380, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __ret_380; \ - __ret_380 = __noswap_splat_laneq_u8(__rev0_380, __p1_380); \ - __ret_380 = __builtin_shufflevector(__ret_380, __ret_380, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_380; \ +#define vdup_laneq_u8(__p0_472, __p1_472) __extension__ ({ \ + uint8x16_t __s0_472 = __p0_472; \ + uint8x16_t __rev0_472; __rev0_472 = __builtin_shufflevector(__s0_472, __s0_472, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __ret_472; \ + __ret_472 = __noswap_splat_laneq_u8(__rev0_472, __p1_472); \ + __ret_472 = __builtin_shufflevector(__ret_472, __ret_472, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_472; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_u32(__p0_381, __p1_381) __extension__ ({ \ - uint32x4_t __s0_381 = __p0_381; \ - uint32x2_t __ret_381; \ - __ret_381 = splat_laneq_u32(__s0_381, __p1_381); \ - __ret_381; \ +#define vdup_laneq_u32(__p0_473, __p1_473) __extension__ ({ \ + uint32x4_t __s0_473 = __p0_473; \ + uint32x2_t __ret_473; \ + __ret_473 = splat_laneq_u32(__s0_473, __p1_473); \ + __ret_473; \ }) #else -#define vdup_laneq_u32(__p0_382, __p1_382) __extension__ ({ \ - uint32x4_t __s0_382 = __p0_382; \ - uint32x4_t __rev0_382; __rev0_382 = __builtin_shufflevector(__s0_382, __s0_382, 3, 2, 1, 0); \ - uint32x2_t __ret_382; \ - __ret_382 = __noswap_splat_laneq_u32(__rev0_382, __p1_382); \ - __ret_382 = __builtin_shufflevector(__ret_382, __ret_382, 1, 0); \ - __ret_382; \ +#define vdup_laneq_u32(__p0_474, __p1_474) __extension__ ({ \ + uint32x4_t __s0_474 = __p0_474; \ + uint32x4_t __rev0_474; __rev0_474 = __builtin_shufflevector(__s0_474, __s0_474, 3, 2, 1, 0); \ + uint32x2_t __ret_474; \ + __ret_474 = __noswap_splat_laneq_u32(__rev0_474, __p1_474); \ + __ret_474 = __builtin_shufflevector(__ret_474, __ret_474, 1, 0); \ + __ret_474; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_u64(__p0_383, __p1_383) __extension__ ({ \ - uint64x2_t __s0_383 = __p0_383; \ - uint64x1_t __ret_383; \ - __ret_383 = splat_laneq_u64(__s0_383, __p1_383); \ - __ret_383; \ +#define vdup_laneq_u64(__p0_475, __p1_475) __extension__ ({ \ + uint64x2_t __s0_475 = __p0_475; \ + uint64x1_t __ret_475; \ + __ret_475 = splat_laneq_u64(__s0_475, __p1_475); \ + __ret_475; \ }) #else -#define vdup_laneq_u64(__p0_384, __p1_384) __extension__ ({ \ - uint64x2_t __s0_384 = __p0_384; \ - uint64x2_t __rev0_384; __rev0_384 = __builtin_shufflevector(__s0_384, __s0_384, 1, 0); \ - uint64x1_t __ret_384; \ - __ret_384 = __noswap_splat_laneq_u64(__rev0_384, __p1_384); \ - __ret_384; \ +#define vdup_laneq_u64(__p0_476, __p1_476) __extension__ ({ \ + uint64x2_t __s0_476 = __p0_476; \ + uint64x2_t __rev0_476; __rev0_476 = __builtin_shufflevector(__s0_476, __s0_476, 1, 0); \ + uint64x1_t __ret_476; \ + __ret_476 = __noswap_splat_laneq_u64(__rev0_476, __p1_476); \ + __ret_476; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_u16(__p0_385, __p1_385) __extension__ ({ \ - uint16x8_t __s0_385 = __p0_385; \ - uint16x4_t __ret_385; \ - __ret_385 = splat_laneq_u16(__s0_385, __p1_385); \ - __ret_385; \ +#define vdup_laneq_u16(__p0_477, __p1_477) __extension__ ({ \ + uint16x8_t __s0_477 = __p0_477; \ + uint16x4_t __ret_477; \ + __ret_477 = splat_laneq_u16(__s0_477, __p1_477); \ + __ret_477; \ }) #else -#define vdup_laneq_u16(__p0_386, __p1_386) __extension__ ({ \ - uint16x8_t __s0_386 = __p0_386; \ - uint16x8_t __rev0_386; __rev0_386 = __builtin_shufflevector(__s0_386, __s0_386, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __ret_386; \ - __ret_386 = __noswap_splat_laneq_u16(__rev0_386, __p1_386); \ - __ret_386 = __builtin_shufflevector(__ret_386, __ret_386, 3, 2, 1, 0); \ - __ret_386; \ +#define vdup_laneq_u16(__p0_478, __p1_478) __extension__ ({ \ + uint16x8_t __s0_478 = __p0_478; \ + uint16x8_t __rev0_478; __rev0_478 = __builtin_shufflevector(__s0_478, __s0_478, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __ret_478; \ + __ret_478 = __noswap_splat_laneq_u16(__rev0_478, __p1_478); \ + __ret_478 = __builtin_shufflevector(__ret_478, __ret_478, 3, 2, 1, 0); \ + __ret_478; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_s8(__p0_387, __p1_387) __extension__ ({ \ - int8x16_t __s0_387 = __p0_387; \ - int8x8_t __ret_387; \ - __ret_387 = splat_laneq_s8(__s0_387, __p1_387); \ - __ret_387; \ +#define vdup_laneq_s8(__p0_479, __p1_479) __extension__ ({ \ + int8x16_t __s0_479 = __p0_479; \ + int8x8_t __ret_479; \ + __ret_479 = splat_laneq_s8(__s0_479, __p1_479); \ + __ret_479; \ }) #else -#define vdup_laneq_s8(__p0_388, __p1_388) __extension__ ({ \ - int8x16_t __s0_388 = __p0_388; \ - int8x16_t __rev0_388; __rev0_388 = __builtin_shufflevector(__s0_388, __s0_388, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x8_t __ret_388; \ - __ret_388 = __noswap_splat_laneq_s8(__rev0_388, __p1_388); \ - __ret_388 = __builtin_shufflevector(__ret_388, __ret_388, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_388; \ +#define vdup_laneq_s8(__p0_480, __p1_480) __extension__ ({ \ + int8x16_t __s0_480 = __p0_480; \ + int8x16_t __rev0_480; __rev0_480 = __builtin_shufflevector(__s0_480, __s0_480, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x8_t __ret_480; \ + __ret_480 = __noswap_splat_laneq_s8(__rev0_480, __p1_480); \ + __ret_480 = __builtin_shufflevector(__ret_480, __ret_480, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_480; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_f64(__p0_389, __p1_389) __extension__ ({ \ - float64x2_t __s0_389 = __p0_389; \ - float64x1_t __ret_389; \ - __ret_389 = splat_laneq_f64(__s0_389, __p1_389); \ - __ret_389; \ +#define vdup_laneq_f64(__p0_481, __p1_481) __extension__ ({ \ + float64x2_t __s0_481 = __p0_481; \ + float64x1_t __ret_481; \ + __ret_481 = splat_laneq_f64(__s0_481, __p1_481); \ + __ret_481; \ }) #else -#define vdup_laneq_f64(__p0_390, __p1_390) __extension__ ({ \ - float64x2_t __s0_390 = __p0_390; \ - float64x2_t __rev0_390; __rev0_390 = __builtin_shufflevector(__s0_390, __s0_390, 1, 0); \ - float64x1_t __ret_390; \ - __ret_390 = __noswap_splat_laneq_f64(__rev0_390, __p1_390); \ - __ret_390; \ +#define vdup_laneq_f64(__p0_482, __p1_482) __extension__ ({ \ + float64x2_t __s0_482 = __p0_482; \ + float64x2_t __rev0_482; __rev0_482 = __builtin_shufflevector(__s0_482, __s0_482, 1, 0); \ + float64x1_t __ret_482; \ + __ret_482 = __noswap_splat_laneq_f64(__rev0_482, __p1_482); \ + __ret_482; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_f32(__p0_391, __p1_391) __extension__ ({ \ - float32x4_t __s0_391 = __p0_391; \ - float32x2_t __ret_391; \ - __ret_391 = splat_laneq_f32(__s0_391, __p1_391); \ - __ret_391; \ +#define vdup_laneq_f32(__p0_483, __p1_483) __extension__ ({ \ + float32x4_t __s0_483 = __p0_483; \ + float32x2_t __ret_483; \ + __ret_483 = splat_laneq_f32(__s0_483, __p1_483); \ + __ret_483; \ }) #else -#define vdup_laneq_f32(__p0_392, __p1_392) __extension__ ({ \ - float32x4_t __s0_392 = __p0_392; \ - float32x4_t __rev0_392; __rev0_392 = __builtin_shufflevector(__s0_392, __s0_392, 3, 2, 1, 0); \ - float32x2_t __ret_392; \ - __ret_392 = __noswap_splat_laneq_f32(__rev0_392, __p1_392); \ - __ret_392 = __builtin_shufflevector(__ret_392, __ret_392, 1, 0); \ - __ret_392; \ +#define vdup_laneq_f32(__p0_484, __p1_484) __extension__ ({ \ + float32x4_t __s0_484 = __p0_484; \ + float32x4_t __rev0_484; __rev0_484 = __builtin_shufflevector(__s0_484, __s0_484, 3, 2, 1, 0); \ + float32x2_t __ret_484; \ + __ret_484 = __noswap_splat_laneq_f32(__rev0_484, __p1_484); \ + __ret_484 = __builtin_shufflevector(__ret_484, __ret_484, 1, 0); \ + __ret_484; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_f16(__p0_393, __p1_393) __extension__ ({ \ - float16x8_t __s0_393 = __p0_393; \ - float16x4_t __ret_393; \ - __ret_393 = splat_laneq_f16(__s0_393, __p1_393); \ - __ret_393; \ +#define vdup_laneq_f16(__p0_485, __p1_485) __extension__ ({ \ + float16x8_t __s0_485 = __p0_485; \ + float16x4_t __ret_485; \ + __ret_485 = splat_laneq_f16(__s0_485, __p1_485); \ + __ret_485; \ }) #else -#define vdup_laneq_f16(__p0_394, __p1_394) __extension__ ({ \ - float16x8_t __s0_394 = __p0_394; \ - float16x8_t __rev0_394; __rev0_394 = __builtin_shufflevector(__s0_394, __s0_394, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __ret_394; \ - __ret_394 = __noswap_splat_laneq_f16(__rev0_394, __p1_394); \ - __ret_394 = __builtin_shufflevector(__ret_394, __ret_394, 3, 2, 1, 0); \ - __ret_394; \ +#define vdup_laneq_f16(__p0_486, __p1_486) __extension__ ({ \ + float16x8_t __s0_486 = __p0_486; \ + float16x8_t __rev0_486; __rev0_486 = __builtin_shufflevector(__s0_486, __s0_486, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __ret_486; \ + __ret_486 = __noswap_splat_laneq_f16(__rev0_486, __p1_486); \ + __ret_486 = __builtin_shufflevector(__ret_486, __ret_486, 3, 2, 1, 0); \ + __ret_486; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_s32(__p0_395, __p1_395) __extension__ ({ \ - int32x4_t __s0_395 = __p0_395; \ - int32x2_t __ret_395; \ - __ret_395 = splat_laneq_s32(__s0_395, __p1_395); \ - __ret_395; \ +#define vdup_laneq_s32(__p0_487, __p1_487) __extension__ ({ \ + int32x4_t __s0_487 = __p0_487; \ + int32x2_t __ret_487; \ + __ret_487 = splat_laneq_s32(__s0_487, __p1_487); \ + __ret_487; \ }) #else -#define vdup_laneq_s32(__p0_396, __p1_396) __extension__ ({ \ - int32x4_t __s0_396 = __p0_396; \ - int32x4_t __rev0_396; __rev0_396 = __builtin_shufflevector(__s0_396, __s0_396, 3, 2, 1, 0); \ - int32x2_t __ret_396; \ - __ret_396 = __noswap_splat_laneq_s32(__rev0_396, __p1_396); \ - __ret_396 = __builtin_shufflevector(__ret_396, __ret_396, 1, 0); \ - __ret_396; \ +#define vdup_laneq_s32(__p0_488, __p1_488) __extension__ ({ \ + int32x4_t __s0_488 = __p0_488; \ + int32x4_t __rev0_488; __rev0_488 = __builtin_shufflevector(__s0_488, __s0_488, 3, 2, 1, 0); \ + int32x2_t __ret_488; \ + __ret_488 = __noswap_splat_laneq_s32(__rev0_488, __p1_488); \ + __ret_488 = __builtin_shufflevector(__ret_488, __ret_488, 1, 0); \ + __ret_488; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_s64(__p0_397, __p1_397) __extension__ ({ \ - int64x2_t __s0_397 = __p0_397; \ - int64x1_t __ret_397; \ - __ret_397 = splat_laneq_s64(__s0_397, __p1_397); \ - __ret_397; \ +#define vdup_laneq_s64(__p0_489, __p1_489) __extension__ ({ \ + int64x2_t __s0_489 = __p0_489; \ + int64x1_t __ret_489; \ + __ret_489 = splat_laneq_s64(__s0_489, __p1_489); \ + __ret_489; \ }) #else -#define vdup_laneq_s64(__p0_398, __p1_398) __extension__ ({ \ - int64x2_t __s0_398 = __p0_398; \ - int64x2_t __rev0_398; __rev0_398 = __builtin_shufflevector(__s0_398, __s0_398, 1, 0); \ - int64x1_t __ret_398; \ - __ret_398 = __noswap_splat_laneq_s64(__rev0_398, __p1_398); \ - __ret_398; \ +#define vdup_laneq_s64(__p0_490, __p1_490) __extension__ ({ \ + int64x2_t __s0_490 = __p0_490; \ + int64x2_t __rev0_490; __rev0_490 = __builtin_shufflevector(__s0_490, __s0_490, 1, 0); \ + int64x1_t __ret_490; \ + __ret_490 = __noswap_splat_laneq_s64(__rev0_490, __p1_490); \ + __ret_490; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vdup_laneq_s16(__p0_399, __p1_399) __extension__ ({ \ - int16x8_t __s0_399 = __p0_399; \ - int16x4_t __ret_399; \ - __ret_399 = splat_laneq_s16(__s0_399, __p1_399); \ - __ret_399; \ +#define vdup_laneq_s16(__p0_491, __p1_491) __extension__ ({ \ + int16x8_t __s0_491 = __p0_491; \ + int16x4_t __ret_491; \ + __ret_491 = splat_laneq_s16(__s0_491, __p1_491); \ + __ret_491; \ }) #else -#define vdup_laneq_s16(__p0_400, __p1_400) __extension__ ({ \ - int16x8_t __s0_400 = __p0_400; \ - int16x8_t __rev0_400; __rev0_400 = __builtin_shufflevector(__s0_400, __s0_400, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __ret_400; \ - __ret_400 = __noswap_splat_laneq_s16(__rev0_400, __p1_400); \ - __ret_400 = __builtin_shufflevector(__ret_400, __ret_400, 3, 2, 1, 0); \ - __ret_400; \ +#define vdup_laneq_s16(__p0_492, __p1_492) __extension__ ({ \ + int16x8_t __s0_492 = __p0_492; \ + int16x8_t __rev0_492; __rev0_492 = __builtin_shufflevector(__s0_492, __s0_492, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __ret_492; \ + __ret_492 = __noswap_splat_laneq_s16(__rev0_492, __p1_492); \ + __ret_492 = __builtin_shufflevector(__ret_492, __ret_492, 3, 2, 1, 0); \ + __ret_492; \ }) #endif @@ -50287,246 +51991,246 @@ __ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) __ret = vfma_f64(__p0, -__p1, __p2); return __ret; } -#define vfmsd_lane_f64(__p0_401, __p1_401, __p2_401, __p3_401) __extension__ ({ \ - float64_t __s0_401 = __p0_401; \ - float64_t __s1_401 = __p1_401; \ - float64x1_t __s2_401 = __p2_401; \ - float64_t __ret_401; \ - __ret_401 = vfmad_lane_f64(__s0_401, -__s1_401, __s2_401, __p3_401); \ - __ret_401; \ +#define vfmsd_lane_f64(__p0_493, __p1_493, __p2_493, __p3_493) __extension__ ({ \ + float64_t __s0_493 = __p0_493; \ + float64_t __s1_493 = __p1_493; \ + float64x1_t __s2_493 = __p2_493; \ + float64_t __ret_493; \ + __ret_493 = vfmad_lane_f64(__s0_493, -__s1_493, __s2_493, __p3_493); \ + __ret_493; \ }) #ifdef __LITTLE_ENDIAN__ -#define vfmss_lane_f32(__p0_402, __p1_402, __p2_402, __p3_402) __extension__ ({ \ - float32_t __s0_402 = __p0_402; \ - float32_t __s1_402 = __p1_402; \ - float32x2_t __s2_402 = __p2_402; \ - float32_t __ret_402; \ - __ret_402 = vfmas_lane_f32(__s0_402, -__s1_402, __s2_402, __p3_402); \ - __ret_402; \ +#define vfmss_lane_f32(__p0_494, __p1_494, __p2_494, __p3_494) __extension__ ({ \ + float32_t __s0_494 = __p0_494; \ + float32_t __s1_494 = __p1_494; \ + float32x2_t __s2_494 = __p2_494; \ + float32_t __ret_494; \ + __ret_494 = vfmas_lane_f32(__s0_494, -__s1_494, __s2_494, __p3_494); \ + __ret_494; \ }) #else -#define vfmss_lane_f32(__p0_403, __p1_403, __p2_403, __p3_403) __extension__ ({ \ - float32_t __s0_403 = __p0_403; \ - float32_t __s1_403 = __p1_403; \ - float32x2_t __s2_403 = __p2_403; \ - float32x2_t __rev2_403; __rev2_403 = __builtin_shufflevector(__s2_403, __s2_403, 1, 0); \ - float32_t __ret_403; \ - __ret_403 = __noswap_vfmas_lane_f32(__s0_403, -__s1_403, __rev2_403, __p3_403); \ - __ret_403; \ +#define vfmss_lane_f32(__p0_495, __p1_495, __p2_495, __p3_495) __extension__ ({ \ + float32_t __s0_495 = __p0_495; \ + float32_t __s1_495 = __p1_495; \ + float32x2_t __s2_495 = __p2_495; \ + float32x2_t __rev2_495; __rev2_495 = __builtin_shufflevector(__s2_495, __s2_495, 1, 0); \ + float32_t __ret_495; \ + __ret_495 = __noswap_vfmas_lane_f32(__s0_495, -__s1_495, __rev2_495, __p3_495); \ + __ret_495; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsq_lane_f64(__p0_404, __p1_404, __p2_404, __p3_404) __extension__ ({ \ - float64x2_t __s0_404 = __p0_404; \ - float64x2_t __s1_404 = __p1_404; \ - float64x1_t __s2_404 = __p2_404; \ - float64x2_t __ret_404; \ - __ret_404 = vfmaq_lane_f64(__s0_404, -__s1_404, __s2_404, __p3_404); \ - __ret_404; \ +#define vfmsq_lane_f64(__p0_496, __p1_496, __p2_496, __p3_496) __extension__ ({ \ + float64x2_t __s0_496 = __p0_496; \ + float64x2_t __s1_496 = __p1_496; \ + float64x1_t __s2_496 = __p2_496; \ + float64x2_t __ret_496; \ + __ret_496 = vfmaq_lane_f64(__s0_496, -__s1_496, __s2_496, __p3_496); \ + __ret_496; \ }) #else -#define vfmsq_lane_f64(__p0_405, __p1_405, __p2_405, __p3_405) __extension__ ({ \ - float64x2_t __s0_405 = __p0_405; \ - float64x2_t __s1_405 = __p1_405; \ - float64x1_t __s2_405 = __p2_405; \ - float64x2_t __rev0_405; __rev0_405 = __builtin_shufflevector(__s0_405, __s0_405, 1, 0); \ - float64x2_t __rev1_405; __rev1_405 = __builtin_shufflevector(__s1_405, __s1_405, 1, 0); \ - float64x2_t __ret_405; \ - __ret_405 = __noswap_vfmaq_lane_f64(__rev0_405, -__rev1_405, __s2_405, __p3_405); \ - __ret_405 = __builtin_shufflevector(__ret_405, __ret_405, 1, 0); \ - __ret_405; \ +#define vfmsq_lane_f64(__p0_497, __p1_497, __p2_497, __p3_497) __extension__ ({ \ + float64x2_t __s0_497 = __p0_497; \ + float64x2_t __s1_497 = __p1_497; \ + float64x1_t __s2_497 = __p2_497; \ + float64x2_t __rev0_497; __rev0_497 = __builtin_shufflevector(__s0_497, __s0_497, 1, 0); \ + float64x2_t __rev1_497; __rev1_497 = __builtin_shufflevector(__s1_497, __s1_497, 1, 0); \ + float64x2_t __ret_497; \ + __ret_497 = __noswap_vfmaq_lane_f64(__rev0_497, -__rev1_497, __s2_497, __p3_497); \ + __ret_497 = __builtin_shufflevector(__ret_497, __ret_497, 1, 0); \ + __ret_497; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsq_lane_f32(__p0_406, __p1_406, __p2_406, __p3_406) __extension__ ({ \ - float32x4_t __s0_406 = __p0_406; \ - float32x4_t __s1_406 = __p1_406; \ - float32x2_t __s2_406 = __p2_406; \ - float32x4_t __ret_406; \ - __ret_406 = vfmaq_lane_f32(__s0_406, -__s1_406, __s2_406, __p3_406); \ - __ret_406; \ +#define vfmsq_lane_f32(__p0_498, __p1_498, __p2_498, __p3_498) __extension__ ({ \ + float32x4_t __s0_498 = __p0_498; \ + float32x4_t __s1_498 = __p1_498; \ + float32x2_t __s2_498 = __p2_498; \ + float32x4_t __ret_498; \ + __ret_498 = vfmaq_lane_f32(__s0_498, -__s1_498, __s2_498, __p3_498); \ + __ret_498; \ }) #else -#define vfmsq_lane_f32(__p0_407, __p1_407, __p2_407, __p3_407) __extension__ ({ \ - float32x4_t __s0_407 = __p0_407; \ - float32x4_t __s1_407 = __p1_407; \ - float32x2_t __s2_407 = __p2_407; \ - float32x4_t __rev0_407; __rev0_407 = __builtin_shufflevector(__s0_407, __s0_407, 3, 2, 1, 0); \ - float32x4_t __rev1_407; __rev1_407 = __builtin_shufflevector(__s1_407, __s1_407, 3, 2, 1, 0); \ - float32x2_t __rev2_407; __rev2_407 = __builtin_shufflevector(__s2_407, __s2_407, 1, 0); \ - float32x4_t __ret_407; \ - __ret_407 = __noswap_vfmaq_lane_f32(__rev0_407, -__rev1_407, __rev2_407, __p3_407); \ - __ret_407 = __builtin_shufflevector(__ret_407, __ret_407, 3, 2, 1, 0); \ - __ret_407; \ +#define vfmsq_lane_f32(__p0_499, __p1_499, __p2_499, __p3_499) __extension__ ({ \ + float32x4_t __s0_499 = __p0_499; \ + float32x4_t __s1_499 = __p1_499; \ + float32x2_t __s2_499 = __p2_499; \ + float32x4_t __rev0_499; __rev0_499 = __builtin_shufflevector(__s0_499, __s0_499, 3, 2, 1, 0); \ + float32x4_t __rev1_499; __rev1_499 = __builtin_shufflevector(__s1_499, __s1_499, 3, 2, 1, 0); \ + float32x2_t __rev2_499; __rev2_499 = __builtin_shufflevector(__s2_499, __s2_499, 1, 0); \ + float32x4_t __ret_499; \ + __ret_499 = __noswap_vfmaq_lane_f32(__rev0_499, -__rev1_499, __rev2_499, __p3_499); \ + __ret_499 = __builtin_shufflevector(__ret_499, __ret_499, 3, 2, 1, 0); \ + __ret_499; \ }) #endif -#define vfms_lane_f64(__p0_408, __p1_408, __p2_408, __p3_408) __extension__ ({ \ - float64x1_t __s0_408 = __p0_408; \ - float64x1_t __s1_408 = __p1_408; \ - float64x1_t __s2_408 = __p2_408; \ - float64x1_t __ret_408; \ - __ret_408 = vfma_lane_f64(__s0_408, -__s1_408, __s2_408, __p3_408); \ - __ret_408; \ +#define vfms_lane_f64(__p0_500, __p1_500, __p2_500, __p3_500) __extension__ ({ \ + float64x1_t __s0_500 = __p0_500; \ + float64x1_t __s1_500 = __p1_500; \ + float64x1_t __s2_500 = __p2_500; \ + float64x1_t __ret_500; \ + __ret_500 = vfma_lane_f64(__s0_500, -__s1_500, __s2_500, __p3_500); \ + __ret_500; \ }) #ifdef __LITTLE_ENDIAN__ -#define vfms_lane_f32(__p0_409, __p1_409, __p2_409, __p3_409) __extension__ ({ \ - float32x2_t __s0_409 = __p0_409; \ - float32x2_t __s1_409 = __p1_409; \ - float32x2_t __s2_409 = __p2_409; \ - float32x2_t __ret_409; \ - __ret_409 = vfma_lane_f32(__s0_409, -__s1_409, __s2_409, __p3_409); \ - __ret_409; \ +#define vfms_lane_f32(__p0_501, __p1_501, __p2_501, __p3_501) __extension__ ({ \ + float32x2_t __s0_501 = __p0_501; \ + float32x2_t __s1_501 = __p1_501; \ + float32x2_t __s2_501 = __p2_501; \ + float32x2_t __ret_501; \ + __ret_501 = vfma_lane_f32(__s0_501, -__s1_501, __s2_501, __p3_501); \ + __ret_501; \ }) #else -#define vfms_lane_f32(__p0_410, __p1_410, __p2_410, __p3_410) __extension__ ({ \ - float32x2_t __s0_410 = __p0_410; \ - float32x2_t __s1_410 = __p1_410; \ - float32x2_t __s2_410 = __p2_410; \ - float32x2_t __rev0_410; __rev0_410 = __builtin_shufflevector(__s0_410, __s0_410, 1, 0); \ - float32x2_t __rev1_410; __rev1_410 = __builtin_shufflevector(__s1_410, __s1_410, 1, 0); \ - float32x2_t __rev2_410; __rev2_410 = __builtin_shufflevector(__s2_410, __s2_410, 1, 0); \ - float32x2_t __ret_410; \ - __ret_410 = __noswap_vfma_lane_f32(__rev0_410, -__rev1_410, __rev2_410, __p3_410); \ - __ret_410 = __builtin_shufflevector(__ret_410, __ret_410, 1, 0); \ - __ret_410; \ +#define vfms_lane_f32(__p0_502, __p1_502, __p2_502, __p3_502) __extension__ ({ \ + float32x2_t __s0_502 = __p0_502; \ + float32x2_t __s1_502 = __p1_502; \ + float32x2_t __s2_502 = __p2_502; \ + float32x2_t __rev0_502; __rev0_502 = __builtin_shufflevector(__s0_502, __s0_502, 1, 0); \ + float32x2_t __rev1_502; __rev1_502 = __builtin_shufflevector(__s1_502, __s1_502, 1, 0); \ + float32x2_t __rev2_502; __rev2_502 = __builtin_shufflevector(__s2_502, __s2_502, 1, 0); \ + float32x2_t __ret_502; \ + __ret_502 = __noswap_vfma_lane_f32(__rev0_502, -__rev1_502, __rev2_502, __p3_502); \ + __ret_502 = __builtin_shufflevector(__ret_502, __ret_502, 1, 0); \ + __ret_502; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsd_laneq_f64(__p0_411, __p1_411, __p2_411, __p3_411) __extension__ ({ \ - float64_t __s0_411 = __p0_411; \ - float64_t __s1_411 = __p1_411; \ - float64x2_t __s2_411 = __p2_411; \ - float64_t __ret_411; \ - __ret_411 = vfmad_laneq_f64(__s0_411, -__s1_411, __s2_411, __p3_411); \ - __ret_411; \ +#define vfmsd_laneq_f64(__p0_503, __p1_503, __p2_503, __p3_503) __extension__ ({ \ + float64_t __s0_503 = __p0_503; \ + float64_t __s1_503 = __p1_503; \ + float64x2_t __s2_503 = __p2_503; \ + float64_t __ret_503; \ + __ret_503 = vfmad_laneq_f64(__s0_503, -__s1_503, __s2_503, __p3_503); \ + __ret_503; \ }) #else -#define vfmsd_laneq_f64(__p0_412, __p1_412, __p2_412, __p3_412) __extension__ ({ \ - float64_t __s0_412 = __p0_412; \ - float64_t __s1_412 = __p1_412; \ - float64x2_t __s2_412 = __p2_412; \ - float64x2_t __rev2_412; __rev2_412 = __builtin_shufflevector(__s2_412, __s2_412, 1, 0); \ - float64_t __ret_412; \ - __ret_412 = __noswap_vfmad_laneq_f64(__s0_412, -__s1_412, __rev2_412, __p3_412); \ - __ret_412; \ +#define vfmsd_laneq_f64(__p0_504, __p1_504, __p2_504, __p3_504) __extension__ ({ \ + float64_t __s0_504 = __p0_504; \ + float64_t __s1_504 = __p1_504; \ + float64x2_t __s2_504 = __p2_504; \ + float64x2_t __rev2_504; __rev2_504 = __builtin_shufflevector(__s2_504, __s2_504, 1, 0); \ + float64_t __ret_504; \ + __ret_504 = __noswap_vfmad_laneq_f64(__s0_504, -__s1_504, __rev2_504, __p3_504); \ + __ret_504; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmss_laneq_f32(__p0_413, __p1_413, __p2_413, __p3_413) __extension__ ({ \ - float32_t __s0_413 = __p0_413; \ - float32_t __s1_413 = __p1_413; \ - float32x4_t __s2_413 = __p2_413; \ - float32_t __ret_413; \ - __ret_413 = vfmas_laneq_f32(__s0_413, -__s1_413, __s2_413, __p3_413); \ - __ret_413; \ +#define vfmss_laneq_f32(__p0_505, __p1_505, __p2_505, __p3_505) __extension__ ({ \ + float32_t __s0_505 = __p0_505; \ + float32_t __s1_505 = __p1_505; \ + float32x4_t __s2_505 = __p2_505; \ + float32_t __ret_505; \ + __ret_505 = vfmas_laneq_f32(__s0_505, -__s1_505, __s2_505, __p3_505); \ + __ret_505; \ }) #else -#define vfmss_laneq_f32(__p0_414, __p1_414, __p2_414, __p3_414) __extension__ ({ \ - float32_t __s0_414 = __p0_414; \ - float32_t __s1_414 = __p1_414; \ - float32x4_t __s2_414 = __p2_414; \ - float32x4_t __rev2_414; __rev2_414 = __builtin_shufflevector(__s2_414, __s2_414, 3, 2, 1, 0); \ - float32_t __ret_414; \ - __ret_414 = __noswap_vfmas_laneq_f32(__s0_414, -__s1_414, __rev2_414, __p3_414); \ - __ret_414; \ +#define vfmss_laneq_f32(__p0_506, __p1_506, __p2_506, __p3_506) __extension__ ({ \ + float32_t __s0_506 = __p0_506; \ + float32_t __s1_506 = __p1_506; \ + float32x4_t __s2_506 = __p2_506; \ + float32x4_t __rev2_506; __rev2_506 = __builtin_shufflevector(__s2_506, __s2_506, 3, 2, 1, 0); \ + float32_t __ret_506; \ + __ret_506 = __noswap_vfmas_laneq_f32(__s0_506, -__s1_506, __rev2_506, __p3_506); \ + __ret_506; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsq_laneq_f64(__p0_415, __p1_415, __p2_415, __p3_415) __extension__ ({ \ - float64x2_t __s0_415 = __p0_415; \ - float64x2_t __s1_415 = __p1_415; \ - float64x2_t __s2_415 = __p2_415; \ - float64x2_t __ret_415; \ - __ret_415 = vfmaq_laneq_f64(__s0_415, -__s1_415, __s2_415, __p3_415); \ - __ret_415; \ +#define vfmsq_laneq_f64(__p0_507, __p1_507, __p2_507, __p3_507) __extension__ ({ \ + float64x2_t __s0_507 = __p0_507; \ + float64x2_t __s1_507 = __p1_507; \ + float64x2_t __s2_507 = __p2_507; \ + float64x2_t __ret_507; \ + __ret_507 = vfmaq_laneq_f64(__s0_507, -__s1_507, __s2_507, __p3_507); \ + __ret_507; \ }) #else -#define vfmsq_laneq_f64(__p0_416, __p1_416, __p2_416, __p3_416) __extension__ ({ \ - float64x2_t __s0_416 = __p0_416; \ - float64x2_t __s1_416 = __p1_416; \ - float64x2_t __s2_416 = __p2_416; \ - float64x2_t __rev0_416; __rev0_416 = __builtin_shufflevector(__s0_416, __s0_416, 1, 0); \ - float64x2_t __rev1_416; __rev1_416 = __builtin_shufflevector(__s1_416, __s1_416, 1, 0); \ - float64x2_t __rev2_416; __rev2_416 = __builtin_shufflevector(__s2_416, __s2_416, 1, 0); \ - float64x2_t __ret_416; \ - __ret_416 = __noswap_vfmaq_laneq_f64(__rev0_416, -__rev1_416, __rev2_416, __p3_416); \ - __ret_416 = __builtin_shufflevector(__ret_416, __ret_416, 1, 0); \ - __ret_416; \ +#define vfmsq_laneq_f64(__p0_508, __p1_508, __p2_508, __p3_508) __extension__ ({ \ + float64x2_t __s0_508 = __p0_508; \ + float64x2_t __s1_508 = __p1_508; \ + float64x2_t __s2_508 = __p2_508; \ + float64x2_t __rev0_508; __rev0_508 = __builtin_shufflevector(__s0_508, __s0_508, 1, 0); \ + float64x2_t __rev1_508; __rev1_508 = __builtin_shufflevector(__s1_508, __s1_508, 1, 0); \ + float64x2_t __rev2_508; __rev2_508 = __builtin_shufflevector(__s2_508, __s2_508, 1, 0); \ + float64x2_t __ret_508; \ + __ret_508 = __noswap_vfmaq_laneq_f64(__rev0_508, -__rev1_508, __rev2_508, __p3_508); \ + __ret_508 = __builtin_shufflevector(__ret_508, __ret_508, 1, 0); \ + __ret_508; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmsq_laneq_f32(__p0_417, __p1_417, __p2_417, __p3_417) __extension__ ({ \ - float32x4_t __s0_417 = __p0_417; \ - float32x4_t __s1_417 = __p1_417; \ - float32x4_t __s2_417 = __p2_417; \ - float32x4_t __ret_417; \ - __ret_417 = vfmaq_laneq_f32(__s0_417, -__s1_417, __s2_417, __p3_417); \ - __ret_417; \ +#define vfmsq_laneq_f32(__p0_509, __p1_509, __p2_509, __p3_509) __extension__ ({ \ + float32x4_t __s0_509 = __p0_509; \ + float32x4_t __s1_509 = __p1_509; \ + float32x4_t __s2_509 = __p2_509; \ + float32x4_t __ret_509; \ + __ret_509 = vfmaq_laneq_f32(__s0_509, -__s1_509, __s2_509, __p3_509); \ + __ret_509; \ }) #else -#define vfmsq_laneq_f32(__p0_418, __p1_418, __p2_418, __p3_418) __extension__ ({ \ - float32x4_t __s0_418 = __p0_418; \ - float32x4_t __s1_418 = __p1_418; \ - float32x4_t __s2_418 = __p2_418; \ - float32x4_t __rev0_418; __rev0_418 = __builtin_shufflevector(__s0_418, __s0_418, 3, 2, 1, 0); \ - float32x4_t __rev1_418; __rev1_418 = __builtin_shufflevector(__s1_418, __s1_418, 3, 2, 1, 0); \ - float32x4_t __rev2_418; __rev2_418 = __builtin_shufflevector(__s2_418, __s2_418, 3, 2, 1, 0); \ - float32x4_t __ret_418; \ - __ret_418 = __noswap_vfmaq_laneq_f32(__rev0_418, -__rev1_418, __rev2_418, __p3_418); \ - __ret_418 = __builtin_shufflevector(__ret_418, __ret_418, 3, 2, 1, 0); \ - __ret_418; \ +#define vfmsq_laneq_f32(__p0_510, __p1_510, __p2_510, __p3_510) __extension__ ({ \ + float32x4_t __s0_510 = __p0_510; \ + float32x4_t __s1_510 = __p1_510; \ + float32x4_t __s2_510 = __p2_510; \ + float32x4_t __rev0_510; __rev0_510 = __builtin_shufflevector(__s0_510, __s0_510, 3, 2, 1, 0); \ + float32x4_t __rev1_510; __rev1_510 = __builtin_shufflevector(__s1_510, __s1_510, 3, 2, 1, 0); \ + float32x4_t __rev2_510; __rev2_510 = __builtin_shufflevector(__s2_510, __s2_510, 3, 2, 1, 0); \ + float32x4_t __ret_510; \ + __ret_510 = __noswap_vfmaq_laneq_f32(__rev0_510, -__rev1_510, __rev2_510, __p3_510); \ + __ret_510 = __builtin_shufflevector(__ret_510, __ret_510, 3, 2, 1, 0); \ + __ret_510; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfms_laneq_f64(__p0_419, __p1_419, __p2_419, __p3_419) __extension__ ({ \ - float64x1_t __s0_419 = __p0_419; \ - float64x1_t __s1_419 = __p1_419; \ - float64x2_t __s2_419 = __p2_419; \ - float64x1_t __ret_419; \ - __ret_419 = vfma_laneq_f64(__s0_419, -__s1_419, __s2_419, __p3_419); \ - __ret_419; \ +#define vfms_laneq_f64(__p0_511, __p1_511, __p2_511, __p3_511) __extension__ ({ \ + float64x1_t __s0_511 = __p0_511; \ + float64x1_t __s1_511 = __p1_511; \ + float64x2_t __s2_511 = __p2_511; \ + float64x1_t __ret_511; \ + __ret_511 = vfma_laneq_f64(__s0_511, -__s1_511, __s2_511, __p3_511); \ + __ret_511; \ }) #else -#define vfms_laneq_f64(__p0_420, __p1_420, __p2_420, __p3_420) __extension__ ({ \ - float64x1_t __s0_420 = __p0_420; \ - float64x1_t __s1_420 = __p1_420; \ - float64x2_t __s2_420 = __p2_420; \ - float64x2_t __rev2_420; __rev2_420 = __builtin_shufflevector(__s2_420, __s2_420, 1, 0); \ - float64x1_t __ret_420; \ - __ret_420 = __noswap_vfma_laneq_f64(__s0_420, -__s1_420, __rev2_420, __p3_420); \ - __ret_420; \ +#define vfms_laneq_f64(__p0_512, __p1_512, __p2_512, __p3_512) __extension__ ({ \ + float64x1_t __s0_512 = __p0_512; \ + float64x1_t __s1_512 = __p1_512; \ + float64x2_t __s2_512 = __p2_512; \ + float64x2_t __rev2_512; __rev2_512 = __builtin_shufflevector(__s2_512, __s2_512, 1, 0); \ + float64x1_t __ret_512; \ + __ret_512 = __noswap_vfma_laneq_f64(__s0_512, -__s1_512, __rev2_512, __p3_512); \ + __ret_512; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfms_laneq_f32(__p0_421, __p1_421, __p2_421, __p3_421) __extension__ ({ \ - float32x2_t __s0_421 = __p0_421; \ - float32x2_t __s1_421 = __p1_421; \ - float32x4_t __s2_421 = __p2_421; \ - float32x2_t __ret_421; \ - __ret_421 = vfma_laneq_f32(__s0_421, -__s1_421, __s2_421, __p3_421); \ - __ret_421; \ +#define vfms_laneq_f32(__p0_513, __p1_513, __p2_513, __p3_513) __extension__ ({ \ + float32x2_t __s0_513 = __p0_513; \ + float32x2_t __s1_513 = __p1_513; \ + float32x4_t __s2_513 = __p2_513; \ + float32x2_t __ret_513; \ + __ret_513 = vfma_laneq_f32(__s0_513, -__s1_513, __s2_513, __p3_513); \ + __ret_513; \ }) #else -#define vfms_laneq_f32(__p0_422, __p1_422, __p2_422, __p3_422) __extension__ ({ \ - float32x2_t __s0_422 = __p0_422; \ - float32x2_t __s1_422 = __p1_422; \ - float32x4_t __s2_422 = __p2_422; \ - float32x2_t __rev0_422; __rev0_422 = __builtin_shufflevector(__s0_422, __s0_422, 1, 0); \ - float32x2_t __rev1_422; __rev1_422 = __builtin_shufflevector(__s1_422, __s1_422, 1, 0); \ - float32x4_t __rev2_422; __rev2_422 = __builtin_shufflevector(__s2_422, __s2_422, 3, 2, 1, 0); \ - float32x2_t __ret_422; \ - __ret_422 = __noswap_vfma_laneq_f32(__rev0_422, -__rev1_422, __rev2_422, __p3_422); \ - __ret_422 = __builtin_shufflevector(__ret_422, __ret_422, 1, 0); \ - __ret_422; \ +#define vfms_laneq_f32(__p0_514, __p1_514, __p2_514, __p3_514) __extension__ ({ \ + float32x2_t __s0_514 = __p0_514; \ + float32x2_t __s1_514 = __p1_514; \ + float32x4_t __s2_514 = __p2_514; \ + float32x2_t __rev0_514; __rev0_514 = __builtin_shufflevector(__s0_514, __s0_514, 1, 0); \ + float32x2_t __rev1_514; __rev1_514 = __builtin_shufflevector(__s1_514, __s1_514, 1, 0); \ + float32x4_t __rev2_514; __rev2_514 = __builtin_shufflevector(__s2_514, __s2_514, 3, 2, 1, 0); \ + float32x2_t __ret_514; \ + __ret_514 = __noswap_vfma_laneq_f32(__rev0_514, -__rev1_514, __rev2_514, __p3_514); \ + __ret_514 = __builtin_shufflevector(__ret_514, __ret_514, 1, 0); \ + __ret_514; \ }) #endif @@ -52548,547 +54252,530 @@ __ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) return __ret; } #ifdef __LITTLE_ENDIAN__ -#define vmlaq_laneq_u32(__p0_423, __p1_423, __p2_423, __p3_423) __extension__ ({ \ - uint32x4_t __s0_423 = __p0_423; \ - uint32x4_t __s1_423 = __p1_423; \ - uint32x4_t __s2_423 = __p2_423; \ - uint32x4_t __ret_423; \ - __ret_423 = __s0_423 + __s1_423 * splatq_laneq_u32(__s2_423, __p3_423); \ - __ret_423; \ +#define vmlaq_laneq_u32(__p0_515, __p1_515, __p2_515, __p3_515) __extension__ ({ \ + uint32x4_t __s0_515 = __p0_515; \ + uint32x4_t __s1_515 = __p1_515; \ + uint32x4_t __s2_515 = __p2_515; \ + uint32x4_t __ret_515; \ + __ret_515 = __s0_515 + __s1_515 * splatq_laneq_u32(__s2_515, __p3_515); \ + __ret_515; \ }) #else -#define vmlaq_laneq_u32(__p0_424, __p1_424, __p2_424, __p3_424) __extension__ ({ \ - uint32x4_t __s0_424 = __p0_424; \ - uint32x4_t __s1_424 = __p1_424; \ - uint32x4_t __s2_424 = __p2_424; \ - uint32x4_t __rev0_424; __rev0_424 = __builtin_shufflevector(__s0_424, __s0_424, 3, 2, 1, 0); \ - uint32x4_t __rev1_424; __rev1_424 = __builtin_shufflevector(__s1_424, __s1_424, 3, 2, 1, 0); \ - uint32x4_t __rev2_424; __rev2_424 = __builtin_shufflevector(__s2_424, __s2_424, 3, 2, 1, 0); \ - uint32x4_t __ret_424; \ - __ret_424 = __rev0_424 + __rev1_424 * __noswap_splatq_laneq_u32(__rev2_424, __p3_424); \ - __ret_424 = __builtin_shufflevector(__ret_424, __ret_424, 3, 2, 1, 0); \ - __ret_424; \ +#define vmlaq_laneq_u32(__p0_516, __p1_516, __p2_516, __p3_516) __extension__ ({ \ + uint32x4_t __s0_516 = __p0_516; \ + uint32x4_t __s1_516 = __p1_516; \ + uint32x4_t __s2_516 = __p2_516; \ + uint32x4_t __rev0_516; __rev0_516 = __builtin_shufflevector(__s0_516, __s0_516, 3, 2, 1, 0); \ + uint32x4_t __rev1_516; __rev1_516 = __builtin_shufflevector(__s1_516, __s1_516, 3, 2, 1, 0); \ + uint32x4_t __rev2_516; __rev2_516 = __builtin_shufflevector(__s2_516, __s2_516, 3, 2, 1, 0); \ + uint32x4_t __ret_516; \ + __ret_516 = __rev0_516 + __rev1_516 * __noswap_splatq_laneq_u32(__rev2_516, __p3_516); \ + __ret_516 = __builtin_shufflevector(__ret_516, __ret_516, 3, 2, 1, 0); \ + __ret_516; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlaq_laneq_u16(__p0_425, __p1_425, __p2_425, __p3_425) __extension__ ({ \ - uint16x8_t __s0_425 = __p0_425; \ - uint16x8_t __s1_425 = __p1_425; \ - uint16x8_t __s2_425 = __p2_425; \ - uint16x8_t __ret_425; \ - __ret_425 = __s0_425 + __s1_425 * splatq_laneq_u16(__s2_425, __p3_425); \ - __ret_425; \ +#define vmlaq_laneq_u16(__p0_517, __p1_517, __p2_517, __p3_517) __extension__ ({ \ + uint16x8_t __s0_517 = __p0_517; \ + uint16x8_t __s1_517 = __p1_517; \ + uint16x8_t __s2_517 = __p2_517; \ + uint16x8_t __ret_517; \ + __ret_517 = __s0_517 + __s1_517 * splatq_laneq_u16(__s2_517, __p3_517); \ + __ret_517; \ }) #else -#define vmlaq_laneq_u16(__p0_426, __p1_426, __p2_426, __p3_426) __extension__ ({ \ - uint16x8_t __s0_426 = __p0_426; \ - uint16x8_t __s1_426 = __p1_426; \ - uint16x8_t __s2_426 = __p2_426; \ - uint16x8_t __rev0_426; __rev0_426 = __builtin_shufflevector(__s0_426, __s0_426, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_426; __rev1_426 = __builtin_shufflevector(__s1_426, __s1_426, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev2_426; __rev2_426 = __builtin_shufflevector(__s2_426, __s2_426, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __ret_426; \ - __ret_426 = __rev0_426 + __rev1_426 * __noswap_splatq_laneq_u16(__rev2_426, __p3_426); \ - __ret_426 = __builtin_shufflevector(__ret_426, __ret_426, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_426; \ +#define vmlaq_laneq_u16(__p0_518, __p1_518, __p2_518, __p3_518) __extension__ ({ \ + uint16x8_t __s0_518 = __p0_518; \ + uint16x8_t __s1_518 = __p1_518; \ + uint16x8_t __s2_518 = __p2_518; \ + uint16x8_t __rev0_518; __rev0_518 = __builtin_shufflevector(__s0_518, __s0_518, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_518; __rev1_518 = __builtin_shufflevector(__s1_518, __s1_518, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev2_518; __rev2_518 = __builtin_shufflevector(__s2_518, __s2_518, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __ret_518; \ + __ret_518 = __rev0_518 + __rev1_518 * __noswap_splatq_laneq_u16(__rev2_518, __p3_518); \ + __ret_518 = __builtin_shufflevector(__ret_518, __ret_518, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_518; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlaq_laneq_f32(__p0_427, __p1_427, __p2_427, __p3_427) __extension__ ({ \ - float32x4_t __s0_427 = __p0_427; \ - float32x4_t __s1_427 = __p1_427; \ - float32x4_t __s2_427 = __p2_427; \ - float32x4_t __ret_427; \ - __ret_427 = __s0_427 + __s1_427 * splatq_laneq_f32(__s2_427, __p3_427); \ - __ret_427; \ +#define vmlaq_laneq_f32(__p0_519, __p1_519, __p2_519, __p3_519) __extension__ ({ \ + float32x4_t __s0_519 = __p0_519; \ + float32x4_t __s1_519 = __p1_519; \ + float32x4_t __s2_519 = __p2_519; \ + float32x4_t __ret_519; \ + __ret_519 = __s0_519 + __s1_519 * splatq_laneq_f32(__s2_519, __p3_519); \ + __ret_519; \ }) #else -#define vmlaq_laneq_f32(__p0_428, __p1_428, __p2_428, __p3_428) __extension__ ({ \ - float32x4_t __s0_428 = __p0_428; \ - float32x4_t __s1_428 = __p1_428; \ - float32x4_t __s2_428 = __p2_428; \ - float32x4_t __rev0_428; __rev0_428 = __builtin_shufflevector(__s0_428, __s0_428, 3, 2, 1, 0); \ - float32x4_t __rev1_428; __rev1_428 = __builtin_shufflevector(__s1_428, __s1_428, 3, 2, 1, 0); \ - float32x4_t __rev2_428; __rev2_428 = __builtin_shufflevector(__s2_428, __s2_428, 3, 2, 1, 0); \ - float32x4_t __ret_428; \ - __ret_428 = __rev0_428 + __rev1_428 * __noswap_splatq_laneq_f32(__rev2_428, __p3_428); \ - __ret_428 = __builtin_shufflevector(__ret_428, __ret_428, 3, 2, 1, 0); \ - __ret_428; \ +#define vmlaq_laneq_f32(__p0_520, __p1_520, __p2_520, __p3_520) __extension__ ({ \ + float32x4_t __s0_520 = __p0_520; \ + float32x4_t __s1_520 = __p1_520; \ + float32x4_t __s2_520 = __p2_520; \ + float32x4_t __rev0_520; __rev0_520 = __builtin_shufflevector(__s0_520, __s0_520, 3, 2, 1, 0); \ + float32x4_t __rev1_520; __rev1_520 = __builtin_shufflevector(__s1_520, __s1_520, 3, 2, 1, 0); \ + float32x4_t __rev2_520; __rev2_520 = __builtin_shufflevector(__s2_520, __s2_520, 3, 2, 1, 0); \ + float32x4_t __ret_520; \ + __ret_520 = __rev0_520 + __rev1_520 * __noswap_splatq_laneq_f32(__rev2_520, __p3_520); \ + __ret_520 = __builtin_shufflevector(__ret_520, __ret_520, 3, 2, 1, 0); \ + __ret_520; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlaq_laneq_s32(__p0_429, __p1_429, __p2_429, __p3_429) __extension__ ({ \ - int32x4_t __s0_429 = __p0_429; \ - int32x4_t __s1_429 = __p1_429; \ - int32x4_t __s2_429 = __p2_429; \ - int32x4_t __ret_429; \ - __ret_429 = __s0_429 + __s1_429 * splatq_laneq_s32(__s2_429, __p3_429); \ - __ret_429; \ +#define vmlaq_laneq_s32(__p0_521, __p1_521, __p2_521, __p3_521) __extension__ ({ \ + int32x4_t __s0_521 = __p0_521; \ + int32x4_t __s1_521 = __p1_521; \ + int32x4_t __s2_521 = __p2_521; \ + int32x4_t __ret_521; \ + __ret_521 = __s0_521 + __s1_521 * splatq_laneq_s32(__s2_521, __p3_521); \ + __ret_521; \ }) #else -#define vmlaq_laneq_s32(__p0_430, __p1_430, __p2_430, __p3_430) __extension__ ({ \ - int32x4_t __s0_430 = __p0_430; \ - int32x4_t __s1_430 = __p1_430; \ - int32x4_t __s2_430 = __p2_430; \ - int32x4_t __rev0_430; __rev0_430 = __builtin_shufflevector(__s0_430, __s0_430, 3, 2, 1, 0); \ - int32x4_t __rev1_430; __rev1_430 = __builtin_shufflevector(__s1_430, __s1_430, 3, 2, 1, 0); \ - int32x4_t __rev2_430; __rev2_430 = __builtin_shufflevector(__s2_430, __s2_430, 3, 2, 1, 0); \ - int32x4_t __ret_430; \ - __ret_430 = __rev0_430 + __rev1_430 * __noswap_splatq_laneq_s32(__rev2_430, __p3_430); \ - __ret_430 = __builtin_shufflevector(__ret_430, __ret_430, 3, 2, 1, 0); \ - __ret_430; \ +#define vmlaq_laneq_s32(__p0_522, __p1_522, __p2_522, __p3_522) __extension__ ({ \ + int32x4_t __s0_522 = __p0_522; \ + int32x4_t __s1_522 = __p1_522; \ + int32x4_t __s2_522 = __p2_522; \ + int32x4_t __rev0_522; __rev0_522 = __builtin_shufflevector(__s0_522, __s0_522, 3, 2, 1, 0); \ + int32x4_t __rev1_522; __rev1_522 = __builtin_shufflevector(__s1_522, __s1_522, 3, 2, 1, 0); \ + int32x4_t __rev2_522; __rev2_522 = __builtin_shufflevector(__s2_522, __s2_522, 3, 2, 1, 0); \ + int32x4_t __ret_522; \ + __ret_522 = __rev0_522 + __rev1_522 * __noswap_splatq_laneq_s32(__rev2_522, __p3_522); \ + __ret_522 = __builtin_shufflevector(__ret_522, __ret_522, 3, 2, 1, 0); \ + __ret_522; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlaq_laneq_s16(__p0_431, __p1_431, __p2_431, __p3_431) __extension__ ({ \ - int16x8_t __s0_431 = __p0_431; \ - int16x8_t __s1_431 = __p1_431; \ - int16x8_t __s2_431 = __p2_431; \ - int16x8_t __ret_431; \ - __ret_431 = __s0_431 + __s1_431 * splatq_laneq_s16(__s2_431, __p3_431); \ - __ret_431; \ +#define vmlaq_laneq_s16(__p0_523, __p1_523, __p2_523, __p3_523) __extension__ ({ \ + int16x8_t __s0_523 = __p0_523; \ + int16x8_t __s1_523 = __p1_523; \ + int16x8_t __s2_523 = __p2_523; \ + int16x8_t __ret_523; \ + __ret_523 = __s0_523 + __s1_523 * splatq_laneq_s16(__s2_523, __p3_523); \ + __ret_523; \ }) #else -#define vmlaq_laneq_s16(__p0_432, __p1_432, __p2_432, __p3_432) __extension__ ({ \ - int16x8_t __s0_432 = __p0_432; \ - int16x8_t __s1_432 = __p1_432; \ - int16x8_t __s2_432 = __p2_432; \ - int16x8_t __rev0_432; __rev0_432 = __builtin_shufflevector(__s0_432, __s0_432, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_432; __rev1_432 = __builtin_shufflevector(__s1_432, __s1_432, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_432; __rev2_432 = __builtin_shufflevector(__s2_432, __s2_432, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_432; \ - __ret_432 = __rev0_432 + __rev1_432 * __noswap_splatq_laneq_s16(__rev2_432, __p3_432); \ - __ret_432 = __builtin_shufflevector(__ret_432, __ret_432, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_432; \ +#define vmlaq_laneq_s16(__p0_524, __p1_524, __p2_524, __p3_524) __extension__ ({ \ + int16x8_t __s0_524 = __p0_524; \ + int16x8_t __s1_524 = __p1_524; \ + int16x8_t __s2_524 = __p2_524; \ + int16x8_t __rev0_524; __rev0_524 = __builtin_shufflevector(__s0_524, __s0_524, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_524; __rev1_524 = __builtin_shufflevector(__s1_524, __s1_524, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_524; __rev2_524 = __builtin_shufflevector(__s2_524, __s2_524, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_524; \ + __ret_524 = __rev0_524 + __rev1_524 * __noswap_splatq_laneq_s16(__rev2_524, __p3_524); \ + __ret_524 = __builtin_shufflevector(__ret_524, __ret_524, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_524; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmla_laneq_u32(__p0_433, __p1_433, __p2_433, __p3_433) __extension__ ({ \ - uint32x2_t __s0_433 = __p0_433; \ - uint32x2_t __s1_433 = __p1_433; \ - uint32x4_t __s2_433 = __p2_433; \ - uint32x2_t __ret_433; \ - __ret_433 = __s0_433 + __s1_433 * splat_laneq_u32(__s2_433, __p3_433); \ - __ret_433; \ +#define vmla_laneq_u32(__p0_525, __p1_525, __p2_525, __p3_525) __extension__ ({ \ + uint32x2_t __s0_525 = __p0_525; \ + uint32x2_t __s1_525 = __p1_525; \ + uint32x4_t __s2_525 = __p2_525; \ + uint32x2_t __ret_525; \ + __ret_525 = __s0_525 + __s1_525 * splat_laneq_u32(__s2_525, __p3_525); \ + __ret_525; \ }) #else -#define vmla_laneq_u32(__p0_434, __p1_434, __p2_434, __p3_434) __extension__ ({ \ - uint32x2_t __s0_434 = __p0_434; \ - uint32x2_t __s1_434 = __p1_434; \ - uint32x4_t __s2_434 = __p2_434; \ - uint32x2_t __rev0_434; __rev0_434 = __builtin_shufflevector(__s0_434, __s0_434, 1, 0); \ - uint32x2_t __rev1_434; __rev1_434 = __builtin_shufflevector(__s1_434, __s1_434, 1, 0); \ - uint32x4_t __rev2_434; __rev2_434 = __builtin_shufflevector(__s2_434, __s2_434, 3, 2, 1, 0); \ - uint32x2_t __ret_434; \ - __ret_434 = __rev0_434 + __rev1_434 * __noswap_splat_laneq_u32(__rev2_434, __p3_434); \ - __ret_434 = __builtin_shufflevector(__ret_434, __ret_434, 1, 0); \ - __ret_434; \ +#define vmla_laneq_u32(__p0_526, __p1_526, __p2_526, __p3_526) __extension__ ({ \ + uint32x2_t __s0_526 = __p0_526; \ + uint32x2_t __s1_526 = __p1_526; \ + uint32x4_t __s2_526 = __p2_526; \ + uint32x2_t __rev0_526; __rev0_526 = __builtin_shufflevector(__s0_526, __s0_526, 1, 0); \ + uint32x2_t __rev1_526; __rev1_526 = __builtin_shufflevector(__s1_526, __s1_526, 1, 0); \ + uint32x4_t __rev2_526; __rev2_526 = __builtin_shufflevector(__s2_526, __s2_526, 3, 2, 1, 0); \ + uint32x2_t __ret_526; \ + __ret_526 = __rev0_526 + __rev1_526 * __noswap_splat_laneq_u32(__rev2_526, __p3_526); \ + __ret_526 = __builtin_shufflevector(__ret_526, __ret_526, 1, 0); \ + __ret_526; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmla_laneq_u16(__p0_435, __p1_435, __p2_435, __p3_435) __extension__ ({ \ - uint16x4_t __s0_435 = __p0_435; \ - uint16x4_t __s1_435 = __p1_435; \ - uint16x8_t __s2_435 = __p2_435; \ - uint16x4_t __ret_435; \ - __ret_435 = __s0_435 + __s1_435 * splat_laneq_u16(__s2_435, __p3_435); \ - __ret_435; \ +#define vmla_laneq_u16(__p0_527, __p1_527, __p2_527, __p3_527) __extension__ ({ \ + uint16x4_t __s0_527 = __p0_527; \ + uint16x4_t __s1_527 = __p1_527; \ + uint16x8_t __s2_527 = __p2_527; \ + uint16x4_t __ret_527; \ + __ret_527 = __s0_527 + __s1_527 * splat_laneq_u16(__s2_527, __p3_527); \ + __ret_527; \ }) #else -#define vmla_laneq_u16(__p0_436, __p1_436, __p2_436, __p3_436) __extension__ ({ \ - uint16x4_t __s0_436 = __p0_436; \ - uint16x4_t __s1_436 = __p1_436; \ - uint16x8_t __s2_436 = __p2_436; \ - uint16x4_t __rev0_436; __rev0_436 = __builtin_shufflevector(__s0_436, __s0_436, 3, 2, 1, 0); \ - uint16x4_t __rev1_436; __rev1_436 = __builtin_shufflevector(__s1_436, __s1_436, 3, 2, 1, 0); \ - uint16x8_t __rev2_436; __rev2_436 = __builtin_shufflevector(__s2_436, __s2_436, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __ret_436; \ - __ret_436 = __rev0_436 + __rev1_436 * __noswap_splat_laneq_u16(__rev2_436, __p3_436); \ - __ret_436 = __builtin_shufflevector(__ret_436, __ret_436, 3, 2, 1, 0); \ - __ret_436; \ +#define vmla_laneq_u16(__p0_528, __p1_528, __p2_528, __p3_528) __extension__ ({ \ + uint16x4_t __s0_528 = __p0_528; \ + uint16x4_t __s1_528 = __p1_528; \ + uint16x8_t __s2_528 = __p2_528; \ + uint16x4_t __rev0_528; __rev0_528 = __builtin_shufflevector(__s0_528, __s0_528, 3, 2, 1, 0); \ + uint16x4_t __rev1_528; __rev1_528 = __builtin_shufflevector(__s1_528, __s1_528, 3, 2, 1, 0); \ + uint16x8_t __rev2_528; __rev2_528 = __builtin_shufflevector(__s2_528, __s2_528, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __ret_528; \ + __ret_528 = __rev0_528 + __rev1_528 * __noswap_splat_laneq_u16(__rev2_528, __p3_528); \ + __ret_528 = __builtin_shufflevector(__ret_528, __ret_528, 3, 2, 1, 0); \ + __ret_528; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmla_laneq_f32(__p0_437, __p1_437, __p2_437, __p3_437) __extension__ ({ \ - float32x2_t __s0_437 = __p0_437; \ - float32x2_t __s1_437 = __p1_437; \ - float32x4_t __s2_437 = __p2_437; \ - float32x2_t __ret_437; \ - __ret_437 = __s0_437 + __s1_437 * splat_laneq_f32(__s2_437, __p3_437); \ - __ret_437; \ +#define vmla_laneq_f32(__p0_529, __p1_529, __p2_529, __p3_529) __extension__ ({ \ + float32x2_t __s0_529 = __p0_529; \ + float32x2_t __s1_529 = __p1_529; \ + float32x4_t __s2_529 = __p2_529; \ + float32x2_t __ret_529; \ + __ret_529 = __s0_529 + __s1_529 * splat_laneq_f32(__s2_529, __p3_529); \ + __ret_529; \ }) #else -#define vmla_laneq_f32(__p0_438, __p1_438, __p2_438, __p3_438) __extension__ ({ \ - float32x2_t __s0_438 = __p0_438; \ - float32x2_t __s1_438 = __p1_438; \ - float32x4_t __s2_438 = __p2_438; \ - float32x2_t __rev0_438; __rev0_438 = __builtin_shufflevector(__s0_438, __s0_438, 1, 0); \ - float32x2_t __rev1_438; __rev1_438 = __builtin_shufflevector(__s1_438, __s1_438, 1, 0); \ - float32x4_t __rev2_438; __rev2_438 = __builtin_shufflevector(__s2_438, __s2_438, 3, 2, 1, 0); \ - float32x2_t __ret_438; \ - __ret_438 = __rev0_438 + __rev1_438 * __noswap_splat_laneq_f32(__rev2_438, __p3_438); \ - __ret_438 = __builtin_shufflevector(__ret_438, __ret_438, 1, 0); \ - __ret_438; \ +#define vmla_laneq_f32(__p0_530, __p1_530, __p2_530, __p3_530) __extension__ ({ \ + float32x2_t __s0_530 = __p0_530; \ + float32x2_t __s1_530 = __p1_530; \ + float32x4_t __s2_530 = __p2_530; \ + float32x2_t __rev0_530; __rev0_530 = __builtin_shufflevector(__s0_530, __s0_530, 1, 0); \ + float32x2_t __rev1_530; __rev1_530 = __builtin_shufflevector(__s1_530, __s1_530, 1, 0); \ + float32x4_t __rev2_530; __rev2_530 = __builtin_shufflevector(__s2_530, __s2_530, 3, 2, 1, 0); \ + float32x2_t __ret_530; \ + __ret_530 = __rev0_530 + __rev1_530 * __noswap_splat_laneq_f32(__rev2_530, __p3_530); \ + __ret_530 = __builtin_shufflevector(__ret_530, __ret_530, 1, 0); \ + __ret_530; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmla_laneq_s32(__p0_439, __p1_439, __p2_439, __p3_439) __extension__ ({ \ - int32x2_t __s0_439 = __p0_439; \ - int32x2_t __s1_439 = __p1_439; \ - int32x4_t __s2_439 = __p2_439; \ - int32x2_t __ret_439; \ - __ret_439 = __s0_439 + __s1_439 * splat_laneq_s32(__s2_439, __p3_439); \ - __ret_439; \ +#define vmla_laneq_s32(__p0_531, __p1_531, __p2_531, __p3_531) __extension__ ({ \ + int32x2_t __s0_531 = __p0_531; \ + int32x2_t __s1_531 = __p1_531; \ + int32x4_t __s2_531 = __p2_531; \ + int32x2_t __ret_531; \ + __ret_531 = __s0_531 + __s1_531 * splat_laneq_s32(__s2_531, __p3_531); \ + __ret_531; \ }) #else -#define vmla_laneq_s32(__p0_440, __p1_440, __p2_440, __p3_440) __extension__ ({ \ - int32x2_t __s0_440 = __p0_440; \ - int32x2_t __s1_440 = __p1_440; \ - int32x4_t __s2_440 = __p2_440; \ - int32x2_t __rev0_440; __rev0_440 = __builtin_shufflevector(__s0_440, __s0_440, 1, 0); \ - int32x2_t __rev1_440; __rev1_440 = __builtin_shufflevector(__s1_440, __s1_440, 1, 0); \ - int32x4_t __rev2_440; __rev2_440 = __builtin_shufflevector(__s2_440, __s2_440, 3, 2, 1, 0); \ - int32x2_t __ret_440; \ - __ret_440 = __rev0_440 + __rev1_440 * __noswap_splat_laneq_s32(__rev2_440, __p3_440); \ - __ret_440 = __builtin_shufflevector(__ret_440, __ret_440, 1, 0); \ - __ret_440; \ +#define vmla_laneq_s32(__p0_532, __p1_532, __p2_532, __p3_532) __extension__ ({ \ + int32x2_t __s0_532 = __p0_532; \ + int32x2_t __s1_532 = __p1_532; \ + int32x4_t __s2_532 = __p2_532; \ + int32x2_t __rev0_532; __rev0_532 = __builtin_shufflevector(__s0_532, __s0_532, 1, 0); \ + int32x2_t __rev1_532; __rev1_532 = __builtin_shufflevector(__s1_532, __s1_532, 1, 0); \ + int32x4_t __rev2_532; __rev2_532 = __builtin_shufflevector(__s2_532, __s2_532, 3, 2, 1, 0); \ + int32x2_t __ret_532; \ + __ret_532 = __rev0_532 + __rev1_532 * __noswap_splat_laneq_s32(__rev2_532, __p3_532); \ + __ret_532 = __builtin_shufflevector(__ret_532, __ret_532, 1, 0); \ + __ret_532; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmla_laneq_s16(__p0_441, __p1_441, __p2_441, __p3_441) __extension__ ({ \ - int16x4_t __s0_441 = __p0_441; \ - int16x4_t __s1_441 = __p1_441; \ - int16x8_t __s2_441 = __p2_441; \ - int16x4_t __ret_441; \ - __ret_441 = __s0_441 + __s1_441 * splat_laneq_s16(__s2_441, __p3_441); \ - __ret_441; \ +#define vmla_laneq_s16(__p0_533, __p1_533, __p2_533, __p3_533) __extension__ ({ \ + int16x4_t __s0_533 = __p0_533; \ + int16x4_t __s1_533 = __p1_533; \ + int16x8_t __s2_533 = __p2_533; \ + int16x4_t __ret_533; \ + __ret_533 = __s0_533 + __s1_533 * splat_laneq_s16(__s2_533, __p3_533); \ + __ret_533; \ }) #else -#define vmla_laneq_s16(__p0_442, __p1_442, __p2_442, __p3_442) __extension__ ({ \ - int16x4_t __s0_442 = __p0_442; \ - int16x4_t __s1_442 = __p1_442; \ - int16x8_t __s2_442 = __p2_442; \ - int16x4_t __rev0_442; __rev0_442 = __builtin_shufflevector(__s0_442, __s0_442, 3, 2, 1, 0); \ - int16x4_t __rev1_442; __rev1_442 = __builtin_shufflevector(__s1_442, __s1_442, 3, 2, 1, 0); \ - int16x8_t __rev2_442; __rev2_442 = __builtin_shufflevector(__s2_442, __s2_442, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __ret_442; \ - __ret_442 = __rev0_442 + __rev1_442 * __noswap_splat_laneq_s16(__rev2_442, __p3_442); \ - __ret_442 = __builtin_shufflevector(__ret_442, __ret_442, 3, 2, 1, 0); \ - __ret_442; \ +#define vmla_laneq_s16(__p0_534, __p1_534, __p2_534, __p3_534) __extension__ ({ \ + int16x4_t __s0_534 = __p0_534; \ + int16x4_t __s1_534 = __p1_534; \ + int16x8_t __s2_534 = __p2_534; \ + int16x4_t __rev0_534; __rev0_534 = __builtin_shufflevector(__s0_534, __s0_534, 3, 2, 1, 0); \ + int16x4_t __rev1_534; __rev1_534 = __builtin_shufflevector(__s1_534, __s1_534, 3, 2, 1, 0); \ + int16x8_t __rev2_534; __rev2_534 = __builtin_shufflevector(__s2_534, __s2_534, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __ret_534; \ + __ret_534 = __rev0_534 + __rev1_534 * __noswap_splat_laneq_s16(__rev2_534, __p3_534); \ + __ret_534 = __builtin_shufflevector(__ret_534, __ret_534, 3, 2, 1, 0); \ + __ret_534; \ }) #endif #ifdef __LITTLE_ENDIAN__ -__ai float64x2_t vmlaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) { - float64x2_t __ret; - __ret = __p0 + __p1 * (float64x2_t) {__p2, __p2}; - return __ret; -} -#else -__ai float64x2_t vmlaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) { - float64x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); - float64x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); - float64x2_t __ret; - __ret = __rev0 + __rev1 * (float64x2_t) {__p2, __p2}; - __ret = __builtin_shufflevector(__ret, __ret, 1, 0); - return __ret; -} -#endif - -#ifdef __LITTLE_ENDIAN__ -#define vmlal_high_lane_u32(__p0_443, __p1_443, __p2_443, __p3_443) __extension__ ({ \ - uint64x2_t __s0_443 = __p0_443; \ - uint32x4_t __s1_443 = __p1_443; \ - uint32x2_t __s2_443 = __p2_443; \ - uint64x2_t __ret_443; \ - __ret_443 = __s0_443 + vmull_u32(vget_high_u32(__s1_443), splat_lane_u32(__s2_443, __p3_443)); \ - __ret_443; \ +#define vmlal_high_lane_u32(__p0_535, __p1_535, __p2_535, __p3_535) __extension__ ({ \ + uint64x2_t __s0_535 = __p0_535; \ + uint32x4_t __s1_535 = __p1_535; \ + uint32x2_t __s2_535 = __p2_535; \ + uint64x2_t __ret_535; \ + __ret_535 = __s0_535 + vmull_u32(vget_high_u32(__s1_535), splat_lane_u32(__s2_535, __p3_535)); \ + __ret_535; \ }) #else -#define vmlal_high_lane_u32(__p0_444, __p1_444, __p2_444, __p3_444) __extension__ ({ \ - uint64x2_t __s0_444 = __p0_444; \ - uint32x4_t __s1_444 = __p1_444; \ - uint32x2_t __s2_444 = __p2_444; \ - uint64x2_t __rev0_444; __rev0_444 = __builtin_shufflevector(__s0_444, __s0_444, 1, 0); \ - uint32x4_t __rev1_444; __rev1_444 = __builtin_shufflevector(__s1_444, __s1_444, 3, 2, 1, 0); \ - uint32x2_t __rev2_444; __rev2_444 = __builtin_shufflevector(__s2_444, __s2_444, 1, 0); \ - uint64x2_t __ret_444; \ - __ret_444 = __rev0_444 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_444), __noswap_splat_lane_u32(__rev2_444, __p3_444)); \ - __ret_444 = __builtin_shufflevector(__ret_444, __ret_444, 1, 0); \ - __ret_444; \ +#define vmlal_high_lane_u32(__p0_536, __p1_536, __p2_536, __p3_536) __extension__ ({ \ + uint64x2_t __s0_536 = __p0_536; \ + uint32x4_t __s1_536 = __p1_536; \ + uint32x2_t __s2_536 = __p2_536; \ + uint64x2_t __rev0_536; __rev0_536 = __builtin_shufflevector(__s0_536, __s0_536, 1, 0); \ + uint32x4_t __rev1_536; __rev1_536 = __builtin_shufflevector(__s1_536, __s1_536, 3, 2, 1, 0); \ + uint32x2_t __rev2_536; __rev2_536 = __builtin_shufflevector(__s2_536, __s2_536, 1, 0); \ + uint64x2_t __ret_536; \ + __ret_536 = __rev0_536 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_536), __noswap_splat_lane_u32(__rev2_536, __p3_536)); \ + __ret_536 = __builtin_shufflevector(__ret_536, __ret_536, 1, 0); \ + __ret_536; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_high_lane_u16(__p0_445, __p1_445, __p2_445, __p3_445) __extension__ ({ \ - uint32x4_t __s0_445 = __p0_445; \ - uint16x8_t __s1_445 = __p1_445; \ - uint16x4_t __s2_445 = __p2_445; \ - uint32x4_t __ret_445; \ - __ret_445 = __s0_445 + vmull_u16(vget_high_u16(__s1_445), splat_lane_u16(__s2_445, __p3_445)); \ - __ret_445; \ +#define vmlal_high_lane_u16(__p0_537, __p1_537, __p2_537, __p3_537) __extension__ ({ \ + uint32x4_t __s0_537 = __p0_537; \ + uint16x8_t __s1_537 = __p1_537; \ + uint16x4_t __s2_537 = __p2_537; \ + uint32x4_t __ret_537; \ + __ret_537 = __s0_537 + vmull_u16(vget_high_u16(__s1_537), splat_lane_u16(__s2_537, __p3_537)); \ + __ret_537; \ }) #else -#define vmlal_high_lane_u16(__p0_446, __p1_446, __p2_446, __p3_446) __extension__ ({ \ - uint32x4_t __s0_446 = __p0_446; \ - uint16x8_t __s1_446 = __p1_446; \ - uint16x4_t __s2_446 = __p2_446; \ - uint32x4_t __rev0_446; __rev0_446 = __builtin_shufflevector(__s0_446, __s0_446, 3, 2, 1, 0); \ - uint16x8_t __rev1_446; __rev1_446 = __builtin_shufflevector(__s1_446, __s1_446, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __rev2_446; __rev2_446 = __builtin_shufflevector(__s2_446, __s2_446, 3, 2, 1, 0); \ - uint32x4_t __ret_446; \ - __ret_446 = __rev0_446 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_446), __noswap_splat_lane_u16(__rev2_446, __p3_446)); \ - __ret_446 = __builtin_shufflevector(__ret_446, __ret_446, 3, 2, 1, 0); \ - __ret_446; \ +#define vmlal_high_lane_u16(__p0_538, __p1_538, __p2_538, __p3_538) __extension__ ({ \ + uint32x4_t __s0_538 = __p0_538; \ + uint16x8_t __s1_538 = __p1_538; \ + uint16x4_t __s2_538 = __p2_538; \ + uint32x4_t __rev0_538; __rev0_538 = __builtin_shufflevector(__s0_538, __s0_538, 3, 2, 1, 0); \ + uint16x8_t __rev1_538; __rev1_538 = __builtin_shufflevector(__s1_538, __s1_538, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __rev2_538; __rev2_538 = __builtin_shufflevector(__s2_538, __s2_538, 3, 2, 1, 0); \ + uint32x4_t __ret_538; \ + __ret_538 = __rev0_538 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_538), __noswap_splat_lane_u16(__rev2_538, __p3_538)); \ + __ret_538 = __builtin_shufflevector(__ret_538, __ret_538, 3, 2, 1, 0); \ + __ret_538; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_high_lane_s32(__p0_447, __p1_447, __p2_447, __p3_447) __extension__ ({ \ - int64x2_t __s0_447 = __p0_447; \ - int32x4_t __s1_447 = __p1_447; \ - int32x2_t __s2_447 = __p2_447; \ - int64x2_t __ret_447; \ - __ret_447 = __s0_447 + vmull_s32(vget_high_s32(__s1_447), splat_lane_s32(__s2_447, __p3_447)); \ - __ret_447; \ +#define vmlal_high_lane_s32(__p0_539, __p1_539, __p2_539, __p3_539) __extension__ ({ \ + int64x2_t __s0_539 = __p0_539; \ + int32x4_t __s1_539 = __p1_539; \ + int32x2_t __s2_539 = __p2_539; \ + int64x2_t __ret_539; \ + __ret_539 = __s0_539 + vmull_s32(vget_high_s32(__s1_539), splat_lane_s32(__s2_539, __p3_539)); \ + __ret_539; \ }) #else -#define vmlal_high_lane_s32(__p0_448, __p1_448, __p2_448, __p3_448) __extension__ ({ \ - int64x2_t __s0_448 = __p0_448; \ - int32x4_t __s1_448 = __p1_448; \ - int32x2_t __s2_448 = __p2_448; \ - int64x2_t __rev0_448; __rev0_448 = __builtin_shufflevector(__s0_448, __s0_448, 1, 0); \ - int32x4_t __rev1_448; __rev1_448 = __builtin_shufflevector(__s1_448, __s1_448, 3, 2, 1, 0); \ - int32x2_t __rev2_448; __rev2_448 = __builtin_shufflevector(__s2_448, __s2_448, 1, 0); \ - int64x2_t __ret_448; \ - __ret_448 = __rev0_448 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_448), __noswap_splat_lane_s32(__rev2_448, __p3_448)); \ - __ret_448 = __builtin_shufflevector(__ret_448, __ret_448, 1, 0); \ - __ret_448; \ +#define vmlal_high_lane_s32(__p0_540, __p1_540, __p2_540, __p3_540) __extension__ ({ \ + int64x2_t __s0_540 = __p0_540; \ + int32x4_t __s1_540 = __p1_540; \ + int32x2_t __s2_540 = __p2_540; \ + int64x2_t __rev0_540; __rev0_540 = __builtin_shufflevector(__s0_540, __s0_540, 1, 0); \ + int32x4_t __rev1_540; __rev1_540 = __builtin_shufflevector(__s1_540, __s1_540, 3, 2, 1, 0); \ + int32x2_t __rev2_540; __rev2_540 = __builtin_shufflevector(__s2_540, __s2_540, 1, 0); \ + int64x2_t __ret_540; \ + __ret_540 = __rev0_540 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_540), __noswap_splat_lane_s32(__rev2_540, __p3_540)); \ + __ret_540 = __builtin_shufflevector(__ret_540, __ret_540, 1, 0); \ + __ret_540; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_high_lane_s16(__p0_449, __p1_449, __p2_449, __p3_449) __extension__ ({ \ - int32x4_t __s0_449 = __p0_449; \ - int16x8_t __s1_449 = __p1_449; \ - int16x4_t __s2_449 = __p2_449; \ - int32x4_t __ret_449; \ - __ret_449 = __s0_449 + vmull_s16(vget_high_s16(__s1_449), splat_lane_s16(__s2_449, __p3_449)); \ - __ret_449; \ +#define vmlal_high_lane_s16(__p0_541, __p1_541, __p2_541, __p3_541) __extension__ ({ \ + int32x4_t __s0_541 = __p0_541; \ + int16x8_t __s1_541 = __p1_541; \ + int16x4_t __s2_541 = __p2_541; \ + int32x4_t __ret_541; \ + __ret_541 = __s0_541 + vmull_s16(vget_high_s16(__s1_541), splat_lane_s16(__s2_541, __p3_541)); \ + __ret_541; \ }) #else -#define vmlal_high_lane_s16(__p0_450, __p1_450, __p2_450, __p3_450) __extension__ ({ \ - int32x4_t __s0_450 = __p0_450; \ - int16x8_t __s1_450 = __p1_450; \ - int16x4_t __s2_450 = __p2_450; \ - int32x4_t __rev0_450; __rev0_450 = __builtin_shufflevector(__s0_450, __s0_450, 3, 2, 1, 0); \ - int16x8_t __rev1_450; __rev1_450 = __builtin_shufflevector(__s1_450, __s1_450, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev2_450; __rev2_450 = __builtin_shufflevector(__s2_450, __s2_450, 3, 2, 1, 0); \ - int32x4_t __ret_450; \ - __ret_450 = __rev0_450 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_450), __noswap_splat_lane_s16(__rev2_450, __p3_450)); \ - __ret_450 = __builtin_shufflevector(__ret_450, __ret_450, 3, 2, 1, 0); \ - __ret_450; \ +#define vmlal_high_lane_s16(__p0_542, __p1_542, __p2_542, __p3_542) __extension__ ({ \ + int32x4_t __s0_542 = __p0_542; \ + int16x8_t __s1_542 = __p1_542; \ + int16x4_t __s2_542 = __p2_542; \ + int32x4_t __rev0_542; __rev0_542 = __builtin_shufflevector(__s0_542, __s0_542, 3, 2, 1, 0); \ + int16x8_t __rev1_542; __rev1_542 = __builtin_shufflevector(__s1_542, __s1_542, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev2_542; __rev2_542 = __builtin_shufflevector(__s2_542, __s2_542, 3, 2, 1, 0); \ + int32x4_t __ret_542; \ + __ret_542 = __rev0_542 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_542), __noswap_splat_lane_s16(__rev2_542, __p3_542)); \ + __ret_542 = __builtin_shufflevector(__ret_542, __ret_542, 3, 2, 1, 0); \ + __ret_542; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_high_laneq_u32(__p0_451, __p1_451, __p2_451, __p3_451) __extension__ ({ \ - uint64x2_t __s0_451 = __p0_451; \ - uint32x4_t __s1_451 = __p1_451; \ - uint32x4_t __s2_451 = __p2_451; \ - uint64x2_t __ret_451; \ - __ret_451 = __s0_451 + vmull_u32(vget_high_u32(__s1_451), splat_laneq_u32(__s2_451, __p3_451)); \ - __ret_451; \ +#define vmlal_high_laneq_u32(__p0_543, __p1_543, __p2_543, __p3_543) __extension__ ({ \ + uint64x2_t __s0_543 = __p0_543; \ + uint32x4_t __s1_543 = __p1_543; \ + uint32x4_t __s2_543 = __p2_543; \ + uint64x2_t __ret_543; \ + __ret_543 = __s0_543 + vmull_u32(vget_high_u32(__s1_543), splat_laneq_u32(__s2_543, __p3_543)); \ + __ret_543; \ }) #else -#define vmlal_high_laneq_u32(__p0_452, __p1_452, __p2_452, __p3_452) __extension__ ({ \ - uint64x2_t __s0_452 = __p0_452; \ - uint32x4_t __s1_452 = __p1_452; \ - uint32x4_t __s2_452 = __p2_452; \ - uint64x2_t __rev0_452; __rev0_452 = __builtin_shufflevector(__s0_452, __s0_452, 1, 0); \ - uint32x4_t __rev1_452; __rev1_452 = __builtin_shufflevector(__s1_452, __s1_452, 3, 2, 1, 0); \ - uint32x4_t __rev2_452; __rev2_452 = __builtin_shufflevector(__s2_452, __s2_452, 3, 2, 1, 0); \ - uint64x2_t __ret_452; \ - __ret_452 = __rev0_452 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_452), __noswap_splat_laneq_u32(__rev2_452, __p3_452)); \ - __ret_452 = __builtin_shufflevector(__ret_452, __ret_452, 1, 0); \ - __ret_452; \ +#define vmlal_high_laneq_u32(__p0_544, __p1_544, __p2_544, __p3_544) __extension__ ({ \ + uint64x2_t __s0_544 = __p0_544; \ + uint32x4_t __s1_544 = __p1_544; \ + uint32x4_t __s2_544 = __p2_544; \ + uint64x2_t __rev0_544; __rev0_544 = __builtin_shufflevector(__s0_544, __s0_544, 1, 0); \ + uint32x4_t __rev1_544; __rev1_544 = __builtin_shufflevector(__s1_544, __s1_544, 3, 2, 1, 0); \ + uint32x4_t __rev2_544; __rev2_544 = __builtin_shufflevector(__s2_544, __s2_544, 3, 2, 1, 0); \ + uint64x2_t __ret_544; \ + __ret_544 = __rev0_544 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_544), __noswap_splat_laneq_u32(__rev2_544, __p3_544)); \ + __ret_544 = __builtin_shufflevector(__ret_544, __ret_544, 1, 0); \ + __ret_544; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_high_laneq_u16(__p0_453, __p1_453, __p2_453, __p3_453) __extension__ ({ \ - uint32x4_t __s0_453 = __p0_453; \ - uint16x8_t __s1_453 = __p1_453; \ - uint16x8_t __s2_453 = __p2_453; \ - uint32x4_t __ret_453; \ - __ret_453 = __s0_453 + vmull_u16(vget_high_u16(__s1_453), splat_laneq_u16(__s2_453, __p3_453)); \ - __ret_453; \ +#define vmlal_high_laneq_u16(__p0_545, __p1_545, __p2_545, __p3_545) __extension__ ({ \ + uint32x4_t __s0_545 = __p0_545; \ + uint16x8_t __s1_545 = __p1_545; \ + uint16x8_t __s2_545 = __p2_545; \ + uint32x4_t __ret_545; \ + __ret_545 = __s0_545 + vmull_u16(vget_high_u16(__s1_545), splat_laneq_u16(__s2_545, __p3_545)); \ + __ret_545; \ }) #else -#define vmlal_high_laneq_u16(__p0_454, __p1_454, __p2_454, __p3_454) __extension__ ({ \ - uint32x4_t __s0_454 = __p0_454; \ - uint16x8_t __s1_454 = __p1_454; \ - uint16x8_t __s2_454 = __p2_454; \ - uint32x4_t __rev0_454; __rev0_454 = __builtin_shufflevector(__s0_454, __s0_454, 3, 2, 1, 0); \ - uint16x8_t __rev1_454; __rev1_454 = __builtin_shufflevector(__s1_454, __s1_454, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev2_454; __rev2_454 = __builtin_shufflevector(__s2_454, __s2_454, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_454; \ - __ret_454 = __rev0_454 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_454), __noswap_splat_laneq_u16(__rev2_454, __p3_454)); \ - __ret_454 = __builtin_shufflevector(__ret_454, __ret_454, 3, 2, 1, 0); \ - __ret_454; \ +#define vmlal_high_laneq_u16(__p0_546, __p1_546, __p2_546, __p3_546) __extension__ ({ \ + uint32x4_t __s0_546 = __p0_546; \ + uint16x8_t __s1_546 = __p1_546; \ + uint16x8_t __s2_546 = __p2_546; \ + uint32x4_t __rev0_546; __rev0_546 = __builtin_shufflevector(__s0_546, __s0_546, 3, 2, 1, 0); \ + uint16x8_t __rev1_546; __rev1_546 = __builtin_shufflevector(__s1_546, __s1_546, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev2_546; __rev2_546 = __builtin_shufflevector(__s2_546, __s2_546, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_546; \ + __ret_546 = __rev0_546 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_546), __noswap_splat_laneq_u16(__rev2_546, __p3_546)); \ + __ret_546 = __builtin_shufflevector(__ret_546, __ret_546, 3, 2, 1, 0); \ + __ret_546; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_high_laneq_s32(__p0_455, __p1_455, __p2_455, __p3_455) __extension__ ({ \ - int64x2_t __s0_455 = __p0_455; \ - int32x4_t __s1_455 = __p1_455; \ - int32x4_t __s2_455 = __p2_455; \ - int64x2_t __ret_455; \ - __ret_455 = __s0_455 + vmull_s32(vget_high_s32(__s1_455), splat_laneq_s32(__s2_455, __p3_455)); \ - __ret_455; \ +#define vmlal_high_laneq_s32(__p0_547, __p1_547, __p2_547, __p3_547) __extension__ ({ \ + int64x2_t __s0_547 = __p0_547; \ + int32x4_t __s1_547 = __p1_547; \ + int32x4_t __s2_547 = __p2_547; \ + int64x2_t __ret_547; \ + __ret_547 = __s0_547 + vmull_s32(vget_high_s32(__s1_547), splat_laneq_s32(__s2_547, __p3_547)); \ + __ret_547; \ }) #else -#define vmlal_high_laneq_s32(__p0_456, __p1_456, __p2_456, __p3_456) __extension__ ({ \ - int64x2_t __s0_456 = __p0_456; \ - int32x4_t __s1_456 = __p1_456; \ - int32x4_t __s2_456 = __p2_456; \ - int64x2_t __rev0_456; __rev0_456 = __builtin_shufflevector(__s0_456, __s0_456, 1, 0); \ - int32x4_t __rev1_456; __rev1_456 = __builtin_shufflevector(__s1_456, __s1_456, 3, 2, 1, 0); \ - int32x4_t __rev2_456; __rev2_456 = __builtin_shufflevector(__s2_456, __s2_456, 3, 2, 1, 0); \ - int64x2_t __ret_456; \ - __ret_456 = __rev0_456 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_456), __noswap_splat_laneq_s32(__rev2_456, __p3_456)); \ - __ret_456 = __builtin_shufflevector(__ret_456, __ret_456, 1, 0); \ - __ret_456; \ +#define vmlal_high_laneq_s32(__p0_548, __p1_548, __p2_548, __p3_548) __extension__ ({ \ + int64x2_t __s0_548 = __p0_548; \ + int32x4_t __s1_548 = __p1_548; \ + int32x4_t __s2_548 = __p2_548; \ + int64x2_t __rev0_548; __rev0_548 = __builtin_shufflevector(__s0_548, __s0_548, 1, 0); \ + int32x4_t __rev1_548; __rev1_548 = __builtin_shufflevector(__s1_548, __s1_548, 3, 2, 1, 0); \ + int32x4_t __rev2_548; __rev2_548 = __builtin_shufflevector(__s2_548, __s2_548, 3, 2, 1, 0); \ + int64x2_t __ret_548; \ + __ret_548 = __rev0_548 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_548), __noswap_splat_laneq_s32(__rev2_548, __p3_548)); \ + __ret_548 = __builtin_shufflevector(__ret_548, __ret_548, 1, 0); \ + __ret_548; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_high_laneq_s16(__p0_457, __p1_457, __p2_457, __p3_457) __extension__ ({ \ - int32x4_t __s0_457 = __p0_457; \ - int16x8_t __s1_457 = __p1_457; \ - int16x8_t __s2_457 = __p2_457; \ - int32x4_t __ret_457; \ - __ret_457 = __s0_457 + vmull_s16(vget_high_s16(__s1_457), splat_laneq_s16(__s2_457, __p3_457)); \ - __ret_457; \ +#define vmlal_high_laneq_s16(__p0_549, __p1_549, __p2_549, __p3_549) __extension__ ({ \ + int32x4_t __s0_549 = __p0_549; \ + int16x8_t __s1_549 = __p1_549; \ + int16x8_t __s2_549 = __p2_549; \ + int32x4_t __ret_549; \ + __ret_549 = __s0_549 + vmull_s16(vget_high_s16(__s1_549), splat_laneq_s16(__s2_549, __p3_549)); \ + __ret_549; \ }) #else -#define vmlal_high_laneq_s16(__p0_458, __p1_458, __p2_458, __p3_458) __extension__ ({ \ - int32x4_t __s0_458 = __p0_458; \ - int16x8_t __s1_458 = __p1_458; \ - int16x8_t __s2_458 = __p2_458; \ - int32x4_t __rev0_458; __rev0_458 = __builtin_shufflevector(__s0_458, __s0_458, 3, 2, 1, 0); \ - int16x8_t __rev1_458; __rev1_458 = __builtin_shufflevector(__s1_458, __s1_458, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_458; __rev2_458 = __builtin_shufflevector(__s2_458, __s2_458, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_458; \ - __ret_458 = __rev0_458 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_458), __noswap_splat_laneq_s16(__rev2_458, __p3_458)); \ - __ret_458 = __builtin_shufflevector(__ret_458, __ret_458, 3, 2, 1, 0); \ - __ret_458; \ +#define vmlal_high_laneq_s16(__p0_550, __p1_550, __p2_550, __p3_550) __extension__ ({ \ + int32x4_t __s0_550 = __p0_550; \ + int16x8_t __s1_550 = __p1_550; \ + int16x8_t __s2_550 = __p2_550; \ + int32x4_t __rev0_550; __rev0_550 = __builtin_shufflevector(__s0_550, __s0_550, 3, 2, 1, 0); \ + int16x8_t __rev1_550; __rev1_550 = __builtin_shufflevector(__s1_550, __s1_550, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_550; __rev2_550 = __builtin_shufflevector(__s2_550, __s2_550, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_550; \ + __ret_550 = __rev0_550 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_550), __noswap_splat_laneq_s16(__rev2_550, __p3_550)); \ + __ret_550 = __builtin_shufflevector(__ret_550, __ret_550, 3, 2, 1, 0); \ + __ret_550; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_laneq_u32(__p0_459, __p1_459, __p2_459, __p3_459) __extension__ ({ \ - uint64x2_t __s0_459 = __p0_459; \ - uint32x2_t __s1_459 = __p1_459; \ - uint32x4_t __s2_459 = __p2_459; \ - uint64x2_t __ret_459; \ - __ret_459 = __s0_459 + vmull_u32(__s1_459, splat_laneq_u32(__s2_459, __p3_459)); \ - __ret_459; \ +#define vmlal_laneq_u32(__p0_551, __p1_551, __p2_551, __p3_551) __extension__ ({ \ + uint64x2_t __s0_551 = __p0_551; \ + uint32x2_t __s1_551 = __p1_551; \ + uint32x4_t __s2_551 = __p2_551; \ + uint64x2_t __ret_551; \ + __ret_551 = __s0_551 + vmull_u32(__s1_551, splat_laneq_u32(__s2_551, __p3_551)); \ + __ret_551; \ }) #else -#define vmlal_laneq_u32(__p0_460, __p1_460, __p2_460, __p3_460) __extension__ ({ \ - uint64x2_t __s0_460 = __p0_460; \ - uint32x2_t __s1_460 = __p1_460; \ - uint32x4_t __s2_460 = __p2_460; \ - uint64x2_t __rev0_460; __rev0_460 = __builtin_shufflevector(__s0_460, __s0_460, 1, 0); \ - uint32x2_t __rev1_460; __rev1_460 = __builtin_shufflevector(__s1_460, __s1_460, 1, 0); \ - uint32x4_t __rev2_460; __rev2_460 = __builtin_shufflevector(__s2_460, __s2_460, 3, 2, 1, 0); \ - uint64x2_t __ret_460; \ - __ret_460 = __rev0_460 + __noswap_vmull_u32(__rev1_460, __noswap_splat_laneq_u32(__rev2_460, __p3_460)); \ - __ret_460 = __builtin_shufflevector(__ret_460, __ret_460, 1, 0); \ - __ret_460; \ +#define vmlal_laneq_u32(__p0_552, __p1_552, __p2_552, __p3_552) __extension__ ({ \ + uint64x2_t __s0_552 = __p0_552; \ + uint32x2_t __s1_552 = __p1_552; \ + uint32x4_t __s2_552 = __p2_552; \ + uint64x2_t __rev0_552; __rev0_552 = __builtin_shufflevector(__s0_552, __s0_552, 1, 0); \ + uint32x2_t __rev1_552; __rev1_552 = __builtin_shufflevector(__s1_552, __s1_552, 1, 0); \ + uint32x4_t __rev2_552; __rev2_552 = __builtin_shufflevector(__s2_552, __s2_552, 3, 2, 1, 0); \ + uint64x2_t __ret_552; \ + __ret_552 = __rev0_552 + __noswap_vmull_u32(__rev1_552, __noswap_splat_laneq_u32(__rev2_552, __p3_552)); \ + __ret_552 = __builtin_shufflevector(__ret_552, __ret_552, 1, 0); \ + __ret_552; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_laneq_u16(__p0_461, __p1_461, __p2_461, __p3_461) __extension__ ({ \ - uint32x4_t __s0_461 = __p0_461; \ - uint16x4_t __s1_461 = __p1_461; \ - uint16x8_t __s2_461 = __p2_461; \ - uint32x4_t __ret_461; \ - __ret_461 = __s0_461 + vmull_u16(__s1_461, splat_laneq_u16(__s2_461, __p3_461)); \ - __ret_461; \ +#define vmlal_laneq_u16(__p0_553, __p1_553, __p2_553, __p3_553) __extension__ ({ \ + uint32x4_t __s0_553 = __p0_553; \ + uint16x4_t __s1_553 = __p1_553; \ + uint16x8_t __s2_553 = __p2_553; \ + uint32x4_t __ret_553; \ + __ret_553 = __s0_553 + vmull_u16(__s1_553, splat_laneq_u16(__s2_553, __p3_553)); \ + __ret_553; \ }) #else -#define vmlal_laneq_u16(__p0_462, __p1_462, __p2_462, __p3_462) __extension__ ({ \ - uint32x4_t __s0_462 = __p0_462; \ - uint16x4_t __s1_462 = __p1_462; \ - uint16x8_t __s2_462 = __p2_462; \ - uint32x4_t __rev0_462; __rev0_462 = __builtin_shufflevector(__s0_462, __s0_462, 3, 2, 1, 0); \ - uint16x4_t __rev1_462; __rev1_462 = __builtin_shufflevector(__s1_462, __s1_462, 3, 2, 1, 0); \ - uint16x8_t __rev2_462; __rev2_462 = __builtin_shufflevector(__s2_462, __s2_462, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_462; \ - __ret_462 = __rev0_462 + __noswap_vmull_u16(__rev1_462, __noswap_splat_laneq_u16(__rev2_462, __p3_462)); \ - __ret_462 = __builtin_shufflevector(__ret_462, __ret_462, 3, 2, 1, 0); \ - __ret_462; \ +#define vmlal_laneq_u16(__p0_554, __p1_554, __p2_554, __p3_554) __extension__ ({ \ + uint32x4_t __s0_554 = __p0_554; \ + uint16x4_t __s1_554 = __p1_554; \ + uint16x8_t __s2_554 = __p2_554; \ + uint32x4_t __rev0_554; __rev0_554 = __builtin_shufflevector(__s0_554, __s0_554, 3, 2, 1, 0); \ + uint16x4_t __rev1_554; __rev1_554 = __builtin_shufflevector(__s1_554, __s1_554, 3, 2, 1, 0); \ + uint16x8_t __rev2_554; __rev2_554 = __builtin_shufflevector(__s2_554, __s2_554, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_554; \ + __ret_554 = __rev0_554 + __noswap_vmull_u16(__rev1_554, __noswap_splat_laneq_u16(__rev2_554, __p3_554)); \ + __ret_554 = __builtin_shufflevector(__ret_554, __ret_554, 3, 2, 1, 0); \ + __ret_554; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_laneq_s32(__p0_463, __p1_463, __p2_463, __p3_463) __extension__ ({ \ - int64x2_t __s0_463 = __p0_463; \ - int32x2_t __s1_463 = __p1_463; \ - int32x4_t __s2_463 = __p2_463; \ - int64x2_t __ret_463; \ - __ret_463 = __s0_463 + vmull_s32(__s1_463, splat_laneq_s32(__s2_463, __p3_463)); \ - __ret_463; \ +#define vmlal_laneq_s32(__p0_555, __p1_555, __p2_555, __p3_555) __extension__ ({ \ + int64x2_t __s0_555 = __p0_555; \ + int32x2_t __s1_555 = __p1_555; \ + int32x4_t __s2_555 = __p2_555; \ + int64x2_t __ret_555; \ + __ret_555 = __s0_555 + vmull_s32(__s1_555, splat_laneq_s32(__s2_555, __p3_555)); \ + __ret_555; \ }) #else -#define vmlal_laneq_s32(__p0_464, __p1_464, __p2_464, __p3_464) __extension__ ({ \ - int64x2_t __s0_464 = __p0_464; \ - int32x2_t __s1_464 = __p1_464; \ - int32x4_t __s2_464 = __p2_464; \ - int64x2_t __rev0_464; __rev0_464 = __builtin_shufflevector(__s0_464, __s0_464, 1, 0); \ - int32x2_t __rev1_464; __rev1_464 = __builtin_shufflevector(__s1_464, __s1_464, 1, 0); \ - int32x4_t __rev2_464; __rev2_464 = __builtin_shufflevector(__s2_464, __s2_464, 3, 2, 1, 0); \ - int64x2_t __ret_464; \ - __ret_464 = __rev0_464 + __noswap_vmull_s32(__rev1_464, __noswap_splat_laneq_s32(__rev2_464, __p3_464)); \ - __ret_464 = __builtin_shufflevector(__ret_464, __ret_464, 1, 0); \ - __ret_464; \ +#define vmlal_laneq_s32(__p0_556, __p1_556, __p2_556, __p3_556) __extension__ ({ \ + int64x2_t __s0_556 = __p0_556; \ + int32x2_t __s1_556 = __p1_556; \ + int32x4_t __s2_556 = __p2_556; \ + int64x2_t __rev0_556; __rev0_556 = __builtin_shufflevector(__s0_556, __s0_556, 1, 0); \ + int32x2_t __rev1_556; __rev1_556 = __builtin_shufflevector(__s1_556, __s1_556, 1, 0); \ + int32x4_t __rev2_556; __rev2_556 = __builtin_shufflevector(__s2_556, __s2_556, 3, 2, 1, 0); \ + int64x2_t __ret_556; \ + __ret_556 = __rev0_556 + __noswap_vmull_s32(__rev1_556, __noswap_splat_laneq_s32(__rev2_556, __p3_556)); \ + __ret_556 = __builtin_shufflevector(__ret_556, __ret_556, 1, 0); \ + __ret_556; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_laneq_s16(__p0_465, __p1_465, __p2_465, __p3_465) __extension__ ({ \ - int32x4_t __s0_465 = __p0_465; \ - int16x4_t __s1_465 = __p1_465; \ - int16x8_t __s2_465 = __p2_465; \ - int32x4_t __ret_465; \ - __ret_465 = __s0_465 + vmull_s16(__s1_465, splat_laneq_s16(__s2_465, __p3_465)); \ - __ret_465; \ +#define vmlal_laneq_s16(__p0_557, __p1_557, __p2_557, __p3_557) __extension__ ({ \ + int32x4_t __s0_557 = __p0_557; \ + int16x4_t __s1_557 = __p1_557; \ + int16x8_t __s2_557 = __p2_557; \ + int32x4_t __ret_557; \ + __ret_557 = __s0_557 + vmull_s16(__s1_557, splat_laneq_s16(__s2_557, __p3_557)); \ + __ret_557; \ }) #else -#define vmlal_laneq_s16(__p0_466, __p1_466, __p2_466, __p3_466) __extension__ ({ \ - int32x4_t __s0_466 = __p0_466; \ - int16x4_t __s1_466 = __p1_466; \ - int16x8_t __s2_466 = __p2_466; \ - int32x4_t __rev0_466; __rev0_466 = __builtin_shufflevector(__s0_466, __s0_466, 3, 2, 1, 0); \ - int16x4_t __rev1_466; __rev1_466 = __builtin_shufflevector(__s1_466, __s1_466, 3, 2, 1, 0); \ - int16x8_t __rev2_466; __rev2_466 = __builtin_shufflevector(__s2_466, __s2_466, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_466; \ - __ret_466 = __rev0_466 + __noswap_vmull_s16(__rev1_466, __noswap_splat_laneq_s16(__rev2_466, __p3_466)); \ - __ret_466 = __builtin_shufflevector(__ret_466, __ret_466, 3, 2, 1, 0); \ - __ret_466; \ +#define vmlal_laneq_s16(__p0_558, __p1_558, __p2_558, __p3_558) __extension__ ({ \ + int32x4_t __s0_558 = __p0_558; \ + int16x4_t __s1_558 = __p1_558; \ + int16x8_t __s2_558 = __p2_558; \ + int32x4_t __rev0_558; __rev0_558 = __builtin_shufflevector(__s0_558, __s0_558, 3, 2, 1, 0); \ + int16x4_t __rev1_558; __rev1_558 = __builtin_shufflevector(__s1_558, __s1_558, 3, 2, 1, 0); \ + int16x8_t __rev2_558; __rev2_558 = __builtin_shufflevector(__s2_558, __s2_558, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_558; \ + __ret_558 = __rev0_558 + __noswap_vmull_s16(__rev1_558, __noswap_splat_laneq_s16(__rev2_558, __p3_558)); \ + __ret_558 = __builtin_shufflevector(__ret_558, __ret_558, 3, 2, 1, 0); \ + __ret_558; \ }) #endif @@ -53116,547 +54803,530 @@ __ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) return __ret; } #ifdef __LITTLE_ENDIAN__ -#define vmlsq_laneq_u32(__p0_467, __p1_467, __p2_467, __p3_467) __extension__ ({ \ - uint32x4_t __s0_467 = __p0_467; \ - uint32x4_t __s1_467 = __p1_467; \ - uint32x4_t __s2_467 = __p2_467; \ - uint32x4_t __ret_467; \ - __ret_467 = __s0_467 - __s1_467 * splatq_laneq_u32(__s2_467, __p3_467); \ - __ret_467; \ +#define vmlsq_laneq_u32(__p0_559, __p1_559, __p2_559, __p3_559) __extension__ ({ \ + uint32x4_t __s0_559 = __p0_559; \ + uint32x4_t __s1_559 = __p1_559; \ + uint32x4_t __s2_559 = __p2_559; \ + uint32x4_t __ret_559; \ + __ret_559 = __s0_559 - __s1_559 * splatq_laneq_u32(__s2_559, __p3_559); \ + __ret_559; \ }) #else -#define vmlsq_laneq_u32(__p0_468, __p1_468, __p2_468, __p3_468) __extension__ ({ \ - uint32x4_t __s0_468 = __p0_468; \ - uint32x4_t __s1_468 = __p1_468; \ - uint32x4_t __s2_468 = __p2_468; \ - uint32x4_t __rev0_468; __rev0_468 = __builtin_shufflevector(__s0_468, __s0_468, 3, 2, 1, 0); \ - uint32x4_t __rev1_468; __rev1_468 = __builtin_shufflevector(__s1_468, __s1_468, 3, 2, 1, 0); \ - uint32x4_t __rev2_468; __rev2_468 = __builtin_shufflevector(__s2_468, __s2_468, 3, 2, 1, 0); \ - uint32x4_t __ret_468; \ - __ret_468 = __rev0_468 - __rev1_468 * __noswap_splatq_laneq_u32(__rev2_468, __p3_468); \ - __ret_468 = __builtin_shufflevector(__ret_468, __ret_468, 3, 2, 1, 0); \ - __ret_468; \ +#define vmlsq_laneq_u32(__p0_560, __p1_560, __p2_560, __p3_560) __extension__ ({ \ + uint32x4_t __s0_560 = __p0_560; \ + uint32x4_t __s1_560 = __p1_560; \ + uint32x4_t __s2_560 = __p2_560; \ + uint32x4_t __rev0_560; __rev0_560 = __builtin_shufflevector(__s0_560, __s0_560, 3, 2, 1, 0); \ + uint32x4_t __rev1_560; __rev1_560 = __builtin_shufflevector(__s1_560, __s1_560, 3, 2, 1, 0); \ + uint32x4_t __rev2_560; __rev2_560 = __builtin_shufflevector(__s2_560, __s2_560, 3, 2, 1, 0); \ + uint32x4_t __ret_560; \ + __ret_560 = __rev0_560 - __rev1_560 * __noswap_splatq_laneq_u32(__rev2_560, __p3_560); \ + __ret_560 = __builtin_shufflevector(__ret_560, __ret_560, 3, 2, 1, 0); \ + __ret_560; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsq_laneq_u16(__p0_469, __p1_469, __p2_469, __p3_469) __extension__ ({ \ - uint16x8_t __s0_469 = __p0_469; \ - uint16x8_t __s1_469 = __p1_469; \ - uint16x8_t __s2_469 = __p2_469; \ - uint16x8_t __ret_469; \ - __ret_469 = __s0_469 - __s1_469 * splatq_laneq_u16(__s2_469, __p3_469); \ - __ret_469; \ +#define vmlsq_laneq_u16(__p0_561, __p1_561, __p2_561, __p3_561) __extension__ ({ \ + uint16x8_t __s0_561 = __p0_561; \ + uint16x8_t __s1_561 = __p1_561; \ + uint16x8_t __s2_561 = __p2_561; \ + uint16x8_t __ret_561; \ + __ret_561 = __s0_561 - __s1_561 * splatq_laneq_u16(__s2_561, __p3_561); \ + __ret_561; \ }) #else -#define vmlsq_laneq_u16(__p0_470, __p1_470, __p2_470, __p3_470) __extension__ ({ \ - uint16x8_t __s0_470 = __p0_470; \ - uint16x8_t __s1_470 = __p1_470; \ - uint16x8_t __s2_470 = __p2_470; \ - uint16x8_t __rev0_470; __rev0_470 = __builtin_shufflevector(__s0_470, __s0_470, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_470; __rev1_470 = __builtin_shufflevector(__s1_470, __s1_470, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev2_470; __rev2_470 = __builtin_shufflevector(__s2_470, __s2_470, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __ret_470; \ - __ret_470 = __rev0_470 - __rev1_470 * __noswap_splatq_laneq_u16(__rev2_470, __p3_470); \ - __ret_470 = __builtin_shufflevector(__ret_470, __ret_470, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_470; \ +#define vmlsq_laneq_u16(__p0_562, __p1_562, __p2_562, __p3_562) __extension__ ({ \ + uint16x8_t __s0_562 = __p0_562; \ + uint16x8_t __s1_562 = __p1_562; \ + uint16x8_t __s2_562 = __p2_562; \ + uint16x8_t __rev0_562; __rev0_562 = __builtin_shufflevector(__s0_562, __s0_562, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_562; __rev1_562 = __builtin_shufflevector(__s1_562, __s1_562, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev2_562; __rev2_562 = __builtin_shufflevector(__s2_562, __s2_562, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __ret_562; \ + __ret_562 = __rev0_562 - __rev1_562 * __noswap_splatq_laneq_u16(__rev2_562, __p3_562); \ + __ret_562 = __builtin_shufflevector(__ret_562, __ret_562, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_562; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsq_laneq_f32(__p0_471, __p1_471, __p2_471, __p3_471) __extension__ ({ \ - float32x4_t __s0_471 = __p0_471; \ - float32x4_t __s1_471 = __p1_471; \ - float32x4_t __s2_471 = __p2_471; \ - float32x4_t __ret_471; \ - __ret_471 = __s0_471 - __s1_471 * splatq_laneq_f32(__s2_471, __p3_471); \ - __ret_471; \ +#define vmlsq_laneq_f32(__p0_563, __p1_563, __p2_563, __p3_563) __extension__ ({ \ + float32x4_t __s0_563 = __p0_563; \ + float32x4_t __s1_563 = __p1_563; \ + float32x4_t __s2_563 = __p2_563; \ + float32x4_t __ret_563; \ + __ret_563 = __s0_563 - __s1_563 * splatq_laneq_f32(__s2_563, __p3_563); \ + __ret_563; \ }) #else -#define vmlsq_laneq_f32(__p0_472, __p1_472, __p2_472, __p3_472) __extension__ ({ \ - float32x4_t __s0_472 = __p0_472; \ - float32x4_t __s1_472 = __p1_472; \ - float32x4_t __s2_472 = __p2_472; \ - float32x4_t __rev0_472; __rev0_472 = __builtin_shufflevector(__s0_472, __s0_472, 3, 2, 1, 0); \ - float32x4_t __rev1_472; __rev1_472 = __builtin_shufflevector(__s1_472, __s1_472, 3, 2, 1, 0); \ - float32x4_t __rev2_472; __rev2_472 = __builtin_shufflevector(__s2_472, __s2_472, 3, 2, 1, 0); \ - float32x4_t __ret_472; \ - __ret_472 = __rev0_472 - __rev1_472 * __noswap_splatq_laneq_f32(__rev2_472, __p3_472); \ - __ret_472 = __builtin_shufflevector(__ret_472, __ret_472, 3, 2, 1, 0); \ - __ret_472; \ +#define vmlsq_laneq_f32(__p0_564, __p1_564, __p2_564, __p3_564) __extension__ ({ \ + float32x4_t __s0_564 = __p0_564; \ + float32x4_t __s1_564 = __p1_564; \ + float32x4_t __s2_564 = __p2_564; \ + float32x4_t __rev0_564; __rev0_564 = __builtin_shufflevector(__s0_564, __s0_564, 3, 2, 1, 0); \ + float32x4_t __rev1_564; __rev1_564 = __builtin_shufflevector(__s1_564, __s1_564, 3, 2, 1, 0); \ + float32x4_t __rev2_564; __rev2_564 = __builtin_shufflevector(__s2_564, __s2_564, 3, 2, 1, 0); \ + float32x4_t __ret_564; \ + __ret_564 = __rev0_564 - __rev1_564 * __noswap_splatq_laneq_f32(__rev2_564, __p3_564); \ + __ret_564 = __builtin_shufflevector(__ret_564, __ret_564, 3, 2, 1, 0); \ + __ret_564; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsq_laneq_s32(__p0_473, __p1_473, __p2_473, __p3_473) __extension__ ({ \ - int32x4_t __s0_473 = __p0_473; \ - int32x4_t __s1_473 = __p1_473; \ - int32x4_t __s2_473 = __p2_473; \ - int32x4_t __ret_473; \ - __ret_473 = __s0_473 - __s1_473 * splatq_laneq_s32(__s2_473, __p3_473); \ - __ret_473; \ +#define vmlsq_laneq_s32(__p0_565, __p1_565, __p2_565, __p3_565) __extension__ ({ \ + int32x4_t __s0_565 = __p0_565; \ + int32x4_t __s1_565 = __p1_565; \ + int32x4_t __s2_565 = __p2_565; \ + int32x4_t __ret_565; \ + __ret_565 = __s0_565 - __s1_565 * splatq_laneq_s32(__s2_565, __p3_565); \ + __ret_565; \ }) #else -#define vmlsq_laneq_s32(__p0_474, __p1_474, __p2_474, __p3_474) __extension__ ({ \ - int32x4_t __s0_474 = __p0_474; \ - int32x4_t __s1_474 = __p1_474; \ - int32x4_t __s2_474 = __p2_474; \ - int32x4_t __rev0_474; __rev0_474 = __builtin_shufflevector(__s0_474, __s0_474, 3, 2, 1, 0); \ - int32x4_t __rev1_474; __rev1_474 = __builtin_shufflevector(__s1_474, __s1_474, 3, 2, 1, 0); \ - int32x4_t __rev2_474; __rev2_474 = __builtin_shufflevector(__s2_474, __s2_474, 3, 2, 1, 0); \ - int32x4_t __ret_474; \ - __ret_474 = __rev0_474 - __rev1_474 * __noswap_splatq_laneq_s32(__rev2_474, __p3_474); \ - __ret_474 = __builtin_shufflevector(__ret_474, __ret_474, 3, 2, 1, 0); \ - __ret_474; \ +#define vmlsq_laneq_s32(__p0_566, __p1_566, __p2_566, __p3_566) __extension__ ({ \ + int32x4_t __s0_566 = __p0_566; \ + int32x4_t __s1_566 = __p1_566; \ + int32x4_t __s2_566 = __p2_566; \ + int32x4_t __rev0_566; __rev0_566 = __builtin_shufflevector(__s0_566, __s0_566, 3, 2, 1, 0); \ + int32x4_t __rev1_566; __rev1_566 = __builtin_shufflevector(__s1_566, __s1_566, 3, 2, 1, 0); \ + int32x4_t __rev2_566; __rev2_566 = __builtin_shufflevector(__s2_566, __s2_566, 3, 2, 1, 0); \ + int32x4_t __ret_566; \ + __ret_566 = __rev0_566 - __rev1_566 * __noswap_splatq_laneq_s32(__rev2_566, __p3_566); \ + __ret_566 = __builtin_shufflevector(__ret_566, __ret_566, 3, 2, 1, 0); \ + __ret_566; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsq_laneq_s16(__p0_475, __p1_475, __p2_475, __p3_475) __extension__ ({ \ - int16x8_t __s0_475 = __p0_475; \ - int16x8_t __s1_475 = __p1_475; \ - int16x8_t __s2_475 = __p2_475; \ - int16x8_t __ret_475; \ - __ret_475 = __s0_475 - __s1_475 * splatq_laneq_s16(__s2_475, __p3_475); \ - __ret_475; \ +#define vmlsq_laneq_s16(__p0_567, __p1_567, __p2_567, __p3_567) __extension__ ({ \ + int16x8_t __s0_567 = __p0_567; \ + int16x8_t __s1_567 = __p1_567; \ + int16x8_t __s2_567 = __p2_567; \ + int16x8_t __ret_567; \ + __ret_567 = __s0_567 - __s1_567 * splatq_laneq_s16(__s2_567, __p3_567); \ + __ret_567; \ }) #else -#define vmlsq_laneq_s16(__p0_476, __p1_476, __p2_476, __p3_476) __extension__ ({ \ - int16x8_t __s0_476 = __p0_476; \ - int16x8_t __s1_476 = __p1_476; \ - int16x8_t __s2_476 = __p2_476; \ - int16x8_t __rev0_476; __rev0_476 = __builtin_shufflevector(__s0_476, __s0_476, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_476; __rev1_476 = __builtin_shufflevector(__s1_476, __s1_476, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_476; __rev2_476 = __builtin_shufflevector(__s2_476, __s2_476, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_476; \ - __ret_476 = __rev0_476 - __rev1_476 * __noswap_splatq_laneq_s16(__rev2_476, __p3_476); \ - __ret_476 = __builtin_shufflevector(__ret_476, __ret_476, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_476; \ +#define vmlsq_laneq_s16(__p0_568, __p1_568, __p2_568, __p3_568) __extension__ ({ \ + int16x8_t __s0_568 = __p0_568; \ + int16x8_t __s1_568 = __p1_568; \ + int16x8_t __s2_568 = __p2_568; \ + int16x8_t __rev0_568; __rev0_568 = __builtin_shufflevector(__s0_568, __s0_568, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_568; __rev1_568 = __builtin_shufflevector(__s1_568, __s1_568, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_568; __rev2_568 = __builtin_shufflevector(__s2_568, __s2_568, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_568; \ + __ret_568 = __rev0_568 - __rev1_568 * __noswap_splatq_laneq_s16(__rev2_568, __p3_568); \ + __ret_568 = __builtin_shufflevector(__ret_568, __ret_568, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_568; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmls_laneq_u32(__p0_477, __p1_477, __p2_477, __p3_477) __extension__ ({ \ - uint32x2_t __s0_477 = __p0_477; \ - uint32x2_t __s1_477 = __p1_477; \ - uint32x4_t __s2_477 = __p2_477; \ - uint32x2_t __ret_477; \ - __ret_477 = __s0_477 - __s1_477 * splat_laneq_u32(__s2_477, __p3_477); \ - __ret_477; \ +#define vmls_laneq_u32(__p0_569, __p1_569, __p2_569, __p3_569) __extension__ ({ \ + uint32x2_t __s0_569 = __p0_569; \ + uint32x2_t __s1_569 = __p1_569; \ + uint32x4_t __s2_569 = __p2_569; \ + uint32x2_t __ret_569; \ + __ret_569 = __s0_569 - __s1_569 * splat_laneq_u32(__s2_569, __p3_569); \ + __ret_569; \ }) #else -#define vmls_laneq_u32(__p0_478, __p1_478, __p2_478, __p3_478) __extension__ ({ \ - uint32x2_t __s0_478 = __p0_478; \ - uint32x2_t __s1_478 = __p1_478; \ - uint32x4_t __s2_478 = __p2_478; \ - uint32x2_t __rev0_478; __rev0_478 = __builtin_shufflevector(__s0_478, __s0_478, 1, 0); \ - uint32x2_t __rev1_478; __rev1_478 = __builtin_shufflevector(__s1_478, __s1_478, 1, 0); \ - uint32x4_t __rev2_478; __rev2_478 = __builtin_shufflevector(__s2_478, __s2_478, 3, 2, 1, 0); \ - uint32x2_t __ret_478; \ - __ret_478 = __rev0_478 - __rev1_478 * __noswap_splat_laneq_u32(__rev2_478, __p3_478); \ - __ret_478 = __builtin_shufflevector(__ret_478, __ret_478, 1, 0); \ - __ret_478; \ +#define vmls_laneq_u32(__p0_570, __p1_570, __p2_570, __p3_570) __extension__ ({ \ + uint32x2_t __s0_570 = __p0_570; \ + uint32x2_t __s1_570 = __p1_570; \ + uint32x4_t __s2_570 = __p2_570; \ + uint32x2_t __rev0_570; __rev0_570 = __builtin_shufflevector(__s0_570, __s0_570, 1, 0); \ + uint32x2_t __rev1_570; __rev1_570 = __builtin_shufflevector(__s1_570, __s1_570, 1, 0); \ + uint32x4_t __rev2_570; __rev2_570 = __builtin_shufflevector(__s2_570, __s2_570, 3, 2, 1, 0); \ + uint32x2_t __ret_570; \ + __ret_570 = __rev0_570 - __rev1_570 * __noswap_splat_laneq_u32(__rev2_570, __p3_570); \ + __ret_570 = __builtin_shufflevector(__ret_570, __ret_570, 1, 0); \ + __ret_570; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmls_laneq_u16(__p0_479, __p1_479, __p2_479, __p3_479) __extension__ ({ \ - uint16x4_t __s0_479 = __p0_479; \ - uint16x4_t __s1_479 = __p1_479; \ - uint16x8_t __s2_479 = __p2_479; \ - uint16x4_t __ret_479; \ - __ret_479 = __s0_479 - __s1_479 * splat_laneq_u16(__s2_479, __p3_479); \ - __ret_479; \ +#define vmls_laneq_u16(__p0_571, __p1_571, __p2_571, __p3_571) __extension__ ({ \ + uint16x4_t __s0_571 = __p0_571; \ + uint16x4_t __s1_571 = __p1_571; \ + uint16x8_t __s2_571 = __p2_571; \ + uint16x4_t __ret_571; \ + __ret_571 = __s0_571 - __s1_571 * splat_laneq_u16(__s2_571, __p3_571); \ + __ret_571; \ }) #else -#define vmls_laneq_u16(__p0_480, __p1_480, __p2_480, __p3_480) __extension__ ({ \ - uint16x4_t __s0_480 = __p0_480; \ - uint16x4_t __s1_480 = __p1_480; \ - uint16x8_t __s2_480 = __p2_480; \ - uint16x4_t __rev0_480; __rev0_480 = __builtin_shufflevector(__s0_480, __s0_480, 3, 2, 1, 0); \ - uint16x4_t __rev1_480; __rev1_480 = __builtin_shufflevector(__s1_480, __s1_480, 3, 2, 1, 0); \ - uint16x8_t __rev2_480; __rev2_480 = __builtin_shufflevector(__s2_480, __s2_480, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __ret_480; \ - __ret_480 = __rev0_480 - __rev1_480 * __noswap_splat_laneq_u16(__rev2_480, __p3_480); \ - __ret_480 = __builtin_shufflevector(__ret_480, __ret_480, 3, 2, 1, 0); \ - __ret_480; \ +#define vmls_laneq_u16(__p0_572, __p1_572, __p2_572, __p3_572) __extension__ ({ \ + uint16x4_t __s0_572 = __p0_572; \ + uint16x4_t __s1_572 = __p1_572; \ + uint16x8_t __s2_572 = __p2_572; \ + uint16x4_t __rev0_572; __rev0_572 = __builtin_shufflevector(__s0_572, __s0_572, 3, 2, 1, 0); \ + uint16x4_t __rev1_572; __rev1_572 = __builtin_shufflevector(__s1_572, __s1_572, 3, 2, 1, 0); \ + uint16x8_t __rev2_572; __rev2_572 = __builtin_shufflevector(__s2_572, __s2_572, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __ret_572; \ + __ret_572 = __rev0_572 - __rev1_572 * __noswap_splat_laneq_u16(__rev2_572, __p3_572); \ + __ret_572 = __builtin_shufflevector(__ret_572, __ret_572, 3, 2, 1, 0); \ + __ret_572; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmls_laneq_f32(__p0_481, __p1_481, __p2_481, __p3_481) __extension__ ({ \ - float32x2_t __s0_481 = __p0_481; \ - float32x2_t __s1_481 = __p1_481; \ - float32x4_t __s2_481 = __p2_481; \ - float32x2_t __ret_481; \ - __ret_481 = __s0_481 - __s1_481 * splat_laneq_f32(__s2_481, __p3_481); \ - __ret_481; \ +#define vmls_laneq_f32(__p0_573, __p1_573, __p2_573, __p3_573) __extension__ ({ \ + float32x2_t __s0_573 = __p0_573; \ + float32x2_t __s1_573 = __p1_573; \ + float32x4_t __s2_573 = __p2_573; \ + float32x2_t __ret_573; \ + __ret_573 = __s0_573 - __s1_573 * splat_laneq_f32(__s2_573, __p3_573); \ + __ret_573; \ }) #else -#define vmls_laneq_f32(__p0_482, __p1_482, __p2_482, __p3_482) __extension__ ({ \ - float32x2_t __s0_482 = __p0_482; \ - float32x2_t __s1_482 = __p1_482; \ - float32x4_t __s2_482 = __p2_482; \ - float32x2_t __rev0_482; __rev0_482 = __builtin_shufflevector(__s0_482, __s0_482, 1, 0); \ - float32x2_t __rev1_482; __rev1_482 = __builtin_shufflevector(__s1_482, __s1_482, 1, 0); \ - float32x4_t __rev2_482; __rev2_482 = __builtin_shufflevector(__s2_482, __s2_482, 3, 2, 1, 0); \ - float32x2_t __ret_482; \ - __ret_482 = __rev0_482 - __rev1_482 * __noswap_splat_laneq_f32(__rev2_482, __p3_482); \ - __ret_482 = __builtin_shufflevector(__ret_482, __ret_482, 1, 0); \ - __ret_482; \ +#define vmls_laneq_f32(__p0_574, __p1_574, __p2_574, __p3_574) __extension__ ({ \ + float32x2_t __s0_574 = __p0_574; \ + float32x2_t __s1_574 = __p1_574; \ + float32x4_t __s2_574 = __p2_574; \ + float32x2_t __rev0_574; __rev0_574 = __builtin_shufflevector(__s0_574, __s0_574, 1, 0); \ + float32x2_t __rev1_574; __rev1_574 = __builtin_shufflevector(__s1_574, __s1_574, 1, 0); \ + float32x4_t __rev2_574; __rev2_574 = __builtin_shufflevector(__s2_574, __s2_574, 3, 2, 1, 0); \ + float32x2_t __ret_574; \ + __ret_574 = __rev0_574 - __rev1_574 * __noswap_splat_laneq_f32(__rev2_574, __p3_574); \ + __ret_574 = __builtin_shufflevector(__ret_574, __ret_574, 1, 0); \ + __ret_574; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmls_laneq_s32(__p0_483, __p1_483, __p2_483, __p3_483) __extension__ ({ \ - int32x2_t __s0_483 = __p0_483; \ - int32x2_t __s1_483 = __p1_483; \ - int32x4_t __s2_483 = __p2_483; \ - int32x2_t __ret_483; \ - __ret_483 = __s0_483 - __s1_483 * splat_laneq_s32(__s2_483, __p3_483); \ - __ret_483; \ +#define vmls_laneq_s32(__p0_575, __p1_575, __p2_575, __p3_575) __extension__ ({ \ + int32x2_t __s0_575 = __p0_575; \ + int32x2_t __s1_575 = __p1_575; \ + int32x4_t __s2_575 = __p2_575; \ + int32x2_t __ret_575; \ + __ret_575 = __s0_575 - __s1_575 * splat_laneq_s32(__s2_575, __p3_575); \ + __ret_575; \ }) #else -#define vmls_laneq_s32(__p0_484, __p1_484, __p2_484, __p3_484) __extension__ ({ \ - int32x2_t __s0_484 = __p0_484; \ - int32x2_t __s1_484 = __p1_484; \ - int32x4_t __s2_484 = __p2_484; \ - int32x2_t __rev0_484; __rev0_484 = __builtin_shufflevector(__s0_484, __s0_484, 1, 0); \ - int32x2_t __rev1_484; __rev1_484 = __builtin_shufflevector(__s1_484, __s1_484, 1, 0); \ - int32x4_t __rev2_484; __rev2_484 = __builtin_shufflevector(__s2_484, __s2_484, 3, 2, 1, 0); \ - int32x2_t __ret_484; \ - __ret_484 = __rev0_484 - __rev1_484 * __noswap_splat_laneq_s32(__rev2_484, __p3_484); \ - __ret_484 = __builtin_shufflevector(__ret_484, __ret_484, 1, 0); \ - __ret_484; \ +#define vmls_laneq_s32(__p0_576, __p1_576, __p2_576, __p3_576) __extension__ ({ \ + int32x2_t __s0_576 = __p0_576; \ + int32x2_t __s1_576 = __p1_576; \ + int32x4_t __s2_576 = __p2_576; \ + int32x2_t __rev0_576; __rev0_576 = __builtin_shufflevector(__s0_576, __s0_576, 1, 0); \ + int32x2_t __rev1_576; __rev1_576 = __builtin_shufflevector(__s1_576, __s1_576, 1, 0); \ + int32x4_t __rev2_576; __rev2_576 = __builtin_shufflevector(__s2_576, __s2_576, 3, 2, 1, 0); \ + int32x2_t __ret_576; \ + __ret_576 = __rev0_576 - __rev1_576 * __noswap_splat_laneq_s32(__rev2_576, __p3_576); \ + __ret_576 = __builtin_shufflevector(__ret_576, __ret_576, 1, 0); \ + __ret_576; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmls_laneq_s16(__p0_485, __p1_485, __p2_485, __p3_485) __extension__ ({ \ - int16x4_t __s0_485 = __p0_485; \ - int16x4_t __s1_485 = __p1_485; \ - int16x8_t __s2_485 = __p2_485; \ - int16x4_t __ret_485; \ - __ret_485 = __s0_485 - __s1_485 * splat_laneq_s16(__s2_485, __p3_485); \ - __ret_485; \ +#define vmls_laneq_s16(__p0_577, __p1_577, __p2_577, __p3_577) __extension__ ({ \ + int16x4_t __s0_577 = __p0_577; \ + int16x4_t __s1_577 = __p1_577; \ + int16x8_t __s2_577 = __p2_577; \ + int16x4_t __ret_577; \ + __ret_577 = __s0_577 - __s1_577 * splat_laneq_s16(__s2_577, __p3_577); \ + __ret_577; \ }) #else -#define vmls_laneq_s16(__p0_486, __p1_486, __p2_486, __p3_486) __extension__ ({ \ - int16x4_t __s0_486 = __p0_486; \ - int16x4_t __s1_486 = __p1_486; \ - int16x8_t __s2_486 = __p2_486; \ - int16x4_t __rev0_486; __rev0_486 = __builtin_shufflevector(__s0_486, __s0_486, 3, 2, 1, 0); \ - int16x4_t __rev1_486; __rev1_486 = __builtin_shufflevector(__s1_486, __s1_486, 3, 2, 1, 0); \ - int16x8_t __rev2_486; __rev2_486 = __builtin_shufflevector(__s2_486, __s2_486, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __ret_486; \ - __ret_486 = __rev0_486 - __rev1_486 * __noswap_splat_laneq_s16(__rev2_486, __p3_486); \ - __ret_486 = __builtin_shufflevector(__ret_486, __ret_486, 3, 2, 1, 0); \ - __ret_486; \ +#define vmls_laneq_s16(__p0_578, __p1_578, __p2_578, __p3_578) __extension__ ({ \ + int16x4_t __s0_578 = __p0_578; \ + int16x4_t __s1_578 = __p1_578; \ + int16x8_t __s2_578 = __p2_578; \ + int16x4_t __rev0_578; __rev0_578 = __builtin_shufflevector(__s0_578, __s0_578, 3, 2, 1, 0); \ + int16x4_t __rev1_578; __rev1_578 = __builtin_shufflevector(__s1_578, __s1_578, 3, 2, 1, 0); \ + int16x8_t __rev2_578; __rev2_578 = __builtin_shufflevector(__s2_578, __s2_578, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __ret_578; \ + __ret_578 = __rev0_578 - __rev1_578 * __noswap_splat_laneq_s16(__rev2_578, __p3_578); \ + __ret_578 = __builtin_shufflevector(__ret_578, __ret_578, 3, 2, 1, 0); \ + __ret_578; \ }) #endif #ifdef __LITTLE_ENDIAN__ -__ai float64x2_t vmlsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) { - float64x2_t __ret; - __ret = __p0 - __p1 * (float64x2_t) {__p2, __p2}; - return __ret; -} -#else -__ai float64x2_t vmlsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) { - float64x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); - float64x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); - float64x2_t __ret; - __ret = __rev0 - __rev1 * (float64x2_t) {__p2, __p2}; - __ret = __builtin_shufflevector(__ret, __ret, 1, 0); - return __ret; -} -#endif - -#ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_lane_u32(__p0_487, __p1_487, __p2_487, __p3_487) __extension__ ({ \ - uint64x2_t __s0_487 = __p0_487; \ - uint32x4_t __s1_487 = __p1_487; \ - uint32x2_t __s2_487 = __p2_487; \ - uint64x2_t __ret_487; \ - __ret_487 = __s0_487 - vmull_u32(vget_high_u32(__s1_487), splat_lane_u32(__s2_487, __p3_487)); \ - __ret_487; \ +#define vmlsl_high_lane_u32(__p0_579, __p1_579, __p2_579, __p3_579) __extension__ ({ \ + uint64x2_t __s0_579 = __p0_579; \ + uint32x4_t __s1_579 = __p1_579; \ + uint32x2_t __s2_579 = __p2_579; \ + uint64x2_t __ret_579; \ + __ret_579 = __s0_579 - vmull_u32(vget_high_u32(__s1_579), splat_lane_u32(__s2_579, __p3_579)); \ + __ret_579; \ }) #else -#define vmlsl_high_lane_u32(__p0_488, __p1_488, __p2_488, __p3_488) __extension__ ({ \ - uint64x2_t __s0_488 = __p0_488; \ - uint32x4_t __s1_488 = __p1_488; \ - uint32x2_t __s2_488 = __p2_488; \ - uint64x2_t __rev0_488; __rev0_488 = __builtin_shufflevector(__s0_488, __s0_488, 1, 0); \ - uint32x4_t __rev1_488; __rev1_488 = __builtin_shufflevector(__s1_488, __s1_488, 3, 2, 1, 0); \ - uint32x2_t __rev2_488; __rev2_488 = __builtin_shufflevector(__s2_488, __s2_488, 1, 0); \ - uint64x2_t __ret_488; \ - __ret_488 = __rev0_488 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_488), __noswap_splat_lane_u32(__rev2_488, __p3_488)); \ - __ret_488 = __builtin_shufflevector(__ret_488, __ret_488, 1, 0); \ - __ret_488; \ +#define vmlsl_high_lane_u32(__p0_580, __p1_580, __p2_580, __p3_580) __extension__ ({ \ + uint64x2_t __s0_580 = __p0_580; \ + uint32x4_t __s1_580 = __p1_580; \ + uint32x2_t __s2_580 = __p2_580; \ + uint64x2_t __rev0_580; __rev0_580 = __builtin_shufflevector(__s0_580, __s0_580, 1, 0); \ + uint32x4_t __rev1_580; __rev1_580 = __builtin_shufflevector(__s1_580, __s1_580, 3, 2, 1, 0); \ + uint32x2_t __rev2_580; __rev2_580 = __builtin_shufflevector(__s2_580, __s2_580, 1, 0); \ + uint64x2_t __ret_580; \ + __ret_580 = __rev0_580 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_580), __noswap_splat_lane_u32(__rev2_580, __p3_580)); \ + __ret_580 = __builtin_shufflevector(__ret_580, __ret_580, 1, 0); \ + __ret_580; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_lane_u16(__p0_489, __p1_489, __p2_489, __p3_489) __extension__ ({ \ - uint32x4_t __s0_489 = __p0_489; \ - uint16x8_t __s1_489 = __p1_489; \ - uint16x4_t __s2_489 = __p2_489; \ - uint32x4_t __ret_489; \ - __ret_489 = __s0_489 - vmull_u16(vget_high_u16(__s1_489), splat_lane_u16(__s2_489, __p3_489)); \ - __ret_489; \ +#define vmlsl_high_lane_u16(__p0_581, __p1_581, __p2_581, __p3_581) __extension__ ({ \ + uint32x4_t __s0_581 = __p0_581; \ + uint16x8_t __s1_581 = __p1_581; \ + uint16x4_t __s2_581 = __p2_581; \ + uint32x4_t __ret_581; \ + __ret_581 = __s0_581 - vmull_u16(vget_high_u16(__s1_581), splat_lane_u16(__s2_581, __p3_581)); \ + __ret_581; \ }) #else -#define vmlsl_high_lane_u16(__p0_490, __p1_490, __p2_490, __p3_490) __extension__ ({ \ - uint32x4_t __s0_490 = __p0_490; \ - uint16x8_t __s1_490 = __p1_490; \ - uint16x4_t __s2_490 = __p2_490; \ - uint32x4_t __rev0_490; __rev0_490 = __builtin_shufflevector(__s0_490, __s0_490, 3, 2, 1, 0); \ - uint16x8_t __rev1_490; __rev1_490 = __builtin_shufflevector(__s1_490, __s1_490, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __rev2_490; __rev2_490 = __builtin_shufflevector(__s2_490, __s2_490, 3, 2, 1, 0); \ - uint32x4_t __ret_490; \ - __ret_490 = __rev0_490 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_490), __noswap_splat_lane_u16(__rev2_490, __p3_490)); \ - __ret_490 = __builtin_shufflevector(__ret_490, __ret_490, 3, 2, 1, 0); \ - __ret_490; \ +#define vmlsl_high_lane_u16(__p0_582, __p1_582, __p2_582, __p3_582) __extension__ ({ \ + uint32x4_t __s0_582 = __p0_582; \ + uint16x8_t __s1_582 = __p1_582; \ + uint16x4_t __s2_582 = __p2_582; \ + uint32x4_t __rev0_582; __rev0_582 = __builtin_shufflevector(__s0_582, __s0_582, 3, 2, 1, 0); \ + uint16x8_t __rev1_582; __rev1_582 = __builtin_shufflevector(__s1_582, __s1_582, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __rev2_582; __rev2_582 = __builtin_shufflevector(__s2_582, __s2_582, 3, 2, 1, 0); \ + uint32x4_t __ret_582; \ + __ret_582 = __rev0_582 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_582), __noswap_splat_lane_u16(__rev2_582, __p3_582)); \ + __ret_582 = __builtin_shufflevector(__ret_582, __ret_582, 3, 2, 1, 0); \ + __ret_582; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_lane_s32(__p0_491, __p1_491, __p2_491, __p3_491) __extension__ ({ \ - int64x2_t __s0_491 = __p0_491; \ - int32x4_t __s1_491 = __p1_491; \ - int32x2_t __s2_491 = __p2_491; \ - int64x2_t __ret_491; \ - __ret_491 = __s0_491 - vmull_s32(vget_high_s32(__s1_491), splat_lane_s32(__s2_491, __p3_491)); \ - __ret_491; \ +#define vmlsl_high_lane_s32(__p0_583, __p1_583, __p2_583, __p3_583) __extension__ ({ \ + int64x2_t __s0_583 = __p0_583; \ + int32x4_t __s1_583 = __p1_583; \ + int32x2_t __s2_583 = __p2_583; \ + int64x2_t __ret_583; \ + __ret_583 = __s0_583 - vmull_s32(vget_high_s32(__s1_583), splat_lane_s32(__s2_583, __p3_583)); \ + __ret_583; \ }) #else -#define vmlsl_high_lane_s32(__p0_492, __p1_492, __p2_492, __p3_492) __extension__ ({ \ - int64x2_t __s0_492 = __p0_492; \ - int32x4_t __s1_492 = __p1_492; \ - int32x2_t __s2_492 = __p2_492; \ - int64x2_t __rev0_492; __rev0_492 = __builtin_shufflevector(__s0_492, __s0_492, 1, 0); \ - int32x4_t __rev1_492; __rev1_492 = __builtin_shufflevector(__s1_492, __s1_492, 3, 2, 1, 0); \ - int32x2_t __rev2_492; __rev2_492 = __builtin_shufflevector(__s2_492, __s2_492, 1, 0); \ - int64x2_t __ret_492; \ - __ret_492 = __rev0_492 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_492), __noswap_splat_lane_s32(__rev2_492, __p3_492)); \ - __ret_492 = __builtin_shufflevector(__ret_492, __ret_492, 1, 0); \ - __ret_492; \ +#define vmlsl_high_lane_s32(__p0_584, __p1_584, __p2_584, __p3_584) __extension__ ({ \ + int64x2_t __s0_584 = __p0_584; \ + int32x4_t __s1_584 = __p1_584; \ + int32x2_t __s2_584 = __p2_584; \ + int64x2_t __rev0_584; __rev0_584 = __builtin_shufflevector(__s0_584, __s0_584, 1, 0); \ + int32x4_t __rev1_584; __rev1_584 = __builtin_shufflevector(__s1_584, __s1_584, 3, 2, 1, 0); \ + int32x2_t __rev2_584; __rev2_584 = __builtin_shufflevector(__s2_584, __s2_584, 1, 0); \ + int64x2_t __ret_584; \ + __ret_584 = __rev0_584 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_584), __noswap_splat_lane_s32(__rev2_584, __p3_584)); \ + __ret_584 = __builtin_shufflevector(__ret_584, __ret_584, 1, 0); \ + __ret_584; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_lane_s16(__p0_493, __p1_493, __p2_493, __p3_493) __extension__ ({ \ - int32x4_t __s0_493 = __p0_493; \ - int16x8_t __s1_493 = __p1_493; \ - int16x4_t __s2_493 = __p2_493; \ - int32x4_t __ret_493; \ - __ret_493 = __s0_493 - vmull_s16(vget_high_s16(__s1_493), splat_lane_s16(__s2_493, __p3_493)); \ - __ret_493; \ +#define vmlsl_high_lane_s16(__p0_585, __p1_585, __p2_585, __p3_585) __extension__ ({ \ + int32x4_t __s0_585 = __p0_585; \ + int16x8_t __s1_585 = __p1_585; \ + int16x4_t __s2_585 = __p2_585; \ + int32x4_t __ret_585; \ + __ret_585 = __s0_585 - vmull_s16(vget_high_s16(__s1_585), splat_lane_s16(__s2_585, __p3_585)); \ + __ret_585; \ }) #else -#define vmlsl_high_lane_s16(__p0_494, __p1_494, __p2_494, __p3_494) __extension__ ({ \ - int32x4_t __s0_494 = __p0_494; \ - int16x8_t __s1_494 = __p1_494; \ - int16x4_t __s2_494 = __p2_494; \ - int32x4_t __rev0_494; __rev0_494 = __builtin_shufflevector(__s0_494, __s0_494, 3, 2, 1, 0); \ - int16x8_t __rev1_494; __rev1_494 = __builtin_shufflevector(__s1_494, __s1_494, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev2_494; __rev2_494 = __builtin_shufflevector(__s2_494, __s2_494, 3, 2, 1, 0); \ - int32x4_t __ret_494; \ - __ret_494 = __rev0_494 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_494), __noswap_splat_lane_s16(__rev2_494, __p3_494)); \ - __ret_494 = __builtin_shufflevector(__ret_494, __ret_494, 3, 2, 1, 0); \ - __ret_494; \ +#define vmlsl_high_lane_s16(__p0_586, __p1_586, __p2_586, __p3_586) __extension__ ({ \ + int32x4_t __s0_586 = __p0_586; \ + int16x8_t __s1_586 = __p1_586; \ + int16x4_t __s2_586 = __p2_586; \ + int32x4_t __rev0_586; __rev0_586 = __builtin_shufflevector(__s0_586, __s0_586, 3, 2, 1, 0); \ + int16x8_t __rev1_586; __rev1_586 = __builtin_shufflevector(__s1_586, __s1_586, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev2_586; __rev2_586 = __builtin_shufflevector(__s2_586, __s2_586, 3, 2, 1, 0); \ + int32x4_t __ret_586; \ + __ret_586 = __rev0_586 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_586), __noswap_splat_lane_s16(__rev2_586, __p3_586)); \ + __ret_586 = __builtin_shufflevector(__ret_586, __ret_586, 3, 2, 1, 0); \ + __ret_586; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_laneq_u32(__p0_495, __p1_495, __p2_495, __p3_495) __extension__ ({ \ - uint64x2_t __s0_495 = __p0_495; \ - uint32x4_t __s1_495 = __p1_495; \ - uint32x4_t __s2_495 = __p2_495; \ - uint64x2_t __ret_495; \ - __ret_495 = __s0_495 - vmull_u32(vget_high_u32(__s1_495), splat_laneq_u32(__s2_495, __p3_495)); \ - __ret_495; \ +#define vmlsl_high_laneq_u32(__p0_587, __p1_587, __p2_587, __p3_587) __extension__ ({ \ + uint64x2_t __s0_587 = __p0_587; \ + uint32x4_t __s1_587 = __p1_587; \ + uint32x4_t __s2_587 = __p2_587; \ + uint64x2_t __ret_587; \ + __ret_587 = __s0_587 - vmull_u32(vget_high_u32(__s1_587), splat_laneq_u32(__s2_587, __p3_587)); \ + __ret_587; \ }) #else -#define vmlsl_high_laneq_u32(__p0_496, __p1_496, __p2_496, __p3_496) __extension__ ({ \ - uint64x2_t __s0_496 = __p0_496; \ - uint32x4_t __s1_496 = __p1_496; \ - uint32x4_t __s2_496 = __p2_496; \ - uint64x2_t __rev0_496; __rev0_496 = __builtin_shufflevector(__s0_496, __s0_496, 1, 0); \ - uint32x4_t __rev1_496; __rev1_496 = __builtin_shufflevector(__s1_496, __s1_496, 3, 2, 1, 0); \ - uint32x4_t __rev2_496; __rev2_496 = __builtin_shufflevector(__s2_496, __s2_496, 3, 2, 1, 0); \ - uint64x2_t __ret_496; \ - __ret_496 = __rev0_496 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_496), __noswap_splat_laneq_u32(__rev2_496, __p3_496)); \ - __ret_496 = __builtin_shufflevector(__ret_496, __ret_496, 1, 0); \ - __ret_496; \ +#define vmlsl_high_laneq_u32(__p0_588, __p1_588, __p2_588, __p3_588) __extension__ ({ \ + uint64x2_t __s0_588 = __p0_588; \ + uint32x4_t __s1_588 = __p1_588; \ + uint32x4_t __s2_588 = __p2_588; \ + uint64x2_t __rev0_588; __rev0_588 = __builtin_shufflevector(__s0_588, __s0_588, 1, 0); \ + uint32x4_t __rev1_588; __rev1_588 = __builtin_shufflevector(__s1_588, __s1_588, 3, 2, 1, 0); \ + uint32x4_t __rev2_588; __rev2_588 = __builtin_shufflevector(__s2_588, __s2_588, 3, 2, 1, 0); \ + uint64x2_t __ret_588; \ + __ret_588 = __rev0_588 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1_588), __noswap_splat_laneq_u32(__rev2_588, __p3_588)); \ + __ret_588 = __builtin_shufflevector(__ret_588, __ret_588, 1, 0); \ + __ret_588; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_laneq_u16(__p0_497, __p1_497, __p2_497, __p3_497) __extension__ ({ \ - uint32x4_t __s0_497 = __p0_497; \ - uint16x8_t __s1_497 = __p1_497; \ - uint16x8_t __s2_497 = __p2_497; \ - uint32x4_t __ret_497; \ - __ret_497 = __s0_497 - vmull_u16(vget_high_u16(__s1_497), splat_laneq_u16(__s2_497, __p3_497)); \ - __ret_497; \ +#define vmlsl_high_laneq_u16(__p0_589, __p1_589, __p2_589, __p3_589) __extension__ ({ \ + uint32x4_t __s0_589 = __p0_589; \ + uint16x8_t __s1_589 = __p1_589; \ + uint16x8_t __s2_589 = __p2_589; \ + uint32x4_t __ret_589; \ + __ret_589 = __s0_589 - vmull_u16(vget_high_u16(__s1_589), splat_laneq_u16(__s2_589, __p3_589)); \ + __ret_589; \ }) #else -#define vmlsl_high_laneq_u16(__p0_498, __p1_498, __p2_498, __p3_498) __extension__ ({ \ - uint32x4_t __s0_498 = __p0_498; \ - uint16x8_t __s1_498 = __p1_498; \ - uint16x8_t __s2_498 = __p2_498; \ - uint32x4_t __rev0_498; __rev0_498 = __builtin_shufflevector(__s0_498, __s0_498, 3, 2, 1, 0); \ - uint16x8_t __rev1_498; __rev1_498 = __builtin_shufflevector(__s1_498, __s1_498, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev2_498; __rev2_498 = __builtin_shufflevector(__s2_498, __s2_498, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_498; \ - __ret_498 = __rev0_498 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_498), __noswap_splat_laneq_u16(__rev2_498, __p3_498)); \ - __ret_498 = __builtin_shufflevector(__ret_498, __ret_498, 3, 2, 1, 0); \ - __ret_498; \ +#define vmlsl_high_laneq_u16(__p0_590, __p1_590, __p2_590, __p3_590) __extension__ ({ \ + uint32x4_t __s0_590 = __p0_590; \ + uint16x8_t __s1_590 = __p1_590; \ + uint16x8_t __s2_590 = __p2_590; \ + uint32x4_t __rev0_590; __rev0_590 = __builtin_shufflevector(__s0_590, __s0_590, 3, 2, 1, 0); \ + uint16x8_t __rev1_590; __rev1_590 = __builtin_shufflevector(__s1_590, __s1_590, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev2_590; __rev2_590 = __builtin_shufflevector(__s2_590, __s2_590, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_590; \ + __ret_590 = __rev0_590 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1_590), __noswap_splat_laneq_u16(__rev2_590, __p3_590)); \ + __ret_590 = __builtin_shufflevector(__ret_590, __ret_590, 3, 2, 1, 0); \ + __ret_590; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_laneq_s32(__p0_499, __p1_499, __p2_499, __p3_499) __extension__ ({ \ - int64x2_t __s0_499 = __p0_499; \ - int32x4_t __s1_499 = __p1_499; \ - int32x4_t __s2_499 = __p2_499; \ - int64x2_t __ret_499; \ - __ret_499 = __s0_499 - vmull_s32(vget_high_s32(__s1_499), splat_laneq_s32(__s2_499, __p3_499)); \ - __ret_499; \ +#define vmlsl_high_laneq_s32(__p0_591, __p1_591, __p2_591, __p3_591) __extension__ ({ \ + int64x2_t __s0_591 = __p0_591; \ + int32x4_t __s1_591 = __p1_591; \ + int32x4_t __s2_591 = __p2_591; \ + int64x2_t __ret_591; \ + __ret_591 = __s0_591 - vmull_s32(vget_high_s32(__s1_591), splat_laneq_s32(__s2_591, __p3_591)); \ + __ret_591; \ }) #else -#define vmlsl_high_laneq_s32(__p0_500, __p1_500, __p2_500, __p3_500) __extension__ ({ \ - int64x2_t __s0_500 = __p0_500; \ - int32x4_t __s1_500 = __p1_500; \ - int32x4_t __s2_500 = __p2_500; \ - int64x2_t __rev0_500; __rev0_500 = __builtin_shufflevector(__s0_500, __s0_500, 1, 0); \ - int32x4_t __rev1_500; __rev1_500 = __builtin_shufflevector(__s1_500, __s1_500, 3, 2, 1, 0); \ - int32x4_t __rev2_500; __rev2_500 = __builtin_shufflevector(__s2_500, __s2_500, 3, 2, 1, 0); \ - int64x2_t __ret_500; \ - __ret_500 = __rev0_500 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_500), __noswap_splat_laneq_s32(__rev2_500, __p3_500)); \ - __ret_500 = __builtin_shufflevector(__ret_500, __ret_500, 1, 0); \ - __ret_500; \ +#define vmlsl_high_laneq_s32(__p0_592, __p1_592, __p2_592, __p3_592) __extension__ ({ \ + int64x2_t __s0_592 = __p0_592; \ + int32x4_t __s1_592 = __p1_592; \ + int32x4_t __s2_592 = __p2_592; \ + int64x2_t __rev0_592; __rev0_592 = __builtin_shufflevector(__s0_592, __s0_592, 1, 0); \ + int32x4_t __rev1_592; __rev1_592 = __builtin_shufflevector(__s1_592, __s1_592, 3, 2, 1, 0); \ + int32x4_t __rev2_592; __rev2_592 = __builtin_shufflevector(__s2_592, __s2_592, 3, 2, 1, 0); \ + int64x2_t __ret_592; \ + __ret_592 = __rev0_592 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1_592), __noswap_splat_laneq_s32(__rev2_592, __p3_592)); \ + __ret_592 = __builtin_shufflevector(__ret_592, __ret_592, 1, 0); \ + __ret_592; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_high_laneq_s16(__p0_501, __p1_501, __p2_501, __p3_501) __extension__ ({ \ - int32x4_t __s0_501 = __p0_501; \ - int16x8_t __s1_501 = __p1_501; \ - int16x8_t __s2_501 = __p2_501; \ - int32x4_t __ret_501; \ - __ret_501 = __s0_501 - vmull_s16(vget_high_s16(__s1_501), splat_laneq_s16(__s2_501, __p3_501)); \ - __ret_501; \ +#define vmlsl_high_laneq_s16(__p0_593, __p1_593, __p2_593, __p3_593) __extension__ ({ \ + int32x4_t __s0_593 = __p0_593; \ + int16x8_t __s1_593 = __p1_593; \ + int16x8_t __s2_593 = __p2_593; \ + int32x4_t __ret_593; \ + __ret_593 = __s0_593 - vmull_s16(vget_high_s16(__s1_593), splat_laneq_s16(__s2_593, __p3_593)); \ + __ret_593; \ }) #else -#define vmlsl_high_laneq_s16(__p0_502, __p1_502, __p2_502, __p3_502) __extension__ ({ \ - int32x4_t __s0_502 = __p0_502; \ - int16x8_t __s1_502 = __p1_502; \ - int16x8_t __s2_502 = __p2_502; \ - int32x4_t __rev0_502; __rev0_502 = __builtin_shufflevector(__s0_502, __s0_502, 3, 2, 1, 0); \ - int16x8_t __rev1_502; __rev1_502 = __builtin_shufflevector(__s1_502, __s1_502, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_502; __rev2_502 = __builtin_shufflevector(__s2_502, __s2_502, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_502; \ - __ret_502 = __rev0_502 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_502), __noswap_splat_laneq_s16(__rev2_502, __p3_502)); \ - __ret_502 = __builtin_shufflevector(__ret_502, __ret_502, 3, 2, 1, 0); \ - __ret_502; \ +#define vmlsl_high_laneq_s16(__p0_594, __p1_594, __p2_594, __p3_594) __extension__ ({ \ + int32x4_t __s0_594 = __p0_594; \ + int16x8_t __s1_594 = __p1_594; \ + int16x8_t __s2_594 = __p2_594; \ + int32x4_t __rev0_594; __rev0_594 = __builtin_shufflevector(__s0_594, __s0_594, 3, 2, 1, 0); \ + int16x8_t __rev1_594; __rev1_594 = __builtin_shufflevector(__s1_594, __s1_594, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_594; __rev2_594 = __builtin_shufflevector(__s2_594, __s2_594, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_594; \ + __ret_594 = __rev0_594 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1_594), __noswap_splat_laneq_s16(__rev2_594, __p3_594)); \ + __ret_594 = __builtin_shufflevector(__ret_594, __ret_594, 3, 2, 1, 0); \ + __ret_594; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_laneq_u32(__p0_503, __p1_503, __p2_503, __p3_503) __extension__ ({ \ - uint64x2_t __s0_503 = __p0_503; \ - uint32x2_t __s1_503 = __p1_503; \ - uint32x4_t __s2_503 = __p2_503; \ - uint64x2_t __ret_503; \ - __ret_503 = __s0_503 - vmull_u32(__s1_503, splat_laneq_u32(__s2_503, __p3_503)); \ - __ret_503; \ +#define vmlsl_laneq_u32(__p0_595, __p1_595, __p2_595, __p3_595) __extension__ ({ \ + uint64x2_t __s0_595 = __p0_595; \ + uint32x2_t __s1_595 = __p1_595; \ + uint32x4_t __s2_595 = __p2_595; \ + uint64x2_t __ret_595; \ + __ret_595 = __s0_595 - vmull_u32(__s1_595, splat_laneq_u32(__s2_595, __p3_595)); \ + __ret_595; \ }) #else -#define vmlsl_laneq_u32(__p0_504, __p1_504, __p2_504, __p3_504) __extension__ ({ \ - uint64x2_t __s0_504 = __p0_504; \ - uint32x2_t __s1_504 = __p1_504; \ - uint32x4_t __s2_504 = __p2_504; \ - uint64x2_t __rev0_504; __rev0_504 = __builtin_shufflevector(__s0_504, __s0_504, 1, 0); \ - uint32x2_t __rev1_504; __rev1_504 = __builtin_shufflevector(__s1_504, __s1_504, 1, 0); \ - uint32x4_t __rev2_504; __rev2_504 = __builtin_shufflevector(__s2_504, __s2_504, 3, 2, 1, 0); \ - uint64x2_t __ret_504; \ - __ret_504 = __rev0_504 - __noswap_vmull_u32(__rev1_504, __noswap_splat_laneq_u32(__rev2_504, __p3_504)); \ - __ret_504 = __builtin_shufflevector(__ret_504, __ret_504, 1, 0); \ - __ret_504; \ +#define vmlsl_laneq_u32(__p0_596, __p1_596, __p2_596, __p3_596) __extension__ ({ \ + uint64x2_t __s0_596 = __p0_596; \ + uint32x2_t __s1_596 = __p1_596; \ + uint32x4_t __s2_596 = __p2_596; \ + uint64x2_t __rev0_596; __rev0_596 = __builtin_shufflevector(__s0_596, __s0_596, 1, 0); \ + uint32x2_t __rev1_596; __rev1_596 = __builtin_shufflevector(__s1_596, __s1_596, 1, 0); \ + uint32x4_t __rev2_596; __rev2_596 = __builtin_shufflevector(__s2_596, __s2_596, 3, 2, 1, 0); \ + uint64x2_t __ret_596; \ + __ret_596 = __rev0_596 - __noswap_vmull_u32(__rev1_596, __noswap_splat_laneq_u32(__rev2_596, __p3_596)); \ + __ret_596 = __builtin_shufflevector(__ret_596, __ret_596, 1, 0); \ + __ret_596; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_laneq_u16(__p0_505, __p1_505, __p2_505, __p3_505) __extension__ ({ \ - uint32x4_t __s0_505 = __p0_505; \ - uint16x4_t __s1_505 = __p1_505; \ - uint16x8_t __s2_505 = __p2_505; \ - uint32x4_t __ret_505; \ - __ret_505 = __s0_505 - vmull_u16(__s1_505, splat_laneq_u16(__s2_505, __p3_505)); \ - __ret_505; \ +#define vmlsl_laneq_u16(__p0_597, __p1_597, __p2_597, __p3_597) __extension__ ({ \ + uint32x4_t __s0_597 = __p0_597; \ + uint16x4_t __s1_597 = __p1_597; \ + uint16x8_t __s2_597 = __p2_597; \ + uint32x4_t __ret_597; \ + __ret_597 = __s0_597 - vmull_u16(__s1_597, splat_laneq_u16(__s2_597, __p3_597)); \ + __ret_597; \ }) #else -#define vmlsl_laneq_u16(__p0_506, __p1_506, __p2_506, __p3_506) __extension__ ({ \ - uint32x4_t __s0_506 = __p0_506; \ - uint16x4_t __s1_506 = __p1_506; \ - uint16x8_t __s2_506 = __p2_506; \ - uint32x4_t __rev0_506; __rev0_506 = __builtin_shufflevector(__s0_506, __s0_506, 3, 2, 1, 0); \ - uint16x4_t __rev1_506; __rev1_506 = __builtin_shufflevector(__s1_506, __s1_506, 3, 2, 1, 0); \ - uint16x8_t __rev2_506; __rev2_506 = __builtin_shufflevector(__s2_506, __s2_506, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_506; \ - __ret_506 = __rev0_506 - __noswap_vmull_u16(__rev1_506, __noswap_splat_laneq_u16(__rev2_506, __p3_506)); \ - __ret_506 = __builtin_shufflevector(__ret_506, __ret_506, 3, 2, 1, 0); \ - __ret_506; \ +#define vmlsl_laneq_u16(__p0_598, __p1_598, __p2_598, __p3_598) __extension__ ({ \ + uint32x4_t __s0_598 = __p0_598; \ + uint16x4_t __s1_598 = __p1_598; \ + uint16x8_t __s2_598 = __p2_598; \ + uint32x4_t __rev0_598; __rev0_598 = __builtin_shufflevector(__s0_598, __s0_598, 3, 2, 1, 0); \ + uint16x4_t __rev1_598; __rev1_598 = __builtin_shufflevector(__s1_598, __s1_598, 3, 2, 1, 0); \ + uint16x8_t __rev2_598; __rev2_598 = __builtin_shufflevector(__s2_598, __s2_598, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_598; \ + __ret_598 = __rev0_598 - __noswap_vmull_u16(__rev1_598, __noswap_splat_laneq_u16(__rev2_598, __p3_598)); \ + __ret_598 = __builtin_shufflevector(__ret_598, __ret_598, 3, 2, 1, 0); \ + __ret_598; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_laneq_s32(__p0_507, __p1_507, __p2_507, __p3_507) __extension__ ({ \ - int64x2_t __s0_507 = __p0_507; \ - int32x2_t __s1_507 = __p1_507; \ - int32x4_t __s2_507 = __p2_507; \ - int64x2_t __ret_507; \ - __ret_507 = __s0_507 - vmull_s32(__s1_507, splat_laneq_s32(__s2_507, __p3_507)); \ - __ret_507; \ +#define vmlsl_laneq_s32(__p0_599, __p1_599, __p2_599, __p3_599) __extension__ ({ \ + int64x2_t __s0_599 = __p0_599; \ + int32x2_t __s1_599 = __p1_599; \ + int32x4_t __s2_599 = __p2_599; \ + int64x2_t __ret_599; \ + __ret_599 = __s0_599 - vmull_s32(__s1_599, splat_laneq_s32(__s2_599, __p3_599)); \ + __ret_599; \ }) #else -#define vmlsl_laneq_s32(__p0_508, __p1_508, __p2_508, __p3_508) __extension__ ({ \ - int64x2_t __s0_508 = __p0_508; \ - int32x2_t __s1_508 = __p1_508; \ - int32x4_t __s2_508 = __p2_508; \ - int64x2_t __rev0_508; __rev0_508 = __builtin_shufflevector(__s0_508, __s0_508, 1, 0); \ - int32x2_t __rev1_508; __rev1_508 = __builtin_shufflevector(__s1_508, __s1_508, 1, 0); \ - int32x4_t __rev2_508; __rev2_508 = __builtin_shufflevector(__s2_508, __s2_508, 3, 2, 1, 0); \ - int64x2_t __ret_508; \ - __ret_508 = __rev0_508 - __noswap_vmull_s32(__rev1_508, __noswap_splat_laneq_s32(__rev2_508, __p3_508)); \ - __ret_508 = __builtin_shufflevector(__ret_508, __ret_508, 1, 0); \ - __ret_508; \ +#define vmlsl_laneq_s32(__p0_600, __p1_600, __p2_600, __p3_600) __extension__ ({ \ + int64x2_t __s0_600 = __p0_600; \ + int32x2_t __s1_600 = __p1_600; \ + int32x4_t __s2_600 = __p2_600; \ + int64x2_t __rev0_600; __rev0_600 = __builtin_shufflevector(__s0_600, __s0_600, 1, 0); \ + int32x2_t __rev1_600; __rev1_600 = __builtin_shufflevector(__s1_600, __s1_600, 1, 0); \ + int32x4_t __rev2_600; __rev2_600 = __builtin_shufflevector(__s2_600, __s2_600, 3, 2, 1, 0); \ + int64x2_t __ret_600; \ + __ret_600 = __rev0_600 - __noswap_vmull_s32(__rev1_600, __noswap_splat_laneq_s32(__rev2_600, __p3_600)); \ + __ret_600 = __builtin_shufflevector(__ret_600, __ret_600, 1, 0); \ + __ret_600; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_laneq_s16(__p0_509, __p1_509, __p2_509, __p3_509) __extension__ ({ \ - int32x4_t __s0_509 = __p0_509; \ - int16x4_t __s1_509 = __p1_509; \ - int16x8_t __s2_509 = __p2_509; \ - int32x4_t __ret_509; \ - __ret_509 = __s0_509 - vmull_s16(__s1_509, splat_laneq_s16(__s2_509, __p3_509)); \ - __ret_509; \ +#define vmlsl_laneq_s16(__p0_601, __p1_601, __p2_601, __p3_601) __extension__ ({ \ + int32x4_t __s0_601 = __p0_601; \ + int16x4_t __s1_601 = __p1_601; \ + int16x8_t __s2_601 = __p2_601; \ + int32x4_t __ret_601; \ + __ret_601 = __s0_601 - vmull_s16(__s1_601, splat_laneq_s16(__s2_601, __p3_601)); \ + __ret_601; \ }) #else -#define vmlsl_laneq_s16(__p0_510, __p1_510, __p2_510, __p3_510) __extension__ ({ \ - int32x4_t __s0_510 = __p0_510; \ - int16x4_t __s1_510 = __p1_510; \ - int16x8_t __s2_510 = __p2_510; \ - int32x4_t __rev0_510; __rev0_510 = __builtin_shufflevector(__s0_510, __s0_510, 3, 2, 1, 0); \ - int16x4_t __rev1_510; __rev1_510 = __builtin_shufflevector(__s1_510, __s1_510, 3, 2, 1, 0); \ - int16x8_t __rev2_510; __rev2_510 = __builtin_shufflevector(__s2_510, __s2_510, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_510; \ - __ret_510 = __rev0_510 - __noswap_vmull_s16(__rev1_510, __noswap_splat_laneq_s16(__rev2_510, __p3_510)); \ - __ret_510 = __builtin_shufflevector(__ret_510, __ret_510, 3, 2, 1, 0); \ - __ret_510; \ +#define vmlsl_laneq_s16(__p0_602, __p1_602, __p2_602, __p3_602) __extension__ ({ \ + int32x4_t __s0_602 = __p0_602; \ + int16x4_t __s1_602 = __p1_602; \ + int16x8_t __s2_602 = __p2_602; \ + int32x4_t __rev0_602; __rev0_602 = __builtin_shufflevector(__s0_602, __s0_602, 3, 2, 1, 0); \ + int16x4_t __rev1_602; __rev1_602 = __builtin_shufflevector(__s1_602, __s1_602, 3, 2, 1, 0); \ + int16x8_t __rev2_602; __rev2_602 = __builtin_shufflevector(__s2_602, __s2_602, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_602; \ + __ret_602 = __rev0_602 - __noswap_vmull_s16(__rev1_602, __noswap_splat_laneq_s16(__rev2_602, __p3_602)); \ + __ret_602 = __builtin_shufflevector(__ret_602, __ret_602, 3, 2, 1, 0); \ + __ret_602; \ }) #endif @@ -53701,146 +55371,146 @@ __ai float64x1_t vmov_n_f64(float64_t __p0) { return __ret; } #ifdef __LITTLE_ENDIAN__ -__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_511) { - uint16x8_t __ret_511; - uint8x8_t __a1_511 = vget_high_u8(__p0_511); - __ret_511 = (uint16x8_t)(vshll_n_u8(__a1_511, 0)); - return __ret_511; +__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_603) { + uint16x8_t __ret_603; + uint8x8_t __a1_603 = vget_high_u8(__p0_603); + __ret_603 = (uint16x8_t)(vshll_n_u8(__a1_603, 0)); + return __ret_603; } #else -__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_512) { - uint8x16_t __rev0_512; __rev0_512 = __builtin_shufflevector(__p0_512, __p0_512, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); - uint16x8_t __ret_512; - uint8x8_t __a1_512 = __noswap_vget_high_u8(__rev0_512); - __ret_512 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_512, 0)); - __ret_512 = __builtin_shufflevector(__ret_512, __ret_512, 7, 6, 5, 4, 3, 2, 1, 0); - return __ret_512; +__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_604) { + uint8x16_t __rev0_604; __rev0_604 = __builtin_shufflevector(__p0_604, __p0_604, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); + uint16x8_t __ret_604; + uint8x8_t __a1_604 = __noswap_vget_high_u8(__rev0_604); + __ret_604 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_604, 0)); + __ret_604 = __builtin_shufflevector(__ret_604, __ret_604, 7, 6, 5, 4, 3, 2, 1, 0); + return __ret_604; } -__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_513) { - uint16x8_t __ret_513; - uint8x8_t __a1_513 = __noswap_vget_high_u8(__p0_513); - __ret_513 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_513, 0)); - return __ret_513; +__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_605) { + uint16x8_t __ret_605; + uint8x8_t __a1_605 = __noswap_vget_high_u8(__p0_605); + __ret_605 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_605, 0)); + return __ret_605; } #endif #ifdef __LITTLE_ENDIAN__ -__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_514) { - uint64x2_t __ret_514; - uint32x2_t __a1_514 = vget_high_u32(__p0_514); - __ret_514 = (uint64x2_t)(vshll_n_u32(__a1_514, 0)); - return __ret_514; +__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_606) { + uint64x2_t __ret_606; + uint32x2_t __a1_606 = vget_high_u32(__p0_606); + __ret_606 = (uint64x2_t)(vshll_n_u32(__a1_606, 0)); + return __ret_606; } #else -__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_515) { - uint32x4_t __rev0_515; __rev0_515 = __builtin_shufflevector(__p0_515, __p0_515, 3, 2, 1, 0); - uint64x2_t __ret_515; - uint32x2_t __a1_515 = __noswap_vget_high_u32(__rev0_515); - __ret_515 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_515, 0)); - __ret_515 = __builtin_shufflevector(__ret_515, __ret_515, 1, 0); - return __ret_515; +__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_607) { + uint32x4_t __rev0_607; __rev0_607 = __builtin_shufflevector(__p0_607, __p0_607, 3, 2, 1, 0); + uint64x2_t __ret_607; + uint32x2_t __a1_607 = __noswap_vget_high_u32(__rev0_607); + __ret_607 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_607, 0)); + __ret_607 = __builtin_shufflevector(__ret_607, __ret_607, 1, 0); + return __ret_607; } -__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_516) { - uint64x2_t __ret_516; - uint32x2_t __a1_516 = __noswap_vget_high_u32(__p0_516); - __ret_516 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_516, 0)); - return __ret_516; +__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_608) { + uint64x2_t __ret_608; + uint32x2_t __a1_608 = __noswap_vget_high_u32(__p0_608); + __ret_608 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_608, 0)); + return __ret_608; } #endif #ifdef __LITTLE_ENDIAN__ -__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_517) { - uint32x4_t __ret_517; - uint16x4_t __a1_517 = vget_high_u16(__p0_517); - __ret_517 = (uint32x4_t)(vshll_n_u16(__a1_517, 0)); - return __ret_517; +__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_609) { + uint32x4_t __ret_609; + uint16x4_t __a1_609 = vget_high_u16(__p0_609); + __ret_609 = (uint32x4_t)(vshll_n_u16(__a1_609, 0)); + return __ret_609; } #else -__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_518) { - uint16x8_t __rev0_518; __rev0_518 = __builtin_shufflevector(__p0_518, __p0_518, 7, 6, 5, 4, 3, 2, 1, 0); - uint32x4_t __ret_518; - uint16x4_t __a1_518 = __noswap_vget_high_u16(__rev0_518); - __ret_518 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_518, 0)); - __ret_518 = __builtin_shufflevector(__ret_518, __ret_518, 3, 2, 1, 0); - return __ret_518; +__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_610) { + uint16x8_t __rev0_610; __rev0_610 = __builtin_shufflevector(__p0_610, __p0_610, 7, 6, 5, 4, 3, 2, 1, 0); + uint32x4_t __ret_610; + uint16x4_t __a1_610 = __noswap_vget_high_u16(__rev0_610); + __ret_610 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_610, 0)); + __ret_610 = __builtin_shufflevector(__ret_610, __ret_610, 3, 2, 1, 0); + return __ret_610; } -__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_519) { - uint32x4_t __ret_519; - uint16x4_t __a1_519 = __noswap_vget_high_u16(__p0_519); - __ret_519 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_519, 0)); - return __ret_519; +__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_611) { + uint32x4_t __ret_611; + uint16x4_t __a1_611 = __noswap_vget_high_u16(__p0_611); + __ret_611 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_611, 0)); + return __ret_611; } #endif #ifdef __LITTLE_ENDIAN__ -__ai int16x8_t vmovl_high_s8(int8x16_t __p0_520) { - int16x8_t __ret_520; - int8x8_t __a1_520 = vget_high_s8(__p0_520); - __ret_520 = (int16x8_t)(vshll_n_s8(__a1_520, 0)); - return __ret_520; +__ai int16x8_t vmovl_high_s8(int8x16_t __p0_612) { + int16x8_t __ret_612; + int8x8_t __a1_612 = vget_high_s8(__p0_612); + __ret_612 = (int16x8_t)(vshll_n_s8(__a1_612, 0)); + return __ret_612; } #else -__ai int16x8_t vmovl_high_s8(int8x16_t __p0_521) { - int8x16_t __rev0_521; __rev0_521 = __builtin_shufflevector(__p0_521, __p0_521, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); - int16x8_t __ret_521; - int8x8_t __a1_521 = __noswap_vget_high_s8(__rev0_521); - __ret_521 = (int16x8_t)(__noswap_vshll_n_s8(__a1_521, 0)); - __ret_521 = __builtin_shufflevector(__ret_521, __ret_521, 7, 6, 5, 4, 3, 2, 1, 0); - return __ret_521; +__ai int16x8_t vmovl_high_s8(int8x16_t __p0_613) { + int8x16_t __rev0_613; __rev0_613 = __builtin_shufflevector(__p0_613, __p0_613, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); + int16x8_t __ret_613; + int8x8_t __a1_613 = __noswap_vget_high_s8(__rev0_613); + __ret_613 = (int16x8_t)(__noswap_vshll_n_s8(__a1_613, 0)); + __ret_613 = __builtin_shufflevector(__ret_613, __ret_613, 7, 6, 5, 4, 3, 2, 1, 0); + return __ret_613; } -__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_522) { - int16x8_t __ret_522; - int8x8_t __a1_522 = __noswap_vget_high_s8(__p0_522); - __ret_522 = (int16x8_t)(__noswap_vshll_n_s8(__a1_522, 0)); - return __ret_522; +__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_614) { + int16x8_t __ret_614; + int8x8_t __a1_614 = __noswap_vget_high_s8(__p0_614); + __ret_614 = (int16x8_t)(__noswap_vshll_n_s8(__a1_614, 0)); + return __ret_614; } #endif #ifdef __LITTLE_ENDIAN__ -__ai int64x2_t vmovl_high_s32(int32x4_t __p0_523) { - int64x2_t __ret_523; - int32x2_t __a1_523 = vget_high_s32(__p0_523); - __ret_523 = (int64x2_t)(vshll_n_s32(__a1_523, 0)); - return __ret_523; +__ai int64x2_t vmovl_high_s32(int32x4_t __p0_615) { + int64x2_t __ret_615; + int32x2_t __a1_615 = vget_high_s32(__p0_615); + __ret_615 = (int64x2_t)(vshll_n_s32(__a1_615, 0)); + return __ret_615; } #else -__ai int64x2_t vmovl_high_s32(int32x4_t __p0_524) { - int32x4_t __rev0_524; __rev0_524 = __builtin_shufflevector(__p0_524, __p0_524, 3, 2, 1, 0); - int64x2_t __ret_524; - int32x2_t __a1_524 = __noswap_vget_high_s32(__rev0_524); - __ret_524 = (int64x2_t)(__noswap_vshll_n_s32(__a1_524, 0)); - __ret_524 = __builtin_shufflevector(__ret_524, __ret_524, 1, 0); - return __ret_524; +__ai int64x2_t vmovl_high_s32(int32x4_t __p0_616) { + int32x4_t __rev0_616; __rev0_616 = __builtin_shufflevector(__p0_616, __p0_616, 3, 2, 1, 0); + int64x2_t __ret_616; + int32x2_t __a1_616 = __noswap_vget_high_s32(__rev0_616); + __ret_616 = (int64x2_t)(__noswap_vshll_n_s32(__a1_616, 0)); + __ret_616 = __builtin_shufflevector(__ret_616, __ret_616, 1, 0); + return __ret_616; } -__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_525) { - int64x2_t __ret_525; - int32x2_t __a1_525 = __noswap_vget_high_s32(__p0_525); - __ret_525 = (int64x2_t)(__noswap_vshll_n_s32(__a1_525, 0)); - return __ret_525; +__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_617) { + int64x2_t __ret_617; + int32x2_t __a1_617 = __noswap_vget_high_s32(__p0_617); + __ret_617 = (int64x2_t)(__noswap_vshll_n_s32(__a1_617, 0)); + return __ret_617; } #endif #ifdef __LITTLE_ENDIAN__ -__ai int32x4_t vmovl_high_s16(int16x8_t __p0_526) { - int32x4_t __ret_526; - int16x4_t __a1_526 = vget_high_s16(__p0_526); - __ret_526 = (int32x4_t)(vshll_n_s16(__a1_526, 0)); - return __ret_526; +__ai int32x4_t vmovl_high_s16(int16x8_t __p0_618) { + int32x4_t __ret_618; + int16x4_t __a1_618 = vget_high_s16(__p0_618); + __ret_618 = (int32x4_t)(vshll_n_s16(__a1_618, 0)); + return __ret_618; } #else -__ai int32x4_t vmovl_high_s16(int16x8_t __p0_527) { - int16x8_t __rev0_527; __rev0_527 = __builtin_shufflevector(__p0_527, __p0_527, 7, 6, 5, 4, 3, 2, 1, 0); - int32x4_t __ret_527; - int16x4_t __a1_527 = __noswap_vget_high_s16(__rev0_527); - __ret_527 = (int32x4_t)(__noswap_vshll_n_s16(__a1_527, 0)); - __ret_527 = __builtin_shufflevector(__ret_527, __ret_527, 3, 2, 1, 0); - return __ret_527; +__ai int32x4_t vmovl_high_s16(int16x8_t __p0_619) { + int16x8_t __rev0_619; __rev0_619 = __builtin_shufflevector(__p0_619, __p0_619, 7, 6, 5, 4, 3, 2, 1, 0); + int32x4_t __ret_619; + int16x4_t __a1_619 = __noswap_vget_high_s16(__rev0_619); + __ret_619 = (int32x4_t)(__noswap_vshll_n_s16(__a1_619, 0)); + __ret_619 = __builtin_shufflevector(__ret_619, __ret_619, 3, 2, 1, 0); + return __ret_619; } -__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_528) { - int32x4_t __ret_528; - int16x4_t __a1_528 = __noswap_vget_high_s16(__p0_528); - __ret_528 = (int32x4_t)(__noswap_vshll_n_s16(__a1_528, 0)); - return __ret_528; +__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_620) { + int32x4_t __ret_620; + int16x4_t __a1_620 = __noswap_vget_high_s16(__p0_620); + __ret_620 = (int32x4_t)(__noswap_vshll_n_s16(__a1_620, 0)); + return __ret_620; } #endif @@ -53968,29 +55638,29 @@ __ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) { __ret = __p0 * __p1; return __ret; } -#define vmuld_lane_f64(__p0_529, __p1_529, __p2_529) __extension__ ({ \ - float64_t __s0_529 = __p0_529; \ - float64x1_t __s1_529 = __p1_529; \ - float64_t __ret_529; \ - __ret_529 = __s0_529 * vget_lane_f64(__s1_529, __p2_529); \ - __ret_529; \ +#define vmuld_lane_f64(__p0_621, __p1_621, __p2_621) __extension__ ({ \ + float64_t __s0_621 = __p0_621; \ + float64x1_t __s1_621 = __p1_621; \ + float64_t __ret_621; \ + __ret_621 = __s0_621 * vget_lane_f64(__s1_621, __p2_621); \ + __ret_621; \ }) #ifdef __LITTLE_ENDIAN__ -#define vmuls_lane_f32(__p0_530, __p1_530, __p2_530) __extension__ ({ \ - float32_t __s0_530 = __p0_530; \ - float32x2_t __s1_530 = __p1_530; \ - float32_t __ret_530; \ - __ret_530 = __s0_530 * vget_lane_f32(__s1_530, __p2_530); \ - __ret_530; \ +#define vmuls_lane_f32(__p0_622, __p1_622, __p2_622) __extension__ ({ \ + float32_t __s0_622 = __p0_622; \ + float32x2_t __s1_622 = __p1_622; \ + float32_t __ret_622; \ + __ret_622 = __s0_622 * vget_lane_f32(__s1_622, __p2_622); \ + __ret_622; \ }) #else -#define vmuls_lane_f32(__p0_531, __p1_531, __p2_531) __extension__ ({ \ - float32_t __s0_531 = __p0_531; \ - float32x2_t __s1_531 = __p1_531; \ - float32x2_t __rev1_531; __rev1_531 = __builtin_shufflevector(__s1_531, __s1_531, 1, 0); \ - float32_t __ret_531; \ - __ret_531 = __s0_531 * __noswap_vget_lane_f32(__rev1_531, __p2_531); \ - __ret_531; \ +#define vmuls_lane_f32(__p0_623, __p1_623, __p2_623) __extension__ ({ \ + float32_t __s0_623 = __p0_623; \ + float32x2_t __s1_623 = __p1_623; \ + float32x2_t __rev1_623; __rev1_623 = __builtin_shufflevector(__s1_623, __s1_623, 1, 0); \ + float32_t __ret_623; \ + __ret_623 = __s0_623 * __noswap_vget_lane_f32(__rev1_623, __p2_623); \ + __ret_623; \ }) #endif @@ -54002,60 +55672,60 @@ __ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) { __ret; \ }) #ifdef __LITTLE_ENDIAN__ -#define vmulq_lane_f64(__p0_532, __p1_532, __p2_532) __extension__ ({ \ - float64x2_t __s0_532 = __p0_532; \ - float64x1_t __s1_532 = __p1_532; \ - float64x2_t __ret_532; \ - __ret_532 = __s0_532 * splatq_lane_f64(__s1_532, __p2_532); \ - __ret_532; \ +#define vmulq_lane_f64(__p0_624, __p1_624, __p2_624) __extension__ ({ \ + float64x2_t __s0_624 = __p0_624; \ + float64x1_t __s1_624 = __p1_624; \ + float64x2_t __ret_624; \ + __ret_624 = __s0_624 * splatq_lane_f64(__s1_624, __p2_624); \ + __ret_624; \ }) #else -#define vmulq_lane_f64(__p0_533, __p1_533, __p2_533) __extension__ ({ \ - float64x2_t __s0_533 = __p0_533; \ - float64x1_t __s1_533 = __p1_533; \ - float64x2_t __rev0_533; __rev0_533 = __builtin_shufflevector(__s0_533, __s0_533, 1, 0); \ - float64x2_t __ret_533; \ - __ret_533 = __rev0_533 * __noswap_splatq_lane_f64(__s1_533, __p2_533); \ - __ret_533 = __builtin_shufflevector(__ret_533, __ret_533, 1, 0); \ - __ret_533; \ +#define vmulq_lane_f64(__p0_625, __p1_625, __p2_625) __extension__ ({ \ + float64x2_t __s0_625 = __p0_625; \ + float64x1_t __s1_625 = __p1_625; \ + float64x2_t __rev0_625; __rev0_625 = __builtin_shufflevector(__s0_625, __s0_625, 1, 0); \ + float64x2_t __ret_625; \ + __ret_625 = __rev0_625 * __noswap_splatq_lane_f64(__s1_625, __p2_625); \ + __ret_625 = __builtin_shufflevector(__ret_625, __ret_625, 1, 0); \ + __ret_625; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmuld_laneq_f64(__p0_534, __p1_534, __p2_534) __extension__ ({ \ - float64_t __s0_534 = __p0_534; \ - float64x2_t __s1_534 = __p1_534; \ - float64_t __ret_534; \ - __ret_534 = __s0_534 * vgetq_lane_f64(__s1_534, __p2_534); \ - __ret_534; \ +#define vmuld_laneq_f64(__p0_626, __p1_626, __p2_626) __extension__ ({ \ + float64_t __s0_626 = __p0_626; \ + float64x2_t __s1_626 = __p1_626; \ + float64_t __ret_626; \ + __ret_626 = __s0_626 * vgetq_lane_f64(__s1_626, __p2_626); \ + __ret_626; \ }) #else -#define vmuld_laneq_f64(__p0_535, __p1_535, __p2_535) __extension__ ({ \ - float64_t __s0_535 = __p0_535; \ - float64x2_t __s1_535 = __p1_535; \ - float64x2_t __rev1_535; __rev1_535 = __builtin_shufflevector(__s1_535, __s1_535, 1, 0); \ - float64_t __ret_535; \ - __ret_535 = __s0_535 * __noswap_vgetq_lane_f64(__rev1_535, __p2_535); \ - __ret_535; \ +#define vmuld_laneq_f64(__p0_627, __p1_627, __p2_627) __extension__ ({ \ + float64_t __s0_627 = __p0_627; \ + float64x2_t __s1_627 = __p1_627; \ + float64x2_t __rev1_627; __rev1_627 = __builtin_shufflevector(__s1_627, __s1_627, 1, 0); \ + float64_t __ret_627; \ + __ret_627 = __s0_627 * __noswap_vgetq_lane_f64(__rev1_627, __p2_627); \ + __ret_627; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmuls_laneq_f32(__p0_536, __p1_536, __p2_536) __extension__ ({ \ - float32_t __s0_536 = __p0_536; \ - float32x4_t __s1_536 = __p1_536; \ - float32_t __ret_536; \ - __ret_536 = __s0_536 * vgetq_lane_f32(__s1_536, __p2_536); \ - __ret_536; \ +#define vmuls_laneq_f32(__p0_628, __p1_628, __p2_628) __extension__ ({ \ + float32_t __s0_628 = __p0_628; \ + float32x4_t __s1_628 = __p1_628; \ + float32_t __ret_628; \ + __ret_628 = __s0_628 * vgetq_lane_f32(__s1_628, __p2_628); \ + __ret_628; \ }) #else -#define vmuls_laneq_f32(__p0_537, __p1_537, __p2_537) __extension__ ({ \ - float32_t __s0_537 = __p0_537; \ - float32x4_t __s1_537 = __p1_537; \ - float32x4_t __rev1_537; __rev1_537 = __builtin_shufflevector(__s1_537, __s1_537, 3, 2, 1, 0); \ - float32_t __ret_537; \ - __ret_537 = __s0_537 * __noswap_vgetq_lane_f32(__rev1_537, __p2_537); \ - __ret_537; \ +#define vmuls_laneq_f32(__p0_629, __p1_629, __p2_629) __extension__ ({ \ + float32_t __s0_629 = __p0_629; \ + float32x4_t __s1_629 = __p1_629; \ + float32x4_t __rev1_629; __rev1_629 = __builtin_shufflevector(__s1_629, __s1_629, 3, 2, 1, 0); \ + float32_t __ret_629; \ + __ret_629 = __s0_629 * __noswap_vgetq_lane_f32(__rev1_629, __p2_629); \ + __ret_629; \ }) #endif @@ -54079,233 +55749,233 @@ __ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_laneq_u32(__p0_538, __p1_538, __p2_538) __extension__ ({ \ - uint32x4_t __s0_538 = __p0_538; \ - uint32x4_t __s1_538 = __p1_538; \ - uint32x4_t __ret_538; \ - __ret_538 = __s0_538 * splatq_laneq_u32(__s1_538, __p2_538); \ - __ret_538; \ +#define vmulq_laneq_u32(__p0_630, __p1_630, __p2_630) __extension__ ({ \ + uint32x4_t __s0_630 = __p0_630; \ + uint32x4_t __s1_630 = __p1_630; \ + uint32x4_t __ret_630; \ + __ret_630 = __s0_630 * splatq_laneq_u32(__s1_630, __p2_630); \ + __ret_630; \ }) #else -#define vmulq_laneq_u32(__p0_539, __p1_539, __p2_539) __extension__ ({ \ - uint32x4_t __s0_539 = __p0_539; \ - uint32x4_t __s1_539 = __p1_539; \ - uint32x4_t __rev0_539; __rev0_539 = __builtin_shufflevector(__s0_539, __s0_539, 3, 2, 1, 0); \ - uint32x4_t __rev1_539; __rev1_539 = __builtin_shufflevector(__s1_539, __s1_539, 3, 2, 1, 0); \ - uint32x4_t __ret_539; \ - __ret_539 = __rev0_539 * __noswap_splatq_laneq_u32(__rev1_539, __p2_539); \ - __ret_539 = __builtin_shufflevector(__ret_539, __ret_539, 3, 2, 1, 0); \ - __ret_539; \ +#define vmulq_laneq_u32(__p0_631, __p1_631, __p2_631) __extension__ ({ \ + uint32x4_t __s0_631 = __p0_631; \ + uint32x4_t __s1_631 = __p1_631; \ + uint32x4_t __rev0_631; __rev0_631 = __builtin_shufflevector(__s0_631, __s0_631, 3, 2, 1, 0); \ + uint32x4_t __rev1_631; __rev1_631 = __builtin_shufflevector(__s1_631, __s1_631, 3, 2, 1, 0); \ + uint32x4_t __ret_631; \ + __ret_631 = __rev0_631 * __noswap_splatq_laneq_u32(__rev1_631, __p2_631); \ + __ret_631 = __builtin_shufflevector(__ret_631, __ret_631, 3, 2, 1, 0); \ + __ret_631; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_laneq_u16(__p0_540, __p1_540, __p2_540) __extension__ ({ \ - uint16x8_t __s0_540 = __p0_540; \ - uint16x8_t __s1_540 = __p1_540; \ - uint16x8_t __ret_540; \ - __ret_540 = __s0_540 * splatq_laneq_u16(__s1_540, __p2_540); \ - __ret_540; \ +#define vmulq_laneq_u16(__p0_632, __p1_632, __p2_632) __extension__ ({ \ + uint16x8_t __s0_632 = __p0_632; \ + uint16x8_t __s1_632 = __p1_632; \ + uint16x8_t __ret_632; \ + __ret_632 = __s0_632 * splatq_laneq_u16(__s1_632, __p2_632); \ + __ret_632; \ }) #else -#define vmulq_laneq_u16(__p0_541, __p1_541, __p2_541) __extension__ ({ \ - uint16x8_t __s0_541 = __p0_541; \ - uint16x8_t __s1_541 = __p1_541; \ - uint16x8_t __rev0_541; __rev0_541 = __builtin_shufflevector(__s0_541, __s0_541, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_541; __rev1_541 = __builtin_shufflevector(__s1_541, __s1_541, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __ret_541; \ - __ret_541 = __rev0_541 * __noswap_splatq_laneq_u16(__rev1_541, __p2_541); \ - __ret_541 = __builtin_shufflevector(__ret_541, __ret_541, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_541; \ +#define vmulq_laneq_u16(__p0_633, __p1_633, __p2_633) __extension__ ({ \ + uint16x8_t __s0_633 = __p0_633; \ + uint16x8_t __s1_633 = __p1_633; \ + uint16x8_t __rev0_633; __rev0_633 = __builtin_shufflevector(__s0_633, __s0_633, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_633; __rev1_633 = __builtin_shufflevector(__s1_633, __s1_633, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __ret_633; \ + __ret_633 = __rev0_633 * __noswap_splatq_laneq_u16(__rev1_633, __p2_633); \ + __ret_633 = __builtin_shufflevector(__ret_633, __ret_633, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_633; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_laneq_f64(__p0_542, __p1_542, __p2_542) __extension__ ({ \ - float64x2_t __s0_542 = __p0_542; \ - float64x2_t __s1_542 = __p1_542; \ - float64x2_t __ret_542; \ - __ret_542 = __s0_542 * splatq_laneq_f64(__s1_542, __p2_542); \ - __ret_542; \ +#define vmulq_laneq_f64(__p0_634, __p1_634, __p2_634) __extension__ ({ \ + float64x2_t __s0_634 = __p0_634; \ + float64x2_t __s1_634 = __p1_634; \ + float64x2_t __ret_634; \ + __ret_634 = __s0_634 * splatq_laneq_f64(__s1_634, __p2_634); \ + __ret_634; \ }) #else -#define vmulq_laneq_f64(__p0_543, __p1_543, __p2_543) __extension__ ({ \ - float64x2_t __s0_543 = __p0_543; \ - float64x2_t __s1_543 = __p1_543; \ - float64x2_t __rev0_543; __rev0_543 = __builtin_shufflevector(__s0_543, __s0_543, 1, 0); \ - float64x2_t __rev1_543; __rev1_543 = __builtin_shufflevector(__s1_543, __s1_543, 1, 0); \ - float64x2_t __ret_543; \ - __ret_543 = __rev0_543 * __noswap_splatq_laneq_f64(__rev1_543, __p2_543); \ - __ret_543 = __builtin_shufflevector(__ret_543, __ret_543, 1, 0); \ - __ret_543; \ +#define vmulq_laneq_f64(__p0_635, __p1_635, __p2_635) __extension__ ({ \ + float64x2_t __s0_635 = __p0_635; \ + float64x2_t __s1_635 = __p1_635; \ + float64x2_t __rev0_635; __rev0_635 = __builtin_shufflevector(__s0_635, __s0_635, 1, 0); \ + float64x2_t __rev1_635; __rev1_635 = __builtin_shufflevector(__s1_635, __s1_635, 1, 0); \ + float64x2_t __ret_635; \ + __ret_635 = __rev0_635 * __noswap_splatq_laneq_f64(__rev1_635, __p2_635); \ + __ret_635 = __builtin_shufflevector(__ret_635, __ret_635, 1, 0); \ + __ret_635; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_laneq_f32(__p0_544, __p1_544, __p2_544) __extension__ ({ \ - float32x4_t __s0_544 = __p0_544; \ - float32x4_t __s1_544 = __p1_544; \ - float32x4_t __ret_544; \ - __ret_544 = __s0_544 * splatq_laneq_f32(__s1_544, __p2_544); \ - __ret_544; \ +#define vmulq_laneq_f32(__p0_636, __p1_636, __p2_636) __extension__ ({ \ + float32x4_t __s0_636 = __p0_636; \ + float32x4_t __s1_636 = __p1_636; \ + float32x4_t __ret_636; \ + __ret_636 = __s0_636 * splatq_laneq_f32(__s1_636, __p2_636); \ + __ret_636; \ }) #else -#define vmulq_laneq_f32(__p0_545, __p1_545, __p2_545) __extension__ ({ \ - float32x4_t __s0_545 = __p0_545; \ - float32x4_t __s1_545 = __p1_545; \ - float32x4_t __rev0_545; __rev0_545 = __builtin_shufflevector(__s0_545, __s0_545, 3, 2, 1, 0); \ - float32x4_t __rev1_545; __rev1_545 = __builtin_shufflevector(__s1_545, __s1_545, 3, 2, 1, 0); \ - float32x4_t __ret_545; \ - __ret_545 = __rev0_545 * __noswap_splatq_laneq_f32(__rev1_545, __p2_545); \ - __ret_545 = __builtin_shufflevector(__ret_545, __ret_545, 3, 2, 1, 0); \ - __ret_545; \ +#define vmulq_laneq_f32(__p0_637, __p1_637, __p2_637) __extension__ ({ \ + float32x4_t __s0_637 = __p0_637; \ + float32x4_t __s1_637 = __p1_637; \ + float32x4_t __rev0_637; __rev0_637 = __builtin_shufflevector(__s0_637, __s0_637, 3, 2, 1, 0); \ + float32x4_t __rev1_637; __rev1_637 = __builtin_shufflevector(__s1_637, __s1_637, 3, 2, 1, 0); \ + float32x4_t __ret_637; \ + __ret_637 = __rev0_637 * __noswap_splatq_laneq_f32(__rev1_637, __p2_637); \ + __ret_637 = __builtin_shufflevector(__ret_637, __ret_637, 3, 2, 1, 0); \ + __ret_637; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_laneq_s32(__p0_546, __p1_546, __p2_546) __extension__ ({ \ - int32x4_t __s0_546 = __p0_546; \ - int32x4_t __s1_546 = __p1_546; \ - int32x4_t __ret_546; \ - __ret_546 = __s0_546 * splatq_laneq_s32(__s1_546, __p2_546); \ - __ret_546; \ +#define vmulq_laneq_s32(__p0_638, __p1_638, __p2_638) __extension__ ({ \ + int32x4_t __s0_638 = __p0_638; \ + int32x4_t __s1_638 = __p1_638; \ + int32x4_t __ret_638; \ + __ret_638 = __s0_638 * splatq_laneq_s32(__s1_638, __p2_638); \ + __ret_638; \ }) #else -#define vmulq_laneq_s32(__p0_547, __p1_547, __p2_547) __extension__ ({ \ - int32x4_t __s0_547 = __p0_547; \ - int32x4_t __s1_547 = __p1_547; \ - int32x4_t __rev0_547; __rev0_547 = __builtin_shufflevector(__s0_547, __s0_547, 3, 2, 1, 0); \ - int32x4_t __rev1_547; __rev1_547 = __builtin_shufflevector(__s1_547, __s1_547, 3, 2, 1, 0); \ - int32x4_t __ret_547; \ - __ret_547 = __rev0_547 * __noswap_splatq_laneq_s32(__rev1_547, __p2_547); \ - __ret_547 = __builtin_shufflevector(__ret_547, __ret_547, 3, 2, 1, 0); \ - __ret_547; \ +#define vmulq_laneq_s32(__p0_639, __p1_639, __p2_639) __extension__ ({ \ + int32x4_t __s0_639 = __p0_639; \ + int32x4_t __s1_639 = __p1_639; \ + int32x4_t __rev0_639; __rev0_639 = __builtin_shufflevector(__s0_639, __s0_639, 3, 2, 1, 0); \ + int32x4_t __rev1_639; __rev1_639 = __builtin_shufflevector(__s1_639, __s1_639, 3, 2, 1, 0); \ + int32x4_t __ret_639; \ + __ret_639 = __rev0_639 * __noswap_splatq_laneq_s32(__rev1_639, __p2_639); \ + __ret_639 = __builtin_shufflevector(__ret_639, __ret_639, 3, 2, 1, 0); \ + __ret_639; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulq_laneq_s16(__p0_548, __p1_548, __p2_548) __extension__ ({ \ - int16x8_t __s0_548 = __p0_548; \ - int16x8_t __s1_548 = __p1_548; \ - int16x8_t __ret_548; \ - __ret_548 = __s0_548 * splatq_laneq_s16(__s1_548, __p2_548); \ - __ret_548; \ +#define vmulq_laneq_s16(__p0_640, __p1_640, __p2_640) __extension__ ({ \ + int16x8_t __s0_640 = __p0_640; \ + int16x8_t __s1_640 = __p1_640; \ + int16x8_t __ret_640; \ + __ret_640 = __s0_640 * splatq_laneq_s16(__s1_640, __p2_640); \ + __ret_640; \ }) #else -#define vmulq_laneq_s16(__p0_549, __p1_549, __p2_549) __extension__ ({ \ - int16x8_t __s0_549 = __p0_549; \ - int16x8_t __s1_549 = __p1_549; \ - int16x8_t __rev0_549; __rev0_549 = __builtin_shufflevector(__s0_549, __s0_549, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_549; __rev1_549 = __builtin_shufflevector(__s1_549, __s1_549, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_549; \ - __ret_549 = __rev0_549 * __noswap_splatq_laneq_s16(__rev1_549, __p2_549); \ - __ret_549 = __builtin_shufflevector(__ret_549, __ret_549, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_549; \ +#define vmulq_laneq_s16(__p0_641, __p1_641, __p2_641) __extension__ ({ \ + int16x8_t __s0_641 = __p0_641; \ + int16x8_t __s1_641 = __p1_641; \ + int16x8_t __rev0_641; __rev0_641 = __builtin_shufflevector(__s0_641, __s0_641, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_641; __rev1_641 = __builtin_shufflevector(__s1_641, __s1_641, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_641; \ + __ret_641 = __rev0_641 * __noswap_splatq_laneq_s16(__rev1_641, __p2_641); \ + __ret_641 = __builtin_shufflevector(__ret_641, __ret_641, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_641; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmul_laneq_u32(__p0_550, __p1_550, __p2_550) __extension__ ({ \ - uint32x2_t __s0_550 = __p0_550; \ - uint32x4_t __s1_550 = __p1_550; \ - uint32x2_t __ret_550; \ - __ret_550 = __s0_550 * splat_laneq_u32(__s1_550, __p2_550); \ - __ret_550; \ +#define vmul_laneq_u32(__p0_642, __p1_642, __p2_642) __extension__ ({ \ + uint32x2_t __s0_642 = __p0_642; \ + uint32x4_t __s1_642 = __p1_642; \ + uint32x2_t __ret_642; \ + __ret_642 = __s0_642 * splat_laneq_u32(__s1_642, __p2_642); \ + __ret_642; \ }) #else -#define vmul_laneq_u32(__p0_551, __p1_551, __p2_551) __extension__ ({ \ - uint32x2_t __s0_551 = __p0_551; \ - uint32x4_t __s1_551 = __p1_551; \ - uint32x2_t __rev0_551; __rev0_551 = __builtin_shufflevector(__s0_551, __s0_551, 1, 0); \ - uint32x4_t __rev1_551; __rev1_551 = __builtin_shufflevector(__s1_551, __s1_551, 3, 2, 1, 0); \ - uint32x2_t __ret_551; \ - __ret_551 = __rev0_551 * __noswap_splat_laneq_u32(__rev1_551, __p2_551); \ - __ret_551 = __builtin_shufflevector(__ret_551, __ret_551, 1, 0); \ - __ret_551; \ +#define vmul_laneq_u32(__p0_643, __p1_643, __p2_643) __extension__ ({ \ + uint32x2_t __s0_643 = __p0_643; \ + uint32x4_t __s1_643 = __p1_643; \ + uint32x2_t __rev0_643; __rev0_643 = __builtin_shufflevector(__s0_643, __s0_643, 1, 0); \ + uint32x4_t __rev1_643; __rev1_643 = __builtin_shufflevector(__s1_643, __s1_643, 3, 2, 1, 0); \ + uint32x2_t __ret_643; \ + __ret_643 = __rev0_643 * __noswap_splat_laneq_u32(__rev1_643, __p2_643); \ + __ret_643 = __builtin_shufflevector(__ret_643, __ret_643, 1, 0); \ + __ret_643; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmul_laneq_u16(__p0_552, __p1_552, __p2_552) __extension__ ({ \ - uint16x4_t __s0_552 = __p0_552; \ - uint16x8_t __s1_552 = __p1_552; \ - uint16x4_t __ret_552; \ - __ret_552 = __s0_552 * splat_laneq_u16(__s1_552, __p2_552); \ - __ret_552; \ +#define vmul_laneq_u16(__p0_644, __p1_644, __p2_644) __extension__ ({ \ + uint16x4_t __s0_644 = __p0_644; \ + uint16x8_t __s1_644 = __p1_644; \ + uint16x4_t __ret_644; \ + __ret_644 = __s0_644 * splat_laneq_u16(__s1_644, __p2_644); \ + __ret_644; \ }) #else -#define vmul_laneq_u16(__p0_553, __p1_553, __p2_553) __extension__ ({ \ - uint16x4_t __s0_553 = __p0_553; \ - uint16x8_t __s1_553 = __p1_553; \ - uint16x4_t __rev0_553; __rev0_553 = __builtin_shufflevector(__s0_553, __s0_553, 3, 2, 1, 0); \ - uint16x8_t __rev1_553; __rev1_553 = __builtin_shufflevector(__s1_553, __s1_553, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __ret_553; \ - __ret_553 = __rev0_553 * __noswap_splat_laneq_u16(__rev1_553, __p2_553); \ - __ret_553 = __builtin_shufflevector(__ret_553, __ret_553, 3, 2, 1, 0); \ - __ret_553; \ +#define vmul_laneq_u16(__p0_645, __p1_645, __p2_645) __extension__ ({ \ + uint16x4_t __s0_645 = __p0_645; \ + uint16x8_t __s1_645 = __p1_645; \ + uint16x4_t __rev0_645; __rev0_645 = __builtin_shufflevector(__s0_645, __s0_645, 3, 2, 1, 0); \ + uint16x8_t __rev1_645; __rev1_645 = __builtin_shufflevector(__s1_645, __s1_645, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __ret_645; \ + __ret_645 = __rev0_645 * __noswap_splat_laneq_u16(__rev1_645, __p2_645); \ + __ret_645 = __builtin_shufflevector(__ret_645, __ret_645, 3, 2, 1, 0); \ + __ret_645; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmul_laneq_f32(__p0_554, __p1_554, __p2_554) __extension__ ({ \ - float32x2_t __s0_554 = __p0_554; \ - float32x4_t __s1_554 = __p1_554; \ - float32x2_t __ret_554; \ - __ret_554 = __s0_554 * splat_laneq_f32(__s1_554, __p2_554); \ - __ret_554; \ +#define vmul_laneq_f32(__p0_646, __p1_646, __p2_646) __extension__ ({ \ + float32x2_t __s0_646 = __p0_646; \ + float32x4_t __s1_646 = __p1_646; \ + float32x2_t __ret_646; \ + __ret_646 = __s0_646 * splat_laneq_f32(__s1_646, __p2_646); \ + __ret_646; \ }) #else -#define vmul_laneq_f32(__p0_555, __p1_555, __p2_555) __extension__ ({ \ - float32x2_t __s0_555 = __p0_555; \ - float32x4_t __s1_555 = __p1_555; \ - float32x2_t __rev0_555; __rev0_555 = __builtin_shufflevector(__s0_555, __s0_555, 1, 0); \ - float32x4_t __rev1_555; __rev1_555 = __builtin_shufflevector(__s1_555, __s1_555, 3, 2, 1, 0); \ - float32x2_t __ret_555; \ - __ret_555 = __rev0_555 * __noswap_splat_laneq_f32(__rev1_555, __p2_555); \ - __ret_555 = __builtin_shufflevector(__ret_555, __ret_555, 1, 0); \ - __ret_555; \ +#define vmul_laneq_f32(__p0_647, __p1_647, __p2_647) __extension__ ({ \ + float32x2_t __s0_647 = __p0_647; \ + float32x4_t __s1_647 = __p1_647; \ + float32x2_t __rev0_647; __rev0_647 = __builtin_shufflevector(__s0_647, __s0_647, 1, 0); \ + float32x4_t __rev1_647; __rev1_647 = __builtin_shufflevector(__s1_647, __s1_647, 3, 2, 1, 0); \ + float32x2_t __ret_647; \ + __ret_647 = __rev0_647 * __noswap_splat_laneq_f32(__rev1_647, __p2_647); \ + __ret_647 = __builtin_shufflevector(__ret_647, __ret_647, 1, 0); \ + __ret_647; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmul_laneq_s32(__p0_556, __p1_556, __p2_556) __extension__ ({ \ - int32x2_t __s0_556 = __p0_556; \ - int32x4_t __s1_556 = __p1_556; \ - int32x2_t __ret_556; \ - __ret_556 = __s0_556 * splat_laneq_s32(__s1_556, __p2_556); \ - __ret_556; \ +#define vmul_laneq_s32(__p0_648, __p1_648, __p2_648) __extension__ ({ \ + int32x2_t __s0_648 = __p0_648; \ + int32x4_t __s1_648 = __p1_648; \ + int32x2_t __ret_648; \ + __ret_648 = __s0_648 * splat_laneq_s32(__s1_648, __p2_648); \ + __ret_648; \ }) #else -#define vmul_laneq_s32(__p0_557, __p1_557, __p2_557) __extension__ ({ \ - int32x2_t __s0_557 = __p0_557; \ - int32x4_t __s1_557 = __p1_557; \ - int32x2_t __rev0_557; __rev0_557 = __builtin_shufflevector(__s0_557, __s0_557, 1, 0); \ - int32x4_t __rev1_557; __rev1_557 = __builtin_shufflevector(__s1_557, __s1_557, 3, 2, 1, 0); \ - int32x2_t __ret_557; \ - __ret_557 = __rev0_557 * __noswap_splat_laneq_s32(__rev1_557, __p2_557); \ - __ret_557 = __builtin_shufflevector(__ret_557, __ret_557, 1, 0); \ - __ret_557; \ +#define vmul_laneq_s32(__p0_649, __p1_649, __p2_649) __extension__ ({ \ + int32x2_t __s0_649 = __p0_649; \ + int32x4_t __s1_649 = __p1_649; \ + int32x2_t __rev0_649; __rev0_649 = __builtin_shufflevector(__s0_649, __s0_649, 1, 0); \ + int32x4_t __rev1_649; __rev1_649 = __builtin_shufflevector(__s1_649, __s1_649, 3, 2, 1, 0); \ + int32x2_t __ret_649; \ + __ret_649 = __rev0_649 * __noswap_splat_laneq_s32(__rev1_649, __p2_649); \ + __ret_649 = __builtin_shufflevector(__ret_649, __ret_649, 1, 0); \ + __ret_649; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmul_laneq_s16(__p0_558, __p1_558, __p2_558) __extension__ ({ \ - int16x4_t __s0_558 = __p0_558; \ - int16x8_t __s1_558 = __p1_558; \ - int16x4_t __ret_558; \ - __ret_558 = __s0_558 * splat_laneq_s16(__s1_558, __p2_558); \ - __ret_558; \ +#define vmul_laneq_s16(__p0_650, __p1_650, __p2_650) __extension__ ({ \ + int16x4_t __s0_650 = __p0_650; \ + int16x8_t __s1_650 = __p1_650; \ + int16x4_t __ret_650; \ + __ret_650 = __s0_650 * splat_laneq_s16(__s1_650, __p2_650); \ + __ret_650; \ }) #else -#define vmul_laneq_s16(__p0_559, __p1_559, __p2_559) __extension__ ({ \ - int16x4_t __s0_559 = __p0_559; \ - int16x8_t __s1_559 = __p1_559; \ - int16x4_t __rev0_559; __rev0_559 = __builtin_shufflevector(__s0_559, __s0_559, 3, 2, 1, 0); \ - int16x8_t __rev1_559; __rev1_559 = __builtin_shufflevector(__s1_559, __s1_559, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __ret_559; \ - __ret_559 = __rev0_559 * __noswap_splat_laneq_s16(__rev1_559, __p2_559); \ - __ret_559 = __builtin_shufflevector(__ret_559, __ret_559, 3, 2, 1, 0); \ - __ret_559; \ +#define vmul_laneq_s16(__p0_651, __p1_651, __p2_651) __extension__ ({ \ + int16x4_t __s0_651 = __p0_651; \ + int16x8_t __s1_651 = __p1_651; \ + int16x4_t __rev0_651; __rev0_651 = __builtin_shufflevector(__s0_651, __s0_651, 3, 2, 1, 0); \ + int16x8_t __rev1_651; __rev1_651 = __builtin_shufflevector(__s1_651, __s1_651, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __ret_651; \ + __ret_651 = __rev0_651 * __noswap_splat_laneq_s16(__rev1_651, __p2_651); \ + __ret_651 = __builtin_shufflevector(__ret_651, __ret_651, 3, 2, 1, 0); \ + __ret_651; \ }) #endif @@ -54471,170 +56141,170 @@ __ai poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_lane_u32(__p0_560, __p1_560, __p2_560) __extension__ ({ \ - uint32x4_t __s0_560 = __p0_560; \ - uint32x2_t __s1_560 = __p1_560; \ - uint64x2_t __ret_560; \ - __ret_560 = vmull_u32(vget_high_u32(__s0_560), splat_lane_u32(__s1_560, __p2_560)); \ - __ret_560; \ +#define vmull_high_lane_u32(__p0_652, __p1_652, __p2_652) __extension__ ({ \ + uint32x4_t __s0_652 = __p0_652; \ + uint32x2_t __s1_652 = __p1_652; \ + uint64x2_t __ret_652; \ + __ret_652 = vmull_u32(vget_high_u32(__s0_652), splat_lane_u32(__s1_652, __p2_652)); \ + __ret_652; \ }) #else -#define vmull_high_lane_u32(__p0_561, __p1_561, __p2_561) __extension__ ({ \ - uint32x4_t __s0_561 = __p0_561; \ - uint32x2_t __s1_561 = __p1_561; \ - uint32x4_t __rev0_561; __rev0_561 = __builtin_shufflevector(__s0_561, __s0_561, 3, 2, 1, 0); \ - uint32x2_t __rev1_561; __rev1_561 = __builtin_shufflevector(__s1_561, __s1_561, 1, 0); \ - uint64x2_t __ret_561; \ - __ret_561 = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0_561), __noswap_splat_lane_u32(__rev1_561, __p2_561)); \ - __ret_561 = __builtin_shufflevector(__ret_561, __ret_561, 1, 0); \ - __ret_561; \ +#define vmull_high_lane_u32(__p0_653, __p1_653, __p2_653) __extension__ ({ \ + uint32x4_t __s0_653 = __p0_653; \ + uint32x2_t __s1_653 = __p1_653; \ + uint32x4_t __rev0_653; __rev0_653 = __builtin_shufflevector(__s0_653, __s0_653, 3, 2, 1, 0); \ + uint32x2_t __rev1_653; __rev1_653 = __builtin_shufflevector(__s1_653, __s1_653, 1, 0); \ + uint64x2_t __ret_653; \ + __ret_653 = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0_653), __noswap_splat_lane_u32(__rev1_653, __p2_653)); \ + __ret_653 = __builtin_shufflevector(__ret_653, __ret_653, 1, 0); \ + __ret_653; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_lane_u16(__p0_562, __p1_562, __p2_562) __extension__ ({ \ - uint16x8_t __s0_562 = __p0_562; \ - uint16x4_t __s1_562 = __p1_562; \ - uint32x4_t __ret_562; \ - __ret_562 = vmull_u16(vget_high_u16(__s0_562), splat_lane_u16(__s1_562, __p2_562)); \ - __ret_562; \ +#define vmull_high_lane_u16(__p0_654, __p1_654, __p2_654) __extension__ ({ \ + uint16x8_t __s0_654 = __p0_654; \ + uint16x4_t __s1_654 = __p1_654; \ + uint32x4_t __ret_654; \ + __ret_654 = vmull_u16(vget_high_u16(__s0_654), splat_lane_u16(__s1_654, __p2_654)); \ + __ret_654; \ }) #else -#define vmull_high_lane_u16(__p0_563, __p1_563, __p2_563) __extension__ ({ \ - uint16x8_t __s0_563 = __p0_563; \ - uint16x4_t __s1_563 = __p1_563; \ - uint16x8_t __rev0_563; __rev0_563 = __builtin_shufflevector(__s0_563, __s0_563, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x4_t __rev1_563; __rev1_563 = __builtin_shufflevector(__s1_563, __s1_563, 3, 2, 1, 0); \ - uint32x4_t __ret_563; \ - __ret_563 = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0_563), __noswap_splat_lane_u16(__rev1_563, __p2_563)); \ - __ret_563 = __builtin_shufflevector(__ret_563, __ret_563, 3, 2, 1, 0); \ - __ret_563; \ +#define vmull_high_lane_u16(__p0_655, __p1_655, __p2_655) __extension__ ({ \ + uint16x8_t __s0_655 = __p0_655; \ + uint16x4_t __s1_655 = __p1_655; \ + uint16x8_t __rev0_655; __rev0_655 = __builtin_shufflevector(__s0_655, __s0_655, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x4_t __rev1_655; __rev1_655 = __builtin_shufflevector(__s1_655, __s1_655, 3, 2, 1, 0); \ + uint32x4_t __ret_655; \ + __ret_655 = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0_655), __noswap_splat_lane_u16(__rev1_655, __p2_655)); \ + __ret_655 = __builtin_shufflevector(__ret_655, __ret_655, 3, 2, 1, 0); \ + __ret_655; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_lane_s32(__p0_564, __p1_564, __p2_564) __extension__ ({ \ - int32x4_t __s0_564 = __p0_564; \ - int32x2_t __s1_564 = __p1_564; \ - int64x2_t __ret_564; \ - __ret_564 = vmull_s32(vget_high_s32(__s0_564), splat_lane_s32(__s1_564, __p2_564)); \ - __ret_564; \ +#define vmull_high_lane_s32(__p0_656, __p1_656, __p2_656) __extension__ ({ \ + int32x4_t __s0_656 = __p0_656; \ + int32x2_t __s1_656 = __p1_656; \ + int64x2_t __ret_656; \ + __ret_656 = vmull_s32(vget_high_s32(__s0_656), splat_lane_s32(__s1_656, __p2_656)); \ + __ret_656; \ }) #else -#define vmull_high_lane_s32(__p0_565, __p1_565, __p2_565) __extension__ ({ \ - int32x4_t __s0_565 = __p0_565; \ - int32x2_t __s1_565 = __p1_565; \ - int32x4_t __rev0_565; __rev0_565 = __builtin_shufflevector(__s0_565, __s0_565, 3, 2, 1, 0); \ - int32x2_t __rev1_565; __rev1_565 = __builtin_shufflevector(__s1_565, __s1_565, 1, 0); \ - int64x2_t __ret_565; \ - __ret_565 = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0_565), __noswap_splat_lane_s32(__rev1_565, __p2_565)); \ - __ret_565 = __builtin_shufflevector(__ret_565, __ret_565, 1, 0); \ - __ret_565; \ +#define vmull_high_lane_s32(__p0_657, __p1_657, __p2_657) __extension__ ({ \ + int32x4_t __s0_657 = __p0_657; \ + int32x2_t __s1_657 = __p1_657; \ + int32x4_t __rev0_657; __rev0_657 = __builtin_shufflevector(__s0_657, __s0_657, 3, 2, 1, 0); \ + int32x2_t __rev1_657; __rev1_657 = __builtin_shufflevector(__s1_657, __s1_657, 1, 0); \ + int64x2_t __ret_657; \ + __ret_657 = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0_657), __noswap_splat_lane_s32(__rev1_657, __p2_657)); \ + __ret_657 = __builtin_shufflevector(__ret_657, __ret_657, 1, 0); \ + __ret_657; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_lane_s16(__p0_566, __p1_566, __p2_566) __extension__ ({ \ - int16x8_t __s0_566 = __p0_566; \ - int16x4_t __s1_566 = __p1_566; \ - int32x4_t __ret_566; \ - __ret_566 = vmull_s16(vget_high_s16(__s0_566), splat_lane_s16(__s1_566, __p2_566)); \ - __ret_566; \ +#define vmull_high_lane_s16(__p0_658, __p1_658, __p2_658) __extension__ ({ \ + int16x8_t __s0_658 = __p0_658; \ + int16x4_t __s1_658 = __p1_658; \ + int32x4_t __ret_658; \ + __ret_658 = vmull_s16(vget_high_s16(__s0_658), splat_lane_s16(__s1_658, __p2_658)); \ + __ret_658; \ }) #else -#define vmull_high_lane_s16(__p0_567, __p1_567, __p2_567) __extension__ ({ \ - int16x8_t __s0_567 = __p0_567; \ - int16x4_t __s1_567 = __p1_567; \ - int16x8_t __rev0_567; __rev0_567 = __builtin_shufflevector(__s0_567, __s0_567, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev1_567; __rev1_567 = __builtin_shufflevector(__s1_567, __s1_567, 3, 2, 1, 0); \ - int32x4_t __ret_567; \ - __ret_567 = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0_567), __noswap_splat_lane_s16(__rev1_567, __p2_567)); \ - __ret_567 = __builtin_shufflevector(__ret_567, __ret_567, 3, 2, 1, 0); \ - __ret_567; \ +#define vmull_high_lane_s16(__p0_659, __p1_659, __p2_659) __extension__ ({ \ + int16x8_t __s0_659 = __p0_659; \ + int16x4_t __s1_659 = __p1_659; \ + int16x8_t __rev0_659; __rev0_659 = __builtin_shufflevector(__s0_659, __s0_659, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev1_659; __rev1_659 = __builtin_shufflevector(__s1_659, __s1_659, 3, 2, 1, 0); \ + int32x4_t __ret_659; \ + __ret_659 = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0_659), __noswap_splat_lane_s16(__rev1_659, __p2_659)); \ + __ret_659 = __builtin_shufflevector(__ret_659, __ret_659, 3, 2, 1, 0); \ + __ret_659; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_laneq_u32(__p0_568, __p1_568, __p2_568) __extension__ ({ \ - uint32x4_t __s0_568 = __p0_568; \ - uint32x4_t __s1_568 = __p1_568; \ - uint64x2_t __ret_568; \ - __ret_568 = vmull_u32(vget_high_u32(__s0_568), splat_laneq_u32(__s1_568, __p2_568)); \ - __ret_568; \ +#define vmull_high_laneq_u32(__p0_660, __p1_660, __p2_660) __extension__ ({ \ + uint32x4_t __s0_660 = __p0_660; \ + uint32x4_t __s1_660 = __p1_660; \ + uint64x2_t __ret_660; \ + __ret_660 = vmull_u32(vget_high_u32(__s0_660), splat_laneq_u32(__s1_660, __p2_660)); \ + __ret_660; \ }) #else -#define vmull_high_laneq_u32(__p0_569, __p1_569, __p2_569) __extension__ ({ \ - uint32x4_t __s0_569 = __p0_569; \ - uint32x4_t __s1_569 = __p1_569; \ - uint32x4_t __rev0_569; __rev0_569 = __builtin_shufflevector(__s0_569, __s0_569, 3, 2, 1, 0); \ - uint32x4_t __rev1_569; __rev1_569 = __builtin_shufflevector(__s1_569, __s1_569, 3, 2, 1, 0); \ - uint64x2_t __ret_569; \ - __ret_569 = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0_569), __noswap_splat_laneq_u32(__rev1_569, __p2_569)); \ - __ret_569 = __builtin_shufflevector(__ret_569, __ret_569, 1, 0); \ - __ret_569; \ +#define vmull_high_laneq_u32(__p0_661, __p1_661, __p2_661) __extension__ ({ \ + uint32x4_t __s0_661 = __p0_661; \ + uint32x4_t __s1_661 = __p1_661; \ + uint32x4_t __rev0_661; __rev0_661 = __builtin_shufflevector(__s0_661, __s0_661, 3, 2, 1, 0); \ + uint32x4_t __rev1_661; __rev1_661 = __builtin_shufflevector(__s1_661, __s1_661, 3, 2, 1, 0); \ + uint64x2_t __ret_661; \ + __ret_661 = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0_661), __noswap_splat_laneq_u32(__rev1_661, __p2_661)); \ + __ret_661 = __builtin_shufflevector(__ret_661, __ret_661, 1, 0); \ + __ret_661; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_laneq_u16(__p0_570, __p1_570, __p2_570) __extension__ ({ \ - uint16x8_t __s0_570 = __p0_570; \ - uint16x8_t __s1_570 = __p1_570; \ - uint32x4_t __ret_570; \ - __ret_570 = vmull_u16(vget_high_u16(__s0_570), splat_laneq_u16(__s1_570, __p2_570)); \ - __ret_570; \ +#define vmull_high_laneq_u16(__p0_662, __p1_662, __p2_662) __extension__ ({ \ + uint16x8_t __s0_662 = __p0_662; \ + uint16x8_t __s1_662 = __p1_662; \ + uint32x4_t __ret_662; \ + __ret_662 = vmull_u16(vget_high_u16(__s0_662), splat_laneq_u16(__s1_662, __p2_662)); \ + __ret_662; \ }) #else -#define vmull_high_laneq_u16(__p0_571, __p1_571, __p2_571) __extension__ ({ \ - uint16x8_t __s0_571 = __p0_571; \ - uint16x8_t __s1_571 = __p1_571; \ - uint16x8_t __rev0_571; __rev0_571 = __builtin_shufflevector(__s0_571, __s0_571, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_571; __rev1_571 = __builtin_shufflevector(__s1_571, __s1_571, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_571; \ - __ret_571 = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0_571), __noswap_splat_laneq_u16(__rev1_571, __p2_571)); \ - __ret_571 = __builtin_shufflevector(__ret_571, __ret_571, 3, 2, 1, 0); \ - __ret_571; \ +#define vmull_high_laneq_u16(__p0_663, __p1_663, __p2_663) __extension__ ({ \ + uint16x8_t __s0_663 = __p0_663; \ + uint16x8_t __s1_663 = __p1_663; \ + uint16x8_t __rev0_663; __rev0_663 = __builtin_shufflevector(__s0_663, __s0_663, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_663; __rev1_663 = __builtin_shufflevector(__s1_663, __s1_663, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_663; \ + __ret_663 = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0_663), __noswap_splat_laneq_u16(__rev1_663, __p2_663)); \ + __ret_663 = __builtin_shufflevector(__ret_663, __ret_663, 3, 2, 1, 0); \ + __ret_663; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_laneq_s32(__p0_572, __p1_572, __p2_572) __extension__ ({ \ - int32x4_t __s0_572 = __p0_572; \ - int32x4_t __s1_572 = __p1_572; \ - int64x2_t __ret_572; \ - __ret_572 = vmull_s32(vget_high_s32(__s0_572), splat_laneq_s32(__s1_572, __p2_572)); \ - __ret_572; \ +#define vmull_high_laneq_s32(__p0_664, __p1_664, __p2_664) __extension__ ({ \ + int32x4_t __s0_664 = __p0_664; \ + int32x4_t __s1_664 = __p1_664; \ + int64x2_t __ret_664; \ + __ret_664 = vmull_s32(vget_high_s32(__s0_664), splat_laneq_s32(__s1_664, __p2_664)); \ + __ret_664; \ }) #else -#define vmull_high_laneq_s32(__p0_573, __p1_573, __p2_573) __extension__ ({ \ - int32x4_t __s0_573 = __p0_573; \ - int32x4_t __s1_573 = __p1_573; \ - int32x4_t __rev0_573; __rev0_573 = __builtin_shufflevector(__s0_573, __s0_573, 3, 2, 1, 0); \ - int32x4_t __rev1_573; __rev1_573 = __builtin_shufflevector(__s1_573, __s1_573, 3, 2, 1, 0); \ - int64x2_t __ret_573; \ - __ret_573 = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0_573), __noswap_splat_laneq_s32(__rev1_573, __p2_573)); \ - __ret_573 = __builtin_shufflevector(__ret_573, __ret_573, 1, 0); \ - __ret_573; \ +#define vmull_high_laneq_s32(__p0_665, __p1_665, __p2_665) __extension__ ({ \ + int32x4_t __s0_665 = __p0_665; \ + int32x4_t __s1_665 = __p1_665; \ + int32x4_t __rev0_665; __rev0_665 = __builtin_shufflevector(__s0_665, __s0_665, 3, 2, 1, 0); \ + int32x4_t __rev1_665; __rev1_665 = __builtin_shufflevector(__s1_665, __s1_665, 3, 2, 1, 0); \ + int64x2_t __ret_665; \ + __ret_665 = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0_665), __noswap_splat_laneq_s32(__rev1_665, __p2_665)); \ + __ret_665 = __builtin_shufflevector(__ret_665, __ret_665, 1, 0); \ + __ret_665; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_high_laneq_s16(__p0_574, __p1_574, __p2_574) __extension__ ({ \ - int16x8_t __s0_574 = __p0_574; \ - int16x8_t __s1_574 = __p1_574; \ - int32x4_t __ret_574; \ - __ret_574 = vmull_s16(vget_high_s16(__s0_574), splat_laneq_s16(__s1_574, __p2_574)); \ - __ret_574; \ +#define vmull_high_laneq_s16(__p0_666, __p1_666, __p2_666) __extension__ ({ \ + int16x8_t __s0_666 = __p0_666; \ + int16x8_t __s1_666 = __p1_666; \ + int32x4_t __ret_666; \ + __ret_666 = vmull_s16(vget_high_s16(__s0_666), splat_laneq_s16(__s1_666, __p2_666)); \ + __ret_666; \ }) #else -#define vmull_high_laneq_s16(__p0_575, __p1_575, __p2_575) __extension__ ({ \ - int16x8_t __s0_575 = __p0_575; \ - int16x8_t __s1_575 = __p1_575; \ - int16x8_t __rev0_575; __rev0_575 = __builtin_shufflevector(__s0_575, __s0_575, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_575; __rev1_575 = __builtin_shufflevector(__s1_575, __s1_575, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_575; \ - __ret_575 = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0_575), __noswap_splat_laneq_s16(__rev1_575, __p2_575)); \ - __ret_575 = __builtin_shufflevector(__ret_575, __ret_575, 3, 2, 1, 0); \ - __ret_575; \ +#define vmull_high_laneq_s16(__p0_667, __p1_667, __p2_667) __extension__ ({ \ + int16x8_t __s0_667 = __p0_667; \ + int16x8_t __s1_667 = __p1_667; \ + int16x8_t __rev0_667; __rev0_667 = __builtin_shufflevector(__s0_667, __s0_667, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_667; __rev1_667 = __builtin_shufflevector(__s1_667, __s1_667, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_667; \ + __ret_667 = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0_667), __noswap_splat_laneq_s16(__rev1_667, __p2_667)); \ + __ret_667 = __builtin_shufflevector(__ret_667, __ret_667, 3, 2, 1, 0); \ + __ret_667; \ }) #endif @@ -54703,86 +56373,86 @@ __ai int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_laneq_u32(__p0_576, __p1_576, __p2_576) __extension__ ({ \ - uint32x2_t __s0_576 = __p0_576; \ - uint32x4_t __s1_576 = __p1_576; \ - uint64x2_t __ret_576; \ - __ret_576 = vmull_u32(__s0_576, splat_laneq_u32(__s1_576, __p2_576)); \ - __ret_576; \ +#define vmull_laneq_u32(__p0_668, __p1_668, __p2_668) __extension__ ({ \ + uint32x2_t __s0_668 = __p0_668; \ + uint32x4_t __s1_668 = __p1_668; \ + uint64x2_t __ret_668; \ + __ret_668 = vmull_u32(__s0_668, splat_laneq_u32(__s1_668, __p2_668)); \ + __ret_668; \ }) #else -#define vmull_laneq_u32(__p0_577, __p1_577, __p2_577) __extension__ ({ \ - uint32x2_t __s0_577 = __p0_577; \ - uint32x4_t __s1_577 = __p1_577; \ - uint32x2_t __rev0_577; __rev0_577 = __builtin_shufflevector(__s0_577, __s0_577, 1, 0); \ - uint32x4_t __rev1_577; __rev1_577 = __builtin_shufflevector(__s1_577, __s1_577, 3, 2, 1, 0); \ - uint64x2_t __ret_577; \ - __ret_577 = __noswap_vmull_u32(__rev0_577, __noswap_splat_laneq_u32(__rev1_577, __p2_577)); \ - __ret_577 = __builtin_shufflevector(__ret_577, __ret_577, 1, 0); \ - __ret_577; \ +#define vmull_laneq_u32(__p0_669, __p1_669, __p2_669) __extension__ ({ \ + uint32x2_t __s0_669 = __p0_669; \ + uint32x4_t __s1_669 = __p1_669; \ + uint32x2_t __rev0_669; __rev0_669 = __builtin_shufflevector(__s0_669, __s0_669, 1, 0); \ + uint32x4_t __rev1_669; __rev1_669 = __builtin_shufflevector(__s1_669, __s1_669, 3, 2, 1, 0); \ + uint64x2_t __ret_669; \ + __ret_669 = __noswap_vmull_u32(__rev0_669, __noswap_splat_laneq_u32(__rev1_669, __p2_669)); \ + __ret_669 = __builtin_shufflevector(__ret_669, __ret_669, 1, 0); \ + __ret_669; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_laneq_u16(__p0_578, __p1_578, __p2_578) __extension__ ({ \ - uint16x4_t __s0_578 = __p0_578; \ - uint16x8_t __s1_578 = __p1_578; \ - uint32x4_t __ret_578; \ - __ret_578 = vmull_u16(__s0_578, splat_laneq_u16(__s1_578, __p2_578)); \ - __ret_578; \ +#define vmull_laneq_u16(__p0_670, __p1_670, __p2_670) __extension__ ({ \ + uint16x4_t __s0_670 = __p0_670; \ + uint16x8_t __s1_670 = __p1_670; \ + uint32x4_t __ret_670; \ + __ret_670 = vmull_u16(__s0_670, splat_laneq_u16(__s1_670, __p2_670)); \ + __ret_670; \ }) #else -#define vmull_laneq_u16(__p0_579, __p1_579, __p2_579) __extension__ ({ \ - uint16x4_t __s0_579 = __p0_579; \ - uint16x8_t __s1_579 = __p1_579; \ - uint16x4_t __rev0_579; __rev0_579 = __builtin_shufflevector(__s0_579, __s0_579, 3, 2, 1, 0); \ - uint16x8_t __rev1_579; __rev1_579 = __builtin_shufflevector(__s1_579, __s1_579, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_579; \ - __ret_579 = __noswap_vmull_u16(__rev0_579, __noswap_splat_laneq_u16(__rev1_579, __p2_579)); \ - __ret_579 = __builtin_shufflevector(__ret_579, __ret_579, 3, 2, 1, 0); \ - __ret_579; \ +#define vmull_laneq_u16(__p0_671, __p1_671, __p2_671) __extension__ ({ \ + uint16x4_t __s0_671 = __p0_671; \ + uint16x8_t __s1_671 = __p1_671; \ + uint16x4_t __rev0_671; __rev0_671 = __builtin_shufflevector(__s0_671, __s0_671, 3, 2, 1, 0); \ + uint16x8_t __rev1_671; __rev1_671 = __builtin_shufflevector(__s1_671, __s1_671, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_671; \ + __ret_671 = __noswap_vmull_u16(__rev0_671, __noswap_splat_laneq_u16(__rev1_671, __p2_671)); \ + __ret_671 = __builtin_shufflevector(__ret_671, __ret_671, 3, 2, 1, 0); \ + __ret_671; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_laneq_s32(__p0_580, __p1_580, __p2_580) __extension__ ({ \ - int32x2_t __s0_580 = __p0_580; \ - int32x4_t __s1_580 = __p1_580; \ - int64x2_t __ret_580; \ - __ret_580 = vmull_s32(__s0_580, splat_laneq_s32(__s1_580, __p2_580)); \ - __ret_580; \ +#define vmull_laneq_s32(__p0_672, __p1_672, __p2_672) __extension__ ({ \ + int32x2_t __s0_672 = __p0_672; \ + int32x4_t __s1_672 = __p1_672; \ + int64x2_t __ret_672; \ + __ret_672 = vmull_s32(__s0_672, splat_laneq_s32(__s1_672, __p2_672)); \ + __ret_672; \ }) #else -#define vmull_laneq_s32(__p0_581, __p1_581, __p2_581) __extension__ ({ \ - int32x2_t __s0_581 = __p0_581; \ - int32x4_t __s1_581 = __p1_581; \ - int32x2_t __rev0_581; __rev0_581 = __builtin_shufflevector(__s0_581, __s0_581, 1, 0); \ - int32x4_t __rev1_581; __rev1_581 = __builtin_shufflevector(__s1_581, __s1_581, 3, 2, 1, 0); \ - int64x2_t __ret_581; \ - __ret_581 = __noswap_vmull_s32(__rev0_581, __noswap_splat_laneq_s32(__rev1_581, __p2_581)); \ - __ret_581 = __builtin_shufflevector(__ret_581, __ret_581, 1, 0); \ - __ret_581; \ +#define vmull_laneq_s32(__p0_673, __p1_673, __p2_673) __extension__ ({ \ + int32x2_t __s0_673 = __p0_673; \ + int32x4_t __s1_673 = __p1_673; \ + int32x2_t __rev0_673; __rev0_673 = __builtin_shufflevector(__s0_673, __s0_673, 1, 0); \ + int32x4_t __rev1_673; __rev1_673 = __builtin_shufflevector(__s1_673, __s1_673, 3, 2, 1, 0); \ + int64x2_t __ret_673; \ + __ret_673 = __noswap_vmull_s32(__rev0_673, __noswap_splat_laneq_s32(__rev1_673, __p2_673)); \ + __ret_673 = __builtin_shufflevector(__ret_673, __ret_673, 1, 0); \ + __ret_673; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmull_laneq_s16(__p0_582, __p1_582, __p2_582) __extension__ ({ \ - int16x4_t __s0_582 = __p0_582; \ - int16x8_t __s1_582 = __p1_582; \ - int32x4_t __ret_582; \ - __ret_582 = vmull_s16(__s0_582, splat_laneq_s16(__s1_582, __p2_582)); \ - __ret_582; \ +#define vmull_laneq_s16(__p0_674, __p1_674, __p2_674) __extension__ ({ \ + int16x4_t __s0_674 = __p0_674; \ + int16x8_t __s1_674 = __p1_674; \ + int32x4_t __ret_674; \ + __ret_674 = vmull_s16(__s0_674, splat_laneq_s16(__s1_674, __p2_674)); \ + __ret_674; \ }) #else -#define vmull_laneq_s16(__p0_583, __p1_583, __p2_583) __extension__ ({ \ - int16x4_t __s0_583 = __p0_583; \ - int16x8_t __s1_583 = __p1_583; \ - int16x4_t __rev0_583; __rev0_583 = __builtin_shufflevector(__s0_583, __s0_583, 3, 2, 1, 0); \ - int16x8_t __rev1_583; __rev1_583 = __builtin_shufflevector(__s1_583, __s1_583, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_583; \ - __ret_583 = __noswap_vmull_s16(__rev0_583, __noswap_splat_laneq_s16(__rev1_583, __p2_583)); \ - __ret_583 = __builtin_shufflevector(__ret_583, __ret_583, 3, 2, 1, 0); \ - __ret_583; \ +#define vmull_laneq_s16(__p0_675, __p1_675, __p2_675) __extension__ ({ \ + int16x4_t __s0_675 = __p0_675; \ + int16x8_t __s1_675 = __p1_675; \ + int16x4_t __rev0_675; __rev0_675 = __builtin_shufflevector(__s0_675, __s0_675, 3, 2, 1, 0); \ + int16x8_t __rev1_675; __rev1_675 = __builtin_shufflevector(__s1_675, __s1_675, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_675; \ + __ret_675 = __noswap_vmull_s16(__rev0_675, __noswap_splat_laneq_s16(__rev1_675, __p2_675)); \ + __ret_675 = __builtin_shufflevector(__ret_675, __ret_675, 3, 2, 1, 0); \ + __ret_675; \ }) #endif @@ -54867,192 +56537,192 @@ __ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) { __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1); return __ret; } -#define vmulxd_lane_f64(__p0_584, __p1_584, __p2_584) __extension__ ({ \ - float64_t __s0_584 = __p0_584; \ - float64x1_t __s1_584 = __p1_584; \ - float64_t __ret_584; \ - __ret_584 = vmulxd_f64(__s0_584, vget_lane_f64(__s1_584, __p2_584)); \ - __ret_584; \ +#define vmulxd_lane_f64(__p0_676, __p1_676, __p2_676) __extension__ ({ \ + float64_t __s0_676 = __p0_676; \ + float64x1_t __s1_676 = __p1_676; \ + float64_t __ret_676; \ + __ret_676 = vmulxd_f64(__s0_676, vget_lane_f64(__s1_676, __p2_676)); \ + __ret_676; \ }) #ifdef __LITTLE_ENDIAN__ -#define vmulxs_lane_f32(__p0_585, __p1_585, __p2_585) __extension__ ({ \ - float32_t __s0_585 = __p0_585; \ - float32x2_t __s1_585 = __p1_585; \ - float32_t __ret_585; \ - __ret_585 = vmulxs_f32(__s0_585, vget_lane_f32(__s1_585, __p2_585)); \ - __ret_585; \ +#define vmulxs_lane_f32(__p0_677, __p1_677, __p2_677) __extension__ ({ \ + float32_t __s0_677 = __p0_677; \ + float32x2_t __s1_677 = __p1_677; \ + float32_t __ret_677; \ + __ret_677 = vmulxs_f32(__s0_677, vget_lane_f32(__s1_677, __p2_677)); \ + __ret_677; \ }) #else -#define vmulxs_lane_f32(__p0_586, __p1_586, __p2_586) __extension__ ({ \ - float32_t __s0_586 = __p0_586; \ - float32x2_t __s1_586 = __p1_586; \ - float32x2_t __rev1_586; __rev1_586 = __builtin_shufflevector(__s1_586, __s1_586, 1, 0); \ - float32_t __ret_586; \ - __ret_586 = vmulxs_f32(__s0_586, __noswap_vget_lane_f32(__rev1_586, __p2_586)); \ - __ret_586; \ +#define vmulxs_lane_f32(__p0_678, __p1_678, __p2_678) __extension__ ({ \ + float32_t __s0_678 = __p0_678; \ + float32x2_t __s1_678 = __p1_678; \ + float32x2_t __rev1_678; __rev1_678 = __builtin_shufflevector(__s1_678, __s1_678, 1, 0); \ + float32_t __ret_678; \ + __ret_678 = vmulxs_f32(__s0_678, __noswap_vget_lane_f32(__rev1_678, __p2_678)); \ + __ret_678; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxq_lane_f64(__p0_587, __p1_587, __p2_587) __extension__ ({ \ - float64x2_t __s0_587 = __p0_587; \ - float64x1_t __s1_587 = __p1_587; \ - float64x2_t __ret_587; \ - __ret_587 = vmulxq_f64(__s0_587, splatq_lane_f64(__s1_587, __p2_587)); \ - __ret_587; \ +#define vmulxq_lane_f64(__p0_679, __p1_679, __p2_679) __extension__ ({ \ + float64x2_t __s0_679 = __p0_679; \ + float64x1_t __s1_679 = __p1_679; \ + float64x2_t __ret_679; \ + __ret_679 = vmulxq_f64(__s0_679, splatq_lane_f64(__s1_679, __p2_679)); \ + __ret_679; \ }) #else -#define vmulxq_lane_f64(__p0_588, __p1_588, __p2_588) __extension__ ({ \ - float64x2_t __s0_588 = __p0_588; \ - float64x1_t __s1_588 = __p1_588; \ - float64x2_t __rev0_588; __rev0_588 = __builtin_shufflevector(__s0_588, __s0_588, 1, 0); \ - float64x2_t __ret_588; \ - __ret_588 = __noswap_vmulxq_f64(__rev0_588, __noswap_splatq_lane_f64(__s1_588, __p2_588)); \ - __ret_588 = __builtin_shufflevector(__ret_588, __ret_588, 1, 0); \ - __ret_588; \ +#define vmulxq_lane_f64(__p0_680, __p1_680, __p2_680) __extension__ ({ \ + float64x2_t __s0_680 = __p0_680; \ + float64x1_t __s1_680 = __p1_680; \ + float64x2_t __rev0_680; __rev0_680 = __builtin_shufflevector(__s0_680, __s0_680, 1, 0); \ + float64x2_t __ret_680; \ + __ret_680 = __noswap_vmulxq_f64(__rev0_680, __noswap_splatq_lane_f64(__s1_680, __p2_680)); \ + __ret_680 = __builtin_shufflevector(__ret_680, __ret_680, 1, 0); \ + __ret_680; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxq_lane_f32(__p0_589, __p1_589, __p2_589) __extension__ ({ \ - float32x4_t __s0_589 = __p0_589; \ - float32x2_t __s1_589 = __p1_589; \ - float32x4_t __ret_589; \ - __ret_589 = vmulxq_f32(__s0_589, splatq_lane_f32(__s1_589, __p2_589)); \ - __ret_589; \ +#define vmulxq_lane_f32(__p0_681, __p1_681, __p2_681) __extension__ ({ \ + float32x4_t __s0_681 = __p0_681; \ + float32x2_t __s1_681 = __p1_681; \ + float32x4_t __ret_681; \ + __ret_681 = vmulxq_f32(__s0_681, splatq_lane_f32(__s1_681, __p2_681)); \ + __ret_681; \ }) #else -#define vmulxq_lane_f32(__p0_590, __p1_590, __p2_590) __extension__ ({ \ - float32x4_t __s0_590 = __p0_590; \ - float32x2_t __s1_590 = __p1_590; \ - float32x4_t __rev0_590; __rev0_590 = __builtin_shufflevector(__s0_590, __s0_590, 3, 2, 1, 0); \ - float32x2_t __rev1_590; __rev1_590 = __builtin_shufflevector(__s1_590, __s1_590, 1, 0); \ - float32x4_t __ret_590; \ - __ret_590 = __noswap_vmulxq_f32(__rev0_590, __noswap_splatq_lane_f32(__rev1_590, __p2_590)); \ - __ret_590 = __builtin_shufflevector(__ret_590, __ret_590, 3, 2, 1, 0); \ - __ret_590; \ +#define vmulxq_lane_f32(__p0_682, __p1_682, __p2_682) __extension__ ({ \ + float32x4_t __s0_682 = __p0_682; \ + float32x2_t __s1_682 = __p1_682; \ + float32x4_t __rev0_682; __rev0_682 = __builtin_shufflevector(__s0_682, __s0_682, 3, 2, 1, 0); \ + float32x2_t __rev1_682; __rev1_682 = __builtin_shufflevector(__s1_682, __s1_682, 1, 0); \ + float32x4_t __ret_682; \ + __ret_682 = __noswap_vmulxq_f32(__rev0_682, __noswap_splatq_lane_f32(__rev1_682, __p2_682)); \ + __ret_682 = __builtin_shufflevector(__ret_682, __ret_682, 3, 2, 1, 0); \ + __ret_682; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulx_lane_f32(__p0_591, __p1_591, __p2_591) __extension__ ({ \ - float32x2_t __s0_591 = __p0_591; \ - float32x2_t __s1_591 = __p1_591; \ - float32x2_t __ret_591; \ - __ret_591 = vmulx_f32(__s0_591, splat_lane_f32(__s1_591, __p2_591)); \ - __ret_591; \ +#define vmulx_lane_f32(__p0_683, __p1_683, __p2_683) __extension__ ({ \ + float32x2_t __s0_683 = __p0_683; \ + float32x2_t __s1_683 = __p1_683; \ + float32x2_t __ret_683; \ + __ret_683 = vmulx_f32(__s0_683, splat_lane_f32(__s1_683, __p2_683)); \ + __ret_683; \ }) #else -#define vmulx_lane_f32(__p0_592, __p1_592, __p2_592) __extension__ ({ \ - float32x2_t __s0_592 = __p0_592; \ - float32x2_t __s1_592 = __p1_592; \ - float32x2_t __rev0_592; __rev0_592 = __builtin_shufflevector(__s0_592, __s0_592, 1, 0); \ - float32x2_t __rev1_592; __rev1_592 = __builtin_shufflevector(__s1_592, __s1_592, 1, 0); \ - float32x2_t __ret_592; \ - __ret_592 = __noswap_vmulx_f32(__rev0_592, __noswap_splat_lane_f32(__rev1_592, __p2_592)); \ - __ret_592 = __builtin_shufflevector(__ret_592, __ret_592, 1, 0); \ - __ret_592; \ +#define vmulx_lane_f32(__p0_684, __p1_684, __p2_684) __extension__ ({ \ + float32x2_t __s0_684 = __p0_684; \ + float32x2_t __s1_684 = __p1_684; \ + float32x2_t __rev0_684; __rev0_684 = __builtin_shufflevector(__s0_684, __s0_684, 1, 0); \ + float32x2_t __rev1_684; __rev1_684 = __builtin_shufflevector(__s1_684, __s1_684, 1, 0); \ + float32x2_t __ret_684; \ + __ret_684 = __noswap_vmulx_f32(__rev0_684, __noswap_splat_lane_f32(__rev1_684, __p2_684)); \ + __ret_684 = __builtin_shufflevector(__ret_684, __ret_684, 1, 0); \ + __ret_684; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxd_laneq_f64(__p0_593, __p1_593, __p2_593) __extension__ ({ \ - float64_t __s0_593 = __p0_593; \ - float64x2_t __s1_593 = __p1_593; \ - float64_t __ret_593; \ - __ret_593 = vmulxd_f64(__s0_593, vgetq_lane_f64(__s1_593, __p2_593)); \ - __ret_593; \ +#define vmulxd_laneq_f64(__p0_685, __p1_685, __p2_685) __extension__ ({ \ + float64_t __s0_685 = __p0_685; \ + float64x2_t __s1_685 = __p1_685; \ + float64_t __ret_685; \ + __ret_685 = vmulxd_f64(__s0_685, vgetq_lane_f64(__s1_685, __p2_685)); \ + __ret_685; \ }) #else -#define vmulxd_laneq_f64(__p0_594, __p1_594, __p2_594) __extension__ ({ \ - float64_t __s0_594 = __p0_594; \ - float64x2_t __s1_594 = __p1_594; \ - float64x2_t __rev1_594; __rev1_594 = __builtin_shufflevector(__s1_594, __s1_594, 1, 0); \ - float64_t __ret_594; \ - __ret_594 = vmulxd_f64(__s0_594, __noswap_vgetq_lane_f64(__rev1_594, __p2_594)); \ - __ret_594; \ +#define vmulxd_laneq_f64(__p0_686, __p1_686, __p2_686) __extension__ ({ \ + float64_t __s0_686 = __p0_686; \ + float64x2_t __s1_686 = __p1_686; \ + float64x2_t __rev1_686; __rev1_686 = __builtin_shufflevector(__s1_686, __s1_686, 1, 0); \ + float64_t __ret_686; \ + __ret_686 = vmulxd_f64(__s0_686, __noswap_vgetq_lane_f64(__rev1_686, __p2_686)); \ + __ret_686; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxs_laneq_f32(__p0_595, __p1_595, __p2_595) __extension__ ({ \ - float32_t __s0_595 = __p0_595; \ - float32x4_t __s1_595 = __p1_595; \ - float32_t __ret_595; \ - __ret_595 = vmulxs_f32(__s0_595, vgetq_lane_f32(__s1_595, __p2_595)); \ - __ret_595; \ +#define vmulxs_laneq_f32(__p0_687, __p1_687, __p2_687) __extension__ ({ \ + float32_t __s0_687 = __p0_687; \ + float32x4_t __s1_687 = __p1_687; \ + float32_t __ret_687; \ + __ret_687 = vmulxs_f32(__s0_687, vgetq_lane_f32(__s1_687, __p2_687)); \ + __ret_687; \ }) #else -#define vmulxs_laneq_f32(__p0_596, __p1_596, __p2_596) __extension__ ({ \ - float32_t __s0_596 = __p0_596; \ - float32x4_t __s1_596 = __p1_596; \ - float32x4_t __rev1_596; __rev1_596 = __builtin_shufflevector(__s1_596, __s1_596, 3, 2, 1, 0); \ - float32_t __ret_596; \ - __ret_596 = vmulxs_f32(__s0_596, __noswap_vgetq_lane_f32(__rev1_596, __p2_596)); \ - __ret_596; \ +#define vmulxs_laneq_f32(__p0_688, __p1_688, __p2_688) __extension__ ({ \ + float32_t __s0_688 = __p0_688; \ + float32x4_t __s1_688 = __p1_688; \ + float32x4_t __rev1_688; __rev1_688 = __builtin_shufflevector(__s1_688, __s1_688, 3, 2, 1, 0); \ + float32_t __ret_688; \ + __ret_688 = vmulxs_f32(__s0_688, __noswap_vgetq_lane_f32(__rev1_688, __p2_688)); \ + __ret_688; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxq_laneq_f64(__p0_597, __p1_597, __p2_597) __extension__ ({ \ - float64x2_t __s0_597 = __p0_597; \ - float64x2_t __s1_597 = __p1_597; \ - float64x2_t __ret_597; \ - __ret_597 = vmulxq_f64(__s0_597, splatq_laneq_f64(__s1_597, __p2_597)); \ - __ret_597; \ +#define vmulxq_laneq_f64(__p0_689, __p1_689, __p2_689) __extension__ ({ \ + float64x2_t __s0_689 = __p0_689; \ + float64x2_t __s1_689 = __p1_689; \ + float64x2_t __ret_689; \ + __ret_689 = vmulxq_f64(__s0_689, splatq_laneq_f64(__s1_689, __p2_689)); \ + __ret_689; \ }) #else -#define vmulxq_laneq_f64(__p0_598, __p1_598, __p2_598) __extension__ ({ \ - float64x2_t __s0_598 = __p0_598; \ - float64x2_t __s1_598 = __p1_598; \ - float64x2_t __rev0_598; __rev0_598 = __builtin_shufflevector(__s0_598, __s0_598, 1, 0); \ - float64x2_t __rev1_598; __rev1_598 = __builtin_shufflevector(__s1_598, __s1_598, 1, 0); \ - float64x2_t __ret_598; \ - __ret_598 = __noswap_vmulxq_f64(__rev0_598, __noswap_splatq_laneq_f64(__rev1_598, __p2_598)); \ - __ret_598 = __builtin_shufflevector(__ret_598, __ret_598, 1, 0); \ - __ret_598; \ +#define vmulxq_laneq_f64(__p0_690, __p1_690, __p2_690) __extension__ ({ \ + float64x2_t __s0_690 = __p0_690; \ + float64x2_t __s1_690 = __p1_690; \ + float64x2_t __rev0_690; __rev0_690 = __builtin_shufflevector(__s0_690, __s0_690, 1, 0); \ + float64x2_t __rev1_690; __rev1_690 = __builtin_shufflevector(__s1_690, __s1_690, 1, 0); \ + float64x2_t __ret_690; \ + __ret_690 = __noswap_vmulxq_f64(__rev0_690, __noswap_splatq_laneq_f64(__rev1_690, __p2_690)); \ + __ret_690 = __builtin_shufflevector(__ret_690, __ret_690, 1, 0); \ + __ret_690; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulxq_laneq_f32(__p0_599, __p1_599, __p2_599) __extension__ ({ \ - float32x4_t __s0_599 = __p0_599; \ - float32x4_t __s1_599 = __p1_599; \ - float32x4_t __ret_599; \ - __ret_599 = vmulxq_f32(__s0_599, splatq_laneq_f32(__s1_599, __p2_599)); \ - __ret_599; \ +#define vmulxq_laneq_f32(__p0_691, __p1_691, __p2_691) __extension__ ({ \ + float32x4_t __s0_691 = __p0_691; \ + float32x4_t __s1_691 = __p1_691; \ + float32x4_t __ret_691; \ + __ret_691 = vmulxq_f32(__s0_691, splatq_laneq_f32(__s1_691, __p2_691)); \ + __ret_691; \ }) #else -#define vmulxq_laneq_f32(__p0_600, __p1_600, __p2_600) __extension__ ({ \ - float32x4_t __s0_600 = __p0_600; \ - float32x4_t __s1_600 = __p1_600; \ - float32x4_t __rev0_600; __rev0_600 = __builtin_shufflevector(__s0_600, __s0_600, 3, 2, 1, 0); \ - float32x4_t __rev1_600; __rev1_600 = __builtin_shufflevector(__s1_600, __s1_600, 3, 2, 1, 0); \ - float32x4_t __ret_600; \ - __ret_600 = __noswap_vmulxq_f32(__rev0_600, __noswap_splatq_laneq_f32(__rev1_600, __p2_600)); \ - __ret_600 = __builtin_shufflevector(__ret_600, __ret_600, 3, 2, 1, 0); \ - __ret_600; \ +#define vmulxq_laneq_f32(__p0_692, __p1_692, __p2_692) __extension__ ({ \ + float32x4_t __s0_692 = __p0_692; \ + float32x4_t __s1_692 = __p1_692; \ + float32x4_t __rev0_692; __rev0_692 = __builtin_shufflevector(__s0_692, __s0_692, 3, 2, 1, 0); \ + float32x4_t __rev1_692; __rev1_692 = __builtin_shufflevector(__s1_692, __s1_692, 3, 2, 1, 0); \ + float32x4_t __ret_692; \ + __ret_692 = __noswap_vmulxq_f32(__rev0_692, __noswap_splatq_laneq_f32(__rev1_692, __p2_692)); \ + __ret_692 = __builtin_shufflevector(__ret_692, __ret_692, 3, 2, 1, 0); \ + __ret_692; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulx_laneq_f32(__p0_601, __p1_601, __p2_601) __extension__ ({ \ - float32x2_t __s0_601 = __p0_601; \ - float32x4_t __s1_601 = __p1_601; \ - float32x2_t __ret_601; \ - __ret_601 = vmulx_f32(__s0_601, splat_laneq_f32(__s1_601, __p2_601)); \ - __ret_601; \ +#define vmulx_laneq_f32(__p0_693, __p1_693, __p2_693) __extension__ ({ \ + float32x2_t __s0_693 = __p0_693; \ + float32x4_t __s1_693 = __p1_693; \ + float32x2_t __ret_693; \ + __ret_693 = vmulx_f32(__s0_693, splat_laneq_f32(__s1_693, __p2_693)); \ + __ret_693; \ }) #else -#define vmulx_laneq_f32(__p0_602, __p1_602, __p2_602) __extension__ ({ \ - float32x2_t __s0_602 = __p0_602; \ - float32x4_t __s1_602 = __p1_602; \ - float32x2_t __rev0_602; __rev0_602 = __builtin_shufflevector(__s0_602, __s0_602, 1, 0); \ - float32x4_t __rev1_602; __rev1_602 = __builtin_shufflevector(__s1_602, __s1_602, 3, 2, 1, 0); \ - float32x2_t __ret_602; \ - __ret_602 = __noswap_vmulx_f32(__rev0_602, __noswap_splat_laneq_f32(__rev1_602, __p2_602)); \ - __ret_602 = __builtin_shufflevector(__ret_602, __ret_602, 1, 0); \ - __ret_602; \ +#define vmulx_laneq_f32(__p0_694, __p1_694, __p2_694) __extension__ ({ \ + float32x2_t __s0_694 = __p0_694; \ + float32x4_t __s1_694 = __p1_694; \ + float32x2_t __rev0_694; __rev0_694 = __builtin_shufflevector(__s0_694, __s0_694, 1, 0); \ + float32x4_t __rev1_694; __rev1_694 = __builtin_shufflevector(__s1_694, __s1_694, 3, 2, 1, 0); \ + float32x2_t __ret_694; \ + __ret_694 = __noswap_vmulx_f32(__rev0_694, __noswap_splat_laneq_f32(__rev1_694, __p2_694)); \ + __ret_694 = __builtin_shufflevector(__ret_694, __ret_694, 1, 0); \ + __ret_694; \ }) #endif @@ -55955,98 +57625,98 @@ __ai int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlal_high_lane_s32(__p0_603, __p1_603, __p2_603, __p3_603) __extension__ ({ \ - int64x2_t __s0_603 = __p0_603; \ - int32x4_t __s1_603 = __p1_603; \ - int32x2_t __s2_603 = __p2_603; \ - int64x2_t __ret_603; \ - __ret_603 = vqdmlal_s32(__s0_603, vget_high_s32(__s1_603), splat_lane_s32(__s2_603, __p3_603)); \ - __ret_603; \ +#define vqdmlal_high_lane_s32(__p0_695, __p1_695, __p2_695, __p3_695) __extension__ ({ \ + int64x2_t __s0_695 = __p0_695; \ + int32x4_t __s1_695 = __p1_695; \ + int32x2_t __s2_695 = __p2_695; \ + int64x2_t __ret_695; \ + __ret_695 = vqdmlal_s32(__s0_695, vget_high_s32(__s1_695), splat_lane_s32(__s2_695, __p3_695)); \ + __ret_695; \ }) #else -#define vqdmlal_high_lane_s32(__p0_604, __p1_604, __p2_604, __p3_604) __extension__ ({ \ - int64x2_t __s0_604 = __p0_604; \ - int32x4_t __s1_604 = __p1_604; \ - int32x2_t __s2_604 = __p2_604; \ - int64x2_t __rev0_604; __rev0_604 = __builtin_shufflevector(__s0_604, __s0_604, 1, 0); \ - int32x4_t __rev1_604; __rev1_604 = __builtin_shufflevector(__s1_604, __s1_604, 3, 2, 1, 0); \ - int32x2_t __rev2_604; __rev2_604 = __builtin_shufflevector(__s2_604, __s2_604, 1, 0); \ - int64x2_t __ret_604; \ - __ret_604 = __noswap_vqdmlal_s32(__rev0_604, __noswap_vget_high_s32(__rev1_604), __noswap_splat_lane_s32(__rev2_604, __p3_604)); \ - __ret_604 = __builtin_shufflevector(__ret_604, __ret_604, 1, 0); \ - __ret_604; \ +#define vqdmlal_high_lane_s32(__p0_696, __p1_696, __p2_696, __p3_696) __extension__ ({ \ + int64x2_t __s0_696 = __p0_696; \ + int32x4_t __s1_696 = __p1_696; \ + int32x2_t __s2_696 = __p2_696; \ + int64x2_t __rev0_696; __rev0_696 = __builtin_shufflevector(__s0_696, __s0_696, 1, 0); \ + int32x4_t __rev1_696; __rev1_696 = __builtin_shufflevector(__s1_696, __s1_696, 3, 2, 1, 0); \ + int32x2_t __rev2_696; __rev2_696 = __builtin_shufflevector(__s2_696, __s2_696, 1, 0); \ + int64x2_t __ret_696; \ + __ret_696 = __noswap_vqdmlal_s32(__rev0_696, __noswap_vget_high_s32(__rev1_696), __noswap_splat_lane_s32(__rev2_696, __p3_696)); \ + __ret_696 = __builtin_shufflevector(__ret_696, __ret_696, 1, 0); \ + __ret_696; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlal_high_lane_s16(__p0_605, __p1_605, __p2_605, __p3_605) __extension__ ({ \ - int32x4_t __s0_605 = __p0_605; \ - int16x8_t __s1_605 = __p1_605; \ - int16x4_t __s2_605 = __p2_605; \ - int32x4_t __ret_605; \ - __ret_605 = vqdmlal_s16(__s0_605, vget_high_s16(__s1_605), splat_lane_s16(__s2_605, __p3_605)); \ - __ret_605; \ +#define vqdmlal_high_lane_s16(__p0_697, __p1_697, __p2_697, __p3_697) __extension__ ({ \ + int32x4_t __s0_697 = __p0_697; \ + int16x8_t __s1_697 = __p1_697; \ + int16x4_t __s2_697 = __p2_697; \ + int32x4_t __ret_697; \ + __ret_697 = vqdmlal_s16(__s0_697, vget_high_s16(__s1_697), splat_lane_s16(__s2_697, __p3_697)); \ + __ret_697; \ }) #else -#define vqdmlal_high_lane_s16(__p0_606, __p1_606, __p2_606, __p3_606) __extension__ ({ \ - int32x4_t __s0_606 = __p0_606; \ - int16x8_t __s1_606 = __p1_606; \ - int16x4_t __s2_606 = __p2_606; \ - int32x4_t __rev0_606; __rev0_606 = __builtin_shufflevector(__s0_606, __s0_606, 3, 2, 1, 0); \ - int16x8_t __rev1_606; __rev1_606 = __builtin_shufflevector(__s1_606, __s1_606, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev2_606; __rev2_606 = __builtin_shufflevector(__s2_606, __s2_606, 3, 2, 1, 0); \ - int32x4_t __ret_606; \ - __ret_606 = __noswap_vqdmlal_s16(__rev0_606, __noswap_vget_high_s16(__rev1_606), __noswap_splat_lane_s16(__rev2_606, __p3_606)); \ - __ret_606 = __builtin_shufflevector(__ret_606, __ret_606, 3, 2, 1, 0); \ - __ret_606; \ +#define vqdmlal_high_lane_s16(__p0_698, __p1_698, __p2_698, __p3_698) __extension__ ({ \ + int32x4_t __s0_698 = __p0_698; \ + int16x8_t __s1_698 = __p1_698; \ + int16x4_t __s2_698 = __p2_698; \ + int32x4_t __rev0_698; __rev0_698 = __builtin_shufflevector(__s0_698, __s0_698, 3, 2, 1, 0); \ + int16x8_t __rev1_698; __rev1_698 = __builtin_shufflevector(__s1_698, __s1_698, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev2_698; __rev2_698 = __builtin_shufflevector(__s2_698, __s2_698, 3, 2, 1, 0); \ + int32x4_t __ret_698; \ + __ret_698 = __noswap_vqdmlal_s16(__rev0_698, __noswap_vget_high_s16(__rev1_698), __noswap_splat_lane_s16(__rev2_698, __p3_698)); \ + __ret_698 = __builtin_shufflevector(__ret_698, __ret_698, 3, 2, 1, 0); \ + __ret_698; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlal_high_laneq_s32(__p0_607, __p1_607, __p2_607, __p3_607) __extension__ ({ \ - int64x2_t __s0_607 = __p0_607; \ - int32x4_t __s1_607 = __p1_607; \ - int32x4_t __s2_607 = __p2_607; \ - int64x2_t __ret_607; \ - __ret_607 = vqdmlal_s32(__s0_607, vget_high_s32(__s1_607), splat_laneq_s32(__s2_607, __p3_607)); \ - __ret_607; \ +#define vqdmlal_high_laneq_s32(__p0_699, __p1_699, __p2_699, __p3_699) __extension__ ({ \ + int64x2_t __s0_699 = __p0_699; \ + int32x4_t __s1_699 = __p1_699; \ + int32x4_t __s2_699 = __p2_699; \ + int64x2_t __ret_699; \ + __ret_699 = vqdmlal_s32(__s0_699, vget_high_s32(__s1_699), splat_laneq_s32(__s2_699, __p3_699)); \ + __ret_699; \ }) #else -#define vqdmlal_high_laneq_s32(__p0_608, __p1_608, __p2_608, __p3_608) __extension__ ({ \ - int64x2_t __s0_608 = __p0_608; \ - int32x4_t __s1_608 = __p1_608; \ - int32x4_t __s2_608 = __p2_608; \ - int64x2_t __rev0_608; __rev0_608 = __builtin_shufflevector(__s0_608, __s0_608, 1, 0); \ - int32x4_t __rev1_608; __rev1_608 = __builtin_shufflevector(__s1_608, __s1_608, 3, 2, 1, 0); \ - int32x4_t __rev2_608; __rev2_608 = __builtin_shufflevector(__s2_608, __s2_608, 3, 2, 1, 0); \ - int64x2_t __ret_608; \ - __ret_608 = __noswap_vqdmlal_s32(__rev0_608, __noswap_vget_high_s32(__rev1_608), __noswap_splat_laneq_s32(__rev2_608, __p3_608)); \ - __ret_608 = __builtin_shufflevector(__ret_608, __ret_608, 1, 0); \ - __ret_608; \ +#define vqdmlal_high_laneq_s32(__p0_700, __p1_700, __p2_700, __p3_700) __extension__ ({ \ + int64x2_t __s0_700 = __p0_700; \ + int32x4_t __s1_700 = __p1_700; \ + int32x4_t __s2_700 = __p2_700; \ + int64x2_t __rev0_700; __rev0_700 = __builtin_shufflevector(__s0_700, __s0_700, 1, 0); \ + int32x4_t __rev1_700; __rev1_700 = __builtin_shufflevector(__s1_700, __s1_700, 3, 2, 1, 0); \ + int32x4_t __rev2_700; __rev2_700 = __builtin_shufflevector(__s2_700, __s2_700, 3, 2, 1, 0); \ + int64x2_t __ret_700; \ + __ret_700 = __noswap_vqdmlal_s32(__rev0_700, __noswap_vget_high_s32(__rev1_700), __noswap_splat_laneq_s32(__rev2_700, __p3_700)); \ + __ret_700 = __builtin_shufflevector(__ret_700, __ret_700, 1, 0); \ + __ret_700; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlal_high_laneq_s16(__p0_609, __p1_609, __p2_609, __p3_609) __extension__ ({ \ - int32x4_t __s0_609 = __p0_609; \ - int16x8_t __s1_609 = __p1_609; \ - int16x8_t __s2_609 = __p2_609; \ - int32x4_t __ret_609; \ - __ret_609 = vqdmlal_s16(__s0_609, vget_high_s16(__s1_609), splat_laneq_s16(__s2_609, __p3_609)); \ - __ret_609; \ +#define vqdmlal_high_laneq_s16(__p0_701, __p1_701, __p2_701, __p3_701) __extension__ ({ \ + int32x4_t __s0_701 = __p0_701; \ + int16x8_t __s1_701 = __p1_701; \ + int16x8_t __s2_701 = __p2_701; \ + int32x4_t __ret_701; \ + __ret_701 = vqdmlal_s16(__s0_701, vget_high_s16(__s1_701), splat_laneq_s16(__s2_701, __p3_701)); \ + __ret_701; \ }) #else -#define vqdmlal_high_laneq_s16(__p0_610, __p1_610, __p2_610, __p3_610) __extension__ ({ \ - int32x4_t __s0_610 = __p0_610; \ - int16x8_t __s1_610 = __p1_610; \ - int16x8_t __s2_610 = __p2_610; \ - int32x4_t __rev0_610; __rev0_610 = __builtin_shufflevector(__s0_610, __s0_610, 3, 2, 1, 0); \ - int16x8_t __rev1_610; __rev1_610 = __builtin_shufflevector(__s1_610, __s1_610, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_610; __rev2_610 = __builtin_shufflevector(__s2_610, __s2_610, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_610; \ - __ret_610 = __noswap_vqdmlal_s16(__rev0_610, __noswap_vget_high_s16(__rev1_610), __noswap_splat_laneq_s16(__rev2_610, __p3_610)); \ - __ret_610 = __builtin_shufflevector(__ret_610, __ret_610, 3, 2, 1, 0); \ - __ret_610; \ +#define vqdmlal_high_laneq_s16(__p0_702, __p1_702, __p2_702, __p3_702) __extension__ ({ \ + int32x4_t __s0_702 = __p0_702; \ + int16x8_t __s1_702 = __p1_702; \ + int16x8_t __s2_702 = __p2_702; \ + int32x4_t __rev0_702; __rev0_702 = __builtin_shufflevector(__s0_702, __s0_702, 3, 2, 1, 0); \ + int16x8_t __rev1_702; __rev1_702 = __builtin_shufflevector(__s1_702, __s1_702, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_702; __rev2_702 = __builtin_shufflevector(__s2_702, __s2_702, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_702; \ + __ret_702 = __noswap_vqdmlal_s16(__rev0_702, __noswap_vget_high_s16(__rev1_702), __noswap_splat_laneq_s16(__rev2_702, __p3_702)); \ + __ret_702 = __builtin_shufflevector(__ret_702, __ret_702, 3, 2, 1, 0); \ + __ret_702; \ }) #endif @@ -56169,50 +57839,50 @@ __ai int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlal_laneq_s32(__p0_611, __p1_611, __p2_611, __p3_611) __extension__ ({ \ - int64x2_t __s0_611 = __p0_611; \ - int32x2_t __s1_611 = __p1_611; \ - int32x4_t __s2_611 = __p2_611; \ - int64x2_t __ret_611; \ - __ret_611 = vqdmlal_s32(__s0_611, __s1_611, splat_laneq_s32(__s2_611, __p3_611)); \ - __ret_611; \ +#define vqdmlal_laneq_s32(__p0_703, __p1_703, __p2_703, __p3_703) __extension__ ({ \ + int64x2_t __s0_703 = __p0_703; \ + int32x2_t __s1_703 = __p1_703; \ + int32x4_t __s2_703 = __p2_703; \ + int64x2_t __ret_703; \ + __ret_703 = vqdmlal_s32(__s0_703, __s1_703, splat_laneq_s32(__s2_703, __p3_703)); \ + __ret_703; \ }) #else -#define vqdmlal_laneq_s32(__p0_612, __p1_612, __p2_612, __p3_612) __extension__ ({ \ - int64x2_t __s0_612 = __p0_612; \ - int32x2_t __s1_612 = __p1_612; \ - int32x4_t __s2_612 = __p2_612; \ - int64x2_t __rev0_612; __rev0_612 = __builtin_shufflevector(__s0_612, __s0_612, 1, 0); \ - int32x2_t __rev1_612; __rev1_612 = __builtin_shufflevector(__s1_612, __s1_612, 1, 0); \ - int32x4_t __rev2_612; __rev2_612 = __builtin_shufflevector(__s2_612, __s2_612, 3, 2, 1, 0); \ - int64x2_t __ret_612; \ - __ret_612 = __noswap_vqdmlal_s32(__rev0_612, __rev1_612, __noswap_splat_laneq_s32(__rev2_612, __p3_612)); \ - __ret_612 = __builtin_shufflevector(__ret_612, __ret_612, 1, 0); \ - __ret_612; \ +#define vqdmlal_laneq_s32(__p0_704, __p1_704, __p2_704, __p3_704) __extension__ ({ \ + int64x2_t __s0_704 = __p0_704; \ + int32x2_t __s1_704 = __p1_704; \ + int32x4_t __s2_704 = __p2_704; \ + int64x2_t __rev0_704; __rev0_704 = __builtin_shufflevector(__s0_704, __s0_704, 1, 0); \ + int32x2_t __rev1_704; __rev1_704 = __builtin_shufflevector(__s1_704, __s1_704, 1, 0); \ + int32x4_t __rev2_704; __rev2_704 = __builtin_shufflevector(__s2_704, __s2_704, 3, 2, 1, 0); \ + int64x2_t __ret_704; \ + __ret_704 = __noswap_vqdmlal_s32(__rev0_704, __rev1_704, __noswap_splat_laneq_s32(__rev2_704, __p3_704)); \ + __ret_704 = __builtin_shufflevector(__ret_704, __ret_704, 1, 0); \ + __ret_704; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlal_laneq_s16(__p0_613, __p1_613, __p2_613, __p3_613) __extension__ ({ \ - int32x4_t __s0_613 = __p0_613; \ - int16x4_t __s1_613 = __p1_613; \ - int16x8_t __s2_613 = __p2_613; \ - int32x4_t __ret_613; \ - __ret_613 = vqdmlal_s16(__s0_613, __s1_613, splat_laneq_s16(__s2_613, __p3_613)); \ - __ret_613; \ +#define vqdmlal_laneq_s16(__p0_705, __p1_705, __p2_705, __p3_705) __extension__ ({ \ + int32x4_t __s0_705 = __p0_705; \ + int16x4_t __s1_705 = __p1_705; \ + int16x8_t __s2_705 = __p2_705; \ + int32x4_t __ret_705; \ + __ret_705 = vqdmlal_s16(__s0_705, __s1_705, splat_laneq_s16(__s2_705, __p3_705)); \ + __ret_705; \ }) #else -#define vqdmlal_laneq_s16(__p0_614, __p1_614, __p2_614, __p3_614) __extension__ ({ \ - int32x4_t __s0_614 = __p0_614; \ - int16x4_t __s1_614 = __p1_614; \ - int16x8_t __s2_614 = __p2_614; \ - int32x4_t __rev0_614; __rev0_614 = __builtin_shufflevector(__s0_614, __s0_614, 3, 2, 1, 0); \ - int16x4_t __rev1_614; __rev1_614 = __builtin_shufflevector(__s1_614, __s1_614, 3, 2, 1, 0); \ - int16x8_t __rev2_614; __rev2_614 = __builtin_shufflevector(__s2_614, __s2_614, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_614; \ - __ret_614 = __noswap_vqdmlal_s16(__rev0_614, __rev1_614, __noswap_splat_laneq_s16(__rev2_614, __p3_614)); \ - __ret_614 = __builtin_shufflevector(__ret_614, __ret_614, 3, 2, 1, 0); \ - __ret_614; \ +#define vqdmlal_laneq_s16(__p0_706, __p1_706, __p2_706, __p3_706) __extension__ ({ \ + int32x4_t __s0_706 = __p0_706; \ + int16x4_t __s1_706 = __p1_706; \ + int16x8_t __s2_706 = __p2_706; \ + int32x4_t __rev0_706; __rev0_706 = __builtin_shufflevector(__s0_706, __s0_706, 3, 2, 1, 0); \ + int16x4_t __rev1_706; __rev1_706 = __builtin_shufflevector(__s1_706, __s1_706, 3, 2, 1, 0); \ + int16x8_t __rev2_706; __rev2_706 = __builtin_shufflevector(__s2_706, __s2_706, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_706; \ + __ret_706 = __noswap_vqdmlal_s16(__rev0_706, __rev1_706, __noswap_splat_laneq_s16(__rev2_706, __p3_706)); \ + __ret_706 = __builtin_shufflevector(__ret_706, __ret_706, 3, 2, 1, 0); \ + __ret_706; \ }) #endif @@ -56263,98 +57933,98 @@ __ai int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlsl_high_lane_s32(__p0_615, __p1_615, __p2_615, __p3_615) __extension__ ({ \ - int64x2_t __s0_615 = __p0_615; \ - int32x4_t __s1_615 = __p1_615; \ - int32x2_t __s2_615 = __p2_615; \ - int64x2_t __ret_615; \ - __ret_615 = vqdmlsl_s32(__s0_615, vget_high_s32(__s1_615), splat_lane_s32(__s2_615, __p3_615)); \ - __ret_615; \ +#define vqdmlsl_high_lane_s32(__p0_707, __p1_707, __p2_707, __p3_707) __extension__ ({ \ + int64x2_t __s0_707 = __p0_707; \ + int32x4_t __s1_707 = __p1_707; \ + int32x2_t __s2_707 = __p2_707; \ + int64x2_t __ret_707; \ + __ret_707 = vqdmlsl_s32(__s0_707, vget_high_s32(__s1_707), splat_lane_s32(__s2_707, __p3_707)); \ + __ret_707; \ }) #else -#define vqdmlsl_high_lane_s32(__p0_616, __p1_616, __p2_616, __p3_616) __extension__ ({ \ - int64x2_t __s0_616 = __p0_616; \ - int32x4_t __s1_616 = __p1_616; \ - int32x2_t __s2_616 = __p2_616; \ - int64x2_t __rev0_616; __rev0_616 = __builtin_shufflevector(__s0_616, __s0_616, 1, 0); \ - int32x4_t __rev1_616; __rev1_616 = __builtin_shufflevector(__s1_616, __s1_616, 3, 2, 1, 0); \ - int32x2_t __rev2_616; __rev2_616 = __builtin_shufflevector(__s2_616, __s2_616, 1, 0); \ - int64x2_t __ret_616; \ - __ret_616 = __noswap_vqdmlsl_s32(__rev0_616, __noswap_vget_high_s32(__rev1_616), __noswap_splat_lane_s32(__rev2_616, __p3_616)); \ - __ret_616 = __builtin_shufflevector(__ret_616, __ret_616, 1, 0); \ - __ret_616; \ +#define vqdmlsl_high_lane_s32(__p0_708, __p1_708, __p2_708, __p3_708) __extension__ ({ \ + int64x2_t __s0_708 = __p0_708; \ + int32x4_t __s1_708 = __p1_708; \ + int32x2_t __s2_708 = __p2_708; \ + int64x2_t __rev0_708; __rev0_708 = __builtin_shufflevector(__s0_708, __s0_708, 1, 0); \ + int32x4_t __rev1_708; __rev1_708 = __builtin_shufflevector(__s1_708, __s1_708, 3, 2, 1, 0); \ + int32x2_t __rev2_708; __rev2_708 = __builtin_shufflevector(__s2_708, __s2_708, 1, 0); \ + int64x2_t __ret_708; \ + __ret_708 = __noswap_vqdmlsl_s32(__rev0_708, __noswap_vget_high_s32(__rev1_708), __noswap_splat_lane_s32(__rev2_708, __p3_708)); \ + __ret_708 = __builtin_shufflevector(__ret_708, __ret_708, 1, 0); \ + __ret_708; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlsl_high_lane_s16(__p0_617, __p1_617, __p2_617, __p3_617) __extension__ ({ \ - int32x4_t __s0_617 = __p0_617; \ - int16x8_t __s1_617 = __p1_617; \ - int16x4_t __s2_617 = __p2_617; \ - int32x4_t __ret_617; \ - __ret_617 = vqdmlsl_s16(__s0_617, vget_high_s16(__s1_617), splat_lane_s16(__s2_617, __p3_617)); \ - __ret_617; \ +#define vqdmlsl_high_lane_s16(__p0_709, __p1_709, __p2_709, __p3_709) __extension__ ({ \ + int32x4_t __s0_709 = __p0_709; \ + int16x8_t __s1_709 = __p1_709; \ + int16x4_t __s2_709 = __p2_709; \ + int32x4_t __ret_709; \ + __ret_709 = vqdmlsl_s16(__s0_709, vget_high_s16(__s1_709), splat_lane_s16(__s2_709, __p3_709)); \ + __ret_709; \ }) #else -#define vqdmlsl_high_lane_s16(__p0_618, __p1_618, __p2_618, __p3_618) __extension__ ({ \ - int32x4_t __s0_618 = __p0_618; \ - int16x8_t __s1_618 = __p1_618; \ - int16x4_t __s2_618 = __p2_618; \ - int32x4_t __rev0_618; __rev0_618 = __builtin_shufflevector(__s0_618, __s0_618, 3, 2, 1, 0); \ - int16x8_t __rev1_618; __rev1_618 = __builtin_shufflevector(__s1_618, __s1_618, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev2_618; __rev2_618 = __builtin_shufflevector(__s2_618, __s2_618, 3, 2, 1, 0); \ - int32x4_t __ret_618; \ - __ret_618 = __noswap_vqdmlsl_s16(__rev0_618, __noswap_vget_high_s16(__rev1_618), __noswap_splat_lane_s16(__rev2_618, __p3_618)); \ - __ret_618 = __builtin_shufflevector(__ret_618, __ret_618, 3, 2, 1, 0); \ - __ret_618; \ +#define vqdmlsl_high_lane_s16(__p0_710, __p1_710, __p2_710, __p3_710) __extension__ ({ \ + int32x4_t __s0_710 = __p0_710; \ + int16x8_t __s1_710 = __p1_710; \ + int16x4_t __s2_710 = __p2_710; \ + int32x4_t __rev0_710; __rev0_710 = __builtin_shufflevector(__s0_710, __s0_710, 3, 2, 1, 0); \ + int16x8_t __rev1_710; __rev1_710 = __builtin_shufflevector(__s1_710, __s1_710, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev2_710; __rev2_710 = __builtin_shufflevector(__s2_710, __s2_710, 3, 2, 1, 0); \ + int32x4_t __ret_710; \ + __ret_710 = __noswap_vqdmlsl_s16(__rev0_710, __noswap_vget_high_s16(__rev1_710), __noswap_splat_lane_s16(__rev2_710, __p3_710)); \ + __ret_710 = __builtin_shufflevector(__ret_710, __ret_710, 3, 2, 1, 0); \ + __ret_710; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlsl_high_laneq_s32(__p0_619, __p1_619, __p2_619, __p3_619) __extension__ ({ \ - int64x2_t __s0_619 = __p0_619; \ - int32x4_t __s1_619 = __p1_619; \ - int32x4_t __s2_619 = __p2_619; \ - int64x2_t __ret_619; \ - __ret_619 = vqdmlsl_s32(__s0_619, vget_high_s32(__s1_619), splat_laneq_s32(__s2_619, __p3_619)); \ - __ret_619; \ +#define vqdmlsl_high_laneq_s32(__p0_711, __p1_711, __p2_711, __p3_711) __extension__ ({ \ + int64x2_t __s0_711 = __p0_711; \ + int32x4_t __s1_711 = __p1_711; \ + int32x4_t __s2_711 = __p2_711; \ + int64x2_t __ret_711; \ + __ret_711 = vqdmlsl_s32(__s0_711, vget_high_s32(__s1_711), splat_laneq_s32(__s2_711, __p3_711)); \ + __ret_711; \ }) #else -#define vqdmlsl_high_laneq_s32(__p0_620, __p1_620, __p2_620, __p3_620) __extension__ ({ \ - int64x2_t __s0_620 = __p0_620; \ - int32x4_t __s1_620 = __p1_620; \ - int32x4_t __s2_620 = __p2_620; \ - int64x2_t __rev0_620; __rev0_620 = __builtin_shufflevector(__s0_620, __s0_620, 1, 0); \ - int32x4_t __rev1_620; __rev1_620 = __builtin_shufflevector(__s1_620, __s1_620, 3, 2, 1, 0); \ - int32x4_t __rev2_620; __rev2_620 = __builtin_shufflevector(__s2_620, __s2_620, 3, 2, 1, 0); \ - int64x2_t __ret_620; \ - __ret_620 = __noswap_vqdmlsl_s32(__rev0_620, __noswap_vget_high_s32(__rev1_620), __noswap_splat_laneq_s32(__rev2_620, __p3_620)); \ - __ret_620 = __builtin_shufflevector(__ret_620, __ret_620, 1, 0); \ - __ret_620; \ +#define vqdmlsl_high_laneq_s32(__p0_712, __p1_712, __p2_712, __p3_712) __extension__ ({ \ + int64x2_t __s0_712 = __p0_712; \ + int32x4_t __s1_712 = __p1_712; \ + int32x4_t __s2_712 = __p2_712; \ + int64x2_t __rev0_712; __rev0_712 = __builtin_shufflevector(__s0_712, __s0_712, 1, 0); \ + int32x4_t __rev1_712; __rev1_712 = __builtin_shufflevector(__s1_712, __s1_712, 3, 2, 1, 0); \ + int32x4_t __rev2_712; __rev2_712 = __builtin_shufflevector(__s2_712, __s2_712, 3, 2, 1, 0); \ + int64x2_t __ret_712; \ + __ret_712 = __noswap_vqdmlsl_s32(__rev0_712, __noswap_vget_high_s32(__rev1_712), __noswap_splat_laneq_s32(__rev2_712, __p3_712)); \ + __ret_712 = __builtin_shufflevector(__ret_712, __ret_712, 1, 0); \ + __ret_712; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlsl_high_laneq_s16(__p0_621, __p1_621, __p2_621, __p3_621) __extension__ ({ \ - int32x4_t __s0_621 = __p0_621; \ - int16x8_t __s1_621 = __p1_621; \ - int16x8_t __s2_621 = __p2_621; \ - int32x4_t __ret_621; \ - __ret_621 = vqdmlsl_s16(__s0_621, vget_high_s16(__s1_621), splat_laneq_s16(__s2_621, __p3_621)); \ - __ret_621; \ +#define vqdmlsl_high_laneq_s16(__p0_713, __p1_713, __p2_713, __p3_713) __extension__ ({ \ + int32x4_t __s0_713 = __p0_713; \ + int16x8_t __s1_713 = __p1_713; \ + int16x8_t __s2_713 = __p2_713; \ + int32x4_t __ret_713; \ + __ret_713 = vqdmlsl_s16(__s0_713, vget_high_s16(__s1_713), splat_laneq_s16(__s2_713, __p3_713)); \ + __ret_713; \ }) #else -#define vqdmlsl_high_laneq_s16(__p0_622, __p1_622, __p2_622, __p3_622) __extension__ ({ \ - int32x4_t __s0_622 = __p0_622; \ - int16x8_t __s1_622 = __p1_622; \ - int16x8_t __s2_622 = __p2_622; \ - int32x4_t __rev0_622; __rev0_622 = __builtin_shufflevector(__s0_622, __s0_622, 3, 2, 1, 0); \ - int16x8_t __rev1_622; __rev1_622 = __builtin_shufflevector(__s1_622, __s1_622, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev2_622; __rev2_622 = __builtin_shufflevector(__s2_622, __s2_622, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_622; \ - __ret_622 = __noswap_vqdmlsl_s16(__rev0_622, __noswap_vget_high_s16(__rev1_622), __noswap_splat_laneq_s16(__rev2_622, __p3_622)); \ - __ret_622 = __builtin_shufflevector(__ret_622, __ret_622, 3, 2, 1, 0); \ - __ret_622; \ +#define vqdmlsl_high_laneq_s16(__p0_714, __p1_714, __p2_714, __p3_714) __extension__ ({ \ + int32x4_t __s0_714 = __p0_714; \ + int16x8_t __s1_714 = __p1_714; \ + int16x8_t __s2_714 = __p2_714; \ + int32x4_t __rev0_714; __rev0_714 = __builtin_shufflevector(__s0_714, __s0_714, 3, 2, 1, 0); \ + int16x8_t __rev1_714; __rev1_714 = __builtin_shufflevector(__s1_714, __s1_714, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev2_714; __rev2_714 = __builtin_shufflevector(__s2_714, __s2_714, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_714; \ + __ret_714 = __noswap_vqdmlsl_s16(__rev0_714, __noswap_vget_high_s16(__rev1_714), __noswap_splat_laneq_s16(__rev2_714, __p3_714)); \ + __ret_714 = __builtin_shufflevector(__ret_714, __ret_714, 3, 2, 1, 0); \ + __ret_714; \ }) #endif @@ -56477,50 +58147,50 @@ __ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlsl_laneq_s32(__p0_623, __p1_623, __p2_623, __p3_623) __extension__ ({ \ - int64x2_t __s0_623 = __p0_623; \ - int32x2_t __s1_623 = __p1_623; \ - int32x4_t __s2_623 = __p2_623; \ - int64x2_t __ret_623; \ - __ret_623 = vqdmlsl_s32(__s0_623, __s1_623, splat_laneq_s32(__s2_623, __p3_623)); \ - __ret_623; \ +#define vqdmlsl_laneq_s32(__p0_715, __p1_715, __p2_715, __p3_715) __extension__ ({ \ + int64x2_t __s0_715 = __p0_715; \ + int32x2_t __s1_715 = __p1_715; \ + int32x4_t __s2_715 = __p2_715; \ + int64x2_t __ret_715; \ + __ret_715 = vqdmlsl_s32(__s0_715, __s1_715, splat_laneq_s32(__s2_715, __p3_715)); \ + __ret_715; \ }) #else -#define vqdmlsl_laneq_s32(__p0_624, __p1_624, __p2_624, __p3_624) __extension__ ({ \ - int64x2_t __s0_624 = __p0_624; \ - int32x2_t __s1_624 = __p1_624; \ - int32x4_t __s2_624 = __p2_624; \ - int64x2_t __rev0_624; __rev0_624 = __builtin_shufflevector(__s0_624, __s0_624, 1, 0); \ - int32x2_t __rev1_624; __rev1_624 = __builtin_shufflevector(__s1_624, __s1_624, 1, 0); \ - int32x4_t __rev2_624; __rev2_624 = __builtin_shufflevector(__s2_624, __s2_624, 3, 2, 1, 0); \ - int64x2_t __ret_624; \ - __ret_624 = __noswap_vqdmlsl_s32(__rev0_624, __rev1_624, __noswap_splat_laneq_s32(__rev2_624, __p3_624)); \ - __ret_624 = __builtin_shufflevector(__ret_624, __ret_624, 1, 0); \ - __ret_624; \ +#define vqdmlsl_laneq_s32(__p0_716, __p1_716, __p2_716, __p3_716) __extension__ ({ \ + int64x2_t __s0_716 = __p0_716; \ + int32x2_t __s1_716 = __p1_716; \ + int32x4_t __s2_716 = __p2_716; \ + int64x2_t __rev0_716; __rev0_716 = __builtin_shufflevector(__s0_716, __s0_716, 1, 0); \ + int32x2_t __rev1_716; __rev1_716 = __builtin_shufflevector(__s1_716, __s1_716, 1, 0); \ + int32x4_t __rev2_716; __rev2_716 = __builtin_shufflevector(__s2_716, __s2_716, 3, 2, 1, 0); \ + int64x2_t __ret_716; \ + __ret_716 = __noswap_vqdmlsl_s32(__rev0_716, __rev1_716, __noswap_splat_laneq_s32(__rev2_716, __p3_716)); \ + __ret_716 = __builtin_shufflevector(__ret_716, __ret_716, 1, 0); \ + __ret_716; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmlsl_laneq_s16(__p0_625, __p1_625, __p2_625, __p3_625) __extension__ ({ \ - int32x4_t __s0_625 = __p0_625; \ - int16x4_t __s1_625 = __p1_625; \ - int16x8_t __s2_625 = __p2_625; \ - int32x4_t __ret_625; \ - __ret_625 = vqdmlsl_s16(__s0_625, __s1_625, splat_laneq_s16(__s2_625, __p3_625)); \ - __ret_625; \ +#define vqdmlsl_laneq_s16(__p0_717, __p1_717, __p2_717, __p3_717) __extension__ ({ \ + int32x4_t __s0_717 = __p0_717; \ + int16x4_t __s1_717 = __p1_717; \ + int16x8_t __s2_717 = __p2_717; \ + int32x4_t __ret_717; \ + __ret_717 = vqdmlsl_s16(__s0_717, __s1_717, splat_laneq_s16(__s2_717, __p3_717)); \ + __ret_717; \ }) #else -#define vqdmlsl_laneq_s16(__p0_626, __p1_626, __p2_626, __p3_626) __extension__ ({ \ - int32x4_t __s0_626 = __p0_626; \ - int16x4_t __s1_626 = __p1_626; \ - int16x8_t __s2_626 = __p2_626; \ - int32x4_t __rev0_626; __rev0_626 = __builtin_shufflevector(__s0_626, __s0_626, 3, 2, 1, 0); \ - int16x4_t __rev1_626; __rev1_626 = __builtin_shufflevector(__s1_626, __s1_626, 3, 2, 1, 0); \ - int16x8_t __rev2_626; __rev2_626 = __builtin_shufflevector(__s2_626, __s2_626, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_626; \ - __ret_626 = __noswap_vqdmlsl_s16(__rev0_626, __rev1_626, __noswap_splat_laneq_s16(__rev2_626, __p3_626)); \ - __ret_626 = __builtin_shufflevector(__ret_626, __ret_626, 3, 2, 1, 0); \ - __ret_626; \ +#define vqdmlsl_laneq_s16(__p0_718, __p1_718, __p2_718, __p3_718) __extension__ ({ \ + int32x4_t __s0_718 = __p0_718; \ + int16x4_t __s1_718 = __p1_718; \ + int16x8_t __s2_718 = __p2_718; \ + int32x4_t __rev0_718; __rev0_718 = __builtin_shufflevector(__s0_718, __s0_718, 3, 2, 1, 0); \ + int16x4_t __rev1_718; __rev1_718 = __builtin_shufflevector(__s1_718, __s1_718, 3, 2, 1, 0); \ + int16x8_t __rev2_718; __rev2_718 = __builtin_shufflevector(__s2_718, __s2_718, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_718; \ + __ret_718 = __noswap_vqdmlsl_s16(__rev0_718, __rev1_718, __noswap_splat_laneq_s16(__rev2_718, __p3_718)); \ + __ret_718 = __builtin_shufflevector(__ret_718, __ret_718, 3, 2, 1, 0); \ + __ret_718; \ }) #endif @@ -56619,78 +58289,78 @@ __ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmulhs_lane_s32(__p0_627, __p1_627, __p2_627) __extension__ ({ \ - int32_t __s0_627 = __p0_627; \ - int32x2_t __s1_627 = __p1_627; \ - int32_t __ret_627; \ - __ret_627 = vqdmulhs_s32(__s0_627, vget_lane_s32(__s1_627, __p2_627)); \ - __ret_627; \ +#define vqdmulhs_lane_s32(__p0_719, __p1_719, __p2_719) __extension__ ({ \ + int32_t __s0_719 = __p0_719; \ + int32x2_t __s1_719 = __p1_719; \ + int32_t __ret_719; \ + __ret_719 = vqdmulhs_s32(__s0_719, vget_lane_s32(__s1_719, __p2_719)); \ + __ret_719; \ }) #else -#define vqdmulhs_lane_s32(__p0_628, __p1_628, __p2_628) __extension__ ({ \ - int32_t __s0_628 = __p0_628; \ - int32x2_t __s1_628 = __p1_628; \ - int32x2_t __rev1_628; __rev1_628 = __builtin_shufflevector(__s1_628, __s1_628, 1, 0); \ - int32_t __ret_628; \ - __ret_628 = vqdmulhs_s32(__s0_628, __noswap_vget_lane_s32(__rev1_628, __p2_628)); \ - __ret_628; \ +#define vqdmulhs_lane_s32(__p0_720, __p1_720, __p2_720) __extension__ ({ \ + int32_t __s0_720 = __p0_720; \ + int32x2_t __s1_720 = __p1_720; \ + int32x2_t __rev1_720; __rev1_720 = __builtin_shufflevector(__s1_720, __s1_720, 1, 0); \ + int32_t __ret_720; \ + __ret_720 = vqdmulhs_s32(__s0_720, __noswap_vget_lane_s32(__rev1_720, __p2_720)); \ + __ret_720; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmulhh_lane_s16(__p0_629, __p1_629, __p2_629) __extension__ ({ \ - int16_t __s0_629 = __p0_629; \ - int16x4_t __s1_629 = __p1_629; \ - int16_t __ret_629; \ - __ret_629 = vqdmulhh_s16(__s0_629, vget_lane_s16(__s1_629, __p2_629)); \ - __ret_629; \ +#define vqdmulhh_lane_s16(__p0_721, __p1_721, __p2_721) __extension__ ({ \ + int16_t __s0_721 = __p0_721; \ + int16x4_t __s1_721 = __p1_721; \ + int16_t __ret_721; \ + __ret_721 = vqdmulhh_s16(__s0_721, vget_lane_s16(__s1_721, __p2_721)); \ + __ret_721; \ }) #else -#define vqdmulhh_lane_s16(__p0_630, __p1_630, __p2_630) __extension__ ({ \ - int16_t __s0_630 = __p0_630; \ - int16x4_t __s1_630 = __p1_630; \ - int16x4_t __rev1_630; __rev1_630 = __builtin_shufflevector(__s1_630, __s1_630, 3, 2, 1, 0); \ - int16_t __ret_630; \ - __ret_630 = vqdmulhh_s16(__s0_630, __noswap_vget_lane_s16(__rev1_630, __p2_630)); \ - __ret_630; \ +#define vqdmulhh_lane_s16(__p0_722, __p1_722, __p2_722) __extension__ ({ \ + int16_t __s0_722 = __p0_722; \ + int16x4_t __s1_722 = __p1_722; \ + int16x4_t __rev1_722; __rev1_722 = __builtin_shufflevector(__s1_722, __s1_722, 3, 2, 1, 0); \ + int16_t __ret_722; \ + __ret_722 = vqdmulhh_s16(__s0_722, __noswap_vget_lane_s16(__rev1_722, __p2_722)); \ + __ret_722; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmulhs_laneq_s32(__p0_631, __p1_631, __p2_631) __extension__ ({ \ - int32_t __s0_631 = __p0_631; \ - int32x4_t __s1_631 = __p1_631; \ - int32_t __ret_631; \ - __ret_631 = vqdmulhs_s32(__s0_631, vgetq_lane_s32(__s1_631, __p2_631)); \ - __ret_631; \ +#define vqdmulhs_laneq_s32(__p0_723, __p1_723, __p2_723) __extension__ ({ \ + int32_t __s0_723 = __p0_723; \ + int32x4_t __s1_723 = __p1_723; \ + int32_t __ret_723; \ + __ret_723 = vqdmulhs_s32(__s0_723, vgetq_lane_s32(__s1_723, __p2_723)); \ + __ret_723; \ }) #else -#define vqdmulhs_laneq_s32(__p0_632, __p1_632, __p2_632) __extension__ ({ \ - int32_t __s0_632 = __p0_632; \ - int32x4_t __s1_632 = __p1_632; \ - int32x4_t __rev1_632; __rev1_632 = __builtin_shufflevector(__s1_632, __s1_632, 3, 2, 1, 0); \ - int32_t __ret_632; \ - __ret_632 = vqdmulhs_s32(__s0_632, __noswap_vgetq_lane_s32(__rev1_632, __p2_632)); \ - __ret_632; \ +#define vqdmulhs_laneq_s32(__p0_724, __p1_724, __p2_724) __extension__ ({ \ + int32_t __s0_724 = __p0_724; \ + int32x4_t __s1_724 = __p1_724; \ + int32x4_t __rev1_724; __rev1_724 = __builtin_shufflevector(__s1_724, __s1_724, 3, 2, 1, 0); \ + int32_t __ret_724; \ + __ret_724 = vqdmulhs_s32(__s0_724, __noswap_vgetq_lane_s32(__rev1_724, __p2_724)); \ + __ret_724; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmulhh_laneq_s16(__p0_633, __p1_633, __p2_633) __extension__ ({ \ - int16_t __s0_633 = __p0_633; \ - int16x8_t __s1_633 = __p1_633; \ - int16_t __ret_633; \ - __ret_633 = vqdmulhh_s16(__s0_633, vgetq_lane_s16(__s1_633, __p2_633)); \ - __ret_633; \ +#define vqdmulhh_laneq_s16(__p0_725, __p1_725, __p2_725) __extension__ ({ \ + int16_t __s0_725 = __p0_725; \ + int16x8_t __s1_725 = __p1_725; \ + int16_t __ret_725; \ + __ret_725 = vqdmulhh_s16(__s0_725, vgetq_lane_s16(__s1_725, __p2_725)); \ + __ret_725; \ }) #else -#define vqdmulhh_laneq_s16(__p0_634, __p1_634, __p2_634) __extension__ ({ \ - int16_t __s0_634 = __p0_634; \ - int16x8_t __s1_634 = __p1_634; \ - int16x8_t __rev1_634; __rev1_634 = __builtin_shufflevector(__s1_634, __s1_634, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16_t __ret_634; \ - __ret_634 = vqdmulhh_s16(__s0_634, __noswap_vgetq_lane_s16(__rev1_634, __p2_634)); \ - __ret_634; \ +#define vqdmulhh_laneq_s16(__p0_726, __p1_726, __p2_726) __extension__ ({ \ + int16_t __s0_726 = __p0_726; \ + int16x8_t __s1_726 = __p1_726; \ + int16x8_t __rev1_726; __rev1_726 = __builtin_shufflevector(__s1_726, __s1_726, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16_t __ret_726; \ + __ret_726 = vqdmulhh_s16(__s0_726, __noswap_vgetq_lane_s16(__rev1_726, __p2_726)); \ + __ret_726; \ }) #endif @@ -56823,86 +58493,86 @@ __ai int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmull_high_lane_s32(__p0_635, __p1_635, __p2_635) __extension__ ({ \ - int32x4_t __s0_635 = __p0_635; \ - int32x2_t __s1_635 = __p1_635; \ - int64x2_t __ret_635; \ - __ret_635 = vqdmull_s32(vget_high_s32(__s0_635), splat_lane_s32(__s1_635, __p2_635)); \ - __ret_635; \ +#define vqdmull_high_lane_s32(__p0_727, __p1_727, __p2_727) __extension__ ({ \ + int32x4_t __s0_727 = __p0_727; \ + int32x2_t __s1_727 = __p1_727; \ + int64x2_t __ret_727; \ + __ret_727 = vqdmull_s32(vget_high_s32(__s0_727), splat_lane_s32(__s1_727, __p2_727)); \ + __ret_727; \ }) #else -#define vqdmull_high_lane_s32(__p0_636, __p1_636, __p2_636) __extension__ ({ \ - int32x4_t __s0_636 = __p0_636; \ - int32x2_t __s1_636 = __p1_636; \ - int32x4_t __rev0_636; __rev0_636 = __builtin_shufflevector(__s0_636, __s0_636, 3, 2, 1, 0); \ - int32x2_t __rev1_636; __rev1_636 = __builtin_shufflevector(__s1_636, __s1_636, 1, 0); \ - int64x2_t __ret_636; \ - __ret_636 = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0_636), __noswap_splat_lane_s32(__rev1_636, __p2_636)); \ - __ret_636 = __builtin_shufflevector(__ret_636, __ret_636, 1, 0); \ - __ret_636; \ +#define vqdmull_high_lane_s32(__p0_728, __p1_728, __p2_728) __extension__ ({ \ + int32x4_t __s0_728 = __p0_728; \ + int32x2_t __s1_728 = __p1_728; \ + int32x4_t __rev0_728; __rev0_728 = __builtin_shufflevector(__s0_728, __s0_728, 3, 2, 1, 0); \ + int32x2_t __rev1_728; __rev1_728 = __builtin_shufflevector(__s1_728, __s1_728, 1, 0); \ + int64x2_t __ret_728; \ + __ret_728 = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0_728), __noswap_splat_lane_s32(__rev1_728, __p2_728)); \ + __ret_728 = __builtin_shufflevector(__ret_728, __ret_728, 1, 0); \ + __ret_728; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmull_high_lane_s16(__p0_637, __p1_637, __p2_637) __extension__ ({ \ - int16x8_t __s0_637 = __p0_637; \ - int16x4_t __s1_637 = __p1_637; \ - int32x4_t __ret_637; \ - __ret_637 = vqdmull_s16(vget_high_s16(__s0_637), splat_lane_s16(__s1_637, __p2_637)); \ - __ret_637; \ +#define vqdmull_high_lane_s16(__p0_729, __p1_729, __p2_729) __extension__ ({ \ + int16x8_t __s0_729 = __p0_729; \ + int16x4_t __s1_729 = __p1_729; \ + int32x4_t __ret_729; \ + __ret_729 = vqdmull_s16(vget_high_s16(__s0_729), splat_lane_s16(__s1_729, __p2_729)); \ + __ret_729; \ }) #else -#define vqdmull_high_lane_s16(__p0_638, __p1_638, __p2_638) __extension__ ({ \ - int16x8_t __s0_638 = __p0_638; \ - int16x4_t __s1_638 = __p1_638; \ - int16x8_t __rev0_638; __rev0_638 = __builtin_shufflevector(__s0_638, __s0_638, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x4_t __rev1_638; __rev1_638 = __builtin_shufflevector(__s1_638, __s1_638, 3, 2, 1, 0); \ - int32x4_t __ret_638; \ - __ret_638 = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0_638), __noswap_splat_lane_s16(__rev1_638, __p2_638)); \ - __ret_638 = __builtin_shufflevector(__ret_638, __ret_638, 3, 2, 1, 0); \ - __ret_638; \ +#define vqdmull_high_lane_s16(__p0_730, __p1_730, __p2_730) __extension__ ({ \ + int16x8_t __s0_730 = __p0_730; \ + int16x4_t __s1_730 = __p1_730; \ + int16x8_t __rev0_730; __rev0_730 = __builtin_shufflevector(__s0_730, __s0_730, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x4_t __rev1_730; __rev1_730 = __builtin_shufflevector(__s1_730, __s1_730, 3, 2, 1, 0); \ + int32x4_t __ret_730; \ + __ret_730 = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0_730), __noswap_splat_lane_s16(__rev1_730, __p2_730)); \ + __ret_730 = __builtin_shufflevector(__ret_730, __ret_730, 3, 2, 1, 0); \ + __ret_730; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmull_high_laneq_s32(__p0_639, __p1_639, __p2_639) __extension__ ({ \ - int32x4_t __s0_639 = __p0_639; \ - int32x4_t __s1_639 = __p1_639; \ - int64x2_t __ret_639; \ - __ret_639 = vqdmull_s32(vget_high_s32(__s0_639), splat_laneq_s32(__s1_639, __p2_639)); \ - __ret_639; \ +#define vqdmull_high_laneq_s32(__p0_731, __p1_731, __p2_731) __extension__ ({ \ + int32x4_t __s0_731 = __p0_731; \ + int32x4_t __s1_731 = __p1_731; \ + int64x2_t __ret_731; \ + __ret_731 = vqdmull_s32(vget_high_s32(__s0_731), splat_laneq_s32(__s1_731, __p2_731)); \ + __ret_731; \ }) #else -#define vqdmull_high_laneq_s32(__p0_640, __p1_640, __p2_640) __extension__ ({ \ - int32x4_t __s0_640 = __p0_640; \ - int32x4_t __s1_640 = __p1_640; \ - int32x4_t __rev0_640; __rev0_640 = __builtin_shufflevector(__s0_640, __s0_640, 3, 2, 1, 0); \ - int32x4_t __rev1_640; __rev1_640 = __builtin_shufflevector(__s1_640, __s1_640, 3, 2, 1, 0); \ - int64x2_t __ret_640; \ - __ret_640 = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0_640), __noswap_splat_laneq_s32(__rev1_640, __p2_640)); \ - __ret_640 = __builtin_shufflevector(__ret_640, __ret_640, 1, 0); \ - __ret_640; \ +#define vqdmull_high_laneq_s32(__p0_732, __p1_732, __p2_732) __extension__ ({ \ + int32x4_t __s0_732 = __p0_732; \ + int32x4_t __s1_732 = __p1_732; \ + int32x4_t __rev0_732; __rev0_732 = __builtin_shufflevector(__s0_732, __s0_732, 3, 2, 1, 0); \ + int32x4_t __rev1_732; __rev1_732 = __builtin_shufflevector(__s1_732, __s1_732, 3, 2, 1, 0); \ + int64x2_t __ret_732; \ + __ret_732 = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0_732), __noswap_splat_laneq_s32(__rev1_732, __p2_732)); \ + __ret_732 = __builtin_shufflevector(__ret_732, __ret_732, 1, 0); \ + __ret_732; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmull_high_laneq_s16(__p0_641, __p1_641, __p2_641) __extension__ ({ \ - int16x8_t __s0_641 = __p0_641; \ - int16x8_t __s1_641 = __p1_641; \ - int32x4_t __ret_641; \ - __ret_641 = vqdmull_s16(vget_high_s16(__s0_641), splat_laneq_s16(__s1_641, __p2_641)); \ - __ret_641; \ +#define vqdmull_high_laneq_s16(__p0_733, __p1_733, __p2_733) __extension__ ({ \ + int16x8_t __s0_733 = __p0_733; \ + int16x8_t __s1_733 = __p1_733; \ + int32x4_t __ret_733; \ + __ret_733 = vqdmull_s16(vget_high_s16(__s0_733), splat_laneq_s16(__s1_733, __p2_733)); \ + __ret_733; \ }) #else -#define vqdmull_high_laneq_s16(__p0_642, __p1_642, __p2_642) __extension__ ({ \ - int16x8_t __s0_642 = __p0_642; \ - int16x8_t __s1_642 = __p1_642; \ - int16x8_t __rev0_642; __rev0_642 = __builtin_shufflevector(__s0_642, __s0_642, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_642; __rev1_642 = __builtin_shufflevector(__s1_642, __s1_642, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_642; \ - __ret_642 = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0_642), __noswap_splat_laneq_s16(__rev1_642, __p2_642)); \ - __ret_642 = __builtin_shufflevector(__ret_642, __ret_642, 3, 2, 1, 0); \ - __ret_642; \ +#define vqdmull_high_laneq_s16(__p0_734, __p1_734, __p2_734) __extension__ ({ \ + int16x8_t __s0_734 = __p0_734; \ + int16x8_t __s1_734 = __p1_734; \ + int16x8_t __rev0_734; __rev0_734 = __builtin_shufflevector(__s0_734, __s0_734, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_734; __rev1_734 = __builtin_shufflevector(__s1_734, __s1_734, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_734; \ + __ret_734 = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0_734), __noswap_splat_laneq_s16(__rev1_734, __p2_734)); \ + __ret_734 = __builtin_shufflevector(__ret_734, __ret_734, 3, 2, 1, 0); \ + __ret_734; \ }) #endif @@ -56939,120 +58609,120 @@ __ai int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmulls_lane_s32(__p0_643, __p1_643, __p2_643) __extension__ ({ \ - int32_t __s0_643 = __p0_643; \ - int32x2_t __s1_643 = __p1_643; \ - int64_t __ret_643; \ - __ret_643 = vqdmulls_s32(__s0_643, vget_lane_s32(__s1_643, __p2_643)); \ - __ret_643; \ +#define vqdmulls_lane_s32(__p0_735, __p1_735, __p2_735) __extension__ ({ \ + int32_t __s0_735 = __p0_735; \ + int32x2_t __s1_735 = __p1_735; \ + int64_t __ret_735; \ + __ret_735 = vqdmulls_s32(__s0_735, vget_lane_s32(__s1_735, __p2_735)); \ + __ret_735; \ }) #else -#define vqdmulls_lane_s32(__p0_644, __p1_644, __p2_644) __extension__ ({ \ - int32_t __s0_644 = __p0_644; \ - int32x2_t __s1_644 = __p1_644; \ - int32x2_t __rev1_644; __rev1_644 = __builtin_shufflevector(__s1_644, __s1_644, 1, 0); \ - int64_t __ret_644; \ - __ret_644 = vqdmulls_s32(__s0_644, __noswap_vget_lane_s32(__rev1_644, __p2_644)); \ - __ret_644; \ +#define vqdmulls_lane_s32(__p0_736, __p1_736, __p2_736) __extension__ ({ \ + int32_t __s0_736 = __p0_736; \ + int32x2_t __s1_736 = __p1_736; \ + int32x2_t __rev1_736; __rev1_736 = __builtin_shufflevector(__s1_736, __s1_736, 1, 0); \ + int64_t __ret_736; \ + __ret_736 = vqdmulls_s32(__s0_736, __noswap_vget_lane_s32(__rev1_736, __p2_736)); \ + __ret_736; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmullh_lane_s16(__p0_645, __p1_645, __p2_645) __extension__ ({ \ - int16_t __s0_645 = __p0_645; \ - int16x4_t __s1_645 = __p1_645; \ - int32_t __ret_645; \ - __ret_645 = vqdmullh_s16(__s0_645, vget_lane_s16(__s1_645, __p2_645)); \ - __ret_645; \ +#define vqdmullh_lane_s16(__p0_737, __p1_737, __p2_737) __extension__ ({ \ + int16_t __s0_737 = __p0_737; \ + int16x4_t __s1_737 = __p1_737; \ + int32_t __ret_737; \ + __ret_737 = vqdmullh_s16(__s0_737, vget_lane_s16(__s1_737, __p2_737)); \ + __ret_737; \ }) #else -#define vqdmullh_lane_s16(__p0_646, __p1_646, __p2_646) __extension__ ({ \ - int16_t __s0_646 = __p0_646; \ - int16x4_t __s1_646 = __p1_646; \ - int16x4_t __rev1_646; __rev1_646 = __builtin_shufflevector(__s1_646, __s1_646, 3, 2, 1, 0); \ - int32_t __ret_646; \ - __ret_646 = vqdmullh_s16(__s0_646, __noswap_vget_lane_s16(__rev1_646, __p2_646)); \ - __ret_646; \ +#define vqdmullh_lane_s16(__p0_738, __p1_738, __p2_738) __extension__ ({ \ + int16_t __s0_738 = __p0_738; \ + int16x4_t __s1_738 = __p1_738; \ + int16x4_t __rev1_738; __rev1_738 = __builtin_shufflevector(__s1_738, __s1_738, 3, 2, 1, 0); \ + int32_t __ret_738; \ + __ret_738 = vqdmullh_s16(__s0_738, __noswap_vget_lane_s16(__rev1_738, __p2_738)); \ + __ret_738; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmulls_laneq_s32(__p0_647, __p1_647, __p2_647) __extension__ ({ \ - int32_t __s0_647 = __p0_647; \ - int32x4_t __s1_647 = __p1_647; \ - int64_t __ret_647; \ - __ret_647 = vqdmulls_s32(__s0_647, vgetq_lane_s32(__s1_647, __p2_647)); \ - __ret_647; \ +#define vqdmulls_laneq_s32(__p0_739, __p1_739, __p2_739) __extension__ ({ \ + int32_t __s0_739 = __p0_739; \ + int32x4_t __s1_739 = __p1_739; \ + int64_t __ret_739; \ + __ret_739 = vqdmulls_s32(__s0_739, vgetq_lane_s32(__s1_739, __p2_739)); \ + __ret_739; \ }) #else -#define vqdmulls_laneq_s32(__p0_648, __p1_648, __p2_648) __extension__ ({ \ - int32_t __s0_648 = __p0_648; \ - int32x4_t __s1_648 = __p1_648; \ - int32x4_t __rev1_648; __rev1_648 = __builtin_shufflevector(__s1_648, __s1_648, 3, 2, 1, 0); \ - int64_t __ret_648; \ - __ret_648 = vqdmulls_s32(__s0_648, __noswap_vgetq_lane_s32(__rev1_648, __p2_648)); \ - __ret_648; \ +#define vqdmulls_laneq_s32(__p0_740, __p1_740, __p2_740) __extension__ ({ \ + int32_t __s0_740 = __p0_740; \ + int32x4_t __s1_740 = __p1_740; \ + int32x4_t __rev1_740; __rev1_740 = __builtin_shufflevector(__s1_740, __s1_740, 3, 2, 1, 0); \ + int64_t __ret_740; \ + __ret_740 = vqdmulls_s32(__s0_740, __noswap_vgetq_lane_s32(__rev1_740, __p2_740)); \ + __ret_740; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmullh_laneq_s16(__p0_649, __p1_649, __p2_649) __extension__ ({ \ - int16_t __s0_649 = __p0_649; \ - int16x8_t __s1_649 = __p1_649; \ - int32_t __ret_649; \ - __ret_649 = vqdmullh_s16(__s0_649, vgetq_lane_s16(__s1_649, __p2_649)); \ - __ret_649; \ +#define vqdmullh_laneq_s16(__p0_741, __p1_741, __p2_741) __extension__ ({ \ + int16_t __s0_741 = __p0_741; \ + int16x8_t __s1_741 = __p1_741; \ + int32_t __ret_741; \ + __ret_741 = vqdmullh_s16(__s0_741, vgetq_lane_s16(__s1_741, __p2_741)); \ + __ret_741; \ }) #else -#define vqdmullh_laneq_s16(__p0_650, __p1_650, __p2_650) __extension__ ({ \ - int16_t __s0_650 = __p0_650; \ - int16x8_t __s1_650 = __p1_650; \ - int16x8_t __rev1_650; __rev1_650 = __builtin_shufflevector(__s1_650, __s1_650, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32_t __ret_650; \ - __ret_650 = vqdmullh_s16(__s0_650, __noswap_vgetq_lane_s16(__rev1_650, __p2_650)); \ - __ret_650; \ +#define vqdmullh_laneq_s16(__p0_742, __p1_742, __p2_742) __extension__ ({ \ + int16_t __s0_742 = __p0_742; \ + int16x8_t __s1_742 = __p1_742; \ + int16x8_t __rev1_742; __rev1_742 = __builtin_shufflevector(__s1_742, __s1_742, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32_t __ret_742; \ + __ret_742 = vqdmullh_s16(__s0_742, __noswap_vgetq_lane_s16(__rev1_742, __p2_742)); \ + __ret_742; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmull_laneq_s32(__p0_651, __p1_651, __p2_651) __extension__ ({ \ - int32x2_t __s0_651 = __p0_651; \ - int32x4_t __s1_651 = __p1_651; \ - int64x2_t __ret_651; \ - __ret_651 = vqdmull_s32(__s0_651, splat_laneq_s32(__s1_651, __p2_651)); \ - __ret_651; \ +#define vqdmull_laneq_s32(__p0_743, __p1_743, __p2_743) __extension__ ({ \ + int32x2_t __s0_743 = __p0_743; \ + int32x4_t __s1_743 = __p1_743; \ + int64x2_t __ret_743; \ + __ret_743 = vqdmull_s32(__s0_743, splat_laneq_s32(__s1_743, __p2_743)); \ + __ret_743; \ }) #else -#define vqdmull_laneq_s32(__p0_652, __p1_652, __p2_652) __extension__ ({ \ - int32x2_t __s0_652 = __p0_652; \ - int32x4_t __s1_652 = __p1_652; \ - int32x2_t __rev0_652; __rev0_652 = __builtin_shufflevector(__s0_652, __s0_652, 1, 0); \ - int32x4_t __rev1_652; __rev1_652 = __builtin_shufflevector(__s1_652, __s1_652, 3, 2, 1, 0); \ - int64x2_t __ret_652; \ - __ret_652 = __noswap_vqdmull_s32(__rev0_652, __noswap_splat_laneq_s32(__rev1_652, __p2_652)); \ - __ret_652 = __builtin_shufflevector(__ret_652, __ret_652, 1, 0); \ - __ret_652; \ +#define vqdmull_laneq_s32(__p0_744, __p1_744, __p2_744) __extension__ ({ \ + int32x2_t __s0_744 = __p0_744; \ + int32x4_t __s1_744 = __p1_744; \ + int32x2_t __rev0_744; __rev0_744 = __builtin_shufflevector(__s0_744, __s0_744, 1, 0); \ + int32x4_t __rev1_744; __rev1_744 = __builtin_shufflevector(__s1_744, __s1_744, 3, 2, 1, 0); \ + int64x2_t __ret_744; \ + __ret_744 = __noswap_vqdmull_s32(__rev0_744, __noswap_splat_laneq_s32(__rev1_744, __p2_744)); \ + __ret_744 = __builtin_shufflevector(__ret_744, __ret_744, 1, 0); \ + __ret_744; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqdmull_laneq_s16(__p0_653, __p1_653, __p2_653) __extension__ ({ \ - int16x4_t __s0_653 = __p0_653; \ - int16x8_t __s1_653 = __p1_653; \ - int32x4_t __ret_653; \ - __ret_653 = vqdmull_s16(__s0_653, splat_laneq_s16(__s1_653, __p2_653)); \ - __ret_653; \ +#define vqdmull_laneq_s16(__p0_745, __p1_745, __p2_745) __extension__ ({ \ + int16x4_t __s0_745 = __p0_745; \ + int16x8_t __s1_745 = __p1_745; \ + int32x4_t __ret_745; \ + __ret_745 = vqdmull_s16(__s0_745, splat_laneq_s16(__s1_745, __p2_745)); \ + __ret_745; \ }) #else -#define vqdmull_laneq_s16(__p0_654, __p1_654, __p2_654) __extension__ ({ \ - int16x4_t __s0_654 = __p0_654; \ - int16x8_t __s1_654 = __p1_654; \ - int16x4_t __rev0_654; __rev0_654 = __builtin_shufflevector(__s0_654, __s0_654, 3, 2, 1, 0); \ - int16x8_t __rev1_654; __rev1_654 = __builtin_shufflevector(__s1_654, __s1_654, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_654; \ - __ret_654 = __noswap_vqdmull_s16(__rev0_654, __noswap_splat_laneq_s16(__rev1_654, __p2_654)); \ - __ret_654 = __builtin_shufflevector(__ret_654, __ret_654, 3, 2, 1, 0); \ - __ret_654; \ +#define vqdmull_laneq_s16(__p0_746, __p1_746, __p2_746) __extension__ ({ \ + int16x4_t __s0_746 = __p0_746; \ + int16x8_t __s1_746 = __p1_746; \ + int16x4_t __rev0_746; __rev0_746 = __builtin_shufflevector(__s0_746, __s0_746, 3, 2, 1, 0); \ + int16x8_t __rev1_746; __rev1_746 = __builtin_shufflevector(__s1_746, __s1_746, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_746; \ + __ret_746 = __noswap_vqdmull_s16(__rev0_746, __noswap_splat_laneq_s16(__rev1_746, __p2_746)); \ + __ret_746 = __builtin_shufflevector(__ret_746, __ret_746, 3, 2, 1, 0); \ + __ret_746; \ }) #endif @@ -57188,30 +58858,30 @@ __ai int8x16_t vqmovn_high_s16(int8x8_t __p0, int16x8_t __p1) { } #endif -__ai int16_t vqmovuns_s32(int32_t __p0) { - int16_t __ret; - __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0); +__ai uint16_t vqmovuns_s32(int32_t __p0) { + uint16_t __ret; + __ret = (uint16_t) __builtin_neon_vqmovuns_s32(__p0); return __ret; } -__ai int32_t vqmovund_s64(int64_t __p0) { - int32_t __ret; - __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0); +__ai uint32_t vqmovund_s64(int64_t __p0) { + uint32_t __ret; + __ret = (uint32_t) __builtin_neon_vqmovund_s64(__p0); return __ret; } -__ai int8_t vqmovunh_s16(int16_t __p0) { - int8_t __ret; - __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0); +__ai uint8_t vqmovunh_s16(int16_t __p0) { + uint8_t __ret; + __ret = (uint8_t) __builtin_neon_vqmovunh_s16(__p0); return __ret; } #ifdef __LITTLE_ENDIAN__ -__ai uint16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) { +__ai uint16x8_t vqmovun_high_s32(uint16x4_t __p0, int32x4_t __p1) { uint16x8_t __ret; __ret = vcombine_u16((uint16x4_t)(__p0), vqmovun_s32(__p1)); return __ret; } #else -__ai uint16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) { - int16x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); +__ai uint16x8_t vqmovun_high_s32(uint16x4_t __p0, int32x4_t __p1) { + uint16x4_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0); int32x4_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0); uint16x8_t __ret; __ret = __noswap_vcombine_u16((uint16x4_t)(__rev0), __noswap_vqmovun_s32(__rev1)); @@ -57221,14 +58891,14 @@ __ai uint16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -__ai uint32x4_t vqmovun_high_s64(int32x2_t __p0, int64x2_t __p1) { +__ai uint32x4_t vqmovun_high_s64(uint32x2_t __p0, int64x2_t __p1) { uint32x4_t __ret; __ret = vcombine_u32((uint32x2_t)(__p0), vqmovun_s64(__p1)); return __ret; } #else -__ai uint32x4_t vqmovun_high_s64(int32x2_t __p0, int64x2_t __p1) { - int32x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); +__ai uint32x4_t vqmovun_high_s64(uint32x2_t __p0, int64x2_t __p1) { + uint32x2_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0); int64x2_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0); uint32x4_t __ret; __ret = __noswap_vcombine_u32((uint32x2_t)(__rev0), __noswap_vqmovun_s64(__rev1)); @@ -57238,14 +58908,14 @@ __ai uint32x4_t vqmovun_high_s64(int32x2_t __p0, int64x2_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -__ai uint8x16_t vqmovun_high_s16(int8x8_t __p0, int16x8_t __p1) { +__ai uint8x16_t vqmovun_high_s16(uint8x8_t __p0, int16x8_t __p1) { uint8x16_t __ret; __ret = vcombine_u8((uint8x8_t)(__p0), vqmovun_s16(__p1)); return __ret; } #else -__ai uint8x16_t vqmovun_high_s16(int8x8_t __p0, int16x8_t __p1) { - int8x8_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0); +__ai uint8x16_t vqmovun_high_s16(uint8x8_t __p0, int16x8_t __p1) { + uint8x8_t __rev0; __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0); int16x8_t __rev1; __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0); uint8x16_t __ret; __ret = __noswap_vcombine_u8((uint8x8_t)(__rev0), __noswap_vqmovun_s16(__rev1)); @@ -57390,78 +59060,78 @@ __ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmulhs_lane_s32(__p0_655, __p1_655, __p2_655) __extension__ ({ \ - int32_t __s0_655 = __p0_655; \ - int32x2_t __s1_655 = __p1_655; \ - int32_t __ret_655; \ - __ret_655 = vqrdmulhs_s32(__s0_655, vget_lane_s32(__s1_655, __p2_655)); \ - __ret_655; \ +#define vqrdmulhs_lane_s32(__p0_747, __p1_747, __p2_747) __extension__ ({ \ + int32_t __s0_747 = __p0_747; \ + int32x2_t __s1_747 = __p1_747; \ + int32_t __ret_747; \ + __ret_747 = vqrdmulhs_s32(__s0_747, vget_lane_s32(__s1_747, __p2_747)); \ + __ret_747; \ }) #else -#define vqrdmulhs_lane_s32(__p0_656, __p1_656, __p2_656) __extension__ ({ \ - int32_t __s0_656 = __p0_656; \ - int32x2_t __s1_656 = __p1_656; \ - int32x2_t __rev1_656; __rev1_656 = __builtin_shufflevector(__s1_656, __s1_656, 1, 0); \ - int32_t __ret_656; \ - __ret_656 = vqrdmulhs_s32(__s0_656, __noswap_vget_lane_s32(__rev1_656, __p2_656)); \ - __ret_656; \ +#define vqrdmulhs_lane_s32(__p0_748, __p1_748, __p2_748) __extension__ ({ \ + int32_t __s0_748 = __p0_748; \ + int32x2_t __s1_748 = __p1_748; \ + int32x2_t __rev1_748; __rev1_748 = __builtin_shufflevector(__s1_748, __s1_748, 1, 0); \ + int32_t __ret_748; \ + __ret_748 = vqrdmulhs_s32(__s0_748, __noswap_vget_lane_s32(__rev1_748, __p2_748)); \ + __ret_748; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmulhh_lane_s16(__p0_657, __p1_657, __p2_657) __extension__ ({ \ - int16_t __s0_657 = __p0_657; \ - int16x4_t __s1_657 = __p1_657; \ - int16_t __ret_657; \ - __ret_657 = vqrdmulhh_s16(__s0_657, vget_lane_s16(__s1_657, __p2_657)); \ - __ret_657; \ +#define vqrdmulhh_lane_s16(__p0_749, __p1_749, __p2_749) __extension__ ({ \ + int16_t __s0_749 = __p0_749; \ + int16x4_t __s1_749 = __p1_749; \ + int16_t __ret_749; \ + __ret_749 = vqrdmulhh_s16(__s0_749, vget_lane_s16(__s1_749, __p2_749)); \ + __ret_749; \ }) #else -#define vqrdmulhh_lane_s16(__p0_658, __p1_658, __p2_658) __extension__ ({ \ - int16_t __s0_658 = __p0_658; \ - int16x4_t __s1_658 = __p1_658; \ - int16x4_t __rev1_658; __rev1_658 = __builtin_shufflevector(__s1_658, __s1_658, 3, 2, 1, 0); \ - int16_t __ret_658; \ - __ret_658 = vqrdmulhh_s16(__s0_658, __noswap_vget_lane_s16(__rev1_658, __p2_658)); \ - __ret_658; \ +#define vqrdmulhh_lane_s16(__p0_750, __p1_750, __p2_750) __extension__ ({ \ + int16_t __s0_750 = __p0_750; \ + int16x4_t __s1_750 = __p1_750; \ + int16x4_t __rev1_750; __rev1_750 = __builtin_shufflevector(__s1_750, __s1_750, 3, 2, 1, 0); \ + int16_t __ret_750; \ + __ret_750 = vqrdmulhh_s16(__s0_750, __noswap_vget_lane_s16(__rev1_750, __p2_750)); \ + __ret_750; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmulhs_laneq_s32(__p0_659, __p1_659, __p2_659) __extension__ ({ \ - int32_t __s0_659 = __p0_659; \ - int32x4_t __s1_659 = __p1_659; \ - int32_t __ret_659; \ - __ret_659 = vqrdmulhs_s32(__s0_659, vgetq_lane_s32(__s1_659, __p2_659)); \ - __ret_659; \ +#define vqrdmulhs_laneq_s32(__p0_751, __p1_751, __p2_751) __extension__ ({ \ + int32_t __s0_751 = __p0_751; \ + int32x4_t __s1_751 = __p1_751; \ + int32_t __ret_751; \ + __ret_751 = vqrdmulhs_s32(__s0_751, vgetq_lane_s32(__s1_751, __p2_751)); \ + __ret_751; \ }) #else -#define vqrdmulhs_laneq_s32(__p0_660, __p1_660, __p2_660) __extension__ ({ \ - int32_t __s0_660 = __p0_660; \ - int32x4_t __s1_660 = __p1_660; \ - int32x4_t __rev1_660; __rev1_660 = __builtin_shufflevector(__s1_660, __s1_660, 3, 2, 1, 0); \ - int32_t __ret_660; \ - __ret_660 = vqrdmulhs_s32(__s0_660, __noswap_vgetq_lane_s32(__rev1_660, __p2_660)); \ - __ret_660; \ +#define vqrdmulhs_laneq_s32(__p0_752, __p1_752, __p2_752) __extension__ ({ \ + int32_t __s0_752 = __p0_752; \ + int32x4_t __s1_752 = __p1_752; \ + int32x4_t __rev1_752; __rev1_752 = __builtin_shufflevector(__s1_752, __s1_752, 3, 2, 1, 0); \ + int32_t __ret_752; \ + __ret_752 = vqrdmulhs_s32(__s0_752, __noswap_vgetq_lane_s32(__rev1_752, __p2_752)); \ + __ret_752; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmulhh_laneq_s16(__p0_661, __p1_661, __p2_661) __extension__ ({ \ - int16_t __s0_661 = __p0_661; \ - int16x8_t __s1_661 = __p1_661; \ - int16_t __ret_661; \ - __ret_661 = vqrdmulhh_s16(__s0_661, vgetq_lane_s16(__s1_661, __p2_661)); \ - __ret_661; \ +#define vqrdmulhh_laneq_s16(__p0_753, __p1_753, __p2_753) __extension__ ({ \ + int16_t __s0_753 = __p0_753; \ + int16x8_t __s1_753 = __p1_753; \ + int16_t __ret_753; \ + __ret_753 = vqrdmulhh_s16(__s0_753, vgetq_lane_s16(__s1_753, __p2_753)); \ + __ret_753; \ }) #else -#define vqrdmulhh_laneq_s16(__p0_662, __p1_662, __p2_662) __extension__ ({ \ - int16_t __s0_662 = __p0_662; \ - int16x8_t __s1_662 = __p1_662; \ - int16x8_t __rev1_662; __rev1_662 = __builtin_shufflevector(__s1_662, __s1_662, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16_t __ret_662; \ - __ret_662 = vqrdmulhh_s16(__s0_662, __noswap_vgetq_lane_s16(__rev1_662, __p2_662)); \ - __ret_662; \ +#define vqrdmulhh_laneq_s16(__p0_754, __p1_754, __p2_754) __extension__ ({ \ + int16_t __s0_754 = __p0_754; \ + int16x8_t __s1_754 = __p1_754; \ + int16x8_t __rev1_754; __rev1_754 = __builtin_shufflevector(__s1_754, __s1_754, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16_t __ret_754; \ + __ret_754 = vqrdmulhh_s16(__s0_754, __noswap_vgetq_lane_s16(__rev1_754, __p2_754)); \ + __ret_754; \ }) #endif @@ -57549,22 +59219,22 @@ __ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) { }) #endif -__ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) { +__ai uint8_t vqrshlb_u8(uint8_t __p0, int8_t __p1) { uint8_t __ret; __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1); return __ret; } -__ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) { +__ai uint32_t vqrshls_u32(uint32_t __p0, int32_t __p1) { uint32_t __ret; __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1); return __ret; } -__ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) { +__ai uint64_t vqrshld_u64(uint64_t __p0, int64_t __p1) { uint64_t __ret; __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1); return __ret; } -__ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) { +__ai uint16_t vqrshlh_u16(uint16_t __p0, int16_t __p1) { uint16_t __ret; __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1); return __ret; @@ -57590,128 +59260,128 @@ __ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) { return __ret; } #ifdef __LITTLE_ENDIAN__ -#define vqrshrn_high_n_u32(__p0_663, __p1_663, __p2_663) __extension__ ({ \ - uint16x4_t __s0_663 = __p0_663; \ - uint32x4_t __s1_663 = __p1_663; \ - uint16x8_t __ret_663; \ - __ret_663 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_663), (uint16x4_t)(vqrshrn_n_u32(__s1_663, __p2_663)))); \ - __ret_663; \ +#define vqrshrn_high_n_u32(__p0_755, __p1_755, __p2_755) __extension__ ({ \ + uint16x4_t __s0_755 = __p0_755; \ + uint32x4_t __s1_755 = __p1_755; \ + uint16x8_t __ret_755; \ + __ret_755 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_755), (uint16x4_t)(vqrshrn_n_u32(__s1_755, __p2_755)))); \ + __ret_755; \ }) #else -#define vqrshrn_high_n_u32(__p0_664, __p1_664, __p2_664) __extension__ ({ \ - uint16x4_t __s0_664 = __p0_664; \ - uint32x4_t __s1_664 = __p1_664; \ - uint16x4_t __rev0_664; __rev0_664 = __builtin_shufflevector(__s0_664, __s0_664, 3, 2, 1, 0); \ - uint32x4_t __rev1_664; __rev1_664 = __builtin_shufflevector(__s1_664, __s1_664, 3, 2, 1, 0); \ - uint16x8_t __ret_664; \ - __ret_664 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_664), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_664, __p2_664)))); \ - __ret_664 = __builtin_shufflevector(__ret_664, __ret_664, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_664; \ +#define vqrshrn_high_n_u32(__p0_756, __p1_756, __p2_756) __extension__ ({ \ + uint16x4_t __s0_756 = __p0_756; \ + uint32x4_t __s1_756 = __p1_756; \ + uint16x4_t __rev0_756; __rev0_756 = __builtin_shufflevector(__s0_756, __s0_756, 3, 2, 1, 0); \ + uint32x4_t __rev1_756; __rev1_756 = __builtin_shufflevector(__s1_756, __s1_756, 3, 2, 1, 0); \ + uint16x8_t __ret_756; \ + __ret_756 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_756), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_756, __p2_756)))); \ + __ret_756 = __builtin_shufflevector(__ret_756, __ret_756, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_756; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrshrn_high_n_u64(__p0_665, __p1_665, __p2_665) __extension__ ({ \ - uint32x2_t __s0_665 = __p0_665; \ - uint64x2_t __s1_665 = __p1_665; \ - uint32x4_t __ret_665; \ - __ret_665 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_665), (uint32x2_t)(vqrshrn_n_u64(__s1_665, __p2_665)))); \ - __ret_665; \ +#define vqrshrn_high_n_u64(__p0_757, __p1_757, __p2_757) __extension__ ({ \ + uint32x2_t __s0_757 = __p0_757; \ + uint64x2_t __s1_757 = __p1_757; \ + uint32x4_t __ret_757; \ + __ret_757 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_757), (uint32x2_t)(vqrshrn_n_u64(__s1_757, __p2_757)))); \ + __ret_757; \ }) #else -#define vqrshrn_high_n_u64(__p0_666, __p1_666, __p2_666) __extension__ ({ \ - uint32x2_t __s0_666 = __p0_666; \ - uint64x2_t __s1_666 = __p1_666; \ - uint32x2_t __rev0_666; __rev0_666 = __builtin_shufflevector(__s0_666, __s0_666, 1, 0); \ - uint64x2_t __rev1_666; __rev1_666 = __builtin_shufflevector(__s1_666, __s1_666, 1, 0); \ - uint32x4_t __ret_666; \ - __ret_666 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_666), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_666, __p2_666)))); \ - __ret_666 = __builtin_shufflevector(__ret_666, __ret_666, 3, 2, 1, 0); \ - __ret_666; \ +#define vqrshrn_high_n_u64(__p0_758, __p1_758, __p2_758) __extension__ ({ \ + uint32x2_t __s0_758 = __p0_758; \ + uint64x2_t __s1_758 = __p1_758; \ + uint32x2_t __rev0_758; __rev0_758 = __builtin_shufflevector(__s0_758, __s0_758, 1, 0); \ + uint64x2_t __rev1_758; __rev1_758 = __builtin_shufflevector(__s1_758, __s1_758, 1, 0); \ + uint32x4_t __ret_758; \ + __ret_758 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_758), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_758, __p2_758)))); \ + __ret_758 = __builtin_shufflevector(__ret_758, __ret_758, 3, 2, 1, 0); \ + __ret_758; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrshrn_high_n_u16(__p0_667, __p1_667, __p2_667) __extension__ ({ \ - uint8x8_t __s0_667 = __p0_667; \ - uint16x8_t __s1_667 = __p1_667; \ - uint8x16_t __ret_667; \ - __ret_667 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_667), (uint8x8_t)(vqrshrn_n_u16(__s1_667, __p2_667)))); \ - __ret_667; \ +#define vqrshrn_high_n_u16(__p0_759, __p1_759, __p2_759) __extension__ ({ \ + uint8x8_t __s0_759 = __p0_759; \ + uint16x8_t __s1_759 = __p1_759; \ + uint8x16_t __ret_759; \ + __ret_759 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_759), (uint8x8_t)(vqrshrn_n_u16(__s1_759, __p2_759)))); \ + __ret_759; \ }) #else -#define vqrshrn_high_n_u16(__p0_668, __p1_668, __p2_668) __extension__ ({ \ - uint8x8_t __s0_668 = __p0_668; \ - uint16x8_t __s1_668 = __p1_668; \ - uint8x8_t __rev0_668; __rev0_668 = __builtin_shufflevector(__s0_668, __s0_668, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_668; __rev1_668 = __builtin_shufflevector(__s1_668, __s1_668, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __ret_668; \ - __ret_668 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_668), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_668, __p2_668)))); \ - __ret_668 = __builtin_shufflevector(__ret_668, __ret_668, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_668; \ +#define vqrshrn_high_n_u16(__p0_760, __p1_760, __p2_760) __extension__ ({ \ + uint8x8_t __s0_760 = __p0_760; \ + uint16x8_t __s1_760 = __p1_760; \ + uint8x8_t __rev0_760; __rev0_760 = __builtin_shufflevector(__s0_760, __s0_760, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_760; __rev1_760 = __builtin_shufflevector(__s1_760, __s1_760, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __ret_760; \ + __ret_760 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_760), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_760, __p2_760)))); \ + __ret_760 = __builtin_shufflevector(__ret_760, __ret_760, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_760; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrshrn_high_n_s32(__p0_669, __p1_669, __p2_669) __extension__ ({ \ - int16x4_t __s0_669 = __p0_669; \ - int32x4_t __s1_669 = __p1_669; \ - int16x8_t __ret_669; \ - __ret_669 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_669), (int16x4_t)(vqrshrn_n_s32(__s1_669, __p2_669)))); \ - __ret_669; \ +#define vqrshrn_high_n_s32(__p0_761, __p1_761, __p2_761) __extension__ ({ \ + int16x4_t __s0_761 = __p0_761; \ + int32x4_t __s1_761 = __p1_761; \ + int16x8_t __ret_761; \ + __ret_761 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_761), (int16x4_t)(vqrshrn_n_s32(__s1_761, __p2_761)))); \ + __ret_761; \ }) #else -#define vqrshrn_high_n_s32(__p0_670, __p1_670, __p2_670) __extension__ ({ \ - int16x4_t __s0_670 = __p0_670; \ - int32x4_t __s1_670 = __p1_670; \ - int16x4_t __rev0_670; __rev0_670 = __builtin_shufflevector(__s0_670, __s0_670, 3, 2, 1, 0); \ - int32x4_t __rev1_670; __rev1_670 = __builtin_shufflevector(__s1_670, __s1_670, 3, 2, 1, 0); \ - int16x8_t __ret_670; \ - __ret_670 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_670), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_670, __p2_670)))); \ - __ret_670 = __builtin_shufflevector(__ret_670, __ret_670, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_670; \ +#define vqrshrn_high_n_s32(__p0_762, __p1_762, __p2_762) __extension__ ({ \ + int16x4_t __s0_762 = __p0_762; \ + int32x4_t __s1_762 = __p1_762; \ + int16x4_t __rev0_762; __rev0_762 = __builtin_shufflevector(__s0_762, __s0_762, 3, 2, 1, 0); \ + int32x4_t __rev1_762; __rev1_762 = __builtin_shufflevector(__s1_762, __s1_762, 3, 2, 1, 0); \ + int16x8_t __ret_762; \ + __ret_762 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_762), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_762, __p2_762)))); \ + __ret_762 = __builtin_shufflevector(__ret_762, __ret_762, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_762; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrshrn_high_n_s64(__p0_671, __p1_671, __p2_671) __extension__ ({ \ - int32x2_t __s0_671 = __p0_671; \ - int64x2_t __s1_671 = __p1_671; \ - int32x4_t __ret_671; \ - __ret_671 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_671), (int32x2_t)(vqrshrn_n_s64(__s1_671, __p2_671)))); \ - __ret_671; \ +#define vqrshrn_high_n_s64(__p0_763, __p1_763, __p2_763) __extension__ ({ \ + int32x2_t __s0_763 = __p0_763; \ + int64x2_t __s1_763 = __p1_763; \ + int32x4_t __ret_763; \ + __ret_763 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_763), (int32x2_t)(vqrshrn_n_s64(__s1_763, __p2_763)))); \ + __ret_763; \ }) #else -#define vqrshrn_high_n_s64(__p0_672, __p1_672, __p2_672) __extension__ ({ \ - int32x2_t __s0_672 = __p0_672; \ - int64x2_t __s1_672 = __p1_672; \ - int32x2_t __rev0_672; __rev0_672 = __builtin_shufflevector(__s0_672, __s0_672, 1, 0); \ - int64x2_t __rev1_672; __rev1_672 = __builtin_shufflevector(__s1_672, __s1_672, 1, 0); \ - int32x4_t __ret_672; \ - __ret_672 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_672), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_672, __p2_672)))); \ - __ret_672 = __builtin_shufflevector(__ret_672, __ret_672, 3, 2, 1, 0); \ - __ret_672; \ +#define vqrshrn_high_n_s64(__p0_764, __p1_764, __p2_764) __extension__ ({ \ + int32x2_t __s0_764 = __p0_764; \ + int64x2_t __s1_764 = __p1_764; \ + int32x2_t __rev0_764; __rev0_764 = __builtin_shufflevector(__s0_764, __s0_764, 1, 0); \ + int64x2_t __rev1_764; __rev1_764 = __builtin_shufflevector(__s1_764, __s1_764, 1, 0); \ + int32x4_t __ret_764; \ + __ret_764 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_764), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_764, __p2_764)))); \ + __ret_764 = __builtin_shufflevector(__ret_764, __ret_764, 3, 2, 1, 0); \ + __ret_764; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrshrn_high_n_s16(__p0_673, __p1_673, __p2_673) __extension__ ({ \ - int8x8_t __s0_673 = __p0_673; \ - int16x8_t __s1_673 = __p1_673; \ - int8x16_t __ret_673; \ - __ret_673 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_673), (int8x8_t)(vqrshrn_n_s16(__s1_673, __p2_673)))); \ - __ret_673; \ +#define vqrshrn_high_n_s16(__p0_765, __p1_765, __p2_765) __extension__ ({ \ + int8x8_t __s0_765 = __p0_765; \ + int16x8_t __s1_765 = __p1_765; \ + int8x16_t __ret_765; \ + __ret_765 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_765), (int8x8_t)(vqrshrn_n_s16(__s1_765, __p2_765)))); \ + __ret_765; \ }) #else -#define vqrshrn_high_n_s16(__p0_674, __p1_674, __p2_674) __extension__ ({ \ - int8x8_t __s0_674 = __p0_674; \ - int16x8_t __s1_674 = __p1_674; \ - int8x8_t __rev0_674; __rev0_674 = __builtin_shufflevector(__s0_674, __s0_674, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_674; __rev1_674 = __builtin_shufflevector(__s1_674, __s1_674, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_674; \ - __ret_674 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_674), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_674, __p2_674)))); \ - __ret_674 = __builtin_shufflevector(__ret_674, __ret_674, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_674; \ +#define vqrshrn_high_n_s16(__p0_766, __p1_766, __p2_766) __extension__ ({ \ + int8x8_t __s0_766 = __p0_766; \ + int16x8_t __s1_766 = __p1_766; \ + int8x8_t __rev0_766; __rev0_766 = __builtin_shufflevector(__s0_766, __s0_766, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_766; __rev1_766 = __builtin_shufflevector(__s1_766, __s1_766, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_766; \ + __ret_766 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_766), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_766, __p2_766)))); \ + __ret_766 = __builtin_shufflevector(__ret_766, __ret_766, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_766; \ }) #endif @@ -57752,65 +59422,65 @@ __ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) { __ret; \ }) #ifdef __LITTLE_ENDIAN__ -#define vqrshrun_high_n_s32(__p0_675, __p1_675, __p2_675) __extension__ ({ \ - int16x4_t __s0_675 = __p0_675; \ - int32x4_t __s1_675 = __p1_675; \ - int16x8_t __ret_675; \ - __ret_675 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_675), (int16x4_t)(vqrshrun_n_s32(__s1_675, __p2_675)))); \ - __ret_675; \ +#define vqrshrun_high_n_s32(__p0_767, __p1_767, __p2_767) __extension__ ({ \ + int16x4_t __s0_767 = __p0_767; \ + int32x4_t __s1_767 = __p1_767; \ + int16x8_t __ret_767; \ + __ret_767 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_767), (int16x4_t)(vqrshrun_n_s32(__s1_767, __p2_767)))); \ + __ret_767; \ }) #else -#define vqrshrun_high_n_s32(__p0_676, __p1_676, __p2_676) __extension__ ({ \ - int16x4_t __s0_676 = __p0_676; \ - int32x4_t __s1_676 = __p1_676; \ - int16x4_t __rev0_676; __rev0_676 = __builtin_shufflevector(__s0_676, __s0_676, 3, 2, 1, 0); \ - int32x4_t __rev1_676; __rev1_676 = __builtin_shufflevector(__s1_676, __s1_676, 3, 2, 1, 0); \ - int16x8_t __ret_676; \ - __ret_676 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_676), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_676, __p2_676)))); \ - __ret_676 = __builtin_shufflevector(__ret_676, __ret_676, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_676; \ +#define vqrshrun_high_n_s32(__p0_768, __p1_768, __p2_768) __extension__ ({ \ + int16x4_t __s0_768 = __p0_768; \ + int32x4_t __s1_768 = __p1_768; \ + int16x4_t __rev0_768; __rev0_768 = __builtin_shufflevector(__s0_768, __s0_768, 3, 2, 1, 0); \ + int32x4_t __rev1_768; __rev1_768 = __builtin_shufflevector(__s1_768, __s1_768, 3, 2, 1, 0); \ + int16x8_t __ret_768; \ + __ret_768 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_768), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_768, __p2_768)))); \ + __ret_768 = __builtin_shufflevector(__ret_768, __ret_768, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_768; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrshrun_high_n_s64(__p0_677, __p1_677, __p2_677) __extension__ ({ \ - int32x2_t __s0_677 = __p0_677; \ - int64x2_t __s1_677 = __p1_677; \ - int32x4_t __ret_677; \ - __ret_677 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_677), (int32x2_t)(vqrshrun_n_s64(__s1_677, __p2_677)))); \ - __ret_677; \ +#define vqrshrun_high_n_s64(__p0_769, __p1_769, __p2_769) __extension__ ({ \ + int32x2_t __s0_769 = __p0_769; \ + int64x2_t __s1_769 = __p1_769; \ + int32x4_t __ret_769; \ + __ret_769 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_769), (int32x2_t)(vqrshrun_n_s64(__s1_769, __p2_769)))); \ + __ret_769; \ }) #else -#define vqrshrun_high_n_s64(__p0_678, __p1_678, __p2_678) __extension__ ({ \ - int32x2_t __s0_678 = __p0_678; \ - int64x2_t __s1_678 = __p1_678; \ - int32x2_t __rev0_678; __rev0_678 = __builtin_shufflevector(__s0_678, __s0_678, 1, 0); \ - int64x2_t __rev1_678; __rev1_678 = __builtin_shufflevector(__s1_678, __s1_678, 1, 0); \ - int32x4_t __ret_678; \ - __ret_678 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_678), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_678, __p2_678)))); \ - __ret_678 = __builtin_shufflevector(__ret_678, __ret_678, 3, 2, 1, 0); \ - __ret_678; \ +#define vqrshrun_high_n_s64(__p0_770, __p1_770, __p2_770) __extension__ ({ \ + int32x2_t __s0_770 = __p0_770; \ + int64x2_t __s1_770 = __p1_770; \ + int32x2_t __rev0_770; __rev0_770 = __builtin_shufflevector(__s0_770, __s0_770, 1, 0); \ + int64x2_t __rev1_770; __rev1_770 = __builtin_shufflevector(__s1_770, __s1_770, 1, 0); \ + int32x4_t __ret_770; \ + __ret_770 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_770), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_770, __p2_770)))); \ + __ret_770 = __builtin_shufflevector(__ret_770, __ret_770, 3, 2, 1, 0); \ + __ret_770; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrshrun_high_n_s16(__p0_679, __p1_679, __p2_679) __extension__ ({ \ - int8x8_t __s0_679 = __p0_679; \ - int16x8_t __s1_679 = __p1_679; \ - int8x16_t __ret_679; \ - __ret_679 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_679), (int8x8_t)(vqrshrun_n_s16(__s1_679, __p2_679)))); \ - __ret_679; \ +#define vqrshrun_high_n_s16(__p0_771, __p1_771, __p2_771) __extension__ ({ \ + int8x8_t __s0_771 = __p0_771; \ + int16x8_t __s1_771 = __p1_771; \ + int8x16_t __ret_771; \ + __ret_771 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_771), (int8x8_t)(vqrshrun_n_s16(__s1_771, __p2_771)))); \ + __ret_771; \ }) #else -#define vqrshrun_high_n_s16(__p0_680, __p1_680, __p2_680) __extension__ ({ \ - int8x8_t __s0_680 = __p0_680; \ - int16x8_t __s1_680 = __p1_680; \ - int8x8_t __rev0_680; __rev0_680 = __builtin_shufflevector(__s0_680, __s0_680, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_680; __rev1_680 = __builtin_shufflevector(__s1_680, __s1_680, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_680; \ - __ret_680 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_680), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_680, __p2_680)))); \ - __ret_680 = __builtin_shufflevector(__ret_680, __ret_680, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_680; \ +#define vqrshrun_high_n_s16(__p0_772, __p1_772, __p2_772) __extension__ ({ \ + int8x8_t __s0_772 = __p0_772; \ + int16x8_t __s1_772 = __p1_772; \ + int8x8_t __rev0_772; __rev0_772 = __builtin_shufflevector(__s0_772, __s0_772, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_772; __rev1_772 = __builtin_shufflevector(__s1_772, __s1_772, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_772; \ + __ret_772 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_772), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_772, __p2_772)))); \ + __ret_772 = __builtin_shufflevector(__ret_772, __ret_772, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_772; \ }) #endif @@ -57832,22 +59502,22 @@ __ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) { __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \ __ret; \ }) -__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) { +__ai uint8_t vqshlb_u8(uint8_t __p0, int8_t __p1) { uint8_t __ret; __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1); return __ret; } -__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) { +__ai uint32_t vqshls_u32(uint32_t __p0, int32_t __p1) { uint32_t __ret; __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1); return __ret; } -__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) { +__ai uint64_t vqshld_u64(uint64_t __p0, int64_t __p1) { uint64_t __ret; __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1); return __ret; } -__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) { +__ai uint16_t vqshlh_u16(uint16_t __p0, int16_t __p1) { uint16_t __ret; __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1); return __ret; @@ -57945,128 +59615,128 @@ __ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) { __ret; \ }) #ifdef __LITTLE_ENDIAN__ -#define vqshrn_high_n_u32(__p0_681, __p1_681, __p2_681) __extension__ ({ \ - uint16x4_t __s0_681 = __p0_681; \ - uint32x4_t __s1_681 = __p1_681; \ - uint16x8_t __ret_681; \ - __ret_681 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_681), (uint16x4_t)(vqshrn_n_u32(__s1_681, __p2_681)))); \ - __ret_681; \ +#define vqshrn_high_n_u32(__p0_773, __p1_773, __p2_773) __extension__ ({ \ + uint16x4_t __s0_773 = __p0_773; \ + uint32x4_t __s1_773 = __p1_773; \ + uint16x8_t __ret_773; \ + __ret_773 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_773), (uint16x4_t)(vqshrn_n_u32(__s1_773, __p2_773)))); \ + __ret_773; \ }) #else -#define vqshrn_high_n_u32(__p0_682, __p1_682, __p2_682) __extension__ ({ \ - uint16x4_t __s0_682 = __p0_682; \ - uint32x4_t __s1_682 = __p1_682; \ - uint16x4_t __rev0_682; __rev0_682 = __builtin_shufflevector(__s0_682, __s0_682, 3, 2, 1, 0); \ - uint32x4_t __rev1_682; __rev1_682 = __builtin_shufflevector(__s1_682, __s1_682, 3, 2, 1, 0); \ - uint16x8_t __ret_682; \ - __ret_682 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_682), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_682, __p2_682)))); \ - __ret_682 = __builtin_shufflevector(__ret_682, __ret_682, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_682; \ +#define vqshrn_high_n_u32(__p0_774, __p1_774, __p2_774) __extension__ ({ \ + uint16x4_t __s0_774 = __p0_774; \ + uint32x4_t __s1_774 = __p1_774; \ + uint16x4_t __rev0_774; __rev0_774 = __builtin_shufflevector(__s0_774, __s0_774, 3, 2, 1, 0); \ + uint32x4_t __rev1_774; __rev1_774 = __builtin_shufflevector(__s1_774, __s1_774, 3, 2, 1, 0); \ + uint16x8_t __ret_774; \ + __ret_774 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_774), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_774, __p2_774)))); \ + __ret_774 = __builtin_shufflevector(__ret_774, __ret_774, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_774; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqshrn_high_n_u64(__p0_683, __p1_683, __p2_683) __extension__ ({ \ - uint32x2_t __s0_683 = __p0_683; \ - uint64x2_t __s1_683 = __p1_683; \ - uint32x4_t __ret_683; \ - __ret_683 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_683), (uint32x2_t)(vqshrn_n_u64(__s1_683, __p2_683)))); \ - __ret_683; \ +#define vqshrn_high_n_u64(__p0_775, __p1_775, __p2_775) __extension__ ({ \ + uint32x2_t __s0_775 = __p0_775; \ + uint64x2_t __s1_775 = __p1_775; \ + uint32x4_t __ret_775; \ + __ret_775 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_775), (uint32x2_t)(vqshrn_n_u64(__s1_775, __p2_775)))); \ + __ret_775; \ }) #else -#define vqshrn_high_n_u64(__p0_684, __p1_684, __p2_684) __extension__ ({ \ - uint32x2_t __s0_684 = __p0_684; \ - uint64x2_t __s1_684 = __p1_684; \ - uint32x2_t __rev0_684; __rev0_684 = __builtin_shufflevector(__s0_684, __s0_684, 1, 0); \ - uint64x2_t __rev1_684; __rev1_684 = __builtin_shufflevector(__s1_684, __s1_684, 1, 0); \ - uint32x4_t __ret_684; \ - __ret_684 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_684), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_684, __p2_684)))); \ - __ret_684 = __builtin_shufflevector(__ret_684, __ret_684, 3, 2, 1, 0); \ - __ret_684; \ +#define vqshrn_high_n_u64(__p0_776, __p1_776, __p2_776) __extension__ ({ \ + uint32x2_t __s0_776 = __p0_776; \ + uint64x2_t __s1_776 = __p1_776; \ + uint32x2_t __rev0_776; __rev0_776 = __builtin_shufflevector(__s0_776, __s0_776, 1, 0); \ + uint64x2_t __rev1_776; __rev1_776 = __builtin_shufflevector(__s1_776, __s1_776, 1, 0); \ + uint32x4_t __ret_776; \ + __ret_776 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_776), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_776, __p2_776)))); \ + __ret_776 = __builtin_shufflevector(__ret_776, __ret_776, 3, 2, 1, 0); \ + __ret_776; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqshrn_high_n_u16(__p0_685, __p1_685, __p2_685) __extension__ ({ \ - uint8x8_t __s0_685 = __p0_685; \ - uint16x8_t __s1_685 = __p1_685; \ - uint8x16_t __ret_685; \ - __ret_685 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_685), (uint8x8_t)(vqshrn_n_u16(__s1_685, __p2_685)))); \ - __ret_685; \ +#define vqshrn_high_n_u16(__p0_777, __p1_777, __p2_777) __extension__ ({ \ + uint8x8_t __s0_777 = __p0_777; \ + uint16x8_t __s1_777 = __p1_777; \ + uint8x16_t __ret_777; \ + __ret_777 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_777), (uint8x8_t)(vqshrn_n_u16(__s1_777, __p2_777)))); \ + __ret_777; \ }) #else -#define vqshrn_high_n_u16(__p0_686, __p1_686, __p2_686) __extension__ ({ \ - uint8x8_t __s0_686 = __p0_686; \ - uint16x8_t __s1_686 = __p1_686; \ - uint8x8_t __rev0_686; __rev0_686 = __builtin_shufflevector(__s0_686, __s0_686, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_686; __rev1_686 = __builtin_shufflevector(__s1_686, __s1_686, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __ret_686; \ - __ret_686 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_686), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_686, __p2_686)))); \ - __ret_686 = __builtin_shufflevector(__ret_686, __ret_686, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_686; \ +#define vqshrn_high_n_u16(__p0_778, __p1_778, __p2_778) __extension__ ({ \ + uint8x8_t __s0_778 = __p0_778; \ + uint16x8_t __s1_778 = __p1_778; \ + uint8x8_t __rev0_778; __rev0_778 = __builtin_shufflevector(__s0_778, __s0_778, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_778; __rev1_778 = __builtin_shufflevector(__s1_778, __s1_778, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __ret_778; \ + __ret_778 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_778), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_778, __p2_778)))); \ + __ret_778 = __builtin_shufflevector(__ret_778, __ret_778, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_778; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqshrn_high_n_s32(__p0_687, __p1_687, __p2_687) __extension__ ({ \ - int16x4_t __s0_687 = __p0_687; \ - int32x4_t __s1_687 = __p1_687; \ - int16x8_t __ret_687; \ - __ret_687 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_687), (int16x4_t)(vqshrn_n_s32(__s1_687, __p2_687)))); \ - __ret_687; \ +#define vqshrn_high_n_s32(__p0_779, __p1_779, __p2_779) __extension__ ({ \ + int16x4_t __s0_779 = __p0_779; \ + int32x4_t __s1_779 = __p1_779; \ + int16x8_t __ret_779; \ + __ret_779 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_779), (int16x4_t)(vqshrn_n_s32(__s1_779, __p2_779)))); \ + __ret_779; \ }) #else -#define vqshrn_high_n_s32(__p0_688, __p1_688, __p2_688) __extension__ ({ \ - int16x4_t __s0_688 = __p0_688; \ - int32x4_t __s1_688 = __p1_688; \ - int16x4_t __rev0_688; __rev0_688 = __builtin_shufflevector(__s0_688, __s0_688, 3, 2, 1, 0); \ - int32x4_t __rev1_688; __rev1_688 = __builtin_shufflevector(__s1_688, __s1_688, 3, 2, 1, 0); \ - int16x8_t __ret_688; \ - __ret_688 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_688), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_688, __p2_688)))); \ - __ret_688 = __builtin_shufflevector(__ret_688, __ret_688, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_688; \ +#define vqshrn_high_n_s32(__p0_780, __p1_780, __p2_780) __extension__ ({ \ + int16x4_t __s0_780 = __p0_780; \ + int32x4_t __s1_780 = __p1_780; \ + int16x4_t __rev0_780; __rev0_780 = __builtin_shufflevector(__s0_780, __s0_780, 3, 2, 1, 0); \ + int32x4_t __rev1_780; __rev1_780 = __builtin_shufflevector(__s1_780, __s1_780, 3, 2, 1, 0); \ + int16x8_t __ret_780; \ + __ret_780 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_780), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_780, __p2_780)))); \ + __ret_780 = __builtin_shufflevector(__ret_780, __ret_780, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_780; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqshrn_high_n_s64(__p0_689, __p1_689, __p2_689) __extension__ ({ \ - int32x2_t __s0_689 = __p0_689; \ - int64x2_t __s1_689 = __p1_689; \ - int32x4_t __ret_689; \ - __ret_689 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_689), (int32x2_t)(vqshrn_n_s64(__s1_689, __p2_689)))); \ - __ret_689; \ +#define vqshrn_high_n_s64(__p0_781, __p1_781, __p2_781) __extension__ ({ \ + int32x2_t __s0_781 = __p0_781; \ + int64x2_t __s1_781 = __p1_781; \ + int32x4_t __ret_781; \ + __ret_781 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_781), (int32x2_t)(vqshrn_n_s64(__s1_781, __p2_781)))); \ + __ret_781; \ }) #else -#define vqshrn_high_n_s64(__p0_690, __p1_690, __p2_690) __extension__ ({ \ - int32x2_t __s0_690 = __p0_690; \ - int64x2_t __s1_690 = __p1_690; \ - int32x2_t __rev0_690; __rev0_690 = __builtin_shufflevector(__s0_690, __s0_690, 1, 0); \ - int64x2_t __rev1_690; __rev1_690 = __builtin_shufflevector(__s1_690, __s1_690, 1, 0); \ - int32x4_t __ret_690; \ - __ret_690 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_690), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_690, __p2_690)))); \ - __ret_690 = __builtin_shufflevector(__ret_690, __ret_690, 3, 2, 1, 0); \ - __ret_690; \ +#define vqshrn_high_n_s64(__p0_782, __p1_782, __p2_782) __extension__ ({ \ + int32x2_t __s0_782 = __p0_782; \ + int64x2_t __s1_782 = __p1_782; \ + int32x2_t __rev0_782; __rev0_782 = __builtin_shufflevector(__s0_782, __s0_782, 1, 0); \ + int64x2_t __rev1_782; __rev1_782 = __builtin_shufflevector(__s1_782, __s1_782, 1, 0); \ + int32x4_t __ret_782; \ + __ret_782 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_782), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_782, __p2_782)))); \ + __ret_782 = __builtin_shufflevector(__ret_782, __ret_782, 3, 2, 1, 0); \ + __ret_782; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqshrn_high_n_s16(__p0_691, __p1_691, __p2_691) __extension__ ({ \ - int8x8_t __s0_691 = __p0_691; \ - int16x8_t __s1_691 = __p1_691; \ - int8x16_t __ret_691; \ - __ret_691 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_691), (int8x8_t)(vqshrn_n_s16(__s1_691, __p2_691)))); \ - __ret_691; \ +#define vqshrn_high_n_s16(__p0_783, __p1_783, __p2_783) __extension__ ({ \ + int8x8_t __s0_783 = __p0_783; \ + int16x8_t __s1_783 = __p1_783; \ + int8x16_t __ret_783; \ + __ret_783 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_783), (int8x8_t)(vqshrn_n_s16(__s1_783, __p2_783)))); \ + __ret_783; \ }) #else -#define vqshrn_high_n_s16(__p0_692, __p1_692, __p2_692) __extension__ ({ \ - int8x8_t __s0_692 = __p0_692; \ - int16x8_t __s1_692 = __p1_692; \ - int8x8_t __rev0_692; __rev0_692 = __builtin_shufflevector(__s0_692, __s0_692, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_692; __rev1_692 = __builtin_shufflevector(__s1_692, __s1_692, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_692; \ - __ret_692 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_692), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_692, __p2_692)))); \ - __ret_692 = __builtin_shufflevector(__ret_692, __ret_692, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_692; \ +#define vqshrn_high_n_s16(__p0_784, __p1_784, __p2_784) __extension__ ({ \ + int8x8_t __s0_784 = __p0_784; \ + int16x8_t __s1_784 = __p1_784; \ + int8x8_t __rev0_784; __rev0_784 = __builtin_shufflevector(__s0_784, __s0_784, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_784; __rev1_784 = __builtin_shufflevector(__s1_784, __s1_784, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_784; \ + __ret_784 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_784), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_784, __p2_784)))); \ + __ret_784 = __builtin_shufflevector(__ret_784, __ret_784, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_784; \ }) #endif @@ -58107,65 +59777,65 @@ __ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) { __ret; \ }) #ifdef __LITTLE_ENDIAN__ -#define vqshrun_high_n_s32(__p0_693, __p1_693, __p2_693) __extension__ ({ \ - int16x4_t __s0_693 = __p0_693; \ - int32x4_t __s1_693 = __p1_693; \ - int16x8_t __ret_693; \ - __ret_693 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_693), (int16x4_t)(vqshrun_n_s32(__s1_693, __p2_693)))); \ - __ret_693; \ +#define vqshrun_high_n_s32(__p0_785, __p1_785, __p2_785) __extension__ ({ \ + int16x4_t __s0_785 = __p0_785; \ + int32x4_t __s1_785 = __p1_785; \ + int16x8_t __ret_785; \ + __ret_785 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_785), (int16x4_t)(vqshrun_n_s32(__s1_785, __p2_785)))); \ + __ret_785; \ }) #else -#define vqshrun_high_n_s32(__p0_694, __p1_694, __p2_694) __extension__ ({ \ - int16x4_t __s0_694 = __p0_694; \ - int32x4_t __s1_694 = __p1_694; \ - int16x4_t __rev0_694; __rev0_694 = __builtin_shufflevector(__s0_694, __s0_694, 3, 2, 1, 0); \ - int32x4_t __rev1_694; __rev1_694 = __builtin_shufflevector(__s1_694, __s1_694, 3, 2, 1, 0); \ - int16x8_t __ret_694; \ - __ret_694 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_694), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_694, __p2_694)))); \ - __ret_694 = __builtin_shufflevector(__ret_694, __ret_694, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_694; \ +#define vqshrun_high_n_s32(__p0_786, __p1_786, __p2_786) __extension__ ({ \ + int16x4_t __s0_786 = __p0_786; \ + int32x4_t __s1_786 = __p1_786; \ + int16x4_t __rev0_786; __rev0_786 = __builtin_shufflevector(__s0_786, __s0_786, 3, 2, 1, 0); \ + int32x4_t __rev1_786; __rev1_786 = __builtin_shufflevector(__s1_786, __s1_786, 3, 2, 1, 0); \ + int16x8_t __ret_786; \ + __ret_786 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_786), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_786, __p2_786)))); \ + __ret_786 = __builtin_shufflevector(__ret_786, __ret_786, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_786; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqshrun_high_n_s64(__p0_695, __p1_695, __p2_695) __extension__ ({ \ - int32x2_t __s0_695 = __p0_695; \ - int64x2_t __s1_695 = __p1_695; \ - int32x4_t __ret_695; \ - __ret_695 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_695), (int32x2_t)(vqshrun_n_s64(__s1_695, __p2_695)))); \ - __ret_695; \ +#define vqshrun_high_n_s64(__p0_787, __p1_787, __p2_787) __extension__ ({ \ + int32x2_t __s0_787 = __p0_787; \ + int64x2_t __s1_787 = __p1_787; \ + int32x4_t __ret_787; \ + __ret_787 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_787), (int32x2_t)(vqshrun_n_s64(__s1_787, __p2_787)))); \ + __ret_787; \ }) #else -#define vqshrun_high_n_s64(__p0_696, __p1_696, __p2_696) __extension__ ({ \ - int32x2_t __s0_696 = __p0_696; \ - int64x2_t __s1_696 = __p1_696; \ - int32x2_t __rev0_696; __rev0_696 = __builtin_shufflevector(__s0_696, __s0_696, 1, 0); \ - int64x2_t __rev1_696; __rev1_696 = __builtin_shufflevector(__s1_696, __s1_696, 1, 0); \ - int32x4_t __ret_696; \ - __ret_696 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_696), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_696, __p2_696)))); \ - __ret_696 = __builtin_shufflevector(__ret_696, __ret_696, 3, 2, 1, 0); \ - __ret_696; \ +#define vqshrun_high_n_s64(__p0_788, __p1_788, __p2_788) __extension__ ({ \ + int32x2_t __s0_788 = __p0_788; \ + int64x2_t __s1_788 = __p1_788; \ + int32x2_t __rev0_788; __rev0_788 = __builtin_shufflevector(__s0_788, __s0_788, 1, 0); \ + int64x2_t __rev1_788; __rev1_788 = __builtin_shufflevector(__s1_788, __s1_788, 1, 0); \ + int32x4_t __ret_788; \ + __ret_788 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_788), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_788, __p2_788)))); \ + __ret_788 = __builtin_shufflevector(__ret_788, __ret_788, 3, 2, 1, 0); \ + __ret_788; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqshrun_high_n_s16(__p0_697, __p1_697, __p2_697) __extension__ ({ \ - int8x8_t __s0_697 = __p0_697; \ - int16x8_t __s1_697 = __p1_697; \ - int8x16_t __ret_697; \ - __ret_697 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_697), (int8x8_t)(vqshrun_n_s16(__s1_697, __p2_697)))); \ - __ret_697; \ +#define vqshrun_high_n_s16(__p0_789, __p1_789, __p2_789) __extension__ ({ \ + int8x8_t __s0_789 = __p0_789; \ + int16x8_t __s1_789 = __p1_789; \ + int8x16_t __ret_789; \ + __ret_789 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_789), (int8x8_t)(vqshrun_n_s16(__s1_789, __p2_789)))); \ + __ret_789; \ }) #else -#define vqshrun_high_n_s16(__p0_698, __p1_698, __p2_698) __extension__ ({ \ - int8x8_t __s0_698 = __p0_698; \ - int16x8_t __s1_698 = __p1_698; \ - int8x8_t __rev0_698; __rev0_698 = __builtin_shufflevector(__s0_698, __s0_698, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_698; __rev1_698 = __builtin_shufflevector(__s1_698, __s1_698, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_698; \ - __ret_698 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_698), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_698, __p2_698)))); \ - __ret_698 = __builtin_shufflevector(__ret_698, __ret_698, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_698; \ +#define vqshrun_high_n_s16(__p0_790, __p1_790, __p2_790) __extension__ ({ \ + int8x8_t __s0_790 = __p0_790; \ + int16x8_t __s1_790 = __p1_790; \ + int8x8_t __rev0_790; __rev0_790 = __builtin_shufflevector(__s0_790, __s0_790, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_790; __rev1_790 = __builtin_shufflevector(__s1_790, __s1_790, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_790; \ + __ret_790 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_790), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_790, __p2_790)))); \ + __ret_790 = __builtin_shufflevector(__ret_790, __ret_790, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_790; \ }) #endif @@ -59452,7 +61122,7 @@ __ai float32_t vrecpxs_f32(float32_t __p0) { __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0); return __ret; } -__ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) { +__ai uint64_t vrshld_u64(uint64_t __p0, int64_t __p1) { uint64_t __ret; __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1); return __ret; @@ -59475,128 +61145,128 @@ __ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) { __ret; \ }) #ifdef __LITTLE_ENDIAN__ -#define vrshrn_high_n_u32(__p0_699, __p1_699, __p2_699) __extension__ ({ \ - uint16x4_t __s0_699 = __p0_699; \ - uint32x4_t __s1_699 = __p1_699; \ - uint16x8_t __ret_699; \ - __ret_699 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_699), (uint16x4_t)(vrshrn_n_u32(__s1_699, __p2_699)))); \ - __ret_699; \ +#define vrshrn_high_n_u32(__p0_791, __p1_791, __p2_791) __extension__ ({ \ + uint16x4_t __s0_791 = __p0_791; \ + uint32x4_t __s1_791 = __p1_791; \ + uint16x8_t __ret_791; \ + __ret_791 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_791), (uint16x4_t)(vrshrn_n_u32(__s1_791, __p2_791)))); \ + __ret_791; \ }) #else -#define vrshrn_high_n_u32(__p0_700, __p1_700, __p2_700) __extension__ ({ \ - uint16x4_t __s0_700 = __p0_700; \ - uint32x4_t __s1_700 = __p1_700; \ - uint16x4_t __rev0_700; __rev0_700 = __builtin_shufflevector(__s0_700, __s0_700, 3, 2, 1, 0); \ - uint32x4_t __rev1_700; __rev1_700 = __builtin_shufflevector(__s1_700, __s1_700, 3, 2, 1, 0); \ - uint16x8_t __ret_700; \ - __ret_700 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_700), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_700, __p2_700)))); \ - __ret_700 = __builtin_shufflevector(__ret_700, __ret_700, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_700; \ +#define vrshrn_high_n_u32(__p0_792, __p1_792, __p2_792) __extension__ ({ \ + uint16x4_t __s0_792 = __p0_792; \ + uint32x4_t __s1_792 = __p1_792; \ + uint16x4_t __rev0_792; __rev0_792 = __builtin_shufflevector(__s0_792, __s0_792, 3, 2, 1, 0); \ + uint32x4_t __rev1_792; __rev1_792 = __builtin_shufflevector(__s1_792, __s1_792, 3, 2, 1, 0); \ + uint16x8_t __ret_792; \ + __ret_792 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_792), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_792, __p2_792)))); \ + __ret_792 = __builtin_shufflevector(__ret_792, __ret_792, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_792; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vrshrn_high_n_u64(__p0_701, __p1_701, __p2_701) __extension__ ({ \ - uint32x2_t __s0_701 = __p0_701; \ - uint64x2_t __s1_701 = __p1_701; \ - uint32x4_t __ret_701; \ - __ret_701 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_701), (uint32x2_t)(vrshrn_n_u64(__s1_701, __p2_701)))); \ - __ret_701; \ +#define vrshrn_high_n_u64(__p0_793, __p1_793, __p2_793) __extension__ ({ \ + uint32x2_t __s0_793 = __p0_793; \ + uint64x2_t __s1_793 = __p1_793; \ + uint32x4_t __ret_793; \ + __ret_793 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_793), (uint32x2_t)(vrshrn_n_u64(__s1_793, __p2_793)))); \ + __ret_793; \ }) #else -#define vrshrn_high_n_u64(__p0_702, __p1_702, __p2_702) __extension__ ({ \ - uint32x2_t __s0_702 = __p0_702; \ - uint64x2_t __s1_702 = __p1_702; \ - uint32x2_t __rev0_702; __rev0_702 = __builtin_shufflevector(__s0_702, __s0_702, 1, 0); \ - uint64x2_t __rev1_702; __rev1_702 = __builtin_shufflevector(__s1_702, __s1_702, 1, 0); \ - uint32x4_t __ret_702; \ - __ret_702 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_702), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_702, __p2_702)))); \ - __ret_702 = __builtin_shufflevector(__ret_702, __ret_702, 3, 2, 1, 0); \ - __ret_702; \ +#define vrshrn_high_n_u64(__p0_794, __p1_794, __p2_794) __extension__ ({ \ + uint32x2_t __s0_794 = __p0_794; \ + uint64x2_t __s1_794 = __p1_794; \ + uint32x2_t __rev0_794; __rev0_794 = __builtin_shufflevector(__s0_794, __s0_794, 1, 0); \ + uint64x2_t __rev1_794; __rev1_794 = __builtin_shufflevector(__s1_794, __s1_794, 1, 0); \ + uint32x4_t __ret_794; \ + __ret_794 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_794), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_794, __p2_794)))); \ + __ret_794 = __builtin_shufflevector(__ret_794, __ret_794, 3, 2, 1, 0); \ + __ret_794; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vrshrn_high_n_u16(__p0_703, __p1_703, __p2_703) __extension__ ({ \ - uint8x8_t __s0_703 = __p0_703; \ - uint16x8_t __s1_703 = __p1_703; \ - uint8x16_t __ret_703; \ - __ret_703 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_703), (uint8x8_t)(vrshrn_n_u16(__s1_703, __p2_703)))); \ - __ret_703; \ +#define vrshrn_high_n_u16(__p0_795, __p1_795, __p2_795) __extension__ ({ \ + uint8x8_t __s0_795 = __p0_795; \ + uint16x8_t __s1_795 = __p1_795; \ + uint8x16_t __ret_795; \ + __ret_795 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_795), (uint8x8_t)(vrshrn_n_u16(__s1_795, __p2_795)))); \ + __ret_795; \ }) #else -#define vrshrn_high_n_u16(__p0_704, __p1_704, __p2_704) __extension__ ({ \ - uint8x8_t __s0_704 = __p0_704; \ - uint16x8_t __s1_704 = __p1_704; \ - uint8x8_t __rev0_704; __rev0_704 = __builtin_shufflevector(__s0_704, __s0_704, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_704; __rev1_704 = __builtin_shufflevector(__s1_704, __s1_704, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __ret_704; \ - __ret_704 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_704), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_704, __p2_704)))); \ - __ret_704 = __builtin_shufflevector(__ret_704, __ret_704, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_704; \ +#define vrshrn_high_n_u16(__p0_796, __p1_796, __p2_796) __extension__ ({ \ + uint8x8_t __s0_796 = __p0_796; \ + uint16x8_t __s1_796 = __p1_796; \ + uint8x8_t __rev0_796; __rev0_796 = __builtin_shufflevector(__s0_796, __s0_796, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_796; __rev1_796 = __builtin_shufflevector(__s1_796, __s1_796, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __ret_796; \ + __ret_796 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_796), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_796, __p2_796)))); \ + __ret_796 = __builtin_shufflevector(__ret_796, __ret_796, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_796; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vrshrn_high_n_s32(__p0_705, __p1_705, __p2_705) __extension__ ({ \ - int16x4_t __s0_705 = __p0_705; \ - int32x4_t __s1_705 = __p1_705; \ - int16x8_t __ret_705; \ - __ret_705 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_705), (int16x4_t)(vrshrn_n_s32(__s1_705, __p2_705)))); \ - __ret_705; \ +#define vrshrn_high_n_s32(__p0_797, __p1_797, __p2_797) __extension__ ({ \ + int16x4_t __s0_797 = __p0_797; \ + int32x4_t __s1_797 = __p1_797; \ + int16x8_t __ret_797; \ + __ret_797 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_797), (int16x4_t)(vrshrn_n_s32(__s1_797, __p2_797)))); \ + __ret_797; \ }) #else -#define vrshrn_high_n_s32(__p0_706, __p1_706, __p2_706) __extension__ ({ \ - int16x4_t __s0_706 = __p0_706; \ - int32x4_t __s1_706 = __p1_706; \ - int16x4_t __rev0_706; __rev0_706 = __builtin_shufflevector(__s0_706, __s0_706, 3, 2, 1, 0); \ - int32x4_t __rev1_706; __rev1_706 = __builtin_shufflevector(__s1_706, __s1_706, 3, 2, 1, 0); \ - int16x8_t __ret_706; \ - __ret_706 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_706), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_706, __p2_706)))); \ - __ret_706 = __builtin_shufflevector(__ret_706, __ret_706, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_706; \ +#define vrshrn_high_n_s32(__p0_798, __p1_798, __p2_798) __extension__ ({ \ + int16x4_t __s0_798 = __p0_798; \ + int32x4_t __s1_798 = __p1_798; \ + int16x4_t __rev0_798; __rev0_798 = __builtin_shufflevector(__s0_798, __s0_798, 3, 2, 1, 0); \ + int32x4_t __rev1_798; __rev1_798 = __builtin_shufflevector(__s1_798, __s1_798, 3, 2, 1, 0); \ + int16x8_t __ret_798; \ + __ret_798 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_798), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_798, __p2_798)))); \ + __ret_798 = __builtin_shufflevector(__ret_798, __ret_798, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_798; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vrshrn_high_n_s64(__p0_707, __p1_707, __p2_707) __extension__ ({ \ - int32x2_t __s0_707 = __p0_707; \ - int64x2_t __s1_707 = __p1_707; \ - int32x4_t __ret_707; \ - __ret_707 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_707), (int32x2_t)(vrshrn_n_s64(__s1_707, __p2_707)))); \ - __ret_707; \ +#define vrshrn_high_n_s64(__p0_799, __p1_799, __p2_799) __extension__ ({ \ + int32x2_t __s0_799 = __p0_799; \ + int64x2_t __s1_799 = __p1_799; \ + int32x4_t __ret_799; \ + __ret_799 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_799), (int32x2_t)(vrshrn_n_s64(__s1_799, __p2_799)))); \ + __ret_799; \ }) #else -#define vrshrn_high_n_s64(__p0_708, __p1_708, __p2_708) __extension__ ({ \ - int32x2_t __s0_708 = __p0_708; \ - int64x2_t __s1_708 = __p1_708; \ - int32x2_t __rev0_708; __rev0_708 = __builtin_shufflevector(__s0_708, __s0_708, 1, 0); \ - int64x2_t __rev1_708; __rev1_708 = __builtin_shufflevector(__s1_708, __s1_708, 1, 0); \ - int32x4_t __ret_708; \ - __ret_708 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_708), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_708, __p2_708)))); \ - __ret_708 = __builtin_shufflevector(__ret_708, __ret_708, 3, 2, 1, 0); \ - __ret_708; \ +#define vrshrn_high_n_s64(__p0_800, __p1_800, __p2_800) __extension__ ({ \ + int32x2_t __s0_800 = __p0_800; \ + int64x2_t __s1_800 = __p1_800; \ + int32x2_t __rev0_800; __rev0_800 = __builtin_shufflevector(__s0_800, __s0_800, 1, 0); \ + int64x2_t __rev1_800; __rev1_800 = __builtin_shufflevector(__s1_800, __s1_800, 1, 0); \ + int32x4_t __ret_800; \ + __ret_800 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_800), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_800, __p2_800)))); \ + __ret_800 = __builtin_shufflevector(__ret_800, __ret_800, 3, 2, 1, 0); \ + __ret_800; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vrshrn_high_n_s16(__p0_709, __p1_709, __p2_709) __extension__ ({ \ - int8x8_t __s0_709 = __p0_709; \ - int16x8_t __s1_709 = __p1_709; \ - int8x16_t __ret_709; \ - __ret_709 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_709), (int8x8_t)(vrshrn_n_s16(__s1_709, __p2_709)))); \ - __ret_709; \ +#define vrshrn_high_n_s16(__p0_801, __p1_801, __p2_801) __extension__ ({ \ + int8x8_t __s0_801 = __p0_801; \ + int16x8_t __s1_801 = __p1_801; \ + int8x16_t __ret_801; \ + __ret_801 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_801), (int8x8_t)(vrshrn_n_s16(__s1_801, __p2_801)))); \ + __ret_801; \ }) #else -#define vrshrn_high_n_s16(__p0_710, __p1_710, __p2_710) __extension__ ({ \ - int8x8_t __s0_710 = __p0_710; \ - int16x8_t __s1_710 = __p1_710; \ - int8x8_t __rev0_710; __rev0_710 = __builtin_shufflevector(__s0_710, __s0_710, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_710; __rev1_710 = __builtin_shufflevector(__s1_710, __s1_710, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_710; \ - __ret_710 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_710), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_710, __p2_710)))); \ - __ret_710 = __builtin_shufflevector(__ret_710, __ret_710, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_710; \ +#define vrshrn_high_n_s16(__p0_802, __p1_802, __p2_802) __extension__ ({ \ + int8x8_t __s0_802 = __p0_802; \ + int16x8_t __s1_802 = __p1_802; \ + int8x8_t __rev0_802; __rev0_802 = __builtin_shufflevector(__s0_802, __s0_802, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_802; __rev1_802 = __builtin_shufflevector(__s1_802, __s1_802, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_802; \ + __ret_802 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_802), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_802, __p2_802)))); \ + __ret_802 = __builtin_shufflevector(__ret_802, __ret_802, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_802; \ }) #endif @@ -59853,7 +61523,7 @@ __ai int8x16_t vrsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) { __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (float64x1_t)__s1, __p2); \ __ret; \ }) -__ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) { +__ai uint64_t vshld_u64(uint64_t __p0, int64_t __p1) { uint64_t __ret; __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1); return __ret; @@ -59876,110 +61546,110 @@ __ai int64_t vshld_s64(int64_t __p0, int64_t __p1) { __ret; \ }) #ifdef __LITTLE_ENDIAN__ -#define vshll_high_n_u8(__p0_711, __p1_711) __extension__ ({ \ - uint8x16_t __s0_711 = __p0_711; \ - uint16x8_t __ret_711; \ - __ret_711 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_711), __p1_711)); \ - __ret_711; \ +#define vshll_high_n_u8(__p0_803, __p1_803) __extension__ ({ \ + uint8x16_t __s0_803 = __p0_803; \ + uint16x8_t __ret_803; \ + __ret_803 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_803), __p1_803)); \ + __ret_803; \ }) #else -#define vshll_high_n_u8(__p0_712, __p1_712) __extension__ ({ \ - uint8x16_t __s0_712 = __p0_712; \ - uint8x16_t __rev0_712; __rev0_712 = __builtin_shufflevector(__s0_712, __s0_712, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __ret_712; \ - __ret_712 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_712), __p1_712)); \ - __ret_712 = __builtin_shufflevector(__ret_712, __ret_712, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_712; \ +#define vshll_high_n_u8(__p0_804, __p1_804) __extension__ ({ \ + uint8x16_t __s0_804 = __p0_804; \ + uint8x16_t __rev0_804; __rev0_804 = __builtin_shufflevector(__s0_804, __s0_804, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __ret_804; \ + __ret_804 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_804), __p1_804)); \ + __ret_804 = __builtin_shufflevector(__ret_804, __ret_804, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_804; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshll_high_n_u32(__p0_713, __p1_713) __extension__ ({ \ - uint32x4_t __s0_713 = __p0_713; \ - uint64x2_t __ret_713; \ - __ret_713 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_713), __p1_713)); \ - __ret_713; \ +#define vshll_high_n_u32(__p0_805, __p1_805) __extension__ ({ \ + uint32x4_t __s0_805 = __p0_805; \ + uint64x2_t __ret_805; \ + __ret_805 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_805), __p1_805)); \ + __ret_805; \ }) #else -#define vshll_high_n_u32(__p0_714, __p1_714) __extension__ ({ \ - uint32x4_t __s0_714 = __p0_714; \ - uint32x4_t __rev0_714; __rev0_714 = __builtin_shufflevector(__s0_714, __s0_714, 3, 2, 1, 0); \ - uint64x2_t __ret_714; \ - __ret_714 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_714), __p1_714)); \ - __ret_714 = __builtin_shufflevector(__ret_714, __ret_714, 1, 0); \ - __ret_714; \ +#define vshll_high_n_u32(__p0_806, __p1_806) __extension__ ({ \ + uint32x4_t __s0_806 = __p0_806; \ + uint32x4_t __rev0_806; __rev0_806 = __builtin_shufflevector(__s0_806, __s0_806, 3, 2, 1, 0); \ + uint64x2_t __ret_806; \ + __ret_806 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_806), __p1_806)); \ + __ret_806 = __builtin_shufflevector(__ret_806, __ret_806, 1, 0); \ + __ret_806; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshll_high_n_u16(__p0_715, __p1_715) __extension__ ({ \ - uint16x8_t __s0_715 = __p0_715; \ - uint32x4_t __ret_715; \ - __ret_715 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_715), __p1_715)); \ - __ret_715; \ +#define vshll_high_n_u16(__p0_807, __p1_807) __extension__ ({ \ + uint16x8_t __s0_807 = __p0_807; \ + uint32x4_t __ret_807; \ + __ret_807 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_807), __p1_807)); \ + __ret_807; \ }) #else -#define vshll_high_n_u16(__p0_716, __p1_716) __extension__ ({ \ - uint16x8_t __s0_716 = __p0_716; \ - uint16x8_t __rev0_716; __rev0_716 = __builtin_shufflevector(__s0_716, __s0_716, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint32x4_t __ret_716; \ - __ret_716 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_716), __p1_716)); \ - __ret_716 = __builtin_shufflevector(__ret_716, __ret_716, 3, 2, 1, 0); \ - __ret_716; \ +#define vshll_high_n_u16(__p0_808, __p1_808) __extension__ ({ \ + uint16x8_t __s0_808 = __p0_808; \ + uint16x8_t __rev0_808; __rev0_808 = __builtin_shufflevector(__s0_808, __s0_808, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint32x4_t __ret_808; \ + __ret_808 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_808), __p1_808)); \ + __ret_808 = __builtin_shufflevector(__ret_808, __ret_808, 3, 2, 1, 0); \ + __ret_808; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshll_high_n_s8(__p0_717, __p1_717) __extension__ ({ \ - int8x16_t __s0_717 = __p0_717; \ - int16x8_t __ret_717; \ - __ret_717 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_717), __p1_717)); \ - __ret_717; \ +#define vshll_high_n_s8(__p0_809, __p1_809) __extension__ ({ \ + int8x16_t __s0_809 = __p0_809; \ + int16x8_t __ret_809; \ + __ret_809 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_809), __p1_809)); \ + __ret_809; \ }) #else -#define vshll_high_n_s8(__p0_718, __p1_718) __extension__ ({ \ - int8x16_t __s0_718 = __p0_718; \ - int8x16_t __rev0_718; __rev0_718 = __builtin_shufflevector(__s0_718, __s0_718, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __ret_718; \ - __ret_718 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_718), __p1_718)); \ - __ret_718 = __builtin_shufflevector(__ret_718, __ret_718, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_718; \ +#define vshll_high_n_s8(__p0_810, __p1_810) __extension__ ({ \ + int8x16_t __s0_810 = __p0_810; \ + int8x16_t __rev0_810; __rev0_810 = __builtin_shufflevector(__s0_810, __s0_810, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __ret_810; \ + __ret_810 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_810), __p1_810)); \ + __ret_810 = __builtin_shufflevector(__ret_810, __ret_810, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_810; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshll_high_n_s32(__p0_719, __p1_719) __extension__ ({ \ - int32x4_t __s0_719 = __p0_719; \ - int64x2_t __ret_719; \ - __ret_719 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_719), __p1_719)); \ - __ret_719; \ +#define vshll_high_n_s32(__p0_811, __p1_811) __extension__ ({ \ + int32x4_t __s0_811 = __p0_811; \ + int64x2_t __ret_811; \ + __ret_811 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_811), __p1_811)); \ + __ret_811; \ }) #else -#define vshll_high_n_s32(__p0_720, __p1_720) __extension__ ({ \ - int32x4_t __s0_720 = __p0_720; \ - int32x4_t __rev0_720; __rev0_720 = __builtin_shufflevector(__s0_720, __s0_720, 3, 2, 1, 0); \ - int64x2_t __ret_720; \ - __ret_720 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_720), __p1_720)); \ - __ret_720 = __builtin_shufflevector(__ret_720, __ret_720, 1, 0); \ - __ret_720; \ +#define vshll_high_n_s32(__p0_812, __p1_812) __extension__ ({ \ + int32x4_t __s0_812 = __p0_812; \ + int32x4_t __rev0_812; __rev0_812 = __builtin_shufflevector(__s0_812, __s0_812, 3, 2, 1, 0); \ + int64x2_t __ret_812; \ + __ret_812 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_812), __p1_812)); \ + __ret_812 = __builtin_shufflevector(__ret_812, __ret_812, 1, 0); \ + __ret_812; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshll_high_n_s16(__p0_721, __p1_721) __extension__ ({ \ - int16x8_t __s0_721 = __p0_721; \ - int32x4_t __ret_721; \ - __ret_721 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_721), __p1_721)); \ - __ret_721; \ +#define vshll_high_n_s16(__p0_813, __p1_813) __extension__ ({ \ + int16x8_t __s0_813 = __p0_813; \ + int32x4_t __ret_813; \ + __ret_813 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_813), __p1_813)); \ + __ret_813; \ }) #else -#define vshll_high_n_s16(__p0_722, __p1_722) __extension__ ({ \ - int16x8_t __s0_722 = __p0_722; \ - int16x8_t __rev0_722; __rev0_722 = __builtin_shufflevector(__s0_722, __s0_722, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_722; \ - __ret_722 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_722), __p1_722)); \ - __ret_722 = __builtin_shufflevector(__ret_722, __ret_722, 3, 2, 1, 0); \ - __ret_722; \ +#define vshll_high_n_s16(__p0_814, __p1_814) __extension__ ({ \ + int16x8_t __s0_814 = __p0_814; \ + int16x8_t __rev0_814; __rev0_814 = __builtin_shufflevector(__s0_814, __s0_814, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_814; \ + __ret_814 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_814), __p1_814)); \ + __ret_814 = __builtin_shufflevector(__ret_814, __ret_814, 3, 2, 1, 0); \ + __ret_814; \ }) #endif @@ -59996,128 +61666,128 @@ __ai int64_t vshld_s64(int64_t __p0, int64_t __p1) { __ret; \ }) #ifdef __LITTLE_ENDIAN__ -#define vshrn_high_n_u32(__p0_723, __p1_723, __p2_723) __extension__ ({ \ - uint16x4_t __s0_723 = __p0_723; \ - uint32x4_t __s1_723 = __p1_723; \ - uint16x8_t __ret_723; \ - __ret_723 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_723), (uint16x4_t)(vshrn_n_u32(__s1_723, __p2_723)))); \ - __ret_723; \ +#define vshrn_high_n_u32(__p0_815, __p1_815, __p2_815) __extension__ ({ \ + uint16x4_t __s0_815 = __p0_815; \ + uint32x4_t __s1_815 = __p1_815; \ + uint16x8_t __ret_815; \ + __ret_815 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_815), (uint16x4_t)(vshrn_n_u32(__s1_815, __p2_815)))); \ + __ret_815; \ }) #else -#define vshrn_high_n_u32(__p0_724, __p1_724, __p2_724) __extension__ ({ \ - uint16x4_t __s0_724 = __p0_724; \ - uint32x4_t __s1_724 = __p1_724; \ - uint16x4_t __rev0_724; __rev0_724 = __builtin_shufflevector(__s0_724, __s0_724, 3, 2, 1, 0); \ - uint32x4_t __rev1_724; __rev1_724 = __builtin_shufflevector(__s1_724, __s1_724, 3, 2, 1, 0); \ - uint16x8_t __ret_724; \ - __ret_724 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_724), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_724, __p2_724)))); \ - __ret_724 = __builtin_shufflevector(__ret_724, __ret_724, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_724; \ +#define vshrn_high_n_u32(__p0_816, __p1_816, __p2_816) __extension__ ({ \ + uint16x4_t __s0_816 = __p0_816; \ + uint32x4_t __s1_816 = __p1_816; \ + uint16x4_t __rev0_816; __rev0_816 = __builtin_shufflevector(__s0_816, __s0_816, 3, 2, 1, 0); \ + uint32x4_t __rev1_816; __rev1_816 = __builtin_shufflevector(__s1_816, __s1_816, 3, 2, 1, 0); \ + uint16x8_t __ret_816; \ + __ret_816 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_816), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_816, __p2_816)))); \ + __ret_816 = __builtin_shufflevector(__ret_816, __ret_816, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_816; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshrn_high_n_u64(__p0_725, __p1_725, __p2_725) __extension__ ({ \ - uint32x2_t __s0_725 = __p0_725; \ - uint64x2_t __s1_725 = __p1_725; \ - uint32x4_t __ret_725; \ - __ret_725 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_725), (uint32x2_t)(vshrn_n_u64(__s1_725, __p2_725)))); \ - __ret_725; \ +#define vshrn_high_n_u64(__p0_817, __p1_817, __p2_817) __extension__ ({ \ + uint32x2_t __s0_817 = __p0_817; \ + uint64x2_t __s1_817 = __p1_817; \ + uint32x4_t __ret_817; \ + __ret_817 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_817), (uint32x2_t)(vshrn_n_u64(__s1_817, __p2_817)))); \ + __ret_817; \ }) #else -#define vshrn_high_n_u64(__p0_726, __p1_726, __p2_726) __extension__ ({ \ - uint32x2_t __s0_726 = __p0_726; \ - uint64x2_t __s1_726 = __p1_726; \ - uint32x2_t __rev0_726; __rev0_726 = __builtin_shufflevector(__s0_726, __s0_726, 1, 0); \ - uint64x2_t __rev1_726; __rev1_726 = __builtin_shufflevector(__s1_726, __s1_726, 1, 0); \ - uint32x4_t __ret_726; \ - __ret_726 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_726), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_726, __p2_726)))); \ - __ret_726 = __builtin_shufflevector(__ret_726, __ret_726, 3, 2, 1, 0); \ - __ret_726; \ +#define vshrn_high_n_u64(__p0_818, __p1_818, __p2_818) __extension__ ({ \ + uint32x2_t __s0_818 = __p0_818; \ + uint64x2_t __s1_818 = __p1_818; \ + uint32x2_t __rev0_818; __rev0_818 = __builtin_shufflevector(__s0_818, __s0_818, 1, 0); \ + uint64x2_t __rev1_818; __rev1_818 = __builtin_shufflevector(__s1_818, __s1_818, 1, 0); \ + uint32x4_t __ret_818; \ + __ret_818 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_818), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_818, __p2_818)))); \ + __ret_818 = __builtin_shufflevector(__ret_818, __ret_818, 3, 2, 1, 0); \ + __ret_818; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshrn_high_n_u16(__p0_727, __p1_727, __p2_727) __extension__ ({ \ - uint8x8_t __s0_727 = __p0_727; \ - uint16x8_t __s1_727 = __p1_727; \ - uint8x16_t __ret_727; \ - __ret_727 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_727), (uint8x8_t)(vshrn_n_u16(__s1_727, __p2_727)))); \ - __ret_727; \ +#define vshrn_high_n_u16(__p0_819, __p1_819, __p2_819) __extension__ ({ \ + uint8x8_t __s0_819 = __p0_819; \ + uint16x8_t __s1_819 = __p1_819; \ + uint8x16_t __ret_819; \ + __ret_819 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_819), (uint8x8_t)(vshrn_n_u16(__s1_819, __p2_819)))); \ + __ret_819; \ }) #else -#define vshrn_high_n_u16(__p0_728, __p1_728, __p2_728) __extension__ ({ \ - uint8x8_t __s0_728 = __p0_728; \ - uint16x8_t __s1_728 = __p1_728; \ - uint8x8_t __rev0_728; __rev0_728 = __builtin_shufflevector(__s0_728, __s0_728, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint16x8_t __rev1_728; __rev1_728 = __builtin_shufflevector(__s1_728, __s1_728, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __ret_728; \ - __ret_728 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_728), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_728, __p2_728)))); \ - __ret_728 = __builtin_shufflevector(__ret_728, __ret_728, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_728; \ +#define vshrn_high_n_u16(__p0_820, __p1_820, __p2_820) __extension__ ({ \ + uint8x8_t __s0_820 = __p0_820; \ + uint16x8_t __s1_820 = __p1_820; \ + uint8x8_t __rev0_820; __rev0_820 = __builtin_shufflevector(__s0_820, __s0_820, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint16x8_t __rev1_820; __rev1_820 = __builtin_shufflevector(__s1_820, __s1_820, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __ret_820; \ + __ret_820 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_820), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_820, __p2_820)))); \ + __ret_820 = __builtin_shufflevector(__ret_820, __ret_820, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_820; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshrn_high_n_s32(__p0_729, __p1_729, __p2_729) __extension__ ({ \ - int16x4_t __s0_729 = __p0_729; \ - int32x4_t __s1_729 = __p1_729; \ - int16x8_t __ret_729; \ - __ret_729 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_729), (int16x4_t)(vshrn_n_s32(__s1_729, __p2_729)))); \ - __ret_729; \ +#define vshrn_high_n_s32(__p0_821, __p1_821, __p2_821) __extension__ ({ \ + int16x4_t __s0_821 = __p0_821; \ + int32x4_t __s1_821 = __p1_821; \ + int16x8_t __ret_821; \ + __ret_821 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_821), (int16x4_t)(vshrn_n_s32(__s1_821, __p2_821)))); \ + __ret_821; \ }) #else -#define vshrn_high_n_s32(__p0_730, __p1_730, __p2_730) __extension__ ({ \ - int16x4_t __s0_730 = __p0_730; \ - int32x4_t __s1_730 = __p1_730; \ - int16x4_t __rev0_730; __rev0_730 = __builtin_shufflevector(__s0_730, __s0_730, 3, 2, 1, 0); \ - int32x4_t __rev1_730; __rev1_730 = __builtin_shufflevector(__s1_730, __s1_730, 3, 2, 1, 0); \ - int16x8_t __ret_730; \ - __ret_730 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_730), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_730, __p2_730)))); \ - __ret_730 = __builtin_shufflevector(__ret_730, __ret_730, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_730; \ +#define vshrn_high_n_s32(__p0_822, __p1_822, __p2_822) __extension__ ({ \ + int16x4_t __s0_822 = __p0_822; \ + int32x4_t __s1_822 = __p1_822; \ + int16x4_t __rev0_822; __rev0_822 = __builtin_shufflevector(__s0_822, __s0_822, 3, 2, 1, 0); \ + int32x4_t __rev1_822; __rev1_822 = __builtin_shufflevector(__s1_822, __s1_822, 3, 2, 1, 0); \ + int16x8_t __ret_822; \ + __ret_822 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_822), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_822, __p2_822)))); \ + __ret_822 = __builtin_shufflevector(__ret_822, __ret_822, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_822; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshrn_high_n_s64(__p0_731, __p1_731, __p2_731) __extension__ ({ \ - int32x2_t __s0_731 = __p0_731; \ - int64x2_t __s1_731 = __p1_731; \ - int32x4_t __ret_731; \ - __ret_731 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_731), (int32x2_t)(vshrn_n_s64(__s1_731, __p2_731)))); \ - __ret_731; \ +#define vshrn_high_n_s64(__p0_823, __p1_823, __p2_823) __extension__ ({ \ + int32x2_t __s0_823 = __p0_823; \ + int64x2_t __s1_823 = __p1_823; \ + int32x4_t __ret_823; \ + __ret_823 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_823), (int32x2_t)(vshrn_n_s64(__s1_823, __p2_823)))); \ + __ret_823; \ }) #else -#define vshrn_high_n_s64(__p0_732, __p1_732, __p2_732) __extension__ ({ \ - int32x2_t __s0_732 = __p0_732; \ - int64x2_t __s1_732 = __p1_732; \ - int32x2_t __rev0_732; __rev0_732 = __builtin_shufflevector(__s0_732, __s0_732, 1, 0); \ - int64x2_t __rev1_732; __rev1_732 = __builtin_shufflevector(__s1_732, __s1_732, 1, 0); \ - int32x4_t __ret_732; \ - __ret_732 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_732), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_732, __p2_732)))); \ - __ret_732 = __builtin_shufflevector(__ret_732, __ret_732, 3, 2, 1, 0); \ - __ret_732; \ +#define vshrn_high_n_s64(__p0_824, __p1_824, __p2_824) __extension__ ({ \ + int32x2_t __s0_824 = __p0_824; \ + int64x2_t __s1_824 = __p1_824; \ + int32x2_t __rev0_824; __rev0_824 = __builtin_shufflevector(__s0_824, __s0_824, 1, 0); \ + int64x2_t __rev1_824; __rev1_824 = __builtin_shufflevector(__s1_824, __s1_824, 1, 0); \ + int32x4_t __ret_824; \ + __ret_824 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_824), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_824, __p2_824)))); \ + __ret_824 = __builtin_shufflevector(__ret_824, __ret_824, 3, 2, 1, 0); \ + __ret_824; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vshrn_high_n_s16(__p0_733, __p1_733, __p2_733) __extension__ ({ \ - int8x8_t __s0_733 = __p0_733; \ - int16x8_t __s1_733 = __p1_733; \ - int8x16_t __ret_733; \ - __ret_733 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_733), (int8x8_t)(vshrn_n_s16(__s1_733, __p2_733)))); \ - __ret_733; \ +#define vshrn_high_n_s16(__p0_825, __p1_825, __p2_825) __extension__ ({ \ + int8x8_t __s0_825 = __p0_825; \ + int16x8_t __s1_825 = __p1_825; \ + int8x16_t __ret_825; \ + __ret_825 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_825), (int8x8_t)(vshrn_n_s16(__s1_825, __p2_825)))); \ + __ret_825; \ }) #else -#define vshrn_high_n_s16(__p0_734, __p1_734, __p2_734) __extension__ ({ \ - int8x8_t __s0_734 = __p0_734; \ - int16x8_t __s1_734 = __p1_734; \ - int8x8_t __rev0_734; __rev0_734 = __builtin_shufflevector(__s0_734, __s0_734, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16x8_t __rev1_734; __rev1_734 = __builtin_shufflevector(__s1_734, __s1_734, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __ret_734; \ - __ret_734 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_734), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_734, __p2_734)))); \ - __ret_734 = __builtin_shufflevector(__ret_734, __ret_734, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_734; \ +#define vshrn_high_n_s16(__p0_826, __p1_826, __p2_826) __extension__ ({ \ + int8x8_t __s0_826 = __p0_826; \ + int16x8_t __s1_826 = __p1_826; \ + int8x8_t __rev0_826; __rev0_826 = __builtin_shufflevector(__s0_826, __s0_826, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16x8_t __rev1_826; __rev1_826 = __builtin_shufflevector(__s1_826, __s1_826, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __ret_826; \ + __ret_826 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_826), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_826, __p2_826)))); \ + __ret_826 = __builtin_shufflevector(__ret_826, __ret_826, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_826; \ }) #endif @@ -61553,54 +63223,54 @@ __ai int32x4_t vsubw_high_s16(int32x4_t __p0, int16x8_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vsudotq_laneq_s32(__p0_735, __p1_735, __p2_735, __p3_735) __extension__ ({ \ - int32x4_t __s0_735 = __p0_735; \ - int8x16_t __s1_735 = __p1_735; \ - uint8x16_t __s2_735 = __p2_735; \ - int32x4_t __ret_735; \ -uint8x16_t __reint_735 = __s2_735; \ - __ret_735 = vusdotq_s32(__s0_735, (uint8x16_t)(splatq_laneq_s32(*(int32x4_t *) &__reint_735, __p3_735)), __s1_735); \ - __ret_735; \ +#define vsudotq_laneq_s32(__p0_827, __p1_827, __p2_827, __p3_827) __extension__ ({ \ + int32x4_t __s0_827 = __p0_827; \ + int8x16_t __s1_827 = __p1_827; \ + uint8x16_t __s2_827 = __p2_827; \ + int32x4_t __ret_827; \ +uint8x16_t __reint_827 = __s2_827; \ + __ret_827 = vusdotq_s32(__s0_827, (uint8x16_t)(splatq_laneq_s32(*(int32x4_t *) &__reint_827, __p3_827)), __s1_827); \ + __ret_827; \ }) #else -#define vsudotq_laneq_s32(__p0_736, __p1_736, __p2_736, __p3_736) __extension__ ({ \ - int32x4_t __s0_736 = __p0_736; \ - int8x16_t __s1_736 = __p1_736; \ - uint8x16_t __s2_736 = __p2_736; \ - int32x4_t __rev0_736; __rev0_736 = __builtin_shufflevector(__s0_736, __s0_736, 3, 2, 1, 0); \ - int8x16_t __rev1_736; __rev1_736 = __builtin_shufflevector(__s1_736, __s1_736, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __rev2_736; __rev2_736 = __builtin_shufflevector(__s2_736, __s2_736, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_736; \ -uint8x16_t __reint_736 = __rev2_736; \ - __ret_736 = __noswap_vusdotq_s32(__rev0_736, (uint8x16_t)(__noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_736, __p3_736)), __rev1_736); \ - __ret_736 = __builtin_shufflevector(__ret_736, __ret_736, 3, 2, 1, 0); \ - __ret_736; \ +#define vsudotq_laneq_s32(__p0_828, __p1_828, __p2_828, __p3_828) __extension__ ({ \ + int32x4_t __s0_828 = __p0_828; \ + int8x16_t __s1_828 = __p1_828; \ + uint8x16_t __s2_828 = __p2_828; \ + int32x4_t __rev0_828; __rev0_828 = __builtin_shufflevector(__s0_828, __s0_828, 3, 2, 1, 0); \ + int8x16_t __rev1_828; __rev1_828 = __builtin_shufflevector(__s1_828, __s1_828, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __rev2_828; __rev2_828 = __builtin_shufflevector(__s2_828, __s2_828, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_828; \ +uint8x16_t __reint_828 = __rev2_828; \ + __ret_828 = __noswap_vusdotq_s32(__rev0_828, (uint8x16_t)(__noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_828, __p3_828)), __rev1_828); \ + __ret_828 = __builtin_shufflevector(__ret_828, __ret_828, 3, 2, 1, 0); \ + __ret_828; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vsudot_laneq_s32(__p0_737, __p1_737, __p2_737, __p3_737) __extension__ ({ \ - int32x2_t __s0_737 = __p0_737; \ - int8x8_t __s1_737 = __p1_737; \ - uint8x16_t __s2_737 = __p2_737; \ - int32x2_t __ret_737; \ -uint8x16_t __reint_737 = __s2_737; \ - __ret_737 = vusdot_s32(__s0_737, (uint8x8_t)(splat_laneq_s32(*(int32x4_t *) &__reint_737, __p3_737)), __s1_737); \ - __ret_737; \ +#define vsudot_laneq_s32(__p0_829, __p1_829, __p2_829, __p3_829) __extension__ ({ \ + int32x2_t __s0_829 = __p0_829; \ + int8x8_t __s1_829 = __p1_829; \ + uint8x16_t __s2_829 = __p2_829; \ + int32x2_t __ret_829; \ +uint8x16_t __reint_829 = __s2_829; \ + __ret_829 = vusdot_s32(__s0_829, (uint8x8_t)(splat_laneq_s32(*(int32x4_t *) &__reint_829, __p3_829)), __s1_829); \ + __ret_829; \ }) #else -#define vsudot_laneq_s32(__p0_738, __p1_738, __p2_738, __p3_738) __extension__ ({ \ - int32x2_t __s0_738 = __p0_738; \ - int8x8_t __s1_738 = __p1_738; \ - uint8x16_t __s2_738 = __p2_738; \ - int32x2_t __rev0_738; __rev0_738 = __builtin_shufflevector(__s0_738, __s0_738, 1, 0); \ - int8x8_t __rev1_738; __rev1_738 = __builtin_shufflevector(__s1_738, __s1_738, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x16_t __rev2_738; __rev2_738 = __builtin_shufflevector(__s2_738, __s2_738, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x2_t __ret_738; \ -uint8x16_t __reint_738 = __rev2_738; \ - __ret_738 = __noswap_vusdot_s32(__rev0_738, (uint8x8_t)(__noswap_splat_laneq_s32(*(int32x4_t *) &__reint_738, __p3_738)), __rev1_738); \ - __ret_738 = __builtin_shufflevector(__ret_738, __ret_738, 1, 0); \ - __ret_738; \ +#define vsudot_laneq_s32(__p0_830, __p1_830, __p2_830, __p3_830) __extension__ ({ \ + int32x2_t __s0_830 = __p0_830; \ + int8x8_t __s1_830 = __p1_830; \ + uint8x16_t __s2_830 = __p2_830; \ + int32x2_t __rev0_830; __rev0_830 = __builtin_shufflevector(__s0_830, __s0_830, 1, 0); \ + int8x8_t __rev1_830; __rev1_830 = __builtin_shufflevector(__s1_830, __s1_830, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x16_t __rev2_830; __rev2_830 = __builtin_shufflevector(__s2_830, __s2_830, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x2_t __ret_830; \ +uint8x16_t __reint_830 = __rev2_830; \ + __ret_830 = __noswap_vusdot_s32(__rev0_830, (uint8x8_t)(__noswap_splat_laneq_s32(*(int32x4_t *) &__reint_830, __p3_830)), __rev1_830); \ + __ret_830 = __builtin_shufflevector(__ret_830, __ret_830, 1, 0); \ + __ret_830; \ }) #endif @@ -62423,9 +64093,9 @@ __ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) { __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1); return __ret; } -__ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) { - int64_t __ret; - __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1); +__ai uint64_t vtstd_s64(int64_t __p0, int64_t __p1) { + uint64_t __ret; + __ret = (uint64_t) __builtin_neon_vtstd_s64(__p0, __p1); return __ret; } __ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) { @@ -62573,54 +64243,54 @@ __ai int16x4_t vuqadd_s16(int16x4_t __p0, uint16x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vusdotq_laneq_s32(__p0_739, __p1_739, __p2_739, __p3_739) __extension__ ({ \ - int32x4_t __s0_739 = __p0_739; \ - uint8x16_t __s1_739 = __p1_739; \ - int8x16_t __s2_739 = __p2_739; \ - int32x4_t __ret_739; \ -int8x16_t __reint_739 = __s2_739; \ - __ret_739 = vusdotq_s32(__s0_739, __s1_739, (int8x16_t)(splatq_laneq_s32(*(int32x4_t *) &__reint_739, __p3_739))); \ - __ret_739; \ +#define vusdotq_laneq_s32(__p0_831, __p1_831, __p2_831, __p3_831) __extension__ ({ \ + int32x4_t __s0_831 = __p0_831; \ + uint8x16_t __s1_831 = __p1_831; \ + int8x16_t __s2_831 = __p2_831; \ + int32x4_t __ret_831; \ +int8x16_t __reint_831 = __s2_831; \ + __ret_831 = vusdotq_s32(__s0_831, __s1_831, (int8x16_t)(splatq_laneq_s32(*(int32x4_t *) &__reint_831, __p3_831))); \ + __ret_831; \ }) #else -#define vusdotq_laneq_s32(__p0_740, __p1_740, __p2_740, __p3_740) __extension__ ({ \ - int32x4_t __s0_740 = __p0_740; \ - uint8x16_t __s1_740 = __p1_740; \ - int8x16_t __s2_740 = __p2_740; \ - int32x4_t __rev0_740; __rev0_740 = __builtin_shufflevector(__s0_740, __s0_740, 3, 2, 1, 0); \ - uint8x16_t __rev1_740; __rev1_740 = __builtin_shufflevector(__s1_740, __s1_740, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __rev2_740; __rev2_740 = __builtin_shufflevector(__s2_740, __s2_740, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_740; \ -int8x16_t __reint_740 = __rev2_740; \ - __ret_740 = __noswap_vusdotq_s32(__rev0_740, __rev1_740, (int8x16_t)(__noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_740, __p3_740))); \ - __ret_740 = __builtin_shufflevector(__ret_740, __ret_740, 3, 2, 1, 0); \ - __ret_740; \ +#define vusdotq_laneq_s32(__p0_832, __p1_832, __p2_832, __p3_832) __extension__ ({ \ + int32x4_t __s0_832 = __p0_832; \ + uint8x16_t __s1_832 = __p1_832; \ + int8x16_t __s2_832 = __p2_832; \ + int32x4_t __rev0_832; __rev0_832 = __builtin_shufflevector(__s0_832, __s0_832, 3, 2, 1, 0); \ + uint8x16_t __rev1_832; __rev1_832 = __builtin_shufflevector(__s1_832, __s1_832, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __rev2_832; __rev2_832 = __builtin_shufflevector(__s2_832, __s2_832, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_832; \ +int8x16_t __reint_832 = __rev2_832; \ + __ret_832 = __noswap_vusdotq_s32(__rev0_832, __rev1_832, (int8x16_t)(__noswap_splatq_laneq_s32(*(int32x4_t *) &__reint_832, __p3_832))); \ + __ret_832 = __builtin_shufflevector(__ret_832, __ret_832, 3, 2, 1, 0); \ + __ret_832; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vusdot_laneq_s32(__p0_741, __p1_741, __p2_741, __p3_741) __extension__ ({ \ - int32x2_t __s0_741 = __p0_741; \ - uint8x8_t __s1_741 = __p1_741; \ - int8x16_t __s2_741 = __p2_741; \ - int32x2_t __ret_741; \ -int8x16_t __reint_741 = __s2_741; \ - __ret_741 = vusdot_s32(__s0_741, __s1_741, (int8x8_t)(splat_laneq_s32(*(int32x4_t *) &__reint_741, __p3_741))); \ - __ret_741; \ +#define vusdot_laneq_s32(__p0_833, __p1_833, __p2_833, __p3_833) __extension__ ({ \ + int32x2_t __s0_833 = __p0_833; \ + uint8x8_t __s1_833 = __p1_833; \ + int8x16_t __s2_833 = __p2_833; \ + int32x2_t __ret_833; \ +int8x16_t __reint_833 = __s2_833; \ + __ret_833 = vusdot_s32(__s0_833, __s1_833, (int8x8_t)(splat_laneq_s32(*(int32x4_t *) &__reint_833, __p3_833))); \ + __ret_833; \ }) #else -#define vusdot_laneq_s32(__p0_742, __p1_742, __p2_742, __p3_742) __extension__ ({ \ - int32x2_t __s0_742 = __p0_742; \ - uint8x8_t __s1_742 = __p1_742; \ - int8x16_t __s2_742 = __p2_742; \ - int32x2_t __rev0_742; __rev0_742 = __builtin_shufflevector(__s0_742, __s0_742, 1, 0); \ - uint8x8_t __rev1_742; __rev1_742 = __builtin_shufflevector(__s1_742, __s1_742, 7, 6, 5, 4, 3, 2, 1, 0); \ - int8x16_t __rev2_742; __rev2_742 = __builtin_shufflevector(__s2_742, __s2_742, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x2_t __ret_742; \ -int8x16_t __reint_742 = __rev2_742; \ - __ret_742 = __noswap_vusdot_s32(__rev0_742, __rev1_742, (int8x8_t)(__noswap_splat_laneq_s32(*(int32x4_t *) &__reint_742, __p3_742))); \ - __ret_742 = __builtin_shufflevector(__ret_742, __ret_742, 1, 0); \ - __ret_742; \ +#define vusdot_laneq_s32(__p0_834, __p1_834, __p2_834, __p3_834) __extension__ ({ \ + int32x2_t __s0_834 = __p0_834; \ + uint8x8_t __s1_834 = __p1_834; \ + int8x16_t __s2_834 = __p2_834; \ + int32x2_t __rev0_834; __rev0_834 = __builtin_shufflevector(__s0_834, __s0_834, 1, 0); \ + uint8x8_t __rev1_834; __rev1_834 = __builtin_shufflevector(__s1_834, __s1_834, 7, 6, 5, 4, 3, 2, 1, 0); \ + int8x16_t __rev2_834; __rev2_834 = __builtin_shufflevector(__s2_834, __s2_834, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x2_t __ret_834; \ +int8x16_t __reint_834 = __rev2_834; \ + __ret_834 = __noswap_vusdot_s32(__rev0_834, __rev1_834, (int8x8_t)(__noswap_splat_laneq_s32(*(int32x4_t *) &__reint_834, __p3_834))); \ + __ret_834 = __builtin_shufflevector(__ret_834, __ret_834, 1, 0); \ + __ret_834; \ }) #endif @@ -64674,60 +66344,60 @@ __ai int32x4_t vaddw_s16(int32x4_t __p0, int16x4_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vget_lane_f16(__p0_743, __p1_743) __extension__ ({ \ - float16x4_t __s0_743 = __p0_743; \ - float16_t __ret_743; \ -float16x4_t __reint_743 = __s0_743; \ -int16_t __reint1_743 = vget_lane_s16(*(int16x4_t *) &__reint_743, __p1_743); \ - __ret_743 = *(float16_t *) &__reint1_743; \ - __ret_743; \ +#define vget_lane_f16(__p0_835, __p1_835) __extension__ ({ \ + float16x4_t __s0_835 = __p0_835; \ + float16_t __ret_835; \ +float16x4_t __reint_835 = __s0_835; \ +int16_t __reint1_835 = vget_lane_s16(*(int16x4_t *) &__reint_835, __p1_835); \ + __ret_835 = *(float16_t *) &__reint1_835; \ + __ret_835; \ }) #else -#define vget_lane_f16(__p0_744, __p1_744) __extension__ ({ \ - float16x4_t __s0_744 = __p0_744; \ - float16x4_t __rev0_744; __rev0_744 = __builtin_shufflevector(__s0_744, __s0_744, 3, 2, 1, 0); \ - float16_t __ret_744; \ -float16x4_t __reint_744 = __rev0_744; \ -int16_t __reint1_744 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_744, __p1_744); \ - __ret_744 = *(float16_t *) &__reint1_744; \ - __ret_744; \ +#define vget_lane_f16(__p0_836, __p1_836) __extension__ ({ \ + float16x4_t __s0_836 = __p0_836; \ + float16x4_t __rev0_836; __rev0_836 = __builtin_shufflevector(__s0_836, __s0_836, 3, 2, 1, 0); \ + float16_t __ret_836; \ +float16x4_t __reint_836 = __rev0_836; \ +int16_t __reint1_836 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_836, __p1_836); \ + __ret_836 = *(float16_t *) &__reint1_836; \ + __ret_836; \ }) -#define __noswap_vget_lane_f16(__p0_745, __p1_745) __extension__ ({ \ - float16x4_t __s0_745 = __p0_745; \ - float16_t __ret_745; \ -float16x4_t __reint_745 = __s0_745; \ -int16_t __reint1_745 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_745, __p1_745); \ - __ret_745 = *(float16_t *) &__reint1_745; \ - __ret_745; \ +#define __noswap_vget_lane_f16(__p0_837, __p1_837) __extension__ ({ \ + float16x4_t __s0_837 = __p0_837; \ + float16_t __ret_837; \ +float16x4_t __reint_837 = __s0_837; \ +int16_t __reint1_837 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_837, __p1_837); \ + __ret_837 = *(float16_t *) &__reint1_837; \ + __ret_837; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vgetq_lane_f16(__p0_746, __p1_746) __extension__ ({ \ - float16x8_t __s0_746 = __p0_746; \ - float16_t __ret_746; \ -float16x8_t __reint_746 = __s0_746; \ -int16_t __reint1_746 = vgetq_lane_s16(*(int16x8_t *) &__reint_746, __p1_746); \ - __ret_746 = *(float16_t *) &__reint1_746; \ - __ret_746; \ +#define vgetq_lane_f16(__p0_838, __p1_838) __extension__ ({ \ + float16x8_t __s0_838 = __p0_838; \ + float16_t __ret_838; \ +float16x8_t __reint_838 = __s0_838; \ +int16_t __reint1_838 = vgetq_lane_s16(*(int16x8_t *) &__reint_838, __p1_838); \ + __ret_838 = *(float16_t *) &__reint1_838; \ + __ret_838; \ }) #else -#define vgetq_lane_f16(__p0_747, __p1_747) __extension__ ({ \ - float16x8_t __s0_747 = __p0_747; \ - float16x8_t __rev0_747; __rev0_747 = __builtin_shufflevector(__s0_747, __s0_747, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16_t __ret_747; \ -float16x8_t __reint_747 = __rev0_747; \ -int16_t __reint1_747 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_747, __p1_747); \ - __ret_747 = *(float16_t *) &__reint1_747; \ - __ret_747; \ +#define vgetq_lane_f16(__p0_839, __p1_839) __extension__ ({ \ + float16x8_t __s0_839 = __p0_839; \ + float16x8_t __rev0_839; __rev0_839 = __builtin_shufflevector(__s0_839, __s0_839, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16_t __ret_839; \ +float16x8_t __reint_839 = __rev0_839; \ +int16_t __reint1_839 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_839, __p1_839); \ + __ret_839 = *(float16_t *) &__reint1_839; \ + __ret_839; \ }) -#define __noswap_vgetq_lane_f16(__p0_748, __p1_748) __extension__ ({ \ - float16x8_t __s0_748 = __p0_748; \ - float16_t __ret_748; \ -float16x8_t __reint_748 = __s0_748; \ -int16_t __reint1_748 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_748, __p1_748); \ - __ret_748 = *(float16_t *) &__reint1_748; \ - __ret_748; \ +#define __noswap_vgetq_lane_f16(__p0_840, __p1_840) __extension__ ({ \ + float16x8_t __s0_840 = __p0_840; \ + float16_t __ret_840; \ +float16x8_t __reint_840 = __s0_840; \ +int16_t __reint1_840 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_840, __p1_840); \ + __ret_840 = *(float16_t *) &__reint1_840; \ + __ret_840; \ }) #endif @@ -64870,98 +66540,98 @@ __ai int32x4_t __noswap_vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2 #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_lane_u32(__p0_749, __p1_749, __p2_749, __p3_749) __extension__ ({ \ - uint64x2_t __s0_749 = __p0_749; \ - uint32x2_t __s1_749 = __p1_749; \ - uint32x2_t __s2_749 = __p2_749; \ - uint64x2_t __ret_749; \ - __ret_749 = __s0_749 + vmull_u32(__s1_749, splat_lane_u32(__s2_749, __p3_749)); \ - __ret_749; \ +#define vmlal_lane_u32(__p0_841, __p1_841, __p2_841, __p3_841) __extension__ ({ \ + uint64x2_t __s0_841 = __p0_841; \ + uint32x2_t __s1_841 = __p1_841; \ + uint32x2_t __s2_841 = __p2_841; \ + uint64x2_t __ret_841; \ + __ret_841 = __s0_841 + vmull_u32(__s1_841, splat_lane_u32(__s2_841, __p3_841)); \ + __ret_841; \ }) #else -#define vmlal_lane_u32(__p0_750, __p1_750, __p2_750, __p3_750) __extension__ ({ \ - uint64x2_t __s0_750 = __p0_750; \ - uint32x2_t __s1_750 = __p1_750; \ - uint32x2_t __s2_750 = __p2_750; \ - uint64x2_t __rev0_750; __rev0_750 = __builtin_shufflevector(__s0_750, __s0_750, 1, 0); \ - uint32x2_t __rev1_750; __rev1_750 = __builtin_shufflevector(__s1_750, __s1_750, 1, 0); \ - uint32x2_t __rev2_750; __rev2_750 = __builtin_shufflevector(__s2_750, __s2_750, 1, 0); \ - uint64x2_t __ret_750; \ - __ret_750 = __rev0_750 + __noswap_vmull_u32(__rev1_750, __noswap_splat_lane_u32(__rev2_750, __p3_750)); \ - __ret_750 = __builtin_shufflevector(__ret_750, __ret_750, 1, 0); \ - __ret_750; \ +#define vmlal_lane_u32(__p0_842, __p1_842, __p2_842, __p3_842) __extension__ ({ \ + uint64x2_t __s0_842 = __p0_842; \ + uint32x2_t __s1_842 = __p1_842; \ + uint32x2_t __s2_842 = __p2_842; \ + uint64x2_t __rev0_842; __rev0_842 = __builtin_shufflevector(__s0_842, __s0_842, 1, 0); \ + uint32x2_t __rev1_842; __rev1_842 = __builtin_shufflevector(__s1_842, __s1_842, 1, 0); \ + uint32x2_t __rev2_842; __rev2_842 = __builtin_shufflevector(__s2_842, __s2_842, 1, 0); \ + uint64x2_t __ret_842; \ + __ret_842 = __rev0_842 + __noswap_vmull_u32(__rev1_842, __noswap_splat_lane_u32(__rev2_842, __p3_842)); \ + __ret_842 = __builtin_shufflevector(__ret_842, __ret_842, 1, 0); \ + __ret_842; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_lane_u16(__p0_751, __p1_751, __p2_751, __p3_751) __extension__ ({ \ - uint32x4_t __s0_751 = __p0_751; \ - uint16x4_t __s1_751 = __p1_751; \ - uint16x4_t __s2_751 = __p2_751; \ - uint32x4_t __ret_751; \ - __ret_751 = __s0_751 + vmull_u16(__s1_751, splat_lane_u16(__s2_751, __p3_751)); \ - __ret_751; \ +#define vmlal_lane_u16(__p0_843, __p1_843, __p2_843, __p3_843) __extension__ ({ \ + uint32x4_t __s0_843 = __p0_843; \ + uint16x4_t __s1_843 = __p1_843; \ + uint16x4_t __s2_843 = __p2_843; \ + uint32x4_t __ret_843; \ + __ret_843 = __s0_843 + vmull_u16(__s1_843, splat_lane_u16(__s2_843, __p3_843)); \ + __ret_843; \ }) #else -#define vmlal_lane_u16(__p0_752, __p1_752, __p2_752, __p3_752) __extension__ ({ \ - uint32x4_t __s0_752 = __p0_752; \ - uint16x4_t __s1_752 = __p1_752; \ - uint16x4_t __s2_752 = __p2_752; \ - uint32x4_t __rev0_752; __rev0_752 = __builtin_shufflevector(__s0_752, __s0_752, 3, 2, 1, 0); \ - uint16x4_t __rev1_752; __rev1_752 = __builtin_shufflevector(__s1_752, __s1_752, 3, 2, 1, 0); \ - uint16x4_t __rev2_752; __rev2_752 = __builtin_shufflevector(__s2_752, __s2_752, 3, 2, 1, 0); \ - uint32x4_t __ret_752; \ - __ret_752 = __rev0_752 + __noswap_vmull_u16(__rev1_752, __noswap_splat_lane_u16(__rev2_752, __p3_752)); \ - __ret_752 = __builtin_shufflevector(__ret_752, __ret_752, 3, 2, 1, 0); \ - __ret_752; \ +#define vmlal_lane_u16(__p0_844, __p1_844, __p2_844, __p3_844) __extension__ ({ \ + uint32x4_t __s0_844 = __p0_844; \ + uint16x4_t __s1_844 = __p1_844; \ + uint16x4_t __s2_844 = __p2_844; \ + uint32x4_t __rev0_844; __rev0_844 = __builtin_shufflevector(__s0_844, __s0_844, 3, 2, 1, 0); \ + uint16x4_t __rev1_844; __rev1_844 = __builtin_shufflevector(__s1_844, __s1_844, 3, 2, 1, 0); \ + uint16x4_t __rev2_844; __rev2_844 = __builtin_shufflevector(__s2_844, __s2_844, 3, 2, 1, 0); \ + uint32x4_t __ret_844; \ + __ret_844 = __rev0_844 + __noswap_vmull_u16(__rev1_844, __noswap_splat_lane_u16(__rev2_844, __p3_844)); \ + __ret_844 = __builtin_shufflevector(__ret_844, __ret_844, 3, 2, 1, 0); \ + __ret_844; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_lane_s32(__p0_753, __p1_753, __p2_753, __p3_753) __extension__ ({ \ - int64x2_t __s0_753 = __p0_753; \ - int32x2_t __s1_753 = __p1_753; \ - int32x2_t __s2_753 = __p2_753; \ - int64x2_t __ret_753; \ - __ret_753 = __s0_753 + vmull_s32(__s1_753, splat_lane_s32(__s2_753, __p3_753)); \ - __ret_753; \ +#define vmlal_lane_s32(__p0_845, __p1_845, __p2_845, __p3_845) __extension__ ({ \ + int64x2_t __s0_845 = __p0_845; \ + int32x2_t __s1_845 = __p1_845; \ + int32x2_t __s2_845 = __p2_845; \ + int64x2_t __ret_845; \ + __ret_845 = __s0_845 + vmull_s32(__s1_845, splat_lane_s32(__s2_845, __p3_845)); \ + __ret_845; \ }) #else -#define vmlal_lane_s32(__p0_754, __p1_754, __p2_754, __p3_754) __extension__ ({ \ - int64x2_t __s0_754 = __p0_754; \ - int32x2_t __s1_754 = __p1_754; \ - int32x2_t __s2_754 = __p2_754; \ - int64x2_t __rev0_754; __rev0_754 = __builtin_shufflevector(__s0_754, __s0_754, 1, 0); \ - int32x2_t __rev1_754; __rev1_754 = __builtin_shufflevector(__s1_754, __s1_754, 1, 0); \ - int32x2_t __rev2_754; __rev2_754 = __builtin_shufflevector(__s2_754, __s2_754, 1, 0); \ - int64x2_t __ret_754; \ - __ret_754 = __rev0_754 + __noswap_vmull_s32(__rev1_754, __noswap_splat_lane_s32(__rev2_754, __p3_754)); \ - __ret_754 = __builtin_shufflevector(__ret_754, __ret_754, 1, 0); \ - __ret_754; \ +#define vmlal_lane_s32(__p0_846, __p1_846, __p2_846, __p3_846) __extension__ ({ \ + int64x2_t __s0_846 = __p0_846; \ + int32x2_t __s1_846 = __p1_846; \ + int32x2_t __s2_846 = __p2_846; \ + int64x2_t __rev0_846; __rev0_846 = __builtin_shufflevector(__s0_846, __s0_846, 1, 0); \ + int32x2_t __rev1_846; __rev1_846 = __builtin_shufflevector(__s1_846, __s1_846, 1, 0); \ + int32x2_t __rev2_846; __rev2_846 = __builtin_shufflevector(__s2_846, __s2_846, 1, 0); \ + int64x2_t __ret_846; \ + __ret_846 = __rev0_846 + __noswap_vmull_s32(__rev1_846, __noswap_splat_lane_s32(__rev2_846, __p3_846)); \ + __ret_846 = __builtin_shufflevector(__ret_846, __ret_846, 1, 0); \ + __ret_846; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlal_lane_s16(__p0_755, __p1_755, __p2_755, __p3_755) __extension__ ({ \ - int32x4_t __s0_755 = __p0_755; \ - int16x4_t __s1_755 = __p1_755; \ - int16x4_t __s2_755 = __p2_755; \ - int32x4_t __ret_755; \ - __ret_755 = __s0_755 + vmull_s16(__s1_755, splat_lane_s16(__s2_755, __p3_755)); \ - __ret_755; \ +#define vmlal_lane_s16(__p0_847, __p1_847, __p2_847, __p3_847) __extension__ ({ \ + int32x4_t __s0_847 = __p0_847; \ + int16x4_t __s1_847 = __p1_847; \ + int16x4_t __s2_847 = __p2_847; \ + int32x4_t __ret_847; \ + __ret_847 = __s0_847 + vmull_s16(__s1_847, splat_lane_s16(__s2_847, __p3_847)); \ + __ret_847; \ }) #else -#define vmlal_lane_s16(__p0_756, __p1_756, __p2_756, __p3_756) __extension__ ({ \ - int32x4_t __s0_756 = __p0_756; \ - int16x4_t __s1_756 = __p1_756; \ - int16x4_t __s2_756 = __p2_756; \ - int32x4_t __rev0_756; __rev0_756 = __builtin_shufflevector(__s0_756, __s0_756, 3, 2, 1, 0); \ - int16x4_t __rev1_756; __rev1_756 = __builtin_shufflevector(__s1_756, __s1_756, 3, 2, 1, 0); \ - int16x4_t __rev2_756; __rev2_756 = __builtin_shufflevector(__s2_756, __s2_756, 3, 2, 1, 0); \ - int32x4_t __ret_756; \ - __ret_756 = __rev0_756 + __noswap_vmull_s16(__rev1_756, __noswap_splat_lane_s16(__rev2_756, __p3_756)); \ - __ret_756 = __builtin_shufflevector(__ret_756, __ret_756, 3, 2, 1, 0); \ - __ret_756; \ +#define vmlal_lane_s16(__p0_848, __p1_848, __p2_848, __p3_848) __extension__ ({ \ + int32x4_t __s0_848 = __p0_848; \ + int16x4_t __s1_848 = __p1_848; \ + int16x4_t __s2_848 = __p2_848; \ + int32x4_t __rev0_848; __rev0_848 = __builtin_shufflevector(__s0_848, __s0_848, 3, 2, 1, 0); \ + int16x4_t __rev1_848; __rev1_848 = __builtin_shufflevector(__s1_848, __s1_848, 3, 2, 1, 0); \ + int16x4_t __rev2_848; __rev2_848 = __builtin_shufflevector(__s2_848, __s2_848, 3, 2, 1, 0); \ + int32x4_t __ret_848; \ + __ret_848 = __rev0_848 + __noswap_vmull_s16(__rev1_848, __noswap_splat_lane_s16(__rev2_848, __p3_848)); \ + __ret_848 = __builtin_shufflevector(__ret_848, __ret_848, 3, 2, 1, 0); \ + __ret_848; \ }) #endif @@ -65192,98 +66862,98 @@ __ai int32x4_t __noswap_vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2 #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_lane_u32(__p0_757, __p1_757, __p2_757, __p3_757) __extension__ ({ \ - uint64x2_t __s0_757 = __p0_757; \ - uint32x2_t __s1_757 = __p1_757; \ - uint32x2_t __s2_757 = __p2_757; \ - uint64x2_t __ret_757; \ - __ret_757 = __s0_757 - vmull_u32(__s1_757, splat_lane_u32(__s2_757, __p3_757)); \ - __ret_757; \ +#define vmlsl_lane_u32(__p0_849, __p1_849, __p2_849, __p3_849) __extension__ ({ \ + uint64x2_t __s0_849 = __p0_849; \ + uint32x2_t __s1_849 = __p1_849; \ + uint32x2_t __s2_849 = __p2_849; \ + uint64x2_t __ret_849; \ + __ret_849 = __s0_849 - vmull_u32(__s1_849, splat_lane_u32(__s2_849, __p3_849)); \ + __ret_849; \ }) #else -#define vmlsl_lane_u32(__p0_758, __p1_758, __p2_758, __p3_758) __extension__ ({ \ - uint64x2_t __s0_758 = __p0_758; \ - uint32x2_t __s1_758 = __p1_758; \ - uint32x2_t __s2_758 = __p2_758; \ - uint64x2_t __rev0_758; __rev0_758 = __builtin_shufflevector(__s0_758, __s0_758, 1, 0); \ - uint32x2_t __rev1_758; __rev1_758 = __builtin_shufflevector(__s1_758, __s1_758, 1, 0); \ - uint32x2_t __rev2_758; __rev2_758 = __builtin_shufflevector(__s2_758, __s2_758, 1, 0); \ - uint64x2_t __ret_758; \ - __ret_758 = __rev0_758 - __noswap_vmull_u32(__rev1_758, __noswap_splat_lane_u32(__rev2_758, __p3_758)); \ - __ret_758 = __builtin_shufflevector(__ret_758, __ret_758, 1, 0); \ - __ret_758; \ +#define vmlsl_lane_u32(__p0_850, __p1_850, __p2_850, __p3_850) __extension__ ({ \ + uint64x2_t __s0_850 = __p0_850; \ + uint32x2_t __s1_850 = __p1_850; \ + uint32x2_t __s2_850 = __p2_850; \ + uint64x2_t __rev0_850; __rev0_850 = __builtin_shufflevector(__s0_850, __s0_850, 1, 0); \ + uint32x2_t __rev1_850; __rev1_850 = __builtin_shufflevector(__s1_850, __s1_850, 1, 0); \ + uint32x2_t __rev2_850; __rev2_850 = __builtin_shufflevector(__s2_850, __s2_850, 1, 0); \ + uint64x2_t __ret_850; \ + __ret_850 = __rev0_850 - __noswap_vmull_u32(__rev1_850, __noswap_splat_lane_u32(__rev2_850, __p3_850)); \ + __ret_850 = __builtin_shufflevector(__ret_850, __ret_850, 1, 0); \ + __ret_850; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_lane_u16(__p0_759, __p1_759, __p2_759, __p3_759) __extension__ ({ \ - uint32x4_t __s0_759 = __p0_759; \ - uint16x4_t __s1_759 = __p1_759; \ - uint16x4_t __s2_759 = __p2_759; \ - uint32x4_t __ret_759; \ - __ret_759 = __s0_759 - vmull_u16(__s1_759, splat_lane_u16(__s2_759, __p3_759)); \ - __ret_759; \ +#define vmlsl_lane_u16(__p0_851, __p1_851, __p2_851, __p3_851) __extension__ ({ \ + uint32x4_t __s0_851 = __p0_851; \ + uint16x4_t __s1_851 = __p1_851; \ + uint16x4_t __s2_851 = __p2_851; \ + uint32x4_t __ret_851; \ + __ret_851 = __s0_851 - vmull_u16(__s1_851, splat_lane_u16(__s2_851, __p3_851)); \ + __ret_851; \ }) #else -#define vmlsl_lane_u16(__p0_760, __p1_760, __p2_760, __p3_760) __extension__ ({ \ - uint32x4_t __s0_760 = __p0_760; \ - uint16x4_t __s1_760 = __p1_760; \ - uint16x4_t __s2_760 = __p2_760; \ - uint32x4_t __rev0_760; __rev0_760 = __builtin_shufflevector(__s0_760, __s0_760, 3, 2, 1, 0); \ - uint16x4_t __rev1_760; __rev1_760 = __builtin_shufflevector(__s1_760, __s1_760, 3, 2, 1, 0); \ - uint16x4_t __rev2_760; __rev2_760 = __builtin_shufflevector(__s2_760, __s2_760, 3, 2, 1, 0); \ - uint32x4_t __ret_760; \ - __ret_760 = __rev0_760 - __noswap_vmull_u16(__rev1_760, __noswap_splat_lane_u16(__rev2_760, __p3_760)); \ - __ret_760 = __builtin_shufflevector(__ret_760, __ret_760, 3, 2, 1, 0); \ - __ret_760; \ +#define vmlsl_lane_u16(__p0_852, __p1_852, __p2_852, __p3_852) __extension__ ({ \ + uint32x4_t __s0_852 = __p0_852; \ + uint16x4_t __s1_852 = __p1_852; \ + uint16x4_t __s2_852 = __p2_852; \ + uint32x4_t __rev0_852; __rev0_852 = __builtin_shufflevector(__s0_852, __s0_852, 3, 2, 1, 0); \ + uint16x4_t __rev1_852; __rev1_852 = __builtin_shufflevector(__s1_852, __s1_852, 3, 2, 1, 0); \ + uint16x4_t __rev2_852; __rev2_852 = __builtin_shufflevector(__s2_852, __s2_852, 3, 2, 1, 0); \ + uint32x4_t __ret_852; \ + __ret_852 = __rev0_852 - __noswap_vmull_u16(__rev1_852, __noswap_splat_lane_u16(__rev2_852, __p3_852)); \ + __ret_852 = __builtin_shufflevector(__ret_852, __ret_852, 3, 2, 1, 0); \ + __ret_852; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_lane_s32(__p0_761, __p1_761, __p2_761, __p3_761) __extension__ ({ \ - int64x2_t __s0_761 = __p0_761; \ - int32x2_t __s1_761 = __p1_761; \ - int32x2_t __s2_761 = __p2_761; \ - int64x2_t __ret_761; \ - __ret_761 = __s0_761 - vmull_s32(__s1_761, splat_lane_s32(__s2_761, __p3_761)); \ - __ret_761; \ +#define vmlsl_lane_s32(__p0_853, __p1_853, __p2_853, __p3_853) __extension__ ({ \ + int64x2_t __s0_853 = __p0_853; \ + int32x2_t __s1_853 = __p1_853; \ + int32x2_t __s2_853 = __p2_853; \ + int64x2_t __ret_853; \ + __ret_853 = __s0_853 - vmull_s32(__s1_853, splat_lane_s32(__s2_853, __p3_853)); \ + __ret_853; \ }) #else -#define vmlsl_lane_s32(__p0_762, __p1_762, __p2_762, __p3_762) __extension__ ({ \ - int64x2_t __s0_762 = __p0_762; \ - int32x2_t __s1_762 = __p1_762; \ - int32x2_t __s2_762 = __p2_762; \ - int64x2_t __rev0_762; __rev0_762 = __builtin_shufflevector(__s0_762, __s0_762, 1, 0); \ - int32x2_t __rev1_762; __rev1_762 = __builtin_shufflevector(__s1_762, __s1_762, 1, 0); \ - int32x2_t __rev2_762; __rev2_762 = __builtin_shufflevector(__s2_762, __s2_762, 1, 0); \ - int64x2_t __ret_762; \ - __ret_762 = __rev0_762 - __noswap_vmull_s32(__rev1_762, __noswap_splat_lane_s32(__rev2_762, __p3_762)); \ - __ret_762 = __builtin_shufflevector(__ret_762, __ret_762, 1, 0); \ - __ret_762; \ +#define vmlsl_lane_s32(__p0_854, __p1_854, __p2_854, __p3_854) __extension__ ({ \ + int64x2_t __s0_854 = __p0_854; \ + int32x2_t __s1_854 = __p1_854; \ + int32x2_t __s2_854 = __p2_854; \ + int64x2_t __rev0_854; __rev0_854 = __builtin_shufflevector(__s0_854, __s0_854, 1, 0); \ + int32x2_t __rev1_854; __rev1_854 = __builtin_shufflevector(__s1_854, __s1_854, 1, 0); \ + int32x2_t __rev2_854; __rev2_854 = __builtin_shufflevector(__s2_854, __s2_854, 1, 0); \ + int64x2_t __ret_854; \ + __ret_854 = __rev0_854 - __noswap_vmull_s32(__rev1_854, __noswap_splat_lane_s32(__rev2_854, __p3_854)); \ + __ret_854 = __builtin_shufflevector(__ret_854, __ret_854, 1, 0); \ + __ret_854; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmlsl_lane_s16(__p0_763, __p1_763, __p2_763, __p3_763) __extension__ ({ \ - int32x4_t __s0_763 = __p0_763; \ - int16x4_t __s1_763 = __p1_763; \ - int16x4_t __s2_763 = __p2_763; \ - int32x4_t __ret_763; \ - __ret_763 = __s0_763 - vmull_s16(__s1_763, splat_lane_s16(__s2_763, __p3_763)); \ - __ret_763; \ +#define vmlsl_lane_s16(__p0_855, __p1_855, __p2_855, __p3_855) __extension__ ({ \ + int32x4_t __s0_855 = __p0_855; \ + int16x4_t __s1_855 = __p1_855; \ + int16x4_t __s2_855 = __p2_855; \ + int32x4_t __ret_855; \ + __ret_855 = __s0_855 - vmull_s16(__s1_855, splat_lane_s16(__s2_855, __p3_855)); \ + __ret_855; \ }) #else -#define vmlsl_lane_s16(__p0_764, __p1_764, __p2_764, __p3_764) __extension__ ({ \ - int32x4_t __s0_764 = __p0_764; \ - int16x4_t __s1_764 = __p1_764; \ - int16x4_t __s2_764 = __p2_764; \ - int32x4_t __rev0_764; __rev0_764 = __builtin_shufflevector(__s0_764, __s0_764, 3, 2, 1, 0); \ - int16x4_t __rev1_764; __rev1_764 = __builtin_shufflevector(__s1_764, __s1_764, 3, 2, 1, 0); \ - int16x4_t __rev2_764; __rev2_764 = __builtin_shufflevector(__s2_764, __s2_764, 3, 2, 1, 0); \ - int32x4_t __ret_764; \ - __ret_764 = __rev0_764 - __noswap_vmull_s16(__rev1_764, __noswap_splat_lane_s16(__rev2_764, __p3_764)); \ - __ret_764 = __builtin_shufflevector(__ret_764, __ret_764, 3, 2, 1, 0); \ - __ret_764; \ +#define vmlsl_lane_s16(__p0_856, __p1_856, __p2_856, __p3_856) __extension__ ({ \ + int32x4_t __s0_856 = __p0_856; \ + int16x4_t __s1_856 = __p1_856; \ + int16x4_t __s2_856 = __p2_856; \ + int32x4_t __rev0_856; __rev0_856 = __builtin_shufflevector(__s0_856, __s0_856, 3, 2, 1, 0); \ + int16x4_t __rev1_856; __rev1_856 = __builtin_shufflevector(__s1_856, __s1_856, 3, 2, 1, 0); \ + int16x4_t __rev2_856; __rev2_856 = __builtin_shufflevector(__s2_856, __s2_856, 3, 2, 1, 0); \ + int32x4_t __ret_856; \ + __ret_856 = __rev0_856 - __noswap_vmull_s16(__rev1_856, __noswap_splat_lane_s16(__rev2_856, __p3_856)); \ + __ret_856 = __builtin_shufflevector(__ret_856, __ret_856, 3, 2, 1, 0); \ + __ret_856; \ }) #endif @@ -65376,151 +67046,151 @@ __ai int32x4_t __noswap_vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2 #endif #ifdef __LITTLE_ENDIAN__ -#define vset_lane_f16(__p0_765, __p1_765, __p2_765) __extension__ ({ \ - float16_t __s0_765 = __p0_765; \ - float16x4_t __s1_765 = __p1_765; \ - float16x4_t __ret_765; \ -float16_t __reint_765 = __s0_765; \ -float16x4_t __reint1_765 = __s1_765; \ -int16x4_t __reint2_765 = vset_lane_s16(*(int16_t *) &__reint_765, *(int16x4_t *) &__reint1_765, __p2_765); \ - __ret_765 = *(float16x4_t *) &__reint2_765; \ - __ret_765; \ +#define vset_lane_f16(__p0_857, __p1_857, __p2_857) __extension__ ({ \ + float16_t __s0_857 = __p0_857; \ + float16x4_t __s1_857 = __p1_857; \ + float16x4_t __ret_857; \ +float16_t __reint_857 = __s0_857; \ +float16x4_t __reint1_857 = __s1_857; \ +int16x4_t __reint2_857 = vset_lane_s16(*(int16_t *) &__reint_857, *(int16x4_t *) &__reint1_857, __p2_857); \ + __ret_857 = *(float16x4_t *) &__reint2_857; \ + __ret_857; \ }) #else -#define vset_lane_f16(__p0_766, __p1_766, __p2_766) __extension__ ({ \ - float16_t __s0_766 = __p0_766; \ - float16x4_t __s1_766 = __p1_766; \ - float16x4_t __rev1_766; __rev1_766 = __builtin_shufflevector(__s1_766, __s1_766, 3, 2, 1, 0); \ - float16x4_t __ret_766; \ -float16_t __reint_766 = __s0_766; \ -float16x4_t __reint1_766 = __rev1_766; \ -int16x4_t __reint2_766 = __noswap_vset_lane_s16(*(int16_t *) &__reint_766, *(int16x4_t *) &__reint1_766, __p2_766); \ - __ret_766 = *(float16x4_t *) &__reint2_766; \ - __ret_766 = __builtin_shufflevector(__ret_766, __ret_766, 3, 2, 1, 0); \ - __ret_766; \ +#define vset_lane_f16(__p0_858, __p1_858, __p2_858) __extension__ ({ \ + float16_t __s0_858 = __p0_858; \ + float16x4_t __s1_858 = __p1_858; \ + float16x4_t __rev1_858; __rev1_858 = __builtin_shufflevector(__s1_858, __s1_858, 3, 2, 1, 0); \ + float16x4_t __ret_858; \ +float16_t __reint_858 = __s0_858; \ +float16x4_t __reint1_858 = __rev1_858; \ +int16x4_t __reint2_858 = __noswap_vset_lane_s16(*(int16_t *) &__reint_858, *(int16x4_t *) &__reint1_858, __p2_858); \ + __ret_858 = *(float16x4_t *) &__reint2_858; \ + __ret_858 = __builtin_shufflevector(__ret_858, __ret_858, 3, 2, 1, 0); \ + __ret_858; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vsetq_lane_f16(__p0_767, __p1_767, __p2_767) __extension__ ({ \ - float16_t __s0_767 = __p0_767; \ - float16x8_t __s1_767 = __p1_767; \ - float16x8_t __ret_767; \ -float16_t __reint_767 = __s0_767; \ -float16x8_t __reint1_767 = __s1_767; \ -int16x8_t __reint2_767 = vsetq_lane_s16(*(int16_t *) &__reint_767, *(int16x8_t *) &__reint1_767, __p2_767); \ - __ret_767 = *(float16x8_t *) &__reint2_767; \ - __ret_767; \ +#define vsetq_lane_f16(__p0_859, __p1_859, __p2_859) __extension__ ({ \ + float16_t __s0_859 = __p0_859; \ + float16x8_t __s1_859 = __p1_859; \ + float16x8_t __ret_859; \ +float16_t __reint_859 = __s0_859; \ +float16x8_t __reint1_859 = __s1_859; \ +int16x8_t __reint2_859 = vsetq_lane_s16(*(int16_t *) &__reint_859, *(int16x8_t *) &__reint1_859, __p2_859); \ + __ret_859 = *(float16x8_t *) &__reint2_859; \ + __ret_859; \ }) #else -#define vsetq_lane_f16(__p0_768, __p1_768, __p2_768) __extension__ ({ \ - float16_t __s0_768 = __p0_768; \ - float16x8_t __s1_768 = __p1_768; \ - float16x8_t __rev1_768; __rev1_768 = __builtin_shufflevector(__s1_768, __s1_768, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __ret_768; \ -float16_t __reint_768 = __s0_768; \ -float16x8_t __reint1_768 = __rev1_768; \ -int16x8_t __reint2_768 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_768, *(int16x8_t *) &__reint1_768, __p2_768); \ - __ret_768 = *(float16x8_t *) &__reint2_768; \ - __ret_768 = __builtin_shufflevector(__ret_768, __ret_768, 7, 6, 5, 4, 3, 2, 1, 0); \ - __ret_768; \ +#define vsetq_lane_f16(__p0_860, __p1_860, __p2_860) __extension__ ({ \ + float16_t __s0_860 = __p0_860; \ + float16x8_t __s1_860 = __p1_860; \ + float16x8_t __rev1_860; __rev1_860 = __builtin_shufflevector(__s1_860, __s1_860, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __ret_860; \ +float16_t __reint_860 = __s0_860; \ +float16x8_t __reint1_860 = __rev1_860; \ +int16x8_t __reint2_860 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_860, *(int16x8_t *) &__reint1_860, __p2_860); \ + __ret_860 = *(float16x8_t *) &__reint2_860; \ + __ret_860 = __builtin_shufflevector(__ret_860, __ret_860, 7, 6, 5, 4, 3, 2, 1, 0); \ + __ret_860; \ }) #endif #if defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) #ifdef __LITTLE_ENDIAN__ -#define vbfmlalbq_lane_f32(__p0_769, __p1_769, __p2_769, __p3_769) __extension__ ({ \ - float32x4_t __s0_769 = __p0_769; \ - bfloat16x8_t __s1_769 = __p1_769; \ - bfloat16x4_t __s2_769 = __p2_769; \ - float32x4_t __ret_769; \ - __ret_769 = vbfmlalbq_f32(__s0_769, __s1_769, (bfloat16x8_t) {vget_lane_bf16(__s2_769, __p3_769), vget_lane_bf16(__s2_769, __p3_769), vget_lane_bf16(__s2_769, __p3_769), vget_lane_bf16(__s2_769, __p3_769), vget_lane_bf16(__s2_769, __p3_769), vget_lane_bf16(__s2_769, __p3_769), vget_lane_bf16(__s2_769, __p3_769), vget_lane_bf16(__s2_769, __p3_769)}); \ - __ret_769; \ +#define vbfmlalbq_lane_f32(__p0_861, __p1_861, __p2_861, __p3_861) __extension__ ({ \ + float32x4_t __s0_861 = __p0_861; \ + bfloat16x8_t __s1_861 = __p1_861; \ + bfloat16x4_t __s2_861 = __p2_861; \ + float32x4_t __ret_861; \ + __ret_861 = vbfmlalbq_f32(__s0_861, __s1_861, (bfloat16x8_t) {vget_lane_bf16(__s2_861, __p3_861), vget_lane_bf16(__s2_861, __p3_861), vget_lane_bf16(__s2_861, __p3_861), vget_lane_bf16(__s2_861, __p3_861), vget_lane_bf16(__s2_861, __p3_861), vget_lane_bf16(__s2_861, __p3_861), vget_lane_bf16(__s2_861, __p3_861), vget_lane_bf16(__s2_861, __p3_861)}); \ + __ret_861; \ }) #else -#define vbfmlalbq_lane_f32(__p0_770, __p1_770, __p2_770, __p3_770) __extension__ ({ \ - float32x4_t __s0_770 = __p0_770; \ - bfloat16x8_t __s1_770 = __p1_770; \ - bfloat16x4_t __s2_770 = __p2_770; \ - float32x4_t __rev0_770; __rev0_770 = __builtin_shufflevector(__s0_770, __s0_770, 3, 2, 1, 0); \ - bfloat16x8_t __rev1_770; __rev1_770 = __builtin_shufflevector(__s1_770, __s1_770, 7, 6, 5, 4, 3, 2, 1, 0); \ - bfloat16x4_t __rev2_770; __rev2_770 = __builtin_shufflevector(__s2_770, __s2_770, 3, 2, 1, 0); \ - float32x4_t __ret_770; \ - __ret_770 = __noswap_vbfmlalbq_f32(__rev0_770, __rev1_770, (bfloat16x8_t) {__noswap_vget_lane_bf16(__rev2_770, __p3_770), __noswap_vget_lane_bf16(__rev2_770, __p3_770), __noswap_vget_lane_bf16(__rev2_770, __p3_770), __noswap_vget_lane_bf16(__rev2_770, __p3_770), __noswap_vget_lane_bf16(__rev2_770, __p3_770), __noswap_vget_lane_bf16(__rev2_770, __p3_770), __noswap_vget_lane_bf16(__rev2_770, __p3_770), __noswap_vget_lane_bf16(__rev2_770, __p3_770)}); \ - __ret_770 = __builtin_shufflevector(__ret_770, __ret_770, 3, 2, 1, 0); \ - __ret_770; \ +#define vbfmlalbq_lane_f32(__p0_862, __p1_862, __p2_862, __p3_862) __extension__ ({ \ + float32x4_t __s0_862 = __p0_862; \ + bfloat16x8_t __s1_862 = __p1_862; \ + bfloat16x4_t __s2_862 = __p2_862; \ + float32x4_t __rev0_862; __rev0_862 = __builtin_shufflevector(__s0_862, __s0_862, 3, 2, 1, 0); \ + bfloat16x8_t __rev1_862; __rev1_862 = __builtin_shufflevector(__s1_862, __s1_862, 7, 6, 5, 4, 3, 2, 1, 0); \ + bfloat16x4_t __rev2_862; __rev2_862 = __builtin_shufflevector(__s2_862, __s2_862, 3, 2, 1, 0); \ + float32x4_t __ret_862; \ + __ret_862 = __noswap_vbfmlalbq_f32(__rev0_862, __rev1_862, (bfloat16x8_t) {__noswap_vget_lane_bf16(__rev2_862, __p3_862), __noswap_vget_lane_bf16(__rev2_862, __p3_862), __noswap_vget_lane_bf16(__rev2_862, __p3_862), __noswap_vget_lane_bf16(__rev2_862, __p3_862), __noswap_vget_lane_bf16(__rev2_862, __p3_862), __noswap_vget_lane_bf16(__rev2_862, __p3_862), __noswap_vget_lane_bf16(__rev2_862, __p3_862), __noswap_vget_lane_bf16(__rev2_862, __p3_862)}); \ + __ret_862 = __builtin_shufflevector(__ret_862, __ret_862, 3, 2, 1, 0); \ + __ret_862; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vbfmlalbq_laneq_f32(__p0_771, __p1_771, __p2_771, __p3_771) __extension__ ({ \ - float32x4_t __s0_771 = __p0_771; \ - bfloat16x8_t __s1_771 = __p1_771; \ - bfloat16x8_t __s2_771 = __p2_771; \ - float32x4_t __ret_771; \ - __ret_771 = vbfmlalbq_f32(__s0_771, __s1_771, (bfloat16x8_t) {vgetq_lane_bf16(__s2_771, __p3_771), vgetq_lane_bf16(__s2_771, __p3_771), vgetq_lane_bf16(__s2_771, __p3_771), vgetq_lane_bf16(__s2_771, __p3_771), vgetq_lane_bf16(__s2_771, __p3_771), vgetq_lane_bf16(__s2_771, __p3_771), vgetq_lane_bf16(__s2_771, __p3_771), vgetq_lane_bf16(__s2_771, __p3_771)}); \ - __ret_771; \ +#define vbfmlalbq_laneq_f32(__p0_863, __p1_863, __p2_863, __p3_863) __extension__ ({ \ + float32x4_t __s0_863 = __p0_863; \ + bfloat16x8_t __s1_863 = __p1_863; \ + bfloat16x8_t __s2_863 = __p2_863; \ + float32x4_t __ret_863; \ + __ret_863 = vbfmlalbq_f32(__s0_863, __s1_863, (bfloat16x8_t) {vgetq_lane_bf16(__s2_863, __p3_863), vgetq_lane_bf16(__s2_863, __p3_863), vgetq_lane_bf16(__s2_863, __p3_863), vgetq_lane_bf16(__s2_863, __p3_863), vgetq_lane_bf16(__s2_863, __p3_863), vgetq_lane_bf16(__s2_863, __p3_863), vgetq_lane_bf16(__s2_863, __p3_863), vgetq_lane_bf16(__s2_863, __p3_863)}); \ + __ret_863; \ }) #else -#define vbfmlalbq_laneq_f32(__p0_772, __p1_772, __p2_772, __p3_772) __extension__ ({ \ - float32x4_t __s0_772 = __p0_772; \ - bfloat16x8_t __s1_772 = __p1_772; \ - bfloat16x8_t __s2_772 = __p2_772; \ - float32x4_t __rev0_772; __rev0_772 = __builtin_shufflevector(__s0_772, __s0_772, 3, 2, 1, 0); \ - bfloat16x8_t __rev1_772; __rev1_772 = __builtin_shufflevector(__s1_772, __s1_772, 7, 6, 5, 4, 3, 2, 1, 0); \ - bfloat16x8_t __rev2_772; __rev2_772 = __builtin_shufflevector(__s2_772, __s2_772, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x4_t __ret_772; \ - __ret_772 = __noswap_vbfmlalbq_f32(__rev0_772, __rev1_772, (bfloat16x8_t) {__noswap_vgetq_lane_bf16(__rev2_772, __p3_772), __noswap_vgetq_lane_bf16(__rev2_772, __p3_772), __noswap_vgetq_lane_bf16(__rev2_772, __p3_772), __noswap_vgetq_lane_bf16(__rev2_772, __p3_772), __noswap_vgetq_lane_bf16(__rev2_772, __p3_772), __noswap_vgetq_lane_bf16(__rev2_772, __p3_772), __noswap_vgetq_lane_bf16(__rev2_772, __p3_772), __noswap_vgetq_lane_bf16(__rev2_772, __p3_772)}); \ - __ret_772 = __builtin_shufflevector(__ret_772, __ret_772, 3, 2, 1, 0); \ - __ret_772; \ +#define vbfmlalbq_laneq_f32(__p0_864, __p1_864, __p2_864, __p3_864) __extension__ ({ \ + float32x4_t __s0_864 = __p0_864; \ + bfloat16x8_t __s1_864 = __p1_864; \ + bfloat16x8_t __s2_864 = __p2_864; \ + float32x4_t __rev0_864; __rev0_864 = __builtin_shufflevector(__s0_864, __s0_864, 3, 2, 1, 0); \ + bfloat16x8_t __rev1_864; __rev1_864 = __builtin_shufflevector(__s1_864, __s1_864, 7, 6, 5, 4, 3, 2, 1, 0); \ + bfloat16x8_t __rev2_864; __rev2_864 = __builtin_shufflevector(__s2_864, __s2_864, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x4_t __ret_864; \ + __ret_864 = __noswap_vbfmlalbq_f32(__rev0_864, __rev1_864, (bfloat16x8_t) {__noswap_vgetq_lane_bf16(__rev2_864, __p3_864), __noswap_vgetq_lane_bf16(__rev2_864, __p3_864), __noswap_vgetq_lane_bf16(__rev2_864, __p3_864), __noswap_vgetq_lane_bf16(__rev2_864, __p3_864), __noswap_vgetq_lane_bf16(__rev2_864, __p3_864), __noswap_vgetq_lane_bf16(__rev2_864, __p3_864), __noswap_vgetq_lane_bf16(__rev2_864, __p3_864), __noswap_vgetq_lane_bf16(__rev2_864, __p3_864)}); \ + __ret_864 = __builtin_shufflevector(__ret_864, __ret_864, 3, 2, 1, 0); \ + __ret_864; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vbfmlaltq_lane_f32(__p0_773, __p1_773, __p2_773, __p3_773) __extension__ ({ \ - float32x4_t __s0_773 = __p0_773; \ - bfloat16x8_t __s1_773 = __p1_773; \ - bfloat16x4_t __s2_773 = __p2_773; \ - float32x4_t __ret_773; \ - __ret_773 = vbfmlaltq_f32(__s0_773, __s1_773, (bfloat16x8_t) {vget_lane_bf16(__s2_773, __p3_773), vget_lane_bf16(__s2_773, __p3_773), vget_lane_bf16(__s2_773, __p3_773), vget_lane_bf16(__s2_773, __p3_773), vget_lane_bf16(__s2_773, __p3_773), vget_lane_bf16(__s2_773, __p3_773), vget_lane_bf16(__s2_773, __p3_773), vget_lane_bf16(__s2_773, __p3_773)}); \ - __ret_773; \ +#define vbfmlaltq_lane_f32(__p0_865, __p1_865, __p2_865, __p3_865) __extension__ ({ \ + float32x4_t __s0_865 = __p0_865; \ + bfloat16x8_t __s1_865 = __p1_865; \ + bfloat16x4_t __s2_865 = __p2_865; \ + float32x4_t __ret_865; \ + __ret_865 = vbfmlaltq_f32(__s0_865, __s1_865, (bfloat16x8_t) {vget_lane_bf16(__s2_865, __p3_865), vget_lane_bf16(__s2_865, __p3_865), vget_lane_bf16(__s2_865, __p3_865), vget_lane_bf16(__s2_865, __p3_865), vget_lane_bf16(__s2_865, __p3_865), vget_lane_bf16(__s2_865, __p3_865), vget_lane_bf16(__s2_865, __p3_865), vget_lane_bf16(__s2_865, __p3_865)}); \ + __ret_865; \ }) #else -#define vbfmlaltq_lane_f32(__p0_774, __p1_774, __p2_774, __p3_774) __extension__ ({ \ - float32x4_t __s0_774 = __p0_774; \ - bfloat16x8_t __s1_774 = __p1_774; \ - bfloat16x4_t __s2_774 = __p2_774; \ - float32x4_t __rev0_774; __rev0_774 = __builtin_shufflevector(__s0_774, __s0_774, 3, 2, 1, 0); \ - bfloat16x8_t __rev1_774; __rev1_774 = __builtin_shufflevector(__s1_774, __s1_774, 7, 6, 5, 4, 3, 2, 1, 0); \ - bfloat16x4_t __rev2_774; __rev2_774 = __builtin_shufflevector(__s2_774, __s2_774, 3, 2, 1, 0); \ - float32x4_t __ret_774; \ - __ret_774 = __noswap_vbfmlaltq_f32(__rev0_774, __rev1_774, (bfloat16x8_t) {__noswap_vget_lane_bf16(__rev2_774, __p3_774), __noswap_vget_lane_bf16(__rev2_774, __p3_774), __noswap_vget_lane_bf16(__rev2_774, __p3_774), __noswap_vget_lane_bf16(__rev2_774, __p3_774), __noswap_vget_lane_bf16(__rev2_774, __p3_774), __noswap_vget_lane_bf16(__rev2_774, __p3_774), __noswap_vget_lane_bf16(__rev2_774, __p3_774), __noswap_vget_lane_bf16(__rev2_774, __p3_774)}); \ - __ret_774 = __builtin_shufflevector(__ret_774, __ret_774, 3, 2, 1, 0); \ - __ret_774; \ +#define vbfmlaltq_lane_f32(__p0_866, __p1_866, __p2_866, __p3_866) __extension__ ({ \ + float32x4_t __s0_866 = __p0_866; \ + bfloat16x8_t __s1_866 = __p1_866; \ + bfloat16x4_t __s2_866 = __p2_866; \ + float32x4_t __rev0_866; __rev0_866 = __builtin_shufflevector(__s0_866, __s0_866, 3, 2, 1, 0); \ + bfloat16x8_t __rev1_866; __rev1_866 = __builtin_shufflevector(__s1_866, __s1_866, 7, 6, 5, 4, 3, 2, 1, 0); \ + bfloat16x4_t __rev2_866; __rev2_866 = __builtin_shufflevector(__s2_866, __s2_866, 3, 2, 1, 0); \ + float32x4_t __ret_866; \ + __ret_866 = __noswap_vbfmlaltq_f32(__rev0_866, __rev1_866, (bfloat16x8_t) {__noswap_vget_lane_bf16(__rev2_866, __p3_866), __noswap_vget_lane_bf16(__rev2_866, __p3_866), __noswap_vget_lane_bf16(__rev2_866, __p3_866), __noswap_vget_lane_bf16(__rev2_866, __p3_866), __noswap_vget_lane_bf16(__rev2_866, __p3_866), __noswap_vget_lane_bf16(__rev2_866, __p3_866), __noswap_vget_lane_bf16(__rev2_866, __p3_866), __noswap_vget_lane_bf16(__rev2_866, __p3_866)}); \ + __ret_866 = __builtin_shufflevector(__ret_866, __ret_866, 3, 2, 1, 0); \ + __ret_866; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vbfmlaltq_laneq_f32(__p0_775, __p1_775, __p2_775, __p3_775) __extension__ ({ \ - float32x4_t __s0_775 = __p0_775; \ - bfloat16x8_t __s1_775 = __p1_775; \ - bfloat16x8_t __s2_775 = __p2_775; \ - float32x4_t __ret_775; \ - __ret_775 = vbfmlaltq_f32(__s0_775, __s1_775, (bfloat16x8_t) {vgetq_lane_bf16(__s2_775, __p3_775), vgetq_lane_bf16(__s2_775, __p3_775), vgetq_lane_bf16(__s2_775, __p3_775), vgetq_lane_bf16(__s2_775, __p3_775), vgetq_lane_bf16(__s2_775, __p3_775), vgetq_lane_bf16(__s2_775, __p3_775), vgetq_lane_bf16(__s2_775, __p3_775), vgetq_lane_bf16(__s2_775, __p3_775)}); \ - __ret_775; \ +#define vbfmlaltq_laneq_f32(__p0_867, __p1_867, __p2_867, __p3_867) __extension__ ({ \ + float32x4_t __s0_867 = __p0_867; \ + bfloat16x8_t __s1_867 = __p1_867; \ + bfloat16x8_t __s2_867 = __p2_867; \ + float32x4_t __ret_867; \ + __ret_867 = vbfmlaltq_f32(__s0_867, __s1_867, (bfloat16x8_t) {vgetq_lane_bf16(__s2_867, __p3_867), vgetq_lane_bf16(__s2_867, __p3_867), vgetq_lane_bf16(__s2_867, __p3_867), vgetq_lane_bf16(__s2_867, __p3_867), vgetq_lane_bf16(__s2_867, __p3_867), vgetq_lane_bf16(__s2_867, __p3_867), vgetq_lane_bf16(__s2_867, __p3_867), vgetq_lane_bf16(__s2_867, __p3_867)}); \ + __ret_867; \ }) #else -#define vbfmlaltq_laneq_f32(__p0_776, __p1_776, __p2_776, __p3_776) __extension__ ({ \ - float32x4_t __s0_776 = __p0_776; \ - bfloat16x8_t __s1_776 = __p1_776; \ - bfloat16x8_t __s2_776 = __p2_776; \ - float32x4_t __rev0_776; __rev0_776 = __builtin_shufflevector(__s0_776, __s0_776, 3, 2, 1, 0); \ - bfloat16x8_t __rev1_776; __rev1_776 = __builtin_shufflevector(__s1_776, __s1_776, 7, 6, 5, 4, 3, 2, 1, 0); \ - bfloat16x8_t __rev2_776; __rev2_776 = __builtin_shufflevector(__s2_776, __s2_776, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x4_t __ret_776; \ - __ret_776 = __noswap_vbfmlaltq_f32(__rev0_776, __rev1_776, (bfloat16x8_t) {__noswap_vgetq_lane_bf16(__rev2_776, __p3_776), __noswap_vgetq_lane_bf16(__rev2_776, __p3_776), __noswap_vgetq_lane_bf16(__rev2_776, __p3_776), __noswap_vgetq_lane_bf16(__rev2_776, __p3_776), __noswap_vgetq_lane_bf16(__rev2_776, __p3_776), __noswap_vgetq_lane_bf16(__rev2_776, __p3_776), __noswap_vgetq_lane_bf16(__rev2_776, __p3_776), __noswap_vgetq_lane_bf16(__rev2_776, __p3_776)}); \ - __ret_776 = __builtin_shufflevector(__ret_776, __ret_776, 3, 2, 1, 0); \ - __ret_776; \ +#define vbfmlaltq_laneq_f32(__p0_868, __p1_868, __p2_868, __p3_868) __extension__ ({ \ + float32x4_t __s0_868 = __p0_868; \ + bfloat16x8_t __s1_868 = __p1_868; \ + bfloat16x8_t __s2_868 = __p2_868; \ + float32x4_t __rev0_868; __rev0_868 = __builtin_shufflevector(__s0_868, __s0_868, 3, 2, 1, 0); \ + bfloat16x8_t __rev1_868; __rev1_868 = __builtin_shufflevector(__s1_868, __s1_868, 7, 6, 5, 4, 3, 2, 1, 0); \ + bfloat16x8_t __rev2_868; __rev2_868 = __builtin_shufflevector(__s2_868, __s2_868, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x4_t __ret_868; \ + __ret_868 = __noswap_vbfmlaltq_f32(__rev0_868, __rev1_868, (bfloat16x8_t) {__noswap_vgetq_lane_bf16(__rev2_868, __p3_868), __noswap_vgetq_lane_bf16(__rev2_868, __p3_868), __noswap_vgetq_lane_bf16(__rev2_868, __p3_868), __noswap_vgetq_lane_bf16(__rev2_868, __p3_868), __noswap_vgetq_lane_bf16(__rev2_868, __p3_868), __noswap_vgetq_lane_bf16(__rev2_868, __p3_868), __noswap_vgetq_lane_bf16(__rev2_868, __p3_868), __noswap_vgetq_lane_bf16(__rev2_868, __p3_868)}); \ + __ret_868 = __builtin_shufflevector(__ret_868, __ret_868, 3, 2, 1, 0); \ + __ret_868; \ }) #endif @@ -65559,480 +67229,480 @@ __ai float32x4_t vcvtq_low_f32_bf16(bfloat16x8_t __p0) { #endif #if defined(__ARM_FEATURE_FP16FML) && defined(__aarch64__) #ifdef __LITTLE_ENDIAN__ -#define vfmlalq_lane_high_f16(__p0_777, __p1_777, __p2_777, __p3_777) __extension__ ({ \ - float32x4_t __s0_777 = __p0_777; \ - float16x8_t __s1_777 = __p1_777; \ - float16x4_t __s2_777 = __p2_777; \ - float32x4_t __ret_777; \ - __ret_777 = vfmlalq_high_f16(__s0_777, __s1_777, (float16x8_t) {vget_lane_f16(__s2_777, __p3_777), vget_lane_f16(__s2_777, __p3_777), vget_lane_f16(__s2_777, __p3_777), vget_lane_f16(__s2_777, __p3_777), vget_lane_f16(__s2_777, __p3_777), vget_lane_f16(__s2_777, __p3_777), vget_lane_f16(__s2_777, __p3_777), vget_lane_f16(__s2_777, __p3_777)}); \ - __ret_777; \ +#define vfmlalq_lane_high_f16(__p0_869, __p1_869, __p2_869, __p3_869) __extension__ ({ \ + float32x4_t __s0_869 = __p0_869; \ + float16x8_t __s1_869 = __p1_869; \ + float16x4_t __s2_869 = __p2_869; \ + float32x4_t __ret_869; \ + __ret_869 = vfmlalq_high_f16(__s0_869, __s1_869, (float16x8_t) {vget_lane_f16(__s2_869, __p3_869), vget_lane_f16(__s2_869, __p3_869), vget_lane_f16(__s2_869, __p3_869), vget_lane_f16(__s2_869, __p3_869), vget_lane_f16(__s2_869, __p3_869), vget_lane_f16(__s2_869, __p3_869), vget_lane_f16(__s2_869, __p3_869), vget_lane_f16(__s2_869, __p3_869)}); \ + __ret_869; \ }) #else -#define vfmlalq_lane_high_f16(__p0_778, __p1_778, __p2_778, __p3_778) __extension__ ({ \ - float32x4_t __s0_778 = __p0_778; \ - float16x8_t __s1_778 = __p1_778; \ - float16x4_t __s2_778 = __p2_778; \ - float32x4_t __rev0_778; __rev0_778 = __builtin_shufflevector(__s0_778, __s0_778, 3, 2, 1, 0); \ - float16x8_t __rev1_778; __rev1_778 = __builtin_shufflevector(__s1_778, __s1_778, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __rev2_778; __rev2_778 = __builtin_shufflevector(__s2_778, __s2_778, 3, 2, 1, 0); \ - float32x4_t __ret_778; \ - __ret_778 = __noswap_vfmlalq_high_f16(__rev0_778, __rev1_778, (float16x8_t) {__noswap_vget_lane_f16(__rev2_778, __p3_778), __noswap_vget_lane_f16(__rev2_778, __p3_778), __noswap_vget_lane_f16(__rev2_778, __p3_778), __noswap_vget_lane_f16(__rev2_778, __p3_778), __noswap_vget_lane_f16(__rev2_778, __p3_778), __noswap_vget_lane_f16(__rev2_778, __p3_778), __noswap_vget_lane_f16(__rev2_778, __p3_778), __noswap_vget_lane_f16(__rev2_778, __p3_778)}); \ - __ret_778 = __builtin_shufflevector(__ret_778, __ret_778, 3, 2, 1, 0); \ - __ret_778; \ +#define vfmlalq_lane_high_f16(__p0_870, __p1_870, __p2_870, __p3_870) __extension__ ({ \ + float32x4_t __s0_870 = __p0_870; \ + float16x8_t __s1_870 = __p1_870; \ + float16x4_t __s2_870 = __p2_870; \ + float32x4_t __rev0_870; __rev0_870 = __builtin_shufflevector(__s0_870, __s0_870, 3, 2, 1, 0); \ + float16x8_t __rev1_870; __rev1_870 = __builtin_shufflevector(__s1_870, __s1_870, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_870; __rev2_870 = __builtin_shufflevector(__s2_870, __s2_870, 3, 2, 1, 0); \ + float32x4_t __ret_870; \ + __ret_870 = __noswap_vfmlalq_high_f16(__rev0_870, __rev1_870, (float16x8_t) {__noswap_vget_lane_f16(__rev2_870, __p3_870), __noswap_vget_lane_f16(__rev2_870, __p3_870), __noswap_vget_lane_f16(__rev2_870, __p3_870), __noswap_vget_lane_f16(__rev2_870, __p3_870), __noswap_vget_lane_f16(__rev2_870, __p3_870), __noswap_vget_lane_f16(__rev2_870, __p3_870), __noswap_vget_lane_f16(__rev2_870, __p3_870), __noswap_vget_lane_f16(__rev2_870, __p3_870)}); \ + __ret_870 = __builtin_shufflevector(__ret_870, __ret_870, 3, 2, 1, 0); \ + __ret_870; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlal_lane_high_f16(__p0_779, __p1_779, __p2_779, __p3_779) __extension__ ({ \ - float32x2_t __s0_779 = __p0_779; \ - float16x4_t __s1_779 = __p1_779; \ - float16x4_t __s2_779 = __p2_779; \ - float32x2_t __ret_779; \ - __ret_779 = vfmlal_high_f16(__s0_779, __s1_779, (float16x4_t) {vget_lane_f16(__s2_779, __p3_779), vget_lane_f16(__s2_779, __p3_779), vget_lane_f16(__s2_779, __p3_779), vget_lane_f16(__s2_779, __p3_779)}); \ - __ret_779; \ +#define vfmlal_lane_high_f16(__p0_871, __p1_871, __p2_871, __p3_871) __extension__ ({ \ + float32x2_t __s0_871 = __p0_871; \ + float16x4_t __s1_871 = __p1_871; \ + float16x4_t __s2_871 = __p2_871; \ + float32x2_t __ret_871; \ + __ret_871 = vfmlal_high_f16(__s0_871, __s1_871, (float16x4_t) {vget_lane_f16(__s2_871, __p3_871), vget_lane_f16(__s2_871, __p3_871), vget_lane_f16(__s2_871, __p3_871), vget_lane_f16(__s2_871, __p3_871)}); \ + __ret_871; \ }) #else -#define vfmlal_lane_high_f16(__p0_780, __p1_780, __p2_780, __p3_780) __extension__ ({ \ - float32x2_t __s0_780 = __p0_780; \ - float16x4_t __s1_780 = __p1_780; \ - float16x4_t __s2_780 = __p2_780; \ - float32x2_t __rev0_780; __rev0_780 = __builtin_shufflevector(__s0_780, __s0_780, 1, 0); \ - float16x4_t __rev1_780; __rev1_780 = __builtin_shufflevector(__s1_780, __s1_780, 3, 2, 1, 0); \ - float16x4_t __rev2_780; __rev2_780 = __builtin_shufflevector(__s2_780, __s2_780, 3, 2, 1, 0); \ - float32x2_t __ret_780; \ - __ret_780 = __noswap_vfmlal_high_f16(__rev0_780, __rev1_780, (float16x4_t) {__noswap_vget_lane_f16(__rev2_780, __p3_780), __noswap_vget_lane_f16(__rev2_780, __p3_780), __noswap_vget_lane_f16(__rev2_780, __p3_780), __noswap_vget_lane_f16(__rev2_780, __p3_780)}); \ - __ret_780 = __builtin_shufflevector(__ret_780, __ret_780, 1, 0); \ - __ret_780; \ +#define vfmlal_lane_high_f16(__p0_872, __p1_872, __p2_872, __p3_872) __extension__ ({ \ + float32x2_t __s0_872 = __p0_872; \ + float16x4_t __s1_872 = __p1_872; \ + float16x4_t __s2_872 = __p2_872; \ + float32x2_t __rev0_872; __rev0_872 = __builtin_shufflevector(__s0_872, __s0_872, 1, 0); \ + float16x4_t __rev1_872; __rev1_872 = __builtin_shufflevector(__s1_872, __s1_872, 3, 2, 1, 0); \ + float16x4_t __rev2_872; __rev2_872 = __builtin_shufflevector(__s2_872, __s2_872, 3, 2, 1, 0); \ + float32x2_t __ret_872; \ + __ret_872 = __noswap_vfmlal_high_f16(__rev0_872, __rev1_872, (float16x4_t) {__noswap_vget_lane_f16(__rev2_872, __p3_872), __noswap_vget_lane_f16(__rev2_872, __p3_872), __noswap_vget_lane_f16(__rev2_872, __p3_872), __noswap_vget_lane_f16(__rev2_872, __p3_872)}); \ + __ret_872 = __builtin_shufflevector(__ret_872, __ret_872, 1, 0); \ + __ret_872; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlalq_lane_low_f16(__p0_781, __p1_781, __p2_781, __p3_781) __extension__ ({ \ - float32x4_t __s0_781 = __p0_781; \ - float16x8_t __s1_781 = __p1_781; \ - float16x4_t __s2_781 = __p2_781; \ - float32x4_t __ret_781; \ - __ret_781 = vfmlalq_low_f16(__s0_781, __s1_781, (float16x8_t) {vget_lane_f16(__s2_781, __p3_781), vget_lane_f16(__s2_781, __p3_781), vget_lane_f16(__s2_781, __p3_781), vget_lane_f16(__s2_781, __p3_781), vget_lane_f16(__s2_781, __p3_781), vget_lane_f16(__s2_781, __p3_781), vget_lane_f16(__s2_781, __p3_781), vget_lane_f16(__s2_781, __p3_781)}); \ - __ret_781; \ +#define vfmlalq_lane_low_f16(__p0_873, __p1_873, __p2_873, __p3_873) __extension__ ({ \ + float32x4_t __s0_873 = __p0_873; \ + float16x8_t __s1_873 = __p1_873; \ + float16x4_t __s2_873 = __p2_873; \ + float32x4_t __ret_873; \ + __ret_873 = vfmlalq_low_f16(__s0_873, __s1_873, (float16x8_t) {vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873), vget_lane_f16(__s2_873, __p3_873)}); \ + __ret_873; \ }) #else -#define vfmlalq_lane_low_f16(__p0_782, __p1_782, __p2_782, __p3_782) __extension__ ({ \ - float32x4_t __s0_782 = __p0_782; \ - float16x8_t __s1_782 = __p1_782; \ - float16x4_t __s2_782 = __p2_782; \ - float32x4_t __rev0_782; __rev0_782 = __builtin_shufflevector(__s0_782, __s0_782, 3, 2, 1, 0); \ - float16x8_t __rev1_782; __rev1_782 = __builtin_shufflevector(__s1_782, __s1_782, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __rev2_782; __rev2_782 = __builtin_shufflevector(__s2_782, __s2_782, 3, 2, 1, 0); \ - float32x4_t __ret_782; \ - __ret_782 = __noswap_vfmlalq_low_f16(__rev0_782, __rev1_782, (float16x8_t) {__noswap_vget_lane_f16(__rev2_782, __p3_782), __noswap_vget_lane_f16(__rev2_782, __p3_782), __noswap_vget_lane_f16(__rev2_782, __p3_782), __noswap_vget_lane_f16(__rev2_782, __p3_782), __noswap_vget_lane_f16(__rev2_782, __p3_782), __noswap_vget_lane_f16(__rev2_782, __p3_782), __noswap_vget_lane_f16(__rev2_782, __p3_782), __noswap_vget_lane_f16(__rev2_782, __p3_782)}); \ - __ret_782 = __builtin_shufflevector(__ret_782, __ret_782, 3, 2, 1, 0); \ - __ret_782; \ +#define vfmlalq_lane_low_f16(__p0_874, __p1_874, __p2_874, __p3_874) __extension__ ({ \ + float32x4_t __s0_874 = __p0_874; \ + float16x8_t __s1_874 = __p1_874; \ + float16x4_t __s2_874 = __p2_874; \ + float32x4_t __rev0_874; __rev0_874 = __builtin_shufflevector(__s0_874, __s0_874, 3, 2, 1, 0); \ + float16x8_t __rev1_874; __rev1_874 = __builtin_shufflevector(__s1_874, __s1_874, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_874; __rev2_874 = __builtin_shufflevector(__s2_874, __s2_874, 3, 2, 1, 0); \ + float32x4_t __ret_874; \ + __ret_874 = __noswap_vfmlalq_low_f16(__rev0_874, __rev1_874, (float16x8_t) {__noswap_vget_lane_f16(__rev2_874, __p3_874), __noswap_vget_lane_f16(__rev2_874, __p3_874), __noswap_vget_lane_f16(__rev2_874, __p3_874), __noswap_vget_lane_f16(__rev2_874, __p3_874), __noswap_vget_lane_f16(__rev2_874, __p3_874), __noswap_vget_lane_f16(__rev2_874, __p3_874), __noswap_vget_lane_f16(__rev2_874, __p3_874), __noswap_vget_lane_f16(__rev2_874, __p3_874)}); \ + __ret_874 = __builtin_shufflevector(__ret_874, __ret_874, 3, 2, 1, 0); \ + __ret_874; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlal_lane_low_f16(__p0_783, __p1_783, __p2_783, __p3_783) __extension__ ({ \ - float32x2_t __s0_783 = __p0_783; \ - float16x4_t __s1_783 = __p1_783; \ - float16x4_t __s2_783 = __p2_783; \ - float32x2_t __ret_783; \ - __ret_783 = vfmlal_low_f16(__s0_783, __s1_783, (float16x4_t) {vget_lane_f16(__s2_783, __p3_783), vget_lane_f16(__s2_783, __p3_783), vget_lane_f16(__s2_783, __p3_783), vget_lane_f16(__s2_783, __p3_783)}); \ - __ret_783; \ +#define vfmlal_lane_low_f16(__p0_875, __p1_875, __p2_875, __p3_875) __extension__ ({ \ + float32x2_t __s0_875 = __p0_875; \ + float16x4_t __s1_875 = __p1_875; \ + float16x4_t __s2_875 = __p2_875; \ + float32x2_t __ret_875; \ + __ret_875 = vfmlal_low_f16(__s0_875, __s1_875, (float16x4_t) {vget_lane_f16(__s2_875, __p3_875), vget_lane_f16(__s2_875, __p3_875), vget_lane_f16(__s2_875, __p3_875), vget_lane_f16(__s2_875, __p3_875)}); \ + __ret_875; \ }) #else -#define vfmlal_lane_low_f16(__p0_784, __p1_784, __p2_784, __p3_784) __extension__ ({ \ - float32x2_t __s0_784 = __p0_784; \ - float16x4_t __s1_784 = __p1_784; \ - float16x4_t __s2_784 = __p2_784; \ - float32x2_t __rev0_784; __rev0_784 = __builtin_shufflevector(__s0_784, __s0_784, 1, 0); \ - float16x4_t __rev1_784; __rev1_784 = __builtin_shufflevector(__s1_784, __s1_784, 3, 2, 1, 0); \ - float16x4_t __rev2_784; __rev2_784 = __builtin_shufflevector(__s2_784, __s2_784, 3, 2, 1, 0); \ - float32x2_t __ret_784; \ - __ret_784 = __noswap_vfmlal_low_f16(__rev0_784, __rev1_784, (float16x4_t) {__noswap_vget_lane_f16(__rev2_784, __p3_784), __noswap_vget_lane_f16(__rev2_784, __p3_784), __noswap_vget_lane_f16(__rev2_784, __p3_784), __noswap_vget_lane_f16(__rev2_784, __p3_784)}); \ - __ret_784 = __builtin_shufflevector(__ret_784, __ret_784, 1, 0); \ - __ret_784; \ +#define vfmlal_lane_low_f16(__p0_876, __p1_876, __p2_876, __p3_876) __extension__ ({ \ + float32x2_t __s0_876 = __p0_876; \ + float16x4_t __s1_876 = __p1_876; \ + float16x4_t __s2_876 = __p2_876; \ + float32x2_t __rev0_876; __rev0_876 = __builtin_shufflevector(__s0_876, __s0_876, 1, 0); \ + float16x4_t __rev1_876; __rev1_876 = __builtin_shufflevector(__s1_876, __s1_876, 3, 2, 1, 0); \ + float16x4_t __rev2_876; __rev2_876 = __builtin_shufflevector(__s2_876, __s2_876, 3, 2, 1, 0); \ + float32x2_t __ret_876; \ + __ret_876 = __noswap_vfmlal_low_f16(__rev0_876, __rev1_876, (float16x4_t) {__noswap_vget_lane_f16(__rev2_876, __p3_876), __noswap_vget_lane_f16(__rev2_876, __p3_876), __noswap_vget_lane_f16(__rev2_876, __p3_876), __noswap_vget_lane_f16(__rev2_876, __p3_876)}); \ + __ret_876 = __builtin_shufflevector(__ret_876, __ret_876, 1, 0); \ + __ret_876; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlalq_laneq_high_f16(__p0_785, __p1_785, __p2_785, __p3_785) __extension__ ({ \ - float32x4_t __s0_785 = __p0_785; \ - float16x8_t __s1_785 = __p1_785; \ - float16x8_t __s2_785 = __p2_785; \ - float32x4_t __ret_785; \ - __ret_785 = vfmlalq_high_f16(__s0_785, __s1_785, (float16x8_t) {vgetq_lane_f16(__s2_785, __p3_785), vgetq_lane_f16(__s2_785, __p3_785), vgetq_lane_f16(__s2_785, __p3_785), vgetq_lane_f16(__s2_785, __p3_785), vgetq_lane_f16(__s2_785, __p3_785), vgetq_lane_f16(__s2_785, __p3_785), vgetq_lane_f16(__s2_785, __p3_785), vgetq_lane_f16(__s2_785, __p3_785)}); \ - __ret_785; \ +#define vfmlalq_laneq_high_f16(__p0_877, __p1_877, __p2_877, __p3_877) __extension__ ({ \ + float32x4_t __s0_877 = __p0_877; \ + float16x8_t __s1_877 = __p1_877; \ + float16x8_t __s2_877 = __p2_877; \ + float32x4_t __ret_877; \ + __ret_877 = vfmlalq_high_f16(__s0_877, __s1_877, (float16x8_t) {vgetq_lane_f16(__s2_877, __p3_877), vgetq_lane_f16(__s2_877, __p3_877), vgetq_lane_f16(__s2_877, __p3_877), vgetq_lane_f16(__s2_877, __p3_877), vgetq_lane_f16(__s2_877, __p3_877), vgetq_lane_f16(__s2_877, __p3_877), vgetq_lane_f16(__s2_877, __p3_877), vgetq_lane_f16(__s2_877, __p3_877)}); \ + __ret_877; \ }) #else -#define vfmlalq_laneq_high_f16(__p0_786, __p1_786, __p2_786, __p3_786) __extension__ ({ \ - float32x4_t __s0_786 = __p0_786; \ - float16x8_t __s1_786 = __p1_786; \ - float16x8_t __s2_786 = __p2_786; \ - float32x4_t __rev0_786; __rev0_786 = __builtin_shufflevector(__s0_786, __s0_786, 3, 2, 1, 0); \ - float16x8_t __rev1_786; __rev1_786 = __builtin_shufflevector(__s1_786, __s1_786, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev2_786; __rev2_786 = __builtin_shufflevector(__s2_786, __s2_786, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x4_t __ret_786; \ - __ret_786 = __noswap_vfmlalq_high_f16(__rev0_786, __rev1_786, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_786, __p3_786), __noswap_vgetq_lane_f16(__rev2_786, __p3_786), __noswap_vgetq_lane_f16(__rev2_786, __p3_786), __noswap_vgetq_lane_f16(__rev2_786, __p3_786), __noswap_vgetq_lane_f16(__rev2_786, __p3_786), __noswap_vgetq_lane_f16(__rev2_786, __p3_786), __noswap_vgetq_lane_f16(__rev2_786, __p3_786), __noswap_vgetq_lane_f16(__rev2_786, __p3_786)}); \ - __ret_786 = __builtin_shufflevector(__ret_786, __ret_786, 3, 2, 1, 0); \ - __ret_786; \ +#define vfmlalq_laneq_high_f16(__p0_878, __p1_878, __p2_878, __p3_878) __extension__ ({ \ + float32x4_t __s0_878 = __p0_878; \ + float16x8_t __s1_878 = __p1_878; \ + float16x8_t __s2_878 = __p2_878; \ + float32x4_t __rev0_878; __rev0_878 = __builtin_shufflevector(__s0_878, __s0_878, 3, 2, 1, 0); \ + float16x8_t __rev1_878; __rev1_878 = __builtin_shufflevector(__s1_878, __s1_878, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_878; __rev2_878 = __builtin_shufflevector(__s2_878, __s2_878, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x4_t __ret_878; \ + __ret_878 = __noswap_vfmlalq_high_f16(__rev0_878, __rev1_878, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_878, __p3_878), __noswap_vgetq_lane_f16(__rev2_878, __p3_878), __noswap_vgetq_lane_f16(__rev2_878, __p3_878), __noswap_vgetq_lane_f16(__rev2_878, __p3_878), __noswap_vgetq_lane_f16(__rev2_878, __p3_878), __noswap_vgetq_lane_f16(__rev2_878, __p3_878), __noswap_vgetq_lane_f16(__rev2_878, __p3_878), __noswap_vgetq_lane_f16(__rev2_878, __p3_878)}); \ + __ret_878 = __builtin_shufflevector(__ret_878, __ret_878, 3, 2, 1, 0); \ + __ret_878; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlal_laneq_high_f16(__p0_787, __p1_787, __p2_787, __p3_787) __extension__ ({ \ - float32x2_t __s0_787 = __p0_787; \ - float16x4_t __s1_787 = __p1_787; \ - float16x8_t __s2_787 = __p2_787; \ - float32x2_t __ret_787; \ - __ret_787 = vfmlal_high_f16(__s0_787, __s1_787, (float16x4_t) {vgetq_lane_f16(__s2_787, __p3_787), vgetq_lane_f16(__s2_787, __p3_787), vgetq_lane_f16(__s2_787, __p3_787), vgetq_lane_f16(__s2_787, __p3_787)}); \ - __ret_787; \ +#define vfmlal_laneq_high_f16(__p0_879, __p1_879, __p2_879, __p3_879) __extension__ ({ \ + float32x2_t __s0_879 = __p0_879; \ + float16x4_t __s1_879 = __p1_879; \ + float16x8_t __s2_879 = __p2_879; \ + float32x2_t __ret_879; \ + __ret_879 = vfmlal_high_f16(__s0_879, __s1_879, (float16x4_t) {vgetq_lane_f16(__s2_879, __p3_879), vgetq_lane_f16(__s2_879, __p3_879), vgetq_lane_f16(__s2_879, __p3_879), vgetq_lane_f16(__s2_879, __p3_879)}); \ + __ret_879; \ }) #else -#define vfmlal_laneq_high_f16(__p0_788, __p1_788, __p2_788, __p3_788) __extension__ ({ \ - float32x2_t __s0_788 = __p0_788; \ - float16x4_t __s1_788 = __p1_788; \ - float16x8_t __s2_788 = __p2_788; \ - float32x2_t __rev0_788; __rev0_788 = __builtin_shufflevector(__s0_788, __s0_788, 1, 0); \ - float16x4_t __rev1_788; __rev1_788 = __builtin_shufflevector(__s1_788, __s1_788, 3, 2, 1, 0); \ - float16x8_t __rev2_788; __rev2_788 = __builtin_shufflevector(__s2_788, __s2_788, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x2_t __ret_788; \ - __ret_788 = __noswap_vfmlal_high_f16(__rev0_788, __rev1_788, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_788, __p3_788), __noswap_vgetq_lane_f16(__rev2_788, __p3_788), __noswap_vgetq_lane_f16(__rev2_788, __p3_788), __noswap_vgetq_lane_f16(__rev2_788, __p3_788)}); \ - __ret_788 = __builtin_shufflevector(__ret_788, __ret_788, 1, 0); \ - __ret_788; \ +#define vfmlal_laneq_high_f16(__p0_880, __p1_880, __p2_880, __p3_880) __extension__ ({ \ + float32x2_t __s0_880 = __p0_880; \ + float16x4_t __s1_880 = __p1_880; \ + float16x8_t __s2_880 = __p2_880; \ + float32x2_t __rev0_880; __rev0_880 = __builtin_shufflevector(__s0_880, __s0_880, 1, 0); \ + float16x4_t __rev1_880; __rev1_880 = __builtin_shufflevector(__s1_880, __s1_880, 3, 2, 1, 0); \ + float16x8_t __rev2_880; __rev2_880 = __builtin_shufflevector(__s2_880, __s2_880, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x2_t __ret_880; \ + __ret_880 = __noswap_vfmlal_high_f16(__rev0_880, __rev1_880, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_880, __p3_880), __noswap_vgetq_lane_f16(__rev2_880, __p3_880), __noswap_vgetq_lane_f16(__rev2_880, __p3_880), __noswap_vgetq_lane_f16(__rev2_880, __p3_880)}); \ + __ret_880 = __builtin_shufflevector(__ret_880, __ret_880, 1, 0); \ + __ret_880; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlalq_laneq_low_f16(__p0_789, __p1_789, __p2_789, __p3_789) __extension__ ({ \ - float32x4_t __s0_789 = __p0_789; \ - float16x8_t __s1_789 = __p1_789; \ - float16x8_t __s2_789 = __p2_789; \ - float32x4_t __ret_789; \ - __ret_789 = vfmlalq_low_f16(__s0_789, __s1_789, (float16x8_t) {vgetq_lane_f16(__s2_789, __p3_789), vgetq_lane_f16(__s2_789, __p3_789), vgetq_lane_f16(__s2_789, __p3_789), vgetq_lane_f16(__s2_789, __p3_789), vgetq_lane_f16(__s2_789, __p3_789), vgetq_lane_f16(__s2_789, __p3_789), vgetq_lane_f16(__s2_789, __p3_789), vgetq_lane_f16(__s2_789, __p3_789)}); \ - __ret_789; \ +#define vfmlalq_laneq_low_f16(__p0_881, __p1_881, __p2_881, __p3_881) __extension__ ({ \ + float32x4_t __s0_881 = __p0_881; \ + float16x8_t __s1_881 = __p1_881; \ + float16x8_t __s2_881 = __p2_881; \ + float32x4_t __ret_881; \ + __ret_881 = vfmlalq_low_f16(__s0_881, __s1_881, (float16x8_t) {vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881), vgetq_lane_f16(__s2_881, __p3_881)}); \ + __ret_881; \ }) #else -#define vfmlalq_laneq_low_f16(__p0_790, __p1_790, __p2_790, __p3_790) __extension__ ({ \ - float32x4_t __s0_790 = __p0_790; \ - float16x8_t __s1_790 = __p1_790; \ - float16x8_t __s2_790 = __p2_790; \ - float32x4_t __rev0_790; __rev0_790 = __builtin_shufflevector(__s0_790, __s0_790, 3, 2, 1, 0); \ - float16x8_t __rev1_790; __rev1_790 = __builtin_shufflevector(__s1_790, __s1_790, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev2_790; __rev2_790 = __builtin_shufflevector(__s2_790, __s2_790, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x4_t __ret_790; \ - __ret_790 = __noswap_vfmlalq_low_f16(__rev0_790, __rev1_790, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_790, __p3_790), __noswap_vgetq_lane_f16(__rev2_790, __p3_790), __noswap_vgetq_lane_f16(__rev2_790, __p3_790), __noswap_vgetq_lane_f16(__rev2_790, __p3_790), __noswap_vgetq_lane_f16(__rev2_790, __p3_790), __noswap_vgetq_lane_f16(__rev2_790, __p3_790), __noswap_vgetq_lane_f16(__rev2_790, __p3_790), __noswap_vgetq_lane_f16(__rev2_790, __p3_790)}); \ - __ret_790 = __builtin_shufflevector(__ret_790, __ret_790, 3, 2, 1, 0); \ - __ret_790; \ +#define vfmlalq_laneq_low_f16(__p0_882, __p1_882, __p2_882, __p3_882) __extension__ ({ \ + float32x4_t __s0_882 = __p0_882; \ + float16x8_t __s1_882 = __p1_882; \ + float16x8_t __s2_882 = __p2_882; \ + float32x4_t __rev0_882; __rev0_882 = __builtin_shufflevector(__s0_882, __s0_882, 3, 2, 1, 0); \ + float16x8_t __rev1_882; __rev1_882 = __builtin_shufflevector(__s1_882, __s1_882, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_882; __rev2_882 = __builtin_shufflevector(__s2_882, __s2_882, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x4_t __ret_882; \ + __ret_882 = __noswap_vfmlalq_low_f16(__rev0_882, __rev1_882, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_882, __p3_882), __noswap_vgetq_lane_f16(__rev2_882, __p3_882), __noswap_vgetq_lane_f16(__rev2_882, __p3_882), __noswap_vgetq_lane_f16(__rev2_882, __p3_882), __noswap_vgetq_lane_f16(__rev2_882, __p3_882), __noswap_vgetq_lane_f16(__rev2_882, __p3_882), __noswap_vgetq_lane_f16(__rev2_882, __p3_882), __noswap_vgetq_lane_f16(__rev2_882, __p3_882)}); \ + __ret_882 = __builtin_shufflevector(__ret_882, __ret_882, 3, 2, 1, 0); \ + __ret_882; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlal_laneq_low_f16(__p0_791, __p1_791, __p2_791, __p3_791) __extension__ ({ \ - float32x2_t __s0_791 = __p0_791; \ - float16x4_t __s1_791 = __p1_791; \ - float16x8_t __s2_791 = __p2_791; \ - float32x2_t __ret_791; \ - __ret_791 = vfmlal_low_f16(__s0_791, __s1_791, (float16x4_t) {vgetq_lane_f16(__s2_791, __p3_791), vgetq_lane_f16(__s2_791, __p3_791), vgetq_lane_f16(__s2_791, __p3_791), vgetq_lane_f16(__s2_791, __p3_791)}); \ - __ret_791; \ +#define vfmlal_laneq_low_f16(__p0_883, __p1_883, __p2_883, __p3_883) __extension__ ({ \ + float32x2_t __s0_883 = __p0_883; \ + float16x4_t __s1_883 = __p1_883; \ + float16x8_t __s2_883 = __p2_883; \ + float32x2_t __ret_883; \ + __ret_883 = vfmlal_low_f16(__s0_883, __s1_883, (float16x4_t) {vgetq_lane_f16(__s2_883, __p3_883), vgetq_lane_f16(__s2_883, __p3_883), vgetq_lane_f16(__s2_883, __p3_883), vgetq_lane_f16(__s2_883, __p3_883)}); \ + __ret_883; \ }) #else -#define vfmlal_laneq_low_f16(__p0_792, __p1_792, __p2_792, __p3_792) __extension__ ({ \ - float32x2_t __s0_792 = __p0_792; \ - float16x4_t __s1_792 = __p1_792; \ - float16x8_t __s2_792 = __p2_792; \ - float32x2_t __rev0_792; __rev0_792 = __builtin_shufflevector(__s0_792, __s0_792, 1, 0); \ - float16x4_t __rev1_792; __rev1_792 = __builtin_shufflevector(__s1_792, __s1_792, 3, 2, 1, 0); \ - float16x8_t __rev2_792; __rev2_792 = __builtin_shufflevector(__s2_792, __s2_792, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x2_t __ret_792; \ - __ret_792 = __noswap_vfmlal_low_f16(__rev0_792, __rev1_792, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_792, __p3_792), __noswap_vgetq_lane_f16(__rev2_792, __p3_792), __noswap_vgetq_lane_f16(__rev2_792, __p3_792), __noswap_vgetq_lane_f16(__rev2_792, __p3_792)}); \ - __ret_792 = __builtin_shufflevector(__ret_792, __ret_792, 1, 0); \ - __ret_792; \ +#define vfmlal_laneq_low_f16(__p0_884, __p1_884, __p2_884, __p3_884) __extension__ ({ \ + float32x2_t __s0_884 = __p0_884; \ + float16x4_t __s1_884 = __p1_884; \ + float16x8_t __s2_884 = __p2_884; \ + float32x2_t __rev0_884; __rev0_884 = __builtin_shufflevector(__s0_884, __s0_884, 1, 0); \ + float16x4_t __rev1_884; __rev1_884 = __builtin_shufflevector(__s1_884, __s1_884, 3, 2, 1, 0); \ + float16x8_t __rev2_884; __rev2_884 = __builtin_shufflevector(__s2_884, __s2_884, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x2_t __ret_884; \ + __ret_884 = __noswap_vfmlal_low_f16(__rev0_884, __rev1_884, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_884, __p3_884), __noswap_vgetq_lane_f16(__rev2_884, __p3_884), __noswap_vgetq_lane_f16(__rev2_884, __p3_884), __noswap_vgetq_lane_f16(__rev2_884, __p3_884)}); \ + __ret_884 = __builtin_shufflevector(__ret_884, __ret_884, 1, 0); \ + __ret_884; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlslq_lane_high_f16(__p0_793, __p1_793, __p2_793, __p3_793) __extension__ ({ \ - float32x4_t __s0_793 = __p0_793; \ - float16x8_t __s1_793 = __p1_793; \ - float16x4_t __s2_793 = __p2_793; \ - float32x4_t __ret_793; \ - __ret_793 = vfmlslq_high_f16(__s0_793, __s1_793, (float16x8_t) {vget_lane_f16(__s2_793, __p3_793), vget_lane_f16(__s2_793, __p3_793), vget_lane_f16(__s2_793, __p3_793), vget_lane_f16(__s2_793, __p3_793), vget_lane_f16(__s2_793, __p3_793), vget_lane_f16(__s2_793, __p3_793), vget_lane_f16(__s2_793, __p3_793), vget_lane_f16(__s2_793, __p3_793)}); \ - __ret_793; \ +#define vfmlslq_lane_high_f16(__p0_885, __p1_885, __p2_885, __p3_885) __extension__ ({ \ + float32x4_t __s0_885 = __p0_885; \ + float16x8_t __s1_885 = __p1_885; \ + float16x4_t __s2_885 = __p2_885; \ + float32x4_t __ret_885; \ + __ret_885 = vfmlslq_high_f16(__s0_885, __s1_885, (float16x8_t) {vget_lane_f16(__s2_885, __p3_885), vget_lane_f16(__s2_885, __p3_885), vget_lane_f16(__s2_885, __p3_885), vget_lane_f16(__s2_885, __p3_885), vget_lane_f16(__s2_885, __p3_885), vget_lane_f16(__s2_885, __p3_885), vget_lane_f16(__s2_885, __p3_885), vget_lane_f16(__s2_885, __p3_885)}); \ + __ret_885; \ }) #else -#define vfmlslq_lane_high_f16(__p0_794, __p1_794, __p2_794, __p3_794) __extension__ ({ \ - float32x4_t __s0_794 = __p0_794; \ - float16x8_t __s1_794 = __p1_794; \ - float16x4_t __s2_794 = __p2_794; \ - float32x4_t __rev0_794; __rev0_794 = __builtin_shufflevector(__s0_794, __s0_794, 3, 2, 1, 0); \ - float16x8_t __rev1_794; __rev1_794 = __builtin_shufflevector(__s1_794, __s1_794, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __rev2_794; __rev2_794 = __builtin_shufflevector(__s2_794, __s2_794, 3, 2, 1, 0); \ - float32x4_t __ret_794; \ - __ret_794 = __noswap_vfmlslq_high_f16(__rev0_794, __rev1_794, (float16x8_t) {__noswap_vget_lane_f16(__rev2_794, __p3_794), __noswap_vget_lane_f16(__rev2_794, __p3_794), __noswap_vget_lane_f16(__rev2_794, __p3_794), __noswap_vget_lane_f16(__rev2_794, __p3_794), __noswap_vget_lane_f16(__rev2_794, __p3_794), __noswap_vget_lane_f16(__rev2_794, __p3_794), __noswap_vget_lane_f16(__rev2_794, __p3_794), __noswap_vget_lane_f16(__rev2_794, __p3_794)}); \ - __ret_794 = __builtin_shufflevector(__ret_794, __ret_794, 3, 2, 1, 0); \ - __ret_794; \ +#define vfmlslq_lane_high_f16(__p0_886, __p1_886, __p2_886, __p3_886) __extension__ ({ \ + float32x4_t __s0_886 = __p0_886; \ + float16x8_t __s1_886 = __p1_886; \ + float16x4_t __s2_886 = __p2_886; \ + float32x4_t __rev0_886; __rev0_886 = __builtin_shufflevector(__s0_886, __s0_886, 3, 2, 1, 0); \ + float16x8_t __rev1_886; __rev1_886 = __builtin_shufflevector(__s1_886, __s1_886, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_886; __rev2_886 = __builtin_shufflevector(__s2_886, __s2_886, 3, 2, 1, 0); \ + float32x4_t __ret_886; \ + __ret_886 = __noswap_vfmlslq_high_f16(__rev0_886, __rev1_886, (float16x8_t) {__noswap_vget_lane_f16(__rev2_886, __p3_886), __noswap_vget_lane_f16(__rev2_886, __p3_886), __noswap_vget_lane_f16(__rev2_886, __p3_886), __noswap_vget_lane_f16(__rev2_886, __p3_886), __noswap_vget_lane_f16(__rev2_886, __p3_886), __noswap_vget_lane_f16(__rev2_886, __p3_886), __noswap_vget_lane_f16(__rev2_886, __p3_886), __noswap_vget_lane_f16(__rev2_886, __p3_886)}); \ + __ret_886 = __builtin_shufflevector(__ret_886, __ret_886, 3, 2, 1, 0); \ + __ret_886; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlsl_lane_high_f16(__p0_795, __p1_795, __p2_795, __p3_795) __extension__ ({ \ - float32x2_t __s0_795 = __p0_795; \ - float16x4_t __s1_795 = __p1_795; \ - float16x4_t __s2_795 = __p2_795; \ - float32x2_t __ret_795; \ - __ret_795 = vfmlsl_high_f16(__s0_795, __s1_795, (float16x4_t) {vget_lane_f16(__s2_795, __p3_795), vget_lane_f16(__s2_795, __p3_795), vget_lane_f16(__s2_795, __p3_795), vget_lane_f16(__s2_795, __p3_795)}); \ - __ret_795; \ +#define vfmlsl_lane_high_f16(__p0_887, __p1_887, __p2_887, __p3_887) __extension__ ({ \ + float32x2_t __s0_887 = __p0_887; \ + float16x4_t __s1_887 = __p1_887; \ + float16x4_t __s2_887 = __p2_887; \ + float32x2_t __ret_887; \ + __ret_887 = vfmlsl_high_f16(__s0_887, __s1_887, (float16x4_t) {vget_lane_f16(__s2_887, __p3_887), vget_lane_f16(__s2_887, __p3_887), vget_lane_f16(__s2_887, __p3_887), vget_lane_f16(__s2_887, __p3_887)}); \ + __ret_887; \ }) #else -#define vfmlsl_lane_high_f16(__p0_796, __p1_796, __p2_796, __p3_796) __extension__ ({ \ - float32x2_t __s0_796 = __p0_796; \ - float16x4_t __s1_796 = __p1_796; \ - float16x4_t __s2_796 = __p2_796; \ - float32x2_t __rev0_796; __rev0_796 = __builtin_shufflevector(__s0_796, __s0_796, 1, 0); \ - float16x4_t __rev1_796; __rev1_796 = __builtin_shufflevector(__s1_796, __s1_796, 3, 2, 1, 0); \ - float16x4_t __rev2_796; __rev2_796 = __builtin_shufflevector(__s2_796, __s2_796, 3, 2, 1, 0); \ - float32x2_t __ret_796; \ - __ret_796 = __noswap_vfmlsl_high_f16(__rev0_796, __rev1_796, (float16x4_t) {__noswap_vget_lane_f16(__rev2_796, __p3_796), __noswap_vget_lane_f16(__rev2_796, __p3_796), __noswap_vget_lane_f16(__rev2_796, __p3_796), __noswap_vget_lane_f16(__rev2_796, __p3_796)}); \ - __ret_796 = __builtin_shufflevector(__ret_796, __ret_796, 1, 0); \ - __ret_796; \ +#define vfmlsl_lane_high_f16(__p0_888, __p1_888, __p2_888, __p3_888) __extension__ ({ \ + float32x2_t __s0_888 = __p0_888; \ + float16x4_t __s1_888 = __p1_888; \ + float16x4_t __s2_888 = __p2_888; \ + float32x2_t __rev0_888; __rev0_888 = __builtin_shufflevector(__s0_888, __s0_888, 1, 0); \ + float16x4_t __rev1_888; __rev1_888 = __builtin_shufflevector(__s1_888, __s1_888, 3, 2, 1, 0); \ + float16x4_t __rev2_888; __rev2_888 = __builtin_shufflevector(__s2_888, __s2_888, 3, 2, 1, 0); \ + float32x2_t __ret_888; \ + __ret_888 = __noswap_vfmlsl_high_f16(__rev0_888, __rev1_888, (float16x4_t) {__noswap_vget_lane_f16(__rev2_888, __p3_888), __noswap_vget_lane_f16(__rev2_888, __p3_888), __noswap_vget_lane_f16(__rev2_888, __p3_888), __noswap_vget_lane_f16(__rev2_888, __p3_888)}); \ + __ret_888 = __builtin_shufflevector(__ret_888, __ret_888, 1, 0); \ + __ret_888; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlslq_lane_low_f16(__p0_797, __p1_797, __p2_797, __p3_797) __extension__ ({ \ - float32x4_t __s0_797 = __p0_797; \ - float16x8_t __s1_797 = __p1_797; \ - float16x4_t __s2_797 = __p2_797; \ - float32x4_t __ret_797; \ - __ret_797 = vfmlslq_low_f16(__s0_797, __s1_797, (float16x8_t) {vget_lane_f16(__s2_797, __p3_797), vget_lane_f16(__s2_797, __p3_797), vget_lane_f16(__s2_797, __p3_797), vget_lane_f16(__s2_797, __p3_797), vget_lane_f16(__s2_797, __p3_797), vget_lane_f16(__s2_797, __p3_797), vget_lane_f16(__s2_797, __p3_797), vget_lane_f16(__s2_797, __p3_797)}); \ - __ret_797; \ +#define vfmlslq_lane_low_f16(__p0_889, __p1_889, __p2_889, __p3_889) __extension__ ({ \ + float32x4_t __s0_889 = __p0_889; \ + float16x8_t __s1_889 = __p1_889; \ + float16x4_t __s2_889 = __p2_889; \ + float32x4_t __ret_889; \ + __ret_889 = vfmlslq_low_f16(__s0_889, __s1_889, (float16x8_t) {vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889), vget_lane_f16(__s2_889, __p3_889)}); \ + __ret_889; \ }) #else -#define vfmlslq_lane_low_f16(__p0_798, __p1_798, __p2_798, __p3_798) __extension__ ({ \ - float32x4_t __s0_798 = __p0_798; \ - float16x8_t __s1_798 = __p1_798; \ - float16x4_t __s2_798 = __p2_798; \ - float32x4_t __rev0_798; __rev0_798 = __builtin_shufflevector(__s0_798, __s0_798, 3, 2, 1, 0); \ - float16x8_t __rev1_798; __rev1_798 = __builtin_shufflevector(__s1_798, __s1_798, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x4_t __rev2_798; __rev2_798 = __builtin_shufflevector(__s2_798, __s2_798, 3, 2, 1, 0); \ - float32x4_t __ret_798; \ - __ret_798 = __noswap_vfmlslq_low_f16(__rev0_798, __rev1_798, (float16x8_t) {__noswap_vget_lane_f16(__rev2_798, __p3_798), __noswap_vget_lane_f16(__rev2_798, __p3_798), __noswap_vget_lane_f16(__rev2_798, __p3_798), __noswap_vget_lane_f16(__rev2_798, __p3_798), __noswap_vget_lane_f16(__rev2_798, __p3_798), __noswap_vget_lane_f16(__rev2_798, __p3_798), __noswap_vget_lane_f16(__rev2_798, __p3_798), __noswap_vget_lane_f16(__rev2_798, __p3_798)}); \ - __ret_798 = __builtin_shufflevector(__ret_798, __ret_798, 3, 2, 1, 0); \ - __ret_798; \ +#define vfmlslq_lane_low_f16(__p0_890, __p1_890, __p2_890, __p3_890) __extension__ ({ \ + float32x4_t __s0_890 = __p0_890; \ + float16x8_t __s1_890 = __p1_890; \ + float16x4_t __s2_890 = __p2_890; \ + float32x4_t __rev0_890; __rev0_890 = __builtin_shufflevector(__s0_890, __s0_890, 3, 2, 1, 0); \ + float16x8_t __rev1_890; __rev1_890 = __builtin_shufflevector(__s1_890, __s1_890, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x4_t __rev2_890; __rev2_890 = __builtin_shufflevector(__s2_890, __s2_890, 3, 2, 1, 0); \ + float32x4_t __ret_890; \ + __ret_890 = __noswap_vfmlslq_low_f16(__rev0_890, __rev1_890, (float16x8_t) {__noswap_vget_lane_f16(__rev2_890, __p3_890), __noswap_vget_lane_f16(__rev2_890, __p3_890), __noswap_vget_lane_f16(__rev2_890, __p3_890), __noswap_vget_lane_f16(__rev2_890, __p3_890), __noswap_vget_lane_f16(__rev2_890, __p3_890), __noswap_vget_lane_f16(__rev2_890, __p3_890), __noswap_vget_lane_f16(__rev2_890, __p3_890), __noswap_vget_lane_f16(__rev2_890, __p3_890)}); \ + __ret_890 = __builtin_shufflevector(__ret_890, __ret_890, 3, 2, 1, 0); \ + __ret_890; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlsl_lane_low_f16(__p0_799, __p1_799, __p2_799, __p3_799) __extension__ ({ \ - float32x2_t __s0_799 = __p0_799; \ - float16x4_t __s1_799 = __p1_799; \ - float16x4_t __s2_799 = __p2_799; \ - float32x2_t __ret_799; \ - __ret_799 = vfmlsl_low_f16(__s0_799, __s1_799, (float16x4_t) {vget_lane_f16(__s2_799, __p3_799), vget_lane_f16(__s2_799, __p3_799), vget_lane_f16(__s2_799, __p3_799), vget_lane_f16(__s2_799, __p3_799)}); \ - __ret_799; \ +#define vfmlsl_lane_low_f16(__p0_891, __p1_891, __p2_891, __p3_891) __extension__ ({ \ + float32x2_t __s0_891 = __p0_891; \ + float16x4_t __s1_891 = __p1_891; \ + float16x4_t __s2_891 = __p2_891; \ + float32x2_t __ret_891; \ + __ret_891 = vfmlsl_low_f16(__s0_891, __s1_891, (float16x4_t) {vget_lane_f16(__s2_891, __p3_891), vget_lane_f16(__s2_891, __p3_891), vget_lane_f16(__s2_891, __p3_891), vget_lane_f16(__s2_891, __p3_891)}); \ + __ret_891; \ }) #else -#define vfmlsl_lane_low_f16(__p0_800, __p1_800, __p2_800, __p3_800) __extension__ ({ \ - float32x2_t __s0_800 = __p0_800; \ - float16x4_t __s1_800 = __p1_800; \ - float16x4_t __s2_800 = __p2_800; \ - float32x2_t __rev0_800; __rev0_800 = __builtin_shufflevector(__s0_800, __s0_800, 1, 0); \ - float16x4_t __rev1_800; __rev1_800 = __builtin_shufflevector(__s1_800, __s1_800, 3, 2, 1, 0); \ - float16x4_t __rev2_800; __rev2_800 = __builtin_shufflevector(__s2_800, __s2_800, 3, 2, 1, 0); \ - float32x2_t __ret_800; \ - __ret_800 = __noswap_vfmlsl_low_f16(__rev0_800, __rev1_800, (float16x4_t) {__noswap_vget_lane_f16(__rev2_800, __p3_800), __noswap_vget_lane_f16(__rev2_800, __p3_800), __noswap_vget_lane_f16(__rev2_800, __p3_800), __noswap_vget_lane_f16(__rev2_800, __p3_800)}); \ - __ret_800 = __builtin_shufflevector(__ret_800, __ret_800, 1, 0); \ - __ret_800; \ +#define vfmlsl_lane_low_f16(__p0_892, __p1_892, __p2_892, __p3_892) __extension__ ({ \ + float32x2_t __s0_892 = __p0_892; \ + float16x4_t __s1_892 = __p1_892; \ + float16x4_t __s2_892 = __p2_892; \ + float32x2_t __rev0_892; __rev0_892 = __builtin_shufflevector(__s0_892, __s0_892, 1, 0); \ + float16x4_t __rev1_892; __rev1_892 = __builtin_shufflevector(__s1_892, __s1_892, 3, 2, 1, 0); \ + float16x4_t __rev2_892; __rev2_892 = __builtin_shufflevector(__s2_892, __s2_892, 3, 2, 1, 0); \ + float32x2_t __ret_892; \ + __ret_892 = __noswap_vfmlsl_low_f16(__rev0_892, __rev1_892, (float16x4_t) {__noswap_vget_lane_f16(__rev2_892, __p3_892), __noswap_vget_lane_f16(__rev2_892, __p3_892), __noswap_vget_lane_f16(__rev2_892, __p3_892), __noswap_vget_lane_f16(__rev2_892, __p3_892)}); \ + __ret_892 = __builtin_shufflevector(__ret_892, __ret_892, 1, 0); \ + __ret_892; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlslq_laneq_high_f16(__p0_801, __p1_801, __p2_801, __p3_801) __extension__ ({ \ - float32x4_t __s0_801 = __p0_801; \ - float16x8_t __s1_801 = __p1_801; \ - float16x8_t __s2_801 = __p2_801; \ - float32x4_t __ret_801; \ - __ret_801 = vfmlslq_high_f16(__s0_801, __s1_801, (float16x8_t) {vgetq_lane_f16(__s2_801, __p3_801), vgetq_lane_f16(__s2_801, __p3_801), vgetq_lane_f16(__s2_801, __p3_801), vgetq_lane_f16(__s2_801, __p3_801), vgetq_lane_f16(__s2_801, __p3_801), vgetq_lane_f16(__s2_801, __p3_801), vgetq_lane_f16(__s2_801, __p3_801), vgetq_lane_f16(__s2_801, __p3_801)}); \ - __ret_801; \ +#define vfmlslq_laneq_high_f16(__p0_893, __p1_893, __p2_893, __p3_893) __extension__ ({ \ + float32x4_t __s0_893 = __p0_893; \ + float16x8_t __s1_893 = __p1_893; \ + float16x8_t __s2_893 = __p2_893; \ + float32x4_t __ret_893; \ + __ret_893 = vfmlslq_high_f16(__s0_893, __s1_893, (float16x8_t) {vgetq_lane_f16(__s2_893, __p3_893), vgetq_lane_f16(__s2_893, __p3_893), vgetq_lane_f16(__s2_893, __p3_893), vgetq_lane_f16(__s2_893, __p3_893), vgetq_lane_f16(__s2_893, __p3_893), vgetq_lane_f16(__s2_893, __p3_893), vgetq_lane_f16(__s2_893, __p3_893), vgetq_lane_f16(__s2_893, __p3_893)}); \ + __ret_893; \ }) #else -#define vfmlslq_laneq_high_f16(__p0_802, __p1_802, __p2_802, __p3_802) __extension__ ({ \ - float32x4_t __s0_802 = __p0_802; \ - float16x8_t __s1_802 = __p1_802; \ - float16x8_t __s2_802 = __p2_802; \ - float32x4_t __rev0_802; __rev0_802 = __builtin_shufflevector(__s0_802, __s0_802, 3, 2, 1, 0); \ - float16x8_t __rev1_802; __rev1_802 = __builtin_shufflevector(__s1_802, __s1_802, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev2_802; __rev2_802 = __builtin_shufflevector(__s2_802, __s2_802, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x4_t __ret_802; \ - __ret_802 = __noswap_vfmlslq_high_f16(__rev0_802, __rev1_802, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_802, __p3_802), __noswap_vgetq_lane_f16(__rev2_802, __p3_802), __noswap_vgetq_lane_f16(__rev2_802, __p3_802), __noswap_vgetq_lane_f16(__rev2_802, __p3_802), __noswap_vgetq_lane_f16(__rev2_802, __p3_802), __noswap_vgetq_lane_f16(__rev2_802, __p3_802), __noswap_vgetq_lane_f16(__rev2_802, __p3_802), __noswap_vgetq_lane_f16(__rev2_802, __p3_802)}); \ - __ret_802 = __builtin_shufflevector(__ret_802, __ret_802, 3, 2, 1, 0); \ - __ret_802; \ +#define vfmlslq_laneq_high_f16(__p0_894, __p1_894, __p2_894, __p3_894) __extension__ ({ \ + float32x4_t __s0_894 = __p0_894; \ + float16x8_t __s1_894 = __p1_894; \ + float16x8_t __s2_894 = __p2_894; \ + float32x4_t __rev0_894; __rev0_894 = __builtin_shufflevector(__s0_894, __s0_894, 3, 2, 1, 0); \ + float16x8_t __rev1_894; __rev1_894 = __builtin_shufflevector(__s1_894, __s1_894, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_894; __rev2_894 = __builtin_shufflevector(__s2_894, __s2_894, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x4_t __ret_894; \ + __ret_894 = __noswap_vfmlslq_high_f16(__rev0_894, __rev1_894, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_894, __p3_894), __noswap_vgetq_lane_f16(__rev2_894, __p3_894), __noswap_vgetq_lane_f16(__rev2_894, __p3_894), __noswap_vgetq_lane_f16(__rev2_894, __p3_894), __noswap_vgetq_lane_f16(__rev2_894, __p3_894), __noswap_vgetq_lane_f16(__rev2_894, __p3_894), __noswap_vgetq_lane_f16(__rev2_894, __p3_894), __noswap_vgetq_lane_f16(__rev2_894, __p3_894)}); \ + __ret_894 = __builtin_shufflevector(__ret_894, __ret_894, 3, 2, 1, 0); \ + __ret_894; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlsl_laneq_high_f16(__p0_803, __p1_803, __p2_803, __p3_803) __extension__ ({ \ - float32x2_t __s0_803 = __p0_803; \ - float16x4_t __s1_803 = __p1_803; \ - float16x8_t __s2_803 = __p2_803; \ - float32x2_t __ret_803; \ - __ret_803 = vfmlsl_high_f16(__s0_803, __s1_803, (float16x4_t) {vgetq_lane_f16(__s2_803, __p3_803), vgetq_lane_f16(__s2_803, __p3_803), vgetq_lane_f16(__s2_803, __p3_803), vgetq_lane_f16(__s2_803, __p3_803)}); \ - __ret_803; \ +#define vfmlsl_laneq_high_f16(__p0_895, __p1_895, __p2_895, __p3_895) __extension__ ({ \ + float32x2_t __s0_895 = __p0_895; \ + float16x4_t __s1_895 = __p1_895; \ + float16x8_t __s2_895 = __p2_895; \ + float32x2_t __ret_895; \ + __ret_895 = vfmlsl_high_f16(__s0_895, __s1_895, (float16x4_t) {vgetq_lane_f16(__s2_895, __p3_895), vgetq_lane_f16(__s2_895, __p3_895), vgetq_lane_f16(__s2_895, __p3_895), vgetq_lane_f16(__s2_895, __p3_895)}); \ + __ret_895; \ }) #else -#define vfmlsl_laneq_high_f16(__p0_804, __p1_804, __p2_804, __p3_804) __extension__ ({ \ - float32x2_t __s0_804 = __p0_804; \ - float16x4_t __s1_804 = __p1_804; \ - float16x8_t __s2_804 = __p2_804; \ - float32x2_t __rev0_804; __rev0_804 = __builtin_shufflevector(__s0_804, __s0_804, 1, 0); \ - float16x4_t __rev1_804; __rev1_804 = __builtin_shufflevector(__s1_804, __s1_804, 3, 2, 1, 0); \ - float16x8_t __rev2_804; __rev2_804 = __builtin_shufflevector(__s2_804, __s2_804, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x2_t __ret_804; \ - __ret_804 = __noswap_vfmlsl_high_f16(__rev0_804, __rev1_804, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_804, __p3_804), __noswap_vgetq_lane_f16(__rev2_804, __p3_804), __noswap_vgetq_lane_f16(__rev2_804, __p3_804), __noswap_vgetq_lane_f16(__rev2_804, __p3_804)}); \ - __ret_804 = __builtin_shufflevector(__ret_804, __ret_804, 1, 0); \ - __ret_804; \ +#define vfmlsl_laneq_high_f16(__p0_896, __p1_896, __p2_896, __p3_896) __extension__ ({ \ + float32x2_t __s0_896 = __p0_896; \ + float16x4_t __s1_896 = __p1_896; \ + float16x8_t __s2_896 = __p2_896; \ + float32x2_t __rev0_896; __rev0_896 = __builtin_shufflevector(__s0_896, __s0_896, 1, 0); \ + float16x4_t __rev1_896; __rev1_896 = __builtin_shufflevector(__s1_896, __s1_896, 3, 2, 1, 0); \ + float16x8_t __rev2_896; __rev2_896 = __builtin_shufflevector(__s2_896, __s2_896, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x2_t __ret_896; \ + __ret_896 = __noswap_vfmlsl_high_f16(__rev0_896, __rev1_896, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_896, __p3_896), __noswap_vgetq_lane_f16(__rev2_896, __p3_896), __noswap_vgetq_lane_f16(__rev2_896, __p3_896), __noswap_vgetq_lane_f16(__rev2_896, __p3_896)}); \ + __ret_896 = __builtin_shufflevector(__ret_896, __ret_896, 1, 0); \ + __ret_896; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlslq_laneq_low_f16(__p0_805, __p1_805, __p2_805, __p3_805) __extension__ ({ \ - float32x4_t __s0_805 = __p0_805; \ - float16x8_t __s1_805 = __p1_805; \ - float16x8_t __s2_805 = __p2_805; \ - float32x4_t __ret_805; \ - __ret_805 = vfmlslq_low_f16(__s0_805, __s1_805, (float16x8_t) {vgetq_lane_f16(__s2_805, __p3_805), vgetq_lane_f16(__s2_805, __p3_805), vgetq_lane_f16(__s2_805, __p3_805), vgetq_lane_f16(__s2_805, __p3_805), vgetq_lane_f16(__s2_805, __p3_805), vgetq_lane_f16(__s2_805, __p3_805), vgetq_lane_f16(__s2_805, __p3_805), vgetq_lane_f16(__s2_805, __p3_805)}); \ - __ret_805; \ +#define vfmlslq_laneq_low_f16(__p0_897, __p1_897, __p2_897, __p3_897) __extension__ ({ \ + float32x4_t __s0_897 = __p0_897; \ + float16x8_t __s1_897 = __p1_897; \ + float16x8_t __s2_897 = __p2_897; \ + float32x4_t __ret_897; \ + __ret_897 = vfmlslq_low_f16(__s0_897, __s1_897, (float16x8_t) {vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897), vgetq_lane_f16(__s2_897, __p3_897)}); \ + __ret_897; \ }) #else -#define vfmlslq_laneq_low_f16(__p0_806, __p1_806, __p2_806, __p3_806) __extension__ ({ \ - float32x4_t __s0_806 = __p0_806; \ - float16x8_t __s1_806 = __p1_806; \ - float16x8_t __s2_806 = __p2_806; \ - float32x4_t __rev0_806; __rev0_806 = __builtin_shufflevector(__s0_806, __s0_806, 3, 2, 1, 0); \ - float16x8_t __rev1_806; __rev1_806 = __builtin_shufflevector(__s1_806, __s1_806, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16x8_t __rev2_806; __rev2_806 = __builtin_shufflevector(__s2_806, __s2_806, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x4_t __ret_806; \ - __ret_806 = __noswap_vfmlslq_low_f16(__rev0_806, __rev1_806, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_806, __p3_806), __noswap_vgetq_lane_f16(__rev2_806, __p3_806), __noswap_vgetq_lane_f16(__rev2_806, __p3_806), __noswap_vgetq_lane_f16(__rev2_806, __p3_806), __noswap_vgetq_lane_f16(__rev2_806, __p3_806), __noswap_vgetq_lane_f16(__rev2_806, __p3_806), __noswap_vgetq_lane_f16(__rev2_806, __p3_806), __noswap_vgetq_lane_f16(__rev2_806, __p3_806)}); \ - __ret_806 = __builtin_shufflevector(__ret_806, __ret_806, 3, 2, 1, 0); \ - __ret_806; \ +#define vfmlslq_laneq_low_f16(__p0_898, __p1_898, __p2_898, __p3_898) __extension__ ({ \ + float32x4_t __s0_898 = __p0_898; \ + float16x8_t __s1_898 = __p1_898; \ + float16x8_t __s2_898 = __p2_898; \ + float32x4_t __rev0_898; __rev0_898 = __builtin_shufflevector(__s0_898, __s0_898, 3, 2, 1, 0); \ + float16x8_t __rev1_898; __rev1_898 = __builtin_shufflevector(__s1_898, __s1_898, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16x8_t __rev2_898; __rev2_898 = __builtin_shufflevector(__s2_898, __s2_898, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x4_t __ret_898; \ + __ret_898 = __noswap_vfmlslq_low_f16(__rev0_898, __rev1_898, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_898, __p3_898), __noswap_vgetq_lane_f16(__rev2_898, __p3_898), __noswap_vgetq_lane_f16(__rev2_898, __p3_898), __noswap_vgetq_lane_f16(__rev2_898, __p3_898), __noswap_vgetq_lane_f16(__rev2_898, __p3_898), __noswap_vgetq_lane_f16(__rev2_898, __p3_898), __noswap_vgetq_lane_f16(__rev2_898, __p3_898), __noswap_vgetq_lane_f16(__rev2_898, __p3_898)}); \ + __ret_898 = __builtin_shufflevector(__ret_898, __ret_898, 3, 2, 1, 0); \ + __ret_898; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vfmlsl_laneq_low_f16(__p0_807, __p1_807, __p2_807, __p3_807) __extension__ ({ \ - float32x2_t __s0_807 = __p0_807; \ - float16x4_t __s1_807 = __p1_807; \ - float16x8_t __s2_807 = __p2_807; \ - float32x2_t __ret_807; \ - __ret_807 = vfmlsl_low_f16(__s0_807, __s1_807, (float16x4_t) {vgetq_lane_f16(__s2_807, __p3_807), vgetq_lane_f16(__s2_807, __p3_807), vgetq_lane_f16(__s2_807, __p3_807), vgetq_lane_f16(__s2_807, __p3_807)}); \ - __ret_807; \ +#define vfmlsl_laneq_low_f16(__p0_899, __p1_899, __p2_899, __p3_899) __extension__ ({ \ + float32x2_t __s0_899 = __p0_899; \ + float16x4_t __s1_899 = __p1_899; \ + float16x8_t __s2_899 = __p2_899; \ + float32x2_t __ret_899; \ + __ret_899 = vfmlsl_low_f16(__s0_899, __s1_899, (float16x4_t) {vgetq_lane_f16(__s2_899, __p3_899), vgetq_lane_f16(__s2_899, __p3_899), vgetq_lane_f16(__s2_899, __p3_899), vgetq_lane_f16(__s2_899, __p3_899)}); \ + __ret_899; \ }) #else -#define vfmlsl_laneq_low_f16(__p0_808, __p1_808, __p2_808, __p3_808) __extension__ ({ \ - float32x2_t __s0_808 = __p0_808; \ - float16x4_t __s1_808 = __p1_808; \ - float16x8_t __s2_808 = __p2_808; \ - float32x2_t __rev0_808; __rev0_808 = __builtin_shufflevector(__s0_808, __s0_808, 1, 0); \ - float16x4_t __rev1_808; __rev1_808 = __builtin_shufflevector(__s1_808, __s1_808, 3, 2, 1, 0); \ - float16x8_t __rev2_808; __rev2_808 = __builtin_shufflevector(__s2_808, __s2_808, 7, 6, 5, 4, 3, 2, 1, 0); \ - float32x2_t __ret_808; \ - __ret_808 = __noswap_vfmlsl_low_f16(__rev0_808, __rev1_808, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_808, __p3_808), __noswap_vgetq_lane_f16(__rev2_808, __p3_808), __noswap_vgetq_lane_f16(__rev2_808, __p3_808), __noswap_vgetq_lane_f16(__rev2_808, __p3_808)}); \ - __ret_808 = __builtin_shufflevector(__ret_808, __ret_808, 1, 0); \ - __ret_808; \ +#define vfmlsl_laneq_low_f16(__p0_900, __p1_900, __p2_900, __p3_900) __extension__ ({ \ + float32x2_t __s0_900 = __p0_900; \ + float16x4_t __s1_900 = __p1_900; \ + float16x8_t __s2_900 = __p2_900; \ + float32x2_t __rev0_900; __rev0_900 = __builtin_shufflevector(__s0_900, __s0_900, 1, 0); \ + float16x4_t __rev1_900; __rev1_900 = __builtin_shufflevector(__s1_900, __s1_900, 3, 2, 1, 0); \ + float16x8_t __rev2_900; __rev2_900 = __builtin_shufflevector(__s2_900, __s2_900, 7, 6, 5, 4, 3, 2, 1, 0); \ + float32x2_t __ret_900; \ + __ret_900 = __noswap_vfmlsl_low_f16(__rev0_900, __rev1_900, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_900, __p3_900), __noswap_vgetq_lane_f16(__rev2_900, __p3_900), __noswap_vgetq_lane_f16(__rev2_900, __p3_900), __noswap_vgetq_lane_f16(__rev2_900, __p3_900)}); \ + __ret_900 = __builtin_shufflevector(__ret_900, __ret_900, 1, 0); \ + __ret_900; \ }) #endif #endif #if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && defined(__aarch64__) #ifdef __LITTLE_ENDIAN__ -#define vmulh_lane_f16(__p0_809, __p1_809, __p2_809) __extension__ ({ \ - float16_t __s0_809 = __p0_809; \ - float16x4_t __s1_809 = __p1_809; \ - float16_t __ret_809; \ - __ret_809 = __s0_809 * vget_lane_f16(__s1_809, __p2_809); \ - __ret_809; \ +#define vmulh_lane_f16(__p0_901, __p1_901, __p2_901) __extension__ ({ \ + float16_t __s0_901 = __p0_901; \ + float16x4_t __s1_901 = __p1_901; \ + float16_t __ret_901; \ + __ret_901 = __s0_901 * vget_lane_f16(__s1_901, __p2_901); \ + __ret_901; \ }) #else -#define vmulh_lane_f16(__p0_810, __p1_810, __p2_810) __extension__ ({ \ - float16_t __s0_810 = __p0_810; \ - float16x4_t __s1_810 = __p1_810; \ - float16x4_t __rev1_810; __rev1_810 = __builtin_shufflevector(__s1_810, __s1_810, 3, 2, 1, 0); \ - float16_t __ret_810; \ - __ret_810 = __s0_810 * __noswap_vget_lane_f16(__rev1_810, __p2_810); \ - __ret_810; \ +#define vmulh_lane_f16(__p0_902, __p1_902, __p2_902) __extension__ ({ \ + float16_t __s0_902 = __p0_902; \ + float16x4_t __s1_902 = __p1_902; \ + float16x4_t __rev1_902; __rev1_902 = __builtin_shufflevector(__s1_902, __s1_902, 3, 2, 1, 0); \ + float16_t __ret_902; \ + __ret_902 = __s0_902 * __noswap_vget_lane_f16(__rev1_902, __p2_902); \ + __ret_902; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vmulh_laneq_f16(__p0_811, __p1_811, __p2_811) __extension__ ({ \ - float16_t __s0_811 = __p0_811; \ - float16x8_t __s1_811 = __p1_811; \ - float16_t __ret_811; \ - __ret_811 = __s0_811 * vgetq_lane_f16(__s1_811, __p2_811); \ - __ret_811; \ +#define vmulh_laneq_f16(__p0_903, __p1_903, __p2_903) __extension__ ({ \ + float16_t __s0_903 = __p0_903; \ + float16x8_t __s1_903 = __p1_903; \ + float16_t __ret_903; \ + __ret_903 = __s0_903 * vgetq_lane_f16(__s1_903, __p2_903); \ + __ret_903; \ }) #else -#define vmulh_laneq_f16(__p0_812, __p1_812, __p2_812) __extension__ ({ \ - float16_t __s0_812 = __p0_812; \ - float16x8_t __s1_812 = __p1_812; \ - float16x8_t __rev1_812; __rev1_812 = __builtin_shufflevector(__s1_812, __s1_812, 7, 6, 5, 4, 3, 2, 1, 0); \ - float16_t __ret_812; \ - __ret_812 = __s0_812 * __noswap_vgetq_lane_f16(__rev1_812, __p2_812); \ - __ret_812; \ +#define vmulh_laneq_f16(__p0_904, __p1_904, __p2_904) __extension__ ({ \ + float16_t __s0_904 = __p0_904; \ + float16x8_t __s1_904 = __p1_904; \ + float16x8_t __rev1_904; __rev1_904 = __builtin_shufflevector(__s1_904, __s1_904, 7, 6, 5, 4, 3, 2, 1, 0); \ + float16_t __ret_904; \ + __ret_904 = __s0_904 * __noswap_vgetq_lane_f16(__rev1_904, __p2_904); \ + __ret_904; \ }) #endif #endif #if defined(__ARM_FEATURE_MATMUL_INT8) #ifdef __LITTLE_ENDIAN__ -#define vsudotq_lane_s32(__p0_813, __p1_813, __p2_813, __p3_813) __extension__ ({ \ - int32x4_t __s0_813 = __p0_813; \ - int8x16_t __s1_813 = __p1_813; \ - uint8x8_t __s2_813 = __p2_813; \ - int32x4_t __ret_813; \ -uint8x8_t __reint_813 = __s2_813; \ - __ret_813 = vusdotq_s32(__s0_813, (uint8x16_t)(splatq_lane_s32(*(int32x2_t *) &__reint_813, __p3_813)), __s1_813); \ - __ret_813; \ +#define vsudotq_lane_s32(__p0_905, __p1_905, __p2_905, __p3_905) __extension__ ({ \ + int32x4_t __s0_905 = __p0_905; \ + int8x16_t __s1_905 = __p1_905; \ + uint8x8_t __s2_905 = __p2_905; \ + int32x4_t __ret_905; \ +uint8x8_t __reint_905 = __s2_905; \ + __ret_905 = vusdotq_s32(__s0_905, (uint8x16_t)(splatq_lane_s32(*(int32x2_t *) &__reint_905, __p3_905)), __s1_905); \ + __ret_905; \ }) #else -#define vsudotq_lane_s32(__p0_814, __p1_814, __p2_814, __p3_814) __extension__ ({ \ - int32x4_t __s0_814 = __p0_814; \ - int8x16_t __s1_814 = __p1_814; \ - uint8x8_t __s2_814 = __p2_814; \ - int32x4_t __rev0_814; __rev0_814 = __builtin_shufflevector(__s0_814, __s0_814, 3, 2, 1, 0); \ - int8x16_t __rev1_814; __rev1_814 = __builtin_shufflevector(__s1_814, __s1_814, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __rev2_814; __rev2_814 = __builtin_shufflevector(__s2_814, __s2_814, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x4_t __ret_814; \ -uint8x8_t __reint_814 = __rev2_814; \ - __ret_814 = __noswap_vusdotq_s32(__rev0_814, (uint8x16_t)(__noswap_splatq_lane_s32(*(int32x2_t *) &__reint_814, __p3_814)), __rev1_814); \ - __ret_814 = __builtin_shufflevector(__ret_814, __ret_814, 3, 2, 1, 0); \ - __ret_814; \ +#define vsudotq_lane_s32(__p0_906, __p1_906, __p2_906, __p3_906) __extension__ ({ \ + int32x4_t __s0_906 = __p0_906; \ + int8x16_t __s1_906 = __p1_906; \ + uint8x8_t __s2_906 = __p2_906; \ + int32x4_t __rev0_906; __rev0_906 = __builtin_shufflevector(__s0_906, __s0_906, 3, 2, 1, 0); \ + int8x16_t __rev1_906; __rev1_906 = __builtin_shufflevector(__s1_906, __s1_906, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __rev2_906; __rev2_906 = __builtin_shufflevector(__s2_906, __s2_906, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x4_t __ret_906; \ +uint8x8_t __reint_906 = __rev2_906; \ + __ret_906 = __noswap_vusdotq_s32(__rev0_906, (uint8x16_t)(__noswap_splatq_lane_s32(*(int32x2_t *) &__reint_906, __p3_906)), __rev1_906); \ + __ret_906 = __builtin_shufflevector(__ret_906, __ret_906, 3, 2, 1, 0); \ + __ret_906; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vsudot_lane_s32(__p0_815, __p1_815, __p2_815, __p3_815) __extension__ ({ \ - int32x2_t __s0_815 = __p0_815; \ - int8x8_t __s1_815 = __p1_815; \ - uint8x8_t __s2_815 = __p2_815; \ - int32x2_t __ret_815; \ -uint8x8_t __reint_815 = __s2_815; \ - __ret_815 = vusdot_s32(__s0_815, (uint8x8_t)(splat_lane_s32(*(int32x2_t *) &__reint_815, __p3_815)), __s1_815); \ - __ret_815; \ +#define vsudot_lane_s32(__p0_907, __p1_907, __p2_907, __p3_907) __extension__ ({ \ + int32x2_t __s0_907 = __p0_907; \ + int8x8_t __s1_907 = __p1_907; \ + uint8x8_t __s2_907 = __p2_907; \ + int32x2_t __ret_907; \ +uint8x8_t __reint_907 = __s2_907; \ + __ret_907 = vusdot_s32(__s0_907, (uint8x8_t)(splat_lane_s32(*(int32x2_t *) &__reint_907, __p3_907)), __s1_907); \ + __ret_907; \ }) #else -#define vsudot_lane_s32(__p0_816, __p1_816, __p2_816, __p3_816) __extension__ ({ \ - int32x2_t __s0_816 = __p0_816; \ - int8x8_t __s1_816 = __p1_816; \ - uint8x8_t __s2_816 = __p2_816; \ - int32x2_t __rev0_816; __rev0_816 = __builtin_shufflevector(__s0_816, __s0_816, 1, 0); \ - int8x8_t __rev1_816; __rev1_816 = __builtin_shufflevector(__s1_816, __s1_816, 7, 6, 5, 4, 3, 2, 1, 0); \ - uint8x8_t __rev2_816; __rev2_816 = __builtin_shufflevector(__s2_816, __s2_816, 7, 6, 5, 4, 3, 2, 1, 0); \ - int32x2_t __ret_816; \ -uint8x8_t __reint_816 = __rev2_816; \ - __ret_816 = __noswap_vusdot_s32(__rev0_816, (uint8x8_t)(__noswap_splat_lane_s32(*(int32x2_t *) &__reint_816, __p3_816)), __rev1_816); \ - __ret_816 = __builtin_shufflevector(__ret_816, __ret_816, 1, 0); \ - __ret_816; \ +#define vsudot_lane_s32(__p0_908, __p1_908, __p2_908, __p3_908) __extension__ ({ \ + int32x2_t __s0_908 = __p0_908; \ + int8x8_t __s1_908 = __p1_908; \ + uint8x8_t __s2_908 = __p2_908; \ + int32x2_t __rev0_908; __rev0_908 = __builtin_shufflevector(__s0_908, __s0_908, 1, 0); \ + int8x8_t __rev1_908; __rev1_908 = __builtin_shufflevector(__s1_908, __s1_908, 7, 6, 5, 4, 3, 2, 1, 0); \ + uint8x8_t __rev2_908; __rev2_908 = __builtin_shufflevector(__s2_908, __s2_908, 7, 6, 5, 4, 3, 2, 1, 0); \ + int32x2_t __ret_908; \ +uint8x8_t __reint_908 = __rev2_908; \ + __ret_908 = __noswap_vusdot_s32(__rev0_908, (uint8x8_t)(__noswap_splat_lane_s32(*(int32x2_t *) &__reint_908, __p3_908)), __rev1_908); \ + __ret_908 = __builtin_shufflevector(__ret_908, __ret_908, 1, 0); \ + __ret_908; \ }) #endif @@ -66049,86 +67719,86 @@ __ai int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) { return __ret; } #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahs_lane_s32(__p0_817, __p1_817, __p2_817, __p3_817) __extension__ ({ \ - int32_t __s0_817 = __p0_817; \ - int32_t __s1_817 = __p1_817; \ - int32x2_t __s2_817 = __p2_817; \ - int32_t __ret_817; \ - __ret_817 = vqadds_s32(__s0_817, vqrdmulhs_s32(__s1_817, vget_lane_s32(__s2_817, __p3_817))); \ - __ret_817; \ +#define vqrdmlahs_lane_s32(__p0_909, __p1_909, __p2_909, __p3_909) __extension__ ({ \ + int32_t __s0_909 = __p0_909; \ + int32_t __s1_909 = __p1_909; \ + int32x2_t __s2_909 = __p2_909; \ + int32_t __ret_909; \ + __ret_909 = vqadds_s32(__s0_909, vqrdmulhs_s32(__s1_909, vget_lane_s32(__s2_909, __p3_909))); \ + __ret_909; \ }) #else -#define vqrdmlahs_lane_s32(__p0_818, __p1_818, __p2_818, __p3_818) __extension__ ({ \ - int32_t __s0_818 = __p0_818; \ - int32_t __s1_818 = __p1_818; \ - int32x2_t __s2_818 = __p2_818; \ - int32x2_t __rev2_818; __rev2_818 = __builtin_shufflevector(__s2_818, __s2_818, 1, 0); \ - int32_t __ret_818; \ - __ret_818 = vqadds_s32(__s0_818, vqrdmulhs_s32(__s1_818, __noswap_vget_lane_s32(__rev2_818, __p3_818))); \ - __ret_818; \ +#define vqrdmlahs_lane_s32(__p0_910, __p1_910, __p2_910, __p3_910) __extension__ ({ \ + int32_t __s0_910 = __p0_910; \ + int32_t __s1_910 = __p1_910; \ + int32x2_t __s2_910 = __p2_910; \ + int32x2_t __rev2_910; __rev2_910 = __builtin_shufflevector(__s2_910, __s2_910, 1, 0); \ + int32_t __ret_910; \ + __ret_910 = vqadds_s32(__s0_910, vqrdmulhs_s32(__s1_910, __noswap_vget_lane_s32(__rev2_910, __p3_910))); \ + __ret_910; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahh_lane_s16(__p0_819, __p1_819, __p2_819, __p3_819) __extension__ ({ \ - int16_t __s0_819 = __p0_819; \ - int16_t __s1_819 = __p1_819; \ - int16x4_t __s2_819 = __p2_819; \ - int16_t __ret_819; \ - __ret_819 = vqaddh_s16(__s0_819, vqrdmulhh_s16(__s1_819, vget_lane_s16(__s2_819, __p3_819))); \ - __ret_819; \ +#define vqrdmlahh_lane_s16(__p0_911, __p1_911, __p2_911, __p3_911) __extension__ ({ \ + int16_t __s0_911 = __p0_911; \ + int16_t __s1_911 = __p1_911; \ + int16x4_t __s2_911 = __p2_911; \ + int16_t __ret_911; \ + __ret_911 = vqaddh_s16(__s0_911, vqrdmulhh_s16(__s1_911, vget_lane_s16(__s2_911, __p3_911))); \ + __ret_911; \ }) #else -#define vqrdmlahh_lane_s16(__p0_820, __p1_820, __p2_820, __p3_820) __extension__ ({ \ - int16_t __s0_820 = __p0_820; \ - int16_t __s1_820 = __p1_820; \ - int16x4_t __s2_820 = __p2_820; \ - int16x4_t __rev2_820; __rev2_820 = __builtin_shufflevector(__s2_820, __s2_820, 3, 2, 1, 0); \ - int16_t __ret_820; \ - __ret_820 = vqaddh_s16(__s0_820, vqrdmulhh_s16(__s1_820, __noswap_vget_lane_s16(__rev2_820, __p3_820))); \ - __ret_820; \ +#define vqrdmlahh_lane_s16(__p0_912, __p1_912, __p2_912, __p3_912) __extension__ ({ \ + int16_t __s0_912 = __p0_912; \ + int16_t __s1_912 = __p1_912; \ + int16x4_t __s2_912 = __p2_912; \ + int16x4_t __rev2_912; __rev2_912 = __builtin_shufflevector(__s2_912, __s2_912, 3, 2, 1, 0); \ + int16_t __ret_912; \ + __ret_912 = vqaddh_s16(__s0_912, vqrdmulhh_s16(__s1_912, __noswap_vget_lane_s16(__rev2_912, __p3_912))); \ + __ret_912; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahs_laneq_s32(__p0_821, __p1_821, __p2_821, __p3_821) __extension__ ({ \ - int32_t __s0_821 = __p0_821; \ - int32_t __s1_821 = __p1_821; \ - int32x4_t __s2_821 = __p2_821; \ - int32_t __ret_821; \ - __ret_821 = vqadds_s32(__s0_821, vqrdmulhs_s32(__s1_821, vgetq_lane_s32(__s2_821, __p3_821))); \ - __ret_821; \ +#define vqrdmlahs_laneq_s32(__p0_913, __p1_913, __p2_913, __p3_913) __extension__ ({ \ + int32_t __s0_913 = __p0_913; \ + int32_t __s1_913 = __p1_913; \ + int32x4_t __s2_913 = __p2_913; \ + int32_t __ret_913; \ + __ret_913 = vqadds_s32(__s0_913, vqrdmulhs_s32(__s1_913, vgetq_lane_s32(__s2_913, __p3_913))); \ + __ret_913; \ }) #else -#define vqrdmlahs_laneq_s32(__p0_822, __p1_822, __p2_822, __p3_822) __extension__ ({ \ - int32_t __s0_822 = __p0_822; \ - int32_t __s1_822 = __p1_822; \ - int32x4_t __s2_822 = __p2_822; \ - int32x4_t __rev2_822; __rev2_822 = __builtin_shufflevector(__s2_822, __s2_822, 3, 2, 1, 0); \ - int32_t __ret_822; \ - __ret_822 = vqadds_s32(__s0_822, vqrdmulhs_s32(__s1_822, __noswap_vgetq_lane_s32(__rev2_822, __p3_822))); \ - __ret_822; \ +#define vqrdmlahs_laneq_s32(__p0_914, __p1_914, __p2_914, __p3_914) __extension__ ({ \ + int32_t __s0_914 = __p0_914; \ + int32_t __s1_914 = __p1_914; \ + int32x4_t __s2_914 = __p2_914; \ + int32x4_t __rev2_914; __rev2_914 = __builtin_shufflevector(__s2_914, __s2_914, 3, 2, 1, 0); \ + int32_t __ret_914; \ + __ret_914 = vqadds_s32(__s0_914, vqrdmulhs_s32(__s1_914, __noswap_vgetq_lane_s32(__rev2_914, __p3_914))); \ + __ret_914; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlahh_laneq_s16(__p0_823, __p1_823, __p2_823, __p3_823) __extension__ ({ \ - int16_t __s0_823 = __p0_823; \ - int16_t __s1_823 = __p1_823; \ - int16x8_t __s2_823 = __p2_823; \ - int16_t __ret_823; \ - __ret_823 = vqaddh_s16(__s0_823, vqrdmulhh_s16(__s1_823, vgetq_lane_s16(__s2_823, __p3_823))); \ - __ret_823; \ +#define vqrdmlahh_laneq_s16(__p0_915, __p1_915, __p2_915, __p3_915) __extension__ ({ \ + int16_t __s0_915 = __p0_915; \ + int16_t __s1_915 = __p1_915; \ + int16x8_t __s2_915 = __p2_915; \ + int16_t __ret_915; \ + __ret_915 = vqaddh_s16(__s0_915, vqrdmulhh_s16(__s1_915, vgetq_lane_s16(__s2_915, __p3_915))); \ + __ret_915; \ }) #else -#define vqrdmlahh_laneq_s16(__p0_824, __p1_824, __p2_824, __p3_824) __extension__ ({ \ - int16_t __s0_824 = __p0_824; \ - int16_t __s1_824 = __p1_824; \ - int16x8_t __s2_824 = __p2_824; \ - int16x8_t __rev2_824; __rev2_824 = __builtin_shufflevector(__s2_824, __s2_824, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16_t __ret_824; \ - __ret_824 = vqaddh_s16(__s0_824, vqrdmulhh_s16(__s1_824, __noswap_vgetq_lane_s16(__rev2_824, __p3_824))); \ - __ret_824; \ +#define vqrdmlahh_laneq_s16(__p0_916, __p1_916, __p2_916, __p3_916) __extension__ ({ \ + int16_t __s0_916 = __p0_916; \ + int16_t __s1_916 = __p1_916; \ + int16x8_t __s2_916 = __p2_916; \ + int16x8_t __rev2_916; __rev2_916 = __builtin_shufflevector(__s2_916, __s2_916, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16_t __ret_916; \ + __ret_916 = vqaddh_s16(__s0_916, vqrdmulhh_s16(__s1_916, __noswap_vgetq_lane_s16(__rev2_916, __p3_916))); \ + __ret_916; \ }) #endif @@ -66143,86 +67813,86 @@ __ai int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) { return __ret; } #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshs_lane_s32(__p0_825, __p1_825, __p2_825, __p3_825) __extension__ ({ \ - int32_t __s0_825 = __p0_825; \ - int32_t __s1_825 = __p1_825; \ - int32x2_t __s2_825 = __p2_825; \ - int32_t __ret_825; \ - __ret_825 = vqsubs_s32(__s0_825, vqrdmulhs_s32(__s1_825, vget_lane_s32(__s2_825, __p3_825))); \ - __ret_825; \ +#define vqrdmlshs_lane_s32(__p0_917, __p1_917, __p2_917, __p3_917) __extension__ ({ \ + int32_t __s0_917 = __p0_917; \ + int32_t __s1_917 = __p1_917; \ + int32x2_t __s2_917 = __p2_917; \ + int32_t __ret_917; \ + __ret_917 = vqsubs_s32(__s0_917, vqrdmulhs_s32(__s1_917, vget_lane_s32(__s2_917, __p3_917))); \ + __ret_917; \ }) #else -#define vqrdmlshs_lane_s32(__p0_826, __p1_826, __p2_826, __p3_826) __extension__ ({ \ - int32_t __s0_826 = __p0_826; \ - int32_t __s1_826 = __p1_826; \ - int32x2_t __s2_826 = __p2_826; \ - int32x2_t __rev2_826; __rev2_826 = __builtin_shufflevector(__s2_826, __s2_826, 1, 0); \ - int32_t __ret_826; \ - __ret_826 = vqsubs_s32(__s0_826, vqrdmulhs_s32(__s1_826, __noswap_vget_lane_s32(__rev2_826, __p3_826))); \ - __ret_826; \ +#define vqrdmlshs_lane_s32(__p0_918, __p1_918, __p2_918, __p3_918) __extension__ ({ \ + int32_t __s0_918 = __p0_918; \ + int32_t __s1_918 = __p1_918; \ + int32x2_t __s2_918 = __p2_918; \ + int32x2_t __rev2_918; __rev2_918 = __builtin_shufflevector(__s2_918, __s2_918, 1, 0); \ + int32_t __ret_918; \ + __ret_918 = vqsubs_s32(__s0_918, vqrdmulhs_s32(__s1_918, __noswap_vget_lane_s32(__rev2_918, __p3_918))); \ + __ret_918; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshh_lane_s16(__p0_827, __p1_827, __p2_827, __p3_827) __extension__ ({ \ - int16_t __s0_827 = __p0_827; \ - int16_t __s1_827 = __p1_827; \ - int16x4_t __s2_827 = __p2_827; \ - int16_t __ret_827; \ - __ret_827 = vqsubh_s16(__s0_827, vqrdmulhh_s16(__s1_827, vget_lane_s16(__s2_827, __p3_827))); \ - __ret_827; \ +#define vqrdmlshh_lane_s16(__p0_919, __p1_919, __p2_919, __p3_919) __extension__ ({ \ + int16_t __s0_919 = __p0_919; \ + int16_t __s1_919 = __p1_919; \ + int16x4_t __s2_919 = __p2_919; \ + int16_t __ret_919; \ + __ret_919 = vqsubh_s16(__s0_919, vqrdmulhh_s16(__s1_919, vget_lane_s16(__s2_919, __p3_919))); \ + __ret_919; \ }) #else -#define vqrdmlshh_lane_s16(__p0_828, __p1_828, __p2_828, __p3_828) __extension__ ({ \ - int16_t __s0_828 = __p0_828; \ - int16_t __s1_828 = __p1_828; \ - int16x4_t __s2_828 = __p2_828; \ - int16x4_t __rev2_828; __rev2_828 = __builtin_shufflevector(__s2_828, __s2_828, 3, 2, 1, 0); \ - int16_t __ret_828; \ - __ret_828 = vqsubh_s16(__s0_828, vqrdmulhh_s16(__s1_828, __noswap_vget_lane_s16(__rev2_828, __p3_828))); \ - __ret_828; \ +#define vqrdmlshh_lane_s16(__p0_920, __p1_920, __p2_920, __p3_920) __extension__ ({ \ + int16_t __s0_920 = __p0_920; \ + int16_t __s1_920 = __p1_920; \ + int16x4_t __s2_920 = __p2_920; \ + int16x4_t __rev2_920; __rev2_920 = __builtin_shufflevector(__s2_920, __s2_920, 3, 2, 1, 0); \ + int16_t __ret_920; \ + __ret_920 = vqsubh_s16(__s0_920, vqrdmulhh_s16(__s1_920, __noswap_vget_lane_s16(__rev2_920, __p3_920))); \ + __ret_920; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshs_laneq_s32(__p0_829, __p1_829, __p2_829, __p3_829) __extension__ ({ \ - int32_t __s0_829 = __p0_829; \ - int32_t __s1_829 = __p1_829; \ - int32x4_t __s2_829 = __p2_829; \ - int32_t __ret_829; \ - __ret_829 = vqsubs_s32(__s0_829, vqrdmulhs_s32(__s1_829, vgetq_lane_s32(__s2_829, __p3_829))); \ - __ret_829; \ +#define vqrdmlshs_laneq_s32(__p0_921, __p1_921, __p2_921, __p3_921) __extension__ ({ \ + int32_t __s0_921 = __p0_921; \ + int32_t __s1_921 = __p1_921; \ + int32x4_t __s2_921 = __p2_921; \ + int32_t __ret_921; \ + __ret_921 = vqsubs_s32(__s0_921, vqrdmulhs_s32(__s1_921, vgetq_lane_s32(__s2_921, __p3_921))); \ + __ret_921; \ }) #else -#define vqrdmlshs_laneq_s32(__p0_830, __p1_830, __p2_830, __p3_830) __extension__ ({ \ - int32_t __s0_830 = __p0_830; \ - int32_t __s1_830 = __p1_830; \ - int32x4_t __s2_830 = __p2_830; \ - int32x4_t __rev2_830; __rev2_830 = __builtin_shufflevector(__s2_830, __s2_830, 3, 2, 1, 0); \ - int32_t __ret_830; \ - __ret_830 = vqsubs_s32(__s0_830, vqrdmulhs_s32(__s1_830, __noswap_vgetq_lane_s32(__rev2_830, __p3_830))); \ - __ret_830; \ +#define vqrdmlshs_laneq_s32(__p0_922, __p1_922, __p2_922, __p3_922) __extension__ ({ \ + int32_t __s0_922 = __p0_922; \ + int32_t __s1_922 = __p1_922; \ + int32x4_t __s2_922 = __p2_922; \ + int32x4_t __rev2_922; __rev2_922 = __builtin_shufflevector(__s2_922, __s2_922, 3, 2, 1, 0); \ + int32_t __ret_922; \ + __ret_922 = vqsubs_s32(__s0_922, vqrdmulhs_s32(__s1_922, __noswap_vgetq_lane_s32(__rev2_922, __p3_922))); \ + __ret_922; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vqrdmlshh_laneq_s16(__p0_831, __p1_831, __p2_831, __p3_831) __extension__ ({ \ - int16_t __s0_831 = __p0_831; \ - int16_t __s1_831 = __p1_831; \ - int16x8_t __s2_831 = __p2_831; \ - int16_t __ret_831; \ - __ret_831 = vqsubh_s16(__s0_831, vqrdmulhh_s16(__s1_831, vgetq_lane_s16(__s2_831, __p3_831))); \ - __ret_831; \ +#define vqrdmlshh_laneq_s16(__p0_923, __p1_923, __p2_923, __p3_923) __extension__ ({ \ + int16_t __s0_923 = __p0_923; \ + int16_t __s1_923 = __p1_923; \ + int16x8_t __s2_923 = __p2_923; \ + int16_t __ret_923; \ + __ret_923 = vqsubh_s16(__s0_923, vqrdmulhh_s16(__s1_923, vgetq_lane_s16(__s2_923, __p3_923))); \ + __ret_923; \ }) #else -#define vqrdmlshh_laneq_s16(__p0_832, __p1_832, __p2_832, __p3_832) __extension__ ({ \ - int16_t __s0_832 = __p0_832; \ - int16_t __s1_832 = __p1_832; \ - int16x8_t __s2_832 = __p2_832; \ - int16x8_t __rev2_832; __rev2_832 = __builtin_shufflevector(__s2_832, __s2_832, 7, 6, 5, 4, 3, 2, 1, 0); \ - int16_t __ret_832; \ - __ret_832 = vqsubh_s16(__s0_832, vqrdmulhh_s16(__s1_832, __noswap_vgetq_lane_s16(__rev2_832, __p3_832))); \ - __ret_832; \ +#define vqrdmlshh_laneq_s16(__p0_924, __p1_924, __p2_924, __p3_924) __extension__ ({ \ + int16_t __s0_924 = __p0_924; \ + int16_t __s1_924 = __p1_924; \ + int16x8_t __s2_924 = __p2_924; \ + int16x8_t __rev2_924; __rev2_924 = __builtin_shufflevector(__s2_924, __s2_924, 7, 6, 5, 4, 3, 2, 1, 0); \ + int16_t __ret_924; \ + __ret_924 = vqsubh_s16(__s0_924, vqrdmulhh_s16(__s1_924, __noswap_vgetq_lane_s16(__rev2_924, __p3_924))); \ + __ret_924; \ }) #endif @@ -66535,136 +68205,136 @@ __ai int32x4_t vaddw_high_s16(int32x4_t __p0, int16x8_t __p1) { #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_p64(__p0_833, __p1_833, __p2_833, __p3_833) __extension__ ({ \ - poly64x2_t __s0_833 = __p0_833; \ - poly64x1_t __s2_833 = __p2_833; \ - poly64x2_t __ret_833; \ - __ret_833 = vsetq_lane_p64(vget_lane_p64(__s2_833, __p3_833), __s0_833, __p1_833); \ - __ret_833; \ +#define vcopyq_lane_p64(__p0_925, __p1_925, __p2_925, __p3_925) __extension__ ({ \ + poly64x2_t __s0_925 = __p0_925; \ + poly64x1_t __s2_925 = __p2_925; \ + poly64x2_t __ret_925; \ + __ret_925 = vsetq_lane_p64(vget_lane_p64(__s2_925, __p3_925), __s0_925, __p1_925); \ + __ret_925; \ }) #else -#define vcopyq_lane_p64(__p0_834, __p1_834, __p2_834, __p3_834) __extension__ ({ \ - poly64x2_t __s0_834 = __p0_834; \ - poly64x1_t __s2_834 = __p2_834; \ - poly64x2_t __rev0_834; __rev0_834 = __builtin_shufflevector(__s0_834, __s0_834, 1, 0); \ - poly64x2_t __ret_834; \ - __ret_834 = __noswap_vsetq_lane_p64(vget_lane_p64(__s2_834, __p3_834), __rev0_834, __p1_834); \ - __ret_834 = __builtin_shufflevector(__ret_834, __ret_834, 1, 0); \ - __ret_834; \ +#define vcopyq_lane_p64(__p0_926, __p1_926, __p2_926, __p3_926) __extension__ ({ \ + poly64x2_t __s0_926 = __p0_926; \ + poly64x1_t __s2_926 = __p2_926; \ + poly64x2_t __rev0_926; __rev0_926 = __builtin_shufflevector(__s0_926, __s0_926, 1, 0); \ + poly64x2_t __ret_926; \ + __ret_926 = __noswap_vsetq_lane_p64(vget_lane_p64(__s2_926, __p3_926), __rev0_926, __p1_926); \ + __ret_926 = __builtin_shufflevector(__ret_926, __ret_926, 1, 0); \ + __ret_926; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_lane_f64(__p0_835, __p1_835, __p2_835, __p3_835) __extension__ ({ \ - float64x2_t __s0_835 = __p0_835; \ - float64x1_t __s2_835 = __p2_835; \ - float64x2_t __ret_835; \ - __ret_835 = vsetq_lane_f64(vget_lane_f64(__s2_835, __p3_835), __s0_835, __p1_835); \ - __ret_835; \ +#define vcopyq_lane_f64(__p0_927, __p1_927, __p2_927, __p3_927) __extension__ ({ \ + float64x2_t __s0_927 = __p0_927; \ + float64x1_t __s2_927 = __p2_927; \ + float64x2_t __ret_927; \ + __ret_927 = vsetq_lane_f64(vget_lane_f64(__s2_927, __p3_927), __s0_927, __p1_927); \ + __ret_927; \ }) #else -#define vcopyq_lane_f64(__p0_836, __p1_836, __p2_836, __p3_836) __extension__ ({ \ - float64x2_t __s0_836 = __p0_836; \ - float64x1_t __s2_836 = __p2_836; \ - float64x2_t __rev0_836; __rev0_836 = __builtin_shufflevector(__s0_836, __s0_836, 1, 0); \ - float64x2_t __ret_836; \ - __ret_836 = __noswap_vsetq_lane_f64(vget_lane_f64(__s2_836, __p3_836), __rev0_836, __p1_836); \ - __ret_836 = __builtin_shufflevector(__ret_836, __ret_836, 1, 0); \ - __ret_836; \ +#define vcopyq_lane_f64(__p0_928, __p1_928, __p2_928, __p3_928) __extension__ ({ \ + float64x2_t __s0_928 = __p0_928; \ + float64x1_t __s2_928 = __p2_928; \ + float64x2_t __rev0_928; __rev0_928 = __builtin_shufflevector(__s0_928, __s0_928, 1, 0); \ + float64x2_t __ret_928; \ + __ret_928 = __noswap_vsetq_lane_f64(vget_lane_f64(__s2_928, __p3_928), __rev0_928, __p1_928); \ + __ret_928 = __builtin_shufflevector(__ret_928, __ret_928, 1, 0); \ + __ret_928; \ }) #endif -#define vcopy_lane_p64(__p0_837, __p1_837, __p2_837, __p3_837) __extension__ ({ \ - poly64x1_t __s0_837 = __p0_837; \ - poly64x1_t __s2_837 = __p2_837; \ - poly64x1_t __ret_837; \ - __ret_837 = vset_lane_p64(vget_lane_p64(__s2_837, __p3_837), __s0_837, __p1_837); \ - __ret_837; \ +#define vcopy_lane_p64(__p0_929, __p1_929, __p2_929, __p3_929) __extension__ ({ \ + poly64x1_t __s0_929 = __p0_929; \ + poly64x1_t __s2_929 = __p2_929; \ + poly64x1_t __ret_929; \ + __ret_929 = vset_lane_p64(vget_lane_p64(__s2_929, __p3_929), __s0_929, __p1_929); \ + __ret_929; \ }) -#define vcopy_lane_f64(__p0_838, __p1_838, __p2_838, __p3_838) __extension__ ({ \ - float64x1_t __s0_838 = __p0_838; \ - float64x1_t __s2_838 = __p2_838; \ - float64x1_t __ret_838; \ - __ret_838 = vset_lane_f64(vget_lane_f64(__s2_838, __p3_838), __s0_838, __p1_838); \ - __ret_838; \ +#define vcopy_lane_f64(__p0_930, __p1_930, __p2_930, __p3_930) __extension__ ({ \ + float64x1_t __s0_930 = __p0_930; \ + float64x1_t __s2_930 = __p2_930; \ + float64x1_t __ret_930; \ + __ret_930 = vset_lane_f64(vget_lane_f64(__s2_930, __p3_930), __s0_930, __p1_930); \ + __ret_930; \ }) #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_p64(__p0_839, __p1_839, __p2_839, __p3_839) __extension__ ({ \ - poly64x2_t __s0_839 = __p0_839; \ - poly64x2_t __s2_839 = __p2_839; \ - poly64x2_t __ret_839; \ - __ret_839 = vsetq_lane_p64(vgetq_lane_p64(__s2_839, __p3_839), __s0_839, __p1_839); \ - __ret_839; \ +#define vcopyq_laneq_p64(__p0_931, __p1_931, __p2_931, __p3_931) __extension__ ({ \ + poly64x2_t __s0_931 = __p0_931; \ + poly64x2_t __s2_931 = __p2_931; \ + poly64x2_t __ret_931; \ + __ret_931 = vsetq_lane_p64(vgetq_lane_p64(__s2_931, __p3_931), __s0_931, __p1_931); \ + __ret_931; \ }) #else -#define vcopyq_laneq_p64(__p0_840, __p1_840, __p2_840, __p3_840) __extension__ ({ \ - poly64x2_t __s0_840 = __p0_840; \ - poly64x2_t __s2_840 = __p2_840; \ - poly64x2_t __rev0_840; __rev0_840 = __builtin_shufflevector(__s0_840, __s0_840, 1, 0); \ - poly64x2_t __rev2_840; __rev2_840 = __builtin_shufflevector(__s2_840, __s2_840, 1, 0); \ - poly64x2_t __ret_840; \ - __ret_840 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_840, __p3_840), __rev0_840, __p1_840); \ - __ret_840 = __builtin_shufflevector(__ret_840, __ret_840, 1, 0); \ - __ret_840; \ +#define vcopyq_laneq_p64(__p0_932, __p1_932, __p2_932, __p3_932) __extension__ ({ \ + poly64x2_t __s0_932 = __p0_932; \ + poly64x2_t __s2_932 = __p2_932; \ + poly64x2_t __rev0_932; __rev0_932 = __builtin_shufflevector(__s0_932, __s0_932, 1, 0); \ + poly64x2_t __rev2_932; __rev2_932 = __builtin_shufflevector(__s2_932, __s2_932, 1, 0); \ + poly64x2_t __ret_932; \ + __ret_932 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_932, __p3_932), __rev0_932, __p1_932); \ + __ret_932 = __builtin_shufflevector(__ret_932, __ret_932, 1, 0); \ + __ret_932; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopyq_laneq_f64(__p0_841, __p1_841, __p2_841, __p3_841) __extension__ ({ \ - float64x2_t __s0_841 = __p0_841; \ - float64x2_t __s2_841 = __p2_841; \ - float64x2_t __ret_841; \ - __ret_841 = vsetq_lane_f64(vgetq_lane_f64(__s2_841, __p3_841), __s0_841, __p1_841); \ - __ret_841; \ +#define vcopyq_laneq_f64(__p0_933, __p1_933, __p2_933, __p3_933) __extension__ ({ \ + float64x2_t __s0_933 = __p0_933; \ + float64x2_t __s2_933 = __p2_933; \ + float64x2_t __ret_933; \ + __ret_933 = vsetq_lane_f64(vgetq_lane_f64(__s2_933, __p3_933), __s0_933, __p1_933); \ + __ret_933; \ }) #else -#define vcopyq_laneq_f64(__p0_842, __p1_842, __p2_842, __p3_842) __extension__ ({ \ - float64x2_t __s0_842 = __p0_842; \ - float64x2_t __s2_842 = __p2_842; \ - float64x2_t __rev0_842; __rev0_842 = __builtin_shufflevector(__s0_842, __s0_842, 1, 0); \ - float64x2_t __rev2_842; __rev2_842 = __builtin_shufflevector(__s2_842, __s2_842, 1, 0); \ - float64x2_t __ret_842; \ - __ret_842 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_842, __p3_842), __rev0_842, __p1_842); \ - __ret_842 = __builtin_shufflevector(__ret_842, __ret_842, 1, 0); \ - __ret_842; \ +#define vcopyq_laneq_f64(__p0_934, __p1_934, __p2_934, __p3_934) __extension__ ({ \ + float64x2_t __s0_934 = __p0_934; \ + float64x2_t __s2_934 = __p2_934; \ + float64x2_t __rev0_934; __rev0_934 = __builtin_shufflevector(__s0_934, __s0_934, 1, 0); \ + float64x2_t __rev2_934; __rev2_934 = __builtin_shufflevector(__s2_934, __s2_934, 1, 0); \ + float64x2_t __ret_934; \ + __ret_934 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_934, __p3_934), __rev0_934, __p1_934); \ + __ret_934 = __builtin_shufflevector(__ret_934, __ret_934, 1, 0); \ + __ret_934; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_p64(__p0_843, __p1_843, __p2_843, __p3_843) __extension__ ({ \ - poly64x1_t __s0_843 = __p0_843; \ - poly64x2_t __s2_843 = __p2_843; \ - poly64x1_t __ret_843; \ - __ret_843 = vset_lane_p64(vgetq_lane_p64(__s2_843, __p3_843), __s0_843, __p1_843); \ - __ret_843; \ +#define vcopy_laneq_p64(__p0_935, __p1_935, __p2_935, __p3_935) __extension__ ({ \ + poly64x1_t __s0_935 = __p0_935; \ + poly64x2_t __s2_935 = __p2_935; \ + poly64x1_t __ret_935; \ + __ret_935 = vset_lane_p64(vgetq_lane_p64(__s2_935, __p3_935), __s0_935, __p1_935); \ + __ret_935; \ }) #else -#define vcopy_laneq_p64(__p0_844, __p1_844, __p2_844, __p3_844) __extension__ ({ \ - poly64x1_t __s0_844 = __p0_844; \ - poly64x2_t __s2_844 = __p2_844; \ - poly64x2_t __rev2_844; __rev2_844 = __builtin_shufflevector(__s2_844, __s2_844, 1, 0); \ - poly64x1_t __ret_844; \ - __ret_844 = vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_844, __p3_844), __s0_844, __p1_844); \ - __ret_844; \ +#define vcopy_laneq_p64(__p0_936, __p1_936, __p2_936, __p3_936) __extension__ ({ \ + poly64x1_t __s0_936 = __p0_936; \ + poly64x2_t __s2_936 = __p2_936; \ + poly64x2_t __rev2_936; __rev2_936 = __builtin_shufflevector(__s2_936, __s2_936, 1, 0); \ + poly64x1_t __ret_936; \ + __ret_936 = vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_936, __p3_936), __s0_936, __p1_936); \ + __ret_936; \ }) #endif #ifdef __LITTLE_ENDIAN__ -#define vcopy_laneq_f64(__p0_845, __p1_845, __p2_845, __p3_845) __extension__ ({ \ - float64x1_t __s0_845 = __p0_845; \ - float64x2_t __s2_845 = __p2_845; \ - float64x1_t __ret_845; \ - __ret_845 = vset_lane_f64(vgetq_lane_f64(__s2_845, __p3_845), __s0_845, __p1_845); \ - __ret_845; \ +#define vcopy_laneq_f64(__p0_937, __p1_937, __p2_937, __p3_937) __extension__ ({ \ + float64x1_t __s0_937 = __p0_937; \ + float64x2_t __s2_937 = __p2_937; \ + float64x1_t __ret_937; \ + __ret_937 = vset_lane_f64(vgetq_lane_f64(__s2_937, __p3_937), __s0_937, __p1_937); \ + __ret_937; \ }) #else -#define vcopy_laneq_f64(__p0_846, __p1_846, __p2_846, __p3_846) __extension__ ({ \ - float64x1_t __s0_846 = __p0_846; \ - float64x2_t __s2_846 = __p2_846; \ - float64x2_t __rev2_846; __rev2_846 = __builtin_shufflevector(__s2_846, __s2_846, 1, 0); \ - float64x1_t __ret_846; \ - __ret_846 = vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_846, __p3_846), __s0_846, __p1_846); \ - __ret_846; \ +#define vcopy_laneq_f64(__p0_938, __p1_938, __p2_938, __p3_938) __extension__ ({ \ + float64x1_t __s0_938 = __p0_938; \ + float64x2_t __s2_938 = __p2_938; \ + float64x2_t __rev2_938; __rev2_938 = __builtin_shufflevector(__s2_938, __s2_938, 1, 0); \ + float64x1_t __ret_938; \ + __ret_938 = vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_938, __p3_938), __s0_938, __p1_938); \ + __ret_938; \ }) #endif @@ -67020,38 +68690,38 @@ __ai int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) { } #endif -#define vmulx_lane_f64(__p0_847, __p1_847, __p2_847) __extension__ ({ \ - float64x1_t __s0_847 = __p0_847; \ - float64x1_t __s1_847 = __p1_847; \ - float64x1_t __ret_847; \ - float64_t __x_847 = vget_lane_f64(__s0_847, 0); \ - float64_t __y_847 = vget_lane_f64(__s1_847, __p2_847); \ - float64_t __z_847 = vmulxd_f64(__x_847, __y_847); \ - __ret_847 = vset_lane_f64(__z_847, __s0_847, __p2_847); \ - __ret_847; \ +#define vmulx_lane_f64(__p0_939, __p1_939, __p2_939) __extension__ ({ \ + float64x1_t __s0_939 = __p0_939; \ + float64x1_t __s1_939 = __p1_939; \ + float64x1_t __ret_939; \ + float64_t __x_939 = vget_lane_f64(__s0_939, 0); \ + float64_t __y_939 = vget_lane_f64(__s1_939, __p2_939); \ + float64_t __z_939 = vmulxd_f64(__x_939, __y_939); \ + __ret_939 = vset_lane_f64(__z_939, __s0_939, __p2_939); \ + __ret_939; \ }) #ifdef __LITTLE_ENDIAN__ -#define vmulx_laneq_f64(__p0_848, __p1_848, __p2_848) __extension__ ({ \ - float64x1_t __s0_848 = __p0_848; \ - float64x2_t __s1_848 = __p1_848; \ - float64x1_t __ret_848; \ - float64_t __x_848 = vget_lane_f64(__s0_848, 0); \ - float64_t __y_848 = vgetq_lane_f64(__s1_848, __p2_848); \ - float64_t __z_848 = vmulxd_f64(__x_848, __y_848); \ - __ret_848 = vset_lane_f64(__z_848, __s0_848, 0); \ - __ret_848; \ +#define vmulx_laneq_f64(__p0_940, __p1_940, __p2_940) __extension__ ({ \ + float64x1_t __s0_940 = __p0_940; \ + float64x2_t __s1_940 = __p1_940; \ + float64x1_t __ret_940; \ + float64_t __x_940 = vget_lane_f64(__s0_940, 0); \ + float64_t __y_940 = vgetq_lane_f64(__s1_940, __p2_940); \ + float64_t __z_940 = vmulxd_f64(__x_940, __y_940); \ + __ret_940 = vset_lane_f64(__z_940, __s0_940, 0); \ + __ret_940; \ }) #else -#define vmulx_laneq_f64(__p0_849, __p1_849, __p2_849) __extension__ ({ \ - float64x1_t __s0_849 = __p0_849; \ - float64x2_t __s1_849 = __p1_849; \ - float64x2_t __rev1_849; __rev1_849 = __builtin_shufflevector(__s1_849, __s1_849, 1, 0); \ - float64x1_t __ret_849; \ - float64_t __x_849 = vget_lane_f64(__s0_849, 0); \ - float64_t __y_849 = __noswap_vgetq_lane_f64(__rev1_849, __p2_849); \ - float64_t __z_849 = vmulxd_f64(__x_849, __y_849); \ - __ret_849 = vset_lane_f64(__z_849, __s0_849, 0); \ - __ret_849; \ +#define vmulx_laneq_f64(__p0_941, __p1_941, __p2_941) __extension__ ({ \ + float64x1_t __s0_941 = __p0_941; \ + float64x2_t __s1_941 = __p1_941; \ + float64x2_t __rev1_941; __rev1_941 = __builtin_shufflevector(__s1_941, __s1_941, 1, 0); \ + float64x1_t __ret_941; \ + float64_t __x_941 = vget_lane_f64(__s0_941, 0); \ + float64_t __y_941 = __noswap_vgetq_lane_f64(__rev1_941, __p2_941); \ + float64_t __z_941 = vmulxd_f64(__x_941, __y_941); \ + __ret_941 = vset_lane_f64(__z_941, __s0_941, 0); \ + __ret_941; \ }) #endif diff --git a/lib/include/arm_sve.h b/lib/include/arm_sve.h index 1035d41811..8a03f9da58 100644 --- a/lib/include/arm_sve.h +++ b/lib/include/arm_sve.h @@ -94,7 +94,7 @@ typedef __clang_svbfloat16x2_t svbfloat16x2_t; typedef __clang_svbfloat16x3_t svbfloat16x3_t; typedef __clang_svbfloat16x4_t svbfloat16x4_t; #endif -typedef enum +enum svpattern { SV_POW2 = 0, SV_VL1 = 1, @@ -113,9 +113,9 @@ typedef enum SV_MUL4 = 29, SV_MUL3 = 30, SV_ALL = 31 -} sv_pattern; +}; -typedef enum +enum svprfop { SV_PLDL1KEEP = 0, SV_PLDL1STRM = 1, @@ -129,7 +129,7 @@ typedef enum SV_PSTL2STRM = 11, SV_PSTL3KEEP = 12, SV_PSTL3STRM = 13 -} sv_prfop; +}; /* Function attributes */ #define __aio static inline __attribute__((__always_inline__, __nodebug__, __overloadable__)) @@ -10013,69 +10013,69 @@ int16_t svorv(svbool_t, svint16_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svpfirst_b))) svbool_t svpfirst(svbool_t, svbool_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_u32base))) -void svprfb_gather(svbool_t, svuint32_t, sv_prfop); +void svprfb_gather(svbool_t, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_u64base))) -void svprfb_gather(svbool_t, svuint64_t, sv_prfop); +void svprfb_gather(svbool_t, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_u32base_offset))) -void svprfb_gather_offset(svbool_t, svuint32_t, int64_t, sv_prfop); +void svprfb_gather_offset(svbool_t, svuint32_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_u64base_offset))) -void svprfb_gather_offset(svbool_t, svuint64_t, int64_t, sv_prfop); +void svprfb_gather_offset(svbool_t, svuint64_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_s32offset))) -void svprfb_gather_offset(svbool_t, void const *, svint32_t, sv_prfop); +void svprfb_gather_offset(svbool_t, void const *, svint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_u32offset))) -void svprfb_gather_offset(svbool_t, void const *, svuint32_t, sv_prfop); +void svprfb_gather_offset(svbool_t, void const *, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_s64offset))) -void svprfb_gather_offset(svbool_t, void const *, svint64_t, sv_prfop); +void svprfb_gather_offset(svbool_t, void const *, svint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfb_gather_u64offset))) -void svprfb_gather_offset(svbool_t, void const *, svuint64_t, sv_prfop); +void svprfb_gather_offset(svbool_t, void const *, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_u32base))) -void svprfd_gather(svbool_t, svuint32_t, sv_prfop); +void svprfd_gather(svbool_t, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_u64base))) -void svprfd_gather(svbool_t, svuint64_t, sv_prfop); +void svprfd_gather(svbool_t, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_u32base_index))) -void svprfd_gather_index(svbool_t, svuint32_t, int64_t, sv_prfop); +void svprfd_gather_index(svbool_t, svuint32_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_u64base_index))) -void svprfd_gather_index(svbool_t, svuint64_t, int64_t, sv_prfop); +void svprfd_gather_index(svbool_t, svuint64_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_s32index))) -void svprfd_gather_index(svbool_t, void const *, svint32_t, sv_prfop); +void svprfd_gather_index(svbool_t, void const *, svint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_u32index))) -void svprfd_gather_index(svbool_t, void const *, svuint32_t, sv_prfop); +void svprfd_gather_index(svbool_t, void const *, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_s64index))) -void svprfd_gather_index(svbool_t, void const *, svint64_t, sv_prfop); +void svprfd_gather_index(svbool_t, void const *, svint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfd_gather_u64index))) -void svprfd_gather_index(svbool_t, void const *, svuint64_t, sv_prfop); +void svprfd_gather_index(svbool_t, void const *, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_u32base))) -void svprfh_gather(svbool_t, svuint32_t, sv_prfop); +void svprfh_gather(svbool_t, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_u64base))) -void svprfh_gather(svbool_t, svuint64_t, sv_prfop); +void svprfh_gather(svbool_t, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_u32base_index))) -void svprfh_gather_index(svbool_t, svuint32_t, int64_t, sv_prfop); +void svprfh_gather_index(svbool_t, svuint32_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_u64base_index))) -void svprfh_gather_index(svbool_t, svuint64_t, int64_t, sv_prfop); +void svprfh_gather_index(svbool_t, svuint64_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_s32index))) -void svprfh_gather_index(svbool_t, void const *, svint32_t, sv_prfop); +void svprfh_gather_index(svbool_t, void const *, svint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_u32index))) -void svprfh_gather_index(svbool_t, void const *, svuint32_t, sv_prfop); +void svprfh_gather_index(svbool_t, void const *, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_s64index))) -void svprfh_gather_index(svbool_t, void const *, svint64_t, sv_prfop); +void svprfh_gather_index(svbool_t, void const *, svint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfh_gather_u64index))) -void svprfh_gather_index(svbool_t, void const *, svuint64_t, sv_prfop); +void svprfh_gather_index(svbool_t, void const *, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_u32base))) -void svprfw_gather(svbool_t, svuint32_t, sv_prfop); +void svprfw_gather(svbool_t, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_u64base))) -void svprfw_gather(svbool_t, svuint64_t, sv_prfop); +void svprfw_gather(svbool_t, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_u32base_index))) -void svprfw_gather_index(svbool_t, svuint32_t, int64_t, sv_prfop); +void svprfw_gather_index(svbool_t, svuint32_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_u64base_index))) -void svprfw_gather_index(svbool_t, svuint64_t, int64_t, sv_prfop); +void svprfw_gather_index(svbool_t, svuint64_t, int64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_s32index))) -void svprfw_gather_index(svbool_t, void const *, svint32_t, sv_prfop); +void svprfw_gather_index(svbool_t, void const *, svint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_u32index))) -void svprfw_gather_index(svbool_t, void const *, svuint32_t, sv_prfop); +void svprfw_gather_index(svbool_t, void const *, svuint32_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_s64index))) -void svprfw_gather_index(svbool_t, void const *, svint64_t, sv_prfop); +void svprfw_gather_index(svbool_t, void const *, svint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svprfw_gather_u64index))) -void svprfw_gather_index(svbool_t, void const *, svuint64_t, sv_prfop); +void svprfw_gather_index(svbool_t, void const *, svuint64_t, enum svprfop); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqadd_n_s8))) svint8_t svqadd(svint8_t, int8_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqadd_n_s32))) @@ -10117,13 +10117,13 @@ uint32_t svqdecb(uint32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecb_n_u64))) uint64_t svqdecb(uint64_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecb_pat_n_s32))) -int32_t svqdecb_pat(int32_t, sv_pattern, uint64_t); +int32_t svqdecb_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecb_pat_n_s64))) -int64_t svqdecb_pat(int64_t, sv_pattern, uint64_t); +int64_t svqdecb_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecb_pat_n_u32))) -uint32_t svqdecb_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqdecb_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecb_pat_n_u64))) -uint64_t svqdecb_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqdecb_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_n_s32))) int32_t svqdecd(int32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_n_s64))) @@ -10137,17 +10137,17 @@ svint64_t svqdecd(svint64_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_u64))) svuint64_t svqdecd(svuint64_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_pat_n_s32))) -int32_t svqdecd_pat(int32_t, sv_pattern, uint64_t); +int32_t svqdecd_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_pat_n_s64))) -int64_t svqdecd_pat(int64_t, sv_pattern, uint64_t); +int64_t svqdecd_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_pat_n_u32))) -uint32_t svqdecd_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqdecd_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_pat_n_u64))) -uint64_t svqdecd_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqdecd_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_pat_s64))) -svint64_t svqdecd_pat(svint64_t, sv_pattern, uint64_t); +svint64_t svqdecd_pat(svint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecd_pat_u64))) -svuint64_t svqdecd_pat(svuint64_t, sv_pattern, uint64_t); +svuint64_t svqdecd_pat(svuint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_n_s32))) int32_t svqdech(int32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_n_s64))) @@ -10161,17 +10161,17 @@ svint16_t svqdech(svint16_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_u16))) svuint16_t svqdech(svuint16_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_pat_n_s32))) -int32_t svqdech_pat(int32_t, sv_pattern, uint64_t); +int32_t svqdech_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_pat_n_s64))) -int64_t svqdech_pat(int64_t, sv_pattern, uint64_t); +int64_t svqdech_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_pat_n_u32))) -uint32_t svqdech_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqdech_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_pat_n_u64))) -uint64_t svqdech_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqdech_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_pat_s16))) -svint16_t svqdech_pat(svint16_t, sv_pattern, uint64_t); +svint16_t svqdech_pat(svint16_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdech_pat_u16))) -svuint16_t svqdech_pat(svuint16_t, sv_pattern, uint64_t); +svuint16_t svqdech_pat(svuint16_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecp_n_s32_b8))) int32_t svqdecp_b8(int32_t, svbool_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecp_n_s32_b32))) @@ -10229,17 +10229,17 @@ svint32_t svqdecw(svint32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecw_u32))) svuint32_t svqdecw(svuint32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecw_pat_n_s32))) -int32_t svqdecw_pat(int32_t, sv_pattern, uint64_t); +int32_t svqdecw_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecw_pat_n_s64))) -int64_t svqdecw_pat(int64_t, sv_pattern, uint64_t); +int64_t svqdecw_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecw_pat_n_u32))) -uint32_t svqdecw_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqdecw_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecw_pat_n_u64))) -uint64_t svqdecw_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqdecw_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecw_pat_s32))) -svint32_t svqdecw_pat(svint32_t, sv_pattern, uint64_t); +svint32_t svqdecw_pat(svint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqdecw_pat_u32))) -svuint32_t svqdecw_pat(svuint32_t, sv_pattern, uint64_t); +svuint32_t svqdecw_pat(svuint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincb_n_s32))) int32_t svqincb(int32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincb_n_s64))) @@ -10249,13 +10249,13 @@ uint32_t svqincb(uint32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincb_n_u64))) uint64_t svqincb(uint64_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincb_pat_n_s32))) -int32_t svqincb_pat(int32_t, sv_pattern, uint64_t); +int32_t svqincb_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincb_pat_n_s64))) -int64_t svqincb_pat(int64_t, sv_pattern, uint64_t); +int64_t svqincb_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincb_pat_n_u32))) -uint32_t svqincb_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqincb_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincb_pat_n_u64))) -uint64_t svqincb_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqincb_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_n_s32))) int32_t svqincd(int32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_n_s64))) @@ -10269,17 +10269,17 @@ svint64_t svqincd(svint64_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_u64))) svuint64_t svqincd(svuint64_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_pat_n_s32))) -int32_t svqincd_pat(int32_t, sv_pattern, uint64_t); +int32_t svqincd_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_pat_n_s64))) -int64_t svqincd_pat(int64_t, sv_pattern, uint64_t); +int64_t svqincd_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_pat_n_u32))) -uint32_t svqincd_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqincd_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_pat_n_u64))) -uint64_t svqincd_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqincd_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_pat_s64))) -svint64_t svqincd_pat(svint64_t, sv_pattern, uint64_t); +svint64_t svqincd_pat(svint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincd_pat_u64))) -svuint64_t svqincd_pat(svuint64_t, sv_pattern, uint64_t); +svuint64_t svqincd_pat(svuint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_n_s32))) int32_t svqinch(int32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_n_s64))) @@ -10293,17 +10293,17 @@ svint16_t svqinch(svint16_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_u16))) svuint16_t svqinch(svuint16_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_pat_n_s32))) -int32_t svqinch_pat(int32_t, sv_pattern, uint64_t); +int32_t svqinch_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_pat_n_s64))) -int64_t svqinch_pat(int64_t, sv_pattern, uint64_t); +int64_t svqinch_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_pat_n_u32))) -uint32_t svqinch_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqinch_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_pat_n_u64))) -uint64_t svqinch_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqinch_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_pat_s16))) -svint16_t svqinch_pat(svint16_t, sv_pattern, uint64_t); +svint16_t svqinch_pat(svint16_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqinch_pat_u16))) -svuint16_t svqinch_pat(svuint16_t, sv_pattern, uint64_t); +svuint16_t svqinch_pat(svuint16_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincp_n_s32_b8))) int32_t svqincp_b8(int32_t, svbool_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincp_n_s32_b32))) @@ -10361,17 +10361,17 @@ svint32_t svqincw(svint32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincw_u32))) svuint32_t svqincw(svuint32_t, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincw_pat_n_s32))) -int32_t svqincw_pat(int32_t, sv_pattern, uint64_t); +int32_t svqincw_pat(int32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincw_pat_n_s64))) -int64_t svqincw_pat(int64_t, sv_pattern, uint64_t); +int64_t svqincw_pat(int64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincw_pat_n_u32))) -uint32_t svqincw_pat(uint32_t, sv_pattern, uint64_t); +uint32_t svqincw_pat(uint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincw_pat_n_u64))) -uint64_t svqincw_pat(uint64_t, sv_pattern, uint64_t); +uint64_t svqincw_pat(uint64_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincw_pat_s32))) -svint32_t svqincw_pat(svint32_t, sv_pattern, uint64_t); +svint32_t svqincw_pat(svint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqincw_pat_u32))) -svuint32_t svqincw_pat(svuint32_t, sv_pattern, uint64_t); +svuint32_t svqincw_pat(svuint32_t, enum svpattern, uint64_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqsub_n_s8))) svint8_t svqsub(svint8_t, int8_t); __aio __attribute__((__clang_arm_builtin_alias(__builtin_sve_svqsub_n_s32))) diff --git a/lib/include/avx512fintrin.h b/lib/include/avx512fintrin.h index fa22ef3fdd..f226382cbb 100644 --- a/lib/include/avx512fintrin.h +++ b/lib/include/avx512fintrin.h @@ -9297,303 +9297,232 @@ _mm512_mask_abs_pd(__m512d __W, __mmask8 __K, __m512d __A) /* Vector-reduction arithmetic accepts vectors as inputs and produces scalars as * outputs. This class of vector operation forms the basis of many scientific - * computations. In vector-reduction arithmetic, the evaluation off is + * computations. In vector-reduction arithmetic, the evaluation order is * independent of the order of the input elements of V. + * For floating point types, we always assume the elements are reassociable even + * if -fast-math is off. + * Used bisection method. At each step, we partition the vector with previous * step in half, and the operation is performed on its two halves. * This takes log2(n) steps where n is the number of elements in the vector. */ -#define _mm512_mask_reduce_operator(op) \ - __v4du __t1 = (__v4du)_mm512_extracti64x4_epi64(__W, 0); \ - __v4du __t2 = (__v4du)_mm512_extracti64x4_epi64(__W, 1); \ - __m256i __t3 = (__m256i)(__t1 op __t2); \ - __v2du __t4 = (__v2du)_mm256_extracti128_si256(__t3, 0); \ - __v2du __t5 = (__v2du)_mm256_extracti128_si256(__t3, 1); \ - __v2du __t6 = __t4 op __t5; \ - __v2du __t7 = __builtin_shufflevector(__t6, __t6, 1, 0); \ - __v2du __t8 = __t6 op __t7; \ - return __t8[0] - static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_add_epi64(__m512i __W) { - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_add_q512(__W); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_mul_epi64(__m512i __W) { - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_mul_q512(__W); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_and_epi64(__m512i __W) { - _mm512_mask_reduce_operator(&); + return __builtin_ia32_reduce_and_q512(__W); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_or_epi64(__m512i __W) { - _mm512_mask_reduce_operator(|); + return __builtin_ia32_reduce_or_q512(__W); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_add_epi64(__mmask8 __M, __m512i __W) { __W = _mm512_maskz_mov_epi64(__M, __W); - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_add_q512(__W); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_mul_epi64(__mmask8 __M, __m512i __W) { __W = _mm512_mask_mov_epi64(_mm512_set1_epi64(1), __M, __W); - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_mul_q512(__W); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_and_epi64(__mmask8 __M, __m512i __W) { __W = _mm512_mask_mov_epi64(_mm512_set1_epi64(~0ULL), __M, __W); - _mm512_mask_reduce_operator(&); + return __builtin_ia32_reduce_and_q512(__W); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_or_epi64(__mmask8 __M, __m512i __W) { __W = _mm512_maskz_mov_epi64(__M, __W); - _mm512_mask_reduce_operator(|); + return __builtin_ia32_reduce_or_q512(__W); } -#undef _mm512_mask_reduce_operator - -#define _mm512_mask_reduce_operator(op) \ - __m256d __t1 = _mm512_extractf64x4_pd(__W, 0); \ - __m256d __t2 = _mm512_extractf64x4_pd(__W, 1); \ - __m256d __t3 = __t1 op __t2; \ - __m128d __t4 = _mm256_extractf128_pd(__t3, 0); \ - __m128d __t5 = _mm256_extractf128_pd(__t3, 1); \ - __m128d __t6 = __t4 op __t5; \ - __m128d __t7 = __builtin_shufflevector(__t6, __t6, 1, 0); \ - __m128d __t8 = __t6 op __t7; \ - return __t8[0] +// -0.0 is used to ignore the start value since it is the neutral value of +// floating point addition. For more information, please refer to +// https://llvm.org/docs/LangRef.html#llvm-vector-reduce-fadd-intrinsic static __inline__ double __DEFAULT_FN_ATTRS512 _mm512_reduce_add_pd(__m512d __W) { - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_fadd_pd512(-0.0, __W); } static __inline__ double __DEFAULT_FN_ATTRS512 _mm512_reduce_mul_pd(__m512d __W) { - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_fmul_pd512(1.0, __W); } static __inline__ double __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_add_pd(__mmask8 __M, __m512d __W) { __W = _mm512_maskz_mov_pd(__M, __W); - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_fadd_pd512(-0.0, __W); } static __inline__ double __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_mul_pd(__mmask8 __M, __m512d __W) { __W = _mm512_mask_mov_pd(_mm512_set1_pd(1.0), __M, __W); - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_fmul_pd512(1.0, __W); } -#undef _mm512_mask_reduce_operator - -#define _mm512_mask_reduce_operator(op) \ - __v8su __t1 = (__v8su)_mm512_extracti64x4_epi64(__W, 0); \ - __v8su __t2 = (__v8su)_mm512_extracti64x4_epi64(__W, 1); \ - __m256i __t3 = (__m256i)(__t1 op __t2); \ - __v4su __t4 = (__v4su)_mm256_extracti128_si256(__t3, 0); \ - __v4su __t5 = (__v4su)_mm256_extracti128_si256(__t3, 1); \ - __v4su __t6 = __t4 op __t5; \ - __v4su __t7 = __builtin_shufflevector(__t6, __t6, 2, 3, 0, 1); \ - __v4su __t8 = __t6 op __t7; \ - __v4su __t9 = __builtin_shufflevector(__t8, __t8, 1, 0, 3, 2); \ - __v4su __t10 = __t8 op __t9; \ - return __t10[0] static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_reduce_add_epi32(__m512i __W) { - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_add_d512((__v16si)__W); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_reduce_mul_epi32(__m512i __W) { - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_mul_d512((__v16si)__W); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_reduce_and_epi32(__m512i __W) { - _mm512_mask_reduce_operator(&); + return __builtin_ia32_reduce_and_d512((__v16si)__W); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_reduce_or_epi32(__m512i __W) { - _mm512_mask_reduce_operator(|); + return __builtin_ia32_reduce_or_d512((__v16si)__W); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_add_epi32( __mmask16 __M, __m512i __W) { __W = _mm512_maskz_mov_epi32(__M, __W); - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_add_d512((__v16si)__W); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_mul_epi32( __mmask16 __M, __m512i __W) { __W = _mm512_mask_mov_epi32(_mm512_set1_epi32(1), __M, __W); - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_mul_d512((__v16si)__W); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_and_epi32( __mmask16 __M, __m512i __W) { __W = _mm512_mask_mov_epi32(_mm512_set1_epi32(~0U), __M, __W); - _mm512_mask_reduce_operator(&); + return __builtin_ia32_reduce_and_d512((__v16si)__W); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_or_epi32(__mmask16 __M, __m512i __W) { __W = _mm512_maskz_mov_epi32(__M, __W); - _mm512_mask_reduce_operator(|); + return __builtin_ia32_reduce_or_d512((__v16si)__W); } -#undef _mm512_mask_reduce_operator - -#define _mm512_mask_reduce_operator(op) \ - __m256 __t1 = (__m256)_mm512_extractf64x4_pd((__m512d)__W, 0); \ - __m256 __t2 = (__m256)_mm512_extractf64x4_pd((__m512d)__W, 1); \ - __m256 __t3 = __t1 op __t2; \ - __m128 __t4 = _mm256_extractf128_ps(__t3, 0); \ - __m128 __t5 = _mm256_extractf128_ps(__t3, 1); \ - __m128 __t6 = __t4 op __t5; \ - __m128 __t7 = __builtin_shufflevector(__t6, __t6, 2, 3, 0, 1); \ - __m128 __t8 = __t6 op __t7; \ - __m128 __t9 = __builtin_shufflevector(__t8, __t8, 1, 0, 3, 2); \ - __m128 __t10 = __t8 op __t9; \ - return __t10[0] static __inline__ float __DEFAULT_FN_ATTRS512 _mm512_reduce_add_ps(__m512 __W) { - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_fadd_ps512(-0.0f, __W); } static __inline__ float __DEFAULT_FN_ATTRS512 _mm512_reduce_mul_ps(__m512 __W) { - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_fmul_ps512(1.0f, __W); } static __inline__ float __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_add_ps(__mmask16 __M, __m512 __W) { __W = _mm512_maskz_mov_ps(__M, __W); - _mm512_mask_reduce_operator(+); + return __builtin_ia32_reduce_fadd_ps512(-0.0f, __W); } static __inline__ float __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_mul_ps(__mmask16 __M, __m512 __W) { __W = _mm512_mask_mov_ps(_mm512_set1_ps(1.0f), __M, __W); - _mm512_mask_reduce_operator(*); + return __builtin_ia32_reduce_fmul_ps512(1.0f, __W); } -#undef _mm512_mask_reduce_operator - -#define _mm512_mask_reduce_operator(op) \ - __m512i __t1 = (__m512i)__builtin_shufflevector((__v8di)__V, (__v8di)__V, 4, 5, 6, 7, 0, 1, 2, 3); \ - __m512i __t2 = _mm512_##op(__V, __t1); \ - __m512i __t3 = (__m512i)__builtin_shufflevector((__v8di)__t2, (__v8di)__t2, 2, 3, 0, 1, 6, 7, 4, 5); \ - __m512i __t4 = _mm512_##op(__t2, __t3); \ - __m512i __t5 = (__m512i)__builtin_shufflevector((__v8di)__t4, (__v8di)__t4, 1, 0, 3, 2, 5, 4, 7, 6); \ - __v8di __t6 = (__v8di)_mm512_##op(__t4, __t5); \ - return __t6[0] static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_max_epi64(__m512i __V) { - _mm512_mask_reduce_operator(max_epi64); + return __builtin_ia32_reduce_smax_q512(__V); } static __inline__ unsigned long long __DEFAULT_FN_ATTRS512 _mm512_reduce_max_epu64(__m512i __V) { - _mm512_mask_reduce_operator(max_epu64); + return __builtin_ia32_reduce_umax_q512(__V); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_min_epi64(__m512i __V) { - _mm512_mask_reduce_operator(min_epi64); + return __builtin_ia32_reduce_smin_q512(__V); } static __inline__ unsigned long long __DEFAULT_FN_ATTRS512 _mm512_reduce_min_epu64(__m512i __V) { - _mm512_mask_reduce_operator(min_epu64); + return __builtin_ia32_reduce_umin_q512(__V); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_max_epi64(__mmask8 __M, __m512i __V) { __V = _mm512_mask_mov_epi64(_mm512_set1_epi64(-__LONG_LONG_MAX__ - 1LL), __M, __V); - _mm512_mask_reduce_operator(max_epi64); + return __builtin_ia32_reduce_smax_q512(__V); } static __inline__ unsigned long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_max_epu64(__mmask8 __M, __m512i __V) { __V = _mm512_maskz_mov_epi64(__M, __V); - _mm512_mask_reduce_operator(max_epu64); + return __builtin_ia32_reduce_umax_q512(__V); } static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_min_epi64(__mmask8 __M, __m512i __V) { __V = _mm512_mask_mov_epi64(_mm512_set1_epi64(__LONG_LONG_MAX__), __M, __V); - _mm512_mask_reduce_operator(min_epi64); + return __builtin_ia32_reduce_smin_q512(__V); } static __inline__ unsigned long long __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_min_epu64(__mmask8 __M, __m512i __V) { __V = _mm512_mask_mov_epi64(_mm512_set1_epi64(~0ULL), __M, __V); - _mm512_mask_reduce_operator(min_epu64); + return __builtin_ia32_reduce_umin_q512(__V); } -#undef _mm512_mask_reduce_operator - -#define _mm512_mask_reduce_operator(op) \ - __m256i __t1 = _mm512_extracti64x4_epi64(__V, 0); \ - __m256i __t2 = _mm512_extracti64x4_epi64(__V, 1); \ - __m256i __t3 = _mm256_##op(__t1, __t2); \ - __m128i __t4 = _mm256_extracti128_si256(__t3, 0); \ - __m128i __t5 = _mm256_extracti128_si256(__t3, 1); \ - __m128i __t6 = _mm_##op(__t4, __t5); \ - __m128i __t7 = (__m128i)__builtin_shufflevector((__v4si)__t6, (__v4si)__t6, 2, 3, 0, 1); \ - __m128i __t8 = _mm_##op(__t6, __t7); \ - __m128i __t9 = (__m128i)__builtin_shufflevector((__v4si)__t8, (__v4si)__t8, 1, 0, 3, 2); \ - __v4si __t10 = (__v4si)_mm_##op(__t8, __t9); \ - return __t10[0] - static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_reduce_max_epi32(__m512i __V) { - _mm512_mask_reduce_operator(max_epi32); + return __builtin_ia32_reduce_smax_d512((__v16si)__V); } static __inline__ unsigned int __DEFAULT_FN_ATTRS512 _mm512_reduce_max_epu32(__m512i __V) { - _mm512_mask_reduce_operator(max_epu32); + return __builtin_ia32_reduce_umax_d512((__v16si)__V); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_reduce_min_epi32(__m512i __V) { - _mm512_mask_reduce_operator(min_epi32); + return __builtin_ia32_reduce_smin_d512((__v16si)__V); } static __inline__ unsigned int __DEFAULT_FN_ATTRS512 _mm512_reduce_min_epu32(__m512i __V) { - _mm512_mask_reduce_operator(min_epu32); + return __builtin_ia32_reduce_umin_d512((__v16si)__V); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_max_epi32(__mmask16 __M, __m512i __V) { __V = _mm512_mask_mov_epi32(_mm512_set1_epi32(-__INT_MAX__ - 1), __M, __V); - _mm512_mask_reduce_operator(max_epi32); + return __builtin_ia32_reduce_smax_d512((__v16si)__V); } static __inline__ unsigned int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_max_epu32(__mmask16 __M, __m512i __V) { __V = _mm512_maskz_mov_epi32(__M, __V); - _mm512_mask_reduce_operator(max_epu32); + return __builtin_ia32_reduce_umax_d512((__v16si)__V); } static __inline__ int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_min_epi32(__mmask16 __M, __m512i __V) { __V = _mm512_mask_mov_epi32(_mm512_set1_epi32(__INT_MAX__), __M, __V); - _mm512_mask_reduce_operator(min_epi32); + return __builtin_ia32_reduce_smin_d512((__v16si)__V); } static __inline__ unsigned int __DEFAULT_FN_ATTRS512 _mm512_mask_reduce_min_epu32(__mmask16 __M, __m512i __V) { __V = _mm512_mask_mov_epi32(_mm512_set1_epi32(~0U), __M, __V); - _mm512_mask_reduce_operator(min_epu32); + return __builtin_ia32_reduce_umin_d512((__v16si)__V); } -#undef _mm512_mask_reduce_operator #define _mm512_mask_reduce_operator(op) \ __m256d __t1 = _mm512_extractf64x4_pd(__V, 0); \ diff --git a/lib/include/avx512vlvnniintrin.h b/lib/include/avx512vlvnniintrin.h index b7c8fa08c6..71ac1b4370 100644 --- a/lib/include/avx512vlvnniintrin.h +++ b/lib/include/avx512vlvnniintrin.h @@ -18,13 +18,157 @@ #define __DEFAULT_FN_ATTRS128 __attribute__((__always_inline__, __nodebug__, __target__("avx512vl,avx512vnni"), __min_vector_width__(128))) #define __DEFAULT_FN_ATTRS256 __attribute__((__always_inline__, __nodebug__, __target__("avx512vl,avx512vnni"), __min_vector_width__(256))) +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with +/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a S, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3])) +/// DST.dword[j] := S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4 +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +#define _mm256_dpbusd_epi32(S, A, B) \ + (__m256i)__builtin_ia32_vpdpbusd256((__v8si)(S), (__v8si)(A), (__v8si)(B)) -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_dpbusd_epi32(__m256i __S, __m256i __A, __m256i __B) -{ - return (__m256i)__builtin_ia32_vpdpbusd256((__v8si)__S, (__v8si)__A, - (__v8si)__B); -} +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with +/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a S using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3])) +/// DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4) +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +#define _mm256_dpbusds_epi32(S, A, B) \ + (__m256i)__builtin_ia32_vpdpbusds256((__v8si)(S), (__v8si)(A), (__v8si)(B)) + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with +/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a S, +/// and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j]) +/// tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1]) +/// DST.dword[j] := S.dword[j] + tmp1 + tmp2 +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +#define _mm256_dpwssd_epi32(S, A, B) \ + (__m256i)__builtin_ia32_vpdpwssd256((__v8si)(S), (__v8si)(A), (__v8si)(B)) + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with +/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a S +/// using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j]) +/// tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1]) +/// DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2) +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +#define _mm256_dpwssds_epi32(S, A, B) \ + (__m256i)__builtin_ia32_vpdpwssds256((__v8si)(S), (__v8si)(A), (__v8si)(B)) + +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with +/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a S, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3])) +/// DST.dword[j] := S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4 +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +#define _mm_dpbusd_epi32(S, A, B) \ + (__m128i)__builtin_ia32_vpdpbusd128((__v4si)(S), (__v4si)(A), (__v4si)(B)) + +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with +/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a S using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3])) +/// DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4) +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +#define _mm_dpbusds_epi32(S, A, B) \ + (__m128i)__builtin_ia32_vpdpbusds128((__v4si)(S), (__v4si)(A), (__v4si)(B)) + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with +/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a S, +/// and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j]) +/// tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1]) +/// DST.dword[j] := S.dword[j] + tmp1 + tmp2 +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +#define _mm_dpwssd_epi32(S, A, B) \ + (__m128i)__builtin_ia32_vpdpwssd128((__v4si)(S), (__v4si)(A), (__v4si)(B)) + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with +/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a S +/// using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j]) +/// tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1]) +/// DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2) +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +#define _mm_dpwssds_epi32(S, A, B) \ + (__m128i)__builtin_ia32_vpdpwssds128((__v4si)(S), (__v4si)(A), (__v4si)(B)) static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpbusd_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B) @@ -43,13 +187,6 @@ _mm256_maskz_dpbusd_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B) } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_dpbusds_epi32(__m256i __S, __m256i __A, __m256i __B) -{ - return (__m256i)__builtin_ia32_vpdpbusds256((__v8si)__S, (__v8si)__A, - (__v8si)__B); -} - -static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpbusds_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectd_256(__U, @@ -66,13 +203,6 @@ _mm256_maskz_dpbusds_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B) } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_dpwssd_epi32(__m256i __S, __m256i __A, __m256i __B) -{ - return (__m256i)__builtin_ia32_vpdpwssd256((__v8si)__S, (__v8si)__A, - (__v8si)__B); -} - -static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpwssd_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectd_256(__U, @@ -89,13 +219,6 @@ _mm256_maskz_dpwssd_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B) } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_dpwssds_epi32(__m256i __S, __m256i __A, __m256i __B) -{ - return (__m256i)__builtin_ia32_vpdpwssds256((__v8si)__S, (__v8si)__A, - (__v8si)__B); -} - -static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpwssds_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectd_256(__U, @@ -112,13 +235,6 @@ _mm256_maskz_dpwssds_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B) } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_dpbusd_epi32(__m128i __S, __m128i __A, __m128i __B) -{ - return (__m128i)__builtin_ia32_vpdpbusd128((__v4si)__S, (__v4si)__A, - (__v4si)__B); -} - -static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpbusd_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128(__U, @@ -135,13 +251,6 @@ _mm_maskz_dpbusd_epi32(__mmask8 __U, __m128i __S, __m128i __A, __m128i __B) } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_dpbusds_epi32(__m128i __S, __m128i __A, __m128i __B) -{ - return (__m128i)__builtin_ia32_vpdpbusds128((__v4si)__S, (__v4si)__A, - (__v4si)__B); -} - -static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpbusds_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128(__U, @@ -158,13 +267,6 @@ _mm_maskz_dpbusds_epi32(__mmask8 __U, __m128i __S, __m128i __A, __m128i __B) } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_dpwssd_epi32(__m128i __S, __m128i __A, __m128i __B) -{ - return (__m128i)__builtin_ia32_vpdpwssd128((__v4si)__S, (__v4si)__A, - (__v4si)__B); -} - -static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpwssd_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128(__U, @@ -181,13 +283,6 @@ _mm_maskz_dpwssd_epi32(__mmask8 __U, __m128i __S, __m128i __A, __m128i __B) } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_dpwssds_epi32(__m128i __S, __m128i __A, __m128i __B) -{ - return (__m128i)__builtin_ia32_vpdpwssds128((__v4si)__S, (__v4si)__A, - (__v4si)__B); -} - -static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpwssds_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128(__U, diff --git a/lib/include/avxintrin.h b/lib/include/avxintrin.h index 84421bf1b9..382b621575 100644 --- a/lib/include/avxintrin.h +++ b/lib/include/avxintrin.h @@ -2245,7 +2245,7 @@ _mm256_cvttps_epi32(__m256 __a) /// Returns the first element of the input vector of [4 x double]. /// -/// \headerfile <avxintrin.h> +/// \headerfile <x86intrin.h> /// /// This intrinsic is a utility function and does not correspond to a specific /// instruction. @@ -2261,7 +2261,7 @@ _mm256_cvtsd_f64(__m256d __a) /// Returns the first element of the input vector of [8 x i32]. /// -/// \headerfile <avxintrin.h> +/// \headerfile <x86intrin.h> /// /// This intrinsic is a utility function and does not correspond to a specific /// instruction. @@ -2278,7 +2278,7 @@ _mm256_cvtsi256_si32(__m256i __a) /// Returns the first element of the input vector of [8 x float]. /// -/// \headerfile <avxintrin.h> +/// \headerfile <x86intrin.h> /// /// This intrinsic is a utility function and does not correspond to a specific /// instruction. diff --git a/lib/include/avxvnniintrin.h b/lib/include/avxvnniintrin.h new file mode 100644 index 0000000000..ad45cb7962 --- /dev/null +++ b/lib/include/avxvnniintrin.h @@ -0,0 +1,225 @@ +/*===--------------- avxvnniintrin.h - VNNI intrinsics --------------------=== + * + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + * + *===-----------------------------------------------------------------------=== + */ +#ifndef __IMMINTRIN_H +#error "Never use <avxvnniintrin.h> directly; include <immintrin.h> instead." +#endif + +#ifndef __AVXVNNIINTRIN_H +#define __AVXVNNIINTRIN_H + +/* Below intrinsics defined in avx512vlvnniintrin.h can be used for AVXVNNI */ +/// \fn __m256i _mm256_dpbusd_epi32(__m256i __S, __m256i __A, __m256i __B) +/// \fn __m256i _mm256_dpbusds_epi32(__m256i __S, __m256i __A, __m256i __B) +/// \fn __m256i _mm256_dpwssd_epi32(__m256i __S, __m256i __A, __m256i __B) +/// \fn __m256i _mm256_dpwssds_epi32(__m256i __S, __m256i __A, __m256i __B) +/// \fn __m128i _mm_dpbusd_epi32(__m128i __S, __m128i __A, __m128i __B) +/// \fn __m128i _mm_dpbusds_epi32(__m128i __S, __m128i __A, __m128i __B) +/// \fn __m128i _mm_dpwssd_epi32(__m128i __S, __m128i __A, __m128i __B) +/// \fn __m128i _mm_dpwssds_epi32(__m128i __S, __m128i __A, __m128i __B) + +/* Intrinsics with _avx_ prefix are for compatibility with msvc. */ +/* Define the default attributes for the functions in this file. */ +#define __DEFAULT_FN_ATTRS256 __attribute__((__always_inline__, __nodebug__, __target__("avxvnni"), __min_vector_width__(256))) +#define __DEFAULT_FN_ATTRS128 __attribute__((__always_inline__, __nodebug__, __target__("avxvnni"), __min_vector_width__(128))) + +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with +/// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a __S, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3])) +/// DST.dword[j] := __S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4 +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +static __inline__ __m256i __DEFAULT_FN_ATTRS256 +_mm256_dpbusd_avx_epi32(__m256i __S, __m256i __A, __m256i __B) +{ + return (__m256i)__builtin_ia32_vpdpbusd256((__v8si)__S, (__v8si)__A, (__v8si)__B); +} + +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with +/// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a __S using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3])) +/// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4) +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +static __inline__ __m256i __DEFAULT_FN_ATTRS256 +_mm256_dpbusds_avx_epi32(__m256i __S, __m256i __A, __m256i __B) +{ + return (__m256i)__builtin_ia32_vpdpbusds256((__v8si)__S, (__v8si)__A, (__v8si)__B); +} + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with +/// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a __S, +/// and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j]) +/// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1]) +/// DST.dword[j] := __S.dword[j] + tmp1 + tmp2 +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +static __inline__ __m256i __DEFAULT_FN_ATTRS256 +_mm256_dpwssd_avx_epi32(__m256i __S, __m256i __A, __m256i __B) +{ + return (__m256i)__builtin_ia32_vpdpwssd256((__v8si)__S, (__v8si)__A, (__v8si)__B); +} + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with +/// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a __S +/// using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 7 +/// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j]) +/// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1]) +/// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2) +/// ENDFOR +/// DST[MAX:256] := 0 +/// \endoperation +static __inline__ __m256i __DEFAULT_FN_ATTRS256 +_mm256_dpwssds_avx_epi32(__m256i __S, __m256i __A, __m256i __B) +{ + return (__m256i)__builtin_ia32_vpdpwssds256((__v8si)__S, (__v8si)__A, (__v8si)__B); +} + +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with +/// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a __S, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3])) +/// DST.dword[j] := __S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4 +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +static __inline__ __m128i __DEFAULT_FN_ATTRS128 +_mm_dpbusd_avx_epi32(__m128i __S, __m128i __A, __m128i __B) +{ + return (__m128i)__builtin_ia32_vpdpbusd128((__v4si)__S, (__v4si)__A, (__v4si)__B); +} + +/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with +/// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed +/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer +/// in \a __S using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPBUSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j])) +/// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1])) +/// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2])) +/// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3])) +/// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4) +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +static __inline__ __m128i __DEFAULT_FN_ATTRS128 +_mm_dpbusds_avx_epi32(__m128i __S, __m128i __A, __m128i __B) +{ + return (__m128i)__builtin_ia32_vpdpbusds128((__v4si)__S, (__v4si)__A, (__v4si)__B); +} + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with +/// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a __S, +/// and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSD </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j]) +/// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1]) +/// DST.dword[j] := __S.dword[j] + tmp1 + tmp2 +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +static __inline__ __m128i __DEFAULT_FN_ATTRS128 +_mm_dpwssd_avx_epi32(__m128i __S, __m128i __A, __m128i __B) +{ + return (__m128i)__builtin_ia32_vpdpwssd128((__v4si)__S, (__v4si)__A, (__v4si)__B); +} + +/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with +/// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit +/// results. Sum these 2 results with the corresponding 32-bit integer in \a __S +/// using signed saturation, and store the packed 32-bit results in DST. +/// +/// This intrinsic corresponds to the <c> VPDPWSSDS </c> instructions. +/// +/// \operation +/// FOR j := 0 to 3 +/// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j]) +/// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1]) +/// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2) +/// ENDFOR +/// DST[MAX:128] := 0 +/// \endoperation +static __inline__ __m128i __DEFAULT_FN_ATTRS128 +_mm_dpwssds_avx_epi32(__m128i __S, __m128i __A, __m128i __B) +{ + return (__m128i)__builtin_ia32_vpdpwssds128((__v4si)__S, (__v4si)__A, (__v4si)__B); +} + +#undef __DEFAULT_FN_ATTRS128 +#undef __DEFAULT_FN_ATTRS256 + +#endif // __AVXVNNIINTRIN_H diff --git a/lib/include/cpuid.h b/lib/include/cpuid.h index 2a88c042d0..34f0e76807 100644 --- a/lib/include/cpuid.h +++ b/lib/include/cpuid.h @@ -7,6 +7,9 @@ *===-----------------------------------------------------------------------=== */ +#ifndef __CPUID_H +#define __CPUID_H + #if !(__x86_64__ || __i386__) #error this header is for x86 only #endif @@ -186,6 +189,7 @@ /* Features in %edx for leaf 7 sub-leaf 0 */ #define bit_AVX5124VNNIW 0x00000004 #define bit_AVX5124FMAPS 0x00000008 +#define bit_UINTR 0x00000020 #define bit_SERIALIZE 0x00004000 #define bit_TSXLDTRK 0x00010000 #define bit_PCONFIG 0x00040000 @@ -195,7 +199,9 @@ #define bit_AMXINT8 0x02000000 /* Features in %eax for leaf 7 sub-leaf 1 */ +#define bit_AVXVNNI 0x00000008 #define bit_AVX512BF16 0x00000020 +#define bit_HRESET 0x00400000 /* Features in %eax for leaf 13 sub-leaf 1 */ #define bit_XSAVEOPT 0x00000001 @@ -309,3 +315,5 @@ static __inline int __get_cpuid_count (unsigned int __leaf, __cpuid_count(__leaf, __subleaf, *__eax, *__ebx, *__ecx, *__edx); return 1; } + +#endif /* __CPUID_H */ diff --git a/lib/include/cuda_wrappers/algorithm b/lib/include/cuda_wrappers/algorithm index 01af18360d..f14a0b00bb 100644 --- a/lib/include/cuda_wrappers/algorithm +++ b/lib/include/cuda_wrappers/algorithm @@ -1,4 +1,4 @@ -/*===---- complex - CUDA wrapper for <algorithm> ----------------------------=== +/*===---- algorithm - CUDA wrapper for <algorithm> -------------------------=== * * Permission is hereby granted, free of charge, to any person obtaining a copy * of this software and associated documentation files (the "Software"), to deal diff --git a/lib/include/cuda_wrappers/new b/lib/include/cuda_wrappers/new index f49811c5a5..d5fb3b7011 100644 --- a/lib/include/cuda_wrappers/new +++ b/lib/include/cuda_wrappers/new @@ -1,4 +1,4 @@ -/*===---- complex - CUDA wrapper for <new> ------------------------------=== +/*===---- new - CUDA wrapper for <new> -------------------------------------=== * * Permission is hereby granted, free of charge, to any person obtaining a copy * of this software and associated documentation files (the "Software"), to deal @@ -26,6 +26,13 @@ #include_next <new> +#if !defined(__device__) +// The header has been included too early from the standard C++ library +// and CUDA-specific macros are not available yet. +// Undo the include guard and try again later. +#undef __CLANG_CUDA_WRAPPERS_NEW +#else + #pragma push_macro("CUDA_NOEXCEPT") #if __cplusplus >= 201103L #define CUDA_NOEXCEPT noexcept @@ -95,4 +102,5 @@ __device__ inline void operator delete[](void *, void *) CUDA_NOEXCEPT {} #pragma pop_macro("CUDA_NOEXCEPT") +#endif // __device__ #endif // include guard diff --git a/lib/include/emmintrin.h b/lib/include/emmintrin.h index 73a777b107..bb759721fa 100644 --- a/lib/include/emmintrin.h +++ b/lib/include/emmintrin.h @@ -4025,7 +4025,7 @@ _mm_storeu_si128(__m128i_u *__p, __m128i __b) /// /// \param __p /// A pointer to a 64-bit memory location. The address of the memory -/// location does not have to be algned. +/// location does not have to be aligned. /// \param __b /// A 128-bit integer vector containing the value to be stored. static __inline__ void __DEFAULT_FN_ATTRS diff --git a/lib/include/gfniintrin.h b/lib/include/gfniintrin.h index 9bff0fcb60..11a321b7c9 100644 --- a/lib/include/gfniintrin.h +++ b/lib/include/gfniintrin.h @@ -14,38 +14,56 @@ #ifndef __GFNIINTRIN_H #define __GFNIINTRIN_H +/* Default attributes for simple form (no masking). */ +#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("gfni"), __min_vector_width__(128))) + +/* Default attributes for YMM unmasked form. */ +#define __DEFAULT_FN_ATTRS_Y __attribute__((__always_inline__, __nodebug__, __target__("avx,gfni"), __min_vector_width__(256))) + +/* Default attributes for ZMM forms. */ +#define __DEFAULT_FN_ATTRS_Z __attribute__((__always_inline__, __nodebug__, __target__("avx512bw,gfni"), __min_vector_width__(512))) + +/* Default attributes for VLX forms. */ +#define __DEFAULT_FN_ATTRS_VL128 __attribute__((__always_inline__, __nodebug__, __target__("avx512bw,avx512vl,gfni"), __min_vector_width__(128))) +#define __DEFAULT_FN_ATTRS_VL256 __attribute__((__always_inline__, __nodebug__, __target__("avx512bw,avx512vl,gfni"), __min_vector_width__(256))) #define _mm_gf2p8affineinv_epi64_epi8(A, B, I) \ (__m128i)__builtin_ia32_vgf2p8affineinvqb_v16qi((__v16qi)(__m128i)(A), \ (__v16qi)(__m128i)(B), \ (char)(I)) -#define _mm_mask_gf2p8affineinv_epi64_epi8(S, U, A, B, I) \ - (__m128i)__builtin_ia32_selectb_128((__mmask16)(U), \ - (__v16qi)_mm_gf2p8affineinv_epi64_epi8(A, B, I), \ - (__v16qi)(__m128i)(S)) - - -#define _mm_maskz_gf2p8affineinv_epi64_epi8(U, A, B, I) \ - (__m128i)_mm_mask_gf2p8affineinv_epi64_epi8((__m128i)_mm_setzero_si128(), \ - U, A, B, I) +#define _mm_gf2p8affine_epi64_epi8(A, B, I) \ + (__m128i)__builtin_ia32_vgf2p8affineqb_v16qi((__v16qi)(__m128i)(A), \ + (__v16qi)(__m128i)(B), \ + (char)(I)) +static __inline__ __m128i __DEFAULT_FN_ATTRS +_mm_gf2p8mul_epi8(__m128i __A, __m128i __B) +{ + return (__m128i) __builtin_ia32_vgf2p8mulb_v16qi((__v16qi) __A, + (__v16qi) __B); +} +#ifdef __AVXINTRIN_H #define _mm256_gf2p8affineinv_epi64_epi8(A, B, I) \ (__m256i)__builtin_ia32_vgf2p8affineinvqb_v32qi((__v32qi)(__m256i)(A), \ (__v32qi)(__m256i)(B), \ (char)(I)) -#define _mm256_mask_gf2p8affineinv_epi64_epi8(S, U, A, B, I) \ - (__m256i)__builtin_ia32_selectb_256((__mmask32)(U), \ - (__v32qi)_mm256_gf2p8affineinv_epi64_epi8(A, B, I), \ - (__v32qi)(__m256i)(S)) - -#define _mm256_maskz_gf2p8affineinv_epi64_epi8(U, A, B, I) \ - (__m256i)_mm256_mask_gf2p8affineinv_epi64_epi8((__m256i)_mm256_setzero_si256(), \ - U, A, B, I) +#define _mm256_gf2p8affine_epi64_epi8(A, B, I) \ + (__m256i)__builtin_ia32_vgf2p8affineqb_v32qi((__v32qi)(__m256i)(A), \ + (__v32qi)(__m256i)(B), \ + (char)(I)) +static __inline__ __m256i __DEFAULT_FN_ATTRS_Y +_mm256_gf2p8mul_epi8(__m256i __A, __m256i __B) +{ + return (__m256i) __builtin_ia32_vgf2p8mulb_v32qi((__v32qi) __A, + (__v32qi) __B); +} +#endif /* __AVXINTRIN_H */ +#ifdef __AVX512BWINTRIN_H #define _mm512_gf2p8affineinv_epi64_epi8(A, B, I) \ (__m512i)__builtin_ia32_vgf2p8affineinvqb_v64qi((__v64qi)(__m512i)(A), \ (__v64qi)(__m512i)(B), \ @@ -60,27 +78,71 @@ (__m512i)_mm512_mask_gf2p8affineinv_epi64_epi8((__m512i)_mm512_setzero_si512(), \ U, A, B, I) -#define _mm_gf2p8affine_epi64_epi8(A, B, I) \ - (__m128i)__builtin_ia32_vgf2p8affineqb_v16qi((__v16qi)(__m128i)(A), \ - (__v16qi)(__m128i)(B), \ +#define _mm512_gf2p8affine_epi64_epi8(A, B, I) \ + (__m512i)__builtin_ia32_vgf2p8affineqb_v64qi((__v64qi)(__m512i)(A), \ + (__v64qi)(__m512i)(B), \ (char)(I)) +#define _mm512_mask_gf2p8affine_epi64_epi8(S, U, A, B, I) \ + (__m512i)__builtin_ia32_selectb_512((__mmask64)(U), \ + (__v64qi)_mm512_gf2p8affine_epi64_epi8(A, B, I), \ + (__v64qi)(__m512i)(S)) + +#define _mm512_maskz_gf2p8affine_epi64_epi8(U, A, B, I) \ + (__m512i)_mm512_mask_gf2p8affine_epi64_epi8((__m512i)_mm512_setzero_si512(), \ + U, A, B, I) + +static __inline__ __m512i __DEFAULT_FN_ATTRS_Z +_mm512_gf2p8mul_epi8(__m512i __A, __m512i __B) +{ + return (__m512i) __builtin_ia32_vgf2p8mulb_v64qi((__v64qi) __A, + (__v64qi) __B); +} + +static __inline__ __m512i __DEFAULT_FN_ATTRS_Z +_mm512_mask_gf2p8mul_epi8(__m512i __S, __mmask64 __U, __m512i __A, __m512i __B) +{ + return (__m512i) __builtin_ia32_selectb_512(__U, + (__v64qi) _mm512_gf2p8mul_epi8(__A, __B), + (__v64qi) __S); +} + +static __inline__ __m512i __DEFAULT_FN_ATTRS_Z +_mm512_maskz_gf2p8mul_epi8(__mmask64 __U, __m512i __A, __m512i __B) +{ + return _mm512_mask_gf2p8mul_epi8((__m512i)_mm512_setzero_si512(), + __U, __A, __B); +} +#endif /* __AVX512BWINTRIN_H */ + +#ifdef __AVX512VLBWINTRIN_H +#define _mm_mask_gf2p8affineinv_epi64_epi8(S, U, A, B, I) \ + (__m128i)__builtin_ia32_selectb_128((__mmask16)(U), \ + (__v16qi)_mm_gf2p8affineinv_epi64_epi8(A, B, I), \ + (__v16qi)(__m128i)(S)) + +#define _mm_maskz_gf2p8affineinv_epi64_epi8(U, A, B, I) \ + (__m128i)_mm_mask_gf2p8affineinv_epi64_epi8((__m128i)_mm_setzero_si128(), \ + U, A, B, I) + +#define _mm256_mask_gf2p8affineinv_epi64_epi8(S, U, A, B, I) \ + (__m256i)__builtin_ia32_selectb_256((__mmask32)(U), \ + (__v32qi)_mm256_gf2p8affineinv_epi64_epi8(A, B, I), \ + (__v32qi)(__m256i)(S)) + +#define _mm256_maskz_gf2p8affineinv_epi64_epi8(U, A, B, I) \ + (__m256i)_mm256_mask_gf2p8affineinv_epi64_epi8((__m256i)_mm256_setzero_si256(), \ + U, A, B, I) + #define _mm_mask_gf2p8affine_epi64_epi8(S, U, A, B, I) \ (__m128i)__builtin_ia32_selectb_128((__mmask16)(U), \ (__v16qi)_mm_gf2p8affine_epi64_epi8(A, B, I), \ (__v16qi)(__m128i)(S)) - #define _mm_maskz_gf2p8affine_epi64_epi8(U, A, B, I) \ (__m128i)_mm_mask_gf2p8affine_epi64_epi8((__m128i)_mm_setzero_si128(), \ U, A, B, I) - -#define _mm256_gf2p8affine_epi64_epi8(A, B, I) \ - (__m256i)__builtin_ia32_vgf2p8affineqb_v32qi((__v32qi)(__m256i)(A), \ - (__v32qi)(__m256i)(B), \ - (char)(I)) - #define _mm256_mask_gf2p8affine_epi64_epi8(S, U, A, B, I) \ (__m256i)__builtin_ia32_selectb_256((__mmask32)(U), \ (__v32qi)_mm256_gf2p8affine_epi64_epi8(A, B, I), \ @@ -90,41 +152,6 @@ (__m256i)_mm256_mask_gf2p8affine_epi64_epi8((__m256i)_mm256_setzero_si256(), \ U, A, B, I) - -#define _mm512_gf2p8affine_epi64_epi8(A, B, I) \ - (__m512i)__builtin_ia32_vgf2p8affineqb_v64qi((__v64qi)(__m512i)(A), \ - (__v64qi)(__m512i)(B), \ - (char)(I)) - -#define _mm512_mask_gf2p8affine_epi64_epi8(S, U, A, B, I) \ - (__m512i)__builtin_ia32_selectb_512((__mmask64)(U), \ - (__v64qi)_mm512_gf2p8affine_epi64_epi8(A, B, I), \ - (__v64qi)(__m512i)(S)) - -#define _mm512_maskz_gf2p8affine_epi64_epi8(U, A, B, I) \ - (__m512i)_mm512_mask_gf2p8affine_epi64_epi8((__m512i)_mm512_setzero_si512(), \ - U, A, B, I) - -/* Default attributes for simple form (no masking). */ -#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("gfni"), __min_vector_width__(128))) - -/* Default attributes for YMM unmasked form. */ -#define __DEFAULT_FN_ATTRS_Y __attribute__((__always_inline__, __nodebug__, __target__("avx,gfni"), __min_vector_width__(256))) - -/* Default attributes for ZMM forms. */ -#define __DEFAULT_FN_ATTRS_Z __attribute__((__always_inline__, __nodebug__, __target__("avx512bw,gfni"), __min_vector_width__(512))) - -/* Default attributes for VLX forms. */ -#define __DEFAULT_FN_ATTRS_VL128 __attribute__((__always_inline__, __nodebug__, __target__("avx512bw,avx512vl,gfni"), __min_vector_width__(128))) -#define __DEFAULT_FN_ATTRS_VL256 __attribute__((__always_inline__, __nodebug__, __target__("avx512bw,avx512vl,gfni"), __min_vector_width__(256))) - -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_gf2p8mul_epi8(__m128i __A, __m128i __B) -{ - return (__m128i) __builtin_ia32_vgf2p8mulb_v16qi((__v16qi) __A, - (__v16qi) __B); -} - static __inline__ __m128i __DEFAULT_FN_ATTRS_VL128 _mm_mask_gf2p8mul_epi8(__m128i __S, __mmask16 __U, __m128i __A, __m128i __B) { @@ -140,13 +167,6 @@ _mm_maskz_gf2p8mul_epi8(__mmask16 __U, __m128i __A, __m128i __B) __U, __A, __B); } -static __inline__ __m256i __DEFAULT_FN_ATTRS_Y -_mm256_gf2p8mul_epi8(__m256i __A, __m256i __B) -{ - return (__m256i) __builtin_ia32_vgf2p8mulb_v32qi((__v32qi) __A, - (__v32qi) __B); -} - static __inline__ __m256i __DEFAULT_FN_ATTRS_VL256 _mm256_mask_gf2p8mul_epi8(__m256i __S, __mmask32 __U, __m256i __A, __m256i __B) { @@ -161,28 +181,7 @@ _mm256_maskz_gf2p8mul_epi8(__mmask32 __U, __m256i __A, __m256i __B) return _mm256_mask_gf2p8mul_epi8((__m256i)_mm256_setzero_si256(), __U, __A, __B); } - -static __inline__ __m512i __DEFAULT_FN_ATTRS_Z -_mm512_gf2p8mul_epi8(__m512i __A, __m512i __B) -{ - return (__m512i) __builtin_ia32_vgf2p8mulb_v64qi((__v64qi) __A, - (__v64qi) __B); -} - -static __inline__ __m512i __DEFAULT_FN_ATTRS_Z -_mm512_mask_gf2p8mul_epi8(__m512i __S, __mmask64 __U, __m512i __A, __m512i __B) -{ - return (__m512i) __builtin_ia32_selectb_512(__U, - (__v64qi) _mm512_gf2p8mul_epi8(__A, __B), - (__v64qi) __S); -} - -static __inline__ __m512i __DEFAULT_FN_ATTRS_Z -_mm512_maskz_gf2p8mul_epi8(__mmask64 __U, __m512i __A, __m512i __B) -{ - return _mm512_mask_gf2p8mul_epi8((__m512i)_mm512_setzero_si512(), - __U, __A, __B); -} +#endif /* __AVX512VLBWINTRIN_H */ #undef __DEFAULT_FN_ATTRS #undef __DEFAULT_FN_ATTRS_Y diff --git a/lib/include/hresetintrin.h b/lib/include/hresetintrin.h new file mode 100644 index 0000000000..13e31a2e03 --- /dev/null +++ b/lib/include/hresetintrin.h @@ -0,0 +1,49 @@ +/*===---------------- hresetintrin.h - HRESET intrinsics -------------------=== + * + * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. + * See https://llvm.org/LICENSE.txt for license information. + * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception + * + *===-----------------------------------------------------------------------=== + */ +#ifndef __X86GPRINTRIN_H +#error "Never use <hresetintrin.h> directly; include <x86gprintrin.h> instead." +#endif + +#ifndef __HRESETINTRIN_H +#define __HRESETINTRIN_H + +#if __has_extension(gnu_asm) + +/* Define the default attributes for the functions in this file. */ +#define __DEFAULT_FN_ATTRS \ + __attribute__((__always_inline__, __nodebug__, __target__("hreset"))) + +/// Provides a hint to the processor to selectively reset the prediction +/// history of the current logical processor specified by a 32-bit integer +/// value \a __eax. +/// +/// This intrinsic corresponds to the <c> HRESET </c> instruction. +/// +/// \operation +/// IF __eax == 0 +/// // nop +/// ELSE +/// FOR i := 0 to 31 +/// IF __eax[i] +/// ResetPredictionFeature(i) +/// FI +/// ENDFOR +/// FI +/// \endoperation +static __inline void __DEFAULT_FN_ATTRS +_hreset(int __eax) +{ + __asm__ ("hreset $0" :: "a"(__eax)); +} + +#undef __DEFAULT_FN_ATTRS + +#endif /* __has_extension(gnu_asm) */ + +#endif /* __HRESETINTRIN_H */ diff --git a/lib/include/ia32intrin.h b/lib/include/ia32intrin.h index 79b7f0655c..00138effd5 100644 --- a/lib/include/ia32intrin.h +++ b/lib/include/ia32intrin.h @@ -14,6 +14,18 @@ #ifndef __IA32INTRIN_H #define __IA32INTRIN_H +/* Define the default attributes for the functions in this file. */ +#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__)) +#define __DEFAULT_FN_ATTRS_SSE42 __attribute__((__always_inline__, __nodebug__, __target__("sse4.2"))) + +#if defined(__cplusplus) && (__cplusplus >= 201103L) +#define __DEFAULT_FN_ATTRS_CAST __attribute__((__always_inline__)) constexpr +#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS constexpr +#else +#define __DEFAULT_FN_ATTRS_CAST __attribute__((__always_inline__)) +#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS +#endif + /** Find the first set bit starting from the lsb. Result is undefined if * input is 0. * @@ -26,7 +38,7 @@ * A 32-bit integer operand. * \returns A 32-bit integer containing the bit number. */ -static __inline__ int __attribute__((__always_inline__, __nodebug__)) +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR __bsfd(int __A) { return __builtin_ctz(__A); } @@ -43,7 +55,7 @@ __bsfd(int __A) { * A 32-bit integer operand. * \returns A 32-bit integer containing the bit number. */ -static __inline__ int __attribute__((__always_inline__, __nodebug__)) +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR __bsrd(int __A) { return 31 - __builtin_clz(__A); } @@ -59,12 +71,12 @@ __bsrd(int __A) { * A 32-bit integer operand. * \returns A 32-bit integer containing the swapped bytes. */ -static __inline__ int __attribute__((__always_inline__, __nodebug__)) +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR __bswapd(int __A) { return __builtin_bswap32(__A); } -static __inline__ int __attribute__((__always_inline__, __nodebug__)) +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR _bswap(int __A) { return __builtin_bswap32(__A); } @@ -85,7 +97,7 @@ _bswap(int __A) { * A 64-bit integer operand. * \returns A 32-bit integer containing the bit number. */ -static __inline__ int __attribute__((__always_inline__, __nodebug__)) +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR __bsfq(long long __A) { return __builtin_ctzll(__A); } @@ -102,7 +114,7 @@ __bsfq(long long __A) { * A 64-bit integer operand. * \returns A 32-bit integer containing the bit number. */ -static __inline__ int __attribute__((__always_inline__, __nodebug__)) +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR __bsrq(long long __A) { return 63 - __builtin_clzll(__A); } @@ -118,7 +130,7 @@ __bsrq(long long __A) { * A 64-bit integer operand. * \returns A 64-bit integer containing the swapped bytes. */ -static __inline__ long long __attribute__((__always_inline__, __nodebug__)) +static __inline__ long long __DEFAULT_FN_ATTRS_CONSTEXPR __bswapq(long long __A) { return __builtin_bswap64(__A); } @@ -138,7 +150,7 @@ __bswapq(long long __A) { * \returns A 32-bit integer containing the number of bits with value 1 in the * source operand. */ -static __inline__ int __attribute__((__always_inline__, __nodebug__)) +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR __popcntd(unsigned int __A) { return __builtin_popcount(__A); @@ -159,7 +171,7 @@ __popcntd(unsigned int __A) * \returns A 64-bit integer containing the number of bits with value 1 in the * source operand. */ -static __inline__ long long __attribute__((__always_inline__, __nodebug__)) +static __inline__ long long __DEFAULT_FN_ATTRS_CONSTEXPR __popcntq(unsigned long long __A) { return __builtin_popcountll(__A); @@ -169,26 +181,26 @@ __popcntq(unsigned long long __A) #endif /* __x86_64__ */ #ifdef __x86_64__ -static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned long long __DEFAULT_FN_ATTRS __readeflags(void) { return __builtin_ia32_readeflags_u64(); } -static __inline__ void __attribute__((__always_inline__, __nodebug__)) +static __inline__ void __DEFAULT_FN_ATTRS __writeeflags(unsigned long long __f) { __builtin_ia32_writeeflags_u64(__f); } #else /* !__x86_64__ */ -static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned int __DEFAULT_FN_ATTRS __readeflags(void) { return __builtin_ia32_readeflags_u32(); } -static __inline__ void __attribute__((__always_inline__, __nodebug__)) +static __inline__ void __DEFAULT_FN_ATTRS __writeeflags(unsigned int __f) { __builtin_ia32_writeeflags_u32(__f); @@ -205,11 +217,9 @@ __writeeflags(unsigned int __f) * A 32-bit float value. * \returns a 32-bit unsigned integer containing the converted value. */ -static __inline__ unsigned int __attribute__((__always_inline__)) +static __inline__ unsigned int __DEFAULT_FN_ATTRS_CAST _castf32_u32(float __A) { - unsigned int D; - __builtin_memcpy(&D, &__A, sizeof(__A)); - return D; + return __builtin_bit_cast(unsigned int, __A); } /** Cast a 64-bit float value to a 64-bit unsigned integer value @@ -222,11 +232,9 @@ _castf32_u32(float __A) { * A 64-bit float value. * \returns a 64-bit unsigned integer containing the converted value. */ -static __inline__ unsigned long long __attribute__((__always_inline__)) +static __inline__ unsigned long long __DEFAULT_FN_ATTRS_CAST _castf64_u64(double __A) { - unsigned long long D; - __builtin_memcpy(&D, &__A, sizeof(__A)); - return D; + return __builtin_bit_cast(unsigned long long, __A); } /** Cast a 32-bit unsigned integer value to a 32-bit float value @@ -239,11 +247,9 @@ _castf64_u64(double __A) { * A 32-bit unsigned integer value. * \returns a 32-bit float value containing the converted value. */ -static __inline__ float __attribute__((__always_inline__)) +static __inline__ float __DEFAULT_FN_ATTRS_CAST _castu32_f32(unsigned int __A) { - float D; - __builtin_memcpy(&D, &__A, sizeof(__A)); - return D; + return __builtin_bit_cast(float, __A); } /** Cast a 64-bit unsigned integer value to a 64-bit float value @@ -256,11 +262,9 @@ _castu32_f32(unsigned int __A) { * A 64-bit unsigned integer value. * \returns a 64-bit float value containing the converted value. */ -static __inline__ double __attribute__((__always_inline__)) +static __inline__ double __DEFAULT_FN_ATTRS_CAST _castu64_f64(unsigned long long __A) { - double D; - __builtin_memcpy(&D, &__A, sizeof(__A)); - return D; + return __builtin_bit_cast(double, __A); } /** Adds the unsigned integer operand to the CRC-32C checksum of the @@ -278,7 +282,7 @@ _castu64_f64(unsigned long long __A) { * \returns The result of adding operand \a __C to the CRC-32C checksum of * operand \a __D. */ -static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__, __target__("sse4.2"))) +static __inline__ unsigned int __DEFAULT_FN_ATTRS_SSE42 __crc32b(unsigned int __C, unsigned char __D) { return __builtin_ia32_crc32qi(__C, __D); @@ -299,7 +303,7 @@ __crc32b(unsigned int __C, unsigned char __D) * \returns The result of adding operand \a __C to the CRC-32C checksum of * operand \a __D. */ -static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__, __target__("sse4.2"))) +static __inline__ unsigned int __DEFAULT_FN_ATTRS_SSE42 __crc32w(unsigned int __C, unsigned short __D) { return __builtin_ia32_crc32hi(__C, __D); @@ -320,7 +324,7 @@ __crc32w(unsigned int __C, unsigned short __D) * \returns The result of adding operand \a __C to the CRC-32C checksum of * operand \a __D. */ -static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__, __target__("sse4.2"))) +static __inline__ unsigned int __DEFAULT_FN_ATTRS_SSE42 __crc32d(unsigned int __C, unsigned int __D) { return __builtin_ia32_crc32si(__C, __D); @@ -342,20 +346,20 @@ __crc32d(unsigned int __C, unsigned int __D) * \returns The result of adding operand \a __C to the CRC-32C checksum of * operand \a __D. */ -static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__, __target__("sse4.2"))) +static __inline__ unsigned long long __DEFAULT_FN_ATTRS_SSE42 __crc32q(unsigned long long __C, unsigned long long __D) { return __builtin_ia32_crc32di(__C, __D); } #endif /* __x86_64__ */ -static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned long long __DEFAULT_FN_ATTRS __rdpmc(int __A) { return __builtin_ia32_rdpmc(__A); } /* __rdtscp */ -static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned long long __DEFAULT_FN_ATTRS __rdtscp(unsigned int *__A) { return __builtin_ia32_rdtscp(__A); } @@ -364,48 +368,48 @@ __rdtscp(unsigned int *__A) { #define _rdpmc(A) __rdpmc(A) -static __inline__ void __attribute__((__always_inline__, __nodebug__)) +static __inline__ void __DEFAULT_FN_ATTRS _wbinvd(void) { __builtin_ia32_wbinvd(); } -static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned char __DEFAULT_FN_ATTRS_CONSTEXPR __rolb(unsigned char __X, int __C) { return __builtin_rotateleft8(__X, __C); } -static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned char __DEFAULT_FN_ATTRS_CONSTEXPR __rorb(unsigned char __X, int __C) { return __builtin_rotateright8(__X, __C); } -static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned short __DEFAULT_FN_ATTRS_CONSTEXPR __rolw(unsigned short __X, int __C) { return __builtin_rotateleft16(__X, __C); } -static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned short __DEFAULT_FN_ATTRS_CONSTEXPR __rorw(unsigned short __X, int __C) { return __builtin_rotateright16(__X, __C); } -static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned int __DEFAULT_FN_ATTRS_CONSTEXPR __rold(unsigned int __X, int __C) { return __builtin_rotateleft32(__X, __C); } -static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned int __DEFAULT_FN_ATTRS_CONSTEXPR __rord(unsigned int __X, int __C) { return __builtin_rotateright32(__X, __C); } #ifdef __x86_64__ -static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned long long __DEFAULT_FN_ATTRS_CONSTEXPR __rolq(unsigned long long __X, int __C) { return __builtin_rotateleft64(__X, __C); } -static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__)) +static __inline__ unsigned long long __DEFAULT_FN_ATTRS_CONSTEXPR __rorq(unsigned long long __X, int __C) { return __builtin_rotateright64(__X, __C); } @@ -429,4 +433,9 @@ __rorq(unsigned long long __X, int __C) { #define _rotwl(a,b) __rolw((a), (b)) #define _rotwr(a,b) __rorw((a), (b)) +#undef __DEFAULT_FN_ATTRS +#undef __DEFAULT_FN_ATTRS_CAST +#undef __DEFAULT_FN_ATTRS_SSE42 +#undef __DEFAULT_FN_ATTRS_CONSTEXPR + #endif /* __IA32INTRIN_H */ diff --git a/lib/include/immintrin.h b/lib/include/immintrin.h index e9dff2310f..22f7a520c9 100644 --- a/lib/include/immintrin.h +++ b/lib/include/immintrin.h @@ -10,6 +10,8 @@ #ifndef __IMMINTRIN_H #define __IMMINTRIN_H +#include <x86gprintrin.h> + #if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ defined(__MMX__) #include <mmintrin.h> @@ -144,6 +146,11 @@ #endif #if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ + defined(__AVXVNNI__) +#include <avxvnniintrin.h> +#endif + +#if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ defined(__AVX512DQ__) #include <avx512dqintrin.h> #endif @@ -472,6 +479,11 @@ _storebe_i64(void * __P, long long __D) { #endif #if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ + defined(__KL__) || defined(__WIDEKL__) +#include <keylockerintrin.h> +#endif + +#if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ defined(__AMXTILE__) || defined(__AMXINT8__) || defined(__AMXBF16__) #include <amxintrin.h> #endif diff --git a/lib/include/intrin.h b/lib/include/intrin.h index 871b47ca82..a78b96997d 100644 --- a/lib/include/intrin.h +++ b/lib/include/intrin.h @@ -57,16 +57,11 @@ void __addfsbyte(unsigned long, unsigned char); void __addfsdword(unsigned long, unsigned long); void __addfsword(unsigned long, unsigned short); void __code_seg(const char *); -static __inline__ void __cpuid(int[4], int); -static __inline__ void __cpuidex(int[4], int, int); -static __inline__ __int64 __emul(int, int); -static __inline__ unsigned __int64 __emulu(unsigned int, unsigned int); unsigned int __getcallerseflags(void); -static __inline__ void __halt(void); unsigned char __inbyte(unsigned short); void __inbytestring(unsigned short, unsigned char *, unsigned long); @@ -82,13 +77,9 @@ void __inwordstring(unsigned short, unsigned short *, unsigned long); void __lidt(void *); unsigned __int64 __ll_lshift(unsigned __int64, int); __int64 __ll_rshift(__int64, int); -static __inline__ void __movsb(unsigned char *, unsigned char const *, size_t); -static __inline__ void __movsd(unsigned long *, unsigned long const *, size_t); -static __inline__ void __movsw(unsigned short *, unsigned short const *, size_t); -static __inline__ void __nop(void); void __nvreg_restore_fence(void); void __nvreg_save_fence(void); @@ -105,23 +96,16 @@ unsigned long __readcr4(void); unsigned long __readcr8(void); unsigned int __readdr(unsigned int); #ifdef __i386__ -static __inline__ unsigned char __readfsbyte(unsigned long); -static __inline__ unsigned __int64 __readfsqword(unsigned long); -static __inline__ unsigned short __readfsword(unsigned long); #endif -static __inline__ unsigned __int64 __readmsr(unsigned long); unsigned __int64 __readpmc(unsigned long); unsigned long __segmentlimit(unsigned long); void __sidt(void *); -static __inline__ void __stosb(unsigned char *, unsigned char, size_t); -static __inline__ void __stosd(unsigned long *, unsigned long, size_t); -static __inline__ void __stosw(unsigned short *, unsigned short, size_t); void __svm_clgi(void); void __svm_invlpga(void *, int); @@ -136,7 +120,6 @@ void __vmx_off(void); void __vmx_vmptrst(unsigned __int64 *); void __wbinvd(void); void __writecr0(unsigned int); -static __inline__ void __writecr3(unsigned __INTPTR_TYPE__); void __writecr4(unsigned int); void __writecr8(unsigned int); @@ -146,11 +129,8 @@ void __writefsdword(unsigned long, unsigned long); void __writefsqword(unsigned long, unsigned __int64); void __writefsword(unsigned long, unsigned short); void __writemsr(unsigned long, unsigned __int64); -static __inline__ void *_AddressOfReturnAddress(void); -static __inline__ unsigned char _BitScanForward(unsigned long *_Index, unsigned long _Mask); -static __inline__ unsigned char _BitScanReverse(unsigned long *_Index, unsigned long _Mask); unsigned char _bittest(long const *, long); unsigned char _bittestandcomplement(long *, long); @@ -169,12 +149,10 @@ long _InterlockedExchangeAdd_HLEAcquire(long volatile *, long); long _InterlockedExchangeAdd_HLERelease(long volatile *, long); __int64 _InterlockedExchangeAdd64_HLEAcquire(__int64 volatile *, __int64); __int64 _InterlockedExchangeAdd64_HLERelease(__int64 volatile *, __int64); -static __inline__ void -__attribute__((__deprecated__("use other intrinsics or C++11 atomics instead"))) -_ReadBarrier(void); -static __inline__ void -__attribute__((__deprecated__("use other intrinsics or C++11 atomics instead"))) -_ReadWriteBarrier(void); +void __attribute__((__deprecated__( + "use other intrinsics or C++11 atomics instead"))) _ReadBarrier(void); +void __attribute__((__deprecated__( + "use other intrinsics or C++11 atomics instead"))) _ReadWriteBarrier(void); unsigned int _rorx_u32(unsigned int, const unsigned int); int _sarx_i32(int, unsigned int); #if __STDC_HOSTED__ @@ -185,9 +163,8 @@ unsigned int _shrx_u32(unsigned int, unsigned int); void _Store_HLERelease(long volatile *, long); void _Store64_HLERelease(__int64 volatile *, __int64); void _StorePointer_HLERelease(void *volatile *, void *); -static __inline__ void -__attribute__((__deprecated__("use other intrinsics or C++11 atomics instead"))) -_WriteBarrier(void); +void __attribute__((__deprecated__( + "use other intrinsics or C++11 atomics instead"))) _WriteBarrier(void); unsigned __int32 xbegin(void); void _xend(void); @@ -197,19 +174,14 @@ void __addgsbyte(unsigned long, unsigned char); void __addgsdword(unsigned long, unsigned long); void __addgsqword(unsigned long, unsigned __int64); void __addgsword(unsigned long, unsigned short); -static __inline__ void __faststorefence(void); void __incgsbyte(unsigned long); void __incgsdword(unsigned long); void __incgsqword(unsigned long); void __incgsword(unsigned long); -static __inline__ void __movsq(unsigned long long *, unsigned long long const *, size_t); -static __inline__ unsigned char __readgsbyte(unsigned long); -static __inline__ unsigned long __readgsdword(unsigned long); -static __inline__ unsigned __int64 __readgsqword(unsigned long); unsigned short __readgsword(unsigned long); unsigned __int64 __shiftleft128(unsigned __int64 _LowPart, @@ -218,7 +190,6 @@ unsigned __int64 __shiftleft128(unsigned __int64 _LowPart, unsigned __int64 __shiftright128(unsigned __int64 _LowPart, unsigned __int64 _HighPart, unsigned char _Shift); -static __inline__ void __stosq(unsigned __int64 *, unsigned __int64, size_t); unsigned char __vmx_on(unsigned __int64 *); unsigned char __vmx_vmclear(unsigned __int64 *); @@ -243,10 +214,6 @@ unsigned char _interlockedbittestandreset64(__int64 volatile *, __int64); unsigned char _interlockedbittestandset64(__int64 volatile *, __int64); long _InterlockedCompareExchange_np(long volatile *_Destination, long _Exchange, long _Comparand); -unsigned char _InterlockedCompareExchange128(__int64 volatile *_Destination, - __int64 _ExchangeHigh, - __int64 _ExchangeLow, - __int64 *_CompareandResult); unsigned char _InterlockedCompareExchange128_np(__int64 volatile *_Destination, __int64 _ExchangeHigh, __int64 _ExchangeLow, @@ -269,13 +236,9 @@ unsigned __int64 _rorx_u64(unsigned __int64, const unsigned int); __int64 _sarx_i64(__int64, unsigned int); unsigned __int64 _shlx_u64(unsigned __int64, unsigned int); unsigned __int64 _shrx_u64(unsigned __int64, unsigned int); -static __inline__ __int64 __mulh(__int64, __int64); -static __inline__ unsigned __int64 __umulh(unsigned __int64, unsigned __int64); -static __inline__ __int64 _mul128(__int64, __int64, __int64*); -static __inline__ unsigned __int64 _umul128(unsigned __int64, unsigned __int64, unsigned __int64*); @@ -284,29 +247,19 @@ unsigned __int64 _umul128(unsigned __int64, #if defined(__x86_64__) || defined(__arm__) || defined(__aarch64__) -static __inline__ unsigned char _BitScanForward64(unsigned long *_Index, unsigned __int64 _Mask); -static __inline__ unsigned char _BitScanReverse64(unsigned long *_Index, unsigned __int64 _Mask); #endif #if defined(__i386__) || defined(__x86_64__) || defined(__arm__) || defined(__aarch64__) -static __inline__ __int64 _InterlockedDecrement64(__int64 volatile *_Addend); -static __inline__ __int64 _InterlockedExchange64(__int64 volatile *_Target, __int64 _Value); -static __inline__ __int64 _InterlockedExchangeAdd64(__int64 volatile *_Addend, __int64 _Value); -static __inline__ __int64 _InterlockedExchangeSub64(__int64 volatile *_Subend, __int64 _Value); -static __inline__ __int64 _InterlockedIncrement64(__int64 volatile *_Addend); -static __inline__ __int64 _InterlockedOr64(__int64 volatile *_Value, __int64 _Mask); -static __inline__ __int64 _InterlockedXor64(__int64 volatile *_Value, __int64 _Mask); -static __inline__ __int64 _InterlockedAnd64(__int64 volatile *_Value, __int64 _Mask); #endif @@ -470,45 +423,81 @@ __int64 _InterlockedCompareExchange64_nf(__int64 volatile *_Destination, __int64 _InterlockedCompareExchange64_rel(__int64 volatile *_Destination, __int64 _Exchange, __int64 _Comparand); #endif +#if defined(__x86_64__) || defined(__aarch64__) +unsigned char _InterlockedCompareExchange128(__int64 volatile *_Destination, + __int64 _ExchangeHigh, + __int64 _ExchangeLow, + __int64 *_ComparandResult); +#endif +#if defined(__aarch64__) +unsigned char _InterlockedCompareExchange128_acq(__int64 volatile *_Destination, + __int64 _ExchangeHigh, + __int64 _ExchangeLow, + __int64 *_ComparandResult); +unsigned char _InterlockedCompareExchange128_nf(__int64 volatile *_Destination, + __int64 _ExchangeHigh, + __int64 _ExchangeLow, + __int64 *_ComparandResult); +unsigned char _InterlockedCompareExchange128_rel(__int64 volatile *_Destination, + __int64 _ExchangeHigh, + __int64 _ExchangeLow, + __int64 *_ComparandResult); +#endif /*----------------------------------------------------------------------------*\ |* movs, stos \*----------------------------------------------------------------------------*/ #if defined(__i386__) || defined(__x86_64__) -static __inline__ void __DEFAULT_FN_ATTRS -__movsb(unsigned char *__dst, unsigned char const *__src, size_t __n) { +static __inline__ void __DEFAULT_FN_ATTRS __movsb(unsigned char *__dst, + unsigned char const *__src, + size_t __n) { __asm__ __volatile__("rep movsb" : "+D"(__dst), "+S"(__src), "+c"(__n) : : "memory"); } -static __inline__ void __DEFAULT_FN_ATTRS -__movsd(unsigned long *__dst, unsigned long const *__src, size_t __n) { - __asm__ __volatile__("rep movsl" : "+D"(__dst), "+S"(__src), "+c"(__n) - : : "memory"); +static __inline__ void __DEFAULT_FN_ATTRS __movsd(unsigned long *__dst, + unsigned long const *__src, + size_t __n) { + __asm__ __volatile__("rep movsl" + : "+D"(__dst), "+S"(__src), "+c"(__n) + : + : "memory"); } -static __inline__ void __DEFAULT_FN_ATTRS -__movsw(unsigned short *__dst, unsigned short const *__src, size_t __n) { - __asm__ __volatile__("rep movsw" : "+D"(__dst), "+S"(__src), "+c"(__n) - : : "memory"); +static __inline__ void __DEFAULT_FN_ATTRS __movsw(unsigned short *__dst, + unsigned short const *__src, + size_t __n) { + __asm__ __volatile__("rep movsw" + : "+D"(__dst), "+S"(__src), "+c"(__n) + : + : "memory"); } -static __inline__ void __DEFAULT_FN_ATTRS -__stosd(unsigned long *__dst, unsigned long __x, size_t __n) { - __asm__ __volatile__("rep stosl" : "+D"(__dst), "+c"(__n) : "a"(__x) +static __inline__ void __DEFAULT_FN_ATTRS __stosd(unsigned long *__dst, + unsigned long __x, + size_t __n) { + __asm__ __volatile__("rep stosl" + : "+D"(__dst), "+c"(__n) + : "a"(__x) : "memory"); } -static __inline__ void __DEFAULT_FN_ATTRS -__stosw(unsigned short *__dst, unsigned short __x, size_t __n) { - __asm__ __volatile__("rep stosw" : "+D"(__dst), "+c"(__n) : "a"(__x) +static __inline__ void __DEFAULT_FN_ATTRS __stosw(unsigned short *__dst, + unsigned short __x, + size_t __n) { + __asm__ __volatile__("rep stosw" + : "+D"(__dst), "+c"(__n) + : "a"(__x) : "memory"); } #endif #ifdef __x86_64__ -static __inline__ void __DEFAULT_FN_ATTRS -__movsq(unsigned long long *__dst, unsigned long long const *__src, size_t __n) { - __asm__ __volatile__("rep movsq" : "+D"(__dst), "+S"(__src), "+c"(__n) - : : "memory"); +static __inline__ void __DEFAULT_FN_ATTRS __movsq( + unsigned long long *__dst, unsigned long long const *__src, size_t __n) { + __asm__ __volatile__("rep movsq" + : "+D"(__dst), "+S"(__src), "+c"(__n) + : + : "memory"); } -static __inline__ void __DEFAULT_FN_ATTRS -__stosq(unsigned __int64 *__dst, unsigned __int64 __x, size_t __n) { +static __inline__ void __DEFAULT_FN_ATTRS __stosq(unsigned __int64 *__dst, + unsigned __int64 __x, + size_t __n) { __asm__ __volatile__("rep stosq" : "+D"(__dst), "+c"(__n) : "a"(__x) : "memory"); } @@ -518,26 +507,25 @@ __stosq(unsigned __int64 *__dst, unsigned __int64 __x, size_t __n) { |* Misc \*----------------------------------------------------------------------------*/ #if defined(__i386__) || defined(__x86_64__) -static __inline__ void __DEFAULT_FN_ATTRS -__cpuid(int __info[4], int __level) { - __asm__ ("cpuid" : "=a"(__info[0]), "=b" (__info[1]), "=c"(__info[2]), "=d"(__info[3]) - : "a"(__level), "c"(0)); +static __inline__ void __DEFAULT_FN_ATTRS __cpuid(int __info[4], int __level) { + __asm__("cpuid" + : "=a"(__info[0]), "=b"(__info[1]), "=c"(__info[2]), "=d"(__info[3]) + : "a"(__level), "c"(0)); } -static __inline__ void __DEFAULT_FN_ATTRS -__cpuidex(int __info[4], int __level, int __ecx) { - __asm__ ("cpuid" : "=a"(__info[0]), "=b" (__info[1]), "=c"(__info[2]), "=d"(__info[3]) - : "a"(__level), "c"(__ecx)); +static __inline__ void __DEFAULT_FN_ATTRS __cpuidex(int __info[4], int __level, + int __ecx) { + __asm__("cpuid" + : "=a"(__info[0]), "=b"(__info[1]), "=c"(__info[2]), "=d"(__info[3]) + : "a"(__level), "c"(__ecx)); } -static __inline__ void __DEFAULT_FN_ATTRS -__halt(void) { - __asm__ volatile ("hlt"); +static __inline__ void __DEFAULT_FN_ATTRS __halt(void) { + __asm__ volatile("hlt"); } #endif #if defined(__i386__) || defined(__x86_64__) || defined(__aarch64__) -static __inline__ void __DEFAULT_FN_ATTRS -__nop(void) { - __asm__ volatile ("nop"); +static __inline__ void __DEFAULT_FN_ATTRS __nop(void) { + __asm__ volatile("nop"); } #endif @@ -574,8 +562,7 @@ __readmsr(unsigned long __register) { } #endif -static __inline__ unsigned __LPTRINT_TYPE__ __DEFAULT_FN_ATTRS -__readcr3(void) { +static __inline__ unsigned __LPTRINT_TYPE__ __DEFAULT_FN_ATTRS __readcr3(void) { unsigned __LPTRINT_TYPE__ __cr3_val; __asm__ __volatile__ ("mov %%cr3, %0" : "=r"(__cr3_val) : : "memory"); return __cr3_val; diff --git a/lib/include/keylockerintrin.h b/lib/include/keylockerintrin.h new file mode 100644 index 0000000000..c15d39c8e3 --- /dev/null +++ b/lib/include/keylockerintrin.h @@ -0,0 +1,506 @@ +/*===----------------- keylockerintrin.h - KL Intrinsics -------------------=== + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + * + *===-----------------------------------------------------------------------=== + */ + +#ifndef __IMMINTRIN_H +#error "Never use <keylockerintrin.h> directly; include <immintrin.h> instead." +#endif + +#ifndef _KEYLOCKERINTRIN_H +#define _KEYLOCKERINTRIN_H + +#if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ + defined(__KL__) + +/* Define the default attributes for the functions in this file. */ +#define __DEFAULT_FN_ATTRS \ + __attribute__((__always_inline__, __nodebug__, __target__("kl"),\ + __min_vector_width__(128))) + +/// Load internal wrapping key from __intkey, __enkey_lo and __enkey_hi. __ctl +/// will assigned to EAX, whch specifies the KeySource and whether backing up +/// the key is permitted. The 256-bit encryption key is loaded from the two +/// explicit operands (__enkey_lo and __enkey_hi). The 128-bit integrity key is +/// loaded from the implicit operand XMM0 which assigned by __intkey. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> LOADIWKEY </c> instructions. +/// +/// \operation +/// IF CPL > 0 // LOADKWKEY only allowed at ring 0 (supervisor mode) +/// GP (0) +/// FI +/// IF “LOADIWKEY exiting” VM execution control set +/// VMexit +/// FI +/// IF __ctl[4:1] > 1 // Reserved KeySource encoding used +/// GP (0) +/// FI +/// IF __ctl[31:5] != 0 // Reserved bit in __ctl is set +/// GP (0) +/// FI +/// IF __ctl[0] AND (CPUID.19H.ECX[0] == 0) // NoBackup is not supported on this part +/// GP (0) +/// FI +/// IF (__ctl[4:1] == 1) AND (CPUID.19H.ECX[1] == 0) // KeySource of 1 is not supported on this part +/// GP (0) +/// FI +/// IF (__ctl[4:1] == 0) // KeySource of 0. +/// IWKey.Encryption Key[127:0] := __enkey_hi[127:0]: +/// IWKey.Encryption Key[255:128] := __enkey_lo[127:0] +/// IWKey.IntegrityKey[127:0] := __intkey[127:0] +/// IWKey.NoBackup := __ctl[0] +/// IWKey.KeySource := __ctl[4:1] +/// ZF := 0 +/// ELSE // KeySource of 1. See RDSEED definition for details of randomness +/// IF HW_NRND_GEN.ready == 1 // Full-entropy random data from RDSEED was received +/// IWKey.Encryption Key[127:0] := __enkey_hi[127:0] XOR HW_NRND_GEN.data[127:0] +/// IWKey.Encryption Key[255:128] := __enkey_lo[127:0] XOR HW_NRND_GEN.data[255:128] +/// IWKey.Encryption Key[255:0] := __enkey_hi[127:0]:__enkey_lo[127:0] XOR HW_NRND_GEN.data[255:0] +/// IWKey.IntegrityKey[127:0] := __intkey[127:0] XOR HW_NRND_GEN.data[383:256] +/// IWKey.NoBackup := __ctl[0] +/// IWKey.KeySource := __ctl[4:1] +/// ZF := 0 +/// ELSE // Random data was not returned from RDSEED. IWKey was not loaded +/// ZF := 1 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ void __DEFAULT_FN_ATTRS +_mm_loadiwkey (unsigned int __ctl, __m128i __intkey, + __m128i __enkey_lo, __m128i __enkey_hi) { + __builtin_ia32_loadiwkey (__intkey, __enkey_lo, __enkey_hi, __ctl); +} + +/// Wrap a 128-bit AES key from __key into a key handle and output in +/// ((__m128i*)__h) to ((__m128i*)__h) + 5 and a 32-bit value as return. +/// The explicit source operand __htype specifies handle restrictions. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> ENCODEKEY128 </c> instructions. +/// +/// \operation +/// InputKey[127:0] := __key[127:0] +/// KeyMetadata[2:0] := __htype[2:0] +/// KeyMetadata[23:3] := 0 // Reserved for future usage +/// KeyMetadata[27:24] := 0 // KeyType is AES-128 (value of 0) +/// KeyMetadata[127:28] := 0 // Reserved for future usage +/// Handle[383:0] := WrapKey128(InputKey[127:0], KeyMetadata[127:0], +/// IWKey.Integrity Key[127:0], IWKey.Encryption Key[255:0]) +/// dst[0] := IWKey.NoBackup +/// dst[4:1] := IWKey.KeySource[3:0] +/// dst[31:5] := 0 +/// MEM[__h+127:__h] := Handle[127:0] // AAD +/// MEM[__h+255:__h+128] := Handle[255:128] // Integrity Tag +/// MEM[__h+383:__h+256] := Handle[383:256] // CipherText +/// MEM[__h+511:__h+384] := 0 // Reserved for future usage +/// MEM[__h+639:__h+512] := 0 // Reserved for future usage +/// MEM[__h+767:__h+640] := 0 // Reserved for future usage +/// OF := 0 +/// SF := 0 +/// ZF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned int __DEFAULT_FN_ATTRS +_mm_encodekey128_u32(unsigned int __htype, __m128i __key, void *__h) { + return __builtin_ia32_encodekey128_u32(__htype, (__v2di)__key, __h); +} + +/// Wrap a 256-bit AES key from __key_hi:__key_lo into a key handle, then +/// output handle in ((__m128i*)__h) to ((__m128i*)__h) + 6 and +/// a 32-bit value as return. +/// The explicit source operand __htype specifies handle restrictions. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> ENCODEKEY256 </c> instructions. +/// +/// \operation +/// InputKey[127:0] := __key_lo[127:0] +/// InputKey[255:128] := __key_hi[255:128] +/// KeyMetadata[2:0] := __htype[2:0] +/// KeyMetadata[23:3] := 0 // Reserved for future usage +/// KeyMetadata[27:24] := 1 // KeyType is AES-256 (value of 1) +/// KeyMetadata[127:28] := 0 // Reserved for future usage +/// Handle[511:0] := WrapKey256(InputKey[255:0], KeyMetadata[127:0], +/// IWKey.Integrity Key[127:0], IWKey.Encryption Key[255:0]) +/// dst[0] := IWKey.NoBackup +/// dst[4:1] := IWKey.KeySource[3:0] +/// dst[31:5] := 0 +/// MEM[__h+127:__h] := Handle[127:0] // AAD +/// MEM[__h+255:__h+128] := Handle[255:128] // Tag +/// MEM[__h+383:__h+256] := Handle[383:256] // CipherText[127:0] +/// MEM[__h+511:__h+384] := Handle[511:384] // CipherText[255:128] +/// MEM[__h+639:__h+512] := 0 // Reserved for future usage +/// MEM[__h+767:__h+640] := 0 // Reserved for future usage +/// MEM[__h+895:__h+768] := 0 Integrity// Reserved for future usage +/// OF := 0 +/// SF := 0 +/// ZF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned int __DEFAULT_FN_ATTRS +_mm_encodekey256_u32(unsigned int __htype, __m128i __key_lo, __m128i __key_hi, + void *__h) { + return __builtin_ia32_encodekey256_u32(__htype, (__v2di)__key_lo, + (__v2di)__key_hi, __h); +} + +/// The AESENC128KL performs 10 rounds of AES to encrypt the __idata using +/// the 128-bit key in the handle from the __h. It stores the result in the +/// __odata. And return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESENC128KL </c> instructions. +/// +/// \operation +/// Handle[383:0] := MEM[__h+383:__h] // Load is not guaranteed to be atomic. +/// IllegalHandle := ( HandleReservedBitSet (Handle[383:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[383:256] || +/// HandleKeyType (Handle[383:0]) != HANDLE_KEY_TYPE_AES128 ) +/// IF (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate384 (Handle[383:0], IWKey) +/// IF (Authentic == 0) +/// ZF := 1 +/// ELSE +/// MEM[__odata+127:__odata] := AES128Encrypt (__idata[127:0], UnwrappedKey) +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesenc128kl_u8(__m128i* __odata, __m128i __idata, const void *__h) { + return __builtin_ia32_aesenc128kl_u8((__v2di *)__odata, (__v2di)__idata, __h); +} + +/// The AESENC256KL performs 14 rounds of AES to encrypt the __idata using +/// the 256-bit key in the handle from the __h. It stores the result in the +/// __odata. And return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESENC256KL </c> instructions. +/// +/// \operation +/// Handle[511:0] := MEM[__h+511:__h] // Load is not guaranteed to be atomic. +/// IllegalHandle := ( HandleReservedBitSet (Handle[511:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[255:128] || +/// HandleKeyType (Handle[511:0]) != HANDLE_KEY_TYPE_AES256 ) +/// IF (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate512 (Handle[511:0], IWKey) +/// IF (Authentic == 0) +/// ZF := 1 +/// ELSE +/// MEM[__odata+127:__odata] := AES256Encrypt (__idata[127:0], UnwrappedKey) +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesenc256kl_u8(__m128i* __odata, __m128i __idata, const void *__h) { + return __builtin_ia32_aesenc256kl_u8((__v2di *)__odata, (__v2di)__idata, __h); +} + +/// The AESDEC128KL performs 10 rounds of AES to decrypt the __idata using +/// the 128-bit key in the handle from the __h. It stores the result in the +/// __odata. And return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESDEC128KL </c> instructions. +/// +/// \operation +/// Handle[383:0] := MEM[__h+383:__h] // Load is not guaranteed to be atomic. +/// IllegalHandle := (HandleReservedBitSet (Handle[383:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[383:256] || +/// HandleKeyType (Handle[383:0]) != HANDLE_KEY_TYPE_AES128) +/// IF (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate384 (Handle[383:0], IWKey) +/// IF (Authentic == 0) +/// ZF := 1 +/// ELSE +/// MEM[__odata+127:__odata] := AES128Decrypt (__idata[127:0], UnwrappedKey) +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesdec128kl_u8(__m128i* __odata, __m128i __idata, const void *__h) { + return __builtin_ia32_aesdec128kl_u8((__v2di *)__odata, (__v2di)__idata, __h); +} + +/// The AESDEC256KL performs 10 rounds of AES to decrypt the __idata using +/// the 256-bit key in the handle from the __h. It stores the result in the +/// __odata. And return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESDEC256KL </c> instructions. +/// +/// \operation +/// Handle[511:0] := MEM[__h+511:__h] +/// IllegalHandle := (HandleReservedBitSet (Handle[511:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[383:256] || +/// HandleKeyType (Handle[511:0]) != HANDLE_KEY_TYPE_AES256) +/// IF (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate512 (Handle[511:0], IWKey) +/// IF (Authentic == 0) +/// ZF := 1 +/// ELSE +/// MEM[__odata+127:__odata] := AES256Decrypt (__idata[127:0], UnwrappedKey) +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesdec256kl_u8(__m128i* __odata, __m128i __idata, const void *__h) { + return __builtin_ia32_aesdec256kl_u8((__v2di *)__odata, (__v2di)__idata, __h); +} + +#undef __DEFAULT_FN_ATTRS + +#endif /* !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) \ + || defined(__KL__) */ + +#if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ + defined(__WIDEKL__) + +/* Define the default attributes for the functions in this file. */ +#define __DEFAULT_FN_ATTRS \ + __attribute__((__always_inline__, __nodebug__, __target__("kl,widekl"),\ + __min_vector_width__(128))) + +/// Encrypt __idata[0] to __idata[7] using 128-bit AES key indicated by handle +/// at __h and store each resultant block back from __odata to __odata+7. And +/// return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESENCWIDE128KL </c> instructions. +/// +/// \operation +/// Handle := MEM[__h+383:__h] +/// IllegalHandle := ( HandleReservedBitSet (Handle[383:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[255:128] || +/// HandleKeyType (Handle[383:0]) != HANDLE_KEY_TYPE_AES128 ) +/// IF (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate384 (Handle[383:0], IWKey) +/// IF Authentic == 0 +/// ZF := 1 +/// ELSE +/// FOR i := 0 to 7 +/// __odata[i] := AES128Encrypt (__idata[i], UnwrappedKey) +/// ENDFOR +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesencwide128kl_u8(__m128i __odata[8], const __m128i __idata[8], const void* __h) { + return __builtin_ia32_aesencwide128kl_u8((__v2di *)__odata, + (const __v2di *)__idata, __h); +} + +/// Encrypt __idata[0] to __idata[7] using 256-bit AES key indicated by handle +/// at __h and store each resultant block back from __odata to __odata+7. And +/// return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESENCWIDE256KL </c> instructions. +/// +/// \operation +/// Handle[511:0] := MEM[__h+511:__h] +/// IllegalHandle := ( HandleReservedBitSet (Handle[511:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[255:128] || +/// HandleKeyType (Handle[511:0]) != HANDLE_KEY_TYPE_AES512 ) +/// IF (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate512 (Handle[511:0], IWKey) +/// IF Authentic == 0 +/// ZF := 1 +/// ELSE +/// FOR i := 0 to 7 +/// __odata[i] := AES256Encrypt (__idata[i], UnwrappedKey) +/// ENDFOR +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesencwide256kl_u8(__m128i __odata[8], const __m128i __idata[8], const void* __h) { + return __builtin_ia32_aesencwide256kl_u8((__v2di *)__odata, + (const __v2di *)__idata, __h); +} + +/// Decrypt __idata[0] to __idata[7] using 128-bit AES key indicated by handle +/// at __h and store each resultant block back from __odata to __odata+7. And +/// return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESDECWIDE128KL </c> instructions. +/// +/// \operation +/// Handle[383:0] := MEM[__h+383:__h] +/// IllegalHandle := ( HandleReservedBitSet (Handle[383:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[255:128] || +/// HandleKeyType (Handle) != HANDLE_KEY_TYPE_AES128 ) +/// IF (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate384 (Handle[383:0], IWKey) +/// IF Authentic == 0 +/// ZF := 1 +/// ELSE +/// FOR i := 0 to 7 +/// __odata[i] := AES128Decrypt (__idata[i], UnwrappedKey) +/// ENDFOR +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesdecwide128kl_u8(__m128i __odata[8], const __m128i __idata[8], const void* __h) { + return __builtin_ia32_aesdecwide128kl_u8((__v2di *)__odata, + (const __v2di *)__idata, __h); +} + +/// Decrypt __idata[0] to __idata[7] using 256-bit AES key indicated by handle +/// at __h and store each resultant block back from __odata to __odata+7. And +/// return the affected ZF flag status. +/// +/// \headerfile <x86intrin.h> +/// +/// This intrinsic corresponds to the <c> AESDECWIDE256KL </c> instructions. +/// +/// \operation +/// Handle[511:0] := MEM[__h+511:__h] +/// IllegalHandle = ( HandleReservedBitSet (Handle[511:0]) || +/// (Handle[127:0] AND (CPL > 0)) || +/// Handle[255:128] || +/// HandleKeyType (Handle) != HANDLE_KEY_TYPE_AES512 ) +/// If (IllegalHandle) +/// ZF := 1 +/// ELSE +/// (UnwrappedKey, Authentic) := UnwrapKeyAndAuthenticate512 (Handle[511:0], IWKey) +/// IF Authentic == 0 +/// ZF := 1 +/// ELSE +/// FOR i := 0 to 7 +/// __odata[i] := AES256Decrypt (__idata[i], UnwrappedKey) +/// ENDFOR +/// ZF := 0 +/// FI +/// FI +/// dst := ZF +/// OF := 0 +/// SF := 0 +/// AF := 0 +/// PF := 0 +/// CF := 0 +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_mm_aesdecwide256kl_u8(__m128i __odata[8], const __m128i __idata[8], const void* __h) { + return __builtin_ia32_aesdecwide256kl_u8((__v2di *)__odata, + (const __v2di *)__idata, __h); +} + +#undef __DEFAULT_FN_ATTRS + +#endif /* !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) \ + || defined(__WIDEKL__) */ + +#endif /* _KEYLOCKERINTRIN_H */ diff --git a/lib/include/mm_malloc.h b/lib/include/mm_malloc.h index 0ea32517ae..933dbaacad 100644 --- a/lib/include/mm_malloc.h +++ b/lib/include/mm_malloc.h @@ -54,7 +54,13 @@ _mm_malloc(size_t __size, size_t __align) static __inline__ void __attribute__((__always_inline__, __nodebug__)) _mm_free(void *__p) { +#if defined(__MINGW32__) + __mingw_aligned_free(__p); +#elif defined(_WIN32) + _aligned_free(__p); +#else free(__p); +#endif } #endif diff --git a/lib/include/opencl-c-base.h b/lib/include/opencl-c-base.h index 430e07d36f..e8dcd70377 100644 --- a/lib/include/opencl-c-base.h +++ b/lib/include/opencl-c-base.h @@ -9,6 +9,21 @@ #ifndef _OPENCL_BASE_H_ #define _OPENCL_BASE_H_ +// Define extension macros + +#if (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) +// For SPIR all extensions are supported. +#if defined(__SPIR__) +#define cl_khr_subgroup_extended_types 1 +#define cl_khr_subgroup_non_uniform_vote 1 +#define cl_khr_subgroup_ballot 1 +#define cl_khr_subgroup_non_uniform_arithmetic 1 +#define cl_khr_subgroup_shuffle 1 +#define cl_khr_subgroup_shuffle_relative 1 +#define cl_khr_subgroup_clustered_reduce 1 +#endif // defined(__SPIR__) +#endif // (defined(__OPENCL_CPP_VERSION__) || __OPENCL_C_VERSION__ >= 200) + // built-in scalar data types: /** @@ -568,4 +583,7 @@ typedef struct { #pragma OPENCL EXTENSION cl_intel_device_side_avc_motion_estimation : end #endif // cl_intel_device_side_avc_motion_estimation +// Disable any extensions we may have enabled previously. +#pragma OPENCL EXTENSION all : disable + #endif //_OPENCL_BASE_H_ diff --git a/lib/include/opencl-c.h b/lib/include/opencl-c.h index 66e18bdd47..ab665628c8 100644 --- a/lib/include/opencl-c.h +++ b/lib/include/opencl-c.h @@ -4633,6 +4633,7 @@ float16 __ovld __cnfn convert_float16(float16); // Conversions with double data type parameters or return value. #ifdef cl_khr_fp64 +#pragma OPENCL EXTENSION cl_khr_fp64 : enable char __ovld __cnfn convert_char(double); char __ovld __cnfn convert_char_rte(double); char __ovld __cnfn convert_char_rtn(double); @@ -5455,6 +5456,7 @@ double16 __ovld __cnfn convert_double16_rtz(ushort16); #endif //cl_khr_fp64 #ifdef cl_khr_fp16 +#pragma OPENCL EXTENSION cl_khr_fp16 : enable // Convert half types to non-double types. uchar __ovld __cnfn convert_uchar(half); uchar __ovld __cnfn convert_uchar_rte(half); diff --git a/lib/include/openmp_wrappers/cmath b/lib/include/openmp_wrappers/cmath index bd6011eb6f..1aff66af7d 100644 --- a/lib/include/openmp_wrappers/cmath +++ b/lib/include/openmp_wrappers/cmath @@ -24,8 +24,11 @@ // which might live in cstdlib. #include <cstdlib> +// We need limits because __clang_cuda_cmath.h below uses `std::numeric_limit`. +#include <limits> + #pragma omp begin declare variant match( \ - device = {arch(nvptx, nvptx64)}, implementation = {extension(match_any)}) + device = {arch(nvptx, nvptx64)}, implementation = {extension(match_any, allow_templates)}) #define __CUDA__ #define __OPENMP_NVPTX__ diff --git a/lib/include/openmp_wrappers/complex b/lib/include/openmp_wrappers/complex index d8dcd41670..142e526b81 100644 --- a/lib/include/openmp_wrappers/complex +++ b/lib/include/openmp_wrappers/complex @@ -25,3 +25,28 @@ // Grab the host header too. #include_next <complex> + + +#ifdef __cplusplus + +// If we are compiling against libc++, the macro _LIBCPP_STD_VER should be set +// after including <cmath> above. Since the complex header we use is a +// simplified version of the libc++, we don't need it in this case. If we +// compile against libstdc++, or any other standard library, we will overload +// the (hopefully template) functions in the <complex> header with the ones we +// got from libc++ which decomposes math functions, like `std::sin`, into +// arithmetic and calls to non-complex functions, all of which we can then +// handle. +#ifndef _LIBCPP_STD_VER + +#pragma omp begin declare variant match( \ + device = {arch(nvptx, nvptx64)}, \ + implementation = {extension(match_any, allow_templates)}) + +#include <complex_cmath.h> + +#pragma omp end declare variant + +#endif + +#endif diff --git a/lib/include/openmp_wrappers/complex_cmath.h b/lib/include/openmp_wrappers/complex_cmath.h new file mode 100644 index 0000000000..e3d9aebbbc --- /dev/null +++ b/lib/include/openmp_wrappers/complex_cmath.h @@ -0,0 +1,388 @@ +//===------------------------- __complex_cmath.h --------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// +// +// std::complex header copied from the libcxx source and simplified for use in +// OpenMP target offload regions. +// +//===----------------------------------------------------------------------===// + +#ifndef _OPENMP +#error "This file is for OpenMP compilation only." +#endif + +#ifndef __cplusplus +#error "This file is for C++ compilation only." +#endif + +#ifndef _LIBCPP_COMPLEX +#define _LIBCPP_COMPLEX + +#include <cmath> +#include <type_traits> + +#define __DEVICE__ static constexpr __attribute__((nothrow)) + +namespace std { + +// abs + +template <class _Tp> __DEVICE__ _Tp abs(const std::complex<_Tp> &__c) { + return hypot(__c.real(), __c.imag()); +} + +// arg + +template <class _Tp> __DEVICE__ _Tp arg(const std::complex<_Tp> &__c) { + return atan2(__c.imag(), __c.real()); +} + +template <class _Tp> +typename enable_if<is_integral<_Tp>::value || is_same<_Tp, double>::value, + double>::type +arg(_Tp __re) { + return atan2(0., __re); +} + +template <class _Tp> +typename enable_if<is_same<_Tp, float>::value, float>::type arg(_Tp __re) { + return atan2f(0.F, __re); +} + +// norm + +template <class _Tp> __DEVICE__ _Tp norm(const std::complex<_Tp> &__c) { + if (std::isinf(__c.real())) + return abs(__c.real()); + if (std::isinf(__c.imag())) + return abs(__c.imag()); + return __c.real() * __c.real() + __c.imag() * __c.imag(); +} + +// conj + +template <class _Tp> std::complex<_Tp> conj(const std::complex<_Tp> &__c) { + return std::complex<_Tp>(__c.real(), -__c.imag()); +} + +// proj + +template <class _Tp> std::complex<_Tp> proj(const std::complex<_Tp> &__c) { + std::complex<_Tp> __r = __c; + if (std::isinf(__c.real()) || std::isinf(__c.imag())) + __r = std::complex<_Tp>(INFINITY, copysign(_Tp(0), __c.imag())); + return __r; +} + +// polar + +template <class _Tp> +complex<_Tp> polar(const _Tp &__rho, const _Tp &__theta = _Tp()) { + if (std::isnan(__rho) || signbit(__rho)) + return std::complex<_Tp>(_Tp(NAN), _Tp(NAN)); + if (std::isnan(__theta)) { + if (std::isinf(__rho)) + return std::complex<_Tp>(__rho, __theta); + return std::complex<_Tp>(__theta, __theta); + } + if (std::isinf(__theta)) { + if (std::isinf(__rho)) + return std::complex<_Tp>(__rho, _Tp(NAN)); + return std::complex<_Tp>(_Tp(NAN), _Tp(NAN)); + } + _Tp __x = __rho * cos(__theta); + if (std::isnan(__x)) + __x = 0; + _Tp __y = __rho * sin(__theta); + if (std::isnan(__y)) + __y = 0; + return std::complex<_Tp>(__x, __y); +} + +// log + +template <class _Tp> std::complex<_Tp> log(const std::complex<_Tp> &__x) { + return std::complex<_Tp>(log(abs(__x)), arg(__x)); +} + +// log10 + +template <class _Tp> std::complex<_Tp> log10(const std::complex<_Tp> &__x) { + return log(__x) / log(_Tp(10)); +} + +// sqrt + +template <class _Tp> +__DEVICE__ std::complex<_Tp> sqrt(const std::complex<_Tp> &__x) { + if (std::isinf(__x.imag())) + return std::complex<_Tp>(_Tp(INFINITY), __x.imag()); + if (std::isinf(__x.real())) { + if (__x.real() > _Tp(0)) + return std::complex<_Tp>(__x.real(), std::isnan(__x.imag()) + ? __x.imag() + : copysign(_Tp(0), __x.imag())); + return std::complex<_Tp>(std::isnan(__x.imag()) ? __x.imag() : _Tp(0), + copysign(__x.real(), __x.imag())); + } + return polar(sqrt(abs(__x)), arg(__x) / _Tp(2)); +} + +// exp + +template <class _Tp> +__DEVICE__ std::complex<_Tp> exp(const std::complex<_Tp> &__x) { + _Tp __i = __x.imag(); + if (std::isinf(__x.real())) { + if (__x.real() < _Tp(0)) { + if (!std::isfinite(__i)) + __i = _Tp(1); + } else if (__i == 0 || !std::isfinite(__i)) { + if (std::isinf(__i)) + __i = _Tp(NAN); + return std::complex<_Tp>(__x.real(), __i); + } + } else if (std::isnan(__x.real()) && __x.imag() == 0) + return __x; + _Tp __e = exp(__x.real()); + return std::complex<_Tp>(__e * cos(__i), __e * sin(__i)); +} + +// pow + +template <class _Tp> +std::complex<_Tp> pow(const std::complex<_Tp> &__x, + const std::complex<_Tp> &__y) { + return exp(__y * log(__x)); +} + +// __sqr, computes pow(x, 2) + +template <class _Tp> std::complex<_Tp> __sqr(const std::complex<_Tp> &__x) { + return std::complex<_Tp>((__x.real() - __x.imag()) * + (__x.real() + __x.imag()), + _Tp(2) * __x.real() * __x.imag()); +} + +// asinh + +template <class _Tp> +__DEVICE__ std::complex<_Tp> asinh(const std::complex<_Tp> &__x) { + const _Tp __pi(atan2(+0., -0.)); + if (std::isinf(__x.real())) { + if (std::isnan(__x.imag())) + return __x; + if (std::isinf(__x.imag())) + return std::complex<_Tp>(__x.real(), + copysign(__pi * _Tp(0.25), __x.imag())); + return std::complex<_Tp>(__x.real(), copysign(_Tp(0), __x.imag())); + } + if (std::isnan(__x.real())) { + if (std::isinf(__x.imag())) + return std::complex<_Tp>(__x.imag(), __x.real()); + if (__x.imag() == 0) + return __x; + return std::complex<_Tp>(__x.real(), __x.real()); + } + if (std::isinf(__x.imag())) + return std::complex<_Tp>(copysign(__x.imag(), __x.real()), + copysign(__pi / _Tp(2), __x.imag())); + std::complex<_Tp> __z = log(__x + sqrt(__sqr(__x) + _Tp(1))); + return std::complex<_Tp>(copysign(__z.real(), __x.real()), + copysign(__z.imag(), __x.imag())); +} + +// acosh + +template <class _Tp> +__DEVICE__ std::complex<_Tp> acosh(const std::complex<_Tp> &__x) { + const _Tp __pi(atan2(+0., -0.)); + if (std::isinf(__x.real())) { + if (std::isnan(__x.imag())) + return std::complex<_Tp>(abs(__x.real()), __x.imag()); + if (std::isinf(__x.imag())) { + if (__x.real() > 0) + return std::complex<_Tp>(__x.real(), + copysign(__pi * _Tp(0.25), __x.imag())); + else + return std::complex<_Tp>(-__x.real(), + copysign(__pi * _Tp(0.75), __x.imag())); + } + if (__x.real() < 0) + return std::complex<_Tp>(-__x.real(), copysign(__pi, __x.imag())); + return std::complex<_Tp>(__x.real(), copysign(_Tp(0), __x.imag())); + } + if (std::isnan(__x.real())) { + if (std::isinf(__x.imag())) + return std::complex<_Tp>(abs(__x.imag()), __x.real()); + return std::complex<_Tp>(__x.real(), __x.real()); + } + if (std::isinf(__x.imag())) + return std::complex<_Tp>(abs(__x.imag()), + copysign(__pi / _Tp(2), __x.imag())); + std::complex<_Tp> __z = log(__x + sqrt(__sqr(__x) - _Tp(1))); + return std::complex<_Tp>(copysign(__z.real(), _Tp(0)), + copysign(__z.imag(), __x.imag())); +} + +// atanh + +template <class _Tp> +__DEVICE__ std::complex<_Tp> atanh(const std::complex<_Tp> &__x) { + const _Tp __pi(atan2(+0., -0.)); + if (std::isinf(__x.imag())) { + return std::complex<_Tp>(copysign(_Tp(0), __x.real()), + copysign(__pi / _Tp(2), __x.imag())); + } + if (std::isnan(__x.imag())) { + if (std::isinf(__x.real()) || __x.real() == 0) + return std::complex<_Tp>(copysign(_Tp(0), __x.real()), __x.imag()); + return std::complex<_Tp>(__x.imag(), __x.imag()); + } + if (std::isnan(__x.real())) { + return std::complex<_Tp>(__x.real(), __x.real()); + } + if (std::isinf(__x.real())) { + return std::complex<_Tp>(copysign(_Tp(0), __x.real()), + copysign(__pi / _Tp(2), __x.imag())); + } + if (abs(__x.real()) == _Tp(1) && __x.imag() == _Tp(0)) { + return std::complex<_Tp>(copysign(_Tp(INFINITY), __x.real()), + copysign(_Tp(0), __x.imag())); + } + std::complex<_Tp> __z = log((_Tp(1) + __x) / (_Tp(1) - __x)) / _Tp(2); + return std::complex<_Tp>(copysign(__z.real(), __x.real()), + copysign(__z.imag(), __x.imag())); +} + +// sinh + +template <class _Tp> +__DEVICE__ std::complex<_Tp> sinh(const std::complex<_Tp> &__x) { + if (std::isinf(__x.real()) && !std::isfinite(__x.imag())) + return std::complex<_Tp>(__x.real(), _Tp(NAN)); + if (__x.real() == 0 && !std::isfinite(__x.imag())) + return std::complex<_Tp>(__x.real(), _Tp(NAN)); + if (__x.imag() == 0 && !std::isfinite(__x.real())) + return __x; + return std::complex<_Tp>(sinh(__x.real()) * cos(__x.imag()), + cosh(__x.real()) * sin(__x.imag())); +} + +// cosh + +template <class _Tp> +__DEVICE__ std::complex<_Tp> cosh(const std::complex<_Tp> &__x) { + if (std::isinf(__x.real()) && !std::isfinite(__x.imag())) + return std::complex<_Tp>(abs(__x.real()), _Tp(NAN)); + if (__x.real() == 0 && !std::isfinite(__x.imag())) + return std::complex<_Tp>(_Tp(NAN), __x.real()); + if (__x.real() == 0 && __x.imag() == 0) + return std::complex<_Tp>(_Tp(1), __x.imag()); + if (__x.imag() == 0 && !std::isfinite(__x.real())) + return std::complex<_Tp>(abs(__x.real()), __x.imag()); + return std::complex<_Tp>(cosh(__x.real()) * cos(__x.imag()), + sinh(__x.real()) * sin(__x.imag())); +} + +// tanh + +template <class _Tp> +__DEVICE__ std::complex<_Tp> tanh(const std::complex<_Tp> &__x) { + if (std::isinf(__x.real())) { + if (!std::isfinite(__x.imag())) + return std::complex<_Tp>(_Tp(1), _Tp(0)); + return std::complex<_Tp>(_Tp(1), + copysign(_Tp(0), sin(_Tp(2) * __x.imag()))); + } + if (std::isnan(__x.real()) && __x.imag() == 0) + return __x; + _Tp __2r(_Tp(2) * __x.real()); + _Tp __2i(_Tp(2) * __x.imag()); + _Tp __d(cosh(__2r) + cos(__2i)); + _Tp __2rsh(sinh(__2r)); + if (std::isinf(__2rsh) && std::isinf(__d)) + return std::complex<_Tp>(__2rsh > _Tp(0) ? _Tp(1) : _Tp(-1), + __2i > _Tp(0) ? _Tp(0) : _Tp(-0.)); + return std::complex<_Tp>(__2rsh / __d, sin(__2i) / __d); +} + +// asin + +template <class _Tp> +__DEVICE__ std::complex<_Tp> asin(const std::complex<_Tp> &__x) { + std::complex<_Tp> __z = asinh(complex<_Tp>(-__x.imag(), __x.real())); + return std::complex<_Tp>(__z.imag(), -__z.real()); +} + +// acos + +template <class _Tp> +__DEVICE__ std::complex<_Tp> acos(const std::complex<_Tp> &__x) { + const _Tp __pi(atan2(+0., -0.)); + if (std::isinf(__x.real())) { + if (std::isnan(__x.imag())) + return std::complex<_Tp>(__x.imag(), __x.real()); + if (std::isinf(__x.imag())) { + if (__x.real() < _Tp(0)) + return std::complex<_Tp>(_Tp(0.75) * __pi, -__x.imag()); + return std::complex<_Tp>(_Tp(0.25) * __pi, -__x.imag()); + } + if (__x.real() < _Tp(0)) + return std::complex<_Tp>(__pi, + signbit(__x.imag()) ? -__x.real() : __x.real()); + return std::complex<_Tp>(_Tp(0), + signbit(__x.imag()) ? __x.real() : -__x.real()); + } + if (std::isnan(__x.real())) { + if (std::isinf(__x.imag())) + return std::complex<_Tp>(__x.real(), -__x.imag()); + return std::complex<_Tp>(__x.real(), __x.real()); + } + if (std::isinf(__x.imag())) + return std::complex<_Tp>(__pi / _Tp(2), -__x.imag()); + if (__x.real() == 0 && (__x.imag() == 0 || isnan(__x.imag()))) + return std::complex<_Tp>(__pi / _Tp(2), -__x.imag()); + std::complex<_Tp> __z = log(__x + sqrt(__sqr(__x) - _Tp(1))); + if (signbit(__x.imag())) + return std::complex<_Tp>(abs(__z.imag()), abs(__z.real())); + return std::complex<_Tp>(abs(__z.imag()), -abs(__z.real())); +} + +// atan + +template <class _Tp> +__DEVICE__ std::complex<_Tp> atan(const std::complex<_Tp> &__x) { + std::complex<_Tp> __z = atanh(complex<_Tp>(-__x.imag(), __x.real())); + return std::complex<_Tp>(__z.imag(), -__z.real()); +} + +// sin + +template <class _Tp> +__DEVICE__ std::complex<_Tp> sin(const std::complex<_Tp> &__x) { + std::complex<_Tp> __z = sinh(complex<_Tp>(-__x.imag(), __x.real())); + return std::complex<_Tp>(__z.imag(), -__z.real()); +} + +// cos + +template <class _Tp> std::complex<_Tp> cos(const std::complex<_Tp> &__x) { + return cosh(complex<_Tp>(-__x.imag(), __x.real())); +} + +// tan + +template <class _Tp> +__DEVICE__ std::complex<_Tp> tan(const std::complex<_Tp> &__x) { + std::complex<_Tp> __z = tanh(complex<_Tp>(-__x.imag(), __x.real())); + return std::complex<_Tp>(__z.imag(), -__z.real()); +} + +} // namespace std + +#endif diff --git a/lib/include/popcntintrin.h b/lib/include/popcntintrin.h index 3129010147..0aa94aecda 100644 --- a/lib/include/popcntintrin.h +++ b/lib/include/popcntintrin.h @@ -13,6 +13,12 @@ /* Define the default attributes for the functions in this file. */ #define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("popcnt"))) +#if defined(__cplusplus) && (__cplusplus >= 201103L) +#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS constexpr +#else +#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS +#endif + /// Counts the number of bits in the source operand having a value of 1. /// /// \headerfile <x86intrin.h> @@ -23,7 +29,7 @@ /// An unsigned 32-bit integer operand. /// \returns A 32-bit integer containing the number of bits with value 1 in the /// source operand. -static __inline__ int __DEFAULT_FN_ATTRS +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR _mm_popcnt_u32(unsigned int __A) { return __builtin_popcount(__A); @@ -40,7 +46,7 @@ _mm_popcnt_u32(unsigned int __A) /// An unsigned 64-bit integer operand. /// \returns A 64-bit integer containing the number of bits with value 1 in the /// source operand. -static __inline__ long long __DEFAULT_FN_ATTRS +static __inline__ long long __DEFAULT_FN_ATTRS_CONSTEXPR _mm_popcnt_u64(unsigned long long __A) { return __builtin_popcountll(__A); @@ -48,5 +54,6 @@ _mm_popcnt_u64(unsigned long long __A) #endif /* __x86_64__ */ #undef __DEFAULT_FN_ATTRS +#undef __DEFAULT_FN_ATTRS_CONSTEXPR #endif /* __POPCNTINTRIN_H */ diff --git a/lib/include/ppc_wrappers/smmintrin.h b/lib/include/ppc_wrappers/smmintrin.h index 56ef6ba76b..64f0c76199 100644 --- a/lib/include/ppc_wrappers/smmintrin.h +++ b/lib/include/ppc_wrappers/smmintrin.h @@ -78,6 +78,30 @@ extern __inline __m128i return (__m128i)vec_sel((__v16qu)__A, (__v16qu)__B, __lmask); } +extern __inline __m128i + __attribute__((__gnu_inline__, __always_inline__, __artificial__)) + _mm_insert_epi8(__m128i const __A, int const __D, int const __N) { + __v16qi result = (__v16qi)__A; + result[__N & 0xf] = __D; + return (__m128i)result; +} + +extern __inline __m128i + __attribute__((__gnu_inline__, __always_inline__, __artificial__)) + _mm_insert_epi32(__m128i const __A, int const __D, int const __N) { + __v4si result = (__v4si)__A; + result[__N & 3] = __D; + return (__m128i)result; +} + +extern __inline __m128i + __attribute__((__gnu_inline__, __always_inline__, __artificial__)) + _mm_insert_epi64(__m128i const __A, long long const __D, int const __N) { + __v2di result = (__v2di)__A; + result[__N & 1] = __D; + return (__m128i)result; +} + #else #include_next <smmintrin.h> #endif /* defined(__linux__) && defined(__ppc64__) */ diff --git a/lib/include/uintrintrin.h b/lib/include/uintrintrin.h new file mode 100644 index 0000000000..78aa8779c3 --- /dev/null +++ b/lib/include/uintrintrin.h @@ -0,0 +1,150 @@ +/*===------------------ uintrintrin.h - UINTR intrinsics -------------------=== + * + * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. + * See https://llvm.org/LICENSE.txt for license information. + * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception + * + *===-----------------------------------------------------------------------=== + */ + +#ifndef __X86GPRINTRIN_H +#error "Never use <uintrintrin.h> directly; include <x86gprintrin.h> instead." +#endif + +#ifndef __UINTRINTRIN_H +#define __UINTRINTRIN_H + +/* Define the default attributes for the functions in this file */ +#define __DEFAULT_FN_ATTRS \ + __attribute__((__always_inline__, __nodebug__, __target__("uintr"))) + +#ifdef __x86_64__ + +/// Clears the user interrupt flag (UIF). Its effect takes place immediately: a +/// user interrupt cannot be delivered on the instruction boundary following +/// CLUI. Can be executed only if CR4.UINT = 1, the logical processor is in +/// 64-bit mode, and software is not executing inside an enclave; otherwise, +/// each causes an invalid-opcode exception. Causes a transactional abort if +/// executed inside a transactional region; the abort loads EAX as it would +/// had it been due to an execution of CLI. +/// +/// \headerfile <x86gprintrin.h> +/// +/// This intrinsic corresponds to the <c> CLUI </c> instruction. +/// +/// \operation +/// UIF := 0 +/// \endoperation +static __inline__ void __DEFAULT_FN_ATTRS +_clui (void) +{ + __builtin_ia32_clui(); +} + +/// Sets the user interrupt flag (UIF). Its effect takes place immediately; a +/// user interrupt may be delivered on the instruction boundary following +/// STUI. Can be executed only if CR4.UINT = 1, the logical processor is in +/// 64-bit mode, and software is not executing inside an enclave; otherwise, +/// each causes an invalid-opcode exception. Causes a transactional abort if +/// executed inside a transactional region; the abort loads EAX as it would +/// had it been due to an execution of STI. +/// +/// \headerfile <x86gprintrin.h> +/// +/// This intrinsic corresponds to the <c> STUI </c> instruction. +/// +/// \operation +/// UIF := 1 +/// \endoperation +static __inline__ void __DEFAULT_FN_ATTRS +_stui (void) +{ + __builtin_ia32_stui(); +} + +/// Get the current value of the user interrupt flag (UIF). Can be executed +/// regardless of CPL and inside a transactional region. Can be executed only +/// if CR4.UINT = 1, the logical processor is in 64-bit mode, and software is +/// not executing inside an enclave; otherwise, it causes an invalid-opcode +/// exception. +/// +/// \headerfile <x86gprintrin.h> +/// +/// This intrinsic corresponds to the <c> TESTUI </c> instruction. +/// +/// \returns The current value of the user interrupt flag (UIF). +/// +/// \operation +/// CF := UIF +/// ZF := 0 +/// AF := 0 +/// OF := 0 +/// PF := 0 +/// SF := 0 +/// dst := CF +/// \endoperation +static __inline__ unsigned char __DEFAULT_FN_ATTRS +_testui (void) +{ + return __builtin_ia32_testui(); +} + +/// Send interprocessor user interrupt. Can be executed only if +/// CR4.UINT = IA32_UINT_TT[0] = 1, the logical processor is in 64-bit mode, +/// and software is not executing inside an enclave; otherwise, it causes an +/// invalid-opcode exception. May be executed at any privilege level, all of +/// its memory accesses are performed with supervisor privilege. +/// +/// \headerfile <x86gprintrin.h> +/// +/// This intrinsic corresponds to the <c> SENDUIPI </c> instruction +/// +/// \param __a +/// Index of user-interrupt target table entry in user-interrupt target +/// table. +/// +/// \operation +/// IF __a > UITTSZ +/// GP (0) +/// FI +/// tempUITTE := MEM[UITTADDR + (a<<4)] +/// // tempUITTE must be valid, and can't have any reserved bit set +/// IF (tempUITTE.V == 0 OR tempUITTE[7:1] != 0) +/// GP (0) +/// FI +/// tempUPID := MEM[tempUITTE.UPIDADDR] // under lock +/// // tempUPID can't have any reserved bit set +/// IF (tempUPID[15:2] != 0 OR tempUPID[31:24] != 0) +/// GP (0) // release lock +/// FI +/// tempUPID.PIR[tempUITTE.UV] := 1; +/// IF (tempUPID.SN == 0 AND tempUPID.ON == 0) +/// tempUPID.ON := 1 +/// sendNotify := 1 +/// ELSE +/// sendNotify := 0 +/// FI +/// MEM[tempUITTE.UPIDADDR] := tempUPID // release lock +/// IF sendNotify == 1 +/// IF IA32_APIC_BASE[10] == 1 // local APIC is in x2APIC mode +/// // send ordinary IPI with vector tempUPID.NV to 32-bit physical APIC +/// // ID tempUPID.NDST +/// SendOrdinaryIPI(tempUPID.NV, tempUPID.NDST) +/// ELSE +/// // send ordinary IPI with vector tempUPID.NV to 8-bit physical APIC +/// // ID tempUPID.NDST[15:8] +/// SendOrdinaryIPI(tempUPID.NV, tempUPID.NDST[15:8]) +/// FI +/// FI +/// \endoperation +static __inline__ void __DEFAULT_FN_ATTRS +_senduipi (unsigned long long __a) +{ + __builtin_ia32_senduipi(__a); +} + +#endif /* __x86_64__ */ + +#undef __DEFAULT_FN_ATTRS + +#endif /* __UINTRINTRIN_H */ diff --git a/lib/include/wasm_simd128.h b/lib/include/wasm_simd128.h index b78123834b..ac88516ac9 100644 --- a/lib/include/wasm_simd128.h +++ b/lib/include/wasm_simd128.h @@ -18,8 +18,7 @@ typedef int32_t v128_t __attribute__((__vector_size__(16), __aligned__(16))); // Internal types determined by clang builtin definitions typedef int32_t __v128_u __attribute__((__vector_size__(16), __aligned__(1))); -typedef char __i8x16 __attribute__((__vector_size__(16), __aligned__(16))); -typedef signed char __s8x16 +typedef signed char __i8x16 __attribute__((__vector_size__(16), __aligned__(16))); typedef unsigned char __u8x16 __attribute__((__vector_size__(16), __aligned__(16))); @@ -35,6 +34,13 @@ typedef unsigned long long __u64x2 typedef float __f32x4 __attribute__((__vector_size__(16), __aligned__(16))); typedef double __f64x2 __attribute__((__vector_size__(16), __aligned__(16))); +typedef signed char __i8x8 __attribute__((__vector_size__(8), __aligned__(8))); +typedef unsigned char __u8x8 + __attribute__((__vector_size__(8), __aligned__(8))); +typedef short __i16x4 __attribute__((__vector_size__(8), __aligned__(8))); +typedef unsigned short __u16x4 + __attribute__((__vector_size__(8), __aligned__(8))); + #define __DEFAULT_FN_ATTRS \ __attribute__((__always_inline__, __nodebug__, __target__("simd128"), \ __min_vector_width__(128))) @@ -273,7 +279,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_splat(int8_t __a) { (__builtin_wasm_extract_lane_s_i8x16((__i8x16)(__a), __i)) #define wasm_u8x16_extract_lane(__a, __i) \ - (__builtin_wasm_extract_lane_u_i8x16((__i8x16)(__a), __i)) + (__builtin_wasm_extract_lane_u_i8x16((__u8x16)(__a), __i)) #define wasm_i8x16_replace_lane(__a, __i, __b) \ ((v128_t)__builtin_wasm_replace_lane_i8x16((__i8x16)(__a), __i, __b)) @@ -286,7 +292,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_splat(int16_t __a) { (__builtin_wasm_extract_lane_s_i16x8((__i16x8)(__a), __i)) #define wasm_u16x8_extract_lane(__a, __i) \ - (__builtin_wasm_extract_lane_u_i16x8((__i16x8)(__a), __i)) + (__builtin_wasm_extract_lane_u_i16x8((__u16x8)(__a), __i)) #define wasm_i16x8_replace_lane(__a, __i, __b) \ ((v128_t)__builtin_wasm_replace_lane_i16x8((__i16x8)(__a), __i, __b)) @@ -333,17 +339,17 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_f64x2_splat(double __a) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_eq(v128_t __a, v128_t __b) { - return (v128_t)((__s8x16)__a == (__s8x16)__b); + return (v128_t)((__i8x16)__a == (__i8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_ne(v128_t __a, v128_t __b) { - return (v128_t)((__s8x16)__a != (__s8x16)__b); + return (v128_t)((__i8x16)__a != (__i8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_lt(v128_t __a, v128_t __b) { - return (v128_t)((__s8x16)__a < (__s8x16)__b); + return (v128_t)((__i8x16)__a < (__i8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_lt(v128_t __a, @@ -353,7 +359,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_lt(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_gt(v128_t __a, v128_t __b) { - return (v128_t)((__s8x16)__a > (__s8x16)__b); + return (v128_t)((__i8x16)__a > (__i8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_gt(v128_t __a, @@ -363,7 +369,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_gt(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_le(v128_t __a, v128_t __b) { - return (v128_t)((__s8x16)__a <= (__s8x16)__b); + return (v128_t)((__i8x16)__a <= (__i8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_le(v128_t __a, @@ -373,7 +379,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_le(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_ge(v128_t __a, v128_t __b) { - return (v128_t)((__s8x16)__a >= (__s8x16)__b); + return (v128_t)((__i8x16)__a >= (__i8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_ge(v128_t __a, @@ -595,7 +601,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_shl(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_shr(v128_t __a, int32_t __b) { - return (v128_t)((__s8x16)__a >> __b); + return (v128_t)((__i8x16)__a >> __b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_shr(v128_t __a, @@ -616,8 +622,8 @@ wasm_i8x16_add_saturate(v128_t __a, v128_t __b) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_add_saturate(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_add_saturate_u_i8x16((__i8x16)__a, - (__i8x16)__b); + return (v128_t)__builtin_wasm_add_saturate_u_i8x16((__u8x16)__a, + (__u8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_sub(v128_t __a, @@ -633,8 +639,8 @@ wasm_i8x16_sub_saturate(v128_t __a, v128_t __b) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_sub_saturate(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_sub_saturate_u_i8x16((__i8x16)__a, - (__i8x16)__b); + return (v128_t)__builtin_wasm_sub_saturate_u_i8x16((__u8x16)__a, + (__u8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_min(v128_t __a, @@ -644,7 +650,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_min(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_min(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_min_u_i8x16((__i8x16)__a, (__i8x16)__b); + return (v128_t)__builtin_wasm_min_u_i8x16((__u8x16)__a, (__u8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_max(v128_t __a, @@ -654,12 +660,12 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_max(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_max(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_max_u_i8x16((__i8x16)__a, (__i8x16)__b); + return (v128_t)__builtin_wasm_max_u_i8x16((__u8x16)__a, (__u8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_avgr(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_avgr_u_i8x16((__i8x16)__a, (__i8x16)__b); + return (v128_t)__builtin_wasm_avgr_u_i8x16((__u8x16)__a, (__u8x16)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_abs(v128_t __a) { @@ -706,8 +712,8 @@ wasm_i16x8_add_saturate(v128_t __a, v128_t __b) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u16x8_add_saturate(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_add_saturate_u_i16x8((__i16x8)__a, - (__i16x8)__b); + return (v128_t)__builtin_wasm_add_saturate_u_i16x8((__u16x8)__a, + (__u16x8)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_sub(v128_t __a, @@ -723,8 +729,8 @@ wasm_i16x8_sub_saturate(v128_t __a, v128_t __b) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u16x8_sub_saturate(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_sub_saturate_u_i16x8((__i16x8)__a, - (__i16x8)__b); + return (v128_t)__builtin_wasm_sub_saturate_u_i16x8((__u16x8)__a, + (__u16x8)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_mul(v128_t __a, @@ -739,7 +745,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_min(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u16x8_min(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_min_u_i16x8((__i16x8)__a, (__i16x8)__b); + return (v128_t)__builtin_wasm_min_u_i16x8((__u16x8)__a, (__u16x8)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_max(v128_t __a, @@ -749,12 +755,12 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_max(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u16x8_max(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_max_u_i16x8((__i16x8)__a, (__i16x8)__b); + return (v128_t)__builtin_wasm_max_u_i16x8((__u16x8)__a, (__u16x8)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u16x8_avgr(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_avgr_u_i16x8((__i16x8)__a, (__i16x8)__b); + return (v128_t)__builtin_wasm_avgr_u_i16x8((__u16x8)__a, (__u16x8)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_abs(v128_t __a) { @@ -810,7 +816,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_min(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u32x4_min(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_min_u_i32x4((__i32x4)__a, (__i32x4)__b); + return (v128_t)__builtin_wasm_min_u_i32x4((__u32x4)__a, (__u32x4)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_max(v128_t __a, @@ -820,7 +826,7 @@ static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_max(v128_t __a, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u32x4_max(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_max_u_i32x4((__i32x4)__a, (__i32x4)__b); + return (v128_t)__builtin_wasm_max_u_i32x4((__u32x4)__a, (__u32x4)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i64x2_neg(v128_t __a) { @@ -1071,8 +1077,8 @@ wasm_i8x16_narrow_i16x8(v128_t __a, v128_t __b) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u8x16_narrow_i16x8(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_narrow_u_i8x16_i16x8((__i16x8)__a, - (__i16x8)__b); + return (v128_t)__builtin_wasm_narrow_u_i8x16_i16x8((__u16x8)__a, + (__u16x8)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS @@ -1083,48 +1089,76 @@ wasm_i16x8_narrow_i32x4(v128_t __a, v128_t __b) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_u16x8_narrow_i32x4(v128_t __a, v128_t __b) { - return (v128_t)__builtin_wasm_narrow_u_i16x8_i32x4((__i32x4)__a, - (__i32x4)__b); + return (v128_t)__builtin_wasm_narrow_u_i16x8_i32x4((__u32x4)__a, + (__u32x4)__b); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_widen_low_i8x16(v128_t __a) { - return (v128_t)__builtin_wasm_widen_low_s_i16x8_i8x16((__i8x16)__a); + return (v128_t) __builtin_convertvector( + (__i8x8){((__i8x16)__a)[0], ((__i8x16)__a)[1], ((__i8x16)__a)[2], + ((__i8x16)__a)[3], ((__i8x16)__a)[4], ((__i8x16)__a)[5], + ((__i8x16)__a)[6], ((__i8x16)__a)[7]}, + __i16x8); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_widen_high_i8x16(v128_t __a) { - return (v128_t)__builtin_wasm_widen_high_s_i16x8_i8x16((__i8x16)__a); + return (v128_t) __builtin_convertvector( + (__i8x8){((__i8x16)__a)[8], ((__i8x16)__a)[9], ((__i8x16)__a)[10], + ((__i8x16)__a)[11], ((__i8x16)__a)[12], ((__i8x16)__a)[13], + ((__i8x16)__a)[14], ((__i8x16)__a)[15]}, + __i16x8); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_widen_low_u8x16(v128_t __a) { - return (v128_t)__builtin_wasm_widen_low_u_i16x8_i8x16((__i8x16)__a); + return (v128_t) __builtin_convertvector( + (__u8x8){((__u8x16)__a)[0], ((__u8x16)__a)[1], ((__u8x16)__a)[2], + ((__u8x16)__a)[3], ((__u8x16)__a)[4], ((__u8x16)__a)[5], + ((__u8x16)__a)[6], ((__u8x16)__a)[7]}, + __u16x8); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i16x8_widen_high_u8x16(v128_t __a) { - return (v128_t)__builtin_wasm_widen_high_u_i16x8_i8x16((__i8x16)__a); + return (v128_t) __builtin_convertvector( + (__u8x8){((__u8x16)__a)[8], ((__u8x16)__a)[9], ((__u8x16)__a)[10], + ((__u8x16)__a)[11], ((__u8x16)__a)[12], ((__u8x16)__a)[13], + ((__u8x16)__a)[14], ((__u8x16)__a)[15]}, + __u16x8); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_widen_low_i16x8(v128_t __a) { - return (v128_t)__builtin_wasm_widen_low_s_i32x4_i16x8((__i16x8)__a); + return (v128_t) __builtin_convertvector( + (__i16x4){((__i16x8)__a)[0], ((__i16x8)__a)[1], ((__i16x8)__a)[2], + ((__i16x8)__a)[3]}, + __i32x4); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_widen_high_i16x8(v128_t __a) { - return (v128_t)__builtin_wasm_widen_high_s_i32x4_i16x8((__i16x8)__a); + return (v128_t) __builtin_convertvector( + (__i16x4){((__i16x8)__a)[4], ((__i16x8)__a)[5], ((__i16x8)__a)[6], + ((__i16x8)__a)[7]}, + __i32x4); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_widen_low_u16x8(v128_t __a) { - return (v128_t)__builtin_wasm_widen_low_u_i32x4_i16x8((__i16x8)__a); + return (v128_t) __builtin_convertvector( + (__u16x4){((__u16x8)__a)[0], ((__u16x8)__a)[1], ((__u16x8)__a)[2], + ((__u16x8)__a)[3]}, + __u32x4); } static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_widen_high_u16x8(v128_t __a) { - return (v128_t)__builtin_wasm_widen_high_u_i32x4_i16x8((__i16x8)__a); + return (v128_t) __builtin_convertvector( + (__u16x4){((__u16x8)__a)[4], ((__u16x8)__a)[5], ((__u16x8)__a)[6], + ((__u16x8)__a)[7]}, + __u32x4); } // Undefine helper macros diff --git a/lib/include/x86gprintrin.h b/lib/include/x86gprintrin.h new file mode 100644 index 0000000000..1fc6cab4b2 --- /dev/null +++ b/lib/include/x86gprintrin.h @@ -0,0 +1,23 @@ +/*===--------------- x86gprintrin.h - X86 GPR intrinsics ------------------=== + * + * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. + * See https://llvm.org/LICENSE.txt for license information. + * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception + * + *===-----------------------------------------------------------------------=== + */ + +#ifndef __X86GPRINTRIN_H +#define __X86GPRINTRIN_H + +#if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ + defined(__HRESET__) +#include <hresetintrin.h> +#endif + +#if !(defined(_MSC_VER) || defined(__SCE__)) || __has_feature(modules) || \ + defined(__UINTR__) +#include <uintrintrin.h> +#endif + +#endif /* __X86GPRINTRIN_H */ diff --git a/lib/libcxx/include/__availability b/lib/libcxx/include/__availability new file mode 100644 index 0000000000..db2267c8eb --- /dev/null +++ b/lib/libcxx/include/__availability @@ -0,0 +1,206 @@ +// -*- C++ -*- +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP___AVAILABILITY +#define _LIBCPP___AVAILABILITY + +#include <__config> + +#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) +# pragma GCC system_header +#endif + +// Libc++ is shipped by various vendors. In particular, it is used as a system +// library on macOS, iOS and other Apple platforms. In order for users to be +// able to compile a binary that is intended to be deployed to an older version +// of a platform, Clang provides availability attributes [1]. These attributes +// can be placed on declarations and are used to describe the life cycle of a +// symbol in the library. +// +// The main goal is to ensure a compile-time error if a symbol that hasn't been +// introduced in a previously released library is used in a program that targets +// that previously released library. Normally, this would be a load-time error +// when one tries to launch the program against the older library. +// +// For example, the filesystem library was introduced in the dylib in macOS 10.15. +// If a user compiles on a macOS 10.15 host but targets macOS 10.13 with their +// program, the compiler would normally not complain (because the required +// declarations are in the headers), but the dynamic loader would fail to find +// the symbols when actually trying to launch the program on macOS 10.13. To +// turn this into a compile-time issue instead, declarations are annotated with +// when they were introduced, and the compiler can produce a diagnostic if the +// program references something that isn't available on the deployment target. +// +// This mechanism is general in nature, and any vendor can add their markup to +// the library (see below). Whenever a new feature is added that requires support +// in the shared library, a macro should be added below to mark this feature +// as unavailable. When vendors decide to ship the feature as part of their +// shared library, they can update the markup appropriately. +// +// Note that this mechanism is disabled by default in the "upstream" libc++. +// Availability annotations are only meaningful when shipping libc++ inside +// a platform (i.e. as a system library), and so vendors that want them should +// turn those annotations on at CMake configuration time. +// +// [1]: https://clang.llvm.org/docs/AttributeReference.html#availability + + +// For backwards compatibility, allow users to define _LIBCPP_DISABLE_AVAILABILITY +// for a while. +#if defined(_LIBCPP_DISABLE_AVAILABILITY) +# if !defined(_LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS) +# define _LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS +# endif +#endif + +// Availability markup is disabled when building the library, or when the compiler +// doesn't support the proper attributes. +#if defined(_LIBCPP_BUILDING_LIBRARY) || \ + defined(_LIBCXXABI_BUILDING_LIBRARY) || \ + !__has_feature(attribute_availability_with_strict) || \ + !__has_feature(attribute_availability_in_templates) || \ + !__has_extension(pragma_clang_attribute_external_declaration) +# if !defined(_LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS) +# define _LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS +# endif +#endif + +#if defined(_LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS) + + // This controls the availability of std::shared_mutex and std::shared_timed_mutex, + // which were added to the dylib later. +# define _LIBCPP_AVAILABILITY_SHARED_MUTEX + + // These macros control the availability of std::bad_optional_access and + // other exception types. These were put in the shared library to prevent + // code bloat from every user program defining the vtable for these exception + // types. +# define _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS +# define _LIBCPP_AVAILABILITY_BAD_VARIANT_ACCESS +# define _LIBCPP_AVAILABILITY_BAD_ANY_CAST + + // This controls the availability of std::uncaught_exceptions(). +# define _LIBCPP_AVAILABILITY_UNCAUGHT_EXCEPTIONS + + // This controls the availability of the sized version of ::operator delete, + // which was added to the dylib later. +# define _LIBCPP_AVAILABILITY_SIZED_NEW_DELETE + + // This controls the availability of the std::future_error exception. +# define _LIBCPP_AVAILABILITY_FUTURE_ERROR + + // This controls the availability of std::type_info's vtable. + // I can't imagine how using std::type_info can work at all if + // this isn't supported. +# define _LIBCPP_AVAILABILITY_TYPEINFO_VTABLE + + // This controls the availability of std::locale::category members + // (e.g. std::locale::collate), which are defined in the dylib. +# define _LIBCPP_AVAILABILITY_LOCALE_CATEGORY + + // This controls the availability of atomic operations on std::shared_ptr + // (e.g. `std::atomic_store(std::shared_ptr)`), which require a shared + // lock table located in the dylib. +# define _LIBCPP_AVAILABILITY_ATOMIC_SHARED_PTR + + // These macros control the availability of all parts of <filesystem> that + // depend on something in the dylib. +# define _LIBCPP_AVAILABILITY_FILESYSTEM +# define _LIBCPP_AVAILABILITY_FILESYSTEM_PUSH +# define _LIBCPP_AVAILABILITY_FILESYSTEM_POP + + // This controls the availability of std::to_chars. +# define _LIBCPP_AVAILABILITY_TO_CHARS + + // This controls the availability of the C++20 synchronization library, + // which requires shared library support for various operations + // (see libcxx/src/atomic.cpp). +# define _LIBCPP_AVAILABILITY_SYNC + +#elif defined(__APPLE__) + +# define _LIBCPP_AVAILABILITY_SHARED_MUTEX \ + __attribute__((availability(macosx,strict,introduced=10.12))) \ + __attribute__((availability(ios,strict,introduced=10.0))) \ + __attribute__((availability(tvos,strict,introduced=10.0))) \ + __attribute__((availability(watchos,strict,introduced=3.0))) +# define _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS \ + __attribute__((availability(macosx,strict,introduced=10.13))) \ + __attribute__((availability(ios,strict,introduced=11.0))) \ + __attribute__((availability(tvos,strict,introduced=11.0))) \ + __attribute__((availability(watchos,strict,introduced=4.0))) +# define _LIBCPP_AVAILABILITY_BAD_VARIANT_ACCESS \ + _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS +# define _LIBCPP_AVAILABILITY_BAD_ANY_CAST \ + _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS +# define _LIBCPP_AVAILABILITY_UNCAUGHT_EXCEPTIONS \ + __attribute__((availability(macosx,strict,introduced=10.12))) \ + __attribute__((availability(ios,strict,introduced=10.0))) \ + __attribute__((availability(tvos,strict,introduced=10.0))) \ + __attribute__((availability(watchos,strict,introduced=3.0))) +# define _LIBCPP_AVAILABILITY_SIZED_NEW_DELETE \ + __attribute__((availability(macosx,strict,introduced=10.12))) \ + __attribute__((availability(ios,strict,introduced=10.0))) \ + __attribute__((availability(tvos,strict,introduced=10.0))) \ + __attribute__((availability(watchos,strict,introduced=3.0))) +# define _LIBCPP_AVAILABILITY_FUTURE_ERROR \ + __attribute__((availability(ios,strict,introduced=6.0))) +# define _LIBCPP_AVAILABILITY_TYPEINFO_VTABLE \ + __attribute__((availability(macosx,strict,introduced=10.9))) \ + __attribute__((availability(ios,strict,introduced=7.0))) +# define _LIBCPP_AVAILABILITY_LOCALE_CATEGORY \ + __attribute__((availability(macosx,strict,introduced=10.9))) \ + __attribute__((availability(ios,strict,introduced=7.0))) +# define _LIBCPP_AVAILABILITY_ATOMIC_SHARED_PTR \ + __attribute__((availability(macosx,strict,introduced=10.9))) \ + __attribute__((availability(ios,strict,introduced=7.0))) +# define _LIBCPP_AVAILABILITY_FILESYSTEM \ + __attribute__((availability(macosx,strict,introduced=10.15))) \ + __attribute__((availability(ios,strict,introduced=13.0))) \ + __attribute__((availability(tvos,strict,introduced=13.0))) \ + __attribute__((availability(watchos,strict,introduced=6.0))) +# define _LIBCPP_AVAILABILITY_FILESYSTEM_PUSH \ + _Pragma("clang attribute push(__attribute__((availability(macosx,strict,introduced=10.15))), apply_to=any(function,record))") \ + _Pragma("clang attribute push(__attribute__((availability(ios,strict,introduced=13.0))), apply_to=any(function,record))") \ + _Pragma("clang attribute push(__attribute__((availability(tvos,strict,introduced=13.0))), apply_to=any(function,record))") \ + _Pragma("clang attribute push(__attribute__((availability(watchos,strict,introduced=6.0))), apply_to=any(function,record))") +# define _LIBCPP_AVAILABILITY_FILESYSTEM_POP \ + _Pragma("clang attribute pop") \ + _Pragma("clang attribute pop") \ + _Pragma("clang attribute pop") \ + _Pragma("clang attribute pop") +# define _LIBCPP_AVAILABILITY_TO_CHARS \ + _LIBCPP_AVAILABILITY_FILESYSTEM +# define _LIBCPP_AVAILABILITY_SYNC \ + __attribute__((unavailable)) + +#else + +// ...New vendors can add availability markup here... + +# error "It looks like you're trying to enable vendor availability markup, but you haven't defined the corresponding macros yet!" + +#endif + +// Define availability attributes that depend on _LIBCPP_NO_EXCEPTIONS. +// Those are defined in terms of the availability attributes above, and +// should not be vendor-specific. +#if defined(_LIBCPP_NO_EXCEPTIONS) +# define _LIBCPP_AVAILABILITY_FUTURE +# define _LIBCPP_AVAILABILITY_THROW_BAD_ANY_CAST +# define _LIBCPP_AVAILABILITY_THROW_BAD_OPTIONAL_ACCESS +# define _LIBCPP_AVAILABILITY_THROW_BAD_VARIANT_ACCESS +#else +# define _LIBCPP_AVAILABILITY_FUTURE _LIBCPP_AVAILABILITY_FUTURE_ERROR +# define _LIBCPP_AVAILABILITY_THROW_BAD_ANY_CAST _LIBCPP_AVAILABILITY_BAD_ANY_CAST +# define _LIBCPP_AVAILABILITY_THROW_BAD_OPTIONAL_ACCESS _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS +# define _LIBCPP_AVAILABILITY_THROW_BAD_VARIANT_ACCESS _LIBCPP_AVAILABILITY_BAD_VARIANT_ACCESS +#endif + +#endif // _LIBCPP___AVAILABILITY diff --git a/lib/libcxx/include/__bit_reference b/lib/libcxx/include/__bit_reference index 4a2b82064b..9cfb4b84e6 100644 --- a/lib/libcxx/include/__bit_reference +++ b/lib/libcxx/include/__bit_reference @@ -11,7 +11,7 @@ #define _LIBCPP___BIT_REFERENCE #include <__config> -#include <bit> +#include <__bits> #include <algorithm> #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) @@ -239,8 +239,8 @@ __bit_iterator<_Cp, _IsConst> find(__bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, const _Tp& __value_) { if (static_cast<bool>(__value_)) - return __find_bool_true(__first, static_cast<typename _Cp::size_type>(__last - __first)); - return __find_bool_false(__first, static_cast<typename _Cp::size_type>(__last - __first)); + return _VSTD::__find_bool_true(__first, static_cast<typename _Cp::size_type>(__last - __first)); + return _VSTD::__find_bool_false(__first, static_cast<typename _Cp::size_type>(__last - __first)); } // count @@ -313,8 +313,8 @@ typename __bit_iterator<_Cp, _IsConst>::difference_type count(__bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, const _Tp& __value_) { if (static_cast<bool>(__value_)) - return __count_bool_true(__first, static_cast<typename _Cp::size_type>(__last - __first)); - return __count_bool_false(__first, static_cast<typename _Cp::size_type>(__last - __first)); + return _VSTD::__count_bool_true(__first, static_cast<typename _Cp::size_type>(__last - __first)); + return _VSTD::__count_bool_false(__first, static_cast<typename _Cp::size_type>(__last - __first)); } // fill_n @@ -387,9 +387,9 @@ fill_n(__bit_iterator<_Cp, false> __first, typename _Cp::size_type __n, bool __v if (__n > 0) { if (__value_) - __fill_n_true(__first, __n); + _VSTD::__fill_n_true(__first, __n); else - __fill_n_false(__first, __n); + _VSTD::__fill_n_false(__first, __n); } } @@ -538,8 +538,8 @@ __bit_iterator<_Cp, false> copy(__bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, __bit_iterator<_Cp, false> __result) { if (__first.__ctz_ == __result.__ctz_) - return __copy_aligned(__first, __last, __result); - return __copy_unaligned(__first, __last, __result); + return _VSTD::__copy_aligned(__first, __last, __result); + return _VSTD::__copy_unaligned(__first, __last, __result); } // copy_backward @@ -685,8 +685,8 @@ __bit_iterator<_Cp, false> copy_backward(__bit_iterator<_Cp, _IsConst> __first, __bit_iterator<_Cp, _IsConst> __last, __bit_iterator<_Cp, false> __result) { if (__last.__ctz_ == __result.__ctz_) - return __copy_backward_aligned(__first, __last, __result); - return __copy_backward_unaligned(__first, __last, __result); + return _VSTD::__copy_backward_aligned(__first, __last, __result); + return _VSTD::__copy_backward_unaligned(__first, __last, __result); } // move @@ -868,8 +868,8 @@ swap_ranges(__bit_iterator<__C1, false> __first1, __bit_iterator<__C1, false> __ __bit_iterator<__C2, false> __first2) { if (__first1.__ctz_ == __first2.__ctz_) - return __swap_ranges_aligned(__first1, __last1, __first2); - return __swap_ranges_unaligned(__first1, __last1, __first2); + return _VSTD::__swap_ranges_aligned(__first1, __last1, __first2); + return _VSTD::__swap_ranges_unaligned(__first1, __last1, __first2); } // rotate @@ -1083,8 +1083,8 @@ bool equal(__bit_iterator<_Cp, _IC1> __first1, __bit_iterator<_Cp, _IC1> __last1, __bit_iterator<_Cp, _IC2> __first2) { if (__first1.__ctz_ == __first2.__ctz_) - return __equal_aligned(__first1, __last1, __first2); - return __equal_unaligned(__first1, __last1, __first2); + return _VSTD::__equal_aligned(__first1, __last1, __first2); + return _VSTD::__equal_unaligned(__first1, __last1, __first2); } template <class _Cp, bool _IsConst, diff --git a/lib/libcxx/include/__bits b/lib/libcxx/include/__bits new file mode 100644 index 0000000000..0d321da922 --- /dev/null +++ b/lib/libcxx/include/__bits @@ -0,0 +1,146 @@ +// -*- C++ -*- +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP___BITS +#define _LIBCPP___BITS + +#include <__config> + +#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) +#pragma GCC system_header +#endif + +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + + +_LIBCPP_BEGIN_NAMESPACE_STD + +#ifndef _LIBCPP_COMPILER_MSVC + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_ctz(unsigned __x) _NOEXCEPT { return __builtin_ctz(__x); } + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_ctz(unsigned long __x) _NOEXCEPT { return __builtin_ctzl(__x); } + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_ctz(unsigned long long __x) _NOEXCEPT { return __builtin_ctzll(__x); } + + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_clz(unsigned __x) _NOEXCEPT { return __builtin_clz(__x); } + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_clz(unsigned long __x) _NOEXCEPT { return __builtin_clzl(__x); } + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_clz(unsigned long long __x) _NOEXCEPT { return __builtin_clzll(__x); } + + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_popcount(unsigned __x) _NOEXCEPT { return __builtin_popcount(__x); } + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_popcount(unsigned long __x) _NOEXCEPT { return __builtin_popcountl(__x); } + +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +int __libcpp_popcount(unsigned long long __x) _NOEXCEPT { return __builtin_popcountll(__x); } + +#else // _LIBCPP_COMPILER_MSVC + +// Precondition: __x != 0 +inline _LIBCPP_INLINE_VISIBILITY +int __libcpp_ctz(unsigned __x) { + static_assert(sizeof(unsigned) == sizeof(unsigned long), ""); + static_assert(sizeof(unsigned long) == 4, ""); + unsigned long __where; + if (_BitScanForward(&__where, __x)) + return static_cast<int>(__where); + return 32; +} + +inline _LIBCPP_INLINE_VISIBILITY +int __libcpp_ctz(unsigned long __x) { + static_assert(sizeof(unsigned long) == sizeof(unsigned), ""); + return __ctz(static_cast<unsigned>(__x)); +} + +inline _LIBCPP_INLINE_VISIBILITY +int __libcpp_ctz(unsigned long long __x) { + unsigned long __where; +#if defined(_LIBCPP_HAS_BITSCAN64) + (defined(_M_AMD64) || defined(__x86_64__)) + if (_BitScanForward64(&__where, __x)) + return static_cast<int>(__where); +#else + // Win32 doesn't have _BitScanForward64 so emulate it with two 32 bit calls. + if (_BitScanForward(&__where, static_cast<unsigned long>(__x))) + return static_cast<int>(__where); + if (_BitScanForward(&__where, static_cast<unsigned long>(__x >> 32))) + return static_cast<int>(__where + 32); +#endif + return 64; +} + +// Precondition: __x != 0 +inline _LIBCPP_INLINE_VISIBILITY +int __libcpp_clz(unsigned __x) { + static_assert(sizeof(unsigned) == sizeof(unsigned long), ""); + static_assert(sizeof(unsigned long) == 4, ""); + unsigned long __where; + if (_BitScanReverse(&__where, __x)) + return static_cast<int>(31 - __where); + return 32; // Undefined Behavior. +} + +inline _LIBCPP_INLINE_VISIBILITY +int __libcpp_clz(unsigned long __x) { + static_assert(sizeof(unsigned) == sizeof(unsigned long), ""); + return __libcpp_clz(static_cast<unsigned>(__x)); +} + +inline _LIBCPP_INLINE_VISIBILITY +int __libcpp_clz(unsigned long long __x) { + unsigned long __where; +#if defined(_LIBCPP_HAS_BITSCAN64) + if (_BitScanReverse64(&__where, __x)) + return static_cast<int>(63 - __where); +#else + // Win32 doesn't have _BitScanReverse64 so emulate it with two 32 bit calls. + if (_BitScanReverse(&__where, static_cast<unsigned long>(__x >> 32))) + return static_cast<int>(63 - (__where + 32)); + if (_BitScanReverse(&__where, static_cast<unsigned long>(__x))) + return static_cast<int>(63 - __where); +#endif + return 64; // Undefined Behavior. +} + +inline _LIBCPP_INLINE_VISIBILITY int __libcpp_popcount(unsigned __x) { + static_assert(sizeof(unsigned) == 4, ""); + return __popcnt(__x); +} + +inline _LIBCPP_INLINE_VISIBILITY int __libcpp_popcount(unsigned long __x) { + static_assert(sizeof(unsigned long) == 4, ""); + return __popcnt(__x); +} + +inline _LIBCPP_INLINE_VISIBILITY int __libcpp_popcount(unsigned long long __x) { + static_assert(sizeof(unsigned long long) == 8, ""); + return __popcnt64(__x); +} + +#endif // _LIBCPP_COMPILER_MSVC + +_LIBCPP_END_NAMESPACE_STD + +_LIBCPP_POP_MACROS + +#endif // _LIBCPP__BITS diff --git a/lib/libcxx/include/__config b/lib/libcxx/include/__config index 575147cead..a3838c89e8 100644 --- a/lib/libcxx/include/__config +++ b/lib/libcxx/include/__config @@ -32,13 +32,13 @@ # define _GNUC_VER_NEW 0 #endif -#define _LIBCPP_VERSION 11000 +#define _LIBCPP_VERSION 12000 #ifndef _LIBCPP_ABI_VERSION # define _LIBCPP_ABI_VERSION 1 #endif -#ifndef __STDC_HOSTED__ +#if __STDC_HOSTED__ == 0 # define _LIBCPP_FREESTANDING #endif @@ -49,8 +49,10 @@ # define _LIBCPP_STD_VER 14 # elif __cplusplus <= 201703L # define _LIBCPP_STD_VER 17 +# elif __cplusplus <= 202002L +# define _LIBCPP_STD_VER 20 # else -# define _LIBCPP_STD_VER 18 // current year, or date of c++2a ratification +# define _LIBCPP_STD_VER 21 // current year, or date of c++2b ratification # endif #endif // _LIBCPP_STD_VER @@ -63,7 +65,7 @@ #elif defined(__wasm__) # define _LIBCPP_OBJECT_FORMAT_WASM 1 #else -# error Unknown object file format + // ... add new file formats here ... #endif #if defined(_LIBCPP_ABI_UNSTABLE) || _LIBCPP_ABI_VERSION >= 2 @@ -105,6 +107,10 @@ // Re-worked external template instantiations for std::string with a focus on // performance and fast-path inlining. # define _LIBCPP_ABI_STRING_OPTIMIZED_EXTERNAL_INSTANTIATION +// Enable clang::trivial_abi on std::unique_ptr. +# define _LIBCPP_ABI_ENABLE_UNIQUE_PTR_TRIVIAL_ABI +// Enable clang::trivial_abi on std::shared_ptr and std::weak_ptr +# define _LIBCPP_ABI_ENABLE_SHARED_PTR_TRIVIAL_ABI #elif _LIBCPP_ABI_VERSION == 1 # if !defined(_LIBCPP_OBJECT_FORMAT_COFF) // Enable compiling copies of now inline methods into the dylib to support @@ -121,9 +127,11 @@ # endif #endif -#ifdef _LIBCPP_TRIVIAL_PAIR_COPY_CTOR -#error "_LIBCPP_TRIVIAL_PAIR_COPY_CTOR" is no longer supported. \ - use _LIBCPP_DEPRECATED_ABI_DISABLE_PAIR_TRIVIAL_COPY_CTOR instead +#if defined(_LIBCPP_BUILDING_LIBRARY) || defined(_LIBCPP_ABI_UNSTABLE) || _LIBCPP_ABI_VERSION >= 2 +// Enable additional explicit instantiations of iostreams components. This +// reduces the number of weak definitions generated in programs that use +// iostreams by providing a single strong definition in the shared library. +# define _LIBCPP_ABI_ENABLE_ADDITIONAL_IOSTREAM_EXPLICIT_INSTANTIATIONS_1 #endif #define _LIBCPP_CONCAT1(_LIBCPP_X,_LIBCPP_Y) _LIBCPP_X##_LIBCPP_Y @@ -256,14 +264,14 @@ # endif // __LONG_LONG_SUPPORTED #endif // __FreeBSD__ -#ifdef __NetBSD__ +#if defined(__NetBSD__) || defined(__OpenBSD__) # include <sys/endian.h> # if _BYTE_ORDER == _LITTLE_ENDIAN # define _LIBCPP_LITTLE_ENDIAN # else // _BYTE_ORDER == _LITTLE_ENDIAN # define _LIBCPP_BIG_ENDIAN # endif // _BYTE_ORDER == _LITTLE_ENDIAN -#endif // __NetBSD__ +#endif // defined(__NetBSD__) || defined(__OpenBSD__) #if defined(_WIN32) # define _LIBCPP_WIN32API @@ -304,7 +312,7 @@ # endif #endif // __sun__ -#if defined(__CloudABI__) +#if defined(__OpenBSD__) || defined(__CloudABI__) // Certain architectures provide arc4random(). Prefer using // arc4random() over /dev/{u,}random to make it possible to obtain // random data even when using sandboxing mechanisms such as chroots, @@ -344,13 +352,11 @@ # if defined(__FreeBSD__) # define _LIBCPP_HAS_ALIGNED_ALLOC # define _LIBCPP_HAS_QUICK_EXIT -# define _LIBCPP_HAS_C11_FEATURES # if __FreeBSD_version >= 1300064 || \ (__FreeBSD_version >= 1201504 && __FreeBSD_version < 1300000) # define _LIBCPP_HAS_TIMESPEC_GET # endif # elif defined(__BIONIC__) -# define _LIBCPP_HAS_C11_FEATURES # if __ANDROID_API__ >= 21 # define _LIBCPP_HAS_QUICK_EXIT # endif @@ -364,7 +370,9 @@ # define _LIBCPP_HAS_ALIGNED_ALLOC # define _LIBCPP_HAS_QUICK_EXIT # define _LIBCPP_HAS_TIMESPEC_GET -# define _LIBCPP_HAS_C11_FEATURES +# elif defined(__OpenBSD__) +# define _LIBCPP_HAS_ALIGNED_ALLOC +# define _LIBCPP_HAS_TIMESPEC_GET # elif defined(__linux__) # if !defined(_LIBCPP_HAS_MUSL_LIBC) # if _LIBCPP_GLIBC_PREREQ(2, 15) || defined(__BIONIC__) @@ -372,16 +380,24 @@ # endif # if _LIBCPP_GLIBC_PREREQ(2, 17) # define _LIBCPP_HAS_ALIGNED_ALLOC -# define _LIBCPP_HAS_C11_FEATURES # define _LIBCPP_HAS_TIMESPEC_GET # endif # else // defined(_LIBCPP_HAS_MUSL_LIBC) # define _LIBCPP_HAS_ALIGNED_ALLOC # define _LIBCPP_HAS_QUICK_EXIT # define _LIBCPP_HAS_TIMESPEC_GET -# define _LIBCPP_HAS_C11_FEATURES # endif -# endif // __linux__ +# elif defined(__APPLE__) + // timespec_get and aligned_alloc were introduced in macOS 10.15 and + // aligned releases +# if (__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ >= 101500 || \ + __ENVIRONMENT_IPHONE_OS_VERSION_MIN_REQUIRED__ >= 130000 || \ + __ENVIRONMENT_TV_OS_VERSION_MIN_REQUIRED__ >= 130000 || \ + __ENVIRONMENT_WATCH_OS_VERSION_MIN_REQUIRED__ >= 60000) +# define _LIBCPP_HAS_ALIGNED_ALLOC +# define _LIBCPP_HAS_TIMESPEC_GET +# endif +# endif // __APPLE__ #endif #ifndef _LIBCPP_CXX03_LANG @@ -389,9 +405,7 @@ #elif defined(_LIBCPP_COMPILER_CLANG) # define _LIBCPP_ALIGNOF(_Tp) _Alignof(_Tp) #else -// This definition is potentially buggy, but it's only taken with GCC in C++03, -// which we barely support anyway. See llvm.org/PR39713 -# define _LIBCPP_ALIGNOF(_Tp) __alignof(_Tp) +# error "We don't know a correct way to implement alignof(T) in C++03 outside of Clang" #endif #define _LIBCPP_PREFERRED_ALIGNOF(_Tp) __alignof(_Tp) @@ -433,10 +447,6 @@ typedef __char32_t char32_t; # define _LIBCPP_NORETURN __attribute__ ((noreturn)) #endif -#if !(__has_feature(cxx_lambdas)) -#define _LIBCPP_HAS_NO_LAMBDAS -#endif - #if !(__has_feature(cxx_nullptr)) # if (__has_extension(cxx_nullptr) || __has_keyword(__nullptr)) && defined(_LIBCPP_ABI_ALWAYS_USE_CXX11_NULLPTR) # define nullptr __nullptr @@ -445,18 +455,6 @@ typedef __char32_t char32_t; # endif #endif -#if !(__has_feature(cxx_rvalue_references)) -#define _LIBCPP_HAS_NO_RVALUE_REFERENCES -#endif - -#if !(__has_feature(cxx_auto_type)) -#define _LIBCPP_HAS_NO_AUTO_TYPE -#endif - -#if !(__has_feature(cxx_variadic_templates)) -#define _LIBCPP_HAS_NO_VARIADICS -#endif - // Objective-C++ features (opt-in) #if __has_feature(objc_arc) #define _LIBCPP_HAS_OBJC_ARC @@ -720,7 +718,7 @@ typedef __char32_t char32_t; #endif #ifndef _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS -# if !defined(_LIBCPP_DISABLE_VISIBILITY_ANNOTATIONS) && __has_attribute(__type_visibility__) +# if !defined(_LIBCPP_DISABLE_VISIBILITY_ANNOTATIONS) # define _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __attribute__ ((__visibility__("default"))) # else # define _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS @@ -754,16 +752,6 @@ typedef __char32_t char32_t; # endif #endif -#ifndef _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION -# ifdef _LIBCPP_OBJECT_FORMAT_COFF // Windows binaries can't merge typeinfos. -# define _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION 2 -# else - // TODO: This isn't strictly correct on ELF platforms due to llvm.org/PR37398 - // And we should consider defaulting to OFF. -# define _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION 1 -# endif -#endif - #ifndef _LIBCPP_HIDE_FROM_ABI # if _LIBCPP_HIDE_FROM_ABI_PER_TU # define _LIBCPP_HIDE_FROM_ABI _LIBCPP_HIDDEN _LIBCPP_INTERNAL_LINKAGE @@ -838,6 +826,12 @@ typedef unsigned int char32_t; # define _LIBCPP_CONSTEXPR constexpr #endif +#ifndef __cpp_consteval +# define _LIBCPP_CONSTEVAL _LIBCPP_CONSTEXPR +#else +# define _LIBCPP_CONSTEVAL consteval +#endif + #ifdef _LIBCPP_CXX03_LANG # define _LIBCPP_DEFAULT {} #else @@ -863,10 +857,6 @@ typedef unsigned int char32_t; # define _LIBCPP_EXPLICIT #endif -#if !__has_builtin(__builtin_operator_new) || !__has_builtin(__builtin_operator_delete) -#define _LIBCPP_HAS_NO_BUILTIN_OPERATOR_NEW_DELETE -#endif - #ifdef _LIBCPP_HAS_NO_STRONG_ENUMS # define _LIBCPP_DECLARE_STRONG_ENUM(x) struct _LIBCPP_TYPE_VIS x { enum __lx # define _LIBCPP_DECLARE_STRONG_ENUM_EPILOG(x) \ @@ -880,34 +870,43 @@ typedef unsigned int char32_t; # define _LIBCPP_DECLARE_STRONG_ENUM_EPILOG(x) #endif // _LIBCPP_HAS_NO_STRONG_ENUMS -#ifdef _LIBCPP_DEBUG -# if _LIBCPP_DEBUG == 0 -# define _LIBCPP_DEBUG_LEVEL 1 -# elif _LIBCPP_DEBUG == 1 -# define _LIBCPP_DEBUG_LEVEL 2 -# else -# error Supported values for _LIBCPP_DEBUG are 0 and 1 -# endif -# if !defined(_LIBCPP_BUILDING_LIBRARY) -# define _LIBCPP_EXTERN_TEMPLATE(...) -# endif +// _LIBCPP_DEBUG potential values: +// - undefined: No assertions. This is the default. +// - 0: Basic assertions +// - 1: Basic assertions + iterator validity checks. +#if !defined(_LIBCPP_DEBUG) +# define _LIBCPP_DEBUG_LEVEL 0 +#elif _LIBCPP_DEBUG == 0 +# define _LIBCPP_DEBUG_LEVEL 1 +#elif _LIBCPP_DEBUG == 1 +# define _LIBCPP_DEBUG_LEVEL 2 +#else +# error Supported values for _LIBCPP_DEBUG are 0 and 1 #endif -#ifndef _LIBCPP_DEBUG_LEVEL -# define _LIBCPP_DEBUG_LEVEL 0 +// _LIBCPP_DEBUG_LEVEL is always defined to one of [0, 1, 2] at this point +#if _LIBCPP_DEBUG_LEVEL >= 1 && !defined(_LIBCPP_DISABLE_EXTERN_TEMPLATE) +# define _LIBCPP_EXTERN_TEMPLATE(...) #endif #ifdef _LIBCPP_DISABLE_EXTERN_TEMPLATE -#define _LIBCPP_EXTERN_TEMPLATE(...) -#define _LIBCPP_EXTERN_TEMPLATE2(...) +# define _LIBCPP_EXTERN_TEMPLATE(...) +# define _LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(...) #endif #ifndef _LIBCPP_EXTERN_TEMPLATE #define _LIBCPP_EXTERN_TEMPLATE(...) extern template __VA_ARGS__; #endif -#ifndef _LIBCPP_EXTERN_TEMPLATE2 -#define _LIBCPP_EXTERN_TEMPLATE2(...) extern template __VA_ARGS__; +// When the Debug mode is enabled, we disable extern declarations because we +// don't want to use the functions compiled in the library, which might not +// have had the debug mode enabled when built. However, some extern declarations +// need to be used, because code correctness depends on it (several instances +// in the <locale>). Those special declarations are declared with +// _LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE, which is enabled even +// when the debug mode is enabled. +#ifndef _LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE +# define _LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(...) extern template __VA_ARGS__; #endif #ifndef _LIBCPP_EXTERN_TEMPLATE_DEFINE @@ -938,6 +937,8 @@ typedef unsigned int char32_t; // We're deferring to Microsoft's STL to provide aligned new et al. We don't // have it unless the language feature test macro is defined. # define _LIBCPP_HAS_NO_LIBRARY_ALIGNED_ALLOCATION +#elif defined(__MVS__) +# define _LIBCPP_HAS_NO_LIBRARY_ALIGNED_ALLOCATION #endif #if defined(__APPLE__) @@ -999,6 +1000,18 @@ typedef unsigned int char32_t; # define _LIBCPP_DEPRECATED_IN_CXX17 #endif +#if _LIBCPP_STD_VER > 17 +# define _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_DEPRECATED +#else +# define _LIBCPP_DEPRECATED_IN_CXX20 +#endif + +#if !defined(_LIBCPP_NO_HAS_CHAR8_T) +# define _LIBCPP_DEPRECATED_WITH_CHAR8_T _LIBCPP_DEPRECATED +#else +# define _LIBCPP_DEPRECATED_WITH_CHAR8_T +#endif + // Macros to enter and leave a state where deprecation warnings are suppressed. #if !defined(_LIBCPP_SUPPRESS_DEPRECATED_PUSH) && \ (defined(_LIBCPP_COMPILER_CLANG) || defined(_LIBCPP_COMPILER_GCC)) @@ -1037,14 +1050,6 @@ typedef unsigned int char32_t; # define _LIBCPP_CONSTEXPR_AFTER_CXX17 #endif -#if _LIBCPP_STD_VER > 17 && \ - !defined(_LIBCPP_HAS_NO_CXX14_CONSTEXPR) && \ - !defined(_LIBCPP_HAS_NO_BUILTIN_IS_CONSTANT_EVALUATED) -# define _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED constexpr -#else -# define _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED -#endif - // The _LIBCPP_NODISCARD_ATTRIBUTE should only be used to define other // NODISCARD macros to the correct attribute. #if __has_cpp_attribute(nodiscard) || defined(_LIBCPP_COMPILER_MSVC) @@ -1079,12 +1084,6 @@ typedef unsigned int char32_t; # define _LIBCPP_INLINE_VAR #endif -#ifdef _LIBCPP_HAS_NO_RVALUE_REFERENCES -# define _LIBCPP_EXPLICIT_MOVE(x) _VSTD::move(x) -#else -# define _LIBCPP_EXPLICIT_MOVE(x) (x) -#endif - #ifndef _LIBCPP_CONSTEXPR_IF_NODEBUG #if defined(_LIBCPP_DEBUG) || defined(_LIBCPP_HAS_NO_CXX14_CONSTEXPR) #define _LIBCPP_CONSTEXPR_IF_NODEBUG @@ -1100,7 +1099,7 @@ typedef unsigned int char32_t; #endif #ifndef _LIBCPP_HAS_NO_ASAN -_LIBCPP_FUNC_VIS extern "C" void __sanitizer_annotate_contiguous_container( +extern "C" _LIBCPP_FUNC_VIS void __sanitizer_annotate_contiguous_container( const void *, const void *, const void *, const void *); #endif @@ -1125,11 +1124,14 @@ _LIBCPP_FUNC_VIS extern "C" void __sanitizer_annotate_contiguous_container( # if defined(__FreeBSD__) || \ defined(__wasi__) || \ defined(__NetBSD__) || \ + defined(__OpenBSD__) || \ + defined(__NuttX__) || \ defined(__linux__) || \ defined(__GNU__) || \ defined(__APPLE__) || \ defined(__CloudABI__) || \ defined(__sun__) || \ + defined(__MVS__) || \ (defined(__MINGW32__) && __has_include(<pthread.h>)) # define _LIBCPP_HAS_THREAD_API_PTHREAD # elif defined(__Fuchsia__) @@ -1167,10 +1169,6 @@ _LIBCPP_FUNC_VIS extern "C" void __sanitizer_annotate_contiguous_container( _LIBCPP_HAS_NO_THREADS is defined. #endif -#if defined(__STDCPP_THREADS__) && defined(_LIBCPP_HAS_NO_THREADS) -#error _LIBCPP_HAS_NO_THREADS cannot be set when __STDCPP_THREADS__ is set. -#endif - #if !defined(_LIBCPP_HAS_NO_THREADS) && !defined(__STDCPP_THREADS__) #define __STDCPP_THREADS__ 1 #endif @@ -1222,13 +1220,15 @@ _LIBCPP_FUNC_VIS extern "C" void __sanitizer_annotate_contiguous_container( // Some systems do not provide gets() in their C library, for security reasons. #ifndef _LIBCPP_C_HAS_NO_GETS # if defined(_LIBCPP_MSVCRT) || \ - (defined(__FreeBSD_version) && __FreeBSD_version >= 1300043) + (defined(__FreeBSD_version) && __FreeBSD_version >= 1300043) || \ + defined(__OpenBSD__) # define _LIBCPP_C_HAS_NO_GETS # endif #endif -#if defined(__BIONIC__) || defined(__CloudABI__) || \ - defined(__Fuchsia__) || defined(__wasi__) || defined(_LIBCPP_HAS_MUSL_LIBC) +#if defined(__BIONIC__) || defined(__CloudABI__) || defined(__NuttX__) || \ + defined(__Fuchsia__) || defined(__wasi__) || defined(_LIBCPP_HAS_MUSL_LIBC) || \ + defined(__MVS__) || defined(__OpenBSD__) #define _LIBCPP_PROVIDES_DEFAULT_RUNE_TABLE #endif @@ -1337,6 +1337,12 @@ _LIBCPP_FUNC_VIS extern "C" void __sanitizer_annotate_contiguous_container( #endif #endif // !defined(_LIBCPP_NODEBUG_TYPE) +#if __has_attribute(__preferred_name__) +#define _LIBCPP_PREFERRED_NAME(x) __attribute__((__preferred_name__(x))) +#else +#define _LIBCPP_PREFERRED_NAME(x) +#endif + #if defined(_LIBCPP_ABI_MICROSOFT) && \ (defined(_LIBCPP_COMPILER_MSVC) || __has_declspec_attribute(empty_bases)) # define _LIBCPP_DECLSPEC_EMPTY_BASES __declspec(empty_bases) @@ -1367,120 +1373,6 @@ _LIBCPP_FUNC_VIS extern "C" void __sanitizer_annotate_contiguous_container( #define _LIBCPP_HAS_NO_SPACESHIP_OPERATOR #endif -// Decide whether to use availability macros. -#if !defined(_LIBCPP_BUILDING_LIBRARY) && \ - !defined(_LIBCXXABI_BUILDING_LIBRARY) && \ - !defined(_LIBCPP_DISABLE_AVAILABILITY) && \ - __has_feature(attribute_availability_with_strict) && \ - __has_feature(attribute_availability_in_templates) && \ - __has_extension(pragma_clang_attribute_external_declaration) -# ifdef __APPLE__ -# define _LIBCPP_USE_AVAILABILITY_APPLE -# endif -#endif - -// Define availability macros. -#if defined(_LIBCPP_USE_AVAILABILITY_APPLE) -# define _LIBCPP_AVAILABILITY_SHARED_MUTEX \ - __attribute__((availability(macosx,strict,introduced=10.12))) \ - __attribute__((availability(ios,strict,introduced=10.0))) \ - __attribute__((availability(tvos,strict,introduced=10.0))) \ - __attribute__((availability(watchos,strict,introduced=3.0))) -# define _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS \ - __attribute__((availability(macosx,strict,introduced=10.13))) \ - __attribute__((availability(ios,strict,introduced=11.0))) \ - __attribute__((availability(tvos,strict,introduced=11.0))) \ - __attribute__((availability(watchos,strict,introduced=4.0))) -# define _LIBCPP_AVAILABILITY_BAD_VARIANT_ACCESS \ - _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS -# define _LIBCPP_AVAILABILITY_BAD_ANY_CAST \ - _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS -# define _LIBCPP_AVAILABILITY_UNCAUGHT_EXCEPTIONS \ - __attribute__((availability(macosx,strict,introduced=10.12))) \ - __attribute__((availability(ios,strict,introduced=10.0))) \ - __attribute__((availability(tvos,strict,introduced=10.0))) \ - __attribute__((availability(watchos,strict,introduced=3.0))) -# define _LIBCPP_AVAILABILITY_SIZED_NEW_DELETE \ - __attribute__((availability(macosx,strict,introduced=10.12))) \ - __attribute__((availability(ios,strict,introduced=10.0))) \ - __attribute__((availability(tvos,strict,introduced=10.0))) \ - __attribute__((availability(watchos,strict,introduced=3.0))) -# define _LIBCPP_AVAILABILITY_FUTURE_ERROR \ - __attribute__((availability(ios,strict,introduced=6.0))) -# define _LIBCPP_AVAILABILITY_TYPEINFO_VTABLE \ - __attribute__((availability(macosx,strict,introduced=10.9))) \ - __attribute__((availability(ios,strict,introduced=7.0))) -# define _LIBCPP_AVAILABILITY_LOCALE_CATEGORY \ - __attribute__((availability(macosx,strict,introduced=10.9))) \ - __attribute__((availability(ios,strict,introduced=7.0))) -# define _LIBCPP_AVAILABILITY_ATOMIC_SHARED_PTR \ - __attribute__((availability(macosx,strict,introduced=10.9))) \ - __attribute__((availability(ios,strict,introduced=7.0))) -# define _LIBCPP_AVAILABILITY_FILESYSTEM \ - __attribute__((availability(macosx,strict,introduced=10.15))) \ - __attribute__((availability(ios,strict,introduced=13.0))) \ - __attribute__((availability(tvos,strict,introduced=13.0))) \ - __attribute__((availability(watchos,strict,introduced=6.0))) -# define _LIBCPP_AVAILABILITY_FILESYSTEM_PUSH \ - _Pragma("clang attribute push(__attribute__((availability(macosx,strict,introduced=10.15))), apply_to=any(function,record))") \ - _Pragma("clang attribute push(__attribute__((availability(ios,strict,introduced=13.0))), apply_to=any(function,record))") \ - _Pragma("clang attribute push(__attribute__((availability(tvos,strict,introduced=13.0))), apply_to=any(function,record))") \ - _Pragma("clang attribute push(__attribute__((availability(watchos,strict,introduced=6.0))), apply_to=any(function,record))") -# define _LIBCPP_AVAILABILITY_FILESYSTEM_POP \ - _Pragma("clang attribute pop") \ - _Pragma("clang attribute pop") \ - _Pragma("clang attribute pop") \ - _Pragma("clang attribute pop") -# define _LIBCPP_AVAILABILITY_TO_CHARS \ - _LIBCPP_AVAILABILITY_FILESYSTEM -# define _LIBCPP_AVAILABILITY_SYNC \ - __attribute__((unavailable)) -#else -# define _LIBCPP_AVAILABILITY_SHARED_MUTEX -# define _LIBCPP_AVAILABILITY_BAD_VARIANT_ACCESS -# define _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS -# define _LIBCPP_AVAILABILITY_BAD_ANY_CAST -# define _LIBCPP_AVAILABILITY_UNCAUGHT_EXCEPTIONS -# define _LIBCPP_AVAILABILITY_SIZED_NEW_DELETE -# define _LIBCPP_AVAILABILITY_FUTURE_ERROR -# define _LIBCPP_AVAILABILITY_TYPEINFO_VTABLE -# define _LIBCPP_AVAILABILITY_LOCALE_CATEGORY -# define _LIBCPP_AVAILABILITY_ATOMIC_SHARED_PTR -# define _LIBCPP_AVAILABILITY_FILESYSTEM -# define _LIBCPP_AVAILABILITY_FILESYSTEM_PUSH -# define _LIBCPP_AVAILABILITY_FILESYSTEM_POP -# define _LIBCPP_AVAILABILITY_TO_CHARS -# define _LIBCPP_AVAILABILITY_SYNC -#endif - -// Define availability that depends on _LIBCPP_NO_EXCEPTIONS. -#ifdef _LIBCPP_NO_EXCEPTIONS -# define _LIBCPP_AVAILABILITY_FUTURE -# define _LIBCPP_AVAILABILITY_THROW_BAD_ANY_CAST -# define _LIBCPP_AVAILABILITY_THROW_BAD_OPTIONAL_ACCESS -# define _LIBCPP_AVAILABILITY_THROW_BAD_VARIANT_ACCESS -#else -# define _LIBCPP_AVAILABILITY_FUTURE _LIBCPP_AVAILABILITY_FUTURE_ERROR -# define _LIBCPP_AVAILABILITY_THROW_BAD_ANY_CAST _LIBCPP_AVAILABILITY_BAD_ANY_CAST -# define _LIBCPP_AVAILABILITY_THROW_BAD_OPTIONAL_ACCESS _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS -# define _LIBCPP_AVAILABILITY_THROW_BAD_VARIANT_ACCESS _LIBCPP_AVAILABILITY_BAD_VARIANT_ACCESS -#endif - -// The stream API was dropped and re-added in the dylib shipped on macOS -// and iOS. We can only assume the dylib to provide these definitions for -// macosx >= 10.9 and ios >= 7.0. Otherwise, the definitions are available -// from the headers, but not from the dylib. Explicit instantiation -// declarations for streams exist conditionally to this; if we provide -// an explicit instantiation declaration and we try to deploy to a dylib -// that does not provide those symbols, we'll get a load-time error. -#if !defined(_LIBCPP_BUILDING_LIBRARY) && \ - ((defined(__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__) && \ - __ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ < 1090) || \ - (defined(__ENVIRONMENT_IPHONE_OS_VERSION_MIN_REQUIRED__) && \ - __ENVIRONMENT_IPHONE_OS_VERSION_MIN_REQUIRED__ < 70000)) -# define _LIBCPP_DO_NOT_ASSUME_STREAMS_EXPLICIT_INSTANTIATION_IN_DYLIB -#endif - #if defined(_LIBCPP_COMPILER_IBM) #define _LIBCPP_HAS_NO_PRAGMA_PUSH_POP_MACRO #endif @@ -1547,6 +1439,12 @@ _LIBCPP_FUNC_VIS extern "C" void __sanitizer_annotate_contiguous_container( #define _LIBCPP_HAS_NO_FGETPOS_FSETPOS #endif +#if __has_attribute(init_priority) +# define _LIBCPP_INIT_PRIORITY_MAX __attribute__((init_priority(101))) +#else +# define _LIBCPP_INIT_PRIORITY_MAX +#endif + #endif // __cplusplus #endif // _LIBCPP_CONFIG diff --git a/lib/libcxx/include/__config_site.in b/lib/libcxx/include/__config_site.in index a6984b2eef..ec4d410bb9 100644 --- a/lib/libcxx/include/__config_site.in +++ b/lib/libcxx/include/__config_site.in @@ -26,12 +26,14 @@ #cmakedefine _LIBCPP_HAS_THREAD_API_WIN32 #cmakedefine _LIBCPP_HAS_THREAD_LIBRARY_EXTERNAL #cmakedefine _LIBCPP_DISABLE_VISIBILITY_ANNOTATIONS +#cmakedefine _LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS #cmakedefine _LIBCPP_NO_VCRUNTIME -#ifndef _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION #cmakedefine _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION @_LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION@ -#endif #cmakedefine _LIBCPP_ABI_NAMESPACE @_LIBCPP_ABI_NAMESPACE@ +#cmakedefine _LIBCPP_HAS_NO_FILESYSTEM_LIBRARY #cmakedefine _LIBCPP_HAS_PARALLEL_ALGORITHMS +#cmakedefine _LIBCPP_HAS_NO_RANDOM_DEVICE +#cmakedefine _LIBCPP_HAS_NO_LOCALIZATION @_LIBCPP_ABI_DEFINES@ diff --git a/lib/libcxx/include/__debug b/lib/libcxx/include/__debug index 11367413fc..7b5bfb3f83 100644 --- a/lib/libcxx/include/__debug +++ b/lib/libcxx/include/__debug @@ -27,26 +27,21 @@ # include <cstddef> #endif -#if _LIBCPP_DEBUG_LEVEL >= 1 && !defined(_LIBCPP_ASSERT) -# define _LIBCPP_ASSERT(x, m) ((x) ? (void)0 : \ - _VSTD::__libcpp_debug_function(_VSTD::__libcpp_debug_info(__FILE__, __LINE__, #x, m))) -#endif - -#if _LIBCPP_DEBUG_LEVEL >= 2 -#ifndef _LIBCPP_DEBUG_ASSERT -#define _LIBCPP_DEBUG_ASSERT(x, m) _LIBCPP_ASSERT(x, m) -#endif -#define _LIBCPP_DEBUG_MODE(...) __VA_ARGS__ -#endif - -#ifndef _LIBCPP_ASSERT -# define _LIBCPP_ASSERT(x, m) ((void)0) -#endif -#ifndef _LIBCPP_DEBUG_ASSERT +#if _LIBCPP_DEBUG_LEVEL == 0 +# define _LIBCPP_DEBUG_ASSERT(x, m) ((void)0) +# define _LIBCPP_ASSERT_IMPL(x, m) ((void)0) +#elif _LIBCPP_DEBUG_LEVEL == 1 # define _LIBCPP_DEBUG_ASSERT(x, m) ((void)0) +# define _LIBCPP_ASSERT_IMPL(x, m) ((x) ? (void)0 : _VSTD::__libcpp_debug_function(_VSTD::__libcpp_debug_info(__FILE__, __LINE__, #x, m))) +#elif _LIBCPP_DEBUG_LEVEL == 2 +# define _LIBCPP_DEBUG_ASSERT(x, m) _LIBCPP_ASSERT(x, m) +# define _LIBCPP_ASSERT_IMPL(x, m) ((x) ? (void)0 : _VSTD::__libcpp_debug_function(_VSTD::__libcpp_debug_info(__FILE__, __LINE__, #x, m))) +#else +# error _LIBCPP_DEBUG_LEVEL must be one of 0, 1, 2 #endif -#ifndef _LIBCPP_DEBUG_MODE -#define _LIBCPP_DEBUG_MODE(...) ((void)0) + +#if !defined(_LIBCPP_ASSERT) +# define _LIBCPP_ASSERT(x, m) _LIBCPP_ASSERT_IMPL(x, m) #endif _LIBCPP_BEGIN_NAMESPACE_STD @@ -59,7 +54,7 @@ struct _LIBCPP_TEMPLATE_VIS __libcpp_debug_info { __libcpp_debug_info(const char* __f, int __l, const char* __p, const char* __m) : __file_(__f), __line_(__l), __pred_(__p), __msg_(__m) {} - _LIBCPP_FUNC_VIS std::string what() const; + _LIBCPP_FUNC_VIS string what() const; const char* __file_; int __line_; @@ -83,7 +78,7 @@ void __libcpp_abort_debug_function(__libcpp_debug_info const&); _LIBCPP_FUNC_VIS bool __libcpp_set_debug_function(__libcpp_debug_function_type __func); -#if _LIBCPP_DEBUG_LEVEL >= 2 || defined(_LIBCPP_BUILDING_LIBRARY) +#if _LIBCPP_DEBUG_LEVEL == 2 || defined(_LIBCPP_BUILDING_LIBRARY) struct _LIBCPP_TYPE_VIS __c_node; @@ -226,7 +221,7 @@ public: template <class _Cont> _LIBCPP_INLINE_VISIBILITY static __c_node* __create_C_node(void *__mem, void *__c, __c_node *__next) { - return ::new(__mem) _C_node<_Cont>(__c, __next); + return ::new (__mem) _C_node<_Cont>(__c, __next); } template <class _Cont> @@ -271,7 +266,7 @@ _LIBCPP_FUNC_VIS __libcpp_db* __get_db(); _LIBCPP_FUNC_VIS const __libcpp_db* __get_const_db(); -#endif // _LIBCPP_DEBUG_LEVEL >= 2 || defined(_LIBCPP_BUILDING_LIBRARY) +#endif // _LIBCPP_DEBUG_LEVEL == 2 || defined(_LIBCPP_BUILDING_LIBRARY) _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/include/__functional_03 b/lib/libcxx/include/__functional_03 index bf86428dea..9616480611 100644 --- a/lib/libcxx/include/__functional_03 +++ b/lib/libcxx/include/__functional_03 @@ -126,7 +126,7 @@ __func<_Fp, _Alloc, _Rp()>::__clone() const _Ap __a(__f_.second()); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__func, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) __func(__f_.first(), _Alloc(__a)); + ::new ((void*)__hold.get()) __func(__f_.first(), _Alloc(__a)); return __hold.release(); } @@ -134,7 +134,7 @@ template<class _Fp, class _Alloc, class _Rp> void __func<_Fp, _Alloc, _Rp()>::__clone(__base<_Rp()>* __p) const { - ::new (__p) __func(__f_.first(), __f_.second()); + ::new ((void*)__p) __func(__f_.first(), __f_.second()); } template<class _Fp, class _Alloc, class _Rp> @@ -212,7 +212,7 @@ __func<_Fp, _Alloc, _Rp(_A0)>::__clone() const _Ap __a(__f_.second()); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__func, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) __func(__f_.first(), _Alloc(__a)); + ::new ((void*)__hold.get()) __func(__f_.first(), _Alloc(__a)); return __hold.release(); } @@ -220,7 +220,7 @@ template<class _Fp, class _Alloc, class _Rp, class _A0> void __func<_Fp, _Alloc, _Rp(_A0)>::__clone(__base<_Rp(_A0)>* __p) const { - ::new (__p) __func(__f_.first(), __f_.second()); + ::new ((void*)__p) __func(__f_.first(), __f_.second()); } template<class _Fp, class _Alloc, class _Rp, class _A0> @@ -298,7 +298,7 @@ __func<_Fp, _Alloc, _Rp(_A0, _A1)>::__clone() const _Ap __a(__f_.second()); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__func, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) __func(__f_.first(), _Alloc(__a)); + ::new ((void*)__hold.get()) __func(__f_.first(), _Alloc(__a)); return __hold.release(); } @@ -306,7 +306,7 @@ template<class _Fp, class _Alloc, class _Rp, class _A0, class _A1> void __func<_Fp, _Alloc, _Rp(_A0, _A1)>::__clone(__base<_Rp(_A0, _A1)>* __p) const { - ::new (__p) __func(__f_.first(), __f_.second()); + ::new ((void*)__p) __func(__f_.first(), __f_.second()); } template<class _Fp, class _Alloc, class _Rp, class _A0, class _A1> @@ -384,7 +384,7 @@ __func<_Fp, _Alloc, _Rp(_A0, _A1, _A2)>::__clone() const _Ap __a(__f_.second()); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__func, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) __func(__f_.first(), _Alloc(__a)); + ::new ((void*)__hold.get()) __func(__f_.first(), _Alloc(__a)); return __hold.release(); } @@ -392,7 +392,7 @@ template<class _Fp, class _Alloc, class _Rp, class _A0, class _A1, class _A2> void __func<_Fp, _Alloc, _Rp(_A0, _A1, _A2)>::__clone(__base<_Rp(_A0, _A1, _A2)>* __p) const { - ::new (__p) __func(__f_.first(), __f_.second()); + ::new ((void*)__p) __func(__f_.first(), __f_.second()); } template<class _Fp, class _Alloc, class _Rp, class _A0, class _A1, class _A2> @@ -554,7 +554,7 @@ function<_Rp()>::function(_Fp __f, if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f); + ::new ((void*)__f_) _FF(__f); } else { @@ -562,7 +562,7 @@ function<_Rp()>::function(_Fp __f, _Ap __a; typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, allocator<_Fp>(__a)); + ::new ((void*)__hold.get()) _FF(__f, allocator<_Fp>(__a)); __f_ = __hold.release(); } } @@ -581,7 +581,7 @@ function<_Rp()>::function(allocator_arg_t, const _Alloc& __a0, _Fp __f, if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f, __a0); + ::new ((void*)__f_) _FF(__f, __a0); } else { @@ -589,7 +589,7 @@ function<_Rp()>::function(allocator_arg_t, const _Alloc& __a0, _Fp __f, _Ap __a(__a0); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, _Alloc(__a)); + ::new ((void*)__hold.get()) _FF(__f, _Alloc(__a)); __f_ = __hold.release(); } } @@ -834,7 +834,7 @@ function<_Rp(_A0)>::function(_Fp __f, if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f); + ::new ((void*)__f_) _FF(__f); } else { @@ -842,7 +842,7 @@ function<_Rp(_A0)>::function(_Fp __f, _Ap __a; typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, allocator<_Fp>(__a)); + ::new ((void*)__hold.get()) _FF(__f, allocator<_Fp>(__a)); __f_ = __hold.release(); } } @@ -861,7 +861,7 @@ function<_Rp(_A0)>::function(allocator_arg_t, const _Alloc& __a0, _Fp __f, if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f, __a0); + ::new ((void*)__f_) _FF(__f, __a0); } else { @@ -869,7 +869,7 @@ function<_Rp(_A0)>::function(allocator_arg_t, const _Alloc& __a0, _Fp __f, _Ap __a(__a0); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, _Alloc(__a)); + ::new ((void*)__hold.get()) _FF(__f, _Alloc(__a)); __f_ = __hold.release(); } } @@ -1114,7 +1114,7 @@ function<_Rp(_A0, _A1)>::function(_Fp __f, if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f); + ::new ((void*)__f_) _FF(__f); } else { @@ -1122,7 +1122,7 @@ function<_Rp(_A0, _A1)>::function(_Fp __f, _Ap __a; typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, allocator<_Fp>(__a)); + ::new ((void*)__hold.get()) _FF(__f, allocator<_Fp>(__a)); __f_ = __hold.release(); } } @@ -1141,7 +1141,7 @@ function<_Rp(_A0, _A1)>::function(allocator_arg_t, const _Alloc& __a0, _Fp __f, if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f, __a0); + ::new ((void*)__f_) _FF(__f, __a0); } else { @@ -1149,7 +1149,7 @@ function<_Rp(_A0, _A1)>::function(allocator_arg_t, const _Alloc& __a0, _Fp __f, _Ap __a(__a0); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, _Alloc(__a)); + ::new ((void*)__hold.get()) _FF(__f, _Alloc(__a)); __f_ = __hold.release(); } } @@ -1394,7 +1394,7 @@ function<_Rp(_A0, _A1, _A2)>::function(_Fp __f, if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f); + ::new ((void*)__f_) _FF(__f); } else { @@ -1402,7 +1402,7 @@ function<_Rp(_A0, _A1, _A2)>::function(_Fp __f, _Ap __a; typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, allocator<_Fp>(__a)); + ::new ((void*)__hold.get()) _FF(__f, allocator<_Fp>(__a)); __f_ = __hold.release(); } } @@ -1421,7 +1421,7 @@ function<_Rp(_A0, _A1, _A2)>::function(allocator_arg_t, const _Alloc& __a0, _Fp if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(__f, __a0); + ::new ((void*)__f_) _FF(__f, __a0); } else { @@ -1429,7 +1429,7 @@ function<_Rp(_A0, _A1, _A2)>::function(allocator_arg_t, const _Alloc& __a0, _Fp _Ap __a(__a0); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(__f, _Alloc(__a)); + ::new ((void*)__hold.get()) _FF(__f, _Alloc(__a)); __f_ = __hold.release(); } } diff --git a/lib/libcxx/include/__functional_base b/lib/libcxx/include/__functional_base index f591bf5a9d..1c02e960d5 100644 --- a/lib/libcxx/include/__functional_base +++ b/lib/libcxx/include/__functional_base @@ -298,7 +298,7 @@ struct __weak_result_type<_Rp (_Cp::*)(_A1, _A2, _A3...) const volatile> template <class _Tp, class ..._Args> struct __invoke_return { - typedef decltype(__invoke(_VSTD::declval<_Tp>(), _VSTD::declval<_Args>()...)) type; + typedef decltype(_VSTD::__invoke(declval<_Tp>(), declval<_Args>()...)) type; }; #else // defined(_LIBCPP_CXX03_LANG) @@ -308,64 +308,64 @@ struct __invoke_return #endif // !defined(_LIBCPP_CXX03_LANG) -template <class _Ret> +template <class _Ret, bool = is_void<_Ret>::value> struct __invoke_void_return_wrapper { #ifndef _LIBCPP_CXX03_LANG template <class ..._Args> static _Ret __call(_Args&&... __args) { - return __invoke(_VSTD::forward<_Args>(__args)...); + return _VSTD::__invoke(_VSTD::forward<_Args>(__args)...); } #else template <class _Fn> static _Ret __call(_Fn __f) { - return __invoke(__f); + return _VSTD::__invoke(__f); } template <class _Fn, class _A0> static _Ret __call(_Fn __f, _A0& __a0) { - return __invoke(__f, __a0); + return _VSTD::__invoke(__f, __a0); } template <class _Fn, class _A0, class _A1> static _Ret __call(_Fn __f, _A0& __a0, _A1& __a1) { - return __invoke(__f, __a0, __a1); + return _VSTD::__invoke(__f, __a0, __a1); } template <class _Fn, class _A0, class _A1, class _A2> static _Ret __call(_Fn __f, _A0& __a0, _A1& __a1, _A2& __a2){ - return __invoke(__f, __a0, __a1, __a2); + return _VSTD::__invoke(__f, __a0, __a1, __a2); } #endif }; -template <> -struct __invoke_void_return_wrapper<void> +template <class _Ret> +struct __invoke_void_return_wrapper<_Ret, true> { #ifndef _LIBCPP_CXX03_LANG template <class ..._Args> static void __call(_Args&&... __args) { - __invoke(_VSTD::forward<_Args>(__args)...); + _VSTD::__invoke(_VSTD::forward<_Args>(__args)...); } #else template <class _Fn> static void __call(_Fn __f) { - __invoke(__f); + _VSTD::__invoke(__f); } template <class _Fn, class _A0> static void __call(_Fn __f, _A0& __a0) { - __invoke(__f, __a0); + _VSTD::__invoke(__f, __a0); } template <class _Fn, class _A0, class _A1> static void __call(_Fn __f, _A0& __a0, _A1& __a1) { - __invoke(__f, __a0, __a1); + _VSTD::__invoke(__f, __a0, __a1); } template <class _Fn, class _A0, class _A1, class _A2> static void __call(_Fn __f, _A0& __a0, _A1& __a1, _A2& __a2) { - __invoke(__f, __a0, __a1, __a2); + _VSTD::__invoke(__f, __a0, __a1, __a2); } #endif }; @@ -382,135 +382,138 @@ private: public: // construct/copy/destroy - _LIBCPP_INLINE_VISIBILITY reference_wrapper(type& __f) _NOEXCEPT + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + reference_wrapper(type& __f) _NOEXCEPT : __f_(_VSTD::addressof(__f)) {} #ifndef _LIBCPP_CXX03_LANG private: reference_wrapper(type&&); public: // = delete; // do not bind to temps #endif // access - _LIBCPP_INLINE_VISIBILITY operator type& () const _NOEXCEPT {return *__f_;} - _LIBCPP_INLINE_VISIBILITY type& get() const _NOEXCEPT {return *__f_;} + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + operator type&() const _NOEXCEPT {return *__f_;} + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + type& get() const _NOEXCEPT {return *__f_;} #ifndef _LIBCPP_CXX03_LANG // invoke template <class... _ArgTypes> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename __invoke_of<type&, _ArgTypes...>::type operator() (_ArgTypes&&... __args) const { - return __invoke(get(), _VSTD::forward<_ArgTypes>(__args)...); + return _VSTD::__invoke(get(), _VSTD::forward<_ArgTypes>(__args)...); } #else _LIBCPP_INLINE_VISIBILITY typename __invoke_return<type>::type operator() () const { - return __invoke(get()); + return _VSTD::__invoke(get()); } template <class _A0> _LIBCPP_INLINE_VISIBILITY typename __invoke_return0<type, _A0>::type operator() (_A0& __a0) const { - return __invoke(get(), __a0); + return _VSTD::__invoke(get(), __a0); } template <class _A0> _LIBCPP_INLINE_VISIBILITY typename __invoke_return0<type, _A0 const>::type operator() (_A0 const& __a0) const { - return __invoke(get(), __a0); + return _VSTD::__invoke(get(), __a0); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0, _A1>::type operator() (_A0& __a0, _A1& __a1) const { - return __invoke(get(), __a0, __a1); + return _VSTD::__invoke(get(), __a0, __a1); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0 const, _A1>::type operator() (_A0 const& __a0, _A1& __a1) const { - return __invoke(get(), __a0, __a1); + return _VSTD::__invoke(get(), __a0, __a1); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0, _A1 const>::type operator() (_A0& __a0, _A1 const& __a1) const { - return __invoke(get(), __a0, __a1); + return _VSTD::__invoke(get(), __a0, __a1); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0 const, _A1 const>::type operator() (_A0 const& __a0, _A1 const& __a1) const { - return __invoke(get(), __a0, __a1); + return _VSTD::__invoke(get(), __a0, __a1); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1, _A2>::type operator() (_A0& __a0, _A1& __a1, _A2& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1, _A2>::type operator() (_A0 const& __a0, _A1& __a1, _A2& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1 const, _A2>::type operator() (_A0& __a0, _A1 const& __a1, _A2& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1, _A2 const>::type operator() (_A0& __a0, _A1& __a1, _A2 const& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1 const, _A2>::type operator() (_A0 const& __a0, _A1 const& __a1, _A2& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1, _A2 const>::type operator() (_A0 const& __a0, _A1& __a1, _A2 const& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1 const, _A2 const>::type operator() (_A0& __a0, _A1 const& __a1, _A2 const& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1 const, _A2 const>::type operator() (_A0 const& __a0, _A1 const& __a1, _A2 const& __a2) const { - return __invoke(get(), __a0, __a1, __a2); + return _VSTD::__invoke(get(), __a0, __a1, __a2); } #endif // _LIBCPP_CXX03_LANG }; template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 reference_wrapper<_Tp> ref(_Tp& __t) _NOEXCEPT { @@ -518,7 +521,7 @@ ref(_Tp& __t) _NOEXCEPT } template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 reference_wrapper<_Tp> ref(reference_wrapper<_Tp> __t) _NOEXCEPT { @@ -526,7 +529,7 @@ ref(reference_wrapper<_Tp> __t) _NOEXCEPT } template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 reference_wrapper<const _Tp> cref(const _Tp& __t) _NOEXCEPT { @@ -534,7 +537,7 @@ cref(const _Tp& __t) _NOEXCEPT } template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 reference_wrapper<const _Tp> cref(reference_wrapper<_Tp> __t) _NOEXCEPT { diff --git a/lib/libcxx/include/__functional_base_03 b/lib/libcxx/include/__functional_base_03 index e6dac90c84..9b08bd26a8 100644 --- a/lib/libcxx/include/__functional_base_03 +++ b/lib/libcxx/include/__functional_base_03 @@ -40,7 +40,7 @@ struct __enable_invoke_imp<_Ret, _T1, false, true> { template <class _Ret, class _T1> struct __enable_invoke_imp<_Ret, _T1, false, false> { typedef typename add_lvalue_reference< - typename __apply_cv<decltype(*_VSTD::declval<_T1>()), _Ret>::type + typename __apply_cv<decltype(*declval<_T1>()), _Ret>::type >::type _Bullet4; typedef _Bullet4 type; }; @@ -142,7 +142,7 @@ __invoke(_Fn __f, _T1& __t1) { template <class _Fp> inline _LIBCPP_INLINE_VISIBILITY -decltype(_VSTD::declval<_Fp&>()()) +decltype(declval<_Fp&>()()) __invoke(_Fp& __f) { return __f(); @@ -150,7 +150,7 @@ __invoke(_Fp& __f) template <class _Fp, class _A0> inline _LIBCPP_INLINE_VISIBILITY -decltype(_VSTD::declval<_Fp&>()(_VSTD::declval<_A0&>())) +decltype(declval<_Fp&>()(declval<_A0&>())) __invoke(_Fp& __f, _A0& __a0) { return __f(__a0); @@ -158,7 +158,7 @@ __invoke(_Fp& __f, _A0& __a0) template <class _Fp, class _A0, class _A1> inline _LIBCPP_INLINE_VISIBILITY -decltype(_VSTD::declval<_Fp&>()(_VSTD::declval<_A0&>(), _VSTD::declval<_A1&>())) +decltype(declval<_Fp&>()(declval<_A0&>(), declval<_A1&>())) __invoke(_Fp& __f, _A0& __a0, _A1& __a1) { return __f(__a0, __a1); @@ -166,7 +166,7 @@ __invoke(_Fp& __f, _A0& __a0, _A1& __a1) template <class _Fp, class _A0, class _A1, class _A2> inline _LIBCPP_INLINE_VISIBILITY -decltype(_VSTD::declval<_Fp&>()(_VSTD::declval<_A0&>(), _VSTD::declval<_A1&>(), _VSTD::declval<_A2&>())) +decltype(declval<_Fp&>()(declval<_A0&>(), declval<_A1&>(), declval<_A2&>())) __invoke(_Fp& __f, _A0& __a0, _A1& __a1, _A2& __a2) { return __f(__a0, __a1, __a2); @@ -181,13 +181,13 @@ struct __invoke_return template <class _Fp> struct __invoke_return<_Fp, false> { - typedef decltype(__invoke(_VSTD::declval<_Fp&>())) type; + typedef decltype(_VSTD::__invoke(declval<_Fp&>())) type; }; template <class _Tp, class _A0> struct __invoke_return0 { - typedef decltype(__invoke(_VSTD::declval<_Tp&>(), _VSTD::declval<_A0&>())) type; + typedef decltype(_VSTD::__invoke(declval<_Tp&>(), declval<_A0&>())) type; }; template <class _Rp, class _Tp, class _A0> @@ -199,8 +199,8 @@ struct __invoke_return0<_Rp _Tp::*, _A0> template <class _Tp, class _A0, class _A1> struct __invoke_return1 { - typedef decltype(__invoke(_VSTD::declval<_Tp&>(), _VSTD::declval<_A0&>(), - _VSTD::declval<_A1&>())) type; + typedef decltype(_VSTD::__invoke(declval<_Tp&>(), declval<_A0&>(), + declval<_A1&>())) type; }; template <class _Rp, class _Class, class _A0, class _A1> @@ -211,9 +211,9 @@ struct __invoke_return1<_Rp _Class::*, _A0, _A1> { template <class _Tp, class _A0, class _A1, class _A2> struct __invoke_return2 { - typedef decltype(__invoke(_VSTD::declval<_Tp&>(), _VSTD::declval<_A0&>(), - _VSTD::declval<_A1&>(), - _VSTD::declval<_A2&>())) type; + typedef decltype(_VSTD::__invoke(declval<_Tp&>(), declval<_A0&>(), + declval<_A1&>(), + declval<_A2&>())) type; }; template <class _Ret, class _Class, class _A0, class _A1, class _A2> diff --git a/lib/libcxx/include/__hash_table b/lib/libcxx/include/__hash_table index 13ff096897..521ebbf2c4 100644 --- a/lib/libcxx/include/__hash_table +++ b/lib/libcxx/include/__hash_table @@ -34,19 +34,17 @@ _LIBCPP_BEGIN_NAMESPACE_STD template <class _Key, class _Tp> struct __hash_value_type; -#ifndef _LIBCPP_CXX03_LANG template <class _Tp> struct __is_hash_value_type_imp : false_type {}; template <class _Key, class _Value> -struct __is_hash_value_type_imp<__hash_value_type<_Key, _Value>> : true_type {}; +struct __is_hash_value_type_imp<__hash_value_type<_Key, _Value> > : true_type {}; template <class ..._Args> struct __is_hash_value_type : false_type {}; template <class _One> struct __is_hash_value_type<_One> : __is_hash_value_type_imp<typename __uncvref<_One>::type> {}; -#endif _LIBCPP_FUNC_VIS size_t __next_prime(size_t __n); @@ -122,7 +120,7 @@ inline _LIBCPP_INLINE_VISIBILITY size_t __next_hash_pow2(size_t __n) { - return __n < 2 ? __n : (size_t(1) << (std::numeric_limits<size_t>::digits - __libcpp_clz(__n-1))); + return __n < 2 ? __n : (size_t(1) << (numeric_limits<size_t>::digits - __libcpp_clz(__n-1))); } @@ -155,12 +153,10 @@ struct __hash_key_value_types { static __container_value_type* __get_ptr(__node_value_type& __n) { return _VSTD::addressof(__n); } -#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY static __container_value_type&& __move(__node_value_type& __v) { return _VSTD::move(__v); } -#endif }; template <class _Key, class _Tp> @@ -197,13 +193,10 @@ struct __hash_key_value_types<__hash_value_type<_Key, _Tp> > { static __container_value_type* __get_ptr(__node_value_type& __n) { return _VSTD::addressof(__n.__get_value()); } -#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY static pair<key_type&&, mapped_type&&> __move(__node_value_type& __v) { return __v.__move(); } -#endif - }; template <class _Tp, class _AllocPtr, class _KVTypes = __hash_key_value_types<_Tp>, @@ -295,10 +288,12 @@ public: typedef typename _NodeTypes::__node_value_type_pointer pointer; _LIBCPP_INLINE_VISIBILITY __hash_iterator() _NOEXCEPT : __node_(nullptr) { - _LIBCPP_DEBUG_MODE(__get_db()->__insert_i(this)); +#if _LIBCPP_DEBUG_LEVEL == 2 + __get_db()->__insert_i(this); +#endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_iterator(const __hash_iterator& __i) : __node_(__i.__node_) @@ -322,7 +317,7 @@ public: } return *this; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY reference operator*() const { @@ -364,7 +359,7 @@ public: {return !(__x == __y);} private: -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_iterator(__next_pointer __node, const void* __c) _NOEXCEPT : __node_(__node) @@ -405,17 +400,21 @@ public: _LIBCPP_INLINE_VISIBILITY __hash_const_iterator() _NOEXCEPT : __node_(nullptr) { - _LIBCPP_DEBUG_MODE(__get_db()->__insert_i(this)); +#if _LIBCPP_DEBUG_LEVEL == 2 + __get_db()->__insert_i(this); +#endif } _LIBCPP_INLINE_VISIBILITY __hash_const_iterator(const __non_const_iterator& __x) _NOEXCEPT : __node_(__x.__node_) { - _LIBCPP_DEBUG_MODE(__get_db()->__iterator_copy(this, &__x)); +#if _LIBCPP_DEBUG_LEVEL == 2 + __get_db()->__iterator_copy(this, &__x); +#endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_const_iterator(const __hash_const_iterator& __i) : __node_(__i.__node_) @@ -439,7 +438,7 @@ public: } return *this; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY reference operator*() const { @@ -480,7 +479,7 @@ public: {return !(__x == __y);} private: -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_const_iterator(__next_pointer __node, const void* __c) _NOEXCEPT : __node_(__node) @@ -518,10 +517,12 @@ public: typedef typename _NodeTypes::__node_value_type_pointer pointer; _LIBCPP_INLINE_VISIBILITY __hash_local_iterator() _NOEXCEPT : __node_(nullptr) { - _LIBCPP_DEBUG_MODE(__get_db()->__insert_i(this)); +#if _LIBCPP_DEBUG_LEVEL == 2 + __get_db()->__insert_i(this); +#endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_local_iterator(const __hash_local_iterator& __i) : __node_(__i.__node_), @@ -549,7 +550,7 @@ public: } return *this; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY reference operator*() const { @@ -593,7 +594,7 @@ public: {return !(__x == __y);} private: -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_local_iterator(__next_pointer __node, size_t __bucket, size_t __bucket_count, const void* __c) _NOEXCEPT @@ -650,7 +651,9 @@ public: _LIBCPP_INLINE_VISIBILITY __hash_const_local_iterator() _NOEXCEPT : __node_(nullptr) { - _LIBCPP_DEBUG_MODE(__get_db()->__insert_i(this)); +#if _LIBCPP_DEBUG_LEVEL == 2 + __get_db()->__insert_i(this); +#endif } _LIBCPP_INLINE_VISIBILITY @@ -659,10 +662,12 @@ public: __bucket_(__x.__bucket_), __bucket_count_(__x.__bucket_count_) { - _LIBCPP_DEBUG_MODE(__get_db()->__iterator_copy(this, &__x)); +#if _LIBCPP_DEBUG_LEVEL == 2 + __get_db()->__iterator_copy(this, &__x); +#endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_const_local_iterator(const __hash_const_local_iterator& __i) : __node_(__i.__node_), @@ -690,7 +695,7 @@ public: } return *this; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY reference operator*() const { @@ -734,7 +739,7 @@ public: {return !(__x == __y);} private: -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __hash_const_local_iterator(__next_pointer __node, size_t __bucket, size_t __bucket_count, const void* __c) _NOEXCEPT @@ -783,7 +788,6 @@ public: _NOEXCEPT_(is_nothrow_copy_constructible<allocator_type>::value) : __data_(__size, __a) {} -#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY __bucket_list_deallocator(__bucket_list_deallocator&& __x) _NOEXCEPT_(is_nothrow_move_constructible<allocator_type>::value) @@ -791,7 +795,6 @@ public: { __x.size() = 0; } -#endif _LIBCPP_INLINE_VISIBILITY size_type& size() _NOEXCEPT {return __data_.first();} @@ -1007,7 +1010,6 @@ public: explicit __hash_table(const allocator_type& __a); __hash_table(const __hash_table& __u); __hash_table(const __hash_table& __u, const allocator_type& __a); -#ifndef _LIBCPP_CXX03_LANG __hash_table(__hash_table&& __u) _NOEXCEPT_( is_nothrow_move_constructible<__bucket_list>::value && @@ -1016,11 +1018,9 @@ public: is_nothrow_move_constructible<hasher>::value && is_nothrow_move_constructible<key_equal>::value); __hash_table(__hash_table&& __u, const allocator_type& __a); -#endif // _LIBCPP_CXX03_LANG ~__hash_table(); __hash_table& operator=(const __hash_table& __u); -#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY __hash_table& operator=(__hash_table&& __u) _NOEXCEPT_( @@ -1028,7 +1028,6 @@ public: is_nothrow_move_assignable<__node_allocator>::value && is_nothrow_move_assignable<hasher>::value && is_nothrow_move_assignable<key_equal>::value); -#endif template <class _InputIterator> void __assign_unique(_InputIterator __first, _InputIterator __last); template <class _InputIterator> @@ -1037,7 +1036,7 @@ public: _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT { - return std::min<size_type>( + return _VSTD::min<size_type>( __node_traits::max_size(__node_alloc()), numeric_limits<difference_type >::max() ); @@ -1066,7 +1065,6 @@ public: iterator __node_insert_multi(const_iterator __p, __node_pointer __nd); -#ifndef _LIBCPP_CXX03_LANG template <class _Key, class ..._Args> _LIBCPP_INLINE_VISIBILITY pair<iterator, bool> __emplace_unique_key_args(_Key const& __k, _Args&&... __args); @@ -1151,15 +1149,6 @@ public: return __emplace_hint_multi(__p, _VSTD::forward<_Pp>(__x)); } -#else // !defined(_LIBCPP_CXX03_LANG) - template <class _Key, class _Args> - _LIBCPP_INLINE_VISIBILITY - pair<iterator, bool> __emplace_unique_key_args(_Key const&, _Args& __args); - - iterator __insert_multi(const __container_value_type& __x); - iterator __insert_multi(const_iterator __p, const __container_value_type& __x); -#endif - _LIBCPP_INLINE_VISIBILITY pair<iterator, bool> __insert_unique(const __container_value_type& __x) { return __emplace_unique_key_args(_NodeTypes::__get_key(__x), __x); @@ -1295,7 +1284,7 @@ public: { _LIBCPP_ASSERT(__n < bucket_count(), "unordered container::begin(n) called with n >= bucket_count()"); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return local_iterator(__bucket_list_[__n], __n, bucket_count(), this); #else return local_iterator(__bucket_list_[__n], __n, bucket_count()); @@ -1308,7 +1297,7 @@ public: { _LIBCPP_ASSERT(__n < bucket_count(), "unordered container::end(n) called with n >= bucket_count()"); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return local_iterator(nullptr, __n, bucket_count(), this); #else return local_iterator(nullptr, __n, bucket_count()); @@ -1321,7 +1310,7 @@ public: { _LIBCPP_ASSERT(__n < bucket_count(), "unordered container::cbegin(n) called with n >= bucket_count()"); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_local_iterator(__bucket_list_[__n], __n, bucket_count(), this); #else return const_local_iterator(__bucket_list_[__n], __n, bucket_count()); @@ -1334,35 +1323,30 @@ public: { _LIBCPP_ASSERT(__n < bucket_count(), "unordered container::cend(n) called with n >= bucket_count()"); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_local_iterator(nullptr, __n, bucket_count(), this); #else return const_local_iterator(nullptr, __n, bucket_count()); #endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const; bool __decrementable(const const_iterator* __i) const; bool __addable(const const_iterator* __i, ptrdiff_t __n) const; bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const; -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 private: void __rehash(size_type __n); -#ifndef _LIBCPP_CXX03_LANG template <class ..._Args> __node_holder __construct_node(_Args&& ...__args); template <class _First, class ..._Rest> __node_holder __construct_node_hash(size_t __hash, _First&& __f, _Rest&&... __rest); -#else // _LIBCPP_CXX03_LANG - __node_holder __construct_node(const __container_value_type& __v); - __node_holder __construct_node_hash(size_t __hash, const __container_value_type& __v); -#endif _LIBCPP_INLINE_VISIBILITY @@ -1373,7 +1357,6 @@ private: _LIBCPP_INLINE_VISIBILITY void __copy_assign_alloc(const __hash_table&, false_type) {} -#ifndef _LIBCPP_CXX03_LANG void __move_assign(__hash_table& __u, false_type); void __move_assign(__hash_table& __u, true_type) _NOEXCEPT_( @@ -1400,7 +1383,6 @@ private: } _LIBCPP_INLINE_VISIBILITY void __move_assign_alloc(__hash_table&, false_type) _NOEXCEPT {} -#endif // _LIBCPP_CXX03_LANG void __deallocate_node(__next_pointer __np) _NOEXCEPT; __next_pointer __detach() _NOEXCEPT; @@ -1477,8 +1459,6 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__hash_table(const __hash_table& __u, { } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Hash, class _Equal, class _Alloc> __hash_table<_Tp, _Hash, _Equal, _Alloc>::__hash_table(__hash_table&& __u) _NOEXCEPT_( @@ -1526,8 +1506,6 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__hash_table(__hash_table&& __u, } } -#endif // _LIBCPP_CXX03_LANG - template <class _Tp, class _Hash, class _Equal, class _Alloc> __hash_table<_Tp, _Hash, _Equal, _Alloc>::~__hash_table() { @@ -1539,7 +1517,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::~__hash_table() #endif __deallocate_node(__p1_.first().__next_); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__erase_c(this); #endif } @@ -1583,7 +1561,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__deallocate_node(__next_pointer __np) while (__np != nullptr) { __next_pointer __next = __np->__next_; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); for (__i_node** __p = __c->end_; __p != __c->beg_; ) { @@ -1593,7 +1571,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__deallocate_node(__next_pointer __np) { (*__p)->__c_ = nullptr; if (--__c->end_ != __p) - memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); } } __get_db()->unlock(); @@ -1618,8 +1596,6 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__detach() _NOEXCEPT return __cache; } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Hash, class _Equal, class _Alloc> void __hash_table<_Tp, _Hash, _Equal, _Alloc>::__move_assign( @@ -1646,7 +1622,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__move_assign( __u.__p1_.first().__next_ = nullptr; __u.size() = 0; } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->swap(this, &__u); #endif } @@ -1714,8 +1690,6 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::operator=(__hash_table&& __u) return *this; } -#endif // _LIBCPP_CXX03_LANG - template <class _Tp, class _Hash, class _Equal, class _Alloc> template <class _InputIterator> void @@ -1800,7 +1774,7 @@ inline typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::begin() _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__p1_.first().__next_, this); #else return iterator(__p1_.first().__next_); @@ -1812,7 +1786,7 @@ inline typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::end() _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(nullptr, this); #else return iterator(nullptr); @@ -1824,7 +1798,7 @@ inline typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::const_iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::begin() const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_iterator(__p1_.first().__next_, this); #else return const_iterator(__p1_.first().__next_); @@ -1836,7 +1810,7 @@ inline typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::const_iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::end() const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_iterator(nullptr, this); #else return const_iterator(nullptr); @@ -1945,7 +1919,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_insert_unique(__node_pointer __ __existing_node = __nd->__ptr(); __inserted = true; } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return pair<iterator, bool>(iterator(__existing_node, this), __inserted); #else return pair<iterator, bool>(iterator(__existing_node), __inserted); @@ -1955,7 +1929,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_insert_unique(__node_pointer __ // Prepare the container for an insertion of the value __cp_val with the hash // __cp_hash. This does a lookup into the container to see if __cp_value is // already present, and performs a rehash if necessary. Returns a pointer to the -// last occurance of __cp_val in the map. +// last occurrence of __cp_val in the map. // // Note that this function does forward exceptions if key_eq() throws, and never // mutates __value or actually inserts into the map. @@ -2043,7 +2017,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_insert_multi(__node_pointer __c __next_pointer __pn = __node_insert_multi_prepare(__cp->__hash(), __cp->__value_); __node_insert_multi_perform(__cp, __pn); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__cp->__ptr(), this); #else return iterator(__cp->__ptr()); @@ -2055,7 +2029,7 @@ typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_insert_multi( const_iterator __p, __node_pointer __cp) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "unordered container::emplace_hint(const_iterator, args...) called with an iterator not" " referring to this unordered container"); @@ -2078,7 +2052,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_insert_multi( __cp->__next_ = __np; __pp->__next_ = static_cast<__next_pointer>(__cp); ++size(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(static_cast<__next_pointer>(__cp), this); #else return iterator(static_cast<__next_pointer>(__cp)); @@ -2089,17 +2063,10 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_insert_multi( -#ifndef _LIBCPP_CXX03_LANG template <class _Tp, class _Hash, class _Equal, class _Alloc> template <class _Key, class ..._Args> pair<typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator, bool> __hash_table<_Tp, _Hash, _Equal, _Alloc>::__emplace_unique_key_args(_Key const& __k, _Args&&... __args) -#else -template <class _Tp, class _Hash, class _Equal, class _Alloc> -template <class _Key, class _Args> -pair<typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator, bool> -__hash_table<_Tp, _Hash, _Equal, _Alloc>::__emplace_unique_key_args(_Key const& __k, _Args& __args) -#endif { size_t __hash = hash_function()(__k); @@ -2123,11 +2090,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__emplace_unique_key_args(_Key const& } } { -#ifndef _LIBCPP_CXX03_LANG __node_holder __h = __construct_node_hash(__hash, _VSTD::forward<_Args>(__args)...); -#else - __node_holder __h = __construct_node_hash(__hash, __args); -#endif if (size()+1 > __bc * max_load_factor() || __bc == 0) { rehash(_VSTD::max<size_type>(2 * __bc + !__is_hash_power2(__bc), @@ -2159,15 +2122,13 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__emplace_unique_key_args(_Key const& __inserted = true; } __done: -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return pair<iterator, bool>(iterator(__nd, this), __inserted); #else return pair<iterator, bool>(iterator(__nd), __inserted); #endif } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Hash, class _Equal, class _Alloc> template <class... _Args> pair<typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator, bool> @@ -2197,7 +2158,7 @@ typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::__emplace_hint_multi( const_iterator __p, _Args&&... __args) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "unordered container::emplace_hint(const_iterator, args...) called with an iterator not" " referring to this unordered container"); @@ -2208,36 +2169,6 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__emplace_hint_multi( return __r; } -#else // _LIBCPP_CXX03_LANG - -template <class _Tp, class _Hash, class _Equal, class _Alloc> -typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator -__hash_table<_Tp, _Hash, _Equal, _Alloc>::__insert_multi(const __container_value_type& __x) -{ - __node_holder __h = __construct_node(__x); - iterator __r = __node_insert_multi(__h.get()); - __h.release(); - return __r; -} - -template <class _Tp, class _Hash, class _Equal, class _Alloc> -typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator -__hash_table<_Tp, _Hash, _Equal, _Alloc>::__insert_multi(const_iterator __p, - const __container_value_type& __x) -{ -#if _LIBCPP_DEBUG_LEVEL >= 2 - _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, - "unordered container::insert(const_iterator, lvalue) called with an iterator not" - " referring to this unordered container"); -#endif - __node_holder __h = __construct_node(__x); - iterator __r = __node_insert_multi(__p, __h.get()); - __h.release(); - return __r; -} - -#endif // _LIBCPP_CXX03_LANG - #if _LIBCPP_STD_VER > 14 template <class _Tp, class _Hash, class _Equal, class _Alloc> template <class _NodeHandle, class _InsertReturnType> @@ -2399,9 +2330,9 @@ template <class _Tp, class _Hash, class _Equal, class _Alloc> void __hash_table<_Tp, _Hash, _Equal, _Alloc>::__rehash(size_type __nbc) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__invalidate_all(this); -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif __pointer_allocator& __npa = __bucket_list_.get_deleter().__alloc(); __bucket_list_.reset(__nbc > 0 ? __pointer_alloc_traits::allocate(__npa, __nbc) : nullptr); @@ -2470,7 +2401,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::find(const _Key& __k) { if ((__nd->__hash() == __hash) && key_eq()(__nd->__upcast()->__value_, __k)) -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__nd, this); #else return iterator(__nd); @@ -2501,7 +2432,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::find(const _Key& __k) const { if ((__nd->__hash() == __hash) && key_eq()(__nd->__upcast()->__value_, __k)) -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_iterator(__nd, this); #else return const_iterator(__nd); @@ -2513,8 +2444,6 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::find(const _Key& __k) const return end(); } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Hash, class _Equal, class _Alloc> template <class ..._Args> typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_holder @@ -2550,43 +2479,12 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__construct_node_hash( return __h; } -#else // _LIBCPP_CXX03_LANG - -template <class _Tp, class _Hash, class _Equal, class _Alloc> -typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_holder -__hash_table<_Tp, _Hash, _Equal, _Alloc>::__construct_node(const __container_value_type& __v) -{ - __node_allocator& __na = __node_alloc(); - __node_holder __h(__node_traits::allocate(__na, 1), _Dp(__na)); - __node_traits::construct(__na, _NodeTypes::__get_ptr(__h->__value_), __v); - __h.get_deleter().__value_constructed = true; - __h->__hash_ = hash_function()(__h->__value_); - __h->__next_ = nullptr; - return _LIBCPP_EXPLICIT_MOVE(__h); // explicitly moved for C++03 -} - -template <class _Tp, class _Hash, class _Equal, class _Alloc> -typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::__node_holder -__hash_table<_Tp, _Hash, _Equal, _Alloc>::__construct_node_hash(size_t __hash, - const __container_value_type& __v) -{ - __node_allocator& __na = __node_alloc(); - __node_holder __h(__node_traits::allocate(__na, 1), _Dp(__na)); - __node_traits::construct(__na, _NodeTypes::__get_ptr(__h->__value_), __v); - __h.get_deleter().__value_constructed = true; - __h->__hash_ = __hash; - __h->__next_ = nullptr; - return _LIBCPP_EXPLICIT_MOVE(__h); // explicitly moved for C++03 -} - -#endif // _LIBCPP_CXX03_LANG - template <class _Tp, class _Hash, class _Equal, class _Alloc> typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::erase(const_iterator __p) { __next_pointer __np = __p.__node_; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "unordered container erase(iterator) called with an iterator not" " referring to this container"); @@ -2606,7 +2504,7 @@ typename __hash_table<_Tp, _Hash, _Equal, _Alloc>::iterator __hash_table<_Tp, _Hash, _Equal, _Alloc>::erase(const_iterator __first, const_iterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__first) == this, "unodered container::erase(iterator, iterator) called with an iterator not" " referring to this unodered container"); @@ -2620,7 +2518,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::erase(const_iterator __first, erase(__p); } __next_pointer __np = __last.__node_; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator (__np, this); #else return iterator (__np); @@ -2691,7 +2589,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::remove(const_iterator __p) _NOEXCEPT __pn->__next_ = __cn->__next_; __cn->__next_ = nullptr; --size(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); for (__i_node** __dp = __c->end_; __dp != __c->beg_; ) { @@ -2701,7 +2599,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::remove(const_iterator __p) _NOEXCEPT { (*__dp)->__c_ = nullptr; if (--__c->end_ != __dp) - memmove(__dp, __dp+1, (__c->end_ - __dp)*sizeof(__i_node*)); + _VSTD::memmove(__dp, __dp+1, (__c->end_ - __dp)*sizeof(__i_node*)); } } __get_db()->unlock(); @@ -2830,9 +2728,9 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::swap(__hash_table& __u) __u.__bucket_list_.reset(__npp); } _VSTD::swap(__bucket_list_.get_deleter().size(), __u.__bucket_list_.get_deleter().size()); - __swap_allocator(__bucket_list_.get_deleter().__alloc(), + _VSTD::__swap_allocator(__bucket_list_.get_deleter().__alloc(), __u.__bucket_list_.get_deleter().__alloc()); - __swap_allocator(__node_alloc(), __u.__node_alloc()); + _VSTD::__swap_allocator(__node_alloc(), __u.__node_alloc()); _VSTD::swap(__p1_.first().__next_, __u.__p1_.first().__next_); __p2_.swap(__u.__p2_); __p3_.swap(__u.__p3_); @@ -2842,7 +2740,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::swap(__hash_table& __u) if (__u.size() > 0) __u.__bucket_list_[__constrain_hash(__u.__p1_.first().__next_->__hash(), __u.bucket_count())] = __u.__p1_.first().__ptr(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->swap(this, &__u); #endif } @@ -2876,7 +2774,7 @@ swap(__hash_table<_Tp, _Hash, _Equal, _Alloc>& __x, __x.swap(__y); } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 template <class _Tp, class _Hash, class _Equal, class _Alloc> bool @@ -2906,7 +2804,7 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>::__subscriptable(const const_iterator*, return false; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/include/__libcpp_version b/lib/libcxx/include/__libcpp_version index 82b3803a20..e334181b40 100644 --- a/lib/libcxx/include/__libcpp_version +++ b/lib/libcxx/include/__libcpp_version @@ -1 +1 @@ -11000 +12000 diff --git a/lib/libcxx/include/__locale b/lib/libcxx/include/__locale index 6d10fa4d3d..77e5faab26 100644 --- a/lib/libcxx/include/__locale +++ b/lib/libcxx/include/__locale @@ -11,6 +11,7 @@ #define _LIBCPP___LOCALE #include <__config> +#include <__availability> #include <string> #include <memory> #include <utility> @@ -20,26 +21,30 @@ #include <locale.h> #if defined(_LIBCPP_MSVCRT_LIKE) # include <cstring> -# include <support/win32/locale_win32.h> -#elif defined(_AIX) -# include <support/ibm/xlocale.h> +# include <__support/win32/locale_win32.h> +#elif defined(__NuttX__) +# include <__support/nuttx/xlocale.h> +#elif defined(_AIX) || defined(__MVS__) +# include <__support/ibm/xlocale.h> #elif defined(__ANDROID__) -# include <support/android/locale_bionic.h> +# include <__support/android/locale_bionic.h> #elif defined(__sun__) # include <xlocale.h> -# include <support/solaris/xlocale.h> +# include <__support/solaris/xlocale.h> #elif defined(_NEWLIB_VERSION) -# include <support/newlib/xlocale.h> +# include <__support/newlib/xlocale.h> +#elif defined(__OpenBSD__) +# include <__support/openbsd/xlocale.h> #elif (defined(__APPLE__) || defined(__FreeBSD__) \ || defined(__EMSCRIPTEN__) || defined(__IBMCPP__)) # include <xlocale.h> #elif defined(__Fuchsia__) -# include <support/fuchsia/xlocale.h> +# include <__support/fuchsia/xlocale.h> #elif defined(__wasi__) // WASI libc uses musl's locales support. -# include <support/musl/xlocale.h> +# include <__support/musl/xlocale.h> #elif defined(_LIBCPP_HAS_MUSL_LIBC) -# include <support/musl/xlocale.h> +# include <__support/musl/xlocale.h> #endif #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) @@ -76,7 +81,7 @@ struct __libcpp_locale_guard { // locale name, otherwise it will be a semicolon-separated string listing // each category. In the second case, we know at least one category won't // be what we want, so we only have to check the first case. - if (strcmp(__l.__get_locale(), __lc) != 0) { + if (_VSTD::strcmp(__l.__get_locale(), __lc) != 0) { __locale_all = _strdup(__lc); if (__locale_all == nullptr) __throw_bad_alloc(); @@ -105,7 +110,6 @@ struct __libcpp_locale_guard { }; #endif - class _LIBCPP_TYPE_VIS locale; template <class _Facet> @@ -335,8 +339,8 @@ collate<_CharT>::do_hash(const char_type* __lo, const char_type* __hi) const return static_cast<long>(__h); } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS collate<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS collate<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS collate<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS collate<wchar_t>) // template <class CharT> class collate_byname; @@ -396,7 +400,26 @@ locale::operator()(const basic_string<_CharT, _Traits, _Allocator>& __x, class _LIBCPP_TYPE_VIS ctype_base { public: -#if defined(__GLIBC__) +#if defined(_LIBCPP_PROVIDES_DEFAULT_RUNE_TABLE) + typedef unsigned long mask; + static const mask space = 1<<0; + static const mask print = 1<<1; + static const mask cntrl = 1<<2; + static const mask upper = 1<<3; + static const mask lower = 1<<4; + static const mask alpha = 1<<5; + static const mask digit = 1<<6; + static const mask punct = 1<<7; + static const mask xdigit = 1<<8; + static const mask blank = 1<<9; +#if defined(__BIONIC__) + // Historically this was a part of regex_traits rather than ctype_base. The + // historical value of the constant is preserved for ABI compatibility. + static const mask __regex_word = 0x8000; +#else + static const mask __regex_word = 1<<10; +#endif // defined(__BIONIC__) +#elif defined(__GLIBC__) typedef unsigned short mask; static const mask space = _ISspace; static const mask print = _ISprint; @@ -485,24 +508,7 @@ public: # define _LIBCPP_CTYPE_MASK_IS_COMPOSITE_ALPHA # define _LIBCPP_CTYPE_MASK_IS_COMPOSITE_XDIGIT #else - typedef unsigned long mask; - static const mask space = 1<<0; - static const mask print = 1<<1; - static const mask cntrl = 1<<2; - static const mask upper = 1<<3; - static const mask lower = 1<<4; - static const mask alpha = 1<<5; - static const mask digit = 1<<6; - static const mask punct = 1<<7; - static const mask xdigit = 1<<8; - static const mask blank = 1<<9; -#if defined(__BIONIC__) - // Historically this was a part of regex_traits rather than ctype_base. The - // historical value of the constant is preserved for ABI compatibility. - static const mask __regex_word = 0x8000; -#else - static const mask __regex_word = 1<<10; -#endif // defined(__BIONIC__) +# error unknown rune table for this platform -- do you mean to define _LIBCPP_PROVIDES_DEFAULT_RUNE_TABLE? #endif static const mask alnum = alpha | digit; static const mask graph = alnum | punct; @@ -623,7 +629,7 @@ class _LIBCPP_TYPE_VIS ctype<char> public: typedef char char_type; - explicit ctype(const mask* __tab = 0, bool __del = false, size_t __refs = 0); + explicit ctype(const mask* __tab = nullptr, bool __del = false, size_t __refs = 0); _LIBCPP_INLINE_VISIBILITY bool is(mask __m, char_type __c) const @@ -1069,10 +1075,10 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; -// template <> class codecvt<char16_t, char, mbstate_t> +// template <> class codecvt<char16_t, char, mbstate_t> // deprecated in C++20 template <> -class _LIBCPP_TYPE_VIS codecvt<char16_t, char, mbstate_t> +class _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_TYPE_VIS codecvt<char16_t, char, mbstate_t> : public locale::facet, public codecvt_base { @@ -1155,10 +1161,100 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; -// template <> class codecvt<char32_t, char, mbstate_t> +#ifndef _LIBCPP_NO_HAS_CHAR8_T + +// template <> class codecvt<char16_t, char8_t, mbstate_t> // C++20 template <> -class _LIBCPP_TYPE_VIS codecvt<char32_t, char, mbstate_t> +class _LIBCPP_TYPE_VIS codecvt<char16_t, char8_t, mbstate_t> + : public locale::facet, + public codecvt_base +{ +public: + typedef char16_t intern_type; + typedef char8_t extern_type; + typedef mbstate_t state_type; + + _LIBCPP_INLINE_VISIBILITY + explicit codecvt(size_t __refs = 0) + : locale::facet(__refs) {} + + _LIBCPP_INLINE_VISIBILITY + result out(state_type& __st, + const intern_type* __frm, const intern_type* __frm_end, const intern_type*& __frm_nxt, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const + { + return do_out(__st, __frm, __frm_end, __frm_nxt, __to, __to_end, __to_nxt); + } + + _LIBCPP_INLINE_VISIBILITY + result unshift(state_type& __st, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const + { + return do_unshift(__st, __to, __to_end, __to_nxt); + } + + _LIBCPP_INLINE_VISIBILITY + result in(state_type& __st, + const extern_type* __frm, const extern_type* __frm_end, const extern_type*& __frm_nxt, + intern_type* __to, intern_type* __to_end, intern_type*& __to_nxt) const + { + return do_in(__st, __frm, __frm_end, __frm_nxt, __to, __to_end, __to_nxt); + } + + _LIBCPP_INLINE_VISIBILITY + int encoding() const _NOEXCEPT + { + return do_encoding(); + } + + _LIBCPP_INLINE_VISIBILITY + bool always_noconv() const _NOEXCEPT + { + return do_always_noconv(); + } + + _LIBCPP_INLINE_VISIBILITY + int length(state_type& __st, const extern_type* __frm, const extern_type* __end, size_t __mx) const + { + return do_length(__st, __frm, __end, __mx); + } + + _LIBCPP_INLINE_VISIBILITY + int max_length() const _NOEXCEPT + { + return do_max_length(); + } + + static locale::id id; + +protected: + _LIBCPP_INLINE_VISIBILITY + explicit codecvt(const char*, size_t __refs = 0) + : locale::facet(__refs) {} + + ~codecvt(); + + virtual result do_out(state_type& __st, + const intern_type* __frm, const intern_type* __frm_end, const intern_type*& __frm_nxt, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const; + virtual result do_in(state_type& __st, + const extern_type* __frm, const extern_type* __frm_end, const extern_type*& __frm_nxt, + intern_type* __to, intern_type* __to_end, intern_type*& __to_nxt) const; + virtual result do_unshift(state_type& __st, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const; + virtual int do_encoding() const _NOEXCEPT; + virtual bool do_always_noconv() const _NOEXCEPT; + virtual int do_length(state_type&, const extern_type* __frm, const extern_type* __end, size_t __mx) const; + virtual int do_max_length() const _NOEXCEPT; +}; + +#endif + +// template <> class codecvt<char32_t, char, mbstate_t> // deprecated in C++20 + +template <> +class _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_TYPE_VIS codecvt<char32_t, char, mbstate_t> : public locale::facet, public codecvt_base { @@ -1241,6 +1337,96 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +#ifndef _LIBCPP_NO_HAS_CHAR8_T + +// template <> class codecvt<char32_t, char8_t, mbstate_t> // C++20 + +template <> +class _LIBCPP_TYPE_VIS codecvt<char32_t, char8_t, mbstate_t> + : public locale::facet, + public codecvt_base +{ +public: + typedef char32_t intern_type; + typedef char8_t extern_type; + typedef mbstate_t state_type; + + _LIBCPP_INLINE_VISIBILITY + explicit codecvt(size_t __refs = 0) + : locale::facet(__refs) {} + + _LIBCPP_INLINE_VISIBILITY + result out(state_type& __st, + const intern_type* __frm, const intern_type* __frm_end, const intern_type*& __frm_nxt, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const + { + return do_out(__st, __frm, __frm_end, __frm_nxt, __to, __to_end, __to_nxt); + } + + _LIBCPP_INLINE_VISIBILITY + result unshift(state_type& __st, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const + { + return do_unshift(__st, __to, __to_end, __to_nxt); + } + + _LIBCPP_INLINE_VISIBILITY + result in(state_type& __st, + const extern_type* __frm, const extern_type* __frm_end, const extern_type*& __frm_nxt, + intern_type* __to, intern_type* __to_end, intern_type*& __to_nxt) const + { + return do_in(__st, __frm, __frm_end, __frm_nxt, __to, __to_end, __to_nxt); + } + + _LIBCPP_INLINE_VISIBILITY + int encoding() const _NOEXCEPT + { + return do_encoding(); + } + + _LIBCPP_INLINE_VISIBILITY + bool always_noconv() const _NOEXCEPT + { + return do_always_noconv(); + } + + _LIBCPP_INLINE_VISIBILITY + int length(state_type& __st, const extern_type* __frm, const extern_type* __end, size_t __mx) const + { + return do_length(__st, __frm, __end, __mx); + } + + _LIBCPP_INLINE_VISIBILITY + int max_length() const _NOEXCEPT + { + return do_max_length(); + } + + static locale::id id; + +protected: + _LIBCPP_INLINE_VISIBILITY + explicit codecvt(const char*, size_t __refs = 0) + : locale::facet(__refs) {} + + ~codecvt(); + + virtual result do_out(state_type& __st, + const intern_type* __frm, const intern_type* __frm_end, const intern_type*& __frm_nxt, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const; + virtual result do_in(state_type& __st, + const extern_type* __frm, const extern_type* __frm_end, const extern_type*& __frm_nxt, + intern_type* __to, intern_type* __to_end, intern_type*& __to_nxt) const; + virtual result do_unshift(state_type& __st, + extern_type* __to, extern_type* __to_end, extern_type*& __to_nxt) const; + virtual int do_encoding() const _NOEXCEPT; + virtual bool do_always_noconv() const _NOEXCEPT; + virtual int do_length(state_type&, const extern_type* __frm, const extern_type* __end, size_t __mx) const; + virtual int do_max_length() const _NOEXCEPT; +}; + +#endif + // template <class _InternT, class _ExternT, class _StateT> class codecvt_byname template <class _InternT, class _ExternT, class _StateT> @@ -1258,15 +1444,21 @@ protected: ~codecvt_byname(); }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <class _InternT, class _ExternT, class _StateT> codecvt_byname<_InternT, _ExternT, _StateT>::~codecvt_byname() { } - -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char, char, mbstate_t>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<wchar_t, char, mbstate_t>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char16_t, char, mbstate_t>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char32_t, char, mbstate_t>) +_LIBCPP_SUPPRESS_DEPRECATED_POP + +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char, char, mbstate_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<wchar_t, char, mbstate_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char16_t, char, mbstate_t>) // deprecated in C++20 +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char32_t, char, mbstate_t>) // deprecated in C++20 +#ifndef _LIBCPP_NO_HAS_CHAR8_T +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char16_t, char8_t, mbstate_t>) // C++20 +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS codecvt_byname<char32_t, char8_t, mbstate_t>) // C++20 +#endif template <size_t _Np> struct __narrow_to_utf8 @@ -1290,12 +1482,14 @@ struct __narrow_to_utf8<8> } }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> struct _LIBCPP_TEMPLATE_VIS __narrow_to_utf8<16> : public codecvt<char16_t, char, mbstate_t> { _LIBCPP_INLINE_VISIBILITY __narrow_to_utf8() : codecvt<char16_t, char, mbstate_t>(1) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP _LIBCPP_EXPORTED_FROM_ABI ~__narrow_to_utf8(); @@ -1324,12 +1518,14 @@ struct _LIBCPP_TEMPLATE_VIS __narrow_to_utf8<16> } }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> struct _LIBCPP_TEMPLATE_VIS __narrow_to_utf8<32> : public codecvt<char32_t, char, mbstate_t> { _LIBCPP_INLINE_VISIBILITY __narrow_to_utf8() : codecvt<char32_t, char, mbstate_t>(1) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP _LIBCPP_EXPORTED_FROM_ABI ~__narrow_to_utf8(); @@ -1380,12 +1576,14 @@ struct __widen_from_utf8<8> } }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> struct _LIBCPP_TEMPLATE_VIS __widen_from_utf8<16> : public codecvt<char16_t, char, mbstate_t> { _LIBCPP_INLINE_VISIBILITY __widen_from_utf8() : codecvt<char16_t, char, mbstate_t>(1) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP _LIBCPP_EXPORTED_FROM_ABI ~__widen_from_utf8(); @@ -1407,19 +1605,21 @@ struct _LIBCPP_TEMPLATE_VIS __widen_from_utf8<16> if (__r == codecvt_base::error || __nn == __nb) __throw_runtime_error("locale not supported"); for (const char16_t* __p = __buf; __p < __bn; ++__p, ++__s) - *__s = (wchar_t)*__p; + *__s = *__p; __nb = __nn; } return __s; } }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> struct _LIBCPP_TEMPLATE_VIS __widen_from_utf8<32> : public codecvt<char32_t, char, mbstate_t> { _LIBCPP_INLINE_VISIBILITY __widen_from_utf8() : codecvt<char32_t, char, mbstate_t>(1) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP _LIBCPP_EXPORTED_FROM_ABI ~__widen_from_utf8(); @@ -1441,7 +1641,7 @@ struct _LIBCPP_TEMPLATE_VIS __widen_from_utf8<32> if (__r == codecvt_base::error || __nn == __nb) __throw_runtime_error("locale not supported"); for (const char32_t* __p = __buf; __p < __bn; ++__p, ++__s) - *__s = (wchar_t)*__p; + *__s = *__p; __nb = __nn; } return __s; diff --git a/lib/libcxx/include/__memory/allocator_traits.h b/lib/libcxx/include/__memory/allocator_traits.h new file mode 100644 index 0000000000..9443f61b71 --- /dev/null +++ b/lib/libcxx/include/__memory/allocator_traits.h @@ -0,0 +1,401 @@ +// -*- C++ -*- +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP___MEMORY_ALLOCATOR_TRAITS_H +#define _LIBCPP___MEMORY_ALLOCATOR_TRAITS_H + +#include <__config> +#include <__memory/base.h> +#include <__memory/pointer_traits.h> +#include <type_traits> + +#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) +#pragma GCC system_header +#endif + +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + +_LIBCPP_BEGIN_NAMESPACE_STD + +#define _LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(NAME, PROPERTY) \ + template <class _Tp, class = void> struct NAME : false_type { }; \ + template <class _Tp> struct NAME<_Tp, typename __void_t<typename _Tp:: PROPERTY >::type> : true_type { } + +// __pointer +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_pointer, pointer); +template <class _Tp, class _Alloc, + class _RawAlloc = typename remove_reference<_Alloc>::type, + bool = __has_pointer<_RawAlloc>::value> +struct __pointer { + using type _LIBCPP_NODEBUG_TYPE = typename _RawAlloc::pointer; +}; +template <class _Tp, class _Alloc, class _RawAlloc> +struct __pointer<_Tp, _Alloc, _RawAlloc, false> { + using type _LIBCPP_NODEBUG_TYPE = _Tp*; +}; + +// __const_pointer +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_const_pointer, const_pointer); +template <class _Tp, class _Ptr, class _Alloc, + bool = __has_const_pointer<_Alloc>::value> +struct __const_pointer { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::const_pointer; +}; +template <class _Tp, class _Ptr, class _Alloc> +struct __const_pointer<_Tp, _Ptr, _Alloc, false> { +#ifdef _LIBCPP_CXX03_LANG + using type = typename pointer_traits<_Ptr>::template rebind<const _Tp>::other; +#else + using type _LIBCPP_NODEBUG_TYPE = typename pointer_traits<_Ptr>::template rebind<const _Tp>; +#endif +}; + +// __void_pointer +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_void_pointer, void_pointer); +template <class _Ptr, class _Alloc, + bool = __has_void_pointer<_Alloc>::value> +struct __void_pointer { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::void_pointer; +}; +template <class _Ptr, class _Alloc> +struct __void_pointer<_Ptr, _Alloc, false> { +#ifdef _LIBCPP_CXX03_LANG + using type _LIBCPP_NODEBUG_TYPE = typename pointer_traits<_Ptr>::template rebind<void>::other; +#else + using type _LIBCPP_NODEBUG_TYPE = typename pointer_traits<_Ptr>::template rebind<void>; +#endif +}; + +// __const_void_pointer +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_const_void_pointer, const_void_pointer); +template <class _Ptr, class _Alloc, + bool = __has_const_void_pointer<_Alloc>::value> +struct __const_void_pointer { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::const_void_pointer; +}; +template <class _Ptr, class _Alloc> +struct __const_void_pointer<_Ptr, _Alloc, false> { +#ifdef _LIBCPP_CXX03_LANG + using type _LIBCPP_NODEBUG_TYPE = typename pointer_traits<_Ptr>::template rebind<const void>::other; +#else + using type _LIBCPP_NODEBUG_TYPE = typename pointer_traits<_Ptr>::template rebind<const void>; +#endif +}; + +// __size_type +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_size_type, size_type); +template <class _Alloc, class _DiffType, bool = __has_size_type<_Alloc>::value> +struct __size_type : make_unsigned<_DiffType> { }; +template <class _Alloc, class _DiffType> +struct __size_type<_Alloc, _DiffType, true> { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::size_type; +}; + +// __alloc_traits_difference_type +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_alloc_traits_difference_type, difference_type); +template <class _Alloc, class _Ptr, bool = __has_alloc_traits_difference_type<_Alloc>::value> +struct __alloc_traits_difference_type { + using type _LIBCPP_NODEBUG_TYPE = typename pointer_traits<_Ptr>::difference_type; +}; +template <class _Alloc, class _Ptr> +struct __alloc_traits_difference_type<_Alloc, _Ptr, true> { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::difference_type; +}; + +// __propagate_on_container_copy_assignment +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_propagate_on_container_copy_assignment, propagate_on_container_copy_assignment); +template <class _Alloc, bool = __has_propagate_on_container_copy_assignment<_Alloc>::value> +struct __propagate_on_container_copy_assignment : false_type { }; +template <class _Alloc> +struct __propagate_on_container_copy_assignment<_Alloc, true> { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::propagate_on_container_copy_assignment; +}; + +// __propagate_on_container_move_assignment +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_propagate_on_container_move_assignment, propagate_on_container_move_assignment); +template <class _Alloc, bool = __has_propagate_on_container_move_assignment<_Alloc>::value> +struct __propagate_on_container_move_assignment : false_type { }; +template <class _Alloc> +struct __propagate_on_container_move_assignment<_Alloc, true> { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::propagate_on_container_move_assignment; +}; + +// __propagate_on_container_swap +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_propagate_on_container_swap, propagate_on_container_swap); +template <class _Alloc, bool = __has_propagate_on_container_swap<_Alloc>::value> +struct __propagate_on_container_swap : false_type { }; +template <class _Alloc> +struct __propagate_on_container_swap<_Alloc, true> { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::propagate_on_container_swap; +}; + +// __is_always_equal +_LIBCPP_ALLOCATOR_TRAITS_HAS_XXX(__has_is_always_equal, is_always_equal); +template <class _Alloc, bool = __has_is_always_equal<_Alloc>::value> +struct __is_always_equal : is_empty<_Alloc> { }; +template <class _Alloc> +struct __is_always_equal<_Alloc, true> { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc::is_always_equal; +}; + +// __allocator_traits_rebind +_LIBCPP_SUPPRESS_DEPRECATED_PUSH +template <class _Tp, class _Up, class = void> +struct __has_rebind_other : false_type { }; +template <class _Tp, class _Up> +struct __has_rebind_other<_Tp, _Up, typename __void_t< + typename _Tp::template rebind<_Up>::other +>::type> : true_type { }; + +template <class _Tp, class _Up, bool = __has_rebind_other<_Tp, _Up>::value> +struct __allocator_traits_rebind { + using type _LIBCPP_NODEBUG_TYPE = typename _Tp::template rebind<_Up>::other; +}; +template <template <class, class...> class _Alloc, class _Tp, class ..._Args, class _Up> +struct __allocator_traits_rebind<_Alloc<_Tp, _Args...>, _Up, true> { + using type _LIBCPP_NODEBUG_TYPE = typename _Alloc<_Tp, _Args...>::template rebind<_Up>::other; +}; +template <template <class, class...> class _Alloc, class _Tp, class ..._Args, class _Up> +struct __allocator_traits_rebind<_Alloc<_Tp, _Args...>, _Up, false> { + using type _LIBCPP_NODEBUG_TYPE = _Alloc<_Up, _Args...>; +}; +_LIBCPP_SUPPRESS_DEPRECATED_POP + +template<class _Alloc, class _Tp> +using __allocator_traits_rebind_t = typename __allocator_traits_rebind<_Alloc, _Tp>::type; + +_LIBCPP_SUPPRESS_DEPRECATED_PUSH + +// __has_allocate_hint +template <class _Alloc, class _SizeType, class _ConstVoidPtr, class = void> +struct __has_allocate_hint : false_type { }; + +template <class _Alloc, class _SizeType, class _ConstVoidPtr> +struct __has_allocate_hint<_Alloc, _SizeType, _ConstVoidPtr, decltype( + (void)declval<_Alloc>().allocate(declval<_SizeType>(), declval<_ConstVoidPtr>()) +)> : true_type { }; + +// __has_construct +template <class, class _Alloc, class ..._Args> +struct __has_construct_impl : false_type { }; + +template <class _Alloc, class ..._Args> +struct __has_construct_impl<decltype( + (void)declval<_Alloc>().construct(declval<_Args>()...) +), _Alloc, _Args...> : true_type { }; + +template <class _Alloc, class ..._Args> +struct __has_construct : __has_construct_impl<void, _Alloc, _Args...> { }; + +// __has_destroy +template <class _Alloc, class _Pointer, class = void> +struct __has_destroy : false_type { }; + +template <class _Alloc, class _Pointer> +struct __has_destroy<_Alloc, _Pointer, decltype( + (void)declval<_Alloc>().destroy(declval<_Pointer>()) +)> : true_type { }; + +// __has_max_size +template <class _Alloc, class = void> +struct __has_max_size : false_type { }; + +template <class _Alloc> +struct __has_max_size<_Alloc, decltype( + (void)declval<_Alloc&>().max_size() +)> : true_type { }; + +// __has_select_on_container_copy_construction +template <class _Alloc, class = void> +struct __has_select_on_container_copy_construction : false_type { }; + +template <class _Alloc> +struct __has_select_on_container_copy_construction<_Alloc, decltype( + (void)declval<_Alloc>().select_on_container_copy_construction() +)> : true_type { }; + +_LIBCPP_SUPPRESS_DEPRECATED_POP + +template <class _Alloc> +struct _LIBCPP_TEMPLATE_VIS allocator_traits +{ + using allocator_type = _Alloc; + using value_type = typename allocator_type::value_type; + using pointer = typename __pointer<value_type, allocator_type>::type; + using const_pointer = typename __const_pointer<value_type, pointer, allocator_type>::type; + using void_pointer = typename __void_pointer<pointer, allocator_type>::type; + using const_void_pointer = typename __const_void_pointer<pointer, allocator_type>::type; + using difference_type = typename __alloc_traits_difference_type<allocator_type, pointer>::type; + using size_type = typename __size_type<allocator_type, difference_type>::type; + using propagate_on_container_copy_assignment = typename __propagate_on_container_copy_assignment<allocator_type>::type; + using propagate_on_container_move_assignment = typename __propagate_on_container_move_assignment<allocator_type>::type; + using propagate_on_container_swap = typename __propagate_on_container_swap<allocator_type>::type; + using is_always_equal = typename __is_always_equal<allocator_type>::type; + +#ifndef _LIBCPP_CXX03_LANG + template <class _Tp> + using rebind_alloc = __allocator_traits_rebind_t<allocator_type, _Tp>; + template <class _Tp> + using rebind_traits = allocator_traits<rebind_alloc<_Tp> >; +#else // _LIBCPP_CXX03_LANG + template <class _Tp> + struct rebind_alloc { + using other = __allocator_traits_rebind_t<allocator_type, _Tp>; + }; + template <class _Tp> + struct rebind_traits { + using other = allocator_traits<typename rebind_alloc<_Tp>::other>; + }; +#endif // _LIBCPP_CXX03_LANG + + _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static pointer allocate(allocator_type& __a, size_type __n) { + return __a.allocate(__n); + } + + template <class _Ap = _Alloc, class = + _EnableIf<__has_allocate_hint<_Ap, size_type, const_void_pointer>::value> > + _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static pointer allocate(allocator_type& __a, size_type __n, const_void_pointer __hint) { + _LIBCPP_SUPPRESS_DEPRECATED_PUSH + return __a.allocate(__n, __hint); + _LIBCPP_SUPPRESS_DEPRECATED_POP + } + template <class _Ap = _Alloc, class = void, class = + _EnableIf<!__has_allocate_hint<_Ap, size_type, const_void_pointer>::value> > + _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static pointer allocate(allocator_type& __a, size_type __n, const_void_pointer) { + return __a.allocate(__n); + } + + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static void deallocate(allocator_type& __a, pointer __p, size_type __n) _NOEXCEPT { + __a.deallocate(__p, __n); + } + + template <class _Tp, class... _Args, class = + _EnableIf<__has_construct<allocator_type, _Tp*, _Args...>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static void construct(allocator_type& __a, _Tp* __p, _Args&&... __args) { + _LIBCPP_SUPPRESS_DEPRECATED_PUSH + __a.construct(__p, _VSTD::forward<_Args>(__args)...); + _LIBCPP_SUPPRESS_DEPRECATED_POP + } + template <class _Tp, class... _Args, class = void, class = + _EnableIf<!__has_construct<allocator_type, _Tp*, _Args...>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static void construct(allocator_type&, _Tp* __p, _Args&&... __args) { +#if _LIBCPP_STD_VER > 17 + _VSTD::construct_at(__p, _VSTD::forward<_Args>(__args)...); +#else + ::new ((void*)__p) _Tp(_VSTD::forward<_Args>(__args)...); +#endif + } + + template <class _Tp, class = + _EnableIf<__has_destroy<allocator_type, _Tp*>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static void destroy(allocator_type& __a, _Tp* __p) { + _LIBCPP_SUPPRESS_DEPRECATED_PUSH + __a.destroy(__p); + _LIBCPP_SUPPRESS_DEPRECATED_POP + } + template <class _Tp, class = void, class = + _EnableIf<!__has_destroy<allocator_type, _Tp*>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static void destroy(allocator_type&, _Tp* __p) { +#if _LIBCPP_STD_VER > 17 + _VSTD::destroy_at(__p); +#else + __p->~_Tp(); +#endif + } + + template <class _Ap = _Alloc, class = + _EnableIf<__has_max_size<const _Ap>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static size_type max_size(const allocator_type& __a) _NOEXCEPT { + _LIBCPP_SUPPRESS_DEPRECATED_PUSH + return __a.max_size(); + _LIBCPP_SUPPRESS_DEPRECATED_POP + } + template <class _Ap = _Alloc, class = void, class = + _EnableIf<!__has_max_size<const _Ap>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static size_type max_size(const allocator_type&) _NOEXCEPT { + return numeric_limits<size_type>::max() / sizeof(value_type); + } + + template <class _Ap = _Alloc, class = + _EnableIf<__has_select_on_container_copy_construction<const _Ap>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static allocator_type select_on_container_copy_construction(const allocator_type& __a) { + return __a.select_on_container_copy_construction(); + } + template <class _Ap = _Alloc, class = void, class = + _EnableIf<!__has_select_on_container_copy_construction<const _Ap>::value> > + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static allocator_type select_on_container_copy_construction(const allocator_type& __a) { + return __a; + } +}; + +template <class _Traits, class _Tp> +struct __rebind_alloc_helper { +#ifndef _LIBCPP_CXX03_LANG + using type _LIBCPP_NODEBUG_TYPE = typename _Traits::template rebind_alloc<_Tp>; +#else + using type = typename _Traits::template rebind_alloc<_Tp>::other; +#endif +}; + +// __is_default_allocator +template <class _Tp> +struct __is_default_allocator : false_type { }; + +template <class _Tp> +struct __is_default_allocator<_VSTD::allocator<_Tp> > : true_type { }; + +// __is_cpp17_move_insertable +template <class _Alloc, class = void> +struct __is_cpp17_move_insertable + : is_move_constructible<typename _Alloc::value_type> +{ }; + +template <class _Alloc> +struct __is_cpp17_move_insertable<_Alloc, _EnableIf< + !__is_default_allocator<_Alloc>::value && + __has_construct<_Alloc, typename _Alloc::value_type*, typename _Alloc::value_type&&>::value +> > : true_type { }; + +// __is_cpp17_copy_insertable +template <class _Alloc, class = void> +struct __is_cpp17_copy_insertable + : integral_constant<bool, + is_copy_constructible<typename _Alloc::value_type>::value && + __is_cpp17_move_insertable<_Alloc>::value + > +{ }; + +template <class _Alloc> +struct __is_cpp17_copy_insertable<_Alloc, _EnableIf< + !__is_default_allocator<_Alloc>::value && + __has_construct<_Alloc, typename _Alloc::value_type*, const typename _Alloc::value_type&>::value +> > + : __is_cpp17_move_insertable<_Alloc> +{ }; + +#undef _LIBCPP_ALLOCATOR_TRAITS_HAS_XXX + +_LIBCPP_END_NAMESPACE_STD + +_LIBCPP_POP_MACROS + +#endif // _LIBCPP___MEMORY_ALLOCATOR_TRAITS_H diff --git a/lib/libcxx/include/__memory/base.h b/lib/libcxx/include/__memory/base.h new file mode 100644 index 0000000000..70728bd7f8 --- /dev/null +++ b/lib/libcxx/include/__memory/base.h @@ -0,0 +1,127 @@ +// -*- C++ -*- +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP___MEMORY_BASE_H +#define _LIBCPP___MEMORY_BASE_H + +#include <__config> +#include <__debug> +#include <type_traits> + +#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) +#pragma GCC system_header +#endif + +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + +_LIBCPP_BEGIN_NAMESPACE_STD + +// addressof +#ifndef _LIBCPP_HAS_NO_BUILTIN_ADDRESSOF + +template <class _Tp> +inline _LIBCPP_CONSTEXPR_AFTER_CXX14 +_LIBCPP_NO_CFI _LIBCPP_INLINE_VISIBILITY +_Tp* +addressof(_Tp& __x) _NOEXCEPT +{ + return __builtin_addressof(__x); +} + +#else + +template <class _Tp> +inline _LIBCPP_NO_CFI _LIBCPP_INLINE_VISIBILITY +_Tp* +addressof(_Tp& __x) _NOEXCEPT +{ + return reinterpret_cast<_Tp *>( + const_cast<char *>(&reinterpret_cast<const volatile char &>(__x))); +} + +#endif // _LIBCPP_HAS_NO_BUILTIN_ADDRESSOF + +#if defined(_LIBCPP_HAS_OBJC_ARC) && !defined(_LIBCPP_PREDEFINED_OBJC_ARC_ADDRESSOF) +// Objective-C++ Automatic Reference Counting uses qualified pointers +// that require special addressof() signatures. When +// _LIBCPP_PREDEFINED_OBJC_ARC_ADDRESSOF is defined, the compiler +// itself is providing these definitions. Otherwise, we provide them. +template <class _Tp> +inline _LIBCPP_INLINE_VISIBILITY +__strong _Tp* +addressof(__strong _Tp& __x) _NOEXCEPT +{ + return &__x; +} + +#ifdef _LIBCPP_HAS_OBJC_ARC_WEAK +template <class _Tp> +inline _LIBCPP_INLINE_VISIBILITY +__weak _Tp* +addressof(__weak _Tp& __x) _NOEXCEPT +{ + return &__x; +} +#endif + +template <class _Tp> +inline _LIBCPP_INLINE_VISIBILITY +__autoreleasing _Tp* +addressof(__autoreleasing _Tp& __x) _NOEXCEPT +{ + return &__x; +} + +template <class _Tp> +inline _LIBCPP_INLINE_VISIBILITY +__unsafe_unretained _Tp* +addressof(__unsafe_unretained _Tp& __x) _NOEXCEPT +{ + return &__x; +} +#endif + +#if !defined(_LIBCPP_CXX03_LANG) +template <class _Tp> _Tp* addressof(const _Tp&&) noexcept = delete; +#endif + +// construct_at + +#if _LIBCPP_STD_VER > 17 + +template<class _Tp, class ..._Args, class = decltype( + ::new (_VSTD::declval<void*>()) _Tp(_VSTD::declval<_Args>()...) +)> +_LIBCPP_INLINE_VISIBILITY +constexpr _Tp* construct_at(_Tp* __location, _Args&& ...__args) { + _LIBCPP_ASSERT(__location, "null pointer given to construct_at"); + return ::new ((void*)__location) _Tp(_VSTD::forward<_Args>(__args)...); +} + +#endif + +// destroy_at + +#if _LIBCPP_STD_VER > 14 + +template <class _Tp> +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 +void destroy_at(_Tp* __loc) { + _LIBCPP_ASSERT(__loc, "null pointer given to destroy_at"); + __loc->~_Tp(); +} + +#endif + +_LIBCPP_END_NAMESPACE_STD + +_LIBCPP_POP_MACROS + +#endif // _LIBCPP___MEMORY_BASE_H diff --git a/lib/libcxx/include/__memory/pointer_traits.h b/lib/libcxx/include/__memory/pointer_traits.h new file mode 100644 index 0000000000..b2c5d34cb0 --- /dev/null +++ b/lib/libcxx/include/__memory/pointer_traits.h @@ -0,0 +1,169 @@ +// -*- C++ -*- +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP___MEMORY_POINTER_TRAITS_H +#define _LIBCPP___MEMORY_POINTER_TRAITS_H + +#include <__config> +#include <type_traits> + +#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) +#pragma GCC system_header +#endif + +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + +_LIBCPP_BEGIN_NAMESPACE_STD + +template <class _Tp, class = void> +struct __has_element_type : false_type {}; + +template <class _Tp> +struct __has_element_type<_Tp, + typename __void_t<typename _Tp::element_type>::type> : true_type {}; + +template <class _Ptr, bool = __has_element_type<_Ptr>::value> +struct __pointer_traits_element_type; + +template <class _Ptr> +struct __pointer_traits_element_type<_Ptr, true> +{ + typedef _LIBCPP_NODEBUG_TYPE typename _Ptr::element_type type; +}; + +template <template <class, class...> class _Sp, class _Tp, class ..._Args> +struct __pointer_traits_element_type<_Sp<_Tp, _Args...>, true> +{ + typedef _LIBCPP_NODEBUG_TYPE typename _Sp<_Tp, _Args...>::element_type type; +}; + +template <template <class, class...> class _Sp, class _Tp, class ..._Args> +struct __pointer_traits_element_type<_Sp<_Tp, _Args...>, false> +{ + typedef _LIBCPP_NODEBUG_TYPE _Tp type; +}; + +template <class _Tp, class = void> +struct __has_difference_type : false_type {}; + +template <class _Tp> +struct __has_difference_type<_Tp, + typename __void_t<typename _Tp::difference_type>::type> : true_type {}; + +template <class _Ptr, bool = __has_difference_type<_Ptr>::value> +struct __pointer_traits_difference_type +{ + typedef _LIBCPP_NODEBUG_TYPE ptrdiff_t type; +}; + +template <class _Ptr> +struct __pointer_traits_difference_type<_Ptr, true> +{ + typedef _LIBCPP_NODEBUG_TYPE typename _Ptr::difference_type type; +}; + +template <class _Tp, class _Up> +struct __has_rebind +{ +private: + struct __two {char __lx; char __lxx;}; + template <class _Xp> static __two __test(...); + _LIBCPP_SUPPRESS_DEPRECATED_PUSH + template <class _Xp> static char __test(typename _Xp::template rebind<_Up>* = 0); + _LIBCPP_SUPPRESS_DEPRECATED_POP +public: + static const bool value = sizeof(__test<_Tp>(0)) == 1; +}; + +template <class _Tp, class _Up, bool = __has_rebind<_Tp, _Up>::value> +struct __pointer_traits_rebind +{ +#ifndef _LIBCPP_CXX03_LANG + typedef _LIBCPP_NODEBUG_TYPE typename _Tp::template rebind<_Up> type; +#else + typedef _LIBCPP_NODEBUG_TYPE typename _Tp::template rebind<_Up>::other type; +#endif +}; + +template <template <class, class...> class _Sp, class _Tp, class ..._Args, class _Up> +struct __pointer_traits_rebind<_Sp<_Tp, _Args...>, _Up, true> +{ +#ifndef _LIBCPP_CXX03_LANG + typedef _LIBCPP_NODEBUG_TYPE typename _Sp<_Tp, _Args...>::template rebind<_Up> type; +#else + typedef _LIBCPP_NODEBUG_TYPE typename _Sp<_Tp, _Args...>::template rebind<_Up>::other type; +#endif +}; + +template <template <class, class...> class _Sp, class _Tp, class ..._Args, class _Up> +struct __pointer_traits_rebind<_Sp<_Tp, _Args...>, _Up, false> +{ + typedef _Sp<_Up, _Args...> type; +}; + +template <class _Ptr> +struct _LIBCPP_TEMPLATE_VIS pointer_traits +{ + typedef _Ptr pointer; + typedef typename __pointer_traits_element_type<pointer>::type element_type; + typedef typename __pointer_traits_difference_type<pointer>::type difference_type; + +#ifndef _LIBCPP_CXX03_LANG + template <class _Up> using rebind = typename __pointer_traits_rebind<pointer, _Up>::type; +#else + template <class _Up> struct rebind + {typedef typename __pointer_traits_rebind<pointer, _Up>::type other;}; +#endif // _LIBCPP_CXX03_LANG + +private: + struct __nat {}; +public: + _LIBCPP_INLINE_VISIBILITY + static pointer pointer_to(typename conditional<is_void<element_type>::value, + __nat, element_type>::type& __r) + {return pointer::pointer_to(__r);} +}; + +template <class _Tp> +struct _LIBCPP_TEMPLATE_VIS pointer_traits<_Tp*> +{ + typedef _Tp* pointer; + typedef _Tp element_type; + typedef ptrdiff_t difference_type; + +#ifndef _LIBCPP_CXX03_LANG + template <class _Up> using rebind = _Up*; +#else + template <class _Up> struct rebind {typedef _Up* other;}; +#endif + +private: + struct __nat {}; +public: + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + static pointer pointer_to(typename conditional<is_void<element_type>::value, + __nat, element_type>::type& __r) _NOEXCEPT + {return _VSTD::addressof(__r);} +}; + +template <class _From, class _To> +struct __rebind_pointer { +#ifndef _LIBCPP_CXX03_LANG + typedef typename pointer_traits<_From>::template rebind<_To> type; +#else + typedef typename pointer_traits<_From>::template rebind<_To>::other type; +#endif +}; + +_LIBCPP_END_NAMESPACE_STD + +_LIBCPP_POP_MACROS + +#endif // _LIBCPP___MEMORY_POINTER_TRAITS_H diff --git a/lib/libcxx/include/__memory/utilities.h b/lib/libcxx/include/__memory/utilities.h new file mode 100644 index 0000000000..aac3d11cab --- /dev/null +++ b/lib/libcxx/include/__memory/utilities.h @@ -0,0 +1,88 @@ +// -*- C++ -*- +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP___MEMORY_UTILITIES_H +#define _LIBCPP___MEMORY_UTILITIES_H + +#include <__config> +#include <__memory/allocator_traits.h> +#include <cstddef> + +#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) +#pragma GCC system_header +#endif + +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + + +_LIBCPP_BEGIN_NAMESPACE_STD + +// Helper class to allocate memory using an Allocator in an exception safe +// manner. +// +// The intended usage of this class is as follows: +// +// 0 +// 1 __allocation_guard<SomeAllocator> guard(alloc, 10); +// 2 do_some_initialization_that_may_throw(guard.__get()); +// 3 save_allocated_pointer_in_a_noexcept_operation(guard.__release_ptr()); +// 4 +// +// If line (2) throws an exception during initialization of the memory, the +// guard's destructor will be called, and the memory will be released using +// Allocator deallocation. Otherwise, we release the memory from the guard on +// line (3) in an operation that can't throw -- after that, the guard is not +// responsible for the memory anymore. +// +// This is similar to a unique_ptr, except it's easier to use with a +// custom allocator. +template<class _Alloc> +struct __allocation_guard { + using _Pointer = typename allocator_traits<_Alloc>::pointer; + using _Size = typename allocator_traits<_Alloc>::size_type; + + template<class _AllocT> // we perform the allocator conversion inside the constructor + _LIBCPP_HIDE_FROM_ABI + explicit __allocation_guard(_AllocT __alloc, _Size __n) + : __alloc_(_VSTD::move(__alloc)) + , __n_(__n) + , __ptr_(allocator_traits<_Alloc>::allocate(__alloc_, __n_)) // initialization order is important + { } + + _LIBCPP_HIDE_FROM_ABI + ~__allocation_guard() _NOEXCEPT { + if (__ptr_ != nullptr) { + allocator_traits<_Alloc>::deallocate(__alloc_, __ptr_, __n_); + } + } + + _LIBCPP_HIDE_FROM_ABI + _Pointer __release_ptr() _NOEXCEPT { // not called __release() because it's a keyword in objective-c++ + _Pointer __tmp = __ptr_; + __ptr_ = nullptr; + return __tmp; + } + + _LIBCPP_HIDE_FROM_ABI + _Pointer __get() const _NOEXCEPT { + return __ptr_; + } + +private: + _Alloc __alloc_; + _Size __n_; + _Pointer __ptr_; +}; + +_LIBCPP_END_NAMESPACE_STD + +_LIBCPP_POP_MACROS + +#endif // _LIBCPP___MEMORY_UTILITIES_H diff --git a/lib/libcxx/include/__mutex_base b/lib/libcxx/include/__mutex_base index 8b4b74802b..96454ace9f 100644 --- a/lib/libcxx/include/__mutex_base +++ b/lib/libcxx/include/__mutex_base @@ -146,7 +146,6 @@ private: unique_lock& operator=(unique_lock const&); // = delete; public: -#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY unique_lock(unique_lock&& __u) _NOEXCEPT : __m_(__u.__m_), __owns_(__u.__owns_) @@ -163,8 +162,6 @@ public: return *this; } -#endif // _LIBCPP_CXX03_LANG - void lock(); bool try_lock(); @@ -382,12 +379,12 @@ __safe_nanosecond_cast(chrono::duration<_Rep, _Period> __d) using __ratio = ratio_divide<_Period, nano>; using __ns_rep = nanoseconds::rep; - __ns_rep __result_max = std::numeric_limits<__ns_rep>::max(); + __ns_rep __result_max = numeric_limits<__ns_rep>::max(); if (__d.count() > 0 && __d.count() > __result_max / __ratio::num) { return nanoseconds::max(); } - __ns_rep __result_min = std::numeric_limits<__ns_rep>::min(); + __ns_rep __result_min = numeric_limits<__ns_rep>::min(); if (__d.count() < 0 && __d.count() < __result_min / __ratio::num) { return nanoseconds::min(); } @@ -421,7 +418,7 @@ condition_variable::wait_until(unique_lock<mutex>& __lk, if (__t <= __now) return cv_status::timeout; - __clock_tp_ns __t_ns = __clock_tp_ns(__safe_nanosecond_cast(__t.time_since_epoch())); + __clock_tp_ns __t_ns = __clock_tp_ns(_VSTD::__safe_nanosecond_cast(__t.time_since_epoch())); __do_timed_wait(__lk, __t_ns); return _Clock::now() < __t ? cv_status::no_timeout : cv_status::timeout; @@ -454,13 +451,13 @@ condition_variable::wait_for(unique_lock<mutex>& __lk, #if defined(_LIBCPP_HAS_COND_CLOCKWAIT) using __clock_tp_ns = time_point<steady_clock, nanoseconds>; - __ns_rep __now_count_ns = __safe_nanosecond_cast(__c_now.time_since_epoch()).count(); + __ns_rep __now_count_ns = _VSTD::__safe_nanosecond_cast(__c_now.time_since_epoch()).count(); #else using __clock_tp_ns = time_point<system_clock, nanoseconds>; - __ns_rep __now_count_ns = __safe_nanosecond_cast(system_clock::now().time_since_epoch()).count(); + __ns_rep __now_count_ns = _VSTD::__safe_nanosecond_cast(system_clock::now().time_since_epoch()).count(); #endif - __ns_rep __d_ns_count = __safe_nanosecond_cast(__d).count(); + __ns_rep __d_ns_count = _VSTD::__safe_nanosecond_cast(__d).count(); if (__now_count_ns > numeric_limits<__ns_rep>::max() - __d_ns_count) { __do_timed_wait(__lk, __clock_tp_ns::max()); diff --git a/lib/libcxx/include/__split_buffer b/lib/libcxx/include/__split_buffer index fce209f828..20480d19d3 100644 --- a/lib/libcxx/include/__split_buffer +++ b/lib/libcxx/include/__split_buffer @@ -68,7 +68,6 @@ public: __split_buffer(size_type __cap, size_type __start, __alloc_rr& __a); ~__split_buffer(); -#ifndef _LIBCPP_CXX03_LANG __split_buffer(__split_buffer&& __c) _NOEXCEPT_(is_nothrow_move_constructible<allocator_type>::value); __split_buffer(__split_buffer&& __c, const __alloc_rr& __a); @@ -76,7 +75,6 @@ public: _NOEXCEPT_((__alloc_traits::propagate_on_container_move_assignment::value && is_nothrow_move_assignable<allocator_type>::value) || !__alloc_traits::propagate_on_container_move_assignment::value); -#endif // _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY iterator begin() _NOEXCEPT {return __begin_;} _LIBCPP_INLINE_VISIBILITY const_iterator begin() const _NOEXCEPT {return __begin_;} @@ -101,12 +99,10 @@ public: void shrink_to_fit() _NOEXCEPT; void push_front(const_reference __x); _LIBCPP_INLINE_VISIBILITY void push_back(const_reference __x); -#ifndef _LIBCPP_CXX03_LANG void push_front(value_type&& __x); void push_back(value_type&& __x); template <class... _Args> void emplace_back(_Args&&... __args); -#endif // !defined(_LIBCPP_CXX03_LANG) _LIBCPP_INLINE_VISIBILITY void pop_front() {__destruct_at_begin(__begin_+1);} _LIBCPP_INLINE_VISIBILITY void pop_back() {__destruct_at_end(__end_-1);} @@ -270,7 +266,7 @@ typename enable_if >::type __split_buffer<_Tp, _Allocator>::__construct_at_end(_ForwardIterator __first, _ForwardIterator __last) { - _ConstructTransaction __tx(&this->__end_, std::distance(__first, __last)); + _ConstructTransaction __tx(&this->__end_, _VSTD::distance(__first, __last)); for (; __tx.__pos_ != __tx.__end_; ++__tx.__pos_, ++__first) { __alloc_traits::construct(this->__alloc(), _VSTD::__to_address(__tx.__pos_), *__first); @@ -283,7 +279,7 @@ void __split_buffer<_Tp, _Allocator>::__destruct_at_begin(pointer __new_begin, false_type) { while (__begin_ != __new_begin) - __alloc_traits::destroy(__alloc(), __to_address(__begin_++)); + __alloc_traits::destroy(__alloc(), _VSTD::__to_address(__begin_++)); } template <class _Tp, class _Allocator> @@ -300,7 +296,7 @@ void __split_buffer<_Tp, _Allocator>::__destruct_at_end(pointer __new_last, false_type) _NOEXCEPT { while (__new_last != __end_) - __alloc_traits::destroy(__alloc(), __to_address(--__end_)); + __alloc_traits::destroy(__alloc(), _VSTD::__to_address(--__end_)); } template <class _Tp, class _Allocator> @@ -350,8 +346,6 @@ __split_buffer<_Tp, _Allocator>::~__split_buffer() __alloc_traits::deallocate(__alloc(), __first_, capacity()); } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Allocator> __split_buffer<_Tp, _Allocator>::__split_buffer(__split_buffer&& __c) _NOEXCEPT_(is_nothrow_move_constructible<allocator_type>::value) @@ -412,8 +406,6 @@ __split_buffer<_Tp, _Allocator>::operator=(__split_buffer&& __c) return *this; } -#endif // _LIBCPP_CXX03_LANG - template <class _Tp, class _Allocator> void __split_buffer<_Tp, _Allocator>::swap(__split_buffer& __x) @@ -424,7 +416,7 @@ __split_buffer<_Tp, _Allocator>::swap(__split_buffer& __x) _VSTD::swap(__begin_, __x.__begin_); _VSTD::swap(__end_, __x.__end_); _VSTD::swap(__end_cap(), __x.__end_cap()); - __swap_allocator(__alloc(), __x.__alloc()); + _VSTD::__swap_allocator(__alloc(), __x.__alloc()); } template <class _Tp, class _Allocator> @@ -499,8 +491,6 @@ __split_buffer<_Tp, _Allocator>::push_front(const_reference __x) --__begin_; } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Allocator> void __split_buffer<_Tp, _Allocator>::push_front(value_type&& __x) @@ -531,8 +521,6 @@ __split_buffer<_Tp, _Allocator>::push_front(value_type&& __x) --__begin_; } -#endif // _LIBCPP_CXX03_LANG - template <class _Tp, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY void @@ -563,8 +551,6 @@ __split_buffer<_Tp, _Allocator>::push_back(const_reference __x) ++__end_; } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Allocator> void __split_buffer<_Tp, _Allocator>::push_back(value_type&& __x) @@ -626,8 +612,6 @@ __split_buffer<_Tp, _Allocator>::emplace_back(_Args&&... __args) ++__end_; } -#endif // _LIBCPP_CXX03_LANG - template <class _Tp, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY void diff --git a/lib/libcxx/include/__sso_allocator b/lib/libcxx/include/__sso_allocator index 3930128738..c50ae24c92 100644 --- a/lib/libcxx/include/__sso_allocator +++ b/lib/libcxx/include/__sso_allocator @@ -11,8 +11,9 @@ #define _LIBCPP___SSO_ALLOCATOR #include <__config> -#include <type_traits> +#include <memory> #include <new> +#include <type_traits> #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header @@ -47,21 +48,21 @@ public: private: __sso_allocator& operator=(const __sso_allocator&); public: - _LIBCPP_INLINE_VISIBILITY pointer allocate(size_type __n, typename __sso_allocator<void, _Np>::const_pointer = 0) + _LIBCPP_INLINE_VISIBILITY pointer allocate(size_type __n, typename __sso_allocator<void, _Np>::const_pointer = nullptr) { if (!__allocated_ && __n <= _Np) { __allocated_ = true; return (pointer)&buf_; } - return static_cast<pointer>(_VSTD::__libcpp_allocate(__n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp))); + return allocator<_Tp>().allocate(__n); } _LIBCPP_INLINE_VISIBILITY void deallocate(pointer __p, size_type __n) { if (__p == (pointer)&buf_) __allocated_ = false; else - _VSTD::__libcpp_deallocate(__p, __n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp)); + allocator<_Tp>().deallocate(__p, __n); } _LIBCPP_INLINE_VISIBILITY size_type max_size() const throw() {return size_type(~0) / sizeof(_Tp);} diff --git a/lib/libcxx/include/__string b/lib/libcxx/include/__string index 9060bf98ad..d8b672e4c1 100644 --- a/lib/libcxx/include/__string +++ b/lib/libcxx/include/__string @@ -55,7 +55,9 @@ template <> struct char_traits<char8_t>; // c++20 #include <__config> #include <algorithm> // for search and min -#include <cstdio> // For EOF. +#include <cstdio> // for EOF +#include <cstring> // for memcpy +#include <cwchar> // for wmemcpy #include <memory> // for __murmur2_or_cityhash #include <__debug> @@ -92,7 +94,7 @@ _LIBCPP_BEGIN_NAMESPACE_STD _Func(_LIBCPP_FUNC_VIS void basic_string<_CharType>::__init(value_type const*, size_type, size_type)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::basic_string(basic_string const&)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>& basic_string<_CharType>::replace(size_type, size_type, value_type const*)) \ - _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::basic_string(basic_string const&, std::allocator<_CharType> const&)) \ + _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::basic_string(basic_string const&, allocator<_CharType> const&)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::size_type basic_string<_CharType>::find_last_not_of(value_type const*, size_type, size_type) const) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::~basic_string()) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::size_type basic_string<_CharType>::find_first_not_of(value_type const*, size_type, size_type) const) \ @@ -108,7 +110,7 @@ _LIBCPP_BEGIN_NAMESPACE_STD _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>& basic_string<_CharType>::append(value_type const*, size_type)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>& basic_string<_CharType>::assign(basic_string const&, size_type, size_type)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::size_type basic_string<_CharType>::copy(value_type*, size_type, size_type) const) \ - _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::basic_string(basic_string const&, size_type, size_type, std::allocator<_CharType> const&)) \ + _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::basic_string(basic_string const&, size_type, size_type, allocator<_CharType> const&)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::size_type basic_string<_CharType>::find(value_type, size_type) const) \ _Func(_LIBCPP_FUNC_VIS void basic_string<_CharType>::__init(size_type, value_type)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>& basic_string<_CharType>::insert(size_type, value_type const*)) \ @@ -158,7 +160,7 @@ _LIBCPP_BEGIN_NAMESPACE_STD _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>& basic_string<_CharType>::append(value_type const*, size_type)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>& basic_string<_CharType>::assign(basic_string const&, size_type, size_type)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::size_type basic_string<_CharType>::copy(value_type*, size_type, size_type) const) \ - _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::basic_string(basic_string const&, size_type, size_type, std::allocator<_CharType> const&)) \ + _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::basic_string(basic_string const&, size_type, size_type, allocator<_CharType> const&)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>::size_type basic_string<_CharType>::find(value_type, size_type) const) \ _Func(_LIBCPP_FUNC_VIS void basic_string<_CharType>::__init(size_type, value_type)) \ _Func(_LIBCPP_FUNC_VIS basic_string<_CharType>& basic_string<_CharType>::insert(size_type, value_type const*)) \ @@ -268,7 +270,7 @@ char_traits<_CharT>::find(const char_type* __s, size_t __n, const char_type& __a return __s; ++__s; } - return 0; + return nullptr; } template <class _CharT> @@ -318,7 +320,7 @@ char_traits<_CharT>::assign(char_type* __s, size_t __n, char_type __a) // constexpr versions of move/copy/assign. template <class _CharT> -static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED +static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 _CharT* __move_constexpr(_CharT* __s1, const _CharT* __s2, size_t __n) _NOEXCEPT { if (__n == 0) return __s1; @@ -331,7 +333,7 @@ _CharT* __move_constexpr(_CharT* __s1, const _CharT* __s2, size_t __n) _NOEXCEPT } template <class _CharT> -static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED +static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 _CharT* __copy_constexpr(_CharT* __s1, const _CharT* __s2, size_t __n) _NOEXCEPT { _VSTD::copy_n(__s2, __n, __s1); @@ -339,7 +341,7 @@ _CharT* __copy_constexpr(_CharT* __s1, const _CharT* __s2, size_t __n) _NOEXCEPT } template <class _CharT> -static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED +static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 _CharT* __assign_constexpr(_CharT* __s, size_t __n, _CharT __a) _NOEXCEPT { _VSTD::fill_n(__s, __n, __a); @@ -370,27 +372,27 @@ struct _LIBCPP_TEMPLATE_VIS char_traits<char> length(const char_type* __s) _NOEXCEPT {return __builtin_strlen(__s);} static _LIBCPP_CONSTEXPR_AFTER_CXX14 const char_type* find(const char_type* __s, size_t __n, const char_type& __a) _NOEXCEPT; - static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* move(char_type* __s1, const char_type* __s2, size_t __n) _NOEXCEPT { return __libcpp_is_constant_evaluated() - ? __move_constexpr(__s1, __s2, __n) - : __n == 0 ? __s1 : (char_type*)memmove(__s1, __s2, __n); + ? _VSTD::__move_constexpr(__s1, __s2, __n) + : __n == 0 ? __s1 : (char_type*)_VSTD::memmove(__s1, __s2, __n); } - static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* copy(char_type* __s1, const char_type* __s2, size_t __n) _NOEXCEPT { _LIBCPP_ASSERT(__s2 < __s1 || __s2 >= __s1+__n, "char_traits::copy overlapped range"); return __libcpp_is_constant_evaluated() - ? __copy_constexpr(__s1, __s2, __n) - : __n == 0 ? __s1 : (char_type*)memcpy(__s1, __s2, __n); + ? _VSTD::__copy_constexpr(__s1, __s2, __n) + : __n == 0 ? __s1 : (char_type*)_VSTD::memcpy(__s1, __s2, __n); } - static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* assign(char_type* __s, size_t __n, char_type __a) _NOEXCEPT { return __libcpp_is_constant_evaluated() - ? __assign_constexpr(__s, __n, __a) - : __n == 0 ? __s : (char_type*)memset(__s, to_int_type(__a), __n); + ? _VSTD::__assign_constexpr(__s, __n, __a) + : __n == 0 ? __s : (char_type*)_VSTD::memset(__s, to_int_type(__a), __n); } static inline _LIBCPP_CONSTEXPR int_type not_eof(int_type __c) _NOEXCEPT @@ -414,7 +416,7 @@ char_traits<char>::compare(const char_type* __s1, const char_type* __s2, size_t #if __has_feature(cxx_constexpr_string_builtins) return __builtin_memcmp(__s1, __s2, __n); #elif _LIBCPP_STD_VER <= 14 - return memcmp(__s1, __s2, __n); + return _VSTD::memcmp(__s1, __s2, __n); #else for (; __n; --__n, ++__s1, ++__s2) { @@ -436,7 +438,7 @@ char_traits<char>::find(const char_type* __s, size_t __n, const char_type& __a) #if __has_feature(cxx_constexpr_string_builtins) return __builtin_char_memchr(__s, to_int_type(__a), __n); #elif _LIBCPP_STD_VER <= 14 - return (const char_type*) memchr(__s, to_int_type(__a), __n); + return (const char_type*) _VSTD::memchr(__s, to_int_type(__a), __n); #else for (; __n; --__n) { @@ -473,27 +475,27 @@ struct _LIBCPP_TEMPLATE_VIS char_traits<wchar_t> size_t length(const char_type* __s) _NOEXCEPT; static _LIBCPP_CONSTEXPR_AFTER_CXX14 const char_type* find(const char_type* __s, size_t __n, const char_type& __a) _NOEXCEPT; - static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* move(char_type* __s1, const char_type* __s2, size_t __n) _NOEXCEPT { return __libcpp_is_constant_evaluated() - ? __move_constexpr(__s1, __s2, __n) - : __n == 0 ? __s1 : wmemmove(__s1, __s2, __n); + ? _VSTD::__move_constexpr(__s1, __s2, __n) + : __n == 0 ? __s1 : _VSTD::wmemmove(__s1, __s2, __n); } - static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* copy(char_type* __s1, const char_type* __s2, size_t __n) _NOEXCEPT { _LIBCPP_ASSERT(__s2 < __s1 || __s2 >= __s1+__n, "char_traits::copy overlapped range"); return __libcpp_is_constant_evaluated() - ? __copy_constexpr(__s1, __s2, __n) - : __n == 0 ? __s1 : wmemcpy(__s1, __s2, __n); + ? _VSTD::__copy_constexpr(__s1, __s2, __n) + : __n == 0 ? __s1 : _VSTD::wmemcpy(__s1, __s2, __n); } - static inline _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static inline _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* assign(char_type* __s, size_t __n, char_type __a) _NOEXCEPT { return __libcpp_is_constant_evaluated() - ? __assign_constexpr(__s, __n, __a) - : __n == 0 ? __s : wmemset(__s, __a, __n); + ? _VSTD::__assign_constexpr(__s, __n, __a) + : __n == 0 ? __s : _VSTD::wmemset(__s, __a, __n); } static inline _LIBCPP_CONSTEXPR int_type not_eof(int_type __c) _NOEXCEPT {return eq_int_type(__c, eof()) ? ~eof() : __c;} @@ -516,7 +518,7 @@ char_traits<wchar_t>::compare(const char_type* __s1, const char_type* __s2, size #if __has_feature(cxx_constexpr_string_builtins) return __builtin_wmemcmp(__s1, __s2, __n); #elif _LIBCPP_STD_VER <= 14 - return wmemcmp(__s1, __s2, __n); + return _VSTD::wmemcmp(__s1, __s2, __n); #else for (; __n; --__n, ++__s1, ++__s2) { @@ -548,7 +550,7 @@ char_traits<wchar_t>::length(const char_type* __s) _NOEXCEPT #if __has_feature(cxx_constexpr_string_builtins) return __builtin_wcslen(__s); #elif _LIBCPP_STD_VER <= 14 - return wcslen(__s); + return _VSTD::wcslen(__s); #else size_t __len = 0; for (; !eq(*__s, char_type(0)); ++__s) @@ -566,7 +568,7 @@ char_traits<wchar_t>::find(const char_type* __s, size_t __n, const char_type& __ #if __has_feature(cxx_constexpr_string_builtins) return __builtin_wmemchr(__s, __a, __n); #elif _LIBCPP_STD_VER <= 14 - return wmemchr(__s, __a, __n); + return _VSTD::wmemchr(__s, __a, __n); #else for (; __n; --__n) { @@ -606,29 +608,29 @@ struct _LIBCPP_TEMPLATE_VIS char_traits<char8_t> _LIBCPP_INLINE_VISIBILITY static constexpr const char_type* find(const char_type* __s, size_t __n, const char_type& __a) _NOEXCEPT; - static _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* move(char_type* __s1, const char_type* __s2, size_t __n) _NOEXCEPT { return __libcpp_is_constant_evaluated() - ? __move_constexpr(__s1, __s2, __n) - : __n == 0 ? __s1 : (char_type*)memmove(__s1, __s2, __n); + ? _VSTD::__move_constexpr(__s1, __s2, __n) + : __n == 0 ? __s1 : (char_type*)_VSTD::memmove(__s1, __s2, __n); } - static _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* copy(char_type* __s1, const char_type* __s2, size_t __n) _NOEXCEPT { _LIBCPP_ASSERT(__s2 < __s1 || __s2 >= __s1+__n, "char_traits::copy overlapped range"); return __libcpp_is_constant_evaluated() - ? __copy_constexpr(__s1, __s2, __n) - : __n == 0 ? __s1 : (char_type*)memcpy(__s1, __s2, __n); + ? _VSTD::__copy_constexpr(__s1, __s2, __n) + : __n == 0 ? __s1 : (char_type*)_VSTD::memcpy(__s1, __s2, __n); } - static _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED + static _LIBCPP_CONSTEXPR_AFTER_CXX17 char_type* assign(char_type* __s, size_t __n, char_type __a) _NOEXCEPT { return __libcpp_is_constant_evaluated() - ? __assign_constexpr(__s, __n, __a) - : __n == 0 ? __s : (char_type*)memset(__s, to_int_type(__a), __n); + ? _VSTD::__assign_constexpr(__s, __n, __a) + : __n == 0 ? __s : (char_type*)_VSTD::memset(__s, to_int_type(__a), __n); } static inline constexpr int_type not_eof(int_type __c) noexcept @@ -683,7 +685,7 @@ char_traits<char8_t>::find(const char_type* __s, size_t __n, const char_type& __ return __s; ++__s; } - return 0; + return nullptr; } #endif // #_LIBCPP_NO_HAS_CHAR8_T @@ -765,7 +767,7 @@ char_traits<char16_t>::find(const char_type* __s, size_t __n, const char_type& _ return __s; ++__s; } - return 0; + return nullptr; } inline _LIBCPP_CONSTEXPR_AFTER_CXX17 @@ -885,7 +887,7 @@ char_traits<char32_t>::find(const char_type* __s, size_t __n, const char_type& _ return __s; ++__s; } - return 0; + return nullptr; } inline _LIBCPP_CONSTEXPR_AFTER_CXX17 @@ -943,7 +945,7 @@ __str_find(const _CharT *__p, _SizeT __sz, if (__pos >= __sz) return __npos; const _CharT* __r = _Traits::find(__p + __pos, __sz - __pos, __c); - if (__r == 0) + if (__r == nullptr) return __npos; return static_cast<_SizeT>(__r - __p); } @@ -972,7 +974,7 @@ __search_substring(const _CharT *__first1, const _CharT *__last1, // Find __f2 the first byte matching in __first1. __first1 = _Traits::find(__first1, __len1 - __len2 + 1, __f2); - if (__first1 == 0) + if (__first1 == nullptr) return __last1; // It is faster to compare from the first byte of __first1 even if we @@ -1095,7 +1097,7 @@ __str_find_first_not_of(const _CharT *__p, _SizeT __sz, { const _CharT* __pe = __p + __sz; for (const _CharT* __ps = __p + __pos; __ps != __pe; ++__ps) - if (_Traits::find(__s, __n, *__ps) == 0) + if (_Traits::find(__s, __n, *__ps) == nullptr) return static_cast<_SizeT>(__ps - __p); } return __npos; @@ -1129,7 +1131,7 @@ __str_find_last_not_of(const _CharT *__p, _SizeT __sz, else __pos = __sz; for (const _CharT* __ps = __p + __pos; __ps != __p;) - if (_Traits::find(__s, __n, *--__ps) == 0) + if (_Traits::find(__s, __n, *--__ps) == nullptr) return static_cast<_SizeT>(__ps - __p); return __npos; } diff --git a/lib/libcxx/include/support/android/locale_bionic.h b/lib/libcxx/include/__support/android/locale_bionic.h index f05a6a0522..8c6d4bd0dc 100644 --- a/lib/libcxx/include/support/android/locale_bionic.h +++ b/lib/libcxx/include/__support/android/locale_bionic.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===------------------- support/android/locale_bionic.h ------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. @@ -28,13 +28,13 @@ extern "C" { #include <android/api-level.h> #include <android/ndk-version.h> #if __ANDROID_API__ < 21 -#include <support/xlocale/__posix_l_fallback.h> +#include <__support/xlocale/__posix_l_fallback.h> #endif // In NDK versions later than 16, locale-aware functions are provided by // legacy_stdlib_inlines.h #if __NDK_MAJOR__ <= 16 #if __ANDROID_API__ < 21 -#include <support/xlocale/__strtonum_fallback.h> +#include <__support/xlocale/__strtonum_fallback.h> #elif __ANDROID_API__ < 26 #if defined(__cplusplus) diff --git a/lib/libcxx/include/support/fuchsia/xlocale.h b/lib/libcxx/include/__support/fuchsia/xlocale.h index b86ce9efbd..e8def81480 100644 --- a/lib/libcxx/include/support/fuchsia/xlocale.h +++ b/lib/libcxx/include/__support/fuchsia/xlocale.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===------------------- support/fuchsia/xlocale.h ------------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. @@ -14,8 +14,8 @@ #include <cstdlib> #include <cwchar> -#include <support/xlocale/__posix_l_fallback.h> -#include <support/xlocale/__strtonum_fallback.h> +#include <__support/xlocale/__posix_l_fallback.h> +#include <__support/xlocale/__strtonum_fallback.h> #endif // defined(__Fuchsia__) diff --git a/lib/libcxx/include/support/ibm/limits.h b/lib/libcxx/include/__support/ibm/limits.h index d1c59f066a..45f1f1e368 100644 --- a/lib/libcxx/include/support/ibm/limits.h +++ b/lib/libcxx/include/__support/ibm/limits.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===--------------------- support/ibm/limits.h ---------------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/support/ibm/locale_mgmt_aix.h b/lib/libcxx/include/__support/ibm/locale_mgmt_aix.h index e452dc3252..4f658c3eee 100644 --- a/lib/libcxx/include/support/ibm/locale_mgmt_aix.h +++ b/lib/libcxx/include/__support/ibm/locale_mgmt_aix.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===------------------- support/ibm/locale_mgmt_aix.h --------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/__support/ibm/nanosleep.h b/lib/libcxx/include/__support/ibm/nanosleep.h new file mode 100644 index 0000000000..c82f4eb0cd --- /dev/null +++ b/lib/libcxx/include/__support/ibm/nanosleep.h @@ -0,0 +1,38 @@ +// -*- C++ -*- +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP_SUPPORT_IBM_NANOSLEEP_H +#define _LIBCPP_SUPPORT_IBM_NANOSLEEP_H + +#include <unistd.h> + +inline int nanosleep(const struct timespec* req, struct timespec* rem) +{ + // The nanosleep() function is not available on z/OS. Therefore, we will call + // sleep() to sleep for whole seconds and usleep() to sleep for any remaining + // fraction of a second. Any remaining nanoseconds will round up to the next + // microsecond. + + useconds_t __micro_sec = (rem->tv_nsec + 999) / 1000; + if (__micro_sec > 999999) + { + ++rem->tv_sec; + __micro_sec -= 1000000; + } + while (rem->tv_sec) + rem->tv_sec = sleep(rem->tv_sec); + if (__micro_sec) { + rem->tv_nsec = __micro_sec * 1000; + return usleep(__micro_sec); + } + rem->tv_nsec = 0; + return 0; +} + +#endif // _LIBCPP_SUPPORT_IBM_NANOSLEEP_H diff --git a/lib/libcxx/include/support/ibm/support.h b/lib/libcxx/include/__support/ibm/support.h index 0569cbe746..a7751b0176 100644 --- a/lib/libcxx/include/support/ibm/support.h +++ b/lib/libcxx/include/__support/ibm/support.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===----------------------- support/ibm/support.h ----------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/support/ibm/xlocale.h b/lib/libcxx/include/__support/ibm/xlocale.h index 431bf62687..ad07a255fc 100644 --- a/lib/libcxx/include/support/ibm/xlocale.h +++ b/lib/libcxx/include/__support/ibm/xlocale.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===--------------------- support/ibm/xlocale.h -------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. @@ -9,15 +9,16 @@ #ifndef _LIBCPP_SUPPORT_IBM_XLOCALE_H #define _LIBCPP_SUPPORT_IBM_XLOCALE_H -#include <support/ibm/locale_mgmt_aix.h> -#if defined(_AIX) +#include <__support/ibm/locale_mgmt_aix.h> + #include "cstdlib" #ifdef __cplusplus extern "C" { #endif +#if defined(_AIX) #if !defined(_AIX71) // AIX 7.1 and higher has these definitions. Definitions and stubs // are provied here as a temporary workaround on AIX 6.1. @@ -207,14 +208,20 @@ size_t wcsxfrm_l(wchar_t *__ws1, const wchar_t *__ws2, size_t __n, } #endif // !defined(_AIX71) -// strftime_l() is defined by POSIX. However, AIX 7.1 does not have it -// implemented yet. +// strftime_l() is defined by POSIX. However, AIX 7.1 and z/OS do not have it +// implemented yet. z/OS retrieves it from the POSIX fallbacks. static inline size_t strftime_l(char *__s, size_t __size, const char *__fmt, const struct tm *__tm, locale_t locale) { return __xstrftime(locale, __s, __size, __fmt, __tm); } +#elif defined(__MVS__) +#include <wctype.h> +// POSIX routines +#include <__support/xlocale/__posix_l_fallback.h> +#endif // defined(__MVS__) + // The following are not POSIX routines. These are quick-and-dirty hacks // to make things pretend to work static inline @@ -266,5 +273,4 @@ int vasprintf(char **strp, const char *fmt, va_list ap) #ifdef __cplusplus } #endif -#endif // defined(_AIX) #endif // _LIBCPP_SUPPORT_IBM_XLOCALE_H diff --git a/lib/libcxx/include/support/musl/xlocale.h b/lib/libcxx/include/__support/musl/xlocale.h index 722d13fa1d..2508a8e8e0 100644 --- a/lib/libcxx/include/support/musl/xlocale.h +++ b/lib/libcxx/include/__support/musl/xlocale.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===------------------- support/musl/xlocale.h ------------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/support/newlib/xlocale.h b/lib/libcxx/include/__support/newlib/xlocale.h index 25fa798b6d..b75f9263a4 100644 --- a/lib/libcxx/include/support/newlib/xlocale.h +++ b/lib/libcxx/include/__support/newlib/xlocale.h @@ -17,9 +17,9 @@ #include <ctype.h> #if !defined(__NEWLIB__) || __NEWLIB__ < 2 || \ __NEWLIB__ == 2 && __NEWLIB_MINOR__ < 5 -#include <support/xlocale/__nop_locale_mgmt.h> -#include <support/xlocale/__posix_l_fallback.h> -#include <support/xlocale/__strtonum_fallback.h> +#include <__support/xlocale/__nop_locale_mgmt.h> +#include <__support/xlocale/__posix_l_fallback.h> +#include <__support/xlocale/__strtonum_fallback.h> #endif #endif // _NEWLIB_VERSION diff --git a/lib/libcxx/include/__support/nuttx/xlocale.h b/lib/libcxx/include/__support/nuttx/xlocale.h new file mode 100644 index 0000000000..be738e3b64 --- /dev/null +++ b/lib/libcxx/include/__support/nuttx/xlocale.h @@ -0,0 +1,18 @@ +// -*- C++ -*- +//===-----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP_SUPPORT_NUTTX_XLOCALE_H +#define _LIBCPP_SUPPORT_NUTTX_XLOCALE_H + +#if defined(__NuttX__) +#include <__support/xlocale/__posix_l_fallback.h> +#include <__support/xlocale/__strtonum_fallback.h> +#endif // __NuttX__ + +#endif diff --git a/lib/libcxx/include/__support/openbsd/xlocale.h b/lib/libcxx/include/__support/openbsd/xlocale.h new file mode 100644 index 0000000000..1136fa327f --- /dev/null +++ b/lib/libcxx/include/__support/openbsd/xlocale.h @@ -0,0 +1,19 @@ +// -*- C++ -*- +//===-----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#ifndef _LIBCPP_SUPPORT_OPENBSD_XLOCALE_H +#define _LIBCPP_SUPPORT_OPENBSD_XLOCALE_H + +#include <cstdlib> +#include <clocale> +#include <cwctype> +#include <ctype.h> +#include <__support/xlocale/__strtonum_fallback.h> + +#endif diff --git a/lib/libcxx/include/support/solaris/floatingpoint.h b/lib/libcxx/include/__support/solaris/floatingpoint.h index 5f1628fbe4..5f1628fbe4 100644 --- a/lib/libcxx/include/support/solaris/floatingpoint.h +++ b/lib/libcxx/include/__support/solaris/floatingpoint.h diff --git a/lib/libcxx/include/support/solaris/wchar.h b/lib/libcxx/include/__support/solaris/wchar.h index f01fd743a2..f01fd743a2 100644 --- a/lib/libcxx/include/support/solaris/wchar.h +++ b/lib/libcxx/include/__support/solaris/wchar.h diff --git a/lib/libcxx/include/support/solaris/xlocale.h b/lib/libcxx/include/__support/solaris/xlocale.h index 05131f0272..05131f0272 100644 --- a/lib/libcxx/include/support/solaris/xlocale.h +++ b/lib/libcxx/include/__support/solaris/xlocale.h diff --git a/lib/libcxx/include/support/win32/limits_msvc_win32.h b/lib/libcxx/include/__support/win32/limits_msvc_win32.h index 7bb835559a..758d24647b 100644 --- a/lib/libcxx/include/support/win32/limits_msvc_win32.h +++ b/lib/libcxx/include/__support/win32/limits_msvc_win32.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===------------------ support/win32/limits_msvc_win32.h -----------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/support/win32/locale_win32.h b/lib/libcxx/include/__support/win32/locale_win32.h index 8d7779e0cc..d32a7a8ad3 100644 --- a/lib/libcxx/include/support/win32/locale_win32.h +++ b/lib/libcxx/include/__support/win32/locale_win32.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===--------------------- support/win32/locale_win32.h -------------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. @@ -201,8 +201,8 @@ decltype(MB_CUR_MAX) MB_CUR_MAX_L( locale_t __l ); #define strtof_l _strtof_l #define strtold_l _strtold_l #else -float strtof_l(const char*, char**, locale_t); -long double strtold_l(const char*, char**, locale_t); +_LIBCPP_FUNC_VIS float strtof_l(const char*, char**, locale_t); +_LIBCPP_FUNC_VIS long double strtold_l(const char*, char**, locale_t); #endif inline _LIBCPP_INLINE_VISIBILITY int diff --git a/lib/libcxx/include/support/xlocale/__nop_locale_mgmt.h b/lib/libcxx/include/__support/xlocale/__nop_locale_mgmt.h index f33d3894c3..57b18842ff 100644 --- a/lib/libcxx/include/support/xlocale/__nop_locale_mgmt.h +++ b/lib/libcxx/include/__support/xlocale/__nop_locale_mgmt.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===------------ support/xlocale/__nop_locale_mgmt.h -----------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/support/xlocale/__posix_l_fallback.h b/lib/libcxx/include/__support/xlocale/__posix_l_fallback.h index f3df6c46fb..00d69d19e8 100644 --- a/lib/libcxx/include/support/xlocale/__posix_l_fallback.h +++ b/lib/libcxx/include/__support/xlocale/__posix_l_fallback.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===--------------- support/xlocale/__posix_l_fallback.h -----------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/support/xlocale/__strtonum_fallback.h b/lib/libcxx/include/__support/xlocale/__strtonum_fallback.h index df38598056..1172a5d572 100644 --- a/lib/libcxx/include/support/xlocale/__strtonum_fallback.h +++ b/lib/libcxx/include/__support/xlocale/__strtonum_fallback.h @@ -1,5 +1,5 @@ // -*- C++ -*- -//===-------------- support/xlocale/__strtonum_fallback.h -----------------===// +//===-----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/include/__threading_support b/lib/libcxx/include/__threading_support index 072c4c7bcc..de572f3ff8 100644 --- a/lib/libcxx/include/__threading_support +++ b/lib/libcxx/include/__threading_support @@ -11,10 +11,15 @@ #define _LIBCPP_THREADING_SUPPORT #include <__config> +#include <__availability> #include <chrono> #include <iosfwd> #include <errno.h> +#ifdef __MVS__ +# include <__support/ibm/nanosleep.h> +#endif + #ifndef _LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER #pragma GCC system_header #endif @@ -26,7 +31,7 @@ #if defined(_LIBCPP_HAS_THREAD_API_PTHREAD) # include <pthread.h> # include <sched.h> -# ifdef __APPLE__ +# if defined(__APPLE__) || defined(__MVS__) # define _LIBCPP_NO_NATIVE_SEMAPHORES # endif # ifndef _LIBCPP_NO_NATIVE_SEMAPHORES @@ -82,11 +87,14 @@ typedef pthread_once_t __libcpp_exec_once_flag; #define _LIBCPP_EXEC_ONCE_INITIALIZER PTHREAD_ONCE_INIT // Thread id -typedef pthread_t __libcpp_thread_id; +#if defined(__MVS__) + typedef unsigned long long __libcpp_thread_id; +#else + typedef pthread_t __libcpp_thread_id; +#endif // Thread -#define _LIBCPP_NULL_THREAD 0U - +#define _LIBCPP_NULL_THREAD ((__libcpp_thread_t())) typedef pthread_t __libcpp_thread_t; // Thread Local Storage @@ -278,24 +286,21 @@ int __libcpp_tls_set(__libcpp_tls_key __key, void *__p); #endif // !defined(_LIBCPP_HAS_THREAD_API_EXTERNAL) struct __libcpp_timed_backoff_policy { - _LIBCPP_THREAD_ABI_VISIBILITY - bool operator()(chrono::nanoseconds __elapsed) const; + _LIBCPP_INLINE_VISIBILITY + bool operator()(chrono::nanoseconds __elapsed) const + { + if(__elapsed > chrono::milliseconds(128)) + __libcpp_thread_sleep_for(chrono::milliseconds(8)); + else if(__elapsed > chrono::microseconds(64)) + __libcpp_thread_sleep_for(__elapsed / 2); + else if(__elapsed > chrono::microseconds(4)) + __libcpp_thread_yield(); + else + {} // poll + return false; + } }; -inline _LIBCPP_INLINE_VISIBILITY -bool __libcpp_timed_backoff_policy::operator()(chrono::nanoseconds __elapsed) const -{ - if(__elapsed > chrono::milliseconds(128)) - __libcpp_thread_sleep_for(chrono::milliseconds(8)); - else if(__elapsed > chrono::microseconds(64)) - __libcpp_thread_sleep_for(__elapsed / 2); - else if(__elapsed > chrono::microseconds(4)) - __libcpp_thread_yield(); - else - ; // poll - return false; -} - static _LIBCPP_CONSTEXPR const int __libcpp_polling_count = 64; template<class _Fn, class _BFn> @@ -484,7 +489,7 @@ int __libcpp_execute_once(__libcpp_exec_once_flag *flag, // Returns non-zero if the thread ids are equal, otherwise 0 bool __libcpp_thread_id_equal(__libcpp_thread_id t1, __libcpp_thread_id t2) { - return pthread_equal(t1, t2) != 0; + return t1 == t2; } // Returns non-zero if t1 < t2, otherwise 0 @@ -495,28 +500,33 @@ bool __libcpp_thread_id_less(__libcpp_thread_id t1, __libcpp_thread_id t2) // Thread bool __libcpp_thread_isnull(const __libcpp_thread_t *__t) { - return *__t == 0; + return *__t == __libcpp_thread_t(); } int __libcpp_thread_create(__libcpp_thread_t *__t, void *(*__func)(void *), void *__arg) { - return pthread_create(__t, 0, __func, __arg); + return pthread_create(__t, nullptr, __func, __arg); } __libcpp_thread_id __libcpp_thread_get_current_id() { - return pthread_self(); + const __libcpp_thread_t thread = pthread_self(); + return __libcpp_thread_get_id(&thread); } __libcpp_thread_id __libcpp_thread_get_id(const __libcpp_thread_t *__t) { +#if defined(__MVS__) + return __t->__; +#else return *__t; +#endif } int __libcpp_thread_join(__libcpp_thread_t *__t) { - return pthread_join(*__t, 0); + return pthread_join(*__t, nullptr); } int __libcpp_thread_detach(__libcpp_thread_t *__t) @@ -651,7 +661,7 @@ bool __libcpp_thread_id_less(__libcpp_thread_id t1, __libcpp_thread_id t2) // Thread bool __libcpp_thread_isnull(const __libcpp_thread_t *__t) { - return *__t == 0; + return __libcpp_thread_get_id(__t) == 0; } int __libcpp_thread_create(__libcpp_thread_t *__t, void *(*__func)(void *), diff --git a/lib/libcxx/include/__tree b/lib/libcxx/include/__tree index cb7a1022e6..0f6e4ec379 100644 --- a/lib/libcxx/include/__tree +++ b/lib/libcxx/include/__tree @@ -108,10 +108,10 @@ __tree_sub_invariant(_NodePtr __x) if (__x->__right_ && !__x->__right_->__is_black_) return 0; } - unsigned __h = __tree_sub_invariant(__x->__left_); + unsigned __h = _VSTD::__tree_sub_invariant(__x->__left_); if (__h == 0) return 0; // invalid left subtree - if (__h != __tree_sub_invariant(__x->__right_)) + if (__h != _VSTD::__tree_sub_invariant(__x->__right_)) return 0; // invalid or different height right subtree return __h + __x->__is_black_; // return black height of this node } @@ -128,13 +128,13 @@ __tree_invariant(_NodePtr __root) // check __x->__parent_ consistency if (__root->__parent_ == nullptr) return false; - if (!__tree_is_left_child(__root)) + if (!_VSTD::__tree_is_left_child(__root)) return false; // root must be black if (!__root->__is_black_) return false; // do normal node checks - return __tree_sub_invariant(__root) != 0; + return _VSTD::__tree_sub_invariant(__root) != 0; } // Returns: pointer to the left-most node under __x. @@ -168,8 +168,8 @@ _NodePtr __tree_next(_NodePtr __x) _NOEXCEPT { if (__x->__right_ != nullptr) - return __tree_min(__x->__right_); - while (!__tree_is_left_child(__x)) + return _VSTD::__tree_min(__x->__right_); + while (!_VSTD::__tree_is_left_child(__x)) __x = __x->__parent_unsafe(); return __x->__parent_unsafe(); } @@ -180,8 +180,8 @@ _EndNodePtr __tree_next_iter(_NodePtr __x) _NOEXCEPT { if (__x->__right_ != nullptr) - return static_cast<_EndNodePtr>(__tree_min(__x->__right_)); - while (!__tree_is_left_child(__x)) + return static_cast<_EndNodePtr>(_VSTD::__tree_min(__x->__right_)); + while (!_VSTD::__tree_is_left_child(__x)) __x = __x->__parent_unsafe(); return static_cast<_EndNodePtr>(__x->__parent_); } @@ -195,9 +195,9 @@ _NodePtr __tree_prev_iter(_EndNodePtr __x) _NOEXCEPT { if (__x->__left_ != nullptr) - return __tree_max(__x->__left_); + return _VSTD::__tree_max(__x->__left_); _NodePtr __xx = static_cast<_NodePtr>(__x); - while (__tree_is_left_child(__xx)) + while (_VSTD::__tree_is_left_child(__xx)) __xx = __xx->__parent_unsafe(); return __xx->__parent_unsafe(); } @@ -237,7 +237,7 @@ __tree_left_rotate(_NodePtr __x) _NOEXCEPT if (__x->__right_ != nullptr) __x->__right_->__set_parent(__x); __y->__parent_ = __x->__parent_; - if (__tree_is_left_child(__x)) + if (_VSTD::__tree_is_left_child(__x)) __x->__parent_->__left_ = __y; else __x->__parent_unsafe()->__right_ = __y; @@ -257,7 +257,7 @@ __tree_right_rotate(_NodePtr __x) _NOEXCEPT if (__x->__left_ != nullptr) __x->__left_->__set_parent(__x); __y->__parent_ = __x->__parent_; - if (__tree_is_left_child(__x)) + if (_VSTD::__tree_is_left_child(__x)) __x->__parent_->__left_ = __y; else __x->__parent_unsafe()->__right_ = __y; @@ -281,7 +281,7 @@ __tree_balance_after_insert(_NodePtr __root, _NodePtr __x) _NOEXCEPT while (__x != __root && !__x->__parent_unsafe()->__is_black_) { // __x->__parent_ != __root because __x->__parent_->__is_black == false - if (__tree_is_left_child(__x->__parent_unsafe())) + if (_VSTD::__tree_is_left_child(__x->__parent_unsafe())) { _NodePtr __y = __x->__parent_unsafe()->__parent_unsafe()->__right_; if (__y != nullptr && !__y->__is_black_) @@ -294,16 +294,16 @@ __tree_balance_after_insert(_NodePtr __root, _NodePtr __x) _NOEXCEPT } else { - if (!__tree_is_left_child(__x)) + if (!_VSTD::__tree_is_left_child(__x)) { __x = __x->__parent_unsafe(); - __tree_left_rotate(__x); + _VSTD::__tree_left_rotate(__x); } __x = __x->__parent_unsafe(); __x->__is_black_ = true; __x = __x->__parent_unsafe(); __x->__is_black_ = false; - __tree_right_rotate(__x); + _VSTD::__tree_right_rotate(__x); break; } } @@ -320,16 +320,16 @@ __tree_balance_after_insert(_NodePtr __root, _NodePtr __x) _NOEXCEPT } else { - if (__tree_is_left_child(__x)) + if (_VSTD::__tree_is_left_child(__x)) { __x = __x->__parent_unsafe(); - __tree_right_rotate(__x); + _VSTD::__tree_right_rotate(__x); } __x = __x->__parent_unsafe(); __x->__is_black_ = true; __x = __x->__parent_unsafe(); __x->__is_black_ = false; - __tree_left_rotate(__x); + _VSTD::__tree_left_rotate(__x); break; } } @@ -352,7 +352,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT // __y will have at most one child. // __y will be the initial hole in the tree (make the hole at a leaf) _NodePtr __y = (__z->__left_ == nullptr || __z->__right_ == nullptr) ? - __z : __tree_next(__z); + __z : _VSTD::__tree_next(__z); // __x is __y's possibly null single child _NodePtr __x = __y->__left_ != nullptr ? __y->__left_ : __y->__right_; // __w is __x's possibly null uncle (will become __x's sibling) @@ -360,7 +360,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT // link __x to __y's parent, and find __w if (__x != nullptr) __x->__parent_ = __y->__parent_; - if (__tree_is_left_child(__y)) + if (_VSTD::__tree_is_left_child(__y)) { __y->__parent_->__left_ = __x; if (__y != __root) @@ -381,7 +381,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT { // __z->__left_ != nulptr but __z->__right_ might == __x == nullptr __y->__parent_ = __z->__parent_; - if (__tree_is_left_child(__z)) + if (_VSTD::__tree_is_left_child(__z)) __y->__parent_->__left_ = __y; else __y->__parent_unsafe()->__right_ = __y; @@ -421,13 +421,13 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT // with a non-null black child). while (true) { - if (!__tree_is_left_child(__w)) // if x is left child + if (!_VSTD::__tree_is_left_child(__w)) // if x is left child { if (!__w->__is_black_) { __w->__is_black_ = true; __w->__parent_unsafe()->__is_black_ = false; - __tree_left_rotate(__w->__parent_unsafe()); + _VSTD::__tree_left_rotate(__w->__parent_unsafe()); // __x is still valid // reset __root only if necessary if (__root == __w->__left_) @@ -448,7 +448,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT break; } // reset sibling, and it still can't be null - __w = __tree_is_left_child(__x) ? + __w = _VSTD::__tree_is_left_child(__x) ? __x->__parent_unsafe()->__right_ : __x->__parent_->__left_; // continue; @@ -460,7 +460,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT // __w left child is non-null and red __w->__left_->__is_black_ = true; __w->__is_black_ = false; - __tree_right_rotate(__w); + _VSTD::__tree_right_rotate(__w); // __w is known not to be root, so root hasn't changed // reset sibling, and it still can't be null __w = __w->__parent_unsafe(); @@ -469,7 +469,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT __w->__is_black_ = __w->__parent_unsafe()->__is_black_; __w->__parent_unsafe()->__is_black_ = true; __w->__right_->__is_black_ = true; - __tree_left_rotate(__w->__parent_unsafe()); + _VSTD::__tree_left_rotate(__w->__parent_unsafe()); break; } } @@ -479,7 +479,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT { __w->__is_black_ = true; __w->__parent_unsafe()->__is_black_ = false; - __tree_right_rotate(__w->__parent_unsafe()); + _VSTD::__tree_right_rotate(__w->__parent_unsafe()); // __x is still valid // reset __root only if necessary if (__root == __w->__right_) @@ -500,7 +500,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT break; } // reset sibling, and it still can't be null - __w = __tree_is_left_child(__x) ? + __w = _VSTD::__tree_is_left_child(__x) ? __x->__parent_unsafe()->__right_ : __x->__parent_->__left_; // continue; @@ -512,7 +512,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT // __w right child is non-null and red __w->__right_->__is_black_ = true; __w->__is_black_ = false; - __tree_left_rotate(__w); + _VSTD::__tree_left_rotate(__w); // __w is known not to be root, so root hasn't changed // reset sibling, and it still can't be null __w = __w->__parent_unsafe(); @@ -521,7 +521,7 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT __w->__is_black_ = __w->__parent_unsafe()->__is_black_; __w->__parent_unsafe()->__is_black_ = true; __w->__left_->__is_black_ = true; - __tree_right_rotate(__w->__parent_unsafe()); + _VSTD::__tree_right_rotate(__w->__parent_unsafe()); break; } } @@ -533,19 +533,17 @@ __tree_remove(_NodePtr __root, _NodePtr __z) _NOEXCEPT // node traits -#ifndef _LIBCPP_CXX03_LANG template <class _Tp> struct __is_tree_value_type_imp : false_type {}; template <class _Key, class _Value> -struct __is_tree_value_type_imp<__value_type<_Key, _Value>> : true_type {}; +struct __is_tree_value_type_imp<__value_type<_Key, _Value> > : true_type {}; template <class ..._Args> struct __is_tree_value_type : false_type {}; template <class _One> struct __is_tree_value_type<_One> : __is_tree_value_type_imp<typename __uncvref<_One>::type> {}; -#endif template <class _Tp> struct __tree_key_value_types { @@ -566,12 +564,10 @@ struct __tree_key_value_types { static __container_value_type* __get_ptr(__node_value_type& __n) { return _VSTD::addressof(__n); } -#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY static __container_value_type&& __move(__node_value_type& __v) { return _VSTD::move(__v); } -#endif }; template <class _Key, class _Tp> @@ -616,12 +612,10 @@ struct __tree_key_value_types<__value_type<_Key, _Tp> > { return _VSTD::addressof(__n.__get_value()); } -#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY static pair<key_type&&, mapped_type&&> __move(__node_value_type& __v) { return __v.__move(); } -#endif }; template <class _VoidPtr> @@ -845,7 +839,7 @@ public: _LIBCPP_INLINE_VISIBILITY __tree_iterator& operator++() { __ptr_ = static_cast<__iter_pointer>( - __tree_next_iter<__end_node_pointer>(static_cast<__node_base_pointer>(__ptr_))); + _VSTD::__tree_next_iter<__end_node_pointer>(static_cast<__node_base_pointer>(__ptr_))); return *this; } _LIBCPP_INLINE_VISIBILITY @@ -854,7 +848,7 @@ public: _LIBCPP_INLINE_VISIBILITY __tree_iterator& operator--() { - __ptr_ = static_cast<__iter_pointer>(__tree_prev_iter<__node_base_pointer>( + __ptr_ = static_cast<__iter_pointer>(_VSTD::__tree_prev_iter<__node_base_pointer>( static_cast<__end_node_pointer>(__ptr_))); return *this; } @@ -926,7 +920,7 @@ public: _LIBCPP_INLINE_VISIBILITY __tree_const_iterator& operator++() { __ptr_ = static_cast<__iter_pointer>( - __tree_next_iter<__end_node_pointer>(static_cast<__node_base_pointer>(__ptr_))); + _VSTD::__tree_next_iter<__end_node_pointer>(static_cast<__node_base_pointer>(__ptr_))); return *this; } @@ -936,7 +930,7 @@ public: _LIBCPP_INLINE_VISIBILITY __tree_const_iterator& operator--() { - __ptr_ = static_cast<__iter_pointer>(__tree_prev_iter<__node_base_pointer>( + __ptr_ = static_cast<__iter_pointer>(_VSTD::__tree_prev_iter<__node_base_pointer>( static_cast<__end_node_pointer>(__ptr_))); return *this; } @@ -973,7 +967,7 @@ private: template<class _Tp, class _Compare> #ifndef _LIBCPP_CXX03_LANG - _LIBCPP_DIAGNOSE_WARNING(!std::__invokable<_Compare const&, _Tp const&, _Tp const&>::value, + _LIBCPP_DIAGNOSE_WARNING(!__invokable<_Compare const&, _Tp const&, _Tp const&>::value, "the specified comparator type does not provide a viable const call operator") #endif int __diagnose_non_const_comparator(); @@ -1103,7 +1097,6 @@ public: void __assign_unique(_ForwardIterator __first, _ForwardIterator __last); template <class _InputIterator> void __assign_multi(_InputIterator __first, _InputIterator __last); -#ifndef _LIBCPP_CXX03_LANG __tree(__tree&& __t) _NOEXCEPT_( is_nothrow_move_constructible<__node_allocator>::value && @@ -1114,8 +1107,6 @@ public: __node_traits::propagate_on_container_move_assignment::value && is_nothrow_move_assignable<value_compare>::value && is_nothrow_move_assignable<__node_allocator>::value); -#endif // _LIBCPP_CXX03_LANG - ~__tree(); _LIBCPP_INLINE_VISIBILITY @@ -1129,7 +1120,7 @@ public: _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT - {return std::min<size_type>( + {return _VSTD::min<size_type>( __node_traits::max_size(__node_alloc()), numeric_limits<difference_type >::max());} @@ -1146,12 +1137,11 @@ public: _NOEXCEPT_(__is_nothrow_swappable<value_compare>::value); #endif -#ifndef _LIBCPP_CXX03_LANG template <class _Key, class ..._Args> pair<iterator, bool> __emplace_unique_key_args(_Key const&, _Args&&... __args); template <class _Key, class ..._Args> - iterator + pair<iterator, bool> __emplace_hint_unique_key_args(const_iterator, _Key const&, _Args&&...); template <class... _Args> @@ -1225,7 +1215,7 @@ public: >::type __emplace_hint_unique(const_iterator __p, _First&& __f, _Second&& __s) { return __emplace_hint_unique_key_args(__p, __f, _VSTD::forward<_First>(__f), - _VSTD::forward<_Second>(__s)); + _VSTD::forward<_Second>(__s)).first; } template <class... _Args> @@ -1245,25 +1235,16 @@ public: _LIBCPP_INLINE_VISIBILITY iterator __emplace_hint_unique_extract_key(const_iterator __p, _Pp&& __x, __extract_key_self_tag) { - return __emplace_hint_unique_key_args(__p, __x, _VSTD::forward<_Pp>(__x)); + return __emplace_hint_unique_key_args(__p, __x, _VSTD::forward<_Pp>(__x)).first; } template <class _Pp> _LIBCPP_INLINE_VISIBILITY iterator __emplace_hint_unique_extract_key(const_iterator __p, _Pp&& __x, __extract_key_first_tag) { - return __emplace_hint_unique_key_args(__p, __x.first, _VSTD::forward<_Pp>(__x)); + return __emplace_hint_unique_key_args(__p, __x.first, _VSTD::forward<_Pp>(__x)).first; } -#else - template <class _Key, class _Args> - _LIBCPP_INLINE_VISIBILITY - pair<iterator, bool> __emplace_unique_key_args(_Key const&, _Args& __args); - template <class _Key, class _Args> - _LIBCPP_INLINE_VISIBILITY - iterator __emplace_hint_unique_key_args(const_iterator, _Key const&, _Args&); -#endif - _LIBCPP_INLINE_VISIBILITY pair<iterator, bool> __insert_unique(const __container_value_type& __v) { return __emplace_unique_key_args(_NodeTypes::__get_key(__v), __v); @@ -1271,15 +1252,9 @@ public: _LIBCPP_INLINE_VISIBILITY iterator __insert_unique(const_iterator __p, const __container_value_type& __v) { - return __emplace_hint_unique_key_args(__p, _NodeTypes::__get_key(__v), __v); + return __emplace_hint_unique_key_args(__p, _NodeTypes::__get_key(__v), __v).first; } -#ifdef _LIBCPP_CXX03_LANG - _LIBCPP_INLINE_VISIBILITY - iterator __insert_multi(const __container_value_type& __v); - _LIBCPP_INLINE_VISIBILITY - iterator __insert_multi(const_iterator __p, const __container_value_type& __v); -#else _LIBCPP_INLINE_VISIBILITY pair<iterator, bool> __insert_unique(__container_value_type&& __v) { return __emplace_unique_key_args(_NodeTypes::__get_key(__v), _VSTD::move(__v)); @@ -1287,7 +1262,7 @@ public: _LIBCPP_INLINE_VISIBILITY iterator __insert_unique(const_iterator __p, __container_value_type&& __v) { - return __emplace_hint_unique_key_args(__p, _NodeTypes::__get_key(__v), _VSTD::move(__v)); + return __emplace_hint_unique_key_args(__p, _NodeTypes::__get_key(__v), _VSTD::move(__v)).first; } template <class _Vp, class = typename enable_if< @@ -1332,8 +1307,6 @@ public: return __emplace_hint_multi(__p, _VSTD::forward<_Vp>(__v)); } -#endif // !_LIBCPP_CXX03_LANG - _LIBCPP_INLINE_VISIBILITY pair<iterator, bool> __node_assign_unique(const __container_value_type& __v, __node_pointer __dest); @@ -1455,7 +1428,7 @@ private: __node_base_pointer& __find_leaf(const_iterator __hint, __parent_pointer& __parent, const key_type& __v); - // FIXME: Make this function const qualified. Unfortunetly doing so + // FIXME: Make this function const qualified. Unfortunately doing so // breaks existing code which uses non-const callable comparators. template <class _Key> __node_base_pointer& @@ -1471,12 +1444,8 @@ private: __node_base_pointer& __dummy, const _Key& __v); -#ifndef _LIBCPP_CXX03_LANG template <class ..._Args> __node_holder __construct_node(_Args&& ...__args); -#else - __node_holder __construct_node(const __container_value_type& __v); -#endif void destroy(__node_pointer __nd) _NOEXCEPT; @@ -1621,20 +1590,20 @@ __tree<_Tp, _Compare, _Allocator>::_DetachedTreeCache::__detach_next(__node_poin { if (__cache->__parent_ == nullptr) return nullptr; - if (__tree_is_left_child(static_cast<__node_base_pointer>(__cache))) + if (_VSTD::__tree_is_left_child(static_cast<__node_base_pointer>(__cache))) { __cache->__parent_->__left_ = nullptr; __cache = static_cast<__node_pointer>(__cache->__parent_); if (__cache->__right_ == nullptr) return __cache; - return static_cast<__node_pointer>(__tree_leaf(__cache->__right_)); + return static_cast<__node_pointer>(_VSTD::__tree_leaf(__cache->__right_)); } // __cache is right child __cache->__parent_unsafe()->__right_ = nullptr; __cache = static_cast<__node_pointer>(__cache->__parent_); if (__cache->__left_ == nullptr) return __cache; - return static_cast<__node_pointer>(__tree_leaf(__cache->__left_)); + return static_cast<__node_pointer>(_VSTD::__tree_leaf(__cache->__left_)); } template <class _Tp, class _Compare, class _Allocator> @@ -1706,8 +1675,6 @@ __tree<_Tp, _Compare, _Allocator>::__tree(const __tree& __t) __begin_node() = __end_node(); } -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Compare, class _Allocator> __tree<_Tp, _Compare, _Allocator>::__tree(__tree&& __t) _NOEXCEPT_( @@ -1814,8 +1781,6 @@ __tree<_Tp, _Compare, _Allocator>::operator=(__tree&& __t) return *this; } -#endif // _LIBCPP_CXX03_LANG - template <class _Tp, class _Compare, class _Allocator> __tree<_Tp, _Compare, _Allocator>::~__tree() { @@ -1854,7 +1819,7 @@ __tree<_Tp, _Compare, _Allocator>::swap(__tree& __t) using _VSTD::swap; swap(__begin_node_, __t.__begin_node_); swap(__pair1_.first(), __t.__pair1_.first()); - __swap_allocator(__node_alloc(), __t.__node_alloc()); + _VSTD::__swap_allocator(__node_alloc(), __t.__node_alloc()); __pair3_.swap(__t.__pair3_); if (size() == 0) __begin_node() = __end_node(); @@ -2113,21 +2078,14 @@ void __tree<_Tp, _Compare, _Allocator>::__insert_node_at( __child = __new_node; if (__begin_node()->__left_ != nullptr) __begin_node() = static_cast<__iter_pointer>(__begin_node()->__left_); - __tree_balance_after_insert(__end_node()->__left_, __child); + _VSTD::__tree_balance_after_insert(__end_node()->__left_, __child); ++size(); } -#ifndef _LIBCPP_CXX03_LANG template <class _Tp, class _Compare, class _Allocator> template <class _Key, class... _Args> pair<typename __tree<_Tp, _Compare, _Allocator>::iterator, bool> __tree<_Tp, _Compare, _Allocator>::__emplace_unique_key_args(_Key const& __k, _Args&&... __args) -#else -template <class _Tp, class _Compare, class _Allocator> -template <class _Key, class _Args> -pair<typename __tree<_Tp, _Compare, _Allocator>::iterator, bool> -__tree<_Tp, _Compare, _Allocator>::__emplace_unique_key_args(_Key const& __k, _Args& __args) -#endif { __parent_pointer __parent; __node_base_pointer& __child = __find_equal(__parent, __k); @@ -2135,11 +2093,7 @@ __tree<_Tp, _Compare, _Allocator>::__emplace_unique_key_args(_Key const& __k, _A bool __inserted = false; if (__child == nullptr) { -#ifndef _LIBCPP_CXX03_LANG __node_holder __h = __construct_node(_VSTD::forward<_Args>(__args)...); -#else - __node_holder __h = __construct_node(__args); -#endif __insert_node_at(__parent, __child, static_cast<__node_base_pointer>(__h.get())); __r = __h.release(); __inserted = true; @@ -2147,41 +2101,27 @@ __tree<_Tp, _Compare, _Allocator>::__emplace_unique_key_args(_Key const& __k, _A return pair<iterator, bool>(iterator(__r), __inserted); } - -#ifndef _LIBCPP_CXX03_LANG template <class _Tp, class _Compare, class _Allocator> template <class _Key, class... _Args> -typename __tree<_Tp, _Compare, _Allocator>::iterator +pair<typename __tree<_Tp, _Compare, _Allocator>::iterator, bool> __tree<_Tp, _Compare, _Allocator>::__emplace_hint_unique_key_args( const_iterator __p, _Key const& __k, _Args&&... __args) -#else -template <class _Tp, class _Compare, class _Allocator> -template <class _Key, class _Args> -typename __tree<_Tp, _Compare, _Allocator>::iterator -__tree<_Tp, _Compare, _Allocator>::__emplace_hint_unique_key_args( - const_iterator __p, _Key const& __k, _Args& __args) -#endif { __parent_pointer __parent; __node_base_pointer __dummy; __node_base_pointer& __child = __find_equal(__p, __parent, __dummy, __k); __node_pointer __r = static_cast<__node_pointer>(__child); + bool __inserted = false; if (__child == nullptr) { -#ifndef _LIBCPP_CXX03_LANG __node_holder __h = __construct_node(_VSTD::forward<_Args>(__args)...); -#else - __node_holder __h = __construct_node(__args); -#endif __insert_node_at(__parent, __child, static_cast<__node_base_pointer>(__h.get())); __r = __h.release(); + __inserted = true; } - return iterator(__r); + return pair<iterator, bool>(iterator(__r), __inserted); } - -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp, class _Compare, class _Allocator> template <class ..._Args> typename __tree<_Tp, _Compare, _Allocator>::__node_holder @@ -2259,46 +2199,6 @@ __tree<_Tp, _Compare, _Allocator>::__emplace_hint_multi(const_iterator __p, return iterator(static_cast<__node_pointer>(__h.release())); } - -#else // _LIBCPP_CXX03_LANG - -template <class _Tp, class _Compare, class _Allocator> -typename __tree<_Tp, _Compare, _Allocator>::__node_holder -__tree<_Tp, _Compare, _Allocator>::__construct_node(const __container_value_type& __v) -{ - __node_allocator& __na = __node_alloc(); - __node_holder __h(__node_traits::allocate(__na, 1), _Dp(__na)); - __node_traits::construct(__na, _NodeTypes::__get_ptr(__h->__value_), __v); - __h.get_deleter().__value_constructed = true; - return _LIBCPP_EXPLICIT_MOVE(__h); // explicitly moved for C++03 -} - -#endif // _LIBCPP_CXX03_LANG - -#ifdef _LIBCPP_CXX03_LANG -template <class _Tp, class _Compare, class _Allocator> -typename __tree<_Tp, _Compare, _Allocator>::iterator -__tree<_Tp, _Compare, _Allocator>::__insert_multi(const __container_value_type& __v) -{ - __parent_pointer __parent; - __node_base_pointer& __child = __find_leaf_high(__parent, _NodeTypes::__get_key(__v)); - __node_holder __h = __construct_node(__v); - __insert_node_at(__parent, __child, static_cast<__node_base_pointer>(__h.get())); - return iterator(__h.release()); -} - -template <class _Tp, class _Compare, class _Allocator> -typename __tree<_Tp, _Compare, _Allocator>::iterator -__tree<_Tp, _Compare, _Allocator>::__insert_multi(const_iterator __p, const __container_value_type& __v) -{ - __parent_pointer __parent; - __node_base_pointer& __child = __find_leaf(__p, __parent, _NodeTypes::__get_key(__v)); - __node_holder __h = __construct_node(__v); - __insert_node_at(__parent, __child, static_cast<__node_base_pointer>(__h.get())); - return iterator(__h.release()); -} -#endif - template <class _Tp, class _Compare, class _Allocator> pair<typename __tree<_Tp, _Compare, _Allocator>::iterator, bool> __tree<_Tp, _Compare, _Allocator>::__node_assign_unique(const __container_value_type& __v, __node_pointer __nd) @@ -2348,8 +2248,8 @@ __tree<_Tp, _Compare, _Allocator>::__remove_node_pointer(__node_pointer __ptr) _ if (__begin_node() == __ptr) __begin_node() = __r.__ptr_; --size(); - __tree_remove(__end_node()->__left_, - static_cast<__node_base_pointer>(__ptr)); + _VSTD::__tree_remove(__end_node()->__left_, + static_cast<__node_base_pointer>(__ptr)); return __r; } @@ -2727,7 +2627,7 @@ __tree<_Tp, _Compare, _Allocator>::__equal_range_unique(const _Key& __k) return _Pp(iterator(__rt), iterator( __rt->__right_ != nullptr ? - static_cast<__iter_pointer>(__tree_min(__rt->__right_)) + static_cast<__iter_pointer>(_VSTD::__tree_min(__rt->__right_)) : __result)); } return _Pp(iterator(__result), iterator(__result)); @@ -2755,7 +2655,7 @@ __tree<_Tp, _Compare, _Allocator>::__equal_range_unique(const _Key& __k) const return _Pp(const_iterator(__rt), const_iterator( __rt->__right_ != nullptr ? - static_cast<__iter_pointer>(__tree_min(__rt->__right_)) + static_cast<__iter_pointer>(_VSTD::__tree_min(__rt->__right_)) : __result)); } return _Pp(const_iterator(__result), const_iterator(__result)); @@ -2824,8 +2724,8 @@ __tree<_Tp, _Compare, _Allocator>::remove(const_iterator __p) _NOEXCEPT __begin_node() = static_cast<__iter_pointer>(__np->__parent_); } --size(); - __tree_remove(__end_node()->__left_, - static_cast<__node_base_pointer>(__np)); + _VSTD::__tree_remove(__end_node()->__left_, + static_cast<__node_base_pointer>(__np)); return __node_holder(__np, _Dp(__node_alloc(), true)); } diff --git a/lib/libcxx/include/algorithm b/lib/libcxx/include/algorithm index 83e49f19ab..f7fb2013a7 100644 --- a/lib/libcxx/include/algorithm +++ b/lib/libcxx/include/algorithm @@ -47,16 +47,16 @@ template <class InputIterator, class Predicate> find_if(InputIterator first, InputIterator last, Predicate pred); template<class InputIterator, class Predicate> - InputIterator // constexpr in C++20 + constexpr InputIterator // constexpr in C++20 find_if_not(InputIterator first, InputIterator last, Predicate pred); template <class ForwardIterator1, class ForwardIterator2> - ForwardIterator1 // constexpr in C++20 + constexpr ForwardIterator1 // constexpr in C++20 find_end(ForwardIterator1 first1, ForwardIterator1 last1, ForwardIterator2 first2, ForwardIterator2 last2); template <class ForwardIterator1, class ForwardIterator2, class BinaryPredicate> - ForwardIterator1 // constexpr in C++20 + constexpr ForwardIterator1 // constexpr in C++20 find_end(ForwardIterator1 first1, ForwardIterator1 last1, ForwardIterator2 first2, ForwardIterator2 last2, BinaryPredicate pred); @@ -185,11 +185,11 @@ template <class BidirectionalIterator1, class BidirectionalIterator2> BidirectionalIterator2 result); template <class ForwardIterator1, class ForwardIterator2> - ForwardIterator2 + constexpr ForwardIterator2 // constexpr in C++20 swap_ranges(ForwardIterator1 first1, ForwardIterator1 last1, ForwardIterator2 first2); template <class ForwardIterator1, class ForwardIterator2> - void + constexpr void // constexpr in C++20 iter_swap(ForwardIterator1 a, ForwardIterator2 b); template <class InputIterator, class OutputIterator, class UnaryOperation> @@ -251,23 +251,23 @@ template <class InputIterator, class OutputIterator, class Predicate> remove_copy_if(InputIterator first, InputIterator last, OutputIterator result, Predicate pred); template <class ForwardIterator> - ForwardIterator + constexpr ForwardIterator // constexpr in C++20 unique(ForwardIterator first, ForwardIterator last); template <class ForwardIterator, class BinaryPredicate> - ForwardIterator + constexpr ForwardIterator // constexpr in C++20 unique(ForwardIterator first, ForwardIterator last, BinaryPredicate pred); template <class InputIterator, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr in C++20 unique_copy(InputIterator first, InputIterator last, OutputIterator result); template <class InputIterator, class OutputIterator, class BinaryPredicate> - OutputIterator + constexpr OutputIterator // constexpr in C++20 unique_copy(InputIterator first, InputIterator last, OutputIterator result, BinaryPredicate pred); template <class BidirectionalIterator> - void + constexpr void // constexpr in C++20 reverse(BidirectionalIterator first, BidirectionalIterator last); template <class BidirectionalIterator, class OutputIterator> @@ -275,11 +275,11 @@ template <class BidirectionalIterator, class OutputIterator> reverse_copy(BidirectionalIterator first, BidirectionalIterator last, OutputIterator result); template <class ForwardIterator> - ForwardIterator + constexpr ForwardIterator // constexpr in C++20 rotate(ForwardIterator first, ForwardIterator middle, ForwardIterator last); template <class ForwardIterator, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr in C++20 rotate_copy(ForwardIterator first, ForwardIterator middle, ForwardIterator last, OutputIterator result); template <class RandomAccessIterator> @@ -301,12 +301,22 @@ template<class RandomAccessIterator, class UniformRandomNumberGenerator> void shuffle(RandomAccessIterator first, RandomAccessIterator last, UniformRandomNumberGenerator&& g); +template<class ForwardIterator> + constexpr ForwardIterator + shift_left(ForwardIterator first, ForwardIterator last, + typename iterator_traits<ForwardIterator>::difference_type n); // C++20 + +template<class ForwardIterator> + constexpr ForwardIterator + shift_right(ForwardIterator first, ForwardIterator last, + typename iterator_traits<ForwardIterator>::difference_type n); // C++20 + template <class InputIterator, class Predicate> constexpr bool // constexpr in C++20 is_partitioned(InputIterator first, InputIterator last, Predicate pred); template <class ForwardIterator, class Predicate> - ForwardIterator + constexpr ForwardIterator // constexpr in C++20 partition(ForwardIterator first, ForwardIterator last, Predicate pred); template <class InputIterator, class OutputIterator1, @@ -329,7 +339,7 @@ template <class ForwardIterator> is_sorted(ForwardIterator first, ForwardIterator last); template <class ForwardIterator, class Compare> - bool + constexpr bool // constexpr in C++20 is_sorted(ForwardIterator first, ForwardIterator last, Compare comp); template<class ForwardIterator> @@ -415,12 +425,12 @@ template <class ForwardIterator, class T, class Compare> binary_search(ForwardIterator first, ForwardIterator last, const T& value, Compare comp); template <class InputIterator1, class InputIterator2, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr in C++20 merge(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result); template <class InputIterator1, class InputIterator2, class OutputIterator, class Compare> - OutputIterator + constexpr OutputIterator // constexpr in C++20 merge(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result, Compare comp); @@ -441,12 +451,12 @@ template <class InputIterator1, class InputIterator2, class Compare> includes(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, Compare comp); template <class InputIterator1, class InputIterator2, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr in C++20 set_union(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result); template <class InputIterator1, class InputIterator2, class OutputIterator, class Compare> - OutputIterator + constexpr OutputIterator // constexpr in C++20 set_union(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result, Compare comp); @@ -461,22 +471,22 @@ template <class InputIterator1, class InputIterator2, class OutputIterator, clas InputIterator2 first2, InputIterator2 last2, OutputIterator result, Compare comp); template <class InputIterator1, class InputIterator2, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr in C++20 set_difference(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result); template <class InputIterator1, class InputIterator2, class OutputIterator, class Compare> - OutputIterator + constexpr OutputIterator // constexpr in C++20 set_difference(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result, Compare comp); template <class InputIterator1, class InputIterator2, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr in C++20 set_symmetric_difference(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result); template <class InputIterator1, class InputIterator2, class OutputIterator, class Compare> - OutputIterator + constexpr OutputIterator // constexpr in C++20 set_symmetric_difference(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, InputIterator2 last2, OutputIterator result, Compare comp); @@ -529,82 +539,82 @@ template <class RandomAccessIterator, class Compare> is_heap_until(RandomAccessIterator first, RandomAccessiterator last, Compare comp); template <class ForwardIterator> - ForwardIterator - min_element(ForwardIterator first, ForwardIterator last); // constexpr in C++14 + constexpr ForwardIterator // constexpr in C++14 + min_element(ForwardIterator first, ForwardIterator last); template <class ForwardIterator, class Compare> - ForwardIterator - min_element(ForwardIterator first, ForwardIterator last, Compare comp); // constexpr in C++14 + constexpr ForwardIterator // constexpr in C++14 + min_element(ForwardIterator first, ForwardIterator last, Compare comp); template <class T> - const T& - min(const T& a, const T& b); // constexpr in C++14 + constexpr const T& // constexpr in C++14 + min(const T& a, const T& b); template <class T, class Compare> - const T& - min(const T& a, const T& b, Compare comp); // constexpr in C++14 + constexpr const T& // constexpr in C++14 + min(const T& a, const T& b, Compare comp); template<class T> - T - min(initializer_list<T> t); // constexpr in C++14 + constexpr T // constexpr in C++14 + min(initializer_list<T> t); template<class T, class Compare> - T - min(initializer_list<T> t, Compare comp); // constexpr in C++14 + constexpr T // constexpr in C++14 + min(initializer_list<T> t, Compare comp); template<class T> - constexpr const T& clamp( const T& v, const T& lo, const T& hi ); // C++17 + constexpr const T& clamp(const T& v, const T& lo, const T& hi); // C++17 template<class T, class Compare> - constexpr const T& clamp( const T& v, const T& lo, const T& hi, Compare comp ); // C++17 + constexpr const T& clamp(const T& v, const T& lo, const T& hi, Compare comp); // C++17 template <class ForwardIterator> - ForwardIterator - max_element(ForwardIterator first, ForwardIterator last); // constexpr in C++14 + constexpr ForwardIterator // constexpr in C++14 + max_element(ForwardIterator first, ForwardIterator last); template <class ForwardIterator, class Compare> - ForwardIterator - max_element(ForwardIterator first, ForwardIterator last, Compare comp); // constexpr in C++14 + constexpr ForwardIterator // constexpr in C++14 + max_element(ForwardIterator first, ForwardIterator last, Compare comp); template <class T> - const T& - max(const T& a, const T& b); // constexpr in C++14 + constexpr const T& // constexpr in C++14 + max(const T& a, const T& b); template <class T, class Compare> - const T& - max(const T& a, const T& b, Compare comp); // constexpr in C++14 + constexpr const T& // constexpr in C++14 + max(const T& a, const T& b, Compare comp); template<class T> - T - max(initializer_list<T> t); // constexpr in C++14 + constexpr T // constexpr in C++14 + max(initializer_list<T> t); template<class T, class Compare> - T - max(initializer_list<T> t, Compare comp); // constexpr in C++14 + constexpr T // constexpr in C++14 + max(initializer_list<T> t, Compare comp); template<class ForwardIterator> - pair<ForwardIterator, ForwardIterator> - minmax_element(ForwardIterator first, ForwardIterator last); // constexpr in C++14 + constexpr pair<ForwardIterator, ForwardIterator> // constexpr in C++14 + minmax_element(ForwardIterator first, ForwardIterator last); template<class ForwardIterator, class Compare> - pair<ForwardIterator, ForwardIterator> - minmax_element(ForwardIterator first, ForwardIterator last, Compare comp); // constexpr in C++14 + constexpr pair<ForwardIterator, ForwardIterator> // constexpr in C++14 + minmax_element(ForwardIterator first, ForwardIterator last, Compare comp); template<class T> - pair<const T&, const T&> - minmax(const T& a, const T& b); // constexpr in C++14 + constexpr pair<const T&, const T&> // constexpr in C++14 + minmax(const T& a, const T& b); template<class T, class Compare> - pair<const T&, const T&> - minmax(const T& a, const T& b, Compare comp); // constexpr in C++14 + constexpr pair<const T&, const T&> // constexpr in C++14 + minmax(const T& a, const T& b, Compare comp); template<class T> - pair<T, T> - minmax(initializer_list<T> t); // constexpr in C++14 + constexpr pair<T, T> // constexpr in C++14 + minmax(initializer_list<T> t); template<class T, class Compare> - pair<T, T> - minmax(initializer_list<T> t, Compare comp); // constexpr in C++14 + constexpr pair<T, T> // constexpr in C++14 + minmax(initializer_list<T> t, Compare comp); template <class InputIterator1, class InputIterator2> constexpr bool // constexpr in C++20 @@ -616,19 +626,19 @@ template <class InputIterator1, class InputIterator2, class Compare> InputIterator2 first2, InputIterator2 last2, Compare comp); template <class BidirectionalIterator> - bool + constexpr bool // constexpr in C++20 next_permutation(BidirectionalIterator first, BidirectionalIterator last); template <class BidirectionalIterator, class Compare> - bool + constexpr bool // constexpr in C++20 next_permutation(BidirectionalIterator first, BidirectionalIterator last, Compare comp); template <class BidirectionalIterator> - bool + constexpr bool // constexpr in C++20 prev_permutation(BidirectionalIterator first, BidirectionalIterator last); template <class BidirectionalIterator, class Compare> - bool + constexpr bool // constexpr in C++20 prev_permutation(BidirectionalIterator first, BidirectionalIterator last, Compare comp); } // std @@ -895,7 +905,7 @@ inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _InputIterator for_each_n(_InputIterator __first, _Size __orig_n, _Function __f) { - typedef decltype(__convert_to_integral(__orig_n)) _IntegralSize; + typedef decltype(_VSTD::__convert_to_integral(__orig_n)) _IntegralSize; _IntegralSize __n = __orig_n; while (__n > 0) { @@ -1614,7 +1624,7 @@ search_n(_ForwardIterator __first, _ForwardIterator __last, _Size __count, const _Tp& __value_, _BinaryPredicate __pred) { return _VSTD::__search_n<typename add_lvalue_reference<_BinaryPredicate>::type> - (__first, __last, __convert_to_integral(__count), __value_, __pred, + (__first, __last, _VSTD::__convert_to_integral(__count), __value_, __pred, typename iterator_traits<_ForwardIterator>::iterator_category()); } @@ -1625,13 +1635,13 @@ _ForwardIterator search_n(_ForwardIterator __first, _ForwardIterator __last, _Size __count, const _Tp& __value_) { typedef typename iterator_traits<_ForwardIterator>::value_type __v; - return _VSTD::search_n(__first, __last, __convert_to_integral(__count), + return _VSTD::search_n(__first, __last, _VSTD::__convert_to_integral(__count), __value_, __equal_to<__v, _Tp>()); } // copy template <class _Iter> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR _Iter __unwrap_iter(_Iter __i) { @@ -1639,7 +1649,7 @@ __unwrap_iter(_Iter __i) } template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1653,7 +1663,7 @@ __unwrap_iter(move_iterator<_Tp*> __i) #if _LIBCPP_DEBUG_LEVEL < 2 template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1665,7 +1675,7 @@ __unwrap_iter(__wrap_iter<_Tp*> __i) } template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1679,7 +1689,7 @@ __unwrap_iter(__wrap_iter<const _Tp*> __i) #else template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1707,7 +1717,7 @@ inline _LIBCPP_INLINE_VISIBILITY _OutputIterator __copy(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { - return __copy_constexpr(__first, __last, __result); + return _VSTD::__copy_constexpr(__first, __last, __result); } template <class _Tp, class _Up> @@ -1727,16 +1737,16 @@ __copy(_Tp* __first, _Tp* __last, _Up* __result) } template <class _InputIterator, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator copy(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { if (__libcpp_is_constant_evaluated()) { return _VSTD::__copy_constexpr( - __unwrap_iter(__first), __unwrap_iter(__last), __unwrap_iter(__result)); + _VSTD::__unwrap_iter(__first), _VSTD::__unwrap_iter(__last), _VSTD::__unwrap_iter(__result)); } else { return _VSTD::__copy( - __unwrap_iter(__first), __unwrap_iter(__last), __unwrap_iter(__result)); + _VSTD::__unwrap_iter(__first), _VSTD::__unwrap_iter(__last), _VSTD::__unwrap_iter(__result)); } } @@ -1757,7 +1767,7 @@ inline _LIBCPP_INLINE_VISIBILITY _OutputIterator __copy_backward(_BidirectionalIterator __first, _BidirectionalIterator __last, _OutputIterator __result) { - return __copy_backward_constexpr(__first, __last, __result); + return _VSTD::__copy_backward_constexpr(__first, __last, __result); } template <class _Tp, class _Up> @@ -1780,19 +1790,19 @@ __copy_backward(_Tp* __first, _Tp* __last, _Up* __result) } template <class _BidirectionalIterator1, class _BidirectionalIterator2> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _BidirectionalIterator2 copy_backward(_BidirectionalIterator1 __first, _BidirectionalIterator1 __last, _BidirectionalIterator2 __result) { if (__libcpp_is_constant_evaluated()) { - return _VSTD::__copy_backward_constexpr(__unwrap_iter(__first), - __unwrap_iter(__last), - __unwrap_iter(__result)); + return _VSTD::__copy_backward_constexpr(_VSTD::__unwrap_iter(__first), + _VSTD::__unwrap_iter(__last), + _VSTD::__unwrap_iter(__result)); } else { - return _VSTD::__copy_backward(__unwrap_iter(__first), - __unwrap_iter(__last), - __unwrap_iter(__result)); + return _VSTD::__copy_backward(_VSTD::__unwrap_iter(__first), + _VSTD::__unwrap_iter(__last), + _VSTD::__unwrap_iter(__result)); } } @@ -1818,7 +1828,7 @@ copy_if(_InputIterator __first, _InputIterator __last, // copy_n template<class _InputIterator, class _Size, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename enable_if < __is_cpp17_input_iterator<_InputIterator>::value && @@ -1827,7 +1837,7 @@ typename enable_if >::type copy_n(_InputIterator __first, _Size __orig_n, _OutputIterator __result) { - typedef decltype(__convert_to_integral(__orig_n)) _IntegralSize; + typedef decltype(_VSTD::__convert_to_integral(__orig_n)) _IntegralSize; _IntegralSize __n = __orig_n; if (__n > 0) { @@ -1844,7 +1854,7 @@ copy_n(_InputIterator __first, _Size __orig_n, _OutputIterator __result) } template<class _InputIterator, class _Size, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17_WITH_IS_CONSTANT_EVALUATED +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename enable_if < __is_cpp17_random_access_iterator<_InputIterator>::value, @@ -1852,25 +1862,35 @@ typename enable_if >::type copy_n(_InputIterator __first, _Size __orig_n, _OutputIterator __result) { - typedef decltype(__convert_to_integral(__orig_n)) _IntegralSize; + typedef decltype(_VSTD::__convert_to_integral(__orig_n)) _IntegralSize; _IntegralSize __n = __orig_n; return _VSTD::copy(__first, __first + __n, __result); } // move +// __move_constexpr exists so that __move doesn't call itself when delegating to the constexpr +// version of __move. template <class _InputIterator, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX14 _OutputIterator -__move(_InputIterator __first, _InputIterator __last, _OutputIterator __result) +__move_constexpr(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { for (; __first != __last; ++__first, (void) ++__result) *__result = _VSTD::move(*__first); return __result; } +template <class _InputIterator, class _OutputIterator> +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX14 +_OutputIterator +__move(_InputIterator __first, _InputIterator __last, _OutputIterator __result) +{ + return _VSTD::__move_constexpr(__first, __last, __result); +} + template <class _Tp, class _Up> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX14 typename enable_if < is_same<typename remove_const<_Tp>::type, _Up>::value && @@ -1879,6 +1899,8 @@ typename enable_if >::type __move(_Tp* __first, _Tp* __last, _Up* __result) { + if (__libcpp_is_constant_evaluated()) + return _VSTD::__move_constexpr(__first, __last, __result); const size_t __n = static_cast<size_t>(__last - __first); if (__n > 0) _VSTD::memmove(__result, __first, __n * sizeof(_Up)); @@ -1886,27 +1908,37 @@ __move(_Tp* __first, _Tp* __last, _Up* __result) } template <class _InputIterator, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator move(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { - return _VSTD::__move(__unwrap_iter(__first), __unwrap_iter(__last), __unwrap_iter(__result)); + return _VSTD::__move(_VSTD::__unwrap_iter(__first), _VSTD::__unwrap_iter(__last), _VSTD::__unwrap_iter(__result)); } // move_backward +// __move_backward_constexpr exists so that __move_backward doesn't call itself when delegating to +// the constexpr version of __move_backward. template <class _InputIterator, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX14 _OutputIterator -__move_backward(_InputIterator __first, _InputIterator __last, _OutputIterator __result) +__move_backward_constexpr(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { while (__first != __last) *--__result = _VSTD::move(*--__last); return __result; } +template <class _InputIterator, class _OutputIterator> +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX14 +_OutputIterator +__move_backward(_InputIterator __first, _InputIterator __last, _OutputIterator __result) +{ + return _VSTD::__move_backward_constexpr(__first, __last, __result); +} + template <class _Tp, class _Up> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX14 typename enable_if < is_same<typename remove_const<_Tp>::type, _Up>::value && @@ -1915,6 +1947,8 @@ typename enable_if >::type __move_backward(_Tp* __first, _Tp* __last, _Up* __result) { + if (__libcpp_is_constant_evaluated()) + return _VSTD::__move_backward_constexpr(__first, __last, __result); const size_t __n = static_cast<size_t>(__last - __first); if (__n > 0) { @@ -1925,12 +1959,12 @@ __move_backward(_Tp* __first, _Tp* __last, _Up* __result) } template <class _BidirectionalIterator1, class _BidirectionalIterator2> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _BidirectionalIterator2 move_backward(_BidirectionalIterator1 __first, _BidirectionalIterator1 __last, _BidirectionalIterator2 __result) { - return _VSTD::__move_backward(__unwrap_iter(__first), __unwrap_iter(__last), __unwrap_iter(__result)); + return _VSTD::__move_backward(_VSTD::__unwrap_iter(__first), _VSTD::__unwrap_iter(__last), _VSTD::__unwrap_iter(__result)); } // iter_swap @@ -2033,7 +2067,7 @@ inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator fill_n(_OutputIterator __first, _Size __n, const _Tp& __value_) { - return _VSTD::__fill_n(__first, __convert_to_integral(__n), __value_); + return _VSTD::__fill_n(__first, _VSTD::__convert_to_integral(__n), __value_); } // fill @@ -2081,7 +2115,7 @@ inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator generate_n(_OutputIterator __first, _Size __orig_n, _Generator __gen) { - typedef decltype(__convert_to_integral(__orig_n)) _IntegralSize; + typedef decltype(_VSTD::__convert_to_integral(__orig_n)) _IntegralSize; _IntegralSize __n = __orig_n; for (; __n > 0; ++__first, (void) --__n) *__first = __gen(); @@ -2287,7 +2321,7 @@ unique_copy(_InputIterator __first, _InputIterator __last, _OutputIterator __res // reverse template <class _BidirectionalIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 void __reverse(_BidirectionalIterator __first, _BidirectionalIterator __last, bidirectional_iterator_tag) { @@ -2301,7 +2335,7 @@ __reverse(_BidirectionalIterator __first, _BidirectionalIterator __last, bidirec } template <class _RandomAccessIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 void __reverse(_RandomAccessIterator __first, _RandomAccessIterator __last, random_access_iterator_tag) { @@ -2311,7 +2345,7 @@ __reverse(_RandomAccessIterator __first, _RandomAccessIterator __last, random_ac } template <class _BidirectionalIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 void reverse(_BidirectionalIterator __first, _BidirectionalIterator __last) { @@ -2333,7 +2367,7 @@ reverse_copy(_BidirectionalIterator __first, _BidirectionalIterator __last, _Out // rotate template <class _ForwardIterator> -_ForwardIterator +_LIBCPP_CONSTEXPR_AFTER_CXX11 _ForwardIterator __rotate_left(_ForwardIterator __first, _ForwardIterator __last) { typedef typename iterator_traits<_ForwardIterator>::value_type value_type; @@ -2344,7 +2378,7 @@ __rotate_left(_ForwardIterator __first, _ForwardIterator __last) } template <class _BidirectionalIterator> -_BidirectionalIterator +_LIBCPP_CONSTEXPR_AFTER_CXX11 _BidirectionalIterator __rotate_right(_BidirectionalIterator __first, _BidirectionalIterator __last) { typedef typename iterator_traits<_BidirectionalIterator>::value_type value_type; @@ -2356,7 +2390,7 @@ __rotate_right(_BidirectionalIterator __first, _BidirectionalIterator __last) } template <class _ForwardIterator> -_ForwardIterator +_LIBCPP_CONSTEXPR_AFTER_CXX14 _ForwardIterator __rotate_forward(_ForwardIterator __first, _ForwardIterator __middle, _ForwardIterator __last) { _ForwardIterator __i = __middle; @@ -2392,7 +2426,7 @@ __rotate_forward(_ForwardIterator __first, _ForwardIterator __middle, _ForwardIt template<typename _Integral> inline _LIBCPP_INLINE_VISIBILITY -_Integral +_LIBCPP_CONSTEXPR_AFTER_CXX14 _Integral __algo_gcd(_Integral __x, _Integral __y) { do @@ -2405,7 +2439,7 @@ __algo_gcd(_Integral __x, _Integral __y) } template<typename _RandomAccessIterator> -_RandomAccessIterator +_LIBCPP_CONSTEXPR_AFTER_CXX14 _RandomAccessIterator __rotate_gcd(_RandomAccessIterator __first, _RandomAccessIterator __middle, _RandomAccessIterator __last) { typedef typename iterator_traits<_RandomAccessIterator>::difference_type difference_type; @@ -2441,7 +2475,7 @@ __rotate_gcd(_RandomAccessIterator __first, _RandomAccessIterator __middle, _Ran template <class _ForwardIterator> inline _LIBCPP_INLINE_VISIBILITY -_ForwardIterator +_LIBCPP_CONSTEXPR_AFTER_CXX11 _ForwardIterator __rotate(_ForwardIterator __first, _ForwardIterator __middle, _ForwardIterator __last, _VSTD::forward_iterator_tag) { @@ -2456,7 +2490,7 @@ __rotate(_ForwardIterator __first, _ForwardIterator __middle, _ForwardIterator _ template <class _BidirectionalIterator> inline _LIBCPP_INLINE_VISIBILITY -_BidirectionalIterator +_LIBCPP_CONSTEXPR_AFTER_CXX11 _BidirectionalIterator __rotate(_BidirectionalIterator __first, _BidirectionalIterator __middle, _BidirectionalIterator __last, _VSTD::bidirectional_iterator_tag) { @@ -2473,7 +2507,7 @@ __rotate(_BidirectionalIterator __first, _BidirectionalIterator __middle, _Bidir template <class _RandomAccessIterator> inline _LIBCPP_INLINE_VISIBILITY -_RandomAccessIterator +_LIBCPP_CONSTEXPR_AFTER_CXX11 _RandomAccessIterator __rotate(_RandomAccessIterator __first, _RandomAccessIterator __middle, _RandomAccessIterator __last, _VSTD::random_access_iterator_tag) { @@ -2491,7 +2525,7 @@ __rotate(_RandomAccessIterator __first, _RandomAccessIterator __middle, _RandomA template <class _ForwardIterator> inline _LIBCPP_INLINE_VISIBILITY -_ForwardIterator +_LIBCPP_CONSTEXPR_AFTER_CXX17 _ForwardIterator rotate(_ForwardIterator __first, _ForwardIterator __middle, _ForwardIterator __last) { if (__first == __middle) @@ -2505,7 +2539,7 @@ rotate(_ForwardIterator __first, _ForwardIterator __middle, _ForwardIterator __l // rotate_copy template <class _ForwardIterator, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator rotate_copy(_ForwardIterator __first, _ForwardIterator __middle, _ForwardIterator __last, _OutputIterator __result) { @@ -2684,12 +2718,12 @@ clamp(const _Tp& __v, const _Tp& __lo, const _Tp& __hi) template <class _ForwardIterator, class _Compare> _LIBCPP_NODISCARD_EXT _LIBCPP_CONSTEXPR_AFTER_CXX11 -std::pair<_ForwardIterator, _ForwardIterator> +pair<_ForwardIterator, _ForwardIterator> minmax_element(_ForwardIterator __first, _ForwardIterator __last, _Compare __comp) { static_assert(__is_cpp17_forward_iterator<_ForwardIterator>::value, "std::minmax_element requires a ForwardIterator"); - std::pair<_ForwardIterator, _ForwardIterator> __result(__first, __first); + pair<_ForwardIterator, _ForwardIterator> __result(__first, __first); if (__first != __last) { if (++__first != __last) @@ -2735,7 +2769,7 @@ minmax_element(_ForwardIterator __first, _ForwardIterator __last, _Compare __com template <class _ForwardIterator> _LIBCPP_NODISCARD_EXT inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX11 -std::pair<_ForwardIterator, _ForwardIterator> +pair<_ForwardIterator, _ForwardIterator> minmax_element(_ForwardIterator __first, _ForwardIterator __last) { return _VSTD::minmax_element(__first, __last, @@ -2774,7 +2808,7 @@ minmax(initializer_list<_Tp> __t, _Compare __comp) typedef typename initializer_list<_Tp>::const_iterator _Iter; _Iter __first = __t.begin(); _Iter __last = __t.end(); - std::pair<_Tp, _Tp> __result(*__first, *__first); + pair<_Tp, _Tp> __result(*__first, *__first); ++__first; if (__t.size() % 2 == 0) @@ -3005,9 +3039,17 @@ private: public: // constructors and reset functions - explicit uniform_int_distribution(result_type __a = 0, - result_type __b = numeric_limits<result_type>::max()) +#ifndef _LIBCPP_CXX03_LANG + uniform_int_distribution() : uniform_int_distribution(0) {} + explicit uniform_int_distribution( + result_type __a, result_type __b = numeric_limits<result_type>::max()) + : __p_(param_type(__a, __b)) {} +#else + explicit uniform_int_distribution( + result_type __a = 0, + result_type __b = numeric_limits<result_type>::max()) : __p_(param_type(__a, __b)) {} +#endif explicit uniform_int_distribution(const param_type& __p) : __p_(__p) {} void reset() {} @@ -3050,7 +3092,7 @@ _LIBCPP_DISABLE_UBSAN_UNSIGNED_INTEGER_CHECK if (_Rp == 0) return static_cast<result_type>(_Eng(__g, _Dt)()); size_t __w = _Dt - __libcpp_clz(_Rp) - 1; - if ((_Rp & (std::numeric_limits<_UIntType>::max() >> (_Dt - __w))) != 0) + if ((_Rp & (numeric_limits<_UIntType>::max() >> (_Dt - __w))) != 0) ++__w; _Eng __e(__g, __w); _UIntType __u; @@ -3227,6 +3269,111 @@ template<class _RandomAccessIterator, class _UniformRandomNumberGenerator> } } +#if _LIBCPP_STD_VER > 17 + +// shift_left, shift_right + +template <class _ForwardIterator> +inline _LIBCPP_INLINE_VISIBILITY constexpr +_ForwardIterator +shift_left(_ForwardIterator __first, _ForwardIterator __last, + typename iterator_traits<_ForwardIterator>::difference_type __n) +{ + if (__n == 0) { + return __last; + } + + _ForwardIterator __m = __first; + if constexpr (__is_cpp17_random_access_iterator<_ForwardIterator>::value) { + if (__n >= __last - __first) { + return __first; + } + __m += __n; + } else { + for (; __n > 0; --__n) { + if (__m == __last) { + return __first; + } + ++__m; + } + } + return _VSTD::move(__m, __last, __first); +} + +template <class _ForwardIterator> +inline _LIBCPP_INLINE_VISIBILITY constexpr +_ForwardIterator +shift_right(_ForwardIterator __first, _ForwardIterator __last, + typename iterator_traits<_ForwardIterator>::difference_type __n) +{ + if (__n == 0) { + return __first; + } + + if constexpr (__is_cpp17_random_access_iterator<_ForwardIterator>::value) { + decltype(__n) __d = __last - __first; + if (__n >= __d) { + return __last; + } + _ForwardIterator __m = __first + (__d - __n); + return _VSTD::move_backward(__first, __m, __last); + } else if constexpr (__is_cpp17_bidirectional_iterator<_ForwardIterator>::value) { + _ForwardIterator __m = __last; + for (; __n > 0; --__n) { + if (__m == __first) { + return __last; + } + --__m; + } + return _VSTD::move_backward(__first, __m, __last); + } else { + _ForwardIterator __ret = __first; + for (; __n > 0; --__n) { + if (__ret == __last) { + return __last; + } + ++__ret; + } + + // We have an __n-element scratch space from __first to __ret. + // Slide an __n-element window [__trail, __lead) from left to right. + // We're essentially doing swap_ranges(__first, __ret, __trail, __lead) + // over and over; but once __lead reaches __last we needn't bother + // to save the values of elements [__trail, __last). + + auto __trail = __first; + auto __lead = __ret; + while (__trail != __ret) { + if (__lead == __last) { + _VSTD::move(__first, __trail, __ret); + return __ret; + } + ++__trail; + ++__lead; + } + + _ForwardIterator __mid = __first; + while (true) { + if (__lead == __last) { + __trail = _VSTD::move(__mid, __ret, __trail); + _VSTD::move(__first, __mid, __trail); + return __ret; + } + swap(*__mid, *__trail); + ++__mid; + ++__trail; + ++__lead; + if (__mid == __ret) { + __mid = __first; + } + } + } +} + +#endif // _LIBCPP_STD_VER > 17 + +// is_partitioned + template <class _InputIterator, class _Predicate> _LIBCPP_NODISCARD_EXT _LIBCPP_CONSTEXPR_AFTER_CXX17 bool is_partitioned(_InputIterator __first, _InputIterator __last, _Predicate __pred) @@ -3246,7 +3393,7 @@ is_partitioned(_InputIterator __first, _InputIterator __last, _Predicate __pred) // partition template <class _Predicate, class _ForwardIterator> -_ForwardIterator +_LIBCPP_CONSTEXPR_AFTER_CXX17 _ForwardIterator __partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate __pred, forward_iterator_tag) { while (true) @@ -3269,7 +3416,7 @@ __partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate __pred } template <class _Predicate, class _BidirectionalIterator> -_BidirectionalIterator +_LIBCPP_CONSTEXPR_AFTER_CXX17 _BidirectionalIterator __partition(_BidirectionalIterator __first, _BidirectionalIterator __last, _Predicate __pred, bidirectional_iterator_tag) { @@ -3294,7 +3441,7 @@ __partition(_BidirectionalIterator __first, _BidirectionalIterator __last, _Pred } template <class _ForwardIterator, class _Predicate> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _ForwardIterator partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate __pred) { @@ -3380,8 +3527,8 @@ __stable_partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate // Move the falses into the temporary buffer, and the trues to the front of the line // Update __first to always point to the end of the trues value_type* __t = __p.first; - ::new(__t) value_type(_VSTD::move(*__first)); - __d.__incr((value_type*)0); + ::new ((void*)__t) value_type(_VSTD::move(*__first)); + __d.template __incr<value_type>(); ++__t; _ForwardIterator __i = __first; while (++__i != __last) @@ -3393,8 +3540,8 @@ __stable_partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate } else { - ::new(__t) value_type(_VSTD::move(*__i)); - __d.__incr((value_type*)0); + ::new ((void*)__t) value_type(_VSTD::move(*__i)); + __d.template __incr<value_type>(); ++__t; } } @@ -3415,7 +3562,7 @@ __stable_partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate // F????????????????? // f m l typedef typename add_lvalue_reference<_Predicate>::type _PredRef; - _ForwardIterator __first_false = __stable_partition<_PredRef>(__first, __m, __pred, __len2, __p, __fit); + _ForwardIterator __first_false = _VSTD::__stable_partition<_PredRef>(__first, __m, __pred, __len2, __p, __fit); // TTTFFFFF?????????? // f ff m l // recurse on [__m, __last], except increase __m until *(__m) is false, *__last know to be true @@ -3430,7 +3577,7 @@ __stable_partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate } // TTTFFFFFTTTF?????? // f ff m m1 l - __second_false = __stable_partition<_PredRef>(__m1, __last, __pred, __len_half, __p, __fit); + __second_false = _VSTD::__stable_partition<_PredRef>(__m1, __last, __pred, __len_half, __p, __fit); __second_half_done: // TTTFFFFFTTTTTFFFFF // f ff m sf l @@ -3472,7 +3619,7 @@ __stable_partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate __p = _VSTD::get_temporary_buffer<value_type>(__len); __h.reset(__p.first); } - return __stable_partition<typename add_lvalue_reference<_Predicate>::type> + return _VSTD::__stable_partition<typename add_lvalue_reference<_Predicate>::type> (__first, __last, __pred, __len, __p, forward_iterator_tag()); } @@ -3510,8 +3657,8 @@ __stable_partition(_BidirectionalIterator __first, _BidirectionalIterator __last // Move the falses into the temporary buffer, and the trues to the front of the line // Update __first to always point to the end of the trues value_type* __t = __p.first; - ::new(__t) value_type(_VSTD::move(*__first)); - __d.__incr((value_type*)0); + ::new ((void*)__t) value_type(_VSTD::move(*__first)); + __d.template __incr<value_type>(); ++__t; _BidirectionalIterator __i = __first; while (++__i != __last) @@ -3523,8 +3670,8 @@ __stable_partition(_BidirectionalIterator __first, _BidirectionalIterator __last } else { - ::new(__t) value_type(_VSTD::move(*__i)); - __d.__incr((value_type*)0); + ::new ((void*)__t) value_type(_VSTD::move(*__i)); + __d.template __incr<value_type>(); ++__t; } } @@ -3558,7 +3705,7 @@ __stable_partition(_BidirectionalIterator __first, _BidirectionalIterator __last // F???TFFF?????????T // f m1 m l typedef typename add_lvalue_reference<_Predicate>::type _PredRef; - __first_false = __stable_partition<_PredRef>(__first, __m1, __pred, __len_half, __p, __bit); + __first_false = _VSTD::__stable_partition<_PredRef>(__first, __m1, __pred, __len_half, __p, __bit); __first_half_done: // TTTFFFFF?????????T // f ff m l @@ -3575,7 +3722,7 @@ __first_half_done: } // TTTFFFFFTTTF?????T // f ff m m1 l - __second_false = __stable_partition<_PredRef>(__m1, __last, __pred, __len_half, __p, __bit); + __second_false = _VSTD::__stable_partition<_PredRef>(__m1, __last, __pred, __len_half, __p, __bit); __second_half_done: // TTTFFFFFTTTTTFFFFF // f ff m sf l @@ -3620,7 +3767,7 @@ __stable_partition(_BidirectionalIterator __first, _BidirectionalIterator __last __p = _VSTD::get_temporary_buffer<value_type>(__len); __h.reset(__p.first); } - return __stable_partition<typename add_lvalue_reference<_Predicate>::type> + return _VSTD::__stable_partition<typename add_lvalue_reference<_Predicate>::type> (__first, __last, __pred, __len, __p, bidirectional_iterator_tag()); } @@ -3629,7 +3776,7 @@ inline _LIBCPP_INLINE_VISIBILITY _ForwardIterator stable_partition(_ForwardIterator __first, _ForwardIterator __last, _Predicate __pred) { - return __stable_partition<typename add_lvalue_reference<_Predicate>::type> + return _VSTD::__stable_partition<typename add_lvalue_reference<_Predicate>::type> (__first, __last, __pred, typename iterator_traits<_ForwardIterator>::iterator_category()); } @@ -3727,7 +3874,7 @@ unsigned __sort4(_ForwardIterator __x1, _ForwardIterator __x2, _ForwardIterator __x3, _ForwardIterator __x4, _Compare __c) { - unsigned __r = __sort3<_Compare>(__x1, __x2, __x3, __c); + unsigned __r = _VSTD::__sort3<_Compare>(__x1, __x2, __x3, __c); if (__c(*__x4, *__x3)) { swap(*__x3, *__x4); @@ -3754,7 +3901,7 @@ unsigned __sort5(_ForwardIterator __x1, _ForwardIterator __x2, _ForwardIterator __x3, _ForwardIterator __x4, _ForwardIterator __x5, _Compare __c) { - unsigned __r = __sort4<_Compare>(__x1, __x2, __x3, __x4, __c); + unsigned __r = _VSTD::__sort4<_Compare>(__x1, __x2, __x3, __x4, __c); if (__c(*__x5, *__x4)) { swap(*__x4, *__x5); @@ -3779,14 +3926,14 @@ __sort5(_ForwardIterator __x1, _ForwardIterator __x2, _ForwardIterator __x3, } // Assumes size > 0 -template <class _Compare, class _BirdirectionalIterator> +template <class _Compare, class _BidirectionalIterator> void -__selection_sort(_BirdirectionalIterator __first, _BirdirectionalIterator __last, _Compare __comp) +__selection_sort(_BidirectionalIterator __first, _BidirectionalIterator __last, _Compare __comp) { - _BirdirectionalIterator __lm1 = __last; + _BidirectionalIterator __lm1 = __last; for (--__lm1; __first != __lm1; ++__first) { - _BirdirectionalIterator __i = _VSTD::min_element<_BirdirectionalIterator, + _BidirectionalIterator __i = _VSTD::min_element<_BidirectionalIterator, typename add_lvalue_reference<_Compare>::type> (__first, __last, __comp); if (__i != __first) @@ -3794,19 +3941,19 @@ __selection_sort(_BirdirectionalIterator __first, _BirdirectionalIterator __last } } -template <class _Compare, class _BirdirectionalIterator> +template <class _Compare, class _BidirectionalIterator> void -__insertion_sort(_BirdirectionalIterator __first, _BirdirectionalIterator __last, _Compare __comp) +__insertion_sort(_BidirectionalIterator __first, _BidirectionalIterator __last, _Compare __comp) { - typedef typename iterator_traits<_BirdirectionalIterator>::value_type value_type; + typedef typename iterator_traits<_BidirectionalIterator>::value_type value_type; if (__first != __last) { - _BirdirectionalIterator __i = __first; + _BidirectionalIterator __i = __first; for (++__i; __i != __last; ++__i) { - _BirdirectionalIterator __j = __i; + _BidirectionalIterator __j = __i; value_type __t(_VSTD::move(*__j)); - for (_BirdirectionalIterator __k = __i; __k != __first && __comp(__t, *--__k); --__j) + for (_BidirectionalIterator __k = __i; __k != __first && __comp(__t, *--__k); --__j) *__j = _VSTD::move(*__k); *__j = _VSTD::move(__t); } @@ -3819,7 +3966,7 @@ __insertion_sort_3(_RandomAccessIterator __first, _RandomAccessIterator __last, { typedef typename iterator_traits<_RandomAccessIterator>::value_type value_type; _RandomAccessIterator __j = __first+2; - __sort3<_Compare>(__first, __first+1, __j, __comp); + _VSTD::__sort3<_Compare>(__first, __first+1, __j, __comp); for (_RandomAccessIterator __i = __j+1; __i != __last; ++__i) { if (__comp(*__i, *__j)) @@ -3863,7 +4010,7 @@ __insertion_sort_incomplete(_RandomAccessIterator __first, _RandomAccessIterator } typedef typename iterator_traits<_RandomAccessIterator>::value_type value_type; _RandomAccessIterator __j = __first+2; - __sort3<_Compare>(__first, __first+1, __j, __comp); + _VSTD::__sort3<_Compare>(__first, __first+1, __j, __comp); const unsigned __limit = 8; unsigned __count = 0; for (_RandomAccessIterator __i = __j+1; __i != __last; ++__i) @@ -3887,35 +4034,35 @@ __insertion_sort_incomplete(_RandomAccessIterator __first, _RandomAccessIterator return true; } -template <class _Compare, class _BirdirectionalIterator> +template <class _Compare, class _BidirectionalIterator> void -__insertion_sort_move(_BirdirectionalIterator __first1, _BirdirectionalIterator __last1, - typename iterator_traits<_BirdirectionalIterator>::value_type* __first2, _Compare __comp) +__insertion_sort_move(_BidirectionalIterator __first1, _BidirectionalIterator __last1, + typename iterator_traits<_BidirectionalIterator>::value_type* __first2, _Compare __comp) { - typedef typename iterator_traits<_BirdirectionalIterator>::value_type value_type; + typedef typename iterator_traits<_BidirectionalIterator>::value_type value_type; if (__first1 != __last1) { __destruct_n __d(0); unique_ptr<value_type, __destruct_n&> __h(__first2, __d); value_type* __last2 = __first2; - ::new(__last2) value_type(_VSTD::move(*__first1)); - __d.__incr((value_type*)0); + ::new ((void*)__last2) value_type(_VSTD::move(*__first1)); + __d.template __incr<value_type>(); for (++__last2; ++__first1 != __last1; ++__last2) { value_type* __j2 = __last2; value_type* __i2 = __j2; if (__comp(*__first1, *--__i2)) { - ::new(__j2) value_type(_VSTD::move(*__i2)); - __d.__incr((value_type*)0); + ::new ((void*)__j2) value_type(_VSTD::move(*__i2)); + __d.template __incr<value_type>(); for (--__j2; __i2 != __first2 && __comp(*__first1, *--__i2); --__j2) *__j2 = _VSTD::move(*__i2); *__j2 = _VSTD::move(*__first1); } else { - ::new(__j2) value_type(_VSTD::move(*__first1)); - __d.__incr((value_type*)0); + ::new ((void*)__j2) value_type(_VSTD::move(*__first1)); + __d.template __incr<value_type>(); } } __h.release(); @@ -4032,7 +4179,7 @@ __sort(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compare __c ++__i; } // [__first, __i) == *__first and *__first < [__i, __last) - // The first part is sorted, sort the secod part + // The first part is sorted, sort the second part // _VSTD::__sort<_Compare>(__i, __last, __comp); __first = __i; goto __restart; @@ -4138,7 +4285,7 @@ inline _LIBCPP_INLINE_VISIBILITY void sort(_Tp** __first, _Tp** __last) { - _VSTD::sort((size_t*)__first, (size_t*)__last, __less<size_t>()); + _VSTD::sort((uintptr_t*)__first, (uintptr_t*)__last, __less<uintptr_t>()); } template <class _Tp> @@ -4223,7 +4370,7 @@ _ForwardIterator lower_bound(_ForwardIterator __first, _ForwardIterator __last, const _Tp& __value_, _Compare __comp) { typedef typename add_lvalue_reference<_Compare>::type _Comp_ref; - return __lower_bound<_Comp_ref>(__first, __last, __value_, __comp); + return _VSTD::__lower_bound<_Comp_ref>(__first, __last, __value_, __comp); } template <class _ForwardIterator, class _Tp> @@ -4267,7 +4414,7 @@ _ForwardIterator upper_bound(_ForwardIterator __first, _ForwardIterator __last, const _Tp& __value_, _Compare __comp) { typedef typename add_lvalue_reference<_Compare>::type _Comp_ref; - return __upper_bound<_Comp_ref>(__first, __last, __value_, __comp); + return _VSTD::__upper_bound<_Comp_ref>(__first, __last, __value_, __comp); } template <class _ForwardIterator, class _Tp> @@ -4308,8 +4455,8 @@ __equal_range(_ForwardIterator __first, _ForwardIterator __last, const _Tp& __va _ForwardIterator __mp1 = __m; return pair<_ForwardIterator, _ForwardIterator> ( - __lower_bound<_Compare>(__first, __m, __value_, __comp), - __upper_bound<_Compare>(++__mp1, __last, __value_, __comp) + _VSTD::__lower_bound<_Compare>(__first, __m, __value_, __comp), + _VSTD::__upper_bound<_Compare>(++__mp1, __last, __value_, __comp) ); } } @@ -4323,7 +4470,7 @@ pair<_ForwardIterator, _ForwardIterator> equal_range(_ForwardIterator __first, _ForwardIterator __last, const _Tp& __value_, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __equal_range<_Comp_ref>(__first, __last, __value_, __comp); + return _VSTD::__equal_range<_Comp_ref>(__first, __last, __value_, __comp); } template <class _ForwardIterator, class _Tp> @@ -4343,7 +4490,7 @@ inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 bool __binary_search(_ForwardIterator __first, _ForwardIterator __last, const _Tp& __value_, _Compare __comp) { - __first = __lower_bound<_Compare>(__first, __last, __value_, __comp); + __first = _VSTD::__lower_bound<_Compare>(__first, __last, __value_, __comp); return __first != __last && !__comp(__value_, *__first); } @@ -4354,7 +4501,7 @@ bool binary_search(_ForwardIterator __first, _ForwardIterator __last, const _Tp& __value_, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __binary_search<_Comp_ref>(__first, __last, __value_, __comp); + return _VSTD::__binary_search<_Comp_ref>(__first, __last, __value_, __comp); } template <class _ForwardIterator, class _Tp> @@ -4370,6 +4517,7 @@ binary_search(_ForwardIterator __first, _ForwardIterator __last, const _Tp& __va // merge template <class _Compare, class _InputIterator1, class _InputIterator2, class _OutputIterator> +_LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator __merge(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) @@ -4393,7 +4541,7 @@ __merge(_InputIterator1 __first1, _InputIterator1 __last1, } template <class _InputIterator1, class _InputIterator2, class _OutputIterator, class _Compare> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator merge(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) @@ -4403,7 +4551,7 @@ merge(_InputIterator1 __first1, _InputIterator1 __last1, } template <class _InputIterator1, class _InputIterator2, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator merge(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result) @@ -4456,20 +4604,21 @@ __buffered_inplace_merge(_BidirectionalIterator __first, _BidirectionalIterator if (__len1 <= __len2) { value_type* __p = __buff; - for (_BidirectionalIterator __i = __first; __i != __middle; __d.__incr((value_type*)0), (void) ++__i, (void) ++__p) - ::new(__p) value_type(_VSTD::move(*__i)); - __half_inplace_merge(__buff, __p, __middle, __last, __first, __comp); + for (_BidirectionalIterator __i = __first; __i != __middle; __d.template __incr<value_type>(), (void) ++__i, (void) ++__p) + ::new ((void*)__p) value_type(_VSTD::move(*__i)); + _VSTD::__half_inplace_merge<_Compare>(__buff, __p, __middle, __last, __first, __comp); } else { value_type* __p = __buff; - for (_BidirectionalIterator __i = __middle; __i != __last; __d.__incr((value_type*)0), (void) ++__i, (void) ++__p) - ::new(__p) value_type(_VSTD::move(*__i)); + for (_BidirectionalIterator __i = __middle; __i != __last; __d.template __incr<value_type>(), (void) ++__i, (void) ++__p) + ::new ((void*)__p) value_type(_VSTD::move(*__i)); typedef reverse_iterator<_BidirectionalIterator> _RBi; typedef reverse_iterator<value_type*> _Rv; - __half_inplace_merge(_Rv(__p), _Rv(__buff), - _RBi(__middle), _RBi(__first), - _RBi(__last), __invert<_Compare>(__comp)); + typedef __invert<_Compare> _Inverted; + _VSTD::__half_inplace_merge<_Inverted>(_Rv(__p), _Rv(__buff), + _RBi(__middle), _RBi(__first), + _RBi(__last), _Inverted(__comp)); } } @@ -4487,7 +4636,7 @@ __inplace_merge(_BidirectionalIterator __first, _BidirectionalIterator __middle, if (__len2 == 0) return; if (__len1 <= __buff_size || __len2 <= __buff_size) - return __buffered_inplace_merge<_Compare> + return _VSTD::__buffered_inplace_merge<_Compare> (__first, __middle, __last, __comp, __len1, __len2, __buff); // shrink [__first, __middle) as much as possible (with no moves), returning if it shrinks to 0 for (; true; ++__first, (void) --__len1) @@ -4515,7 +4664,7 @@ __inplace_merge(_BidirectionalIterator __first, _BidirectionalIterator __middle, __len21 = __len2 / 2; __m2 = __middle; _VSTD::advance(__m2, __len21); - __m1 = __upper_bound<_Compare>(__first, __middle, *__m2, __comp); + __m1 = _VSTD::__upper_bound<_Compare>(__first, __middle, *__m2, __comp); __len11 = _VSTD::distance(__first, __m1); } else @@ -4530,7 +4679,7 @@ __inplace_merge(_BidirectionalIterator __first, _BidirectionalIterator __middle, __len11 = __len1 / 2; __m1 = __first; _VSTD::advance(__m1, __len11); - __m2 = __lower_bound<_Compare>(__middle, __last, *__m1, __comp); + __m2 = _VSTD::__lower_bound<_Compare>(__middle, __last, *__m1, __comp); __len21 = _VSTD::distance(__middle, __m2); } difference_type __len12 = __len1 - __len11; // distance(__m1, __middle) @@ -4539,11 +4688,11 @@ __inplace_merge(_BidirectionalIterator __first, _BidirectionalIterator __middle, // swap middle two partitions __middle = _VSTD::rotate(__m1, __middle, __m2); // __len12 and __len21 now have swapped meanings - // merge smaller range with recurisve call and larger with tail recursion elimination + // merge smaller range with recursive call and larger with tail recursion elimination if (__len11 + __len21 < __len12 + __len22) { - __inplace_merge<_Compare>(__first, __m1, __middle, __comp, __len11, __len21, __buff, __buff_size); -// __inplace_merge<_Compare>(__middle, __m2, __last, __comp, __len12, __len22, __buff, __buff_size); + _VSTD::__inplace_merge<_Compare>(__first, __m1, __middle, __comp, __len11, __len21, __buff, __buff_size); +// _VSTD::__inplace_merge<_Compare>(__middle, __m2, __last, __comp, __len12, __len22, __buff, __buff_size); __first = __middle; __middle = __m2; __len1 = __len12; @@ -4551,8 +4700,8 @@ __inplace_merge(_BidirectionalIterator __first, _BidirectionalIterator __middle, } else { - __inplace_merge<_Compare>(__middle, __m2, __last, __comp, __len12, __len22, __buff, __buff_size); -// __inplace_merge<_Compare>(__first, __m1, __middle, __comp, __len11, __len21, __buff, __buff_size); + _VSTD::__inplace_merge<_Compare>(__middle, __m2, __last, __comp, __len12, __len22, __buff, __buff_size); +// _VSTD::__inplace_merge<_Compare>(__first, __m1, __middle, __comp, __len11, __len21, __buff, __buff_size); __last = __middle; __middle = __m1; __len1 = __len11; @@ -4603,28 +4752,28 @@ __merge_move_construct(_InputIterator1 __first1, _InputIterator1 __last1, { if (__first1 == __last1) { - for (; __first2 != __last2; ++__first2, ++__result, (void) __d.__incr((value_type*)0)) - ::new (__result) value_type(_VSTD::move(*__first2)); + for (; __first2 != __last2; ++__first2, ++__result, (void)__d.template __incr<value_type>()) + ::new ((void*)__result) value_type(_VSTD::move(*__first2)); __h.release(); return; } if (__first2 == __last2) { - for (; __first1 != __last1; ++__first1, ++__result, (void) __d.__incr((value_type*)0)) - ::new (__result) value_type(_VSTD::move(*__first1)); + for (; __first1 != __last1; ++__first1, ++__result, (void)__d.template __incr<value_type>()) + ::new ((void*)__result) value_type(_VSTD::move(*__first1)); __h.release(); return; } if (__comp(*__first2, *__first1)) { - ::new (__result) value_type(_VSTD::move(*__first2)); - __d.__incr((value_type*)0); + ::new ((void*)__result) value_type(_VSTD::move(*__first2)); + __d.template __incr<value_type>(); ++__first2; } else { - ::new (__result) value_type(_VSTD::move(*__first1)); - __d.__incr((value_type*)0); + ::new ((void*)__result) value_type(_VSTD::move(*__first1)); + __d.template __incr<value_type>(); ++__first1; } } @@ -4677,38 +4826,38 @@ __stable_sort_move(_RandomAccessIterator __first1, _RandomAccessIterator __last1 case 0: return; case 1: - ::new(__first2) value_type(_VSTD::move(*__first1)); + ::new ((void*)__first2) value_type(_VSTD::move(*__first1)); return; case 2: __destruct_n __d(0); unique_ptr<value_type, __destruct_n&> __h2(__first2, __d); if (__comp(*--__last1, *__first1)) { - ::new(__first2) value_type(_VSTD::move(*__last1)); - __d.__incr((value_type*)0); + ::new ((void*)__first2) value_type(_VSTD::move(*__last1)); + __d.template __incr<value_type>(); ++__first2; - ::new(__first2) value_type(_VSTD::move(*__first1)); + ::new ((void*)__first2) value_type(_VSTD::move(*__first1)); } else { - ::new(__first2) value_type(_VSTD::move(*__first1)); - __d.__incr((value_type*)0); + ::new ((void*)__first2) value_type(_VSTD::move(*__first1)); + __d.template __incr<value_type>(); ++__first2; - ::new(__first2) value_type(_VSTD::move(*__last1)); + ::new ((void*)__first2) value_type(_VSTD::move(*__last1)); } __h2.release(); return; } if (__len <= 8) { - __insertion_sort_move<_Compare>(__first1, __last1, __first2, __comp); + _VSTD::__insertion_sort_move<_Compare>(__first1, __last1, __first2, __comp); return; } typename iterator_traits<_RandomAccessIterator>::difference_type __l2 = __len / 2; _RandomAccessIterator __m = __first1 + __l2; - __stable_sort<_Compare>(__first1, __m, __comp, __l2, __first2, __l2); - __stable_sort<_Compare>(__m, __last1, __comp, __len - __l2, __first2 + __l2, __len - __l2); - __merge_move_construct<_Compare>(__first1, __m, __m, __last1, __first2, __comp); + _VSTD::__stable_sort<_Compare>(__first1, __m, __comp, __l2, __first2, __l2); + _VSTD::__stable_sort<_Compare>(__m, __last1, __comp, __len - __l2, __first2 + __l2, __len - __l2); + _VSTD::__merge_move_construct<_Compare>(__first1, __m, __m, __last1, __first2, __comp); } template <class _Tp> @@ -4737,7 +4886,7 @@ __stable_sort(_RandomAccessIterator __first, _RandomAccessIterator __last, _Comp } if (__len <= static_cast<difference_type>(__stable_sort_switch<value_type>::value)) { - __insertion_sort<_Compare>(__first, __last, __comp); + _VSTD::__insertion_sort<_Compare>(__first, __last, __comp); return; } typename iterator_traits<_RandomAccessIterator>::difference_type __l2 = __len / 2; @@ -4746,21 +4895,21 @@ __stable_sort(_RandomAccessIterator __first, _RandomAccessIterator __last, _Comp { __destruct_n __d(0); unique_ptr<value_type, __destruct_n&> __h2(__buff, __d); - __stable_sort_move<_Compare>(__first, __m, __comp, __l2, __buff); - __d.__set(__l2, (value_type*)0); - __stable_sort_move<_Compare>(__m, __last, __comp, __len - __l2, __buff + __l2); - __d.__set(__len, (value_type*)0); - __merge_move_assign<_Compare>(__buff, __buff + __l2, __buff + __l2, __buff + __len, __first, __comp); -// __merge<_Compare>(move_iterator<value_type*>(__buff), -// move_iterator<value_type*>(__buff + __l2), -// move_iterator<_RandomAccessIterator>(__buff + __l2), -// move_iterator<_RandomAccessIterator>(__buff + __len), -// __first, __comp); + _VSTD::__stable_sort_move<_Compare>(__first, __m, __comp, __l2, __buff); + __d.__set(__l2, (value_type*)nullptr); + _VSTD::__stable_sort_move<_Compare>(__m, __last, __comp, __len - __l2, __buff + __l2); + __d.__set(__len, (value_type*)nullptr); + _VSTD::__merge_move_assign<_Compare>(__buff, __buff + __l2, __buff + __l2, __buff + __len, __first, __comp); +// _VSTD::__merge<_Compare>(move_iterator<value_type*>(__buff), +// move_iterator<value_type*>(__buff + __l2), +// move_iterator<_RandomAccessIterator>(__buff + __l2), +// move_iterator<_RandomAccessIterator>(__buff + __len), +// __first, __comp); return; } - __stable_sort<_Compare>(__first, __m, __comp, __l2, __buff, __buff_size); - __stable_sort<_Compare>(__m, __last, __comp, __len - __l2, __buff, __buff_size); - __inplace_merge<_Compare>(__first, __m, __last, __comp, __l2, __len - __l2, __buff, __buff_size); + _VSTD::__stable_sort<_Compare>(__first, __m, __comp, __l2, __buff, __buff_size); + _VSTD::__stable_sort<_Compare>(__m, __last, __comp, __len - __l2, __buff, __buff_size); + _VSTD::__inplace_merge<_Compare>(__first, __m, __last, __comp, __l2, __len - __l2, __buff, __buff_size); } template <class _RandomAccessIterator, class _Compare> @@ -4779,7 +4928,7 @@ stable_sort(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compar __h.reset(__buf.first); } typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - __stable_sort<_Comp_ref>(__first, __last, __comp, __len, __buf.first, __buf.second); + _VSTD::__stable_sort<_Comp_ref>(__first, __last, __comp, __len, __buf.first, __buf.second); } template <class _RandomAccessIterator> @@ -4883,7 +5032,7 @@ void push_heap(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - __sift_up<_Comp_ref>(__first, __last, __comp, __last - __first); + _VSTD::__sift_up<_Comp_ref>(__first, __last, __comp, __last - __first); } template <class _RandomAccessIterator> @@ -4960,7 +5109,7 @@ __pop_heap(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compare if (__len > 1) { swap(*__first, *--__last); - __sift_down<_Compare>(__first, __last, __comp, __len - 1, __first); + _VSTD::__sift_down<_Compare>(__first, __last, __comp, __len - 1, __first); } } @@ -4970,7 +5119,7 @@ void pop_heap(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - __pop_heap<_Comp_ref>(__first, __last, __comp, __last - __first); + _VSTD::__pop_heap<_Comp_ref>(__first, __last, __comp, __last - __first); } template <class _RandomAccessIterator> @@ -4994,7 +5143,7 @@ __make_heap(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compar // start from the first parent, there is no need to consider children for (difference_type __start = (__n - 2) / 2; __start >= 0; --__start) { - __sift_down<_Compare>(__first, __last, __comp, __n, __first + __start); + _VSTD::__sift_down<_Compare>(__first, __last, __comp, __n, __first + __start); } } } @@ -5005,7 +5154,7 @@ void make_heap(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - __make_heap<_Comp_ref>(__first, __last, __comp); + _VSTD::__make_heap<_Comp_ref>(__first, __last, __comp); } template <class _RandomAccessIterator> @@ -5024,7 +5173,7 @@ __sort_heap(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compar { typedef typename iterator_traits<_RandomAccessIterator>::difference_type difference_type; for (difference_type __n = __last - __first; __n > 1; --__last, (void) --__n) - __pop_heap<_Compare>(__first, __last, __comp, __n); + _VSTD::__pop_heap<_Compare>(__first, __last, __comp, __n); } template <class _RandomAccessIterator, class _Compare> @@ -5033,7 +5182,7 @@ void sort_heap(_RandomAccessIterator __first, _RandomAccessIterator __last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - __sort_heap<_Comp_ref>(__first, __last, __comp); + _VSTD::__sort_heap<_Comp_ref>(__first, __last, __comp); } template <class _RandomAccessIterator> @@ -5051,17 +5200,17 @@ void __partial_sort(_RandomAccessIterator __first, _RandomAccessIterator __middle, _RandomAccessIterator __last, _Compare __comp) { - __make_heap<_Compare>(__first, __middle, __comp); + _VSTD::__make_heap<_Compare>(__first, __middle, __comp); typename iterator_traits<_RandomAccessIterator>::difference_type __len = __middle - __first; for (_RandomAccessIterator __i = __middle; __i != __last; ++__i) { if (__comp(*__i, *__first)) { swap(*__i, *__first); - __sift_down<_Compare>(__first, __middle, __comp, __len, __first); + _VSTD::__sift_down<_Compare>(__first, __middle, __comp, __len, __first); } } - __sort_heap<_Compare>(__first, __middle, __comp); + _VSTD::__sort_heap<_Compare>(__first, __middle, __comp); } template <class _RandomAccessIterator, class _Compare> @@ -5071,7 +5220,7 @@ partial_sort(_RandomAccessIterator __first, _RandomAccessIterator __middle, _Ran _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - __partial_sort<_Comp_ref>(__first, __middle, __last, __comp); + _VSTD::__partial_sort<_Comp_ref>(__first, __middle, __last, __comp); } template <class _RandomAccessIterator> @@ -5095,15 +5244,15 @@ __partial_sort_copy(_InputIterator __first, _InputIterator __last, { for (; __first != __last && __r != __result_last; ++__first, (void) ++__r) *__r = *__first; - __make_heap<_Compare>(__result_first, __r, __comp); + _VSTD::__make_heap<_Compare>(__result_first, __r, __comp); typename iterator_traits<_RandomAccessIterator>::difference_type __len = __r - __result_first; for (; __first != __last; ++__first) if (__comp(*__first, *__result_first)) { *__result_first = *__first; - __sift_down<_Compare>(__result_first, __r, __comp, __len, __result_first); + _VSTD::__sift_down<_Compare>(__result_first, __r, __comp, __len, __result_first); } - __sort_heap<_Compare>(__result_first, __r, __comp); + _VSTD::__sort_heap<_Compare>(__result_first, __r, __comp); } return __r; } @@ -5115,7 +5264,7 @@ partial_sort_copy(_InputIterator __first, _InputIterator __last, _RandomAccessIterator __result_first, _RandomAccessIterator __result_last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __partial_sort_copy<_Comp_ref>(__first, __last, __result_first, __result_last, __comp); + return _VSTD::__partial_sort_copy<_Comp_ref>(__first, __last, __result_first, __result_last, __comp); } template <class _InputIterator, class _RandomAccessIterator> @@ -5161,7 +5310,7 @@ __nth_element(_RandomAccessIterator __first, _RandomAccessIterator __nth, _Rando } if (__len <= __limit) { - __selection_sort<_Compare>(__first, __last, __comp); + _VSTD::__selection_sort<_Compare>(__first, __last, __comp); return; } // __len > __limit >= 3 @@ -5185,7 +5334,7 @@ __nth_element(_RandomAccessIterator __first, _RandomAccessIterator __nth, _Rando if (__i == --__j) { // *__first == *__m, *__m <= all other elements - // Parition instead into [__first, __i) == *__first and *__first < [__i, __last) + // Partition instead into [__first, __i) == *__first and *__first < [__i, __last) ++__i; // __first + 1 __j = __last; if (!__comp(*__first, *--__j)) // we need a guard if *__first == *(__last-1) @@ -5223,8 +5372,8 @@ __nth_element(_RandomAccessIterator __first, _RandomAccessIterator __nth, _Rando // The first part is sorted, if (__nth < __i) return; - // __nth_element the secod part - // __nth_element<_Compare>(__i, __nth, __last, __comp); + // __nth_element the second part + // _VSTD::__nth_element<_Compare>(__i, __nth, __last, __comp); __first = __i; goto __restart; } @@ -5306,12 +5455,12 @@ not_sorted: // __nth_element on range containing __nth if (__nth < __i) { - // __nth_element<_Compare>(__first, __nth, __i, __comp); + // _VSTD::__nth_element<_Compare>(__first, __nth, __i, __comp); __last = __i; } else { - // __nth_element<_Compare>(__i+1, __nth, __last, __comp); + // _VSTD::__nth_element<_Compare>(__i+1, __nth, __last, __comp); __first = ++__i; } } @@ -5323,7 +5472,7 @@ void nth_element(_RandomAccessIterator __first, _RandomAccessIterator __nth, _RandomAccessIterator __last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - __nth_element<_Comp_ref>(__first, __nth, __last, __comp); + _VSTD::__nth_element<_Comp_ref>(__first, __nth, __last, __comp); } template <class _RandomAccessIterator> @@ -5359,7 +5508,7 @@ includes(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __fi _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __includes<_Comp_ref>(__first1, __last1, __first2, __last2, __comp); + return _VSTD::__includes<_Comp_ref>(__first1, __last1, __first2, __last2, __comp); } template <class _InputIterator1, class _InputIterator2> @@ -5376,7 +5525,7 @@ includes(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __fi // set_union template <class _Compare, class _InputIterator1, class _InputIterator2, class _OutputIterator> -_OutputIterator +_LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator __set_union(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) { @@ -5401,17 +5550,17 @@ __set_union(_InputIterator1 __first1, _InputIterator1 __last1, } template <class _InputIterator1, class _InputIterator2, class _OutputIterator, class _Compare> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator set_union(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __set_union<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); + return _VSTD::__set_union<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); } template <class _InputIterator1, class _InputIterator2, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator set_union(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result) @@ -5453,7 +5602,7 @@ set_intersection(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __set_intersection<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); + return _VSTD::__set_intersection<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); } template <class _InputIterator1, class _InputIterator2, class _OutputIterator> @@ -5470,7 +5619,7 @@ set_intersection(_InputIterator1 __first1, _InputIterator1 __last1, // set_difference template <class _Compare, class _InputIterator1, class _InputIterator2, class _OutputIterator> -_OutputIterator +_LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator __set_difference(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) { @@ -5495,17 +5644,17 @@ __set_difference(_InputIterator1 __first1, _InputIterator1 __last1, } template <class _InputIterator1, class _InputIterator2, class _OutputIterator, class _Compare> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator set_difference(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __set_difference<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); + return _VSTD::__set_difference<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); } template <class _InputIterator1, class _InputIterator2, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator set_difference(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result) @@ -5518,7 +5667,7 @@ set_difference(_InputIterator1 __first1, _InputIterator1 __last1, // set_symmetric_difference template <class _Compare, class _InputIterator1, class _InputIterator2, class _OutputIterator> -_OutputIterator +_LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator __set_symmetric_difference(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) { @@ -5548,17 +5697,17 @@ __set_symmetric_difference(_InputIterator1 __first1, _InputIterator1 __last1, } template <class _InputIterator1, class _InputIterator2, class _OutputIterator, class _Compare> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator set_symmetric_difference(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __set_symmetric_difference<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); + return _VSTD::__set_symmetric_difference<_Comp_ref>(__first1, __last1, __first2, __last2, __result, __comp); } template <class _InputIterator1, class _InputIterator2, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator set_symmetric_difference(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _OutputIterator __result) @@ -5593,7 +5742,7 @@ lexicographical_compare(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _InputIterator2 __last2, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __lexicographical_compare<_Comp_ref>(__first1, __last1, __first2, __last2, __comp); + return _VSTD::__lexicographical_compare<_Comp_ref>(__first1, __last1, __first2, __last2, __comp); } template <class _InputIterator1, class _InputIterator2> @@ -5611,7 +5760,7 @@ lexicographical_compare(_InputIterator1 __first1, _InputIterator1 __last1, // next_permutation template <class _Compare, class _BidirectionalIterator> -bool +_LIBCPP_CONSTEXPR_AFTER_CXX17 bool __next_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last, _Compare __comp) { _BidirectionalIterator __i = __last; @@ -5638,16 +5787,16 @@ __next_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last } template <class _BidirectionalIterator, class _Compare> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 bool next_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __next_permutation<_Comp_ref>(__first, __last, __comp); + return _VSTD::__next_permutation<_Comp_ref>(__first, __last, __comp); } template <class _BidirectionalIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 bool next_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last) { @@ -5658,7 +5807,7 @@ next_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last) // prev_permutation template <class _Compare, class _BidirectionalIterator> -bool +_LIBCPP_CONSTEXPR_AFTER_CXX17 bool __prev_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last, _Compare __comp) { _BidirectionalIterator __i = __last; @@ -5685,16 +5834,16 @@ __prev_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last } template <class _BidirectionalIterator, class _Compare> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 bool prev_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last, _Compare __comp) { typedef typename __comp_ref_type<_Compare>::type _Comp_ref; - return __prev_permutation<_Comp_ref>(__first, __last, __comp); + return _VSTD::__prev_permutation<_Comp_ref>(__first, __last, __comp); } template <class _BidirectionalIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 bool prev_permutation(_BidirectionalIterator __first, _BidirectionalIterator __last) { diff --git a/lib/libcxx/include/any b/lib/libcxx/include/any index 36b07c9d7e..968c9769ee 100644 --- a/lib/libcxx/include/any +++ b/lib/libcxx/include/any @@ -81,8 +81,8 @@ namespace std { */ #include <experimental/__config> +#include <__availability> #include <memory> -#include <new> #include <typeinfo> #include <type_traits> #include <cstdlib> @@ -157,7 +157,7 @@ namespace __any_imp template <class _Tp> inline _LIBCPP_INLINE_VISIBILITY constexpr const void* __get_fallback_typeid() { - return &__unique_typeinfo<decay_t<_Tp>>::__id; + return &__unique_typeinfo<remove_cv_t<remove_reference_t<_Tp>>>::__id; } template <class _Tp> @@ -368,7 +368,11 @@ namespace __any_imp template <class ..._Args> _LIBCPP_INLINE_VISIBILITY static _Tp& __create(any & __dest, _Args&&... __args) { - _Tp* __ret = ::new (static_cast<void*>(&__dest.__s.__buf)) _Tp(_VSTD::forward<_Args>(__args)...); + typedef allocator<_Tp> _Alloc; + typedef allocator_traits<_Alloc> _ATraits; + _Alloc __a; + _Tp * __ret = static_cast<_Tp*>(static_cast<void*>(&__dest.__s.__buf)); + _ATraits::construct(__a, __ret, _VSTD::forward<_Args>(__args)...); __dest.__h = &_SmallHandler::__handle; return *__ret; } @@ -376,8 +380,11 @@ namespace __any_imp private: _LIBCPP_INLINE_VISIBILITY static void __destroy(any & __this) { - _Tp & __value = *static_cast<_Tp *>(static_cast<void*>(&__this.__s.__buf)); - __value.~_Tp(); + typedef allocator<_Tp> _Alloc; + typedef allocator_traits<_Alloc> _ATraits; + _Alloc __a; + _Tp * __p = static_cast<_Tp *>(static_cast<void*>(&__this.__s.__buf)); + _ATraits::destroy(__a, __p); __this.__h = nullptr; } @@ -445,10 +452,12 @@ namespace __any_imp _LIBCPP_INLINE_VISIBILITY static _Tp& __create(any & __dest, _Args&&... __args) { typedef allocator<_Tp> _Alloc; + typedef allocator_traits<_Alloc> _ATraits; typedef __allocator_destructor<_Alloc> _Dp; _Alloc __a; - unique_ptr<_Tp, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - _Tp* __ret = ::new ((void*)__hold.get()) _Tp(_VSTD::forward<_Args>(__args)...); + unique_ptr<_Tp, _Dp> __hold(_ATraits::allocate(__a, 1), _Dp(__a, 1)); + _Tp * __ret = __hold.get(); + _ATraits::construct(__a, __ret, _VSTD::forward<_Args>(__args)...); __dest.__s.__ptr = __hold.release(); __dest.__h = &_LargeHandler::__handle; return *__ret; @@ -458,7 +467,12 @@ namespace __any_imp _LIBCPP_INLINE_VISIBILITY static void __destroy(any & __this){ - delete static_cast<_Tp*>(__this.__s.__ptr); + typedef allocator<_Tp> _Alloc; + typedef allocator_traits<_Alloc> _ATraits; + _Alloc __a; + _Tp * __p = static_cast<_Tp *>(__this.__s.__ptr); + _ATraits::destroy(__a, __p); + _ATraits::deallocate(__a, __p, 1); __this.__h = nullptr; } diff --git a/lib/libcxx/include/array b/lib/libcxx/include/array index e73bbe7fea..9a479f7cd1 100644 --- a/lib/libcxx/include/array +++ b/lib/libcxx/include/array @@ -142,8 +142,8 @@ struct _LIBCPP_TEMPLATE_VIS array typedef const value_type* const_pointer; typedef size_t size_type; typedef ptrdiff_t difference_type; - typedef std::reverse_iterator<iterator> reverse_iterator; - typedef std::reverse_iterator<const_iterator> const_reverse_iterator; + typedef _VSTD::reverse_iterator<iterator> reverse_iterator; + typedef _VSTD::reverse_iterator<const_iterator> const_reverse_iterator; _Tp __elems_[_Size]; @@ -155,7 +155,7 @@ struct _LIBCPP_TEMPLATE_VIS array _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 void swap(array& __a) _NOEXCEPT_(__is_nothrow_swappable<_Tp>::value) { - std::swap_ranges(data(), data() + _Size, __a.data()); + _VSTD::swap_ranges(data(), data() + _Size, __a.data()); } // iterators: @@ -245,8 +245,8 @@ struct _LIBCPP_TEMPLATE_VIS array<_Tp, 0> typedef const value_type* const_pointer; typedef size_t size_type; typedef ptrdiff_t difference_type; - typedef std::reverse_iterator<iterator> reverse_iterator; - typedef std::reverse_iterator<const_iterator> const_reverse_iterator; + typedef _VSTD::reverse_iterator<iterator> reverse_iterator; + typedef _VSTD::reverse_iterator<const_iterator> const_reverse_iterator; typedef typename conditional<is_const<_Tp>::value, const char, char>::type _CharType; @@ -459,8 +459,6 @@ get(const array<_Tp, _Size>& __a) _NOEXCEPT return __a.__elems_[_Ip]; } -#ifndef _LIBCPP_CXX03_LANG - template <size_t _Ip, class _Tp, size_t _Size> inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX11 _Tp&& @@ -479,8 +477,6 @@ get(const array<_Tp, _Size>&& __a) _NOEXCEPT return _VSTD::move(__a.__elems_[_Ip]); } -#endif // !_LIBCPP_CXX03_LANG - #if _LIBCPP_STD_VER > 17 template <typename _Tp, size_t _Size, size_t... _Index> @@ -504,7 +500,7 @@ to_array(_Tp (&__arr)[_Size]) noexcept(is_nothrow_constructible_v<_Tp, _Tp&>) { static_assert( is_constructible_v<_Tp, _Tp&>, "[array.creation]/1: to_array requires copy constructible elements."); - return __to_array_lvalue_impl(__arr, make_index_sequence<_Size>()); + return _VSTD::__to_array_lvalue_impl(__arr, make_index_sequence<_Size>()); } template <typename _Tp, size_t _Size> @@ -516,8 +512,8 @@ to_array(_Tp(&&__arr)[_Size]) noexcept(is_nothrow_move_constructible_v<_Tp>) { static_assert( is_move_constructible_v<_Tp>, "[array.creation]/4: to_array requires move constructible elements."); - return __to_array_rvalue_impl(_VSTD::move(__arr), - make_index_sequence<_Size>()); + return _VSTD::__to_array_rvalue_impl(_VSTD::move(__arr), + make_index_sequence<_Size>()); } #endif // _LIBCPP_STD_VER > 17 diff --git a/lib/libcxx/include/atomic b/lib/libcxx/include/atomic index 9c28986537..0fc799a243 100644 --- a/lib/libcxx/include/atomic +++ b/lib/libcxx/include/atomic @@ -16,9 +16,12 @@ namespace std { -// feature test macro +// feature test macro [version.syn] -#define __cpp_lib_atomic_is_always_lock_free // as specified by SG10 +#define __cpp_lib_atomic_is_always_lock_free +#define __cpp_lib_atomic_flag_test +#define __cpp_lib_atomic_lock_free_type_aliases +#define __cpp_lib_atomic_wait // order and consistency @@ -45,6 +48,7 @@ template <class T> T kill_dependency(T y) noexcept; #define ATOMIC_BOOL_LOCK_FREE unspecified #define ATOMIC_CHAR_LOCK_FREE unspecified +#define ATOMIC_CHAR8_T_LOCK_FREE unspecified // C++20 #define ATOMIC_CHAR16_T_LOCK_FREE unspecified #define ATOMIC_CHAR32_T_LOCK_FREE unspecified #define ATOMIC_WCHAR_T_LOCK_FREE unspecified @@ -108,6 +112,7 @@ template <> struct atomic<integral> { using value_type = integral; + using difference_type = value_type; static constexpr bool is_always_lock_free; bool is_lock_free() const volatile noexcept; @@ -190,6 +195,7 @@ template <class T> struct atomic<T*> { using value_type = T*; + using difference_type = ptrdiff_t; static constexpr bool is_always_lock_free; bool is_lock_free() const volatile noexcept; @@ -460,6 +466,7 @@ typedef atomic<long> atomic_long; typedef atomic<unsigned long> atomic_ulong; typedef atomic<long long> atomic_llong; typedef atomic<unsigned long long> atomic_ullong; +typedef atomic<char8_t> atomic_char8_t; // C++20 typedef atomic<char16_t> atomic_char16_t; typedef atomic<char32_t> atomic_char32_t; typedef atomic<wchar_t> atomic_wchar_t; @@ -477,7 +484,7 @@ typedef atomic<int_fast8_t> atomic_int_fast8_t; typedef atomic<uint_fast8_t> atomic_uint_fast8_t; typedef atomic<int_fast16_t> atomic_int_fast16_t; typedef atomic<uint_fast16_t> atomic_uint_fast16_t; -typedef atomic<int_fast32_t> atomic_int_fast32_t; +typedef atomic<int_fast32_t> atomic_int_fast32_t; typedef atomic<uint_fast32_t> atomic_uint_fast32_t; typedef atomic<int_fast64_t> atomic_int_fast64_t; typedef atomic<uint_fast64_t> atomic_uint_fast64_t; @@ -568,6 +575,7 @@ template <class T> */ #include <__config> +#include <__availability> #include <__threading_support> #include <cstddef> #include <cstdint> @@ -654,7 +662,7 @@ typedef enum memory_order { template <typename _Tp> _LIBCPP_INLINE_VISIBILITY bool __cxx_nonatomic_compare_equal(_Tp const& __lhs, _Tp const& __rhs) { - return memcmp(&__lhs, &__rhs, sizeof(_Tp)) == 0; + return _VSTD::memcmp(&__lhs, &__rhs, sizeof(_Tp)) == 0; } static_assert((is_same<underlying_type<memory_order>::type, __memory_order_underlying_t>::value), @@ -1119,6 +1127,9 @@ _Tp kill_dependency(_Tp __y) _NOEXCEPT #if defined(__CLANG_ATOMIC_BOOL_LOCK_FREE) # define ATOMIC_BOOL_LOCK_FREE __CLANG_ATOMIC_BOOL_LOCK_FREE # define ATOMIC_CHAR_LOCK_FREE __CLANG_ATOMIC_CHAR_LOCK_FREE +#ifndef _LIBCPP_NO_HAS_CHAR8_T +# define ATOMIC_CHAR8_T_LOCK_FREE __CLANG_ATOMIC_CHAR8_T_LOCK_FREE +#endif # define ATOMIC_CHAR16_T_LOCK_FREE __CLANG_ATOMIC_CHAR16_T_LOCK_FREE # define ATOMIC_CHAR32_T_LOCK_FREE __CLANG_ATOMIC_CHAR32_T_LOCK_FREE # define ATOMIC_WCHAR_T_LOCK_FREE __CLANG_ATOMIC_WCHAR_T_LOCK_FREE @@ -1130,6 +1141,9 @@ _Tp kill_dependency(_Tp __y) _NOEXCEPT #elif defined(__GCC_ATOMIC_BOOL_LOCK_FREE) # define ATOMIC_BOOL_LOCK_FREE __GCC_ATOMIC_BOOL_LOCK_FREE # define ATOMIC_CHAR_LOCK_FREE __GCC_ATOMIC_CHAR_LOCK_FREE +#ifndef _LIBCPP_NO_HAS_CHAR8_T +# define ATOMIC_CHAR8_T_LOCK_FREE __GCC_ATOMIC_CHAR8_T_LOCK_FREE +#endif # define ATOMIC_CHAR16_T_LOCK_FREE __GCC_ATOMIC_CHAR16_T_LOCK_FREE # define ATOMIC_CHAR32_T_LOCK_FREE __GCC_ATOMIC_CHAR32_T_LOCK_FREE # define ATOMIC_WCHAR_T_LOCK_FREE __GCC_ATOMIC_WCHAR_T_LOCK_FREE @@ -1245,10 +1259,10 @@ template <typename _Tp> _LIBCPP_INLINE_VISIBILITY bool __cxx_atomic_compare_exchange_strong(volatile __cxx_atomic_lock_impl<_Tp>* __a, _Tp* __expected, _Tp __value, memory_order, memory_order) { - __a->__lock(); _Tp __temp; + __a->__lock(); __cxx_atomic_assign_volatile(__temp, __a->__a_value); - bool __ret = __temp == *__expected; + bool __ret = (_VSTD::memcmp(&__temp, __expected, sizeof(_Tp)) == 0); if(__ret) __cxx_atomic_assign_volatile(__a->__a_value, __value); else @@ -1261,11 +1275,11 @@ _LIBCPP_INLINE_VISIBILITY bool __cxx_atomic_compare_exchange_strong(__cxx_atomic_lock_impl<_Tp>* __a, _Tp* __expected, _Tp __value, memory_order, memory_order) { __a->__lock(); - bool __ret = __a->__a_value == *__expected; + bool __ret = (_VSTD::memcmp(&__a->__a_value, __expected, sizeof(_Tp)) == 0); if(__ret) - __a->__a_value = __value; + _VSTD::memcpy(&__a->__a_value, &__value, sizeof(_Tp)); else - *__expected = __a->__a_value; + _VSTD::memcpy(__expected, &__a->__a_value, sizeof(_Tp)); __a->__unlock(); return __ret; } @@ -1274,10 +1288,10 @@ template <typename _Tp> _LIBCPP_INLINE_VISIBILITY bool __cxx_atomic_compare_exchange_weak(volatile __cxx_atomic_lock_impl<_Tp>* __a, _Tp* __expected, _Tp __value, memory_order, memory_order) { - __a->__lock(); _Tp __temp; + __a->__lock(); __cxx_atomic_assign_volatile(__temp, __a->__a_value); - bool __ret = __temp == *__expected; + bool __ret = (_VSTD::memcmp(&__temp, __expected, sizeof(_Tp)) == 0); if(__ret) __cxx_atomic_assign_volatile(__a->__a_value, __value); else @@ -1290,11 +1304,11 @@ _LIBCPP_INLINE_VISIBILITY bool __cxx_atomic_compare_exchange_weak(__cxx_atomic_lock_impl<_Tp>* __a, _Tp* __expected, _Tp __value, memory_order, memory_order) { __a->__lock(); - bool __ret = __a->__a_value == *__expected; + bool __ret = (_VSTD::memcmp(&__a->__a_value, __expected, sizeof(_Tp)) == 0); if(__ret) - __a->__a_value = __value; + _VSTD::memcpy(&__a->__a_value, &__value, sizeof(_Tp)); else - *__expected = __a->__a_value; + _VSTD::memcpy(__expected, &__a->__a_value, sizeof(_Tp)); __a->__unlock(); return __ret; } @@ -1444,6 +1458,9 @@ template<> struct __cxx_is_always_lock_free<bool> { enum { __value = 2 == ATOMIC template<> struct __cxx_is_always_lock_free<char> { enum { __value = 2 == ATOMIC_CHAR_LOCK_FREE }; }; template<> struct __cxx_is_always_lock_free<signed char> { enum { __value = 2 == ATOMIC_CHAR_LOCK_FREE }; }; template<> struct __cxx_is_always_lock_free<unsigned char> { enum { __value = 2 == ATOMIC_CHAR_LOCK_FREE }; }; +#ifndef _LIBCPP_NO_HAS_CHAR8_T +template<> struct __cxx_is_always_lock_free<char8_t> { enum { __value = 2 == ATOMIC_CHAR8_T_LOCK_FREE }; }; +#endif template<> struct __cxx_is_always_lock_free<char16_t> { enum { __value = 2 == ATOMIC_CHAR16_T_LOCK_FREE }; }; template<> struct __cxx_is_always_lock_free<char32_t> { enum { __value = 2 == ATOMIC_CHAR32_T_LOCK_FREE }; }; template<> struct __cxx_is_always_lock_free<wchar_t> { enum { __value = 2 == ATOMIC_WCHAR_T_LOCK_FREE }; }; @@ -1486,8 +1503,6 @@ struct __cxx_atomic_impl : public _Base { using __cxx_contention_t = int64_t; #endif //__linux__ -#if _LIBCPP_STD_VER >= 11 - using __cxx_atomic_contention_t = __cxx_atomic_impl<__cxx_contention_t>; #ifndef _LIBCPP_HAS_NO_PLATFORM_WAIT @@ -1519,7 +1534,7 @@ struct __libcpp_atomic_wait_backoff_impl { else if(__elapsed > chrono::microseconds(4)) __libcpp_thread_yield(); else - ; // poll + {} // poll return false; } }; @@ -1565,8 +1580,6 @@ _LIBCPP_INLINE_VISIBILITY bool __cxx_atomic_wait(_Atp* __a, _Tp const __val, mem return __cxx_atomic_wait(__a, __test_fn); } -#endif //_LIBCPP_STD_VER >= 11 - // general atomic<T> template <class _Tp, bool = is_integral<_Tp>::value && !is_same<_Tp, bool>::value> @@ -1775,6 +1788,7 @@ struct atomic { typedef __atomic_base<_Tp> __base; typedef _Tp value_type; + typedef value_type difference_type; _LIBCPP_INLINE_VISIBILITY atomic() _NOEXCEPT _LIBCPP_DEFAULT _LIBCPP_INLINE_VISIBILITY @@ -1796,6 +1810,7 @@ struct atomic<_Tp*> { typedef __atomic_base<_Tp*> __base; typedef _Tp* value_type; + typedef ptrdiff_t difference_type; _LIBCPP_INLINE_VISIBILITY atomic() _NOEXCEPT _LIBCPP_DEFAULT _LIBCPP_INLINE_VISIBILITY @@ -1872,7 +1887,7 @@ atomic_is_lock_free(const atomic<_Tp>* __o) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY void -atomic_init(volatile atomic<_Tp>* __o, _Tp __d) _NOEXCEPT +atomic_init(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d) _NOEXCEPT { __cxx_atomic_init(&__o->__a_, __d); } @@ -1880,7 +1895,7 @@ atomic_init(volatile atomic<_Tp>* __o, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY void -atomic_init(atomic<_Tp>* __o, _Tp __d) _NOEXCEPT +atomic_init(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d) _NOEXCEPT { __cxx_atomic_init(&__o->__a_, __d); } @@ -1890,7 +1905,7 @@ atomic_init(atomic<_Tp>* __o, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY void -atomic_store(volatile atomic<_Tp>* __o, _Tp __d) _NOEXCEPT +atomic_store(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d) _NOEXCEPT { __o->store(__d); } @@ -1898,7 +1913,7 @@ atomic_store(volatile atomic<_Tp>* __o, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY void -atomic_store(atomic<_Tp>* __o, _Tp __d) _NOEXCEPT +atomic_store(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d) _NOEXCEPT { __o->store(__d); } @@ -1908,7 +1923,7 @@ atomic_store(atomic<_Tp>* __o, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY void -atomic_store_explicit(volatile atomic<_Tp>* __o, _Tp __d, memory_order __m) _NOEXCEPT +atomic_store_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d, memory_order __m) _NOEXCEPT _LIBCPP_CHECK_STORE_MEMORY_ORDER(__m) { __o->store(__d, __m); @@ -1917,7 +1932,7 @@ atomic_store_explicit(volatile atomic<_Tp>* __o, _Tp __d, memory_order __m) _NOE template <class _Tp> _LIBCPP_INLINE_VISIBILITY void -atomic_store_explicit(atomic<_Tp>* __o, _Tp __d, memory_order __m) _NOEXCEPT +atomic_store_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d, memory_order __m) _NOEXCEPT _LIBCPP_CHECK_STORE_MEMORY_ORDER(__m) { __o->store(__d, __m); @@ -1966,7 +1981,7 @@ atomic_load_explicit(const atomic<_Tp>* __o, memory_order __m) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp -atomic_exchange(volatile atomic<_Tp>* __o, _Tp __d) _NOEXCEPT +atomic_exchange(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d) _NOEXCEPT { return __o->exchange(__d); } @@ -1974,7 +1989,7 @@ atomic_exchange(volatile atomic<_Tp>* __o, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp -atomic_exchange(atomic<_Tp>* __o, _Tp __d) _NOEXCEPT +atomic_exchange(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d) _NOEXCEPT { return __o->exchange(__d); } @@ -1984,7 +1999,7 @@ atomic_exchange(atomic<_Tp>* __o, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp -atomic_exchange_explicit(volatile atomic<_Tp>* __o, _Tp __d, memory_order __m) _NOEXCEPT +atomic_exchange_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d, memory_order __m) _NOEXCEPT { return __o->exchange(__d, __m); } @@ -1992,7 +2007,7 @@ atomic_exchange_explicit(volatile atomic<_Tp>* __o, _Tp __d, memory_order __m) _ template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp -atomic_exchange_explicit(atomic<_Tp>* __o, _Tp __d, memory_order __m) _NOEXCEPT +atomic_exchange_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __d, memory_order __m) _NOEXCEPT { return __o->exchange(__d, __m); } @@ -2002,7 +2017,7 @@ atomic_exchange_explicit(atomic<_Tp>* __o, _Tp __d, memory_order __m) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool -atomic_compare_exchange_weak(volatile atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NOEXCEPT +atomic_compare_exchange_weak(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type* __e, typename atomic<_Tp>::value_type __d) _NOEXCEPT { return __o->compare_exchange_weak(*__e, __d); } @@ -2010,7 +2025,7 @@ atomic_compare_exchange_weak(volatile atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NOEX template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool -atomic_compare_exchange_weak(atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NOEXCEPT +atomic_compare_exchange_weak(atomic<_Tp>* __o, typename atomic<_Tp>::value_type* __e, typename atomic<_Tp>::value_type __d) _NOEXCEPT { return __o->compare_exchange_weak(*__e, __d); } @@ -2020,7 +2035,7 @@ atomic_compare_exchange_weak(atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool -atomic_compare_exchange_strong(volatile atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NOEXCEPT +atomic_compare_exchange_strong(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type* __e, typename atomic<_Tp>::value_type __d) _NOEXCEPT { return __o->compare_exchange_strong(*__e, __d); } @@ -2028,7 +2043,7 @@ atomic_compare_exchange_strong(volatile atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NO template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool -atomic_compare_exchange_strong(atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NOEXCEPT +atomic_compare_exchange_strong(atomic<_Tp>* __o, typename atomic<_Tp>::value_type* __e, typename atomic<_Tp>::value_type __d) _NOEXCEPT { return __o->compare_exchange_strong(*__e, __d); } @@ -2038,8 +2053,8 @@ atomic_compare_exchange_strong(atomic<_Tp>* __o, _Tp* __e, _Tp __d) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool -atomic_compare_exchange_weak_explicit(volatile atomic<_Tp>* __o, _Tp* __e, - _Tp __d, +atomic_compare_exchange_weak_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type* __e, + typename atomic<_Tp>::value_type __d, memory_order __s, memory_order __f) _NOEXCEPT _LIBCPP_CHECK_EXCHANGE_MEMORY_ORDER(__s, __f) { @@ -2049,7 +2064,7 @@ atomic_compare_exchange_weak_explicit(volatile atomic<_Tp>* __o, _Tp* __e, template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool -atomic_compare_exchange_weak_explicit(atomic<_Tp>* __o, _Tp* __e, _Tp __d, +atomic_compare_exchange_weak_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::value_type* __e, typename atomic<_Tp>::value_type __d, memory_order __s, memory_order __f) _NOEXCEPT _LIBCPP_CHECK_EXCHANGE_MEMORY_ORDER(__s, __f) { @@ -2062,7 +2077,7 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool atomic_compare_exchange_strong_explicit(volatile atomic<_Tp>* __o, - _Tp* __e, _Tp __d, + typename atomic<_Tp>::value_type* __e, typename atomic<_Tp>::value_type __d, memory_order __s, memory_order __f) _NOEXCEPT _LIBCPP_CHECK_EXCHANGE_MEMORY_ORDER(__s, __f) { @@ -2072,8 +2087,8 @@ atomic_compare_exchange_strong_explicit(volatile atomic<_Tp>* __o, template <class _Tp> _LIBCPP_INLINE_VISIBILITY bool -atomic_compare_exchange_strong_explicit(atomic<_Tp>* __o, _Tp* __e, - _Tp __d, +atomic_compare_exchange_strong_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::value_type* __e, + typename atomic<_Tp>::value_type __d, memory_order __s, memory_order __f) _NOEXCEPT _LIBCPP_CHECK_EXCHANGE_MEMORY_ORDER(__s, __f) { @@ -2156,10 +2171,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_add(volatile atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_add(volatile atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op) _NOEXCEPT { return __o->fetch_add(__op); } @@ -2168,10 +2183,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_add(atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_add(atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op) _NOEXCEPT { return __o->fetch_add(__op); } @@ -2179,7 +2194,7 @@ atomic_fetch_add(atomic<_Tp>* __o, _Tp __op) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_add(volatile atomic<_Tp*>* __o, ptrdiff_t __op) _NOEXCEPT +atomic_fetch_add(volatile atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op) _NOEXCEPT { return __o->fetch_add(__op); } @@ -2187,7 +2202,7 @@ atomic_fetch_add(volatile atomic<_Tp*>* __o, ptrdiff_t __op) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_add(atomic<_Tp*>* __o, ptrdiff_t __op) _NOEXCEPT +atomic_fetch_add(atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op) _NOEXCEPT { return __o->fetch_add(__op); } @@ -2198,10 +2213,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_add_explicit(volatile atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_add_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_add(__op, __m); } @@ -2210,10 +2225,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_add_explicit(atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_add_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_add(__op, __m); } @@ -2221,8 +2236,7 @@ atomic_fetch_add_explicit(atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEP template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_add_explicit(volatile atomic<_Tp*>* __o, ptrdiff_t __op, - memory_order __m) _NOEXCEPT +atomic_fetch_add_explicit(volatile atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_add(__op, __m); } @@ -2230,7 +2244,7 @@ atomic_fetch_add_explicit(volatile atomic<_Tp*>* __o, ptrdiff_t __op, template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_add_explicit(atomic<_Tp*>* __o, ptrdiff_t __op, memory_order __m) _NOEXCEPT +atomic_fetch_add_explicit(atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_add(__op, __m); } @@ -2241,10 +2255,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_sub(volatile atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_sub(volatile atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op) _NOEXCEPT { return __o->fetch_sub(__op); } @@ -2253,10 +2267,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_sub(atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_sub(atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op) _NOEXCEPT { return __o->fetch_sub(__op); } @@ -2264,7 +2278,7 @@ atomic_fetch_sub(atomic<_Tp>* __o, _Tp __op) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_sub(volatile atomic<_Tp*>* __o, ptrdiff_t __op) _NOEXCEPT +atomic_fetch_sub(volatile atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op) _NOEXCEPT { return __o->fetch_sub(__op); } @@ -2272,7 +2286,7 @@ atomic_fetch_sub(volatile atomic<_Tp*>* __o, ptrdiff_t __op) _NOEXCEPT template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_sub(atomic<_Tp*>* __o, ptrdiff_t __op) _NOEXCEPT +atomic_fetch_sub(atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op) _NOEXCEPT { return __o->fetch_sub(__op); } @@ -2283,10 +2297,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_sub_explicit(volatile atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_sub_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_sub(__op, __m); } @@ -2295,10 +2309,10 @@ template <class _Tp> _LIBCPP_INLINE_VISIBILITY typename enable_if < - is_integral<_Tp>::value && !is_same<_Tp, bool>::value, + is_integral<_Tp>::value && !is_same<_Tp, bool>::value && !is_const<_Tp>::value, _Tp >::type -atomic_fetch_sub_explicit(atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_sub_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_sub(__op, __m); } @@ -2306,8 +2320,7 @@ atomic_fetch_sub_explicit(atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEP template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_sub_explicit(volatile atomic<_Tp*>* __o, ptrdiff_t __op, - memory_order __m) _NOEXCEPT +atomic_fetch_sub_explicit(volatile atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_sub(__op, __m); } @@ -2315,7 +2328,7 @@ atomic_fetch_sub_explicit(volatile atomic<_Tp*>* __o, ptrdiff_t __op, template <class _Tp> _LIBCPP_INLINE_VISIBILITY _Tp* -atomic_fetch_sub_explicit(atomic<_Tp*>* __o, ptrdiff_t __op, memory_order __m) _NOEXCEPT +atomic_fetch_sub_explicit(atomic<_Tp*>* __o, typename atomic<_Tp*>::difference_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_sub(__op, __m); } @@ -2329,7 +2342,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_and(volatile atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_and(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op) _NOEXCEPT { return __o->fetch_and(__op); } @@ -2341,7 +2354,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_and(atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_and(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op) _NOEXCEPT { return __o->fetch_and(__op); } @@ -2355,7 +2368,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_and_explicit(volatile atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_and_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_and(__op, __m); } @@ -2367,7 +2380,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_and_explicit(atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_and_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_and(__op, __m); } @@ -2381,7 +2394,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_or(volatile atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_or(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op) _NOEXCEPT { return __o->fetch_or(__op); } @@ -2393,7 +2406,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_or(atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_or(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op) _NOEXCEPT { return __o->fetch_or(__op); } @@ -2407,7 +2420,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_or_explicit(volatile atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_or_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_or(__op, __m); } @@ -2419,7 +2432,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_or_explicit(atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_or_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_or(__op, __m); } @@ -2433,7 +2446,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_xor(volatile atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_xor(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op) _NOEXCEPT { return __o->fetch_xor(__op); } @@ -2445,7 +2458,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_xor(atomic<_Tp>* __o, _Tp __op) _NOEXCEPT +atomic_fetch_xor(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op) _NOEXCEPT { return __o->fetch_xor(__op); } @@ -2459,7 +2472,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_xor_explicit(volatile atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_xor_explicit(volatile atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_xor(__op, __m); } @@ -2471,7 +2484,7 @@ typename enable_if is_integral<_Tp>::value && !is_same<_Tp, bool>::value, _Tp >::type -atomic_fetch_xor_explicit(atomic<_Tp>* __o, _Tp __op, memory_order __m) _NOEXCEPT +atomic_fetch_xor_explicit(atomic<_Tp>* __o, typename atomic<_Tp>::value_type __op, memory_order __m) _NOEXCEPT { return __o->fetch_xor(__op, __m); } @@ -2715,6 +2728,9 @@ typedef atomic<long> atomic_long; typedef atomic<unsigned long> atomic_ulong; typedef atomic<long long> atomic_llong; typedef atomic<unsigned long long> atomic_ullong; +#ifndef _LIBCPP_NO_HAS_CHAR8_T +typedef atomic<char8_t> atomic_char8_t; +#endif typedef atomic<char16_t> atomic_char16_t; typedef atomic<char32_t> atomic_char32_t; typedef atomic<wchar_t> atomic_wchar_t; diff --git a/lib/libcxx/include/barrier b/lib/libcxx/include/barrier index 58e3eef9cf..be213a6895 100644 --- a/lib/libcxx/include/barrier +++ b/lib/libcxx/include/barrier @@ -22,6 +22,8 @@ namespace std public: using arrival_token = see below; + static constexpr ptrdiff_t max() noexcept; + constexpr explicit barrier(ptrdiff_t phase_count, CompletionFunction f = CompletionFunction()); ~barrier(); @@ -44,6 +46,7 @@ namespace std */ #include <__config> +#include <__availability> #include <atomic> #ifndef _LIBCPP_HAS_NO_TREE_BARRIER # include <memory> @@ -57,6 +60,9 @@ namespace std # error <barrier> is not supported on this single threaded system #endif +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + #if _LIBCPP_STD_VER >= 14 _LIBCPP_BEGIN_NAMESPACE_STD @@ -287,7 +293,7 @@ public: _LIBCPP_AVAILABILITY_SYNC _LIBCPP_INLINE_VISIBILITY barrier(ptrdiff_t __count, _CompletionF __completion = _CompletionF()) - : __b(__count, std::move(__completion)) { + : __b(__count, _VSTD::move(__completion)) { } barrier(barrier const&) = delete; @@ -301,7 +307,7 @@ public: _LIBCPP_AVAILABILITY_SYNC _LIBCPP_INLINE_VISIBILITY void wait(arrival_token&& __phase) const { - __b.wait(std::move(__phase)); + __b.wait(_VSTD::move(__phase)); } _LIBCPP_AVAILABILITY_SYNC _LIBCPP_INLINE_VISIBILITY void arrive_and_wait() @@ -319,4 +325,6 @@ _LIBCPP_END_NAMESPACE_STD #endif // _LIBCPP_STD_VER >= 14 +_LIBCPP_POP_MACROS + #endif //_LIBCPP_BARRIER diff --git a/lib/libcxx/include/bit b/lib/libcxx/include/bit index ae4605b191..f8c37c3d6b 100644 --- a/lib/libcxx/include/bit +++ b/lib/libcxx/include/bit @@ -17,13 +17,13 @@ namespace std { // [bit.pow.two], integral powers of 2 template <class T> - constexpr bool ispow2(T x) noexcept; // C++20 + constexpr bool has_single_bit(T x) noexcept; // C++20 template <class T> - constexpr T ceil2(T x); // C++20 + constexpr T bit_ceil(T x); // C++20 template <class T> - constexpr T floor2(T x) noexcept; // C++20 + constexpr T bit_floor(T x) noexcept; // C++20 template <class T> - constexpr T log2p1(T x) noexcept; // C++20 + constexpr T bit_width(T x) noexcept; // C++20 // [bit.rotate], rotating template<class T> @@ -55,13 +55,14 @@ namespace std { */ #include <__config> +#include <__bits> #include <limits> #include <type_traits> #include <version> #include <__debug> #if defined(__IBMCPP__) -#include "support/ibm/support.h" +#include "__support/ibm/support.h" #endif #if defined(_LIBCPP_COMPILER_MSVC) #include <intrin.h> @@ -76,122 +77,6 @@ _LIBCPP_PUSH_MACROS _LIBCPP_BEGIN_NAMESPACE_STD -#ifndef _LIBCPP_COMPILER_MSVC - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_ctz(unsigned __x) _NOEXCEPT { return __builtin_ctz(__x); } - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_ctz(unsigned long __x) _NOEXCEPT { return __builtin_ctzl(__x); } - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_ctz(unsigned long long __x) _NOEXCEPT { return __builtin_ctzll(__x); } - - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_clz(unsigned __x) _NOEXCEPT { return __builtin_clz(__x); } - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_clz(unsigned long __x) _NOEXCEPT { return __builtin_clzl(__x); } - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_clz(unsigned long long __x) _NOEXCEPT { return __builtin_clzll(__x); } - - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_popcount(unsigned __x) _NOEXCEPT { return __builtin_popcount(__x); } - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_popcount(unsigned long __x) _NOEXCEPT { return __builtin_popcountl(__x); } - -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -int __libcpp_popcount(unsigned long long __x) _NOEXCEPT { return __builtin_popcountll(__x); } - -#else // _LIBCPP_COMPILER_MSVC - -// Precondition: __x != 0 -inline _LIBCPP_INLINE_VISIBILITY -int __libcpp_ctz(unsigned __x) { - static_assert(sizeof(unsigned) == sizeof(unsigned long), ""); - static_assert(sizeof(unsigned long) == 4, ""); - unsigned long __where; - if (_BitScanForward(&__where, __x)) - return static_cast<int>(__where); - return 32; -} - -inline _LIBCPP_INLINE_VISIBILITY -int __libcpp_ctz(unsigned long __x) { - static_assert(sizeof(unsigned long) == sizeof(unsigned), ""); - return __ctz(static_cast<unsigned>(__x)); -} - -inline _LIBCPP_INLINE_VISIBILITY -int __libcpp_ctz(unsigned long long __x) { - unsigned long __where; -#if defined(_LIBCPP_HAS_BITSCAN64) - (defined(_M_AMD64) || defined(__x86_64__)) - if (_BitScanForward64(&__where, __x)) - return static_cast<int>(__where); -#else - // Win32 doesn't have _BitScanForward64 so emulate it with two 32 bit calls. - if (_BitScanForward(&__where, static_cast<unsigned long>(__x))) - return static_cast<int>(__where); - if (_BitScanForward(&__where, static_cast<unsigned long>(__x >> 32))) - return static_cast<int>(__where + 32); -#endif - return 64; -} - -// Precondition: __x != 0 -inline _LIBCPP_INLINE_VISIBILITY -int __libcpp_clz(unsigned __x) { - static_assert(sizeof(unsigned) == sizeof(unsigned long), ""); - static_assert(sizeof(unsigned long) == 4, ""); - unsigned long __where; - if (_BitScanReverse(&__where, __x)) - return static_cast<int>(31 - __where); - return 32; // Undefined Behavior. -} - -inline _LIBCPP_INLINE_VISIBILITY -int __libcpp_clz(unsigned long __x) { - static_assert(sizeof(unsigned) == sizeof(unsigned long), ""); - return __libcpp_clz(static_cast<unsigned>(__x)); -} - -inline _LIBCPP_INLINE_VISIBILITY -int __libcpp_clz(unsigned long long __x) { - unsigned long __where; -#if defined(_LIBCPP_HAS_BITSCAN64) - if (_BitScanReverse64(&__where, __x)) - return static_cast<int>(63 - __where); -#else - // Win32 doesn't have _BitScanReverse64 so emulate it with two 32 bit calls. - if (_BitScanReverse(&__where, static_cast<unsigned long>(__x >> 32))) - return static_cast<int>(63 - (__where + 32)); - if (_BitScanReverse(&__where, static_cast<unsigned long>(__x))) - return static_cast<int>(63 - __where); -#endif - return 64; // Undefined Behavior. -} - -inline _LIBCPP_INLINE_VISIBILITY int __libcpp_popcount(unsigned __x) { - static_assert(sizeof(unsigned) == 4, ""); - return __popcnt(__x); -} - -inline _LIBCPP_INLINE_VISIBILITY int __libcpp_popcount(unsigned long __x) { - static_assert(sizeof(unsigned long) == 4, ""); - return __popcnt(__x); -} - -inline _LIBCPP_INLINE_VISIBILITY int __libcpp_popcount(unsigned long long __x) { - static_assert(sizeof(unsigned long long) == 8, ""); - return __popcnt64(__x); -} - -#endif // _LIBCPP_COMPILER_MSVC template <class _Tp> using __bitop_unsigned_integer _LIBCPP_NODEBUG_TYPE = integral_constant<bool, @@ -343,14 +228,14 @@ _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX11 unsigned __bit_log2(_Tp __t) _NOEXCEPT { static_assert(__bitop_unsigned_integer<_Tp>::value, "__bit_log2 requires unsigned"); - return std::numeric_limits<_Tp>::digits - 1 - __countl_zero(__t); + return numeric_limits<_Tp>::digits - 1 - __countl_zero(__t); } template <class _Tp> _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR -bool __ispow2(_Tp __t) _NOEXCEPT +bool __has_single_bit(_Tp __t) _NOEXCEPT { - static_assert(__bitop_unsigned_integer<_Tp>::value, "__ispow2 requires unsigned"); + static_assert(__bitop_unsigned_integer<_Tp>::value, "__has_single_bit requires unsigned"); return __t != 0 && (((__t & (__t - 1)) == 0)); } @@ -399,7 +284,7 @@ _LIBCPP_INLINE_VISIBILITY constexpr enable_if_t<__bitop_unsigned_integer<_Tp>::value, int> countr_zero(_Tp __t) noexcept { - return __countr_zero(__t); + return __countr_zero(__t); } @@ -424,15 +309,15 @@ popcount(_Tp __t) noexcept template <class _Tp> _LIBCPP_INLINE_VISIBILITY constexpr enable_if_t<__bitop_unsigned_integer<_Tp>::value, bool> -ispow2(_Tp __t) noexcept +has_single_bit(_Tp __t) noexcept { - return __ispow2(__t); + return __has_single_bit(__t); } template <class _Tp> _LIBCPP_INLINE_VISIBILITY constexpr enable_if_t<__bitop_unsigned_integer<_Tp>::value, _Tp> -floor2(_Tp __t) noexcept +bit_floor(_Tp __t) noexcept { return __t == 0 ? 0 : _Tp{1} << __bit_log2(__t); } @@ -440,11 +325,11 @@ floor2(_Tp __t) noexcept template <class _Tp> _LIBCPP_INLINE_VISIBILITY constexpr enable_if_t<__bitop_unsigned_integer<_Tp>::value, _Tp> -ceil2(_Tp __t) noexcept +bit_ceil(_Tp __t) noexcept { if (__t < 2) return 1; const unsigned __n = numeric_limits<_Tp>::digits - countl_zero((_Tp)(__t - 1u)); - _LIBCPP_DEBUG_ASSERT(__libcpp_is_constant_evaluated() || __n != numeric_limits<_Tp>::digits, "Bad input to ceil2"); + _LIBCPP_DEBUG_ASSERT(__libcpp_is_constant_evaluated() || __n != numeric_limits<_Tp>::digits, "Bad input to bit_ceil"); if constexpr (sizeof(_Tp) >= sizeof(unsigned)) return _Tp{1} << __n; @@ -459,12 +344,11 @@ ceil2(_Tp __t) noexcept template <class _Tp> _LIBCPP_INLINE_VISIBILITY constexpr enable_if_t<__bitop_unsigned_integer<_Tp>::value, _Tp> -log2p1(_Tp __t) noexcept +bit_width(_Tp __t) noexcept { return __t == 0 ? 0 : __bit_log2(__t) + 1; } - enum class endian { little = 0xDEAD, diff --git a/lib/libcxx/include/bitset b/lib/libcxx/include/bitset index 4755fbeb21..00503fe1c1 100644 --- a/lib/libcxx/include/bitset +++ b/lib/libcxx/include/bitset @@ -380,7 +380,7 @@ unsigned long long __bitset<_N_words, _Size>::to_ullong(true_type, true_type) const { unsigned long long __r = __first_[0]; - for (std::size_t __i = 1; __i < sizeof(unsigned long long) / sizeof(__storage_type); ++__i) + for (size_t __i = 1; __i < sizeof(unsigned long long) / sizeof(__storage_type); ++__i) __r |= static_cast<unsigned long long>(__first_[__i]) << (sizeof(__storage_type) * CHAR_BIT); return __r; } @@ -625,13 +625,13 @@ protected: explicit _LIBCPP_CONSTEXPR __bitset(unsigned long long) _NOEXCEPT; _LIBCPP_INLINE_VISIBILITY reference __make_ref(size_t) _NOEXCEPT - {return reference(0, 1);} + {return reference(nullptr, 1);} _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR const_reference __make_ref(size_t) const _NOEXCEPT - {return const_reference(0, 1);} + {return const_reference(nullptr, 1);} _LIBCPP_INLINE_VISIBILITY iterator __make_iter(size_t) _NOEXCEPT - {return iterator(0, 0);} + {return iterator(nullptr, 0);} _LIBCPP_INLINE_VISIBILITY const_iterator __make_iter(size_t) const _NOEXCEPT - {return const_iterator(0, 0);} + {return const_iterator(nullptr, 0);} _LIBCPP_INLINE_VISIBILITY void operator&=(const __bitset&) _NOEXCEPT {} _LIBCPP_INLINE_VISIBILITY void operator|=(const __bitset&) _NOEXCEPT {} @@ -990,7 +990,7 @@ inline size_t bitset<_Size>::count() const _NOEXCEPT { - return static_cast<size_t>(__count_bool_true(base::__make_iter(0), _Size)); + return static_cast<size_t>(_VSTD::__count_bool_true(base::__make_iter(0), _Size)); } template <size_t _Size> diff --git a/lib/libcxx/include/charconv b/lib/libcxx/include/charconv index b64000242a..4666c5c51d 100644 --- a/lib/libcxx/include/charconv +++ b/lib/libcxx/include/charconv @@ -74,12 +74,13 @@ namespace std { */ #include <__config> +#include <__availability> #include <__errc> -#include <type_traits> +#include <cmath> // for log2f +#include <cstdint> +#include <cstring> #include <limits> -#include <stdint.h> -#include <string.h> -#include <math.h> +#include <type_traits> #include <__debug> @@ -206,7 +207,7 @@ __mul_overflowed(unsigned char __a, _Tp __b, unsigned char& __r) { auto __c = __a * __b; __r = __c; - return __c > (numeric_limits<unsigned char>::max)(); + return __c > numeric_limits<unsigned char>::max(); } template <typename _Tp> @@ -215,7 +216,7 @@ __mul_overflowed(unsigned short __a, _Tp __b, unsigned short& __r) { auto __c = __a * __b; __r = __c; - return __c > (numeric_limits<unsigned short>::max)(); + return __c > numeric_limits<unsigned short>::max(); } template <typename _Tp> @@ -226,7 +227,7 @@ __mul_overflowed(_Tp __a, _Tp __b, _Tp& __r) #if !defined(_LIBCPP_COMPILER_MSVC) return __builtin_mul_overflow(__a, __b, &__r); #else - bool __did = __b && ((numeric_limits<_Tp>::max)() / __b) < __a; + bool __did = __b && (numeric_limits<_Tp>::max() / __b) < __a; __r = __a * __b; return __did; #endif @@ -332,7 +333,7 @@ __to_chars_itoa(char* __first, char* __last, _Tp __value, false_type) auto __len = __p - __buf; if (__len <= __diff) { - memcpy(__first, __buf, __len); + _VSTD::memcpy(__first, __buf, __len); return {__first + __len, {}}; } else @@ -381,7 +382,7 @@ __to_chars_integral(char* __first, char* __last, _Tp __value, int __base, return {__last, errc::value_too_large}; else { - memmove(__first, __p, __len); + _VSTD::memmove(__first, __p, __len); return {__first + __len, {}}; } } @@ -428,13 +429,13 @@ __sign_combinator(_It __first, _It __last, _Tp& __value, _Fn __f, _Ts... __args) if (__x <= __complement(__to_unsigned(__tl::min()))) { __x = __complement(__x); - memcpy(&__value, &__x, sizeof(__x)); + _VSTD::memcpy(&__value, &__x, sizeof(__x)); return __r; } } else { - if (__x <= (__tl::max)()) + if (__x <= __tl::max()) { __value = __x; return __r; @@ -525,7 +526,7 @@ __from_chars_atoi(const char* __first, const char* __last, _Tp& __value) auto __p = __tx::__read(__first, __last, __a, __b); if (__p == __last || !__in_pattern(*__p)) { - __output_type __m = (numeric_limits<_Tp>::max)(); + __output_type __m = numeric_limits<_Tp>::max(); if (__m >= __a && __m - __a >= __b) { __value = __a + __b; @@ -580,7 +581,7 @@ __from_chars_integral(const char* __first, const char* __last, _Tp& __value, if (__p == __last || !__in_pattern(*__p, __base)) { - if ((__tl::max)() - __a >= __b) + if (__tl::max() - __a >= __b) { __value = __a + __b; return {__p, {}}; diff --git a/lib/libcxx/include/chrono b/lib/libcxx/include/chrono index 117aab3190..53e4546010 100644 --- a/lib/libcxx/include/chrono +++ b/lib/libcxx/include/chrono @@ -824,6 +824,7 @@ constexpr chrono::year operator ""y(unsigned lo */ #include <__config> +#include <__availability> #include <ctime> #include <type_traits> #include <ratio> @@ -1076,7 +1077,7 @@ public: is_convertible<_Rep2, rep>::value && (treat_as_floating_point<rep>::value || !treat_as_floating_point<_Rep2>::value) - >::type* = 0) + >::type* = nullptr) : __rep_(__r) {} // conversions @@ -1089,7 +1090,7 @@ public: treat_as_floating_point<rep>::value || (__no_overflow<_Period2, period>::type::den == 1 && !treat_as_floating_point<_Rep2>::value)) - >::type* = 0) + >::type* = nullptr) : __rep_(_VSTD::chrono::duration_cast<duration>(__d).count()) {} // observer @@ -1375,7 +1376,7 @@ public: typename enable_if < is_convertible<_Duration2, duration>::value - >::type* = 0) + >::type* = nullptr) : __d_(t.time_since_epoch()) {} // observer diff --git a/lib/libcxx/include/cmath b/lib/libcxx/include/cmath index 0901a23a24..138ae6f99a 100644 --- a/lib/libcxx/include/cmath +++ b/lib/libcxx/include/cmath @@ -660,8 +660,8 @@ _LIBCPP_CONSTEXPR _IntT __max_representable_int_for_float() _NOEXCEPT { template <class _IntT, class _RealT> _LIBCPP_INLINE_VISIBILITY _IntT __clamp_to_integral(_RealT __r) _NOEXCEPT { - using _Lim = std::numeric_limits<_IntT>; - const _IntT _MaxVal = std::__max_representable_int_for_float<_IntT, _RealT>(); + using _Lim = numeric_limits<_IntT>; + const _IntT _MaxVal = __max_representable_int_for_float<_IntT, _RealT>(); if (__r >= ::nextafter(static_cast<_RealT>(_MaxVal), INFINITY)) { return _Lim::max(); } else if (__r <= _Lim::lowest()) { diff --git a/lib/libcxx/include/codecvt b/lib/libcxx/include/codecvt index 05fa765c31..2befa1b0af 100644 --- a/lib/libcxx/include/codecvt +++ b/lib/libcxx/include/codecvt @@ -109,6 +109,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf8<char16_t> : public codecvt<char16_t, char, mbstate_t> @@ -125,6 +126,8 @@ public: codecvt_mode _Mode) : codecvt<char16_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, @@ -144,6 +147,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf8<char32_t> : public codecvt<char32_t, char, mbstate_t> @@ -160,6 +164,8 @@ public: codecvt_mode _Mode) : codecvt<char32_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, @@ -267,6 +273,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf16<char16_t, false> : public codecvt<char16_t, char, mbstate_t> @@ -283,6 +290,8 @@ public: codecvt_mode _Mode) : codecvt<char16_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, @@ -302,6 +311,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf16<char16_t, true> : public codecvt<char16_t, char, mbstate_t> @@ -318,6 +328,8 @@ public: codecvt_mode _Mode) : codecvt<char16_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, @@ -337,6 +349,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf16<char32_t, false> : public codecvt<char32_t, char, mbstate_t> @@ -353,6 +366,8 @@ public: codecvt_mode _Mode) : codecvt<char32_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, @@ -372,6 +387,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf16<char32_t, true> : public codecvt<char32_t, char, mbstate_t> @@ -388,6 +404,8 @@ public: codecvt_mode _Mode) : codecvt<char32_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, @@ -460,6 +478,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf8_utf16<char32_t> : public codecvt<char32_t, char, mbstate_t> @@ -476,6 +495,8 @@ public: codecvt_mode _Mode) : codecvt<char32_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, @@ -495,6 +516,7 @@ protected: virtual int do_max_length() const _NOEXCEPT; }; +_LIBCPP_SUPPRESS_DEPRECATED_PUSH template <> class _LIBCPP_TYPE_VIS __codecvt_utf8_utf16<char16_t> : public codecvt<char16_t, char, mbstate_t> @@ -511,6 +533,8 @@ public: codecvt_mode _Mode) : codecvt<char16_t, char, mbstate_t>(__refs), _Maxcode_(_Maxcode), _Mode_(_Mode) {} +_LIBCPP_SUPPRESS_DEPRECATED_POP + protected: virtual result do_out(state_type& __st, diff --git a/lib/libcxx/include/compare b/lib/libcxx/include/compare index 717859a1e3..048f4821dd 100644 --- a/lib/libcxx/include/compare +++ b/lib/libcxx/include/compare @@ -154,8 +154,13 @@ enum class _LIBCPP_ENUM_VIS _NCmpResult : signed char { __unordered = -127 }; -struct _CmpUnspecifiedType; -using _CmpUnspecifiedParam = void (_CmpUnspecifiedType::*)(); +struct _CmpUnspecifiedParam { + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEVAL + _CmpUnspecifiedParam(int _CmpUnspecifiedParam::*) noexcept {} + + template<typename _Tp, typename = _VSTD::enable_if_t<!_VSTD::is_same_v<_Tp, int>>> + _CmpUnspecifiedParam(_Tp) = delete; +}; class weak_equality { _LIBCPP_INLINE_VISIBILITY @@ -696,8 +701,8 @@ constexpr _ClassifyCompCategory __type_to_enum() noexcept { template <size_t _Size> constexpr _ClassifyCompCategory -__compute_comp_type(std::array<_ClassifyCompCategory, _Size> __types) { - std::array<int, _CCC_Size> __seen = {}; +__compute_comp_type(array<_ClassifyCompCategory, _Size> __types) { + array<int, _CCC_Size> __seen = {}; for (auto __type : __types) ++__seen[__type]; if (__seen[_None]) diff --git a/lib/libcxx/include/complex b/lib/libcxx/include/complex index 36c66db50e..93b7bb5dd5 100644 --- a/lib/libcxx/include/complex +++ b/lib/libcxx/include/complex @@ -227,14 +227,6 @@ template<class T> complex<T> sqrt (const complex<T>&); template<class T> complex<T> tan (const complex<T>&); template<class T> complex<T> tanh (const complex<T>&); -template<class T, class charT, class traits> - basic_istream<charT, traits>& - operator>>(basic_istream<charT, traits>& is, complex<T>& x); - -template<class T, class charT, class traits> - basic_ostream<charT, traits>& - operator<<(basic_ostream<charT, traits>& o, const complex<T>& x); - } // std */ @@ -244,9 +236,12 @@ template<class T, class charT, class traits> #include <stdexcept> #include <cmath> #include <iosfwd> -#include <sstream> #include <version> +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) +# include <sstream> // for std::basic_ostringstream +#endif + #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header #endif @@ -955,7 +950,7 @@ inline _LIBCPP_INLINE_VISIBILITY complex<_Tp> proj(const complex<_Tp>& __c) { - std::complex<_Tp> __r = __c; + complex<_Tp> __r = __c; if (__libcpp_isinf_or_builtin(__c.real()) || __libcpp_isinf_or_builtin(__c.imag())) __r = complex<_Tp>(INFINITY, copysign(_Tp(0), __c.imag())); return __r; @@ -1438,6 +1433,7 @@ operator>>(basic_istream<_CharT, _Traits>& __is, complex<_Tp>& __x) return __is; } +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) template<class _Tp, class _CharT, class _Traits> basic_ostream<_CharT, _Traits>& operator<<(basic_ostream<_CharT, _Traits>& __os, const complex<_Tp>& __x) @@ -1449,6 +1445,7 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const complex<_Tp>& __x) __s << '(' << __x.real() << ',' << __x.imag() << ')'; return __os << __s.str(); } +#endif // !_LIBCPP_HAS_NO_LOCALIZATION #if _LIBCPP_STD_VER > 11 // Literal suffix for complex number literals [complex.literals] diff --git a/lib/libcxx/include/concepts b/lib/libcxx/include/concepts index 047e2c290f..cf5f9d6397 100644 --- a/lib/libcxx/include/concepts +++ b/lib/libcxx/include/concepts @@ -157,6 +157,11 @@ concept __same_as_impl = _VSTD::_IsSame<_Tp, _Up>::value; template<class _Tp, class _Up> concept same_as = __same_as_impl<_Tp, _Up> && __same_as_impl<_Up, _Tp>; +// [concept.destructible] + +template<class _Tp> +concept destructible = _VSTD::is_nothrow_destructible_v<_Tp>; + #endif //_LIBCPP_STD_VER > 17 && defined(__cpp_concepts) && __cpp_concepts >= 201811L _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/include/ctime b/lib/libcxx/include/ctime index f9f2f1659d..b0e6c65af5 100644 --- a/lib/libcxx/include/ctime +++ b/lib/libcxx/include/ctime @@ -52,6 +52,20 @@ int timespec_get( struct timespec *ts, int base); // C++17 #pragma GCC system_header #endif +// FIXME: +// Apple SDKs don't define ::timespec_get unconditionally in C++ mode. This +// should be fixed in future SDKs, but for the time being we need to avoid +// trying to use that declaration when the SDK doesn't provide it. Note that +// we're detecting this here instead of in <__config> because we can't include +// system headers from <__config>, since it leads to circular module dependencies. +// This is also meant to be a very temporary workaround until the SDKs are fixed. +#if defined(__APPLE__) +# include <sys/cdefs.h> +# if defined(_LIBCPP_HAS_TIMESPEC_GET) && (__DARWIN_C_LEVEL < __DARWIN_C_FULL) +# define _LIBCPP_HAS_TIMESPEC_GET_NOT_ACTUALLY_PROVIDED +# endif +#endif + _LIBCPP_BEGIN_NAMESPACE_STD using ::clock_t; @@ -72,7 +86,7 @@ using ::gmtime; using ::localtime; #endif using ::strftime; -#if _LIBCPP_STD_VER > 14 && defined(_LIBCPP_HAS_TIMESPEC_GET) +#if _LIBCPP_STD_VER > 14 && defined(_LIBCPP_HAS_TIMESPEC_GET) && !defined(_LIBCPP_HAS_TIMESPEC_GET_NOT_ACTUALLY_PROVIDED) using ::timespec_get; #endif diff --git a/lib/libcxx/include/deque b/lib/libcxx/include/deque index c2ea5f2dbe..c6517d170f 100644 --- a/lib/libcxx/include/deque +++ b/lib/libcxx/include/deque @@ -1237,7 +1237,7 @@ __deque_base<_Tp, _Allocator>::swap(__deque_base& __c) __map_.swap(__c.__map_); _VSTD::swap(__start_, __c.__start_); _VSTD::swap(size(), __c.size()); - __swap_allocator(__alloc(), __c.__alloc()); + _VSTD::__swap_allocator(__alloc(), __c.__alloc()); } template <class _Tp, class _Allocator> @@ -1393,7 +1393,7 @@ public: size_type size() const _NOEXCEPT {return __base::size();} _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT - {return std::min<size_type>( + {return _VSTD::min<size_type>( __alloc_traits::max_size(__base::__alloc()), numeric_limits<difference_type>::max());} void resize(size_type __n); @@ -1586,7 +1586,7 @@ public: #ifndef _LIBCPP_HAS_NO_DEDUCTION_GUIDES template<class _InputIterator, - class _Alloc = typename std::allocator<typename iterator_traits<_InputIterator>::value_type>, + class _Alloc = allocator<typename iterator_traits<_InputIterator>::value_type>, class = typename enable_if<__is_allocator<_Alloc>::value, void>::type > deque(_InputIterator, _InputIterator) @@ -2376,7 +2376,7 @@ deque<_Tp, _Allocator>::__append(_ForIter __f, _ForIter __l, for (__deque_block_range __br : __deque_range(__base::end(), __base::end() + __n)) { _ConstructTransaction __tx(this, __br); for (; __tx.__pos_ != __tx.__end_; ++__tx.__pos_, (void)++__f) { - __alloc_traits::construct(__a, std::__to_address(__tx.__pos_), *__f); + __alloc_traits::construct(__a, _VSTD::__to_address(__tx.__pos_), *__f); } } } @@ -2393,7 +2393,7 @@ deque<_Tp, _Allocator>::__append(size_type __n) for (__deque_block_range __br : __deque_range(__base::end(), __base::end() + __n)) { _ConstructTransaction __tx(this, __br); for (; __tx.__pos_ != __tx.__end_; ++__tx.__pos_) { - __alloc_traits::construct(__a, std::__to_address(__tx.__pos_)); + __alloc_traits::construct(__a, _VSTD::__to_address(__tx.__pos_)); } } } @@ -2410,7 +2410,7 @@ deque<_Tp, _Allocator>::__append(size_type __n, const value_type& __v) for (__deque_block_range __br : __deque_range(__base::end(), __base::end() + __n)) { _ConstructTransaction __tx(this, __br); for (; __tx.__pos_ != __tx.__end_; ++__tx.__pos_) { - __alloc_traits::construct(__a, std::__to_address(__tx.__pos_), __v); + __alloc_traits::construct(__a, _VSTD::__to_address(__tx.__pos_), __v); } } @@ -2708,7 +2708,7 @@ void deque<_Tp, _Allocator>::pop_front() { allocator_type& __a = __base::__alloc(); - __alloc_traits::destroy(__a, __to_address(*(__base::__map_.begin() + + __alloc_traits::destroy(__a, _VSTD::__to_address(*(__base::__map_.begin() + __base::__start_ / __base::__block_size) + __base::__start_ % __base::__block_size)); --__base::size(); @@ -2723,7 +2723,7 @@ deque<_Tp, _Allocator>::pop_back() _LIBCPP_ASSERT(!empty(), "deque::pop_back called for empty deque"); allocator_type& __a = __base::__alloc(); size_type __p = __base::size() + __base::__start_ - 1; - __alloc_traits::destroy(__a, __to_address(*(__base::__map_.begin() + + __alloc_traits::destroy(__a, _VSTD::__to_address(*(__base::__map_.begin() + __p / __base::__block_size) + __p % __base::__block_size)); --__base::size(); diff --git a/lib/libcxx/include/exception b/lib/libcxx/include/exception index 8e32979f57..4bf4049f33 100644 --- a/lib/libcxx/include/exception +++ b/lib/libcxx/include/exception @@ -77,6 +77,8 @@ template <class E> void rethrow_if_nested(const E& e); */ #include <__config> +#include <__availability> +#include <__memory/base.h> #include <cstddef> #include <cstdlib> #include <type_traits> diff --git a/lib/libcxx/include/experimental/memory_resource b/lib/libcxx/include/experimental/memory_resource index f999fb9bef..816d21f513 100644 --- a/lib/libcxx/include/experimental/memory_resource +++ b/lib/libcxx/include/experimental/memory_resource @@ -116,7 +116,7 @@ public: { return do_is_equal(__other); } // 8.5.3, memory.resource.priv -protected: +private: virtual void* do_allocate(size_t, size_t) = 0; virtual void do_deallocate(void*, size_t, size_t) = 0; virtual bool do_is_equal(memory_resource const &) const _NOEXCEPT = 0; @@ -381,7 +381,7 @@ public: { return __alloc_; } // 8.7.3, memory.resource.adaptor.mem -protected: +private: virtual void * do_allocate(size_t __bytes, size_t) { if (__bytes > __max_size()) { @@ -407,7 +407,6 @@ protected: return __p ? __alloc_ == __p->__alloc_ : false; } -private: _LIBCPP_INLINE_VISIBILITY size_t __max_size() const _NOEXCEPT { return numeric_limits<size_t>::max() - _MaxAlign; diff --git a/lib/libcxx/include/experimental/simd b/lib/libcxx/include/experimental/simd index 39ac35e4eb..41f8f799a0 100644 --- a/lib/libcxx/include/experimental/simd +++ b/lib/libcxx/include/experimental/simd @@ -659,6 +659,9 @@ public: #pragma GCC system_header #endif +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + _LIBCPP_BEGIN_NAMESPACE_EXPERIMENTAL_SIMD #if _LIBCPP_STD_VER >= 17 @@ -1566,4 +1569,6 @@ public: _LIBCPP_END_NAMESPACE_EXPERIMENTAL_SIMD +_LIBCPP_POP_MACROS + #endif /* _LIBCPP_EXPERIMENTAL_SIMD */ diff --git a/lib/libcxx/include/ext/hash_map b/lib/libcxx/include/ext/hash_map index 7478d74100..2d6024cb90 100644 --- a/lib/libcxx/include/ext/hash_map +++ b/lib/libcxx/include/ext/hash_map @@ -671,7 +671,7 @@ hash_map<_Key, _Tp, _Hash, _Pred, _Alloc>::__construct_node(const key_type& __k) __h.get_deleter().__first_constructed = true; __node_traits::construct(__na, _VSTD::addressof(__h->__value_.second)); __h.get_deleter().__second_constructed = true; - return _LIBCPP_EXPLICIT_MOVE(__h); // explicitly moved for C++03 + return __h; } template <class _Key, class _Tp, class _Hash, class _Pred, class _Alloc> diff --git a/lib/libcxx/include/filesystem b/lib/libcxx/include/filesystem index 1363b63027..92e37e183d 100644 --- a/lib/libcxx/include/filesystem +++ b/lib/libcxx/include/filesystem @@ -230,23 +230,31 @@ */ #include <__config> +#include <__availability> #include <cstddef> #include <cstdlib> #include <chrono> #include <iterator> #include <iosfwd> -#include <locale> #include <memory> #include <stack> #include <string> #include <system_error> #include <utility> -#include <iomanip> // for quoted #include <string_view> #include <version> +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) +# include <locale> +# include <iomanip> // for quoted +#endif + #include <__debug> +#if defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) +# error "The Filesystem library is not supported by this configuration of libc++" +#endif + #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header #endif @@ -543,6 +551,13 @@ struct __can_convert_char<wchar_t> { static const bool value = true; using __char_type = wchar_t; }; +#ifndef _LIBCPP_NO_HAS_CHAR8_T +template <> +struct __can_convert_char<char8_t> { + static const bool value = true; + using __char_type = char8_t; +}; +#endif template <> struct __can_convert_char<char16_t> { static const bool value = true; @@ -557,10 +572,20 @@ struct __can_convert_char<char32_t> { template <class _ECharT> typename enable_if<__can_convert_char<_ECharT>::value, bool>::type __is_separator(_ECharT __e) { +#if defined(_LIBCPP_WIN32API) + return __e == _ECharT('/') || __e == _ECharT('\\'); +#else return __e == _ECharT('/'); +#endif } -struct _NullSentinal {}; +#ifndef _LIBCPP_NO_HAS_CHAR8_T +typedef u8string __u8_string; +#else +typedef string __u8_string; +#endif + +struct _NullSentinel {}; template <class _Tp> using _Void = void; @@ -615,9 +640,9 @@ struct __is_pathable_char_array<_Source, _ECharT*, _UPtr, true> static _ECharT const* __range_begin(const _ECharT* __b) { return __b; } static _ECharT const* __range_end(const _ECharT* __b) { using _Iter = const _ECharT*; - const _ECharT __sentinal = _ECharT{}; + const _ECharT __sentinel = _ECharT{}; _Iter __e = __b; - for (; *__e != __sentinal; ++__e) + for (; *__e != __sentinel; ++__e) ; return __e; } @@ -639,7 +664,7 @@ struct __is_pathable_iter< using _Base = __can_convert_char<_ECharT>; static _Iter __range_begin(_Iter __b) { return __b; } - static _NullSentinal __range_end(_Iter) { return _NullSentinal{}; } + static _NullSentinel __range_end(_Iter) { return _NullSentinel{}; } static _ECharT __first_or_null(_Iter __b) { return *__b; } }; @@ -661,80 +686,219 @@ struct __is_pathable<_Tp, false, true, false> : __is_pathable_char_array<_Tp> { template <class _Tp> struct __is_pathable<_Tp, false, false, true> : __is_pathable_iter<_Tp> {}; +#if defined(_LIBCPP_WIN32API) +typedef wstring __path_string; +typedef wchar_t __path_value; +#else +typedef string __path_string; +typedef char __path_value; +#endif + +#if defined(_LIBCPP_WIN32API) +_LIBCPP_FUNC_VIS +size_t __wide_to_char(const wstring&, char*, size_t); +_LIBCPP_FUNC_VIS +size_t __char_to_wide(const string&, wchar_t*, size_t); +#endif + +template <class _ECharT> +struct _PathCVT; + +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) template <class _ECharT> struct _PathCVT { static_assert(__can_convert_char<_ECharT>::value, "Char type not convertible"); typedef __narrow_to_utf8<sizeof(_ECharT) * __CHAR_BIT__> _Narrower; +#if defined(_LIBCPP_WIN32API) + typedef __widen_from_utf8<sizeof(wchar_t) * __CHAR_BIT__> _Widener; +#endif - static void __append_range(string& __dest, _ECharT const* __b, + static void __append_range(__path_string& __dest, _ECharT const* __b, _ECharT const* __e) { +#if defined(_LIBCPP_WIN32API) + string __utf8; + _Narrower()(back_inserter(__utf8), __b, __e); + _Widener()(back_inserter(__dest), __utf8.data(), __utf8.data() + __utf8.size()); +#else _Narrower()(back_inserter(__dest), __b, __e); +#endif } template <class _Iter> - static void __append_range(string& __dest, _Iter __b, _Iter __e) { + static void __append_range(__path_string& __dest, _Iter __b, _Iter __e) { static_assert(!is_same<_Iter, _ECharT*>::value, "Call const overload"); if (__b == __e) return; basic_string<_ECharT> __tmp(__b, __e); +#if defined(_LIBCPP_WIN32API) + string __utf8; + _Narrower()(back_inserter(__utf8), __tmp.data(), + __tmp.data() + __tmp.length()); + _Widener()(back_inserter(__dest), __utf8.data(), __utf8.data() + __utf8.size()); +#else _Narrower()(back_inserter(__dest), __tmp.data(), __tmp.data() + __tmp.length()); +#endif } template <class _Iter> - static void __append_range(string& __dest, _Iter __b, _NullSentinal) { + static void __append_range(__path_string& __dest, _Iter __b, _NullSentinel) { static_assert(!is_same<_Iter, _ECharT*>::value, "Call const overload"); - const _ECharT __sentinal = _ECharT{}; - if (*__b == __sentinal) + const _ECharT __sentinel = _ECharT{}; + if (*__b == __sentinel) return; basic_string<_ECharT> __tmp; - for (; *__b != __sentinal; ++__b) + for (; *__b != __sentinel; ++__b) __tmp.push_back(*__b); +#if defined(_LIBCPP_WIN32API) + string __utf8; + _Narrower()(back_inserter(__utf8), __tmp.data(), + __tmp.data() + __tmp.length()); + _Widener()(back_inserter(__dest), __utf8.data(), __utf8.data() + __utf8.size()); +#else _Narrower()(back_inserter(__dest), __tmp.data(), __tmp.data() + __tmp.length()); +#endif } template <class _Source> - static void __append_source(string& __dest, _Source const& __s) { + static void __append_source(__path_string& __dest, _Source const& __s) { using _Traits = __is_pathable<_Source>; __append_range(__dest, _Traits::__range_begin(__s), _Traits::__range_end(__s)); } }; +#endif // !_LIBCPP_HAS_NO_LOCALIZATION template <> -struct _PathCVT<char> { +struct _PathCVT<__path_value> { template <class _Iter> static typename enable_if<__is_exactly_cpp17_input_iterator<_Iter>::value>::type - __append_range(string& __dest, _Iter __b, _Iter __e) { + __append_range(__path_string& __dest, _Iter __b, _Iter __e) { for (; __b != __e; ++__b) __dest.push_back(*__b); } template <class _Iter> static typename enable_if<__is_cpp17_forward_iterator<_Iter>::value>::type - __append_range(string& __dest, _Iter __b, _Iter __e) { + __append_range(__path_string& __dest, _Iter __b, _Iter __e) { __dest.__append_forward_unsafe(__b, __e); } template <class _Iter> - static void __append_range(string& __dest, _Iter __b, _NullSentinal) { - const char __sentinal = char{}; - for (; *__b != __sentinal; ++__b) + static void __append_range(__path_string& __dest, _Iter __b, _NullSentinel) { + const char __sentinel = char{}; + for (; *__b != __sentinel; ++__b) __dest.push_back(*__b); } template <class _Source> - static void __append_source(string& __dest, _Source const& __s) { + static void __append_source(__path_string& __dest, _Source const& __s) { + using _Traits = __is_pathable<_Source>; + __append_range(__dest, _Traits::__range_begin(__s), + _Traits::__range_end(__s)); + } +}; + +#if defined(_LIBCPP_WIN32API) +template <> +struct _PathCVT<char> { + + static void + __append_string(__path_string& __dest, const basic_string<char> &__str) { + size_t __size = __char_to_wide(__str, nullptr, 0); + size_t __pos = __dest.size(); + __dest.resize(__pos + __size); + __char_to_wide(__str, const_cast<__path_value*>(__dest.data()) + __pos, __size); + } + + template <class _Iter> + static typename enable_if<__is_exactly_cpp17_input_iterator<_Iter>::value>::type + __append_range(__path_string& __dest, _Iter __b, _Iter __e) { + basic_string<char> __tmp(__b, __e); + __append_string(__dest, __tmp); + } + + template <class _Iter> + static typename enable_if<__is_cpp17_forward_iterator<_Iter>::value>::type + __append_range(__path_string& __dest, _Iter __b, _Iter __e) { + basic_string<char> __tmp(__b, __e); + __append_string(__dest, __tmp); + } + + template <class _Iter> + static void __append_range(__path_string& __dest, _Iter __b, _NullSentinel) { + const char __sentinel = char{}; + basic_string<char> __tmp; + for (; *__b != __sentinel; ++__b) + __tmp.push_back(*__b); + __append_string(__dest, __tmp); + } + + template <class _Source> + static void __append_source(__path_string& __dest, _Source const& __s) { using _Traits = __is_pathable<_Source>; __append_range(__dest, _Traits::__range_begin(__s), _Traits::__range_end(__s)); } }; +template <class _ECharT> +struct _PathExport { + typedef __narrow_to_utf8<sizeof(wchar_t) * __CHAR_BIT__> _Narrower; + typedef __widen_from_utf8<sizeof(_ECharT) * __CHAR_BIT__> _Widener; + + template <class _Str> + static void __append(_Str& __dest, const __path_string& __src) { + string __utf8; + _Narrower()(back_inserter(__utf8), __src.data(), __src.data() + __src.size()); + _Widener()(back_inserter(__dest), __utf8.data(), __utf8.data() + __utf8.size()); + } +}; + +template <> +struct _PathExport<char> { + template <class _Str> + static void __append(_Str& __dest, const __path_string& __src) { + size_t __size = __wide_to_char(__src, nullptr, 0); + size_t __pos = __dest.size(); + __dest.resize(__size); + __wide_to_char(__src, const_cast<char*>(__dest.data()) + __pos, __size); + } +}; + +template <> +struct _PathExport<wchar_t> { + template <class _Str> + static void __append(_Str& __dest, const __path_string& __src) { + __dest.append(__src.begin(), __src.end()); + } +}; + +template <> +struct _PathExport<char16_t> { + template <class _Str> + static void __append(_Str& __dest, const __path_string& __src) { + __dest.append(__src.begin(), __src.end()); + } +}; + +#ifndef _LIBCPP_NO_HAS_CHAR8_T +template <> +struct _PathExport<char8_t> { + typedef __narrow_to_utf8<sizeof(wchar_t) * __CHAR_BIT__> _Narrower; + + template <class _Str> + static void __append(_Str& __dest, const __path_string& __src) { + _Narrower()(back_inserter(__dest), __src.data(), __src.data() + __src.size()); + } +}; +#endif /* !_LIBCPP_NO_HAS_CHAR8_T */ +#endif /* _LIBCPP_WIN32API */ + class _LIBCPP_TYPE_VIS path { template <class _SourceOrIter, class _Tp = path&> using _EnableIfPathable = @@ -747,10 +911,15 @@ class _LIBCPP_TYPE_VIS path { using _SourceCVT = _PathCVT<_SourceChar<_Tp> >; public: +#if defined(_LIBCPP_WIN32API) + typedef wchar_t value_type; + static constexpr value_type preferred_separator = L'\\'; +#else typedef char value_type; - typedef basic_string<value_type> string_type; - typedef _VSTD::string_view __string_view; static constexpr value_type preferred_separator = '/'; +#endif + typedef basic_string<value_type> string_type; + typedef basic_string_view<value_type> __string_view; enum class _LIBCPP_ENUM_VIS format : unsigned char { auto_format, @@ -779,12 +948,14 @@ public: _PathCVT<_ItVal>::__append_range(__pn_, __first, __last); } +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) // TODO Implement locale conversions. template <class _Source, class = _EnableIfPathable<_Source, void> > path(const _Source& __src, const locale& __loc, format = format::auto_format); template <class _InputIt> path(_InputIt __first, _InputIt _last, const locale& __loc, format = format::auto_format); +#endif _LIBCPP_INLINE_VISIBILITY ~path() = default; @@ -922,9 +1093,8 @@ public: template <class _ECharT> typename enable_if<__can_convert_char<_ECharT>::value, path&>::type operator+=(_ECharT __x) { - basic_string<_ECharT> __tmp; - __tmp += __x; - _PathCVT<_ECharT>::__append_source(__pn_, __tmp); + _PathCVT<_ECharT>::__append_source(__pn_, + basic_string_view<_ECharT>(&__x, 1)); return *this; } @@ -950,7 +1120,12 @@ public: _LIBCPP_INLINE_VISIBILITY void clear() noexcept { __pn_.clear(); } - path& make_preferred() { return *this; } + path& make_preferred() { +#if defined(_LIBCPP_WIN32API) + _VSTD::replace(__pn_.begin(), __pn_.end(), L'/', L'\\'); +#endif + return *this; + } _LIBCPP_INLINE_VISIBILITY path& remove_filename() { @@ -983,6 +1158,64 @@ public: _LIBCPP_INLINE_VISIBILITY operator string_type() const { return __pn_; } +#if defined(_LIBCPP_WIN32API) + _LIBCPP_INLINE_VISIBILITY _VSTD::wstring wstring() const { return __pn_; } + + _VSTD::wstring generic_wstring() const { return __pn_; } + +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) + template <class _ECharT, class _Traits = char_traits<_ECharT>, + class _Allocator = allocator<_ECharT> > + basic_string<_ECharT, _Traits, _Allocator> + string(const _Allocator& __a = _Allocator()) const { + using _Str = basic_string<_ECharT, _Traits, _Allocator>; + _Str __s(__a); + __s.reserve(__pn_.size()); + _PathExport<_ECharT>::__append(__s, __pn_); + return __s; + } + + _LIBCPP_INLINE_VISIBILITY _VSTD::string string() const { + return string<char>(); + } + _LIBCPP_INLINE_VISIBILITY __u8_string u8string() const { + using _CVT = __narrow_to_utf8<sizeof(wchar_t) * __CHAR_BIT__>; + __u8_string __s; + __s.reserve(__pn_.size()); + _CVT()(back_inserter(__s), __pn_.data(), __pn_.data() + __pn_.size()); + return __s; + } + + _LIBCPP_INLINE_VISIBILITY _VSTD::u16string u16string() const { + return string<char16_t>(); + } + _LIBCPP_INLINE_VISIBILITY _VSTD::u32string u32string() const { + return string<char32_t>(); + } + + // generic format observers + template <class _ECharT, class _Traits = char_traits<_ECharT>, + class _Allocator = allocator<_ECharT> > + basic_string<_ECharT, _Traits, _Allocator> + generic_string(const _Allocator& __a = _Allocator()) const { + return string<_ECharT, _Traits, _Allocator>(__a); + } + + _VSTD::string generic_string() const { return generic_string<char>(); } + _VSTD::u16string generic_u16string() const { return generic_string<char16_t>(); } + _VSTD::u32string generic_u32string() const { return generic_string<char32_t>(); } + __u8_string generic_u8string() const { return u8string(); } +#endif /* !_LIBCPP_HAS_NO_LOCALIZATION */ +#else /* _LIBCPP_WIN32API */ + + _LIBCPP_INLINE_VISIBILITY _VSTD::string string() const { return __pn_; } +#ifndef _LIBCPP_NO_HAS_CHAR8_T + _LIBCPP_INLINE_VISIBILITY _VSTD::u8string u8string() const { return _VSTD::u8string(__pn_.begin(), __pn_.end()); } +#else + _LIBCPP_INLINE_VISIBILITY _VSTD::string u8string() const { return __pn_; } +#endif + +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) template <class _ECharT, class _Traits = char_traits<_ECharT>, class _Allocator = allocator<_ECharT> > basic_string<_ECharT, _Traits, _Allocator> @@ -995,19 +1228,26 @@ public: return __s; } - _LIBCPP_INLINE_VISIBILITY std::string string() const { return __pn_; } - _LIBCPP_INLINE_VISIBILITY std::wstring wstring() const { + _LIBCPP_INLINE_VISIBILITY _VSTD::wstring wstring() const { return string<wchar_t>(); } - _LIBCPP_INLINE_VISIBILITY std::string u8string() const { return __pn_; } - _LIBCPP_INLINE_VISIBILITY std::u16string u16string() const { + _LIBCPP_INLINE_VISIBILITY _VSTD::u16string u16string() const { return string<char16_t>(); } - _LIBCPP_INLINE_VISIBILITY std::u32string u32string() const { + _LIBCPP_INLINE_VISIBILITY _VSTD::u32string u32string() const { return string<char32_t>(); } +#endif /* !_LIBCPP_HAS_NO_LOCALIZATION */ // generic format observers + _VSTD::string generic_string() const { return __pn_; } +#ifndef _LIBCPP_NO_HAS_CHAR8_T + _VSTD::u8string generic_u8string() const { return _VSTD::u8string(__pn_.begin(), __pn_.end()); } +#else + _VSTD::string generic_u8string() const { return __pn_; } +#endif + +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) template <class _ECharT, class _Traits = char_traits<_ECharT>, class _Allocator = allocator<_ECharT> > basic_string<_ECharT, _Traits, _Allocator> @@ -1015,11 +1255,11 @@ public: return string<_ECharT, _Traits, _Allocator>(__a); } - std::string generic_string() const { return __pn_; } - std::wstring generic_wstring() const { return string<wchar_t>(); } - std::string generic_u8string() const { return __pn_; } - std::u16string generic_u16string() const { return string<char16_t>(); } - std::u32string generic_u32string() const { return string<char32_t>(); } + _VSTD::wstring generic_wstring() const { return string<wchar_t>(); } + _VSTD::u16string generic_u16string() const { return string<char16_t>(); } + _VSTD::u32string generic_u32string() const { return string<char32_t>(); } +#endif /* !_LIBCPP_HAS_NO_LOCALIZATION */ +#endif /* !_LIBCPP_WIN32API */ private: int __compare(__string_view) const; @@ -1123,23 +1363,24 @@ public: iterator begin() const; iterator end() const; +#if !defined(_LIBCPP_HAS_NO_LOCALIZATION) template <class _CharT, class _Traits> _LIBCPP_INLINE_VISIBILITY friend - typename enable_if<is_same<_CharT, char>::value && - is_same<_Traits, char_traits<char> >::value, + typename enable_if<is_same<_CharT, value_type>::value && + is_same<_Traits, char_traits<value_type> >::value, basic_ostream<_CharT, _Traits>&>::type operator<<(basic_ostream<_CharT, _Traits>& __os, const path& __p) { - __os << std::__quoted(__p.native()); + __os << _VSTD::__quoted(__p.native()); return __os; } template <class _CharT, class _Traits> _LIBCPP_INLINE_VISIBILITY friend - typename enable_if<!is_same<_CharT, char>::value || - !is_same<_Traits, char_traits<char> >::value, + typename enable_if<!is_same<_CharT, value_type>::value || + !is_same<_Traits, char_traits<value_type> >::value, basic_ostream<_CharT, _Traits>&>::type operator<<(basic_ostream<_CharT, _Traits>& __os, const path& __p) { - __os << std::__quoted(__p.string<_CharT, _Traits>()); + __os << _VSTD::__quoted(__p.string<_CharT, _Traits>()); return __os; } @@ -1151,6 +1392,7 @@ public: __p = __tmp; return __is; } +#endif // !_LIBCPP_HAS_NO_LOCALIZATION friend _LIBCPP_INLINE_VISIBILITY bool operator==(const path& __lhs, const path& __rhs) noexcept { return __lhs.compare(__rhs) == 0; @@ -1193,25 +1435,70 @@ inline _LIBCPP_INLINE_VISIBILITY void swap(path& __lhs, path& __rhs) noexcept { _LIBCPP_FUNC_VIS size_t hash_value(const path& __p) noexcept; -template <class _Source> -_LIBCPP_INLINE_VISIBILITY - typename enable_if<__is_pathable<_Source>::value, path>::type - u8path(const _Source& __s) { +template <class _InputIt> +_LIBCPP_INLINE_VISIBILITY _LIBCPP_DEPRECATED_WITH_CHAR8_T + typename enable_if<__is_pathable<_InputIt>::value, path>::type + u8path(_InputIt __f, _InputIt __l) { static_assert( - is_same<typename __is_pathable<_Source>::__char_type, char>::value, - "u8path(Source const&) requires Source have a character type of type " - "'char'"); - return path(__s); +#ifndef _LIBCPP_NO_HAS_CHAR8_T + is_same<typename __is_pathable<_InputIt>::__char_type, char8_t>::value || +#endif + is_same<typename __is_pathable<_InputIt>::__char_type, char>::value, + "u8path(Iter, Iter) requires Iter have a value_type of type 'char'" + " or 'char8_t'"); +#if defined(_LIBCPP_WIN32API) + string __tmp(__f, __l); + using _CVT = __widen_from_utf8<sizeof(wchar_t) * __CHAR_BIT__>; + _VSTD::wstring __w; + __w.reserve(__tmp.size()); + _CVT()(back_inserter(__w), __tmp.data(), __tmp.data() + __tmp.size()); + return path(__w); +#else + return path(__f, __l); +#endif /* !_LIBCPP_WIN32API */ } +#if defined(_LIBCPP_WIN32API) template <class _InputIt> -_LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_DEPRECATED_WITH_CHAR8_T typename enable_if<__is_pathable<_InputIt>::value, path>::type - u8path(_InputIt __f, _InputIt __l) { + u8path(_InputIt __f, _NullSentinel) { static_assert( +#ifndef _LIBCPP_NO_HAS_CHAR8_T + is_same<typename __is_pathable<_InputIt>::__char_type, char8_t>::value || +#endif is_same<typename __is_pathable<_InputIt>::__char_type, char>::value, - "u8path(Iter, Iter) requires Iter have a value_type of type 'char'"); - return path(__f, __l); + "u8path(Iter, Iter) requires Iter have a value_type of type 'char'" + " or 'char8_t'"); + string __tmp; + const char __sentinel = char{}; + for (; *__f != __sentinel; ++__f) + __tmp.push_back(*__f); + using _CVT = __widen_from_utf8<sizeof(wchar_t) * __CHAR_BIT__>; + _VSTD::wstring __w; + __w.reserve(__tmp.size()); + _CVT()(back_inserter(__w), __tmp.data(), __tmp.data() + __tmp.size()); + return path(__w); +} +#endif /* _LIBCPP_WIN32API */ + +template <class _Source> +_LIBCPP_INLINE_VISIBILITY _LIBCPP_DEPRECATED_WITH_CHAR8_T + typename enable_if<__is_pathable<_Source>::value, path>::type + u8path(const _Source& __s) { + static_assert( +#ifndef _LIBCPP_NO_HAS_CHAR8_T + is_same<typename __is_pathable<_Source>::__char_type, char8_t>::value || +#endif + is_same<typename __is_pathable<_Source>::__char_type, char>::value, + "u8path(Source const&) requires Source have a character type of type " + "'char' or 'char8_t'"); +#if defined(_LIBCPP_WIN32API) + using _Traits = __is_pathable<_Source>; + return u8path(__unwrap_iter(_Traits::__range_begin(__s)), __unwrap_iter(_Traits::__range_end(__s))); +#else + return path(__s); +#endif } class _LIBCPP_TYPE_VIS path::iterator { @@ -1230,7 +1517,7 @@ public: typedef bidirectional_iterator_tag iterator_category; typedef path value_type; - typedef std::ptrdiff_t difference_type; + typedef ptrdiff_t difference_type; typedef const path* pointer; typedef const path& reference; @@ -1374,7 +1661,7 @@ template <class... _Args> _LIBCPP_NORETURN inline _LIBCPP_INLINE_VISIBILITY #ifndef _LIBCPP_NO_EXCEPTIONS void __throw_filesystem_error(_Args&&... __args) { - throw filesystem_error(std::forward<_Args>(__args)...); + throw filesystem_error(_VSTD::forward<_Args>(__args)...); } #else void __throw_filesystem_error(_Args&&...) { @@ -2204,7 +2491,7 @@ private: _LIBCPP_INLINE_VISIBILITY void __assign_iter_entry(_Path&& __p, __cached_data __dt) { - __p_ = std::move(__p); + __p_ = _VSTD::move(__p); __data_ = __dt; } @@ -2505,10 +2792,10 @@ end(const directory_iterator&) noexcept { class recursive_directory_iterator { public: using value_type = directory_entry; - using difference_type = std::ptrdiff_t; + using difference_type = ptrdiff_t; using pointer = directory_entry const*; using reference = directory_entry const&; - using iterator_category = std::input_iterator_tag; + using iterator_category = input_iterator_tag; public: // constructors and destructor diff --git a/lib/libcxx/include/forward_list b/lib/libcxx/include/forward_list index 3bd8db8b7d..d3d6b8238f 100644 --- a/lib/libcxx/include/forward_list +++ b/lib/libcxx/include/forward_list @@ -603,7 +603,7 @@ __forward_list_base<_Tp, _Alloc>::swap(__forward_list_base& __x) __is_nothrow_swappable<__node_allocator>::value) #endif { - __swap_allocator(__alloc(), __x.__alloc(), + _VSTD::__swap_allocator(__alloc(), __x.__alloc(), integral_constant<bool, __node_traits::propagate_on_container_swap::value>()); using _VSTD::swap; swap(__before_begin()->__next_, __x.__before_begin()->__next_); @@ -758,7 +758,7 @@ public: {return base::__before_begin()->__next_ == nullptr;} _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT { - return std::min<size_type>( + return _VSTD::min<size_type>( __node_traits::max_size(base::__alloc()), numeric_limits<difference_type>::max()); } @@ -871,7 +871,7 @@ private: #ifndef _LIBCPP_HAS_NO_DEDUCTION_GUIDES template<class _InputIterator, - class _Alloc = typename std::allocator<typename iterator_traits<_InputIterator>::value_type>, + class _Alloc = allocator<typename iterator_traits<_InputIterator>::value_type>, class = typename enable_if<__is_allocator<_Alloc>::value, void>::type > forward_list(_InputIterator, _InputIterator) diff --git a/lib/libcxx/include/fstream b/lib/libcxx/include/fstream index e9138998bf..d7d6b46c32 100644 --- a/lib/libcxx/include/fstream +++ b/lib/libcxx/include/fstream @@ -180,12 +180,16 @@ typedef basic_fstream<wchar_t> wfstream; */ #include <__config> +#include <__availability> #include <ostream> #include <istream> #include <__locale> #include <cstdio> #include <cstdlib> -#include <filesystem> + +#if !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) +# include <filesystem> +#endif #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header @@ -234,13 +238,13 @@ public: _LIBCPP_INLINE_VISIBILITY basic_filebuf* open(const string& __s, ios_base::openmode __mode); -#if _LIBCPP_STD_VER >= 17 +#if _LIBCPP_STD_VER >= 17 && !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) _LIBCPP_AVAILABILITY_FILESYSTEM _LIBCPP_INLINE_VISIBILITY basic_filebuf* open(const _VSTD_FS::path& __p, ios_base::openmode __mode) { return open(__p.c_str(), __mode); } #endif - _LIBCPP_INLINE_VISIBILITY + inline _LIBCPP_INLINE_VISIBILITY basic_filebuf* __open(int __fd, ios_base::openmode __mode); #endif basic_filebuf* close(); @@ -286,13 +290,13 @@ private: template <class _CharT, class _Traits> basic_filebuf<_CharT, _Traits>::basic_filebuf() - : __extbuf_(0), - __extbufnext_(0), - __extbufend_(0), + : __extbuf_(nullptr), + __extbufnext_(nullptr), + __extbufend_(nullptr), __ebs_(0), - __intbuf_(0), + __intbuf_(nullptr), __ibs_(0), - __file_(0), + __file_(nullptr), __cv_(nullptr), __st_(), __st_last_(), @@ -307,7 +311,7 @@ basic_filebuf<_CharT, _Traits>::basic_filebuf() __cv_ = &use_facet<codecvt<char_type, char, state_type> >(this->getloc()); __always_noconv_ = __cv_->always_noconv(); } - setbuf(0, 4096); + setbuf(nullptr, 4096); } #ifndef _LIBCPP_CXX03_LANG @@ -359,13 +363,13 @@ basic_filebuf<_CharT, _Traits>::basic_filebuf(basic_filebuf&& __rhs) (char_type*)__extbuf_ + (__rhs.gptr() - __rhs.eback()), (char_type*)__extbuf_ + (__rhs.egptr() - __rhs.eback())); } - __rhs.__extbuf_ = 0; - __rhs.__extbufnext_ = 0; - __rhs.__extbufend_ = 0; + __rhs.__extbuf_ = nullptr; + __rhs.__extbufnext_ = nullptr; + __rhs.__extbufend_ = nullptr; __rhs.__ebs_ = 0; __rhs.__intbuf_ = 0; __rhs.__ibs_ = 0; - __rhs.__file_ = 0; + __rhs.__file_ = nullptr; __rhs.__st_ = state_type(); __rhs.__st_last_ = state_type(); __rhs.__om_ = 0; @@ -499,7 +503,7 @@ inline bool basic_filebuf<_CharT, _Traits>::is_open() const { - return __file_ != 0; + return __file_ != nullptr; } template <class _CharT, class _Traits> @@ -547,8 +551,8 @@ template <class _CharT, class _Traits> basic_filebuf<_CharT, _Traits>* basic_filebuf<_CharT, _Traits>::open(const char* __s, ios_base::openmode __mode) { - basic_filebuf<_CharT, _Traits>* __rt = 0; - if (__file_ == 0) + basic_filebuf<_CharT, _Traits>* __rt = nullptr; + if (__file_ == nullptr) { if (const char* __mdstr = __make_mdstring(__mode)) { __rt = this; @@ -558,22 +562,23 @@ basic_filebuf<_CharT, _Traits>::open(const char* __s, ios_base::openmode __mode) if (__mode & ios_base::ate) { if (fseek(__file_, 0, SEEK_END)) { fclose(__file_); - __file_ = 0; - __rt = 0; + __file_ = nullptr; + __rt = nullptr; } } } else - __rt = 0; + __rt = nullptr; } } return __rt; } template <class _CharT, class _Traits> -_LIBCPP_INLINE_VISIBILITY basic_filebuf<_CharT, _Traits>* +inline _LIBCPP_INLINE_VISIBILITY +basic_filebuf<_CharT, _Traits>* basic_filebuf<_CharT, _Traits>::__open(int __fd, ios_base::openmode __mode) { - basic_filebuf<_CharT, _Traits>* __rt = 0; - if (__file_ == 0) { + basic_filebuf<_CharT, _Traits>* __rt = nullptr; + if (__file_ == nullptr) { if (const char* __mdstr = __make_mdstring(__mode)) { __rt = this; __file_ = fdopen(__fd, __mdstr); @@ -582,12 +587,12 @@ basic_filebuf<_CharT, _Traits>::__open(int __fd, ios_base::openmode __mode) { if (__mode & ios_base::ate) { if (fseek(__file_, 0, SEEK_END)) { fclose(__file_); - __file_ = 0; - __rt = 0; + __file_ = nullptr; + __rt = nullptr; } } } else - __rt = 0; + __rt = nullptr; } } return __rt; @@ -600,8 +605,8 @@ template <class _CharT, class _Traits> basic_filebuf<_CharT, _Traits>* basic_filebuf<_CharT, _Traits>::open(const wchar_t* __s, ios_base::openmode __mode) { - basic_filebuf<_CharT, _Traits>* __rt = 0; - if (__file_ == 0) + basic_filebuf<_CharT, _Traits>* __rt = nullptr; + if (__file_ == nullptr) { __rt = this; const wchar_t* __mdstr; @@ -650,7 +655,7 @@ basic_filebuf<_CharT, _Traits>::open(const wchar_t* __s, ios_base::openmode __mo __mdstr = L"a+b"; break; default: - __rt = 0; + __rt = nullptr; break; } if (__rt) @@ -664,13 +669,13 @@ basic_filebuf<_CharT, _Traits>::open(const wchar_t* __s, ios_base::openmode __mo if (fseek(__file_, 0, SEEK_END)) { fclose(__file_); - __file_ = 0; - __rt = 0; + __file_ = nullptr; + __rt = nullptr; } } } else - __rt = 0; + __rt = nullptr; } } return __rt; @@ -690,16 +695,16 @@ template <class _CharT, class _Traits> basic_filebuf<_CharT, _Traits>* basic_filebuf<_CharT, _Traits>::close() { - basic_filebuf<_CharT, _Traits>* __rt = 0; + basic_filebuf<_CharT, _Traits>* __rt = nullptr; if (__file_) { __rt = this; unique_ptr<FILE, int(*)(FILE*)> __h(__file_, fclose); if (sync()) - __rt = 0; + __rt = nullptr; if (fclose(__h.release())) - __rt = 0; - __file_ = 0; + __rt = nullptr; + __file_ = nullptr; setbuf(0, 0); } return __rt; @@ -709,17 +714,17 @@ template <class _CharT, class _Traits> typename basic_filebuf<_CharT, _Traits>::int_type basic_filebuf<_CharT, _Traits>::underflow() { - if (__file_ == 0) + if (__file_ == nullptr) return traits_type::eof(); bool __initial = __read_mode(); char_type __1buf; - if (this->gptr() == 0) + if (this->gptr() == nullptr) this->setg(&__1buf, &__1buf+1, &__1buf+1); const size_t __unget_sz = __initial ? 0 : min<size_t>((this->egptr() - this->eback()) / 2, 4); int_type __c = traits_type::eof(); if (this->gptr() == this->egptr()) { - memmove(this->eback(), this->egptr() - __unget_sz, __unget_sz * sizeof(char_type)); + _VSTD::memmove(this->eback(), this->egptr() - __unget_sz, __unget_sz * sizeof(char_type)); if (__always_noconv_) { size_t __nmemb = static_cast<size_t>(this->egptr() - this->eback() - __unget_sz); @@ -736,7 +741,7 @@ basic_filebuf<_CharT, _Traits>::underflow() { _LIBCPP_ASSERT ( !(__extbufnext_ == NULL && (__extbufend_ != __extbufnext_)), "underflow moving from NULL" ); if (__extbufend_ != __extbufnext_) - memmove(__extbuf_, __extbufnext_, __extbufend_ - __extbufnext_); + _VSTD::memmove(__extbuf_, __extbufnext_, __extbufend_ - __extbufnext_); __extbufnext_ = __extbuf_ + (__extbufend_ - __extbufnext_); __extbufend_ = __extbuf_ + (__extbuf_ == __extbuf_min_ ? sizeof(__extbuf_min_) : __ebs_); size_t __nmemb = _VSTD::min(static_cast<size_t>(__ibs_ - __unget_sz), @@ -771,7 +776,7 @@ basic_filebuf<_CharT, _Traits>::underflow() else __c = traits_type::to_int_type(*this->gptr()); if (this->eback() == &__1buf) - this->setg(0, 0, 0); + this->setg(nullptr, nullptr, nullptr); return __c; } @@ -801,7 +806,7 @@ template <class _CharT, class _Traits> typename basic_filebuf<_CharT, _Traits>::int_type basic_filebuf<_CharT, _Traits>::overflow(int_type __c) { - if (__file_ == 0) + if (__file_ == nullptr) return traits_type::eof(); __write_mode(); char_type __1buf; @@ -809,7 +814,7 @@ basic_filebuf<_CharT, _Traits>::overflow(int_type __c) char_type* __epb_save = this->epptr(); if (!traits_type::eq_int_type(__c, traits_type::eof())) { - if (this->pptr() == 0) + if (this->pptr() == nullptr) this->setp(&__1buf, &__1buf+1); *this->pptr() = traits_type::to_char_type(__c); this->pbump(1); @@ -866,8 +871,8 @@ template <class _CharT, class _Traits> basic_streambuf<_CharT, _Traits>* basic_filebuf<_CharT, _Traits>::setbuf(char_type* __s, streamsize __n) { - this->setg(0, 0, 0); - this->setp(0, 0); + this->setg(nullptr, nullptr, nullptr); + this->setp(nullptr, nullptr); if (__owns_eb_) delete [] __extbuf_; if (__owns_ib_) @@ -909,7 +914,7 @@ basic_filebuf<_CharT, _Traits>::setbuf(char_type* __s, streamsize __n) else { __ibs_ = 0; - __intbuf_ = 0; + __intbuf_ = nullptr; __owns_ib_ = false; } return this; @@ -924,7 +929,7 @@ basic_filebuf<_CharT, _Traits>::seekoff(off_type __off, ios_base::seekdir __way, __throw_bad_cast(); int __width = __cv_->encoding(); - if (__file_ == 0 || (__width <= 0 && __off != 0) || sync()) + if (__file_ == nullptr || (__width <= 0 && __off != 0) || sync()) return pos_type(off_type(-1)); // __width > 0 || __off == 0 int __whence; @@ -959,7 +964,7 @@ template <class _CharT, class _Traits> typename basic_filebuf<_CharT, _Traits>::pos_type basic_filebuf<_CharT, _Traits>::seekpos(pos_type __sp, ios_base::openmode) { - if (__file_ == 0 || sync()) + if (__file_ == nullptr || sync()) return pos_type(off_type(-1)); #if defined(_LIBCPP_HAS_NO_OFF_T_FUNCTIONS) if (fseek(__file_, __sp, SEEK_SET)) @@ -976,7 +981,7 @@ template <class _CharT, class _Traits> int basic_filebuf<_CharT, _Traits>::sync() { - if (__file_ == 0) + if (__file_ == nullptr) return 0; if (!__cv_) __throw_bad_cast(); @@ -1035,7 +1040,7 @@ basic_filebuf<_CharT, _Traits>::sync() if (__update_st) __st_ = __state; __extbufnext_ = __extbufend_ = __extbuf_; - this->setg(0, 0, 0); + this->setg(nullptr, nullptr, nullptr); __cm_ = 0; } return 0; @@ -1051,8 +1056,8 @@ basic_filebuf<_CharT, _Traits>::imbue(const locale& __loc) __always_noconv_ = __cv_->always_noconv(); if (__old_anc != __always_noconv_) { - this->setg(0, 0, 0); - this->setp(0, 0); + this->setg(nullptr, nullptr, nullptr); + this->setp(nullptr, nullptr); // invariant, char_type is char, else we couldn't get here if (__always_noconv_) // need to dump __intbuf_ { @@ -1062,7 +1067,7 @@ basic_filebuf<_CharT, _Traits>::imbue(const locale& __loc) __ebs_ = __ibs_; __extbuf_ = (char*)__intbuf_; __ibs_ = 0; - __intbuf_ = 0; + __intbuf_ = nullptr; __owns_ib_ = false; } else // need to obtain an __intbuf_. @@ -1091,7 +1096,7 @@ basic_filebuf<_CharT, _Traits>::__read_mode() { if (!(__cm_ & ios_base::in)) { - this->setp(0, 0); + this->setp(nullptr, nullptr); if (__always_noconv_) this->setg((char_type*)__extbuf_, (char_type*)__extbuf_ + __ebs_, @@ -1110,7 +1115,7 @@ basic_filebuf<_CharT, _Traits>::__write_mode() { if (!(__cm_ & ios_base::out)) { - this->setg(0, 0, 0); + this->setg(nullptr, nullptr, nullptr); if (__ebs_ > sizeof(__extbuf_min_)) { if (__always_noconv_) @@ -1120,7 +1125,7 @@ basic_filebuf<_CharT, _Traits>::__write_mode() this->setp(__intbuf_, __intbuf_ + (__ibs_ - 1)); } else - this->setp(0, 0); + this->setp(nullptr, nullptr); __cm_ = ios_base::out; } } @@ -1149,7 +1154,7 @@ public: #endif _LIBCPP_INLINE_VISIBILITY explicit basic_ifstream(const string& __s, ios_base::openmode __mode = ios_base::in); -#if _LIBCPP_STD_VER >= 17 +#if _LIBCPP_STD_VER >= 17 && !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) _LIBCPP_AVAILABILITY_FILESYSTEM _LIBCPP_INLINE_VISIBILITY explicit basic_ifstream(const filesystem::path& __p, ios_base::openmode __mode = ios_base::in) : basic_ifstream(__p.c_str(), __mode) {} @@ -1175,7 +1180,7 @@ public: void open(const wchar_t* __s, ios_base::openmode __mode = ios_base::in); #endif void open(const string& __s, ios_base::openmode __mode = ios_base::in); -#if _LIBCPP_STD_VER >= 17 +#if _LIBCPP_STD_VER >= 17 && !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) _LIBCPP_AVAILABILITY_FILESYSTEM _LIBCPP_INLINE_VISIBILITY void open(const filesystem::path& __p, ios_base::openmode __mode = ios_base::in) { @@ -1206,7 +1211,7 @@ inline basic_ifstream<_CharT, _Traits>::basic_ifstream(const char* __s, ios_base::openmode __mode) : basic_istream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode | ios_base::in) == 0) + if (__sb_.open(__s, __mode | ios_base::in) == nullptr) this->setstate(ios_base::failbit); } @@ -1216,7 +1221,7 @@ inline basic_ifstream<_CharT, _Traits>::basic_ifstream(const wchar_t* __s, ios_base::openmode __mode) : basic_istream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode | ios_base::in) == 0) + if (__sb_.open(__s, __mode | ios_base::in) == nullptr) this->setstate(ios_base::failbit); } #endif @@ -1226,7 +1231,7 @@ inline basic_ifstream<_CharT, _Traits>::basic_ifstream(const string& __s, ios_base::openmode __mode) : basic_istream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode | ios_base::in) == 0) + if (__sb_.open(__s, __mode | ios_base::in) == nullptr) this->setstate(ios_base::failbit); } #endif @@ -1363,7 +1368,7 @@ public: _LIBCPP_INLINE_VISIBILITY explicit basic_ofstream(const string& __s, ios_base::openmode __mode = ios_base::out); -#if _LIBCPP_STD_VER >= 17 +#if _LIBCPP_STD_VER >= 17 && !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) _LIBCPP_AVAILABILITY_FILESYSTEM _LIBCPP_INLINE_VISIBILITY explicit basic_ofstream(const filesystem::path& __p, ios_base::openmode __mode = ios_base::out) : basic_ofstream(__p.c_str(), __mode) {} @@ -1390,7 +1395,7 @@ public: #endif void open(const string& __s, ios_base::openmode __mode = ios_base::out); -#if _LIBCPP_STD_VER >= 17 +#if _LIBCPP_STD_VER >= 17 && !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) _LIBCPP_AVAILABILITY_FILESYSTEM _LIBCPP_INLINE_VISIBILITY void open(const filesystem::path& __p, ios_base::openmode __mode = ios_base::out) { return open(__p.c_str(), __mode); } @@ -1419,7 +1424,7 @@ inline basic_ofstream<_CharT, _Traits>::basic_ofstream(const char* __s, ios_base::openmode __mode) : basic_ostream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode | ios_base::out) == 0) + if (__sb_.open(__s, __mode | ios_base::out) == nullptr) this->setstate(ios_base::failbit); } @@ -1429,7 +1434,7 @@ inline basic_ofstream<_CharT, _Traits>::basic_ofstream(const wchar_t* __s, ios_base::openmode __mode) : basic_ostream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode | ios_base::out) == 0) + if (__sb_.open(__s, __mode | ios_base::out) == nullptr) this->setstate(ios_base::failbit); } #endif @@ -1439,7 +1444,7 @@ inline basic_ofstream<_CharT, _Traits>::basic_ofstream(const string& __s, ios_base::openmode __mode) : basic_ostream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode | ios_base::out) == 0) + if (__sb_.open(__s, __mode | ios_base::out) == nullptr) this->setstate(ios_base::failbit); } #endif @@ -1548,7 +1553,7 @@ inline void basic_ofstream<_CharT, _Traits>::close() { - if (__sb_.close() == 0) + if (__sb_.close() == nullptr) this->setstate(ios_base::failbit); } @@ -1577,7 +1582,7 @@ public: _LIBCPP_INLINE_VISIBILITY explicit basic_fstream(const string& __s, ios_base::openmode __mode = ios_base::in | ios_base::out); -#if _LIBCPP_STD_VER >= 17 +#if _LIBCPP_STD_VER >= 17 && !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) _LIBCPP_AVAILABILITY_FILESYSTEM _LIBCPP_INLINE_VISIBILITY explicit basic_fstream(const filesystem::path& __p, ios_base::openmode __mode = ios_base::in | ios_base::out) : basic_fstream(__p.c_str(), __mode) {} @@ -1605,7 +1610,7 @@ public: #endif void open(const string& __s, ios_base::openmode __mode = ios_base::in | ios_base::out); -#if _LIBCPP_STD_VER >= 17 +#if _LIBCPP_STD_VER >= 17 && !defined(_LIBCPP_HAS_NO_FILESYSTEM_LIBRARY) _LIBCPP_AVAILABILITY_FILESYSTEM _LIBCPP_INLINE_VISIBILITY void open(const filesystem::path& __p, ios_base::openmode __mode = ios_base::in|ios_base::out) { return open(__p.c_str(), __mode); } @@ -1632,7 +1637,7 @@ inline basic_fstream<_CharT, _Traits>::basic_fstream(const char* __s, ios_base::openmode __mode) : basic_iostream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode) == 0) + if (__sb_.open(__s, __mode) == nullptr) this->setstate(ios_base::failbit); } @@ -1642,7 +1647,7 @@ inline basic_fstream<_CharT, _Traits>::basic_fstream(const wchar_t* __s, ios_base::openmode __mode) : basic_iostream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode) == 0) + if (__sb_.open(__s, __mode) == nullptr) this->setstate(ios_base::failbit); } #endif @@ -1652,7 +1657,7 @@ inline basic_fstream<_CharT, _Traits>::basic_fstream(const string& __s, ios_base::openmode __mode) : basic_iostream<char_type, traits_type>(&__sb_) { - if (__sb_.open(__s, __mode) == 0) + if (__sb_.open(__s, __mode) == nullptr) this->setstate(ios_base::failbit); } #endif @@ -1752,10 +1757,16 @@ inline void basic_fstream<_CharT, _Traits>::close() { - if (__sb_.close() == 0) + if (__sb_.close() == nullptr) this->setstate(ios_base::failbit); } +#if defined(_LIBCPP_ABI_ENABLE_ADDITIONAL_IOSTREAM_EXPLICIT_INSTANTIATIONS_1) +_LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_ifstream<char>) +_LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_ofstream<char>) +_LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_filebuf<char>) +#endif + _LIBCPP_END_NAMESPACE_STD _LIBCPP_POP_MACROS diff --git a/lib/libcxx/include/functional b/lib/libcxx/include/functional index 3e9425320f..67baa5bd4b 100644 --- a/lib/libcxx/include/functional +++ b/lib/libcxx/include/functional @@ -213,7 +213,8 @@ public: template <class Predicate> // deprecated in C++17 binary_negate<Predicate> not2(const Predicate& pred); -template <class F> unspecified not_fn(F&& f); // C++17 +template <class F> +constexpr unspecified not_fn(F&& f); // C++17, constexpr in C++20 template<class T> struct is_bind_expression; template<class T> struct is_placeholder; @@ -226,11 +227,12 @@ template <class T> inline constexpr int is_placeholder_v template<class Fn, class... BoundArgs> - unspecified bind(Fn&&, BoundArgs&&...); + constexpr unspecified bind(Fn&&, BoundArgs&&...); // constexpr in C++20 template<class R, class Fn, class... BoundArgs> - unspecified bind(Fn&&, BoundArgs&&...); + constexpr unspecified bind(Fn&&, BoundArgs&&...); // constexpr in C++20 template<class F, class... Args> + constexpr // constexpr in C++20 invoke_result_t<F, Args...> invoke(F&& f, Args&&... args) // C++17 noexcept(is_nothrow_invocable_v<F, Args...>); @@ -376,7 +378,8 @@ public: template <class S, class T> const_mem_fun_ref_t<S,T> mem_fun_ref(S (T::*f)() const); // deprecated in C++11, removed in C++17 template <class S, class T, class A> const_mem_fun1_ref_t<S,T,A> mem_fun_ref(S (T::*f)(A) const); // deprecated in C++11, removed in C++17 -template<class R, class T> unspecified mem_fn(R T::*); +template<class R, class T> +constexpr unspecified mem_fn(R T::*); // constexpr in C++20 class bad_function_call : public exception @@ -470,6 +473,7 @@ template <> struct hash<bool>; template <> struct hash<char>; template <> struct hash<signed char>; template <> struct hash<unsigned char>; +template <> struct hash<char8_t>; // since C++20 template <> struct hash<char16_t>; template <> struct hash<char32_t>; template <> struct hash<wchar_t>; @@ -508,10 +512,6 @@ POLICY: For non-variadic implementations, the number of arguments is limited #include <__functional_base> -#if defined(_LIBCPP_HAS_BLOCKS_RUNTIME) && !defined(_LIBCPP_HAS_OBJC_ARC) -#include <Block.h> -#endif - #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header #endif @@ -1291,15 +1291,16 @@ private: type __f_; public: - _LIBCPP_INLINE_VISIBILITY __mem_fn(type __f) _NOEXCEPT : __f_(__f) {} + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + __mem_fn(type __f) _NOEXCEPT : __f_(__f) {} #ifndef _LIBCPP_CXX03_LANG // invoke template <class... _ArgTypes> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename __invoke_return<type, _ArgTypes...>::type operator() (_ArgTypes&&... __args) const { - return __invoke(__f_, _VSTD::forward<_ArgTypes>(__args)...); + return _VSTD::__invoke(__f_, _VSTD::forward<_ArgTypes>(__args)...); } #else @@ -1307,104 +1308,104 @@ public: _LIBCPP_INLINE_VISIBILITY typename __invoke_return0<type, _A0>::type operator() (_A0& __a0) const { - return __invoke(__f_, __a0); + return _VSTD::__invoke(__f_, __a0); } template <class _A0> _LIBCPP_INLINE_VISIBILITY typename __invoke_return0<type, _A0 const>::type operator() (_A0 const& __a0) const { - return __invoke(__f_, __a0); + return _VSTD::__invoke(__f_, __a0); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0, _A1>::type operator() (_A0& __a0, _A1& __a1) const { - return __invoke(__f_, __a0, __a1); + return _VSTD::__invoke(__f_, __a0, __a1); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0 const, _A1>::type operator() (_A0 const& __a0, _A1& __a1) const { - return __invoke(__f_, __a0, __a1); + return _VSTD::__invoke(__f_, __a0, __a1); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0, _A1 const>::type operator() (_A0& __a0, _A1 const& __a1) const { - return __invoke(__f_, __a0, __a1); + return _VSTD::__invoke(__f_, __a0, __a1); } template <class _A0, class _A1> _LIBCPP_INLINE_VISIBILITY typename __invoke_return1<type, _A0 const, _A1 const>::type operator() (_A0 const& __a0, _A1 const& __a1) const { - return __invoke(__f_, __a0, __a1); + return _VSTD::__invoke(__f_, __a0, __a1); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1, _A2>::type operator() (_A0& __a0, _A1& __a1, _A2& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1, _A2>::type operator() (_A0 const& __a0, _A1& __a1, _A2& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1 const, _A2>::type operator() (_A0& __a0, _A1 const& __a1, _A2& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1, _A2 const>::type operator() (_A0& __a0, _A1& __a1, _A2 const& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1 const, _A2>::type operator() (_A0 const& __a0, _A1 const& __a1, _A2& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1, _A2 const>::type operator() (_A0 const& __a0, _A1& __a1, _A2 const& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0, _A1 const, _A2 const>::type operator() (_A0& __a0, _A1 const& __a1, _A2 const& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } template <class _A0, class _A1, class _A2> _LIBCPP_INLINE_VISIBILITY typename __invoke_return2<type, _A0 const, _A1 const, _A2 const>::type operator() (_A0 const& __a0, _A1 const& __a1, _A2 const& __a2) const { - return __invoke(__f_, __a0, __a1, __a2); + return _VSTD::__invoke(__f_, __a0, __a1, __a2); } #endif }; template<class _Rp, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 __mem_fn<_Rp _Tp::*> mem_fn(_Rp _Tp::* __pm) _NOEXCEPT { @@ -1596,7 +1597,7 @@ public: const _Target& __target() const { return __f_; } _LIBCPP_INLINE_VISIBILITY - explicit __default_alloc_func(_Target&& __f) : __f_(std::move(__f)) {} + explicit __default_alloc_func(_Target&& __f) : __f_(_VSTD::move(__f)) {} _LIBCPP_INLINE_VISIBILITY explicit __default_alloc_func(const _Target& __f) : __f_(__f) {} @@ -1612,7 +1613,7 @@ public: __builtin_new_allocator::__holder_t __hold = __builtin_new_allocator::__allocate_type<__default_alloc_func>(1); __default_alloc_func* __res = - ::new (__hold.get()) __default_alloc_func(__f_); + ::new ((void*)__hold.get()) __default_alloc_func(__f_); (void)__hold.release(); return __res; } @@ -1703,7 +1704,7 @@ template<class _Fp, class _Alloc, class _Rp, class ..._ArgTypes> void __func<_Fp, _Alloc, _Rp(_ArgTypes...)>::__clone(__base<_Rp(_ArgTypes...)>* __p) const { - ::new (__p) __func(__f_.__target(), __f_.__get_allocator()); + ::new ((void*)__p) __func(__f_.__target(), __f_.__get_allocator()); } template<class _Fp, class _Alloc, class _Rp, class ..._ArgTypes> @@ -1739,7 +1740,7 @@ __func<_Fp, _Alloc, _Rp(_ArgTypes...)>::target(const type_info& __ti) const _NOE { if (__ti == typeid(_Fp)) return &__f_.__target(); - return (const void*)0; + return nullptr; } template<class _Fp, class _Alloc, class _Rp, class ..._ArgTypes> @@ -1769,11 +1770,11 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> public: _LIBCPP_INLINE_VISIBILITY - __value_func() _NOEXCEPT : __f_(0) {} + __value_func() _NOEXCEPT : __f_(nullptr) {} template <class _Fp, class _Alloc> _LIBCPP_INLINE_VISIBILITY __value_func(_Fp&& __f, const _Alloc& __a) - : __f_(0) + : __f_(nullptr) { typedef allocator_traits<_Alloc> __alloc_traits; typedef __function::__func<_Fp, _Alloc, _Rp(_ArgTypes...)> _Fun; @@ -1803,13 +1804,13 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> template <class _Fp, class = typename enable_if<!is_same<typename decay<_Fp>::type, __value_func>::value>::type> _LIBCPP_INLINE_VISIBILITY explicit __value_func(_Fp&& __f) - : __value_func(std::forward<_Fp>(__f), allocator<_Fp>()) {} + : __value_func(_VSTD::forward<_Fp>(__f), allocator<_Fp>()) {} _LIBCPP_INLINE_VISIBILITY __value_func(const __value_func& __f) { - if (__f.__f_ == 0) - __f_ = 0; + if (__f.__f_ == nullptr) + __f_ = nullptr; else if ((void*)__f.__f_ == &__f.__buf_) { __f_ = __as_base(&__buf_); @@ -1822,8 +1823,8 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> _LIBCPP_INLINE_VISIBILITY __value_func(__value_func&& __f) _NOEXCEPT { - if (__f.__f_ == 0) - __f_ = 0; + if (__f.__f_ == nullptr) + __f_ = nullptr; else if ((void*)__f.__f_ == &__f.__buf_) { __f_ = __as_base(&__buf_); @@ -1832,7 +1833,7 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> else { __f_ = __f.__f_; - __f.__f_ = 0; + __f.__f_ = nullptr; } } @@ -1849,8 +1850,8 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> __value_func& operator=(__value_func&& __f) { *this = nullptr; - if (__f.__f_ == 0) - __f_ = 0; + if (__f.__f_ == nullptr) + __f_ = nullptr; else if ((void*)__f.__f_ == &__f.__buf_) { __f_ = __as_base(&__buf_); @@ -1859,7 +1860,7 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> else { __f_ = __f.__f_; - __f.__f_ = 0; + __f.__f_ = nullptr; } return *this; } @@ -1868,7 +1869,7 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> __value_func& operator=(nullptr_t) { __func* __f = __f_; - __f_ = 0; + __f_ = nullptr; if ((void*)__f == &__buf_) __f->destroy(); else if (__f) @@ -1879,7 +1880,7 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> _LIBCPP_INLINE_VISIBILITY _Rp operator()(_ArgTypes&&... __args) const { - if (__f_ == 0) + if (__f_ == nullptr) __throw_bad_function_call(); return (*__f_)(_VSTD::forward<_ArgTypes>(__args)...); } @@ -1895,10 +1896,10 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> __func* __t = __as_base(&__tempbuf); __f_->__clone(__t); __f_->destroy(); - __f_ = 0; + __f_ = nullptr; __f.__f_->__clone(__as_base(&__buf_)); __f.__f_->destroy(); - __f.__f_ = 0; + __f.__f_ = nullptr; __f_ = __as_base(&__buf_); __t->__clone(__as_base(&__f.__buf_)); __t->destroy(); @@ -1923,13 +1924,13 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> } _LIBCPP_INLINE_VISIBILITY - _LIBCPP_EXPLICIT operator bool() const _NOEXCEPT { return __f_ != 0; } + _LIBCPP_EXPLICIT operator bool() const _NOEXCEPT { return __f_ != nullptr; } #ifndef _LIBCPP_NO_RTTI _LIBCPP_INLINE_VISIBILITY const std::type_info& target_type() const _NOEXCEPT { - if (__f_ == 0) + if (__f_ == nullptr) return typeid(void); return __f_->target_type(); } @@ -1937,8 +1938,8 @@ template <class _Rp, class... _ArgTypes> class __value_func<_Rp(_ArgTypes...)> template <typename _Tp> _LIBCPP_INLINE_VISIBILITY const _Tp* target() const _NOEXCEPT { - if (__f_ == 0) - return 0; + if (__f_ == nullptr) + return nullptr; return (const _Tp*)__f_->target(typeid(_Tp)); } #endif // _LIBCPP_NO_RTTI @@ -2157,7 +2158,7 @@ template <class _Rp, class... _ArgTypes> class __policy_func<_Rp(_ArgTypes...)> } else { __builtin_new_allocator::__holder_t __hold = __builtin_new_allocator::__allocate_type<_Fun>(1); - __buf_.__large = ::new (__hold.get()) _Fun(_VSTD::move(__f)); + __buf_.__large = ::new ((void*)__hold.get()) _Fun(_VSTD::move(__f)); (void)__hold.release(); } } @@ -2257,6 +2258,9 @@ template <class _Rp, class... _ArgTypes> class __policy_func<_Rp(_ArgTypes...)> #if defined(_LIBCPP_HAS_BLOCKS_RUNTIME) && !defined(_LIBCPP_HAS_OBJC_ARC) +extern "C" void *_Block_copy(const void *); +extern "C" void _Block_release(const void *); + template<class _Rp1, class ..._ArgTypes1, class _Alloc, class _Rp, class ..._ArgTypes> class __func<_Rp1(^)(_ArgTypes1...), _Alloc, _Rp(_ArgTypes...)> : public __base<_Rp(_ArgTypes...)> @@ -2267,14 +2271,14 @@ class __func<_Rp1(^)(_ArgTypes1...), _Alloc, _Rp(_ArgTypes...)> public: _LIBCPP_INLINE_VISIBILITY explicit __func(__block_type const& __f) - : __f_(__f ? Block_copy(__f) : (__block_type)0) + : __f_(reinterpret_cast<__block_type>(__f ? _Block_copy(__f) : nullptr)) { } // [TODO] add && to save on a retain _LIBCPP_INLINE_VISIBILITY explicit __func(__block_type __f, const _Alloc& /* unused */) - : __f_(__f ? Block_copy(__f) : (__block_type)0) + : __f_(reinterpret_cast<__block_type>(__f ? _Block_copy(__f) : nullptr)) { } virtual __base<_Rp(_ArgTypes...)>* __clone() const { @@ -2286,12 +2290,12 @@ public: } virtual void __clone(__base<_Rp(_ArgTypes...)>* __p) const { - ::new (__p) __func(__f_); + ::new ((void*)__p) __func(__f_); } virtual void destroy() _NOEXCEPT { if (__f_) - Block_release(__f_); + _Block_release(__f_); __f_ = 0; } @@ -2303,7 +2307,7 @@ public: } virtual _Rp operator()(_ArgTypes&& ... __arg) { - return __invoke(__f_, _VSTD::forward<_ArgTypes>(__arg)...); + return _VSTD::__invoke(__f_, _VSTD::forward<_ArgTypes>(__arg)...); } #ifndef _LIBCPP_NO_RTTI @@ -2344,9 +2348,9 @@ class _LIBCPP_TEMPLATE_VIS function<_Rp(_ArgTypes...)> template <class _Fp> struct __callable<_Fp, true> { - static const bool value = is_same<void, _Rp>::value || - is_convertible<typename __invoke_of<_Fp, _ArgTypes...>::type, - _Rp>::value; + static const bool value = is_void<_Rp>::value || + __is_core_convertible<typename __invoke_of<_Fp, _ArgTypes...>::type, + _Rp>::value; }; template <class _Fp> struct __callable<_Fp, false> @@ -2518,7 +2522,7 @@ template<class _Rp, class ..._ArgTypes> function<_Rp(_ArgTypes...)>& function<_Rp(_ArgTypes...)>::operator=(function&& __f) _NOEXCEPT { - __f_ = std::move(__f.__f_); + __f_ = _VSTD::move(__f.__f_); return *this; } @@ -2701,7 +2705,7 @@ typename _EnableIf __mu(_Ti& __ti, tuple<_Uj...>& __uj) { typedef typename __make_tuple_indices<sizeof...(_Uj)>::type __indices; - return __mu_expand(__ti, __uj, __indices()); + return _VSTD::__mu_expand(__ti, __uj, __indices()); } template <bool IsPh, class _Ti, class _Uj> @@ -2873,13 +2877,13 @@ public: !is_same<typename remove_reference<_Gp>::type, __bind>::value >::type> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 explicit __bind(_Gp&& __f, _BA&& ...__bound_args) : __f_(_VSTD::forward<_Gp>(__f)), __bound_args_(_VSTD::forward<_BA>(__bound_args)...) {} template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename __bind_return<_Fd, _Td, tuple<_Args&&...> >::type operator()(_Args&& ...__args) { @@ -2888,7 +2892,7 @@ public: } template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename __bind_return<const _Fd, const _Td, tuple<_Args&&...> >::type operator()(_Args&& ...__args) const { @@ -2918,13 +2922,13 @@ public: !is_same<typename remove_reference<_Gp>::type, __bind_r>::value >::type> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 explicit __bind_r(_Gp&& __f, _BA&& ...__bound_args) : base(_VSTD::forward<_Gp>(__f), _VSTD::forward<_BA>(__bound_args)...) {} template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename enable_if < is_convertible<typename __bind_return<_Fd, _Td, tuple<_Args&&...> >::type, @@ -2938,7 +2942,7 @@ public: } template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename enable_if < is_convertible<typename __bind_return<const _Fd, const _Td, tuple<_Args&&...> >::type, @@ -2956,7 +2960,7 @@ template<class _Rp, class _Fp, class ..._BoundArgs> struct __is_bind_expression<__bind_r<_Rp, _Fp, _BoundArgs...> > : public true_type {}; template<class _Fp, class ..._BoundArgs> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 __bind<_Fp, _BoundArgs...> bind(_Fp&& __f, _BoundArgs&&... __bound_args) { @@ -2965,7 +2969,7 @@ bind(_Fp&& __f, _BoundArgs&&... __bound_args) } template<class _Rp, class _Fp, class ..._BoundArgs> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 __bind_r<_Rp, _Fp, _BoundArgs...> bind(_Fp&& __f, _BoundArgs&&... __bound_args) { @@ -2978,7 +2982,7 @@ bind(_Fp&& __f, _BoundArgs&&... __bound_args) #if _LIBCPP_STD_VER > 14 template <class _Fn, class ..._Args> -invoke_result_t<_Fn, _Args...> +_LIBCPP_CONSTEXPR_AFTER_CXX17 invoke_result_t<_Fn, _Args...> invoke(_Fn&& __f, _Args&&... __args) noexcept(is_nothrow_invocable_v<_Fn, _Args...>) { @@ -2993,21 +2997,21 @@ public: __not_fn_imp() = delete; template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 auto operator()(_Args&& ...__args) & noexcept(noexcept(!_VSTD::invoke(__fd, _VSTD::forward<_Args>(__args)...))) -> decltype( !_VSTD::invoke(__fd, _VSTD::forward<_Args>(__args)...)) { return !_VSTD::invoke(__fd, _VSTD::forward<_Args>(__args)...); } template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 auto operator()(_Args&& ...__args) && noexcept(noexcept(!_VSTD::invoke(_VSTD::move(__fd), _VSTD::forward<_Args>(__args)...))) -> decltype( !_VSTD::invoke(_VSTD::move(__fd), _VSTD::forward<_Args>(__args)...)) { return !_VSTD::invoke(_VSTD::move(__fd), _VSTD::forward<_Args>(__args)...); } template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 auto operator()(_Args&& ...__args) const& noexcept(noexcept(!_VSTD::invoke(__fd, _VSTD::forward<_Args>(__args)...))) -> decltype( !_VSTD::invoke(__fd, _VSTD::forward<_Args>(__args)...)) @@ -3015,7 +3019,7 @@ public: template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 auto operator()(_Args&& ...__args) const&& noexcept(noexcept(!_VSTD::invoke(_VSTD::move(__fd), _VSTD::forward<_Args>(__args)...))) -> decltype( !_VSTD::invoke(_VSTD::move(__fd), _VSTD::forward<_Args>(__args)...)) @@ -3024,17 +3028,17 @@ public: private: template <class _RawFunc, class = enable_if_t<!is_same<decay_t<_RawFunc>, __not_fn_imp>::value>> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 explicit __not_fn_imp(_RawFunc&& __rf) : __fd(_VSTD::forward<_RawFunc>(__rf)) {} template <class _RawFunc> - friend inline _LIBCPP_INLINE_VISIBILITY + friend inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 __not_fn_imp<decay_t<_RawFunc>> not_fn(_RawFunc&&); }; template <class _RawFunc> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 __not_fn_imp<decay_t<_RawFunc>> not_fn(_RawFunc&& __fn) { return __not_fn_imp<decay_t<_RawFunc>>(_VSTD::forward<_RawFunc>(__fn)); } @@ -3131,13 +3135,13 @@ __search(_RandomAccessIterator1 __first1, _RandomAccessIterator1 __last1, template<class _ForwardIterator, class _BinaryPredicate = equal_to<>> class _LIBCPP_TYPE_VIS default_searcher { public: - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 default_searcher(_ForwardIterator __f, _ForwardIterator __l, _BinaryPredicate __p = _BinaryPredicate()) : __first_(__f), __last_(__l), __pred_(__p) {} template <typename _ForwardIterator2> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 pair<_ForwardIterator2, _ForwardIterator2> operator () (_ForwardIterator2 __f, _ForwardIterator2 __l) const { diff --git a/lib/libcxx/include/future b/lib/libcxx/include/future index bdf74e3055..db60ab69ec 100644 --- a/lib/libcxx/include/future +++ b/lib/libcxx/include/future @@ -362,6 +362,7 @@ template <class R, class Alloc> struct uses_allocator<packaged_task<R>, Alloc>; */ #include <__config> +#include <__availability> #include <system_error> #include <memory> #include <chrono> @@ -624,18 +625,10 @@ protected: public: template <class _Arg> -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - void set_value(_Arg&& __arg); -#else - void set_value(_Arg& __arg); -#endif + void set_value(_Arg&& __arg); template <class _Arg> -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - void set_value_at_thread_exit(_Arg&& __arg); -#else - void set_value_at_thread_exit(_Arg& __arg); -#endif + void set_value_at_thread_exit(_Arg&& __arg); _Rp move(); typename add_lvalue_reference<_Rp>::type copy(); @@ -654,16 +647,12 @@ template <class _Rp> template <class _Arg> _LIBCPP_AVAILABILITY_FUTURE void -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES __assoc_state<_Rp>::set_value(_Arg&& __arg) -#else -__assoc_state<_Rp>::set_value(_Arg& __arg) -#endif { unique_lock<mutex> __lk(this->__mut_); if (this->__has_value()) __throw_future_error(future_errc::promise_already_satisfied); - ::new(&__value_) _Rp(_VSTD::forward<_Arg>(__arg)); + ::new ((void*)&__value_) _Rp(_VSTD::forward<_Arg>(__arg)); this->__state_ |= base::__constructed | base::ready; __cv_.notify_all(); } @@ -671,16 +660,12 @@ __assoc_state<_Rp>::set_value(_Arg& __arg) template <class _Rp> template <class _Arg> void -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES __assoc_state<_Rp>::set_value_at_thread_exit(_Arg&& __arg) -#else -__assoc_state<_Rp>::set_value_at_thread_exit(_Arg& __arg) -#endif { unique_lock<mutex> __lk(this->__mut_); if (this->__has_value()) __throw_future_error(future_errc::promise_already_satisfied); - ::new(&__value_) _Rp(_VSTD::forward<_Arg>(__arg)); + ::new ((void*)&__value_) _Rp(_VSTD::forward<_Arg>(__arg)); this->__state_ |= base::__constructed; __thread_local_data()->__make_ready_at_thread_exit(this); } @@ -856,16 +841,12 @@ class _LIBCPP_AVAILABILITY_FUTURE __deferred_assoc_state _Fp __func_; public: -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY explicit __deferred_assoc_state(_Fp&& __f); -#endif virtual void __execute(); }; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp, class _Fp> inline __deferred_assoc_state<_Rp, _Fp>::__deferred_assoc_state(_Fp&& __f) @@ -874,8 +855,6 @@ __deferred_assoc_state<_Rp, _Fp>::__deferred_assoc_state(_Fp&& __f) this->__set_deferred(); } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp, class _Fp> void __deferred_assoc_state<_Rp, _Fp>::__execute() @@ -903,16 +882,12 @@ class _LIBCPP_AVAILABILITY_FUTURE __deferred_assoc_state<void, _Fp> _Fp __func_; public: -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY explicit __deferred_assoc_state(_Fp&& __f); -#endif virtual void __execute(); }; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Fp> inline __deferred_assoc_state<void, _Fp>::__deferred_assoc_state(_Fp&& __f) @@ -921,8 +896,6 @@ __deferred_assoc_state<void, _Fp>::__deferred_assoc_state(_Fp&& __f) this->__set_deferred(); } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Fp> void __deferred_assoc_state<void, _Fp>::__execute() @@ -952,16 +925,12 @@ class _LIBCPP_AVAILABILITY_FUTURE __async_assoc_state virtual void __on_zero_shared() _NOEXCEPT; public: -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY explicit __async_assoc_state(_Fp&& __f); -#endif virtual void __execute(); }; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp, class _Fp> inline __async_assoc_state<_Rp, _Fp>::__async_assoc_state(_Fp&& __f) @@ -969,8 +938,6 @@ __async_assoc_state<_Rp, _Fp>::__async_assoc_state(_Fp&& __f) { } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp, class _Fp> void __async_assoc_state<_Rp, _Fp>::__execute() @@ -1007,16 +974,12 @@ class _LIBCPP_AVAILABILITY_FUTURE __async_assoc_state<void, _Fp> virtual void __on_zero_shared() _NOEXCEPT; public: -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY explicit __async_assoc_state(_Fp&& __f); -#endif virtual void __execute(); }; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Fp> inline __async_assoc_state<void, _Fp>::__async_assoc_state(_Fp&& __f) @@ -1024,8 +987,6 @@ __async_assoc_state<void, _Fp>::__async_assoc_state(_Fp&& __f) { } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Fp> void __async_assoc_state<void, _Fp>::__execute() @@ -1062,19 +1023,11 @@ template <class _Rp> class _LIBCPP_TEMPLATE_VIS future; template <class _Rp, class _Fp> _LIBCPP_INLINE_VISIBILITY future<_Rp> -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES __make_deferred_assoc_state(_Fp&& __f); -#else -__make_deferred_assoc_state(_Fp __f); -#endif template <class _Rp, class _Fp> _LIBCPP_INLINE_VISIBILITY future<_Rp> -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES __make_async_assoc_state(_Fp&& __f); -#else -__make_async_assoc_state(_Fp __f); -#endif template <class _Rp> class _LIBCPP_TEMPLATE_VIS _LIBCPP_AVAILABILITY_FUTURE future @@ -1086,22 +1039,14 @@ class _LIBCPP_TEMPLATE_VIS _LIBCPP_AVAILABILITY_FUTURE future template <class> friend class promise; template <class> friend class shared_future; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES template <class _R1, class _Fp> friend future<_R1> __make_deferred_assoc_state(_Fp&& __f); template <class _R1, class _Fp> friend future<_R1> __make_async_assoc_state(_Fp&& __f); -#else - template <class _R1, class _Fp> - friend future<_R1> __make_deferred_assoc_state(_Fp __f); - template <class _R1, class _Fp> - friend future<_R1> __make_async_assoc_state(_Fp __f); -#endif public: _LIBCPP_INLINE_VISIBILITY future() _NOEXCEPT : __state_(nullptr) {} -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY future(future&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} @@ -1110,15 +1055,10 @@ public: _LIBCPP_INLINE_VISIBILITY future& operator=(future&& __rhs) _NOEXCEPT { - future(std::move(__rhs)).swap(*this); + future(_VSTD::move(__rhs)).swap(*this); return *this; } -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - future(const future&); - future& operator=(const future&); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES + ~future(); _LIBCPP_INLINE_VISIBILITY shared_future<_Rp> share() _NOEXCEPT; @@ -1186,22 +1126,14 @@ class _LIBCPP_TEMPLATE_VIS _LIBCPP_AVAILABILITY_FUTURE future<_Rp&> template <class> friend class promise; template <class> friend class shared_future; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES template <class _R1, class _Fp> friend future<_R1> __make_deferred_assoc_state(_Fp&& __f); template <class _R1, class _Fp> friend future<_R1> __make_async_assoc_state(_Fp&& __f); -#else - template <class _R1, class _Fp> - friend future<_R1> __make_deferred_assoc_state(_Fp __f); - template <class _R1, class _Fp> - friend future<_R1> __make_async_assoc_state(_Fp __f); -#endif public: _LIBCPP_INLINE_VISIBILITY future() _NOEXCEPT : __state_(nullptr) {} -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY future(future&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} @@ -1210,15 +1142,10 @@ public: _LIBCPP_INLINE_VISIBILITY future& operator=(future&& __rhs) _NOEXCEPT { - future(std::move(__rhs)).swap(*this); + future(_VSTD::move(__rhs)).swap(*this); return *this; } -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - future(const future&); - future& operator=(const future&); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES + ~future(); _LIBCPP_INLINE_VISIBILITY shared_future<_Rp&> share() _NOEXCEPT; @@ -1281,22 +1208,14 @@ class _LIBCPP_TYPE_VIS _LIBCPP_AVAILABILITY_FUTURE future<void> template <class> friend class promise; template <class> friend class shared_future; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES template <class _R1, class _Fp> friend future<_R1> __make_deferred_assoc_state(_Fp&& __f); template <class _R1, class _Fp> friend future<_R1> __make_async_assoc_state(_Fp&& __f); -#else - template <class _R1, class _Fp> - friend future<_R1> __make_deferred_assoc_state(_Fp __f); - template <class _R1, class _Fp> - friend future<_R1> __make_async_assoc_state(_Fp __f); -#endif public: _LIBCPP_INLINE_VISIBILITY future() _NOEXCEPT : __state_(nullptr) {} -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY future(future&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} @@ -1305,15 +1224,10 @@ public: _LIBCPP_INLINE_VISIBILITY future& operator=(future&& __rhs) _NOEXCEPT { - future(std::move(__rhs)).swap(*this); + future(_VSTD::move(__rhs)).swap(*this); return *this; } -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - future(const future&); - future& operator=(const future&); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES + ~future(); _LIBCPP_INLINE_VISIBILITY shared_future<void> share() _NOEXCEPT; @@ -1367,32 +1281,21 @@ public: promise(); template <class _Alloc> promise(allocator_arg_t, const _Alloc& __a); -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY promise(promise&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} promise(const promise& __rhs) = delete; -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - promise(const promise& __rhs); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES ~promise(); // assignment -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY promise& operator=(promise&& __rhs) _NOEXCEPT { - promise(std::move(__rhs)).swap(*this); + promise(_VSTD::move(__rhs)).swap(*this); return *this; } promise& operator=(const promise& __rhs) = delete; -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - promise& operator=(const promise& __rhs); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES + _LIBCPP_INLINE_VISIBILITY void swap(promise& __rhs) _NOEXCEPT {_VSTD::swap(__state_, __rhs.__state_);} @@ -1401,16 +1304,12 @@ public: // setting the result void set_value(const _Rp& __r); -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES void set_value(_Rp&& __r); -#endif void set_exception(exception_ptr __p); // setting the result with deferred notification void set_value_at_thread_exit(const _Rp& __r); -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES void set_value_at_thread_exit(_Rp&& __r); -#endif void set_exception_at_thread_exit(exception_ptr __p); }; @@ -1429,7 +1328,7 @@ promise<_Rp>::promise(allocator_arg_t, const _Alloc& __a0) typedef __allocator_destructor<_A2> _D2; _A2 __a(__a0); unique_ptr<_State, _D2> __hold(__a.allocate(1), _D2(__a, 1)); - ::new(static_cast<void*>(_VSTD::addressof(*__hold.get()))) _State(__a0); + ::new ((void*)_VSTD::addressof(*__hold.get())) _State(__a0); __state_ = _VSTD::addressof(*__hold.release()); } @@ -1464,8 +1363,6 @@ promise<_Rp>::set_value(const _Rp& __r) __state_->set_value(__r); } -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp> void promise<_Rp>::set_value(_Rp&& __r) @@ -1475,8 +1372,6 @@ promise<_Rp>::set_value(_Rp&& __r) __state_->set_value(_VSTD::move(__r)); } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp> void promise<_Rp>::set_exception(exception_ptr __p) @@ -1496,8 +1391,6 @@ promise<_Rp>::set_value_at_thread_exit(const _Rp& __r) __state_->set_value_at_thread_exit(__r); } -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp> void promise<_Rp>::set_value_at_thread_exit(_Rp&& __r) @@ -1507,8 +1400,6 @@ promise<_Rp>::set_value_at_thread_exit(_Rp&& __r) __state_->set_value_at_thread_exit(_VSTD::move(__r)); } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template <class _Rp> void promise<_Rp>::set_exception_at_thread_exit(exception_ptr __p) @@ -1535,32 +1426,21 @@ public: promise(); template <class _Allocator> promise(allocator_arg_t, const _Allocator& __a); -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY promise(promise&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} promise(const promise& __rhs) = delete; -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - promise(const promise& __rhs); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES ~promise(); // assignment -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY promise& operator=(promise&& __rhs) _NOEXCEPT { - promise(std::move(__rhs)).swap(*this); + promise(_VSTD::move(__rhs)).swap(*this); return *this; } promise& operator=(const promise& __rhs) = delete; -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - promise& operator=(const promise& __rhs); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES + _LIBCPP_INLINE_VISIBILITY void swap(promise& __rhs) _NOEXCEPT {_VSTD::swap(__state_, __rhs.__state_);} @@ -1591,7 +1471,7 @@ promise<_Rp&>::promise(allocator_arg_t, const _Alloc& __a0) typedef __allocator_destructor<_A2> _D2; _A2 __a(__a0); unique_ptr<_State, _D2> __hold(__a.allocate(1), _D2(__a, 1)); - ::new(static_cast<void*>(_VSTD::addressof(*__hold.get()))) _State(__a0); + ::new ((void*)_VSTD::addressof(*__hold.get())) _State(__a0); __state_ = _VSTD::addressof(*__hold.release()); } @@ -1672,32 +1552,21 @@ public: template <class _Allocator> _LIBCPP_METHOD_TEMPLATE_IMPLICIT_INSTANTIATION_VIS promise(allocator_arg_t, const _Allocator& __a); -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY promise(promise&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} promise(const promise& __rhs) = delete; -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - promise(const promise& __rhs); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES ~promise(); // assignment -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY promise& operator=(promise&& __rhs) _NOEXCEPT { - promise(std::move(__rhs)).swap(*this); + promise(_VSTD::move(__rhs)).swap(*this); return *this; } promise& operator=(const promise& __rhs) = delete; -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -private: - promise& operator=(const promise& __rhs); -public: -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES + _LIBCPP_INLINE_VISIBILITY void swap(promise& __rhs) _NOEXCEPT {_VSTD::swap(__state_, __rhs.__state_);} @@ -1721,7 +1590,7 @@ promise<void>::promise(allocator_arg_t, const _Alloc& __a0) typedef __allocator_destructor<_A2> _D2; _A2 __a(__a0); unique_ptr<_State, _D2> __hold(__a.allocate(1), _D2(__a, 1)); - ::new(static_cast<void*>(_VSTD::addressof(*__hold.get()))) _State(__a0); + ::new ((void*)_VSTD::addressof(*__hold.get())) _State(__a0); __state_ = _VSTD::addressof(*__hold.release()); } @@ -1737,8 +1606,6 @@ template <class _Rp, class _Alloc> struct _LIBCPP_TEMPLATE_VIS uses_allocator<promise<_Rp>, _Alloc> : public true_type {}; -#ifndef _LIBCPP_HAS_NO_VARIADICS - // packaged_task template<class _Fp> class __packaged_task_base; @@ -1788,7 +1655,7 @@ void __packaged_task_func<_Fp, _Alloc, _Rp(_ArgTypes...)>::__move_to( __packaged_task_base<_Rp(_ArgTypes...)>* __p) _NOEXCEPT { - ::new (__p) __packaged_task_func(_VSTD::move(__f_.first()), _VSTD::move(__f_.second())); + ::new ((void*)__p) __packaged_task_func(_VSTD::move(__f_.first()), _VSTD::move(__f_.second())); } template<class _Fp, class _Alloc, class _Rp, class ..._ArgTypes> @@ -1814,7 +1681,7 @@ template<class _Fp, class _Alloc, class _Rp, class ..._ArgTypes> _Rp __packaged_task_func<_Fp, _Alloc, _Rp(_ArgTypes...)>::operator()(_ArgTypes&& ... __arg) { - return __invoke(__f_.first(), _VSTD::forward<_ArgTypes>(__arg)...); + return _VSTD::__invoke(__f_.first(), _VSTD::forward<_ArgTypes>(__arg)...); } template <class _Callable> class __packaged_task_function; @@ -1823,6 +1690,10 @@ template<class _Rp, class ..._ArgTypes> class _LIBCPP_AVAILABILITY_FUTURE __packaged_task_function<_Rp(_ArgTypes...)> { typedef __packaged_task_base<_Rp(_ArgTypes...)> __base; + + _LIBCPP_INLINE_VISIBILITY _LIBCPP_NO_CFI + __base* __get_buf() { return (__base*)&__buf_; } + typename aligned_storage<3*sizeof(void*)>::type __buf_; __base* __f_; @@ -1856,10 +1727,10 @@ __packaged_task_function<_Rp(_ArgTypes...)>::__packaged_task_function(__packaged { if (__f.__f_ == nullptr) __f_ = nullptr; - else if (__f.__f_ == (__base*)&__f.__buf_) + else if (__f.__f_ == __f.__get_buf()) { + __f.__f_->__move_to(__get_buf()); __f_ = (__base*)&__buf_; - __f.__f_->__move_to(__f_); } else { @@ -1877,8 +1748,8 @@ __packaged_task_function<_Rp(_ArgTypes...)>::__packaged_task_function(_Fp&& __f) typedef __packaged_task_func<_FR, allocator<_FR>, _Rp(_ArgTypes...)> _FF; if (sizeof(_FF) <= sizeof(__buf_)) { + ::new ((void*)&__buf_) _FF(_VSTD::forward<_Fp>(__f)); __f_ = (__base*)&__buf_; - ::new (__f_) _FF(_VSTD::forward<_Fp>(__f)); } else { @@ -1886,7 +1757,7 @@ __packaged_task_function<_Rp(_ArgTypes...)>::__packaged_task_function(_Fp&& __f) _Ap __a; typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (__hold.get()) _FF(_VSTD::forward<_Fp>(__f), allocator<_FR>(__a)); + ::new ((void*)__hold.get()) _FF(_VSTD::forward<_Fp>(__f), allocator<_FR>(__a)); __f_ = __hold.release(); } } @@ -1902,7 +1773,7 @@ __packaged_task_function<_Rp(_ArgTypes...)>::__packaged_task_function( if (sizeof(_FF) <= sizeof(__buf_)) { __f_ = (__base*)&__buf_; - ::new (__f_) _FF(_VSTD::forward<_Fp>(__f)); + ::new ((void*)__f_) _FF(_VSTD::forward<_Fp>(__f)); } else { @@ -1910,7 +1781,7 @@ __packaged_task_function<_Rp(_ArgTypes...)>::__packaged_task_function( _Ap __a(__a0); typedef __allocator_destructor<_Ap> _Dp; unique_ptr<__base, _Dp> __hold(__a.allocate(1), _Dp(__a, 1)); - ::new (static_cast<void*>(_VSTD::addressof(*__hold.get()))) + ::new ((void*)_VSTD::addressof(*__hold.get())) _FF(_VSTD::forward<_Fp>(__f), _Alloc(__a)); __f_ = _VSTD::addressof(*__hold.release()); } @@ -1920,17 +1791,17 @@ template<class _Rp, class ..._ArgTypes> __packaged_task_function<_Rp(_ArgTypes...)>& __packaged_task_function<_Rp(_ArgTypes...)>::operator=(__packaged_task_function&& __f) _NOEXCEPT { - if (__f_ == (__base*)&__buf_) + if (__f_ == __get_buf()) __f_->destroy(); else if (__f_) __f_->destroy_deallocate(); __f_ = nullptr; if (__f.__f_ == nullptr) __f_ = nullptr; - else if (__f.__f_ == (__base*)&__f.__buf_) + else if (__f.__f_ == __f.__get_buf()) { - __f_ = (__base*)&__buf_; - __f.__f_->__move_to(__f_); + __f.__f_->__move_to(__get_buf()); + __f_ = __get_buf(); } else { @@ -1943,13 +1814,14 @@ __packaged_task_function<_Rp(_ArgTypes...)>::operator=(__packaged_task_function& template<class _Rp, class ..._ArgTypes> __packaged_task_function<_Rp(_ArgTypes...)>::~__packaged_task_function() { - if (__f_ == (__base*)&__buf_) + if (__f_ == __get_buf()) __f_->destroy(); else if (__f_) __f_->destroy_deallocate(); } template<class _Rp, class ..._ArgTypes> +_LIBCPP_NO_CFI void __packaged_task_function<_Rp(_ArgTypes...)>::swap(__packaged_task_function& __f) _NOEXCEPT { @@ -2268,11 +2140,7 @@ struct _LIBCPP_TEMPLATE_VIS uses_allocator<packaged_task<_Callable>, _Alloc> template <class _Rp, class _Fp> _LIBCPP_INLINE_VISIBILITY future<_Rp> -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES __make_deferred_assoc_state(_Fp&& __f) -#else -__make_deferred_assoc_state(_Fp __f) -#endif { unique_ptr<__deferred_assoc_state<_Rp, _Fp>, __release_shared_count> __h(new __deferred_assoc_state<_Rp, _Fp>(_VSTD::forward<_Fp>(__f))); @@ -2281,11 +2149,7 @@ __make_deferred_assoc_state(_Fp __f) template <class _Rp, class _Fp> _LIBCPP_INLINE_VISIBILITY future<_Rp> -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES __make_async_assoc_state(_Fp&& __f) -#else -__make_async_assoc_state(_Fp __f) -#endif { unique_ptr<__async_assoc_state<_Rp, _Fp>, __release_shared_count> __h(new __async_assoc_state<_Rp, _Fp>(_VSTD::forward<_Fp>(__f))); @@ -2293,6 +2157,8 @@ __make_async_assoc_state(_Fp __f) return future<_Rp>(__h.get()); } +#ifndef _LIBCPP_CXX03_LANG + template <class _Fp, class... _Args> class _LIBCPP_HIDDEN __async_func { @@ -2318,7 +2184,7 @@ private: _Rp __execute(__tuple_indices<_Indices...>) { - return __invoke(_VSTD::move(_VSTD::get<0>(__f_)), _VSTD::move(_VSTD::get<_Indices>(__f_))...); + return _VSTD::__invoke(_VSTD::move(_VSTD::get<0>(__f_)), _VSTD::move(_VSTD::get<_Indices>(__f_))...); } }; @@ -2338,16 +2204,16 @@ async(launch __policy, _Fp&& __f, _Args&&... __args) { #endif if (__does_policy_contain(__policy, launch::async)) - return _VSTD::__make_async_assoc_state<_Rp>(_BF(__decay_copy(_VSTD::forward<_Fp>(__f)), - __decay_copy(_VSTD::forward<_Args>(__args))...)); + return _VSTD::__make_async_assoc_state<_Rp>(_BF(_VSTD::__decay_copy(_VSTD::forward<_Fp>(__f)), + _VSTD::__decay_copy(_VSTD::forward<_Args>(__args))...)); #ifndef _LIBCPP_NO_EXCEPTIONS } catch ( ... ) { if (__policy == launch::async) throw ; } #endif if (__does_policy_contain(__policy, launch::deferred)) - return _VSTD::__make_deferred_assoc_state<_Rp>(_BF(__decay_copy(_VSTD::forward<_Fp>(__f)), - __decay_copy(_VSTD::forward<_Args>(__args))...)); + return _VSTD::__make_deferred_assoc_state<_Rp>(_BF(_VSTD::__decay_copy(_VSTD::forward<_Fp>(__f)), + _VSTD::__decay_copy(_VSTD::forward<_Args>(__args))...)); return future<_Rp>{}; } @@ -2360,7 +2226,7 @@ async(_Fp&& __f, _Args&&... __args) _VSTD::forward<_Args>(__args)...); } -#endif // _LIBCPP_HAS_NO_VARIADICS +#endif // C++03 // shared_future @@ -2375,24 +2241,20 @@ public: _LIBCPP_INLINE_VISIBILITY shared_future(const shared_future& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {if (__state_) __state_->__add_shared();} -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_future(future<_Rp>&& __f) _NOEXCEPT : __state_(__f.__state_) {__f.__state_ = nullptr;} _LIBCPP_INLINE_VISIBILITY shared_future(shared_future&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES ~shared_future(); shared_future& operator=(const shared_future& __rhs) _NOEXCEPT; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_future& operator=(shared_future&& __rhs) _NOEXCEPT { - shared_future(std::move(__rhs)).swap(*this); + shared_future(_VSTD::move(__rhs)).swap(*this); return *this; } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES // retrieving the value _LIBCPP_INLINE_VISIBILITY @@ -2449,24 +2311,20 @@ public: _LIBCPP_INLINE_VISIBILITY shared_future(const shared_future& __rhs) : __state_(__rhs.__state_) {if (__state_) __state_->__add_shared();} -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_future(future<_Rp&>&& __f) _NOEXCEPT : __state_(__f.__state_) {__f.__state_ = nullptr;} _LIBCPP_INLINE_VISIBILITY shared_future(shared_future&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES ~shared_future(); shared_future& operator=(const shared_future& __rhs); -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_future& operator=(shared_future&& __rhs) _NOEXCEPT { - shared_future(std::move(__rhs)).swap(*this); + shared_future(_VSTD::move(__rhs)).swap(*this); return *this; } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES // retrieving the value _LIBCPP_INLINE_VISIBILITY @@ -2523,24 +2381,20 @@ public: _LIBCPP_INLINE_VISIBILITY shared_future(const shared_future& __rhs) : __state_(__rhs.__state_) {if (__state_) __state_->__add_shared();} -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_future(future<void>&& __f) _NOEXCEPT : __state_(__f.__state_) {__f.__state_ = nullptr;} _LIBCPP_INLINE_VISIBILITY shared_future(shared_future&& __rhs) _NOEXCEPT : __state_(__rhs.__state_) {__rhs.__state_ = nullptr;} -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES ~shared_future(); shared_future& operator=(const shared_future& __rhs); -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_future& operator=(shared_future&& __rhs) _NOEXCEPT { - shared_future(std::move(__rhs)).swap(*this); + shared_future(_VSTD::move(__rhs)).swap(*this); return *this; } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES // retrieving the value _LIBCPP_INLINE_VISIBILITY @@ -2591,8 +2445,6 @@ future<_Rp&>::share() _NOEXCEPT return shared_future<_Rp&>(_VSTD::move(*this)); } -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - inline shared_future<void> future<void>::share() _NOEXCEPT @@ -2600,8 +2452,6 @@ future<void>::share() _NOEXCEPT return shared_future<void>(_VSTD::move(*this)); } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - _LIBCPP_END_NAMESPACE_STD #endif // !_LIBCPP_HAS_NO_THREADS diff --git a/lib/libcxx/include/iomanip b/lib/libcxx/include/iomanip index 3f78f4d02b..4b6418bc0b 100644 --- a/lib/libcxx/include/iomanip +++ b/lib/libcxx/include/iomanip @@ -514,7 +514,7 @@ put_time(const tm* __tm, const _CharT* __fmt) } template <class _CharT, class _Traits, class _ForwardIterator> -std::basic_ostream<_CharT, _Traits> & +basic_ostream<_CharT, _Traits> & __quoted_output ( basic_ostream<_CharT, _Traits> &__os, _ForwardIterator __first, _ForwardIterator __last, _CharT __delim, _CharT __escape ) { @@ -527,7 +527,7 @@ __quoted_output ( basic_ostream<_CharT, _Traits> &__os, __str.push_back(*__first); } __str.push_back(__delim); - return __put_character_sequence(__os, __str.data(), __str.size()); + return _VSTD::__put_character_sequence(__os, __str.data(), __str.size()); } template <class _CharT, class _Traits, class _String> diff --git a/lib/libcxx/include/ios b/lib/libcxx/include/ios index 7f0e2d65e6..653e3a95fc 100644 --- a/lib/libcxx/include/ios +++ b/lib/libcxx/include/ios @@ -710,7 +710,7 @@ void basic_ios<_CharT, _Traits>::init(basic_streambuf<char_type, traits_type>* __sb) { ios_base::init(__sb); - __tie_ = 0; + __tie_ = nullptr; __fill_ = traits_type::eof(); } @@ -821,7 +821,7 @@ basic_ios<_CharT, _Traits>::move(basic_ios& __rhs) { ios_base::move(__rhs); __tie_ = __rhs.__tie_; - __rhs.__tie_ = 0; + __rhs.__tie_ = nullptr; __fill_ = __rhs.__fill_; } @@ -1035,33 +1035,6 @@ defaultfloat(ios_base& __str) return __str; } -template <class _CharT, class _Traits> -class __save_flags -{ - typedef basic_ios<_CharT, _Traits> __stream_type; - typedef typename __stream_type::fmtflags fmtflags; - - __stream_type& __stream_; - fmtflags __fmtflags_; - _CharT __fill_; - - __save_flags(const __save_flags&); - __save_flags& operator=(const __save_flags&); -public: - _LIBCPP_INLINE_VISIBILITY - explicit __save_flags(__stream_type& __stream) - : __stream_(__stream), - __fmtflags_(__stream.flags()), - __fill_(__stream.fill()) - {} - _LIBCPP_INLINE_VISIBILITY - ~__save_flags() - { - __stream_.flags(__fmtflags_); - __stream_.fill(__fill_); - } -}; - _LIBCPP_END_NAMESPACE_STD #endif // _LIBCPP_IOS diff --git a/lib/libcxx/include/iosfwd b/lib/libcxx/include/iosfwd index 0ffe75f197..0a0de99ff3 100644 --- a/lib/libcxx/include/iosfwd +++ b/lib/libcxx/include/iosfwd @@ -185,6 +185,36 @@ typedef basic_ifstream<wchar_t> wifstream; typedef basic_ofstream<wchar_t> wofstream; typedef basic_fstream<wchar_t> wfstream; +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(ios) _LIBCPP_PREFERRED_NAME(wios) basic_ios; + +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(streambuf) _LIBCPP_PREFERRED_NAME(wstreambuf) basic_streambuf; +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(istream) _LIBCPP_PREFERRED_NAME(wistream) basic_istream; +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(ostream) _LIBCPP_PREFERRED_NAME(wostream) basic_ostream; +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(iostream) _LIBCPP_PREFERRED_NAME(wiostream) basic_iostream; + +template <class _CharT, class _Traits, class _Allocator> + class _LIBCPP_PREFERRED_NAME(stringbuf) _LIBCPP_PREFERRED_NAME(wstringbuf) basic_stringbuf; +template <class _CharT, class _Traits, class _Allocator> + class _LIBCPP_PREFERRED_NAME(istringstream) _LIBCPP_PREFERRED_NAME(wistringstream) basic_istringstream; +template <class _CharT, class _Traits, class _Allocator> + class _LIBCPP_PREFERRED_NAME(ostringstream) _LIBCPP_PREFERRED_NAME(wostringstream) basic_ostringstream; +template <class _CharT, class _Traits, class _Allocator> + class _LIBCPP_PREFERRED_NAME(stringstream) _LIBCPP_PREFERRED_NAME(wstringstream) basic_stringstream; + +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(filebuf) _LIBCPP_PREFERRED_NAME(wfilebuf) basic_filebuf; +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(ifstream) _LIBCPP_PREFERRED_NAME(wifstream) basic_ifstream; +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(ofstream) _LIBCPP_PREFERRED_NAME(wofstream) basic_ofstream; +template <class _CharT, class _Traits> + class _LIBCPP_PREFERRED_NAME(fstream) _LIBCPP_PREFERRED_NAME(wfstream) basic_fstream; + template <class _State> class _LIBCPP_TEMPLATE_VIS fpos; typedef fpos<mbstate_t> streampos; typedef fpos<mbstate_t> wstreampos; @@ -210,11 +240,40 @@ template <class _CharT, // for <stdexcept> typedef basic_string<char, char_traits<char>, allocator<char> > string; typedef basic_string<wchar_t, char_traits<wchar_t>, allocator<wchar_t> > wstring; +template <class _CharT, class _Traits, class _Allocator> + class _LIBCPP_PREFERRED_NAME(string) _LIBCPP_PREFERRED_NAME(wstring) basic_string; // Include other forward declarations here template <class _Tp, class _Alloc = allocator<_Tp> > class _LIBCPP_TEMPLATE_VIS vector; +template <class _CharT, class _Traits> +class __save_flags +{ + typedef basic_ios<_CharT, _Traits> __stream_type; + typedef typename __stream_type::fmtflags fmtflags; + + __stream_type& __stream_; + fmtflags __fmtflags_; + _CharT __fill_; + + __save_flags(const __save_flags&); + __save_flags& operator=(const __save_flags&); +public: + _LIBCPP_INLINE_VISIBILITY + explicit __save_flags(__stream_type& __stream) + : __stream_(__stream), + __fmtflags_(__stream.flags()), + __fill_(__stream.fill()) + {} + _LIBCPP_INLINE_VISIBILITY + ~__save_flags() + { + __stream_.flags(__fmtflags_); + __stream_.fill(__fill_); + } +}; + _LIBCPP_END_NAMESPACE_STD #endif // _LIBCPP_IOSFWD diff --git a/lib/libcxx/include/istream b/lib/libcxx/include/istream index bfbe5f2472..5e984909f8 100644 --- a/lib/libcxx/include/istream +++ b/lib/libcxx/include/istream @@ -150,9 +150,9 @@ template <class charT, class traits> basic_istream<charT,traits>& ws(basic_istream<charT,traits>& is); -template <class charT, class traits, class T> - basic_istream<charT, traits>& - operator>>(basic_istream<charT, traits>&& is, T& x); +// rvalue stream extraction +template <class Stream, class T> + Stream&& operator>>(Stream&& is, T&& x); } // std @@ -1142,7 +1142,7 @@ basic_istream<_CharT, _Traits>::putback(char_type __c) try { #endif // _LIBCPP_NO_EXCEPTIONS - if (this->rdbuf() == 0 || this->rdbuf()->sputbackc(__c) == traits_type::eof()) + if (this->rdbuf() == nullptr || this->rdbuf()->sputbackc(__c) == traits_type::eof()) __state |= ios_base::badbit; #ifndef _LIBCPP_NO_EXCEPTIONS } @@ -1179,7 +1179,7 @@ basic_istream<_CharT, _Traits>::unget() try { #endif // _LIBCPP_NO_EXCEPTIONS - if (this->rdbuf() == 0 || this->rdbuf()->sungetc() == traits_type::eof()) + if (this->rdbuf() == nullptr || this->rdbuf()->sungetc() == traits_type::eof()) __state |= ios_base::badbit; #ifndef _LIBCPP_NO_EXCEPTIONS } @@ -1215,7 +1215,7 @@ basic_istream<_CharT, _Traits>::sync() try { #endif // _LIBCPP_NO_EXCEPTIONS - if (this->rdbuf() == 0) + if (this->rdbuf() == nullptr) return -1; if (this->rdbuf()->pubsync() == -1) { @@ -1378,13 +1378,23 @@ ws(basic_istream<_CharT, _Traits>& __is) #ifndef _LIBCPP_CXX03_LANG -template <class _CharT, class _Traits, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY -basic_istream<_CharT, _Traits>& -operator>>(basic_istream<_CharT, _Traits>&& __is, _Tp&& __x) +template <class _Stream, class _Tp, class = void> +struct __is_istreamable : false_type { }; + +template <class _Stream, class _Tp> +struct __is_istreamable<_Stream, _Tp, decltype( + _VSTD::declval<_Stream>() >> _VSTD::declval<_Tp>(), void() +)> : true_type { }; + +template <class _Stream, class _Tp, class = typename enable_if< + _And<is_base_of<ios_base, _Stream>, + __is_istreamable<_Stream&, _Tp&&>>::value +>::type> +_LIBCPP_INLINE_VISIBILITY +_Stream&& operator>>(_Stream&& __is, _Tp&& __x) { __is >> _VSTD::forward<_Tp>(__x); - return __is; + return _VSTD::move(__is); } #endif // _LIBCPP_CXX03_LANG @@ -1638,11 +1648,9 @@ operator>>(basic_istream<_CharT, _Traits>& __is, bitset<_Size>& __x) return __is; } -#ifndef _LIBCPP_DO_NOT_ASSUME_STREAMS_EXPLICIT_INSTANTIATION_IN_DYLIB _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_istream<char>) _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_istream<wchar_t>) _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_iostream<char>) -#endif _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/include/iterator b/lib/libcxx/include/iterator index a13214fca5..00a3451e7a 100644 --- a/lib/libcxx/include/iterator +++ b/lib/libcxx/include/iterator @@ -420,10 +420,8 @@ template <class E> constexpr const E* data(initializer_list<E> il) noexcept; #include <type_traits> #include <cstddef> #include <initializer_list> +#include <__memory/base.h> #include <version> -#ifdef __APPLE__ -#include <Availability.h> -#endif #include <__debug> @@ -498,12 +496,11 @@ struct __has_iterator_typedefs private: struct __two {char __lx; char __lxx;}; template <class _Up> static __two __test(...); - template <class _Up> static char __test(typename std::__void_t<typename _Up::iterator_category>::type* = 0, - typename std::__void_t<typename _Up::difference_type>::type* = 0, - typename std::__void_t<typename _Up::value_type>::type* = 0, - typename std::__void_t<typename _Up::reference>::type* = 0, - typename std::__void_t<typename _Up::pointer>::type* = 0 - ); + template <class _Up> static char __test(typename __void_t<typename _Up::iterator_category>::type* = 0, + typename __void_t<typename _Up::difference_type>::type* = 0, + typename __void_t<typename _Up::value_type>::type* = 0, + typename __void_t<typename _Up::reference>::type* = 0, + typename __void_t<typename _Up::pointer>::type* = 0); public: static const bool value = sizeof(__test<_Tp>(0,0,0,0,0)) == 1; }; @@ -515,9 +512,9 @@ struct __has_iterator_category private: struct __two {char __lx; char __lxx;}; template <class _Up> static __two __test(...); - template <class _Up> static char __test(typename _Up::iterator_category* = 0); + template <class _Up> static char __test(typename _Up::iterator_category* = nullptr); public: - static const bool value = sizeof(__test<_Tp>(0)) == 1; + static const bool value = sizeof(__test<_Tp>(nullptr)) == 1; }; template <class _Iter, bool> struct __iterator_traits_impl {}; @@ -667,9 +664,9 @@ void advance(_InputIter& __i, _Distance __orig_n) { _LIBCPP_ASSERT(__orig_n >= 0 || __is_cpp17_bidirectional_iterator<_InputIter>::value, "Attempt to advance(it, n) with negative n on a non-bidirectional iterator"); - typedef decltype(__convert_to_integral(__orig_n)) _IntegralSize; + typedef decltype(_VSTD::__convert_to_integral(__orig_n)) _IntegralSize; _IntegralSize __n = __orig_n; - __advance(__i, __n, typename iterator_traits<_InputIter>::iterator_category()); + _VSTD::__advance(__i, __n, typename iterator_traits<_InputIter>::iterator_category()); } template <class _InputIter> @@ -696,7 +693,7 @@ inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX14 typename iterator_traits<_InputIter>::difference_type distance(_InputIter __first, _InputIter __last) { - return __distance(__first, __last, typename iterator_traits<_InputIter>::iterator_category()); + return _VSTD::__distance(__first, __last, typename iterator_traits<_InputIter>::iterator_category()); } template <class _InputIter> @@ -998,11 +995,11 @@ private: istream_type* __in_stream_; _Tp __value_; public: - _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR istream_iterator() : __in_stream_(0), __value_() {} + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR istream_iterator() : __in_stream_(nullptr), __value_() {} _LIBCPP_INLINE_VISIBILITY istream_iterator(istream_type& __s) : __in_stream_(_VSTD::addressof(__s)) { if (!(*__in_stream_ >> __value_)) - __in_stream_ = 0; + __in_stream_ = nullptr; } _LIBCPP_INLINE_VISIBILITY const _Tp& operator*() const {return __value_;} @@ -1010,7 +1007,7 @@ public: _LIBCPP_INLINE_VISIBILITY istream_iterator& operator++() { if (!(*__in_stream_ >> __value_)) - __in_stream_ = 0; + __in_stream_ = nullptr; return *this; } _LIBCPP_INLINE_VISIBILITY istream_iterator operator++(int) @@ -1052,15 +1049,25 @@ class _LIBCPP_TEMPLATE_VIS ostream_iterator : public iterator<output_iterator_tag, void, void, void, void> { public: - typedef _CharT char_type; - typedef _Traits traits_type; - typedef basic_ostream<_CharT,_Traits> ostream_type; + typedef output_iterator_tag iterator_category; + typedef void value_type; +#if _LIBCPP_STD_VER > 17 + typedef std::ptrdiff_t difference_type; +#else + typedef void difference_type; +#endif + typedef void pointer; + typedef void reference; + typedef _CharT char_type; + typedef _Traits traits_type; + typedef basic_ostream<_CharT, _Traits> ostream_type; + private: ostream_type* __out_stream_; const char_type* __delim_; public: _LIBCPP_INLINE_VISIBILITY ostream_iterator(ostream_type& __s) _NOEXCEPT - : __out_stream_(_VSTD::addressof(__s)), __delim_(0) {} + : __out_stream_(_VSTD::addressof(__s)), __delim_(nullptr) {} _LIBCPP_INLINE_VISIBILITY ostream_iterator(ostream_type& __s, const _CharT* __delimiter) _NOEXCEPT : __out_stream_(_VSTD::addressof(__s)), __delim_(__delimiter) {} _LIBCPP_INLINE_VISIBILITY ostream_iterator& operator=(const _Tp& __value_) @@ -1106,11 +1113,11 @@ private: bool __test_for_eof() const { if (__sbuf_ && traits_type::eq_int_type(__sbuf_->sgetc(), traits_type::eof())) - __sbuf_ = 0; - return __sbuf_ == 0; + __sbuf_ = nullptr; + return __sbuf_ == nullptr; } public: - _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR istreambuf_iterator() _NOEXCEPT : __sbuf_(0) {} + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR istreambuf_iterator() _NOEXCEPT : __sbuf_(nullptr) {} _LIBCPP_INLINE_VISIBILITY istreambuf_iterator(istream_type& __s) _NOEXCEPT : __sbuf_(__s.rdbuf()) {} _LIBCPP_INLINE_VISIBILITY istreambuf_iterator(streambuf_type* __s) _NOEXCEPT @@ -1151,10 +1158,20 @@ class _LIBCPP_TEMPLATE_VIS ostreambuf_iterator : public iterator<output_iterator_tag, void, void, void, void> { public: - typedef _CharT char_type; - typedef _Traits traits_type; - typedef basic_streambuf<_CharT,_Traits> streambuf_type; - typedef basic_ostream<_CharT,_Traits> ostream_type; + typedef output_iterator_tag iterator_category; + typedef void value_type; +#if _LIBCPP_STD_VER > 17 + typedef std::ptrdiff_t difference_type; +#else + typedef void difference_type; +#endif + typedef void pointer; + typedef void reference; + typedef _CharT char_type; + typedef _Traits traits_type; + typedef basic_streambuf<_CharT, _Traits> streambuf_type; + typedef basic_ostream<_CharT, _Traits> ostream_type; + private: streambuf_type* __sbuf_; public: @@ -1165,13 +1182,13 @@ public: _LIBCPP_INLINE_VISIBILITY ostreambuf_iterator& operator=(_CharT __c) { if (__sbuf_ && traits_type::eq_int_type(__sbuf_->sputc(__c), traits_type::eof())) - __sbuf_ = 0; + __sbuf_ = nullptr; return *this; } _LIBCPP_INLINE_VISIBILITY ostreambuf_iterator& operator*() {return *this;} _LIBCPP_INLINE_VISIBILITY ostreambuf_iterator& operator++() {return *this;} _LIBCPP_INLINE_VISIBILITY ostreambuf_iterator& operator++(int) {return *this;} - _LIBCPP_INLINE_VISIBILITY bool failed() const _NOEXCEPT {return __sbuf_ == 0;} + _LIBCPP_INLINE_VISIBILITY bool failed() const _NOEXCEPT {return __sbuf_ == nullptr;} template <class _Ch, class _Tr> friend @@ -1373,13 +1390,13 @@ operator+(typename __wrap_iter<_Iter>::difference_type, __wrap_iter<_Iter>) _NOE template <class _Ip, class _Op> _Op _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 copy(_Ip, _Ip, _Op); template <class _B1, class _B2> _B2 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 copy_backward(_B1, _B1, _B2); -template <class _Ip, class _Op> _Op _LIBCPP_INLINE_VISIBILITY move(_Ip, _Ip, _Op); -template <class _B1, class _B2> _B2 _LIBCPP_INLINE_VISIBILITY move_backward(_B1, _B1, _B2); +template <class _Ip, class _Op> _Op _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 move(_Ip, _Ip, _Op); +template <class _B1, class _B2> _B2 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 move_backward(_B1, _B1, _B2); #if _LIBCPP_DEBUG_LEVEL < 2 template <class _Tp> -_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1390,7 +1407,7 @@ __unwrap_iter(__wrap_iter<_Tp*>); #else template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1418,20 +1435,20 @@ public: : __i{} #endif { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_i(this); #endif } template <class _Up> _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG __wrap_iter(const __wrap_iter<_Up>& __u, - typename enable_if<is_convertible<_Up, iterator_type>::value>::type* = 0) _NOEXCEPT + typename enable_if<is_convertible<_Up, iterator_type>::value>::type* = nullptr) _NOEXCEPT : __i(__u.base()) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__iterator_copy(this, &__u); #endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG __wrap_iter(const __wrap_iter& __x) : __i(__x.base()) @@ -1456,7 +1473,7 @@ public: #endif _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG reference operator*() const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to dereference a non-dereferenceable iterator"); #endif @@ -1464,7 +1481,7 @@ public: } _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG pointer operator->() const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to dereference a non-dereferenceable iterator"); #endif @@ -1472,7 +1489,7 @@ public: } _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG __wrap_iter& operator++() _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to increment non-incrementable iterator"); #endif @@ -1484,7 +1501,7 @@ public: _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG __wrap_iter& operator--() _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__decrementable(this), "Attempted to decrement non-decrementable iterator"); #endif @@ -1497,7 +1514,7 @@ public: {__wrap_iter __w(*this); __w += __n; return __w;} _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG __wrap_iter& operator+=(difference_type __n) _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__addable(this, __n), "Attempted to add/subtract iterator outside of valid range"); #endif @@ -1510,7 +1527,7 @@ public: {*this += -__n; return *this;} _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG reference operator[](difference_type __n) const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__subscriptable(this, __n), "Attempted to subscript iterator outside of valid range"); #endif @@ -1520,7 +1537,7 @@ public: _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG iterator_type base() const _NOEXCEPT {return __i;} private: -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG __wrap_iter(const void* __p, iterator_type __x) : __i(__x) { __get_db()->__insert_ic(this, __p); @@ -1584,12 +1601,12 @@ private: template <class _Ip, class _Op> friend _LIBCPP_CONSTEXPR_AFTER_CXX17 _Op copy(_Ip, _Ip, _Op); template <class _B1, class _B2> friend _LIBCPP_CONSTEXPR_AFTER_CXX17 _B2 copy_backward(_B1, _B1, _B2); - template <class _Ip, class _Op> friend _Op move(_Ip, _Ip, _Op); - template <class _B1, class _B2> friend _B2 move_backward(_B1, _B1, _B2); + template <class _Ip, class _Op> friend _LIBCPP_CONSTEXPR_AFTER_CXX17 _Op move(_Ip, _Ip, _Op); + template <class _B1, class _B2> friend _LIBCPP_CONSTEXPR_AFTER_CXX17 _B2 move_backward(_B1, _B1, _B2); #if _LIBCPP_DEBUG_LEVEL < 2 template <class _Tp> - _LIBCPP_CONSTEXPR_IF_NODEBUG friend + _LIBCPP_CONSTEXPR friend typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1598,7 +1615,7 @@ private: __unwrap_iter(__wrap_iter<_Tp*>); #else template <class _Tp> - inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG + inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR friend typename enable_if < is_trivially_copy_assignable<_Tp>::value, @@ -1621,7 +1638,7 @@ inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG bool operator<(const __wrap_iter<_Iter1>& __x, const __wrap_iter<_Iter2>& __y) _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__less_than_comparable(&__x, &__y), "Attempted to compare incomparable iterators"); #endif @@ -1699,7 +1716,7 @@ auto operator-(const __wrap_iter<_Iter1>& __x, const __wrap_iter<_Iter2>& __y) _NOEXCEPT -> decltype(__x.base() - __y.base()) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__less_than_comparable(&__x, &__y), "Attempted to subtract incompatible iterators"); #endif @@ -1711,7 +1728,7 @@ inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_IF_NODEBUG typename __wrap_iter<_Iter1>::difference_type operator-(const __wrap_iter<_Iter1>& __x, const __wrap_iter<_Iter2>& __y) _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__less_than_comparable(&__x, &__y), "Attempted to subtract incompatible iterators"); #endif diff --git a/lib/libcxx/include/latch b/lib/libcxx/include/latch index f669f5860d..a894f8cafd 100644 --- a/lib/libcxx/include/latch +++ b/lib/libcxx/include/latch @@ -19,6 +19,8 @@ namespace std class latch { public: + static constexpr ptrdiff_t max() noexcept; + constexpr explicit latch(ptrdiff_t __expected); ~latch(); @@ -39,6 +41,7 @@ namespace std */ #include <__config> +#include <__availability> #include <atomic> #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) @@ -49,6 +52,9 @@ namespace std # error <latch> is not supported on this single threaded system #endif +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + #if _LIBCPP_STD_VER >= 14 _LIBCPP_BEGIN_NAMESPACE_STD @@ -101,4 +107,6 @@ _LIBCPP_END_NAMESPACE_STD #endif // _LIBCPP_STD_VER >= 14 +_LIBCPP_POP_MACROS + #endif //_LIBCPP_LATCH diff --git a/lib/libcxx/include/limits b/lib/libcxx/include/limits index 6d5d1e1aca..8f97cd10a8 100644 --- a/lib/libcxx/include/limits +++ b/lib/libcxx/include/limits @@ -105,11 +105,11 @@ template<> class numeric_limits<cv long double>; #include <type_traits> #if defined(_LIBCPP_COMPILER_MSVC) -#include "support/win32/limits_msvc_win32.h" +#include "__support/win32/limits_msvc_win32.h" #endif // _LIBCPP_MSVCRT #if defined(__IBMCPP__) -#include "support/ibm/limits.h" +#include "__support/ibm/limits.h" #endif // __IBMCPP__ #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) diff --git a/lib/libcxx/include/list b/lib/libcxx/include/list index 55b45f1a67..a18514d74e 100644 --- a/lib/libcxx/include/list +++ b/lib/libcxx/include/list @@ -293,7 +293,7 @@ class _LIBCPP_TEMPLATE_VIS __list_iterator __link_pointer __ptr_; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY explicit __list_iterator(__link_pointer __p, const void* __c) _NOEXCEPT : __ptr_(__p) @@ -320,12 +320,12 @@ public: _LIBCPP_INLINE_VISIBILITY __list_iterator() _NOEXCEPT : __ptr_(nullptr) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_i(this); #endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __list_iterator(const __list_iterator& __p) @@ -351,12 +351,12 @@ public: return *this; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY reference operator*() const { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to dereference a non-dereferenceable list::iterator"); #endif @@ -365,7 +365,7 @@ public: _LIBCPP_INLINE_VISIBILITY pointer operator->() const { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to dereference a non-dereferenceable list::iterator"); #endif @@ -375,7 +375,7 @@ public: _LIBCPP_INLINE_VISIBILITY __list_iterator& operator++() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to increment non-incrementable list::iterator"); #endif @@ -388,7 +388,7 @@ public: _LIBCPP_INLINE_VISIBILITY __list_iterator& operator--() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__decrementable(this), "Attempted to decrement non-decrementable list::iterator"); #endif @@ -416,7 +416,7 @@ class _LIBCPP_TEMPLATE_VIS __list_const_iterator __link_pointer __ptr_; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY explicit __list_const_iterator(__link_pointer __p, const void* __c) _NOEXCEPT : __ptr_(__p) @@ -440,7 +440,7 @@ public: _LIBCPP_INLINE_VISIBILITY __list_const_iterator() _NOEXCEPT : __ptr_(nullptr) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_i(this); #endif } @@ -448,12 +448,12 @@ public: __list_const_iterator(const __list_iterator<_Tp, _VoidPtr>& __p) _NOEXCEPT : __ptr_(__p.__ptr_) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__iterator_copy(this, &__p); #endif } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY __list_const_iterator(const __list_const_iterator& __p) @@ -479,11 +479,11 @@ public: return *this; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY reference operator*() const { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to dereference a non-dereferenceable list::const_iterator"); #endif @@ -492,7 +492,7 @@ public: _LIBCPP_INLINE_VISIBILITY pointer operator->() const { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to dereference a non-dereferenceable list::const_iterator"); #endif @@ -502,7 +502,7 @@ public: _LIBCPP_INLINE_VISIBILITY __list_const_iterator& operator++() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(this), "Attempted to increment non-incrementable list::const_iterator"); #endif @@ -515,7 +515,7 @@ public: _LIBCPP_INLINE_VISIBILITY __list_const_iterator& operator--() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__decrementable(this), "Attempted to decrement non-decrementable list::const_iterator"); #endif @@ -614,7 +614,7 @@ protected: _LIBCPP_INLINE_VISIBILITY iterator begin() _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__end_.__next_, this); #else return iterator(__end_.__next_); @@ -623,7 +623,7 @@ protected: _LIBCPP_INLINE_VISIBILITY const_iterator begin() const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_iterator(__end_.__next_, this); #else return const_iterator(__end_.__next_); @@ -632,7 +632,7 @@ protected: _LIBCPP_INLINE_VISIBILITY iterator end() _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__end_as_link(), this); #else return iterator(__end_as_link()); @@ -641,7 +641,7 @@ protected: _LIBCPP_INLINE_VISIBILITY const_iterator end() const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_iterator(__end_as_link(), this); #else return const_iterator(__end_as_link()); @@ -696,7 +696,7 @@ private: _LIBCPP_INLINE_VISIBILITY void __invalidate_all_iterators() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__invalidate_all(this); #endif } @@ -735,13 +735,13 @@ inline __list_imp<_Tp, _Alloc>::__list_imp(const __node_allocator& __a) #ifndef _LIBCPP_CXX03_LANG template <class _Tp, class _Alloc> inline __list_imp<_Tp, _Alloc>::__list_imp(__node_allocator&& __a) _NOEXCEPT - : __size_alloc_(0, std::move(__a)) {} + : __size_alloc_(0, _VSTD::move(__a)) {} #endif template <class _Tp, class _Alloc> __list_imp<_Tp, _Alloc>::~__list_imp() { clear(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__erase_c(this); #endif } @@ -783,7 +783,7 @@ __list_imp<_Tp, _Alloc>::swap(__list_imp& __c) "list::swap: Either propagate_on_container_swap must be true" " or the allocators must compare equal"); using _VSTD::swap; - __swap_allocator(__node_alloc(), __c.__node_alloc()); + _VSTD::__swap_allocator(__node_alloc(), __c.__node_alloc()); swap(__sz(), __c.__sz()); swap(__end_, __c.__end_); if (__sz() == 0) @@ -795,13 +795,13 @@ __list_imp<_Tp, _Alloc>::swap(__list_imp& __c) else __c.__end_.__prev_->__next_ = __c.__end_.__next_->__prev_ = __c.__end_as_link(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __libcpp_db* __db = __get_db(); __c_node* __cn1 = __db->__find_c_and_lock(this); __c_node* __cn2 = __db->__find_c(&__c); - std::swap(__cn1->beg_, __cn2->beg_); - std::swap(__cn1->end_, __cn2->end_); - std::swap(__cn1->cap_, __cn2->cap_); + _VSTD::swap(__cn1->beg_, __cn2->beg_); + _VSTD::swap(__cn1->end_, __cn2->end_); + _VSTD::swap(__cn1->cap_, __cn2->cap_); for (__i_node** __p = __cn1->end_; __p != __cn1->beg_;) { --__p; @@ -810,7 +810,7 @@ __list_imp<_Tp, _Alloc>::swap(__list_imp& __c) { __cn2->__add(*__p); if (--__cn1->end_ != __p) - memmove(__p, __p+1, (__cn1->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__cn1->end_ - __p)*sizeof(__i_node*)); } else (*__p)->__c_ = __cn1; @@ -823,7 +823,7 @@ __list_imp<_Tp, _Alloc>::swap(__list_imp& __c) { __cn1->__add(*__p); if (--__cn2->end_ != __p) - memmove(__p, __p+1, (__cn2->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__cn2->end_ - __p)*sizeof(__i_node*)); } else (*__p)->__c_ = __cn2; @@ -870,14 +870,14 @@ public: list() _NOEXCEPT_(is_nothrow_default_constructible<__node_allocator>::value) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } _LIBCPP_INLINE_VISIBILITY explicit list(const allocator_type& __a) : base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -937,7 +937,7 @@ public: _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT { - return std::min<size_type>( + return _VSTD::min<size_type>( base::__node_alloc_max_size(), numeric_limits<difference_type >::max()); } @@ -1117,14 +1117,14 @@ public: return __hold_pointer(__p, __node_destructor(__na, 1)); } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const; bool __decrementable(const const_iterator* __i) const; bool __addable(const const_iterator* __i, ptrdiff_t __n) const; bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const; -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 private: _LIBCPP_INLINE_VISIBILITY @@ -1144,7 +1144,7 @@ private: #ifndef _LIBCPP_HAS_NO_DEDUCTION_GUIDES template<class _InputIterator, - class _Alloc = typename std::allocator<typename iterator_traits<_InputIterator>::value_type>, + class _Alloc = allocator<typename iterator_traits<_InputIterator>::value_type>, class = typename enable_if<__is_allocator<_Alloc>::value, void>::type > list(_InputIterator, _InputIterator) @@ -1207,7 +1207,7 @@ list<_Tp, _Alloc>::__iterator(size_type __n) template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(size_type __n) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __n > 0; --__n) @@ -1222,7 +1222,7 @@ list<_Tp, _Alloc>::list(size_type __n) template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(size_type __n, const allocator_type& __a) : base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __n > 0; --__n) @@ -1233,7 +1233,7 @@ list<_Tp, _Alloc>::list(size_type __n, const allocator_type& __a) : base(__a) template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(size_type __n, const value_type& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __n > 0; --__n) @@ -1244,7 +1244,7 @@ template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(size_type __n, const value_type& __x, const allocator_type& __a) : base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __n > 0; --__n) @@ -1256,7 +1256,7 @@ template <class _InpIter> list<_Tp, _Alloc>::list(_InpIter __f, _InpIter __l, typename enable_if<__is_cpp17_input_iterator<_InpIter>::value>::type*) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __f != __l; ++__f) @@ -1269,7 +1269,7 @@ list<_Tp, _Alloc>::list(_InpIter __f, _InpIter __l, const allocator_type& __a, typename enable_if<__is_cpp17_input_iterator<_InpIter>::value>::type*) : base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __f != __l; ++__f) @@ -1280,7 +1280,7 @@ template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(const list& __c) : base(__node_alloc_traits::select_on_container_copy_construction( __c.__node_alloc())) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (const_iterator __i = __c.begin(), __e = __c.end(); __i != __e; ++__i) @@ -1291,7 +1291,7 @@ template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(const list& __c, const allocator_type& __a) : base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (const_iterator __i = __c.begin(), __e = __c.end(); __i != __e; ++__i) @@ -1304,7 +1304,7 @@ template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(initializer_list<value_type> __il, const allocator_type& __a) : base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (typename initializer_list<value_type>::const_iterator __i = __il.begin(), @@ -1315,7 +1315,7 @@ list<_Tp, _Alloc>::list(initializer_list<value_type> __il, const allocator_type& template <class _Tp, class _Alloc> list<_Tp, _Alloc>::list(initializer_list<value_type> __il) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (typename initializer_list<value_type>::const_iterator __i = __il.begin(), @@ -1327,7 +1327,7 @@ template <class _Tp, class _Alloc> inline list<_Tp, _Alloc>::list(list&& __c) _NOEXCEPT_(is_nothrow_move_constructible<__node_allocator>::value) : base(_VSTD::move(__c.__node_alloc())) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif splice(end(), __c); @@ -1338,7 +1338,7 @@ inline list<_Tp, _Alloc>::list(list&& __c, const allocator_type& __a) : base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__a == __c.get_allocator()) @@ -1415,7 +1415,7 @@ list<_Tp, _Alloc>::assign(_InpIter __f, _InpIter __l, insert(__e, __f, __l); else erase(__i, __e); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__invalidate_all(this); #endif } @@ -1432,7 +1432,7 @@ list<_Tp, _Alloc>::assign(size_type __n, const value_type& __x) insert(__e, __n, __x); else erase(__i, __e); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__invalidate_all(this); #endif } @@ -1449,7 +1449,7 @@ template <class _Tp, class _Alloc> typename list<_Tp, _Alloc>::iterator list<_Tp, _Alloc>::insert(const_iterator __p, const value_type& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::insert(iterator, x) called with an iterator not" " referring to this list"); @@ -1459,7 +1459,7 @@ list<_Tp, _Alloc>::insert(const_iterator __p, const value_type& __x) __node_alloc_traits::construct(__na, _VSTD::addressof(__hold->__value_), __x); __link_nodes(__p.__ptr_, __hold->__as_link(), __hold->__as_link()); ++base::__sz(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__hold.release()->__as_link(), this); #else return iterator(__hold.release()->__as_link()); @@ -1470,7 +1470,7 @@ template <class _Tp, class _Alloc> typename list<_Tp, _Alloc>::iterator list<_Tp, _Alloc>::insert(const_iterator __p, size_type __n, const value_type& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::insert(iterator, n, x) called with an iterator not" " referring to this list"); @@ -1485,7 +1485,7 @@ list<_Tp, _Alloc>::insert(const_iterator __p, size_type __n, const value_type& _ __hold_pointer __hold = __allocate_node(__na); __node_alloc_traits::construct(__na, _VSTD::addressof(__hold->__value_), __x); ++__ds; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __r = iterator(__hold->__as_link(), this); #else __r = iterator(__hold->__as_link()); @@ -1515,7 +1515,7 @@ list<_Tp, _Alloc>::insert(const_iterator __p, size_type __n, const value_type& _ __node_alloc_traits::deallocate(__na, __e.__ptr_->__as_node(), 1); if (__prev == 0) break; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __e = iterator(__prev, this); #else __e = iterator(__prev); @@ -1536,7 +1536,7 @@ typename list<_Tp, _Alloc>::iterator list<_Tp, _Alloc>::insert(const_iterator __p, _InpIter __f, _InpIter __l, typename enable_if<__is_cpp17_input_iterator<_InpIter>::value>::type*) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::insert(iterator, range) called with an iterator not" " referring to this list"); @@ -1551,7 +1551,7 @@ list<_Tp, _Alloc>::insert(const_iterator __p, _InpIter __f, _InpIter __l, __hold_pointer __hold = __allocate_node(__na); __node_alloc_traits::construct(__na, _VSTD::addressof(__hold->__value_), *__f); ++__ds; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __r = iterator(__hold.get()->__as_link(), this); #else __r = iterator(__hold.get()->__as_link()); @@ -1581,7 +1581,7 @@ list<_Tp, _Alloc>::insert(const_iterator __p, _InpIter __f, _InpIter __l, __node_alloc_traits::deallocate(__na, __e.__ptr_->__as_node(), 1); if (__prev == 0) break; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __e = iterator(__prev, this); #else __e = iterator(__prev); @@ -1695,7 +1695,7 @@ template <class... _Args> typename list<_Tp, _Alloc>::iterator list<_Tp, _Alloc>::emplace(const_iterator __p, _Args&&... __args) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::emplace(iterator, args...) called with an iterator not" " referring to this list"); @@ -1707,7 +1707,7 @@ list<_Tp, _Alloc>::emplace(const_iterator __p, _Args&&... __args) __link_nodes(__p.__ptr_, __nl, __nl); ++base::__sz(); __hold.release(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__nl, this); #else return iterator(__nl); @@ -1718,7 +1718,7 @@ template <class _Tp, class _Alloc> typename list<_Tp, _Alloc>::iterator list<_Tp, _Alloc>::insert(const_iterator __p, value_type&& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::insert(iterator, x) called with an iterator not" " referring to this list"); @@ -1730,7 +1730,7 @@ list<_Tp, _Alloc>::insert(const_iterator __p, value_type&& __x) __link_nodes(__p.__ptr_, __nl, __nl); ++base::__sz(); __hold.release(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__nl, this); #else return iterator(__nl); @@ -1748,7 +1748,7 @@ list<_Tp, _Alloc>::pop_front() __link_pointer __n = base::__end_.__next_; base::__unlink_nodes(__n, __n); --base::__sz(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); for (__i_node** __p = __c->end_; __p != __c->beg_; ) { @@ -1758,7 +1758,7 @@ list<_Tp, _Alloc>::pop_front() { (*__p)->__c_ = nullptr; if (--__c->end_ != __p) - memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); } } __get_db()->unlock(); @@ -1777,7 +1777,7 @@ list<_Tp, _Alloc>::pop_back() __link_pointer __n = base::__end_.__prev_; base::__unlink_nodes(__n, __n); --base::__sz(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); for (__i_node** __p = __c->end_; __p != __c->beg_; ) { @@ -1787,7 +1787,7 @@ list<_Tp, _Alloc>::pop_back() { (*__p)->__c_ = nullptr; if (--__c->end_ != __p) - memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); } } __get_db()->unlock(); @@ -1801,7 +1801,7 @@ template <class _Tp, class _Alloc> typename list<_Tp, _Alloc>::iterator list<_Tp, _Alloc>::erase(const_iterator __p) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::erase(iterator) called with an iterator not" " referring to this list"); @@ -1813,7 +1813,7 @@ list<_Tp, _Alloc>::erase(const_iterator __p) __link_pointer __r = __n->__next_; base::__unlink_nodes(__n, __n); --base::__sz(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); for (__i_node** __ip = __c->end_; __ip != __c->beg_; ) { @@ -1823,7 +1823,7 @@ list<_Tp, _Alloc>::erase(const_iterator __p) { (*__ip)->__c_ = nullptr; if (--__c->end_ != __ip) - memmove(__ip, __ip+1, (__c->end_ - __ip)*sizeof(__i_node*)); + _VSTD::memmove(__ip, __ip+1, (__c->end_ - __ip)*sizeof(__i_node*)); } } __get_db()->unlock(); @@ -1831,7 +1831,7 @@ list<_Tp, _Alloc>::erase(const_iterator __p) __node_pointer __np = __n->__as_node(); __node_alloc_traits::destroy(__na, _VSTD::addressof(__np->__value_)); __node_alloc_traits::deallocate(__na, __np, 1); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__r, this); #else return iterator(__r); @@ -1842,7 +1842,7 @@ template <class _Tp, class _Alloc> typename list<_Tp, _Alloc>::iterator list<_Tp, _Alloc>::erase(const_iterator __f, const_iterator __l) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__f) == this, "list::erase(iterator, iterator) called with an iterator not" " referring to this list"); @@ -1859,7 +1859,7 @@ list<_Tp, _Alloc>::erase(const_iterator __f, const_iterator __l) __link_pointer __n = __f.__ptr_; ++__f; --base::__sz(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); for (__i_node** __p = __c->end_; __p != __c->beg_; ) { @@ -1869,7 +1869,7 @@ list<_Tp, _Alloc>::erase(const_iterator __f, const_iterator __l) { (*__p)->__c_ = nullptr; if (--__c->end_ != __p) - memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); } } __get_db()->unlock(); @@ -1879,7 +1879,7 @@ list<_Tp, _Alloc>::erase(const_iterator __f, const_iterator __l) __node_alloc_traits::deallocate(__na, __np, 1); } } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(__l.__ptr_, this); #else return iterator(__l.__ptr_); @@ -1900,7 +1900,7 @@ list<_Tp, _Alloc>::resize(size_type __n) __hold_pointer __hold = __allocate_node(__na); __node_alloc_traits::construct(__na, _VSTD::addressof(__hold->__value_)); ++__ds; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 iterator __r = iterator(__hold.release()->__as_link(), this); #else iterator __r = iterator(__hold.release()->__as_link()); @@ -1929,7 +1929,7 @@ list<_Tp, _Alloc>::resize(size_type __n) __node_alloc_traits::deallocate(__na, __e.__ptr_->__as_node(), 1); if (__prev == 0) break; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __e = iterator(__prev, this); #else __e = iterator(__prev); @@ -1958,7 +1958,7 @@ list<_Tp, _Alloc>::resize(size_type __n, const value_type& __x) __node_alloc_traits::construct(__na, _VSTD::addressof(__hold->__value_), __x); ++__ds; __link_pointer __nl = __hold.release()->__as_link(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 iterator __r = iterator(__nl, this); #else iterator __r = iterator(__nl); @@ -1987,7 +1987,7 @@ list<_Tp, _Alloc>::resize(size_type __n, const value_type& __x) __node_alloc_traits::deallocate(__na, __e.__ptr_->__as_node(), 1); if (__prev == 0) break; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __e = iterator(__prev, this); #else __e = iterator(__prev); @@ -2007,7 +2007,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c) { _LIBCPP_ASSERT(this != &__c, "list::splice(iterator, list) called with this == &list"); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::splice(iterator, list) called with an iterator not" " referring to this list"); @@ -2020,7 +2020,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c) __link_nodes(__p.__ptr_, __f, __l); base::__sz() += __c.__sz(); __c.__sz() = 0; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 if (&__c != this) { __libcpp_db* __db = __get_db(); __c_node* __cn1 = __db->__find_c_and_lock(this); @@ -2034,7 +2034,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c) __cn1->__add(*__ip); (*__ip)->__c_ = __cn1; if (--__cn2->end_ != __ip) - memmove(__ip, __ip+1, (__cn2->end_ - __ip)*sizeof(__i_node*)); + _VSTD::memmove(__ip, __ip+1, (__cn2->end_ - __ip)*sizeof(__i_node*)); } } __db->unlock(); @@ -2047,7 +2047,7 @@ template <class _Tp, class _Alloc> void list<_Tp, _Alloc>::splice(const_iterator __p, list& __c, const_iterator __i) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::splice(iterator, list, iterator) called with first iterator not" " referring to this list"); @@ -2056,7 +2056,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c, const_iterator __i) " referring to list argument"); _LIBCPP_ASSERT(__get_const_db()->__dereferenceable(&__i), "list::splice(iterator, list, iterator) called with second iterator not" - " derefereceable"); + " dereferenceable"); #endif if (__p.__ptr_ != __i.__ptr_ && __p.__ptr_ != __i.__ptr_->__next_) { @@ -2065,7 +2065,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c, const_iterator __i) __link_nodes(__p.__ptr_, __f, __f); --__c.__sz(); ++base::__sz(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 if (&__c != this) { __libcpp_db* __db = __get_db(); __c_node* __cn1 = __db->__find_c_and_lock(this); @@ -2079,7 +2079,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c, const_iterator __i) __cn1->__add(*__ip); (*__ip)->__c_ = __cn1; if (--__cn2->end_ != __ip) - memmove(__ip, __ip+1, (__cn2->end_ - __ip)*sizeof(__i_node*)); + _VSTD::memmove(__ip, __ip+1, (__cn2->end_ - __ip)*sizeof(__i_node*)); } } __db->unlock(); @@ -2092,7 +2092,7 @@ template <class _Tp, class _Alloc> void list<_Tp, _Alloc>::splice(const_iterator __p, list& __c, const_iterator __f, const_iterator __l) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "list::splice(iterator, list, iterator, iterator) called with first iterator not" " referring to this list"); @@ -2121,7 +2121,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c, const_iterator __f, con } base::__unlink_nodes(__first, __last); __link_nodes(__p.__ptr_, __first, __last); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 if (&__c != this) { __libcpp_db* __db = __get_db(); __c_node* __cn1 = __db->__find_c_and_lock(this); @@ -2138,7 +2138,7 @@ list<_Tp, _Alloc>::splice(const_iterator __p, list& __c, const_iterator __f, con __cn1->__add(*__ip); (*__ip)->__c_ = __cn1; if (--__cn2->end_ != __ip) - memmove(__ip, __ip+1, (__cn2->end_ - __ip)*sizeof(__i_node*)); + _VSTD::memmove(__ip, __ip+1, (__cn2->end_ - __ip)*sizeof(__i_node*)); } } } @@ -2258,7 +2258,7 @@ list<_Tp, _Alloc>::merge(list& __c, _Comp __comp) ++__f1; } splice(__e1, __c); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __libcpp_db* __db = __get_db(); __c_node* __cn1 = __db->__find_c_and_lock(this); __c_node* __cn2 = __db->__find_c(&__c); @@ -2271,7 +2271,7 @@ list<_Tp, _Alloc>::merge(list& __c, _Comp __comp) __cn1->__add(*__p); (*__p)->__c_ = __cn1; if (--__cn2->end_ != __p) - memmove(__p, __p+1, (__cn2->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__cn2->end_ - __p)*sizeof(__i_node*)); } } __db->unlock(); @@ -2382,7 +2382,7 @@ list<_Tp, _Alloc>::__invariants() const return size() == _VSTD::distance(begin(), end()); } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 template <class _Tp, class _Alloc> bool @@ -2412,7 +2412,7 @@ list<_Tp, _Alloc>::__subscriptable(const const_iterator*, ptrdiff_t) const return false; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 template <class _Tp, class _Alloc> inline _LIBCPP_INLINE_VISIBILITY diff --git a/lib/libcxx/include/locale b/lib/libcxx/include/locale index 3fe4430022..9a705c77cf 100644 --- a/lib/libcxx/include/locale +++ b/lib/libcxx/include/locale @@ -92,7 +92,11 @@ public: typedef typename Codecvt::state_type state_type; typedef typename wide_string::traits_type::int_type int_type; - explicit wstring_convert(Codecvt* pcvt = new Codecvt); // explicit in C++14 + wstring_convert(Codecvt* pcvt = new Codecvt); // before C++14 + explicit wstring_convert(Codecvt* pcvt = new Codecvt); // before C++20 + wstring_convert() : wstring_convert(new Codecvt) {} // C++20 + explicit wstring_convert(Codecvt* pcvt); // C++20 + wstring_convert(Codecvt* pcvt, state_type state); explicit wstring_convert(const byte_string& byte_err, // explicit in C++14 const wide_string& wide_err = wide_string()); @@ -121,8 +125,14 @@ class wbuffer_convert public: typedef typename Tr::state_type state_type; - explicit wbuffer_convert(streambuf* bytebuf = 0, Codecvt* pcvt = new Codecvt, - state_type state = state_type()); // explicit in C++14 + wbuffer_convert(streambuf* bytebuf = 0, Codecvt* pcvt = new Codecvt, + state_type state = state_type()); // before C++14 + explicit wbuffer_convert(streambuf* bytebuf = nullptr, Codecvt* pcvt = new Codecvt, + state_type state = state_type()); // before C++20 + wbuffer_convert() : wbuffer_convert(nullptr) {} // C++20 + explicit wbuffer_convert(streambuf* bytebuf, Codecvt* pcvt = new Codecvt, + state_type state = state_type()); // C++20 + wbuffer_convert(const wbuffer_convert&) = delete; // C++14 wbuffer_convert & operator=(const wbuffer_convert &) = delete; // C++14 ~wbuffer_convert(); // C++14 @@ -197,10 +207,6 @@ template <class charT> class messages_byname; #include <nl_types.h> #endif -#ifdef __APPLE__ -#include <Availability.h> -#endif - #ifdef _LIBCPP_LOCALE__L_EXTENSIONS #include <__bsd_locale_defaults.h> #else @@ -261,11 +267,11 @@ __scan_keyword(_InputIterator& __b, _InputIterator __e, const unsigned char __does_match = '\2'; unsigned char __statbuf[100]; unsigned char* __status = __statbuf; - unique_ptr<unsigned char, void(*)(void*)> __stat_hold(0, free); + unique_ptr<unsigned char, void(*)(void*)> __stat_hold(nullptr, free); if (__nkw > sizeof(__statbuf)) { __status = (unsigned char*)malloc(__nkw); - if (__status == 0) + if (__status == nullptr) __throw_bad_alloc(); __stat_hold.reset(__status); } @@ -561,8 +567,8 @@ __num_get<_CharT>::__stage2_float_loop(_CharT __ct, bool& __in_units, char& __ex return 0; } -_LIBCPP_EXTERN_TEMPLATE2(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_get<char>) -_LIBCPP_EXTERN_TEMPLATE2(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_get<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_get<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_get<wchar_t>) template <class _CharT, class _InputIterator = istreambuf_iterator<_CharT> > class _LIBCPP_TEMPLATE_VIS num_get @@ -877,8 +883,8 @@ num_get<_CharT, _InputIterator>::do_get(iter_type __b, iter_type __e, const numpunct<_CharT>& __np = use_facet<numpunct<_CharT> >(__iob.getloc()); typedef typename numpunct<_CharT>::string_type string_type; const string_type __names[2] = {__np.truename(), __np.falsename()}; - const string_type* __i = __scan_keyword(__b, __e, __names, __names+2, - __ct, __err); + const string_type* __i = _VSTD::__scan_keyword(__b, __e, __names, __names+2, + __ct, __err); __v = __i == __names; return __b; } @@ -1099,8 +1105,8 @@ num_get<_CharT, _InputIterator>::do_get(iter_type __b, iter_type __e, return __b; } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_get<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_get<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_get<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_get<wchar_t>) struct _LIBCPP_TYPE_VIS __num_put_base { @@ -1249,8 +1255,8 @@ __num_put<_CharT>::__widen_and_group_float(char* __nb, char* __np, char* __ne, __op = __ob + (__np - __nb); } -_LIBCPP_EXTERN_TEMPLATE2(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_put<char>) -_LIBCPP_EXTERN_TEMPLATE2(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_put<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_put<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(struct _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __num_put<wchar_t>) template <class _CharT, class _OutputIterator = ostreambuf_iterator<_CharT> > class _LIBCPP_TEMPLATE_VIS num_put @@ -1427,7 +1433,7 @@ num_put<_CharT, _OutputIterator>::do_put(iter_type __s, ios_base& __iob, return do_put(__s, __iob, __fl, (unsigned long)__v); const numpunct<char_type>& __np = use_facet<numpunct<char_type> >(__iob.getloc()); typedef typename numpunct<char_type>::string_type string_type; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 string_type __tmp(__v ? __np.truename() : __np.falsename()); string_type __nm = _VSTD::move(__tmp); #else @@ -1564,14 +1570,14 @@ num_put<_CharT, _OutputIterator>::do_put(iter_type __s, ios_base& __iob, (int)__iob.precision(), __v); else __nc = __libcpp_snprintf_l(__nb, __nbuf, _LIBCPP_GET_C_LOCALE, __fmt, __v); - unique_ptr<char, void(*)(void*)> __nbh(0, free); + unique_ptr<char, void(*)(void*)> __nbh(nullptr, free); if (__nc > static_cast<int>(__nbuf-1)) { if (__specify_precision) __nc = __libcpp_asprintf_l(&__nb, _LIBCPP_GET_C_LOCALE, __fmt, (int)__iob.precision(), __v); else __nc = __libcpp_asprintf_l(&__nb, _LIBCPP_GET_C_LOCALE, __fmt, __v); - if (__nb == 0) + if (__nc == -1) __throw_bad_alloc(); __nbh.reset(__nb); } @@ -1615,14 +1621,14 @@ num_put<_CharT, _OutputIterator>::do_put(iter_type __s, ios_base& __iob, (int)__iob.precision(), __v); else __nc = __libcpp_snprintf_l(__nb, __nbuf, _LIBCPP_GET_C_LOCALE, __fmt, __v); - unique_ptr<char, void(*)(void*)> __nbh(0, free); + unique_ptr<char, void(*)(void*)> __nbh(nullptr, free); if (__nc > static_cast<int>(__nbuf-1)) { if (__specify_precision) __nc = __libcpp_asprintf_l(&__nb, _LIBCPP_GET_C_LOCALE, __fmt, (int)__iob.precision(), __v); else __nc = __libcpp_asprintf_l(&__nb, _LIBCPP_GET_C_LOCALE, __fmt, __v); - if (__nb == 0) + if (__nc == -1) __throw_bad_alloc(); __nbh.reset(__nb); } @@ -1676,8 +1682,8 @@ num_put<_CharT, _OutputIterator>::do_put(iter_type __s, ios_base& __iob, return __pad_and_output(__s, __o, __op, __oe, __iob, __fl); } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_put<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_put<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_put<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS num_put<wchar_t>) template <class _CharT, class _InputIterator> _LIBCPP_HIDDEN @@ -1916,7 +1922,7 @@ time_get<_CharT, _InputIterator>::__get_weekdayname(int& __w, { // Note: ignoring case comes from the POSIX strptime spec const string_type* __wk = this->__weeks(); - ptrdiff_t __i = __scan_keyword(__b, __e, __wk, __wk+14, __ct, __err, false) - __wk; + ptrdiff_t __i = _VSTD::__scan_keyword(__b, __e, __wk, __wk+14, __ct, __err, false) - __wk; if (__i < 14) __w = __i % 7; } @@ -1930,7 +1936,7 @@ time_get<_CharT, _InputIterator>::__get_monthname(int& __m, { // Note: ignoring case comes from the POSIX strptime spec const string_type* __month = this->__months(); - ptrdiff_t __i = __scan_keyword(__b, __e, __month, __month+24, __ct, __err, false) - __month; + ptrdiff_t __i = _VSTD::__scan_keyword(__b, __e, __month, __month+24, __ct, __err, false) - __month; if (__i < 24) __m = __i % 12; } @@ -1942,7 +1948,7 @@ time_get<_CharT, _InputIterator>::__get_day(int& __d, ios_base::iostate& __err, const ctype<char_type>& __ct) const { - int __t = __get_up_to_n_digits(__b, __e, __err, __ct, 2); + int __t = _VSTD::__get_up_to_n_digits(__b, __e, __err, __ct, 2); if (!(__err & ios_base::failbit) && 1 <= __t && __t <= 31) __d = __t; else @@ -2102,7 +2108,7 @@ time_get<_CharT, _InputIterator>::__get_am_pm(int& __h, __err |= ios_base::failbit; return; } - ptrdiff_t __i = __scan_keyword(__b, __e, __ap, __ap+2, __ct, __err, false) - __ap; + ptrdiff_t __i = _VSTD::__scan_keyword(__b, __e, __ap, __ap+2, __ct, __err, false) - __ap; if (__i == 0 && __h == 12) __h = 0; else if (__i == 1 && __h < 12) @@ -2362,8 +2368,8 @@ time_get<_CharT, _InputIterator>::do_get(iter_type __b, iter_type __e, return __b; } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get<wchar_t>) class _LIBCPP_TYPE_VIS __time_get { @@ -2462,8 +2468,8 @@ private: virtual const string_type& __X() const {return this->__X_;} }; -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get_byname<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get_byname<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get_byname<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_get_byname<wchar_t>) class _LIBCPP_TYPE_VIS __time_put { @@ -2575,8 +2581,8 @@ time_put<_CharT, _OutputIterator>::do_put(iter_type __s, ios_base&, return _VSTD::copy(__nb, __ne, __s); } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put<wchar_t>) template <class _CharT, class _OutputIterator = ostreambuf_iterator<_CharT> > class _LIBCPP_TEMPLATE_VIS time_put_byname @@ -2596,8 +2602,8 @@ protected: ~time_put_byname() {} }; -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put_byname<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put_byname<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put_byname<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS time_put_byname<wchar_t>) // money_base @@ -2663,10 +2669,10 @@ template <class _CharT, bool _International> const bool moneypunct<_CharT, _International>::intl; -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<char, false>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<char, true>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<wchar_t, false>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<wchar_t, true>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<char, false>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<char, true>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<wchar_t, false>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct<wchar_t, true>) // moneypunct_byname @@ -2720,10 +2726,10 @@ template<> _LIBCPP_FUNC_VIS void moneypunct_byname<char, true>::init(const char* template<> _LIBCPP_FUNC_VIS void moneypunct_byname<wchar_t, false>::init(const char*); template<> _LIBCPP_FUNC_VIS void moneypunct_byname<wchar_t, true>::init(const char*); -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<char, false>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<char, true>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<wchar_t, false>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<wchar_t, true>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<char, false>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<char, true>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<wchar_t, false>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS moneypunct_byname<wchar_t, true>) // money_get @@ -2779,8 +2785,8 @@ __money_get<_CharT>::__gather_info(bool __intl, const locale& __loc, } } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_get<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_get<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_get<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_get<wchar_t>) template <class _CharT, class _InputIterator = istreambuf_iterator<_CharT> > class _LIBCPP_TEMPLATE_VIS money_get @@ -3108,11 +3114,11 @@ money_get<_CharT, _InputIterator>::do_get(iter_type __b, iter_type __e, __ct.widen(__src, __src + (sizeof(__src)-1), __atoms); char __nbuf[__bz]; char* __nc = __nbuf; - unique_ptr<char, void(*)(void*)> __h(0, free); + unique_ptr<char, void(*)(void*)> __h(nullptr, free); if (__wn - __wb.get() > __bz-2) { __h.reset((char*)malloc(static_cast<size_t>(__wn - __wb.get() + 2))); - if (__h.get() == 0) + if (__h.get() == nullptr) __throw_bad_alloc(); __nc = __h.get(); } @@ -3162,8 +3168,8 @@ money_get<_CharT, _InputIterator>::do_get(iter_type __b, iter_type __e, return __b; } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_get<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_get<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_get<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_get<wchar_t>) // money_put @@ -3337,8 +3343,8 @@ __money_put<_CharT>::__format(char_type* __mb, char_type*& __mi, char_type*& __m __mi = __mb; } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_put<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_put<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_put<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS __money_put<wchar_t>) template <class _CharT, class _OutputIterator = ostreambuf_iterator<_CharT> > class _LIBCPP_TEMPLATE_VIS money_put @@ -3396,17 +3402,17 @@ money_put<_CharT, _OutputIterator>::do_put(iter_type __s, bool __intl, char* __bb = __buf; char_type __digits[__bs]; char_type* __db = __digits; - size_t __n = static_cast<size_t>(snprintf(__bb, __bs, "%.0Lf", __units)); - unique_ptr<char, void(*)(void*)> __hn(0, free); + int __n = snprintf(__bb, __bs, "%.0Lf", __units); + unique_ptr<char, void(*)(void*)> __hn(nullptr, free); unique_ptr<char_type, void(*)(void*)> __hd(0, free); // secure memory for digit storage - if (__n > __bs-1) + if (static_cast<size_t>(__n) > __bs-1) { - __n = static_cast<size_t>(__libcpp_asprintf_l(&__bb, _LIBCPP_GET_C_LOCALE, "%.0Lf", __units)); - if (__bb == 0) + __n = __libcpp_asprintf_l(&__bb, _LIBCPP_GET_C_LOCALE, "%.0Lf", __units); + if (__n == -1) __throw_bad_alloc(); __hn.reset(__bb); - __hd.reset((char_type*)malloc(__n * sizeof(char_type))); + __hd.reset((char_type*)malloc(static_cast<size_t>(__n) * sizeof(char_type))); if (__hd == nullptr) __throw_bad_alloc(); __db = __hd.get(); @@ -3428,9 +3434,9 @@ money_put<_CharT, _OutputIterator>::do_put(iter_type __s, bool __intl, char_type __mbuf[__bs]; char_type* __mb = __mbuf; unique_ptr<char_type, void(*)(void*)> __hw(0, free); - size_t __exn = static_cast<int>(__n) > __fd ? - (__n - static_cast<size_t>(__fd)) * 2 + __sn.size() + - __sym.size() + static_cast<size_t>(__fd) + 1 + size_t __exn = __n > __fd ? + (static_cast<size_t>(__n) - static_cast<size_t>(__fd)) * 2 + + __sn.size() + __sym.size() + static_cast<size_t>(__fd) + 1 : __sn.size() + __sym.size() + static_cast<size_t>(__fd) + 2; if (__exn > __bs) { @@ -3490,8 +3496,8 @@ money_put<_CharT, _OutputIterator>::do_put(iter_type __s, bool __intl, return __pad_and_output(__s, __mb, __mi, __me, __iob, __fl); } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_put<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_put<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_put<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS money_put<wchar_t>) // messages @@ -3582,7 +3588,7 @@ messages<_CharT>::do_get(catalog __c, int __set, int __msgid, char* __n = catgets(__cat, __set, __msgid, __ndflt.c_str()); string_type __w; __widen_from_utf8<sizeof(char_type)*__CHAR_BIT__>()(back_inserter(__w), - __n, __n + strlen(__n)); + __n, __n + _VSTD::strlen(__n)); return __w; #else // !_LIBCPP_HAS_CATOPEN _LIBCPP_UNUSED_VAR(__c); @@ -3606,8 +3612,8 @@ messages<_CharT>::do_close(catalog __c) const #endif // _LIBCPP_HAS_CATOPEN } -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages<wchar_t>) template <class _CharT> class _LIBCPP_TEMPLATE_VIS messages_byname @@ -3630,8 +3636,8 @@ protected: ~messages_byname() {} }; -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages_byname<char>) -_LIBCPP_EXTERN_TEMPLATE2(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages_byname<wchar_t>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages_byname<char>) +_LIBCPP_EXTERN_TEMPLATE_EVEN_IN_DEBUG_MODE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS messages_byname<wchar_t>) template<class _Codecvt, class _Elem = wchar_t, class _Wide_alloc = allocator<_Elem>, @@ -3654,8 +3660,17 @@ private: wstring_convert(const wstring_convert& __wc); wstring_convert& operator=(const wstring_convert& __wc); public: +#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY - _LIBCPP_EXPLICIT_AFTER_CXX11 wstring_convert(_Codecvt* __pcvt = new _Codecvt); + wstring_convert() : wstring_convert(new _Codecvt) {} + _LIBCPP_INLINE_VISIBILITY + explicit wstring_convert(_Codecvt* __pcvt); +#else + _LIBCPP_INLINE_VISIBILITY + _LIBCPP_EXPLICIT_AFTER_CXX11 + wstring_convert(_Codecvt* __pcvt = new _Codecvt); +#endif + _LIBCPP_INLINE_VISIBILITY wstring_convert(_Codecvt* __pcvt, state_type __state); _LIBCPP_EXPLICIT_AFTER_CXX11 wstring_convert(const byte_string& __byte_err, @@ -3922,9 +3937,20 @@ private: wbuffer_convert(const wbuffer_convert&); wbuffer_convert& operator=(const wbuffer_convert&); + public: - _LIBCPP_EXPLICIT_AFTER_CXX11 wbuffer_convert(streambuf* __bytebuf = 0, - _Codecvt* __pcvt = new _Codecvt, state_type __state = state_type()); +#ifndef _LIBCPP_CXX03_LANG + wbuffer_convert() : wbuffer_convert(nullptr) {} + explicit wbuffer_convert(streambuf* __bytebuf, + _Codecvt* __pcvt = new _Codecvt, + state_type __state = state_type()); +#else + _LIBCPP_EXPLICIT_AFTER_CXX11 + wbuffer_convert(streambuf* __bytebuf = nullptr, + _Codecvt* __pcvt = new _Codecvt, + state_type __state = state_type()); +#endif + ~wbuffer_convert(); _LIBCPP_INLINE_VISIBILITY @@ -3961,9 +3987,9 @@ private: template <class _Codecvt, class _Elem, class _Tr> wbuffer_convert<_Codecvt, _Elem, _Tr>:: wbuffer_convert(streambuf* __bytebuf, _Codecvt* __pcvt, state_type __state) - : __extbuf_(0), - __extbufnext_(0), - __extbufend_(0), + : __extbuf_(nullptr), + __extbufnext_(nullptr), + __extbufend_(nullptr), __ebs_(0), __intbuf_(0), __ibs_(0), @@ -4003,7 +4029,7 @@ wbuffer_convert<_Codecvt, _Elem, _Tr>::underflow() int_type __c = traits_type::eof(); if (this->gptr() == this->egptr()) { - memmove(this->eback(), this->egptr() - __unget_sz, __unget_sz * sizeof(char_type)); + _VSTD::memmove(this->eback(), this->egptr() - __unget_sz, __unget_sz * sizeof(char_type)); if (__always_noconv_) { streamsize __nmemb = static_cast<streamsize>(this->egptr() - this->eback() - __unget_sz); @@ -4020,7 +4046,7 @@ wbuffer_convert<_Codecvt, _Elem, _Tr>::underflow() { _LIBCPP_ASSERT(!(__extbufnext_ == NULL && (__extbufend_ != __extbufnext_)), "underflow moving from NULL" ); if (__extbufend_ != __extbufnext_) - memmove(__extbuf_, __extbufnext_, __extbufend_ - __extbufnext_); + _VSTD::memmove(__extbuf_, __extbufnext_, __extbufend_ - __extbufnext_); __extbufnext_ = __extbuf_ + (__extbufend_ - __extbufnext_); __extbufend_ = __extbuf_ + (__extbuf_ == __extbuf_min_ ? sizeof(__extbuf_min_) : __ebs_); streamsize __nmemb = _VSTD::min(static_cast<streamsize>(this->egptr() - this->eback() - __unget_sz), @@ -4336,12 +4362,12 @@ template <class _Codecvt, class _Elem, class _Tr> wbuffer_convert<_Codecvt, _Elem, _Tr>* wbuffer_convert<_Codecvt, _Elem, _Tr>::__close() { - wbuffer_convert* __rt = 0; - if (__cv_ != 0 && __bufptr_ != 0) + wbuffer_convert* __rt = nullptr; + if (__cv_ != nullptr && __bufptr_ != nullptr) { __rt = this; if ((__cm_ & ios_base::out) && sync()) - __rt = 0; + __rt = nullptr; } return __rt; } diff --git a/lib/libcxx/include/locale.h b/lib/libcxx/include/locale.h index a21ee385c3..81cfcee57f 100644 --- a/lib/libcxx/include/locale.h +++ b/lib/libcxx/include/locale.h @@ -35,8 +35,12 @@ Functions: #include <__config> +#if defined(_LIBCPP_HAS_NO_LOCALIZATION) +# error "Localization is not supported by this configuration of libc++" +#endif + #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) -#pragma GCC system_header +# pragma GCC system_header #endif #include_next <locale.h> diff --git a/lib/libcxx/include/map b/lib/libcxx/include/map index d2b8259136..a27002c279 100644 --- a/lib/libcxx/include/map +++ b/lib/libcxx/include/map @@ -1230,7 +1230,7 @@ public: return __tree_.__emplace_hint_unique_key_args(__h.__i_, __k, _VSTD::piecewise_construct, _VSTD::forward_as_tuple(__k), - _VSTD::forward_as_tuple(_VSTD::forward<_Args>(__args)...)); + _VSTD::forward_as_tuple(_VSTD::forward<_Args>(__args)...)).first; } template <class... _Args> @@ -1240,7 +1240,7 @@ public: return __tree_.__emplace_hint_unique_key_args(__h.__i_, __k, _VSTD::piecewise_construct, _VSTD::forward_as_tuple(_VSTD::move(__k)), - _VSTD::forward_as_tuple(_VSTD::forward<_Args>(__args)...)); + _VSTD::forward_as_tuple(_VSTD::forward<_Args>(__args)...)).first; } template <class _Vp> @@ -1270,30 +1270,30 @@ public: } template <class _Vp> - _LIBCPP_INLINE_VISIBILITY - iterator insert_or_assign(const_iterator __h, const key_type& __k, _Vp&& __v) - { - iterator __p = lower_bound(__k); - if ( __p != end() && !key_comp()(__k, __p->first)) - { - __p->second = _VSTD::forward<_Vp>(__v); - return __p; - } - return emplace_hint(__h, __k, _VSTD::forward<_Vp>(__v)); - } + _LIBCPP_INLINE_VISIBILITY iterator insert_or_assign(const_iterator __h, + const key_type& __k, + _Vp&& __v) { + auto [__r, __inserted] = __tree_.__emplace_hint_unique_key_args( + __h.__i_, __k, __k, _VSTD::forward<_Vp>(__v)); + + if (!__inserted) + __r->__get_value().second = _VSTD::forward<_Vp>(__v); + + return __r; + } template <class _Vp> - _LIBCPP_INLINE_VISIBILITY - iterator insert_or_assign(const_iterator __h, key_type&& __k, _Vp&& __v) - { - iterator __p = lower_bound(__k); - if ( __p != end() && !key_comp()(__k, __p->first)) - { - __p->second = _VSTD::forward<_Vp>(__v); - return __p; - } - return emplace_hint(__h, _VSTD::move(__k), _VSTD::forward<_Vp>(__v)); - } + _LIBCPP_INLINE_VISIBILITY iterator insert_or_assign(const_iterator __h, + key_type&& __k, + _Vp&& __v) { + auto [__r, __inserted] = __tree_.__emplace_hint_unique_key_args( + __h.__i_, __k, _VSTD::move(__k), _VSTD::forward<_Vp>(__v)); + + if (!__inserted) + __r->__get_value().second = _VSTD::forward<_Vp>(__v); + + return __r; + } #endif // _LIBCPP_STD_VER > 14 @@ -1546,7 +1546,7 @@ map<_Key, _Tp, _Compare, _Allocator>::__construct_node_with_key(const key_type& __h.get_deleter().__first_constructed = true; __node_traits::construct(__na, _VSTD::addressof(__h->__value_.__get_value().second)); __h.get_deleter().__second_constructed = true; - return _LIBCPP_EXPLICIT_MOVE(__h); // explicitly moved for C++03 + return __h; } template <class _Key, class _Tp, class _Compare, class _Allocator> diff --git a/lib/libcxx/include/memory b/lib/libcxx/include/memory index 1f9f36c5bb..39d0f5bee6 100644 --- a/lib/libcxx/include/memory +++ b/lib/libcxx/include/memory @@ -46,7 +46,7 @@ struct pointer_traits<T*> }; template <class T> constexpr T* to_address(T* p) noexcept; // C++20 -template <class Ptr> auto to_address(const Ptr& p) noexcept; // C++20 +template <class Ptr> constexpr auto to_address(const Ptr& p) noexcept; // C++20 template <class Alloc> struct allocator_traits @@ -83,21 +83,19 @@ struct allocator_traits template <class T> using rebind_alloc = Alloc::rebind<T>::other | Alloc<T, Args...>; template <class T> using rebind_traits = allocator_traits<rebind_alloc<T>>; - static pointer allocate(allocator_type& a, size_type n); // [[nodiscard]] in C++20 - static pointer allocate(allocator_type& a, size_type n, const_void_pointer hint); // [[nodiscard]] in C++20 + static pointer allocate(allocator_type& a, size_type n); // constexpr and [[nodiscard]] in C++20 + static pointer allocate(allocator_type& a, size_type n, const_void_pointer hint); // constexpr and [[nodiscard]] in C++20 - static void deallocate(allocator_type& a, pointer p, size_type n) noexcept; + static void deallocate(allocator_type& a, pointer p, size_type n) noexcept; // constexpr in C++20 template <class T, class... Args> - static void construct(allocator_type& a, T* p, Args&&... args); + static void construct(allocator_type& a, T* p, Args&&... args); // constexpr in C++20 template <class T> - static void destroy(allocator_type& a, T* p); + static void destroy(allocator_type& a, T* p); // constexpr in C++20 - static size_type max_size(const allocator_type& a); // noexcept in C++14 - - static allocator_type - select_on_container_copy_construction(const allocator_type& a); + static size_type max_size(const allocator_type& a); // noexcept in C++14, constexpr in C++20 + static allocator_type select_on_container_copy_construction(const allocator_type& a); // constexpr in C++20 }; template <> @@ -115,8 +113,8 @@ template <class T> class allocator { public: - typedef size_t size_type; // deprecated in C++17, removed in C++20 - typedef ptrdiff_t difference_type; // deprecated in C++17, removed in C++20 + typedef size_t size_type; + typedef ptrdiff_t difference_type; typedef T* pointer; // deprecated in C++17, removed in C++20 typedef const T* const_pointer; // deprecated in C++17, removed in C++20 typedef typename add_lvalue_reference<T>::type @@ -135,12 +133,12 @@ public: constexpr allocator(const allocator&) noexcept; // constexpr in C++20 template <class U> constexpr allocator(const allocator<U>&) noexcept; // constexpr in C++20 - ~allocator(); + ~allocator(); // constexpr in C++20 pointer address(reference x) const noexcept; // deprecated in C++17, removed in C++20 const_pointer address(const_reference x) const noexcept; // deprecated in C++17, removed in C++20 T* allocate(size_t n, const void* hint); // deprecated in C++17, removed in C++20 - T* allocate(size_t n); - void deallocate(T* p, size_t n) noexcept; + T* allocate(size_t n); // constexpr in C++20 + void deallocate(T* p, size_t n) noexcept; // constexpr in C++20 size_type max_size() const noexcept; // deprecated in C++17, removed in C++20 template<class U, class... Args> void construct(U* p, Args&&... args); // deprecated in C++17, removed in C++20 @@ -149,10 +147,10 @@ public: }; template <class T, class U> -bool operator==(const allocator<T>&, const allocator<U>&) noexcept; +bool operator==(const allocator<T>&, const allocator<U>&) noexcept; // constexpr in C++20 template <class T, class U> -bool operator!=(const allocator<T>&, const allocator<U>&) noexcept; +bool operator!=(const allocator<T>&, const allocator<U>&) noexcept; // constexpr in C++20 template <class OutputIterator, class T> class raw_storage_iterator @@ -191,14 +189,17 @@ template <class ForwardIterator, class Size, class T> ForwardIterator uninitialized_fill_n(ForwardIterator first, Size n, const T& x); +template <class T, class ...Args> +constexpr T* construct_at(T* location, Args&& ...args); // since C++20 + template <class T> -void destroy_at(T* location); +void destroy_at(T* location); // constexpr in C++20 template <class ForwardIterator> - void destroy(ForwardIterator first, ForwardIterator last); +void destroy(ForwardIterator first, ForwardIterator last); // constexpr in C++20 template <class ForwardIterator, class Size> - ForwardIterator destroy_n(ForwardIterator first, Size n); +ForwardIterator destroy_n(ForwardIterator first, Size n); // constexpr in C++20 template <class InputIterator, class ForwardIterator> ForwardIterator uninitialized_move(InputIterator first, InputIterator last, ForwardIterator result); @@ -338,7 +339,7 @@ public: pointer release() noexcept; void reset(pointer p = pointer()) noexcept; void reset(nullptr_t) noexcept; - template <class U> void reset(U) = delete; + template <class U> void reset(U) = delete; void swap(unique_ptr& u) noexcept; }; @@ -664,6 +665,7 @@ void* align(size_t alignment, size_t size, void*& ptr, size_t& space); */ #include <__config> +#include <__availability> #include <type_traits> #include <typeinfo> #include <cstddef> @@ -677,6 +679,10 @@ void* align(size_t alignment, size_t size, void*& ptr, size_t& space); #include <tuple> #include <stdexcept> #include <cstring> +#include <__memory/allocator_traits.h> +#include <__memory/base.h> +#include <__memory/pointer_traits.h> +#include <__memory/utilities.h> #if !defined(_LIBCPP_HAS_NO_ATOMIC_HEADER) # include <atomic> #endif @@ -716,413 +722,11 @@ _ValueType __libcpp_acquire_load(_ValueType const* __value) { #endif } -// addressof moved to <type_traits> - -template <class _Tp> class allocator; - -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) -template <> -class _LIBCPP_TEMPLATE_VIS _LIBCPP_DEPRECATED_IN_CXX17 allocator<void> -{ -public: - typedef void* pointer; - typedef const void* const_pointer; - typedef void value_type; - - template <class _Up> struct rebind {typedef allocator<_Up> other;}; -}; - -template <> -class _LIBCPP_TEMPLATE_VIS _LIBCPP_DEPRECATED_IN_CXX17 allocator<const void> -{ -public: - typedef const void* pointer; - typedef const void* const_pointer; - typedef const void value_type; - - template <class _Up> struct rebind {typedef allocator<_Up> other;}; -}; -#endif - -// pointer_traits - -template <class _Tp, class = void> -struct __has_element_type : false_type {}; - -template <class _Tp> -struct __has_element_type<_Tp, - typename __void_t<typename _Tp::element_type>::type> : true_type {}; - -template <class _Ptr, bool = __has_element_type<_Ptr>::value> -struct __pointer_traits_element_type; - -template <class _Ptr> -struct __pointer_traits_element_type<_Ptr, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Ptr::element_type type; -}; - -#ifndef _LIBCPP_HAS_NO_VARIADICS - -template <template <class, class...> class _Sp, class _Tp, class ..._Args> -struct __pointer_traits_element_type<_Sp<_Tp, _Args...>, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Sp<_Tp, _Args...>::element_type type; -}; - -template <template <class, class...> class _Sp, class _Tp, class ..._Args> -struct __pointer_traits_element_type<_Sp<_Tp, _Args...>, false> -{ - typedef _LIBCPP_NODEBUG_TYPE _Tp type; -}; - -#else // _LIBCPP_HAS_NO_VARIADICS - -template <template <class> class _Sp, class _Tp> -struct __pointer_traits_element_type<_Sp<_Tp>, true> -{ - typedef typename _Sp<_Tp>::element_type type; -}; - -template <template <class> class _Sp, class _Tp> -struct __pointer_traits_element_type<_Sp<_Tp>, false> -{ - typedef _Tp type; -}; - -template <template <class, class> class _Sp, class _Tp, class _A0> -struct __pointer_traits_element_type<_Sp<_Tp, _A0>, true> -{ - typedef typename _Sp<_Tp, _A0>::element_type type; -}; - -template <template <class, class> class _Sp, class _Tp, class _A0> -struct __pointer_traits_element_type<_Sp<_Tp, _A0>, false> -{ - typedef _Tp type; -}; - -template <template <class, class, class> class _Sp, class _Tp, class _A0, class _A1> -struct __pointer_traits_element_type<_Sp<_Tp, _A0, _A1>, true> -{ - typedef typename _Sp<_Tp, _A0, _A1>::element_type type; -}; - -template <template <class, class, class> class _Sp, class _Tp, class _A0, class _A1> -struct __pointer_traits_element_type<_Sp<_Tp, _A0, _A1>, false> -{ - typedef _Tp type; -}; - -template <template <class, class, class, class> class _Sp, class _Tp, class _A0, - class _A1, class _A2> -struct __pointer_traits_element_type<_Sp<_Tp, _A0, _A1, _A2>, true> -{ - typedef typename _Sp<_Tp, _A0, _A1, _A2>::element_type type; -}; - -template <template <class, class, class, class> class _Sp, class _Tp, class _A0, - class _A1, class _A2> -struct __pointer_traits_element_type<_Sp<_Tp, _A0, _A1, _A2>, false> -{ - typedef _Tp type; -}; - -#endif // _LIBCPP_HAS_NO_VARIADICS - -template <class _Tp, class = void> -struct __has_difference_type : false_type {}; - -template <class _Tp> -struct __has_difference_type<_Tp, - typename __void_t<typename _Tp::difference_type>::type> : true_type {}; - -template <class _Ptr, bool = __has_difference_type<_Ptr>::value> -struct __pointer_traits_difference_type -{ - typedef _LIBCPP_NODEBUG_TYPE ptrdiff_t type; -}; - -template <class _Ptr> -struct __pointer_traits_difference_type<_Ptr, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Ptr::difference_type type; -}; - -template <class _Tp, class _Up> -struct __has_rebind -{ -private: - struct __two {char __lx; char __lxx;}; - template <class _Xp> static __two __test(...); - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - template <class _Xp> static char __test(typename _Xp::template rebind<_Up>* = 0); - _LIBCPP_SUPPRESS_DEPRECATED_POP -public: - static const bool value = sizeof(__test<_Tp>(0)) == 1; -}; - -template <class _Tp, class _Up, bool = __has_rebind<_Tp, _Up>::value> -struct __pointer_traits_rebind -{ -#ifndef _LIBCPP_CXX03_LANG - typedef _LIBCPP_NODEBUG_TYPE typename _Tp::template rebind<_Up> type; -#else - typedef _LIBCPP_NODEBUG_TYPE typename _Tp::template rebind<_Up>::other type; -#endif -}; - -#ifndef _LIBCPP_HAS_NO_VARIADICS - -template <template <class, class...> class _Sp, class _Tp, class ..._Args, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _Args...>, _Up, true> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef _LIBCPP_NODEBUG_TYPE typename _Sp<_Tp, _Args...>::template rebind<_Up> type; -#else - typedef _LIBCPP_NODEBUG_TYPE typename _Sp<_Tp, _Args...>::template rebind<_Up>::other type; -#endif -}; - -template <template <class, class...> class _Sp, class _Tp, class ..._Args, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _Args...>, _Up, false> -{ - typedef _Sp<_Up, _Args...> type; -}; - -#else // _LIBCPP_HAS_NO_VARIADICS - -template <template <class> class _Sp, class _Tp, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp>, _Up, true> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef typename _Sp<_Tp>::template rebind<_Up> type; -#else - typedef typename _Sp<_Tp>::template rebind<_Up>::other type; -#endif -}; - -template <template <class> class _Sp, class _Tp, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp>, _Up, false> -{ - typedef _Sp<_Up> type; -}; - -template <template <class, class> class _Sp, class _Tp, class _A0, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _A0>, _Up, true> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef typename _Sp<_Tp, _A0>::template rebind<_Up> type; -#else - typedef typename _Sp<_Tp, _A0>::template rebind<_Up>::other type; -#endif -}; - -template <template <class, class> class _Sp, class _Tp, class _A0, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _A0>, _Up, false> -{ - typedef _Sp<_Up, _A0> type; -}; - -template <template <class, class, class> class _Sp, class _Tp, class _A0, - class _A1, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _A0, _A1>, _Up, true> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef typename _Sp<_Tp, _A0, _A1>::template rebind<_Up> type; -#else - typedef typename _Sp<_Tp, _A0, _A1>::template rebind<_Up>::other type; -#endif -}; - -template <template <class, class, class> class _Sp, class _Tp, class _A0, - class _A1, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _A0, _A1>, _Up, false> -{ - typedef _Sp<_Up, _A0, _A1> type; -}; - -template <template <class, class, class, class> class _Sp, class _Tp, class _A0, - class _A1, class _A2, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _A0, _A1, _A2>, _Up, true> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef typename _Sp<_Tp, _A0, _A1, _A2>::template rebind<_Up> type; -#else - typedef typename _Sp<_Tp, _A0, _A1, _A2>::template rebind<_Up>::other type; -#endif -}; - -template <template <class, class, class, class> class _Sp, class _Tp, class _A0, - class _A1, class _A2, class _Up> -struct __pointer_traits_rebind<_Sp<_Tp, _A0, _A1, _A2>, _Up, false> -{ - typedef _Sp<_Up, _A0, _A1, _A2> type; -}; - -#endif // _LIBCPP_HAS_NO_VARIADICS - -template <class _Ptr> -struct _LIBCPP_TEMPLATE_VIS pointer_traits -{ - typedef _Ptr pointer; - typedef typename __pointer_traits_element_type<pointer>::type element_type; - typedef typename __pointer_traits_difference_type<pointer>::type difference_type; - -#ifndef _LIBCPP_CXX03_LANG - template <class _Up> using rebind = typename __pointer_traits_rebind<pointer, _Up>::type; -#else - template <class _Up> struct rebind - {typedef typename __pointer_traits_rebind<pointer, _Up>::type other;}; -#endif // _LIBCPP_CXX03_LANG - -private: - struct __nat {}; -public: - _LIBCPP_INLINE_VISIBILITY - static pointer pointer_to(typename conditional<is_void<element_type>::value, - __nat, element_type>::type& __r) - {return pointer::pointer_to(__r);} -}; - -template <class _Tp> -struct _LIBCPP_TEMPLATE_VIS pointer_traits<_Tp*> -{ - typedef _Tp* pointer; - typedef _Tp element_type; - typedef ptrdiff_t difference_type; - -#ifndef _LIBCPP_CXX03_LANG - template <class _Up> using rebind = _Up*; -#else - template <class _Up> struct rebind {typedef _Up* other;}; -#endif - -private: - struct __nat {}; -public: - _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 - static pointer pointer_to(typename conditional<is_void<element_type>::value, - __nat, element_type>::type& __r) _NOEXCEPT - {return _VSTD::addressof(__r);} -}; - -template <class _From, class _To> -struct __rebind_pointer { -#ifndef _LIBCPP_CXX03_LANG - typedef typename pointer_traits<_From>::template rebind<_To> type; -#else - typedef typename pointer_traits<_From>::template rebind<_To>::other type; -#endif -}; - -// allocator_traits - -template <class _Tp, class = void> -struct __has_pointer_type : false_type {}; - -template <class _Tp> -struct __has_pointer_type<_Tp, - typename __void_t<typename _Tp::pointer>::type> : true_type {}; - -namespace __pointer_type_imp -{ - -template <class _Tp, class _Dp, bool = __has_pointer_type<_Dp>::value> -struct __pointer_type -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Dp::pointer type; -}; - -template <class _Tp, class _Dp> -struct __pointer_type<_Tp, _Dp, false> -{ - typedef _LIBCPP_NODEBUG_TYPE _Tp* type; -}; - -} // __pointer_type_imp - -template <class _Tp, class _Dp> -struct __pointer_type -{ - typedef _LIBCPP_NODEBUG_TYPE typename __pointer_type_imp::__pointer_type<_Tp, typename remove_reference<_Dp>::type>::type type; -}; - -template <class _Tp, class = void> -struct __has_const_pointer : false_type {}; - -template <class _Tp> -struct __has_const_pointer<_Tp, - typename __void_t<typename _Tp::const_pointer>::type> : true_type {}; - -template <class _Tp, class _Ptr, class _Alloc, bool = __has_const_pointer<_Alloc>::value> -struct __const_pointer -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::const_pointer type; -}; - -template <class _Tp, class _Ptr, class _Alloc> -struct __const_pointer<_Tp, _Ptr, _Alloc, false> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef _LIBCPP_NODEBUG_TYPE typename pointer_traits<_Ptr>::template rebind<const _Tp> type; -#else - typedef typename pointer_traits<_Ptr>::template rebind<const _Tp>::other type; -#endif -}; - -template <class _Tp, class = void> -struct __has_void_pointer : false_type {}; - -template <class _Tp> -struct __has_void_pointer<_Tp, - typename __void_t<typename _Tp::void_pointer>::type> : true_type {}; - -template <class _Ptr, class _Alloc, bool = __has_void_pointer<_Alloc>::value> -struct __void_pointer -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::void_pointer type; -}; - -template <class _Ptr, class _Alloc> -struct __void_pointer<_Ptr, _Alloc, false> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef _LIBCPP_NODEBUG_TYPE typename pointer_traits<_Ptr>::template rebind<void> type; -#else - typedef _LIBCPP_NODEBUG_TYPE typename pointer_traits<_Ptr>::template rebind<void>::other type; -#endif -}; - -template <class _Tp, class = void> -struct __has_const_void_pointer : false_type {}; - -template <class _Tp> -struct __has_const_void_pointer<_Tp, - typename __void_t<typename _Tp::const_void_pointer>::type> : true_type {}; - -template <class _Ptr, class _Alloc, bool = __has_const_void_pointer<_Alloc>::value> -struct __const_void_pointer -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::const_void_pointer type; -}; - -template <class _Ptr, class _Alloc> -struct __const_void_pointer<_Ptr, _Alloc, false> -{ -#ifndef _LIBCPP_CXX03_LANG - typedef _LIBCPP_NODEBUG_TYPE typename pointer_traits<_Ptr>::template rebind<const void> type; -#else - typedef _LIBCPP_NODEBUG_TYPE typename pointer_traits<_Ptr>::template rebind<const void>::other type; -#endif -}; - - template <bool _UsePointerTraits> struct __to_address_helper; template <> struct __to_address_helper<true> { template <class _Pointer> - using __return_type = decltype(pointer_traits<_Pointer>::to_address(std::declval<const _Pointer&>())); + using __return_type = decltype(pointer_traits<_Pointer>::to_address(_VSTD::declval<const _Pointer&>())); template <class _Pointer> _LIBCPP_CONSTEXPR @@ -1157,7 +761,7 @@ template <> struct __to_address_helper<false> { template <class _Pointer> _LIBCPP_CONSTEXPR static __return_type<_Pointer> - __do_it(const _Pointer &__p) _NOEXCEPT { return std::__to_address(__p.operator->()); } + __do_it(const _Pointer &__p) _NOEXCEPT { return _VSTD::__to_address(__p.operator->()); } }; @@ -1172,7 +776,7 @@ to_address(_Tp* __p) _NOEXCEPT } template <class _Pointer> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY constexpr auto to_address(const _Pointer& __p) _NOEXCEPT { @@ -1180,765 +784,287 @@ to_address(const _Pointer& __p) _NOEXCEPT } #endif -template <class _Tp, class = void> -struct __has_size_type : false_type {}; - -template <class _Tp> -struct __has_size_type<_Tp, - typename __void_t<typename _Tp::size_type>::type> : true_type {}; - -template <class _Alloc, class _DiffType, bool = __has_size_type<_Alloc>::value> -struct __size_type -{ - typedef _LIBCPP_NODEBUG_TYPE typename make_unsigned<_DiffType>::type type; -}; - -template <class _Alloc, class _DiffType> -struct __size_type<_Alloc, _DiffType, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::size_type type; -}; - -template <class _Tp, class = void> -struct __has_propagate_on_container_copy_assignment : false_type {}; - -template <class _Tp> -struct __has_propagate_on_container_copy_assignment<_Tp, - typename __void_t<typename _Tp::propagate_on_container_copy_assignment>::type> - : true_type {}; - -template <class _Alloc, bool = __has_propagate_on_container_copy_assignment<_Alloc>::value> -struct __propagate_on_container_copy_assignment -{ - typedef _LIBCPP_NODEBUG_TYPE false_type type; -}; - -template <class _Alloc> -struct __propagate_on_container_copy_assignment<_Alloc, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::propagate_on_container_copy_assignment type; -}; - -template <class _Tp, class = void> -struct __has_propagate_on_container_move_assignment : false_type {}; - -template <class _Tp> -struct __has_propagate_on_container_move_assignment<_Tp, - typename __void_t<typename _Tp::propagate_on_container_move_assignment>::type> - : true_type {}; - -template <class _Alloc, bool = __has_propagate_on_container_move_assignment<_Alloc>::value> -struct __propagate_on_container_move_assignment -{ - typedef false_type type; -}; - -template <class _Alloc> -struct __propagate_on_container_move_assignment<_Alloc, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::propagate_on_container_move_assignment type; -}; - -template <class _Tp, class = void> -struct __has_propagate_on_container_swap : false_type {}; - -template <class _Tp> -struct __has_propagate_on_container_swap<_Tp, - typename __void_t<typename _Tp::propagate_on_container_swap>::type> - : true_type {}; - -template <class _Alloc, bool = __has_propagate_on_container_swap<_Alloc>::value> -struct __propagate_on_container_swap -{ - typedef false_type type; -}; - -template <class _Alloc> -struct __propagate_on_container_swap<_Alloc, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::propagate_on_container_swap type; -}; - -template <class _Tp, class = void> -struct __has_is_always_equal : false_type {}; - -template <class _Tp> -struct __has_is_always_equal<_Tp, - typename __void_t<typename _Tp::is_always_equal>::type> - : true_type {}; - -template <class _Alloc, bool = __has_is_always_equal<_Alloc>::value> -struct __is_always_equal -{ - typedef _LIBCPP_NODEBUG_TYPE typename _VSTD::is_empty<_Alloc>::type type; -}; - -template <class _Alloc> -struct __is_always_equal<_Alloc, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::is_always_equal type; -}; +template <class _Tp> class allocator; -template <class _Tp, class _Up, bool = __has_rebind<_Tp, _Up>::value> -struct __has_rebind_other +#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) +template <> +class _LIBCPP_TEMPLATE_VIS _LIBCPP_DEPRECATED_IN_CXX17 allocator<void> { -private: - struct __two {char __lx; char __lxx;}; - template <class _Xp> static __two __test(...); - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - template <class _Xp> static char __test(typename _Xp::template rebind<_Up>::other* = 0); - _LIBCPP_SUPPRESS_DEPRECATED_POP public: - static const bool value = sizeof(__test<_Tp>(0)) == 1; -}; - -template <class _Tp, class _Up> -struct __has_rebind_other<_Tp, _Up, false> -{ - static const bool value = false; -}; - -template <class _Tp, class _Up, bool = __has_rebind_other<_Tp, _Up>::value> -struct __allocator_traits_rebind -{ - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - typedef _LIBCPP_NODEBUG_TYPE typename _Tp::template rebind<_Up>::other type; - _LIBCPP_SUPPRESS_DEPRECATED_POP -}; - -template <template <class, class...> class _Alloc, class _Tp, class ..._Args, class _Up> -struct __allocator_traits_rebind<_Alloc<_Tp, _Args...>, _Up, true> -{ - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc<_Tp, _Args...>::template rebind<_Up>::other type; - _LIBCPP_SUPPRESS_DEPRECATED_POP -}; - -template <template <class, class...> class _Alloc, class _Tp, class ..._Args, class _Up> -struct __allocator_traits_rebind<_Alloc<_Tp, _Args...>, _Up, false> -{ - typedef _LIBCPP_NODEBUG_TYPE _Alloc<_Up, _Args...> type; -}; - -#ifndef _LIBCPP_CXX03_LANG - -_LIBCPP_SUPPRESS_DEPRECATED_PUSH -template <class _Alloc, class _SizeType, class _ConstVoidPtr> -auto -__has_allocate_hint_test(_Alloc&& __a, _SizeType&& __sz, _ConstVoidPtr&& __p) - -> decltype((void)__a.allocate(__sz, __p), true_type()); -_LIBCPP_SUPPRESS_DEPRECATED_POP - -template <class _Alloc, class _SizeType, class _ConstVoidPtr> -auto -__has_allocate_hint_test(const _Alloc& __a, _SizeType&& __sz, _ConstVoidPtr&& __p) - -> false_type; - -template <class _Alloc, class _SizeType, class _ConstVoidPtr> -struct __has_allocate_hint - : decltype(_VSTD::__has_allocate_hint_test(declval<_Alloc>(), - declval<_SizeType>(), - declval<_ConstVoidPtr>())) -{ -}; - -#else // _LIBCPP_CXX03_LANG - -template <class _Alloc, class _SizeType, class _ConstVoidPtr> -struct __has_allocate_hint - : true_type -{ -}; - -#endif // _LIBCPP_CXX03_LANG - -_LIBCPP_SUPPRESS_DEPRECATED_PUSH -template <class _Alloc, class ..._Args, - class = decltype(_VSTD::declval<_Alloc>().construct(_VSTD::declval<_Args>()...))> -static true_type __test_has_construct(int); -_LIBCPP_SUPPRESS_DEPRECATED_POP - -template <class _Alloc, class...> -static false_type __test_has_construct(...); - -template <class _Alloc, class ..._Args> -struct __has_construct : decltype(__test_has_construct<_Alloc, _Args...>(0)) {}; - -#if !defined(_LIBCPP_CXX03_LANG) - -_LIBCPP_SUPPRESS_DEPRECATED_PUSH -template <class _Alloc, class _Pointer> -auto -__has_destroy_test(_Alloc&& __a, _Pointer&& __p) - -> decltype(__a.destroy(__p), true_type()); -_LIBCPP_SUPPRESS_DEPRECATED_POP - -template <class _Alloc, class _Pointer> -auto -__has_destroy_test(const _Alloc& __a, _Pointer&& __p) - -> false_type; - -template <class _Alloc, class _Pointer> -struct __has_destroy - : decltype(_VSTD::__has_destroy_test(declval<_Alloc>(), - declval<_Pointer>())) -{ -}; - -_LIBCPP_SUPPRESS_DEPRECATED_PUSH -template <class _Alloc> -auto -__has_max_size_test(_Alloc&& __a) - -> decltype(__a.max_size(), true_type()); -_LIBCPP_SUPPRESS_DEPRECATED_POP - -template <class _Alloc> -auto -__has_max_size_test(const volatile _Alloc& __a) - -> false_type; - -template <class _Alloc> -struct __has_max_size - : decltype(_VSTD::__has_max_size_test(declval<_Alloc&>())) -{ -}; - -template <class _Alloc> -auto -__has_select_on_container_copy_construction_test(_Alloc&& __a) - -> decltype(__a.select_on_container_copy_construction(), true_type()); - -template <class _Alloc> -auto -__has_select_on_container_copy_construction_test(const volatile _Alloc& __a) - -> false_type; - -template <class _Alloc> -struct __has_select_on_container_copy_construction - : decltype(_VSTD::__has_select_on_container_copy_construction_test(declval<_Alloc&>())) -{ -}; - -#else // _LIBCPP_CXX03_LANG - -template <class _Alloc, class _Pointer, class = void> -struct __has_destroy : false_type {}; - -template <class _Alloc, class _Pointer> -struct __has_destroy<_Alloc, _Pointer, typename __void_t< - decltype(_VSTD::declval<_Alloc>().destroy(_VSTD::declval<_Pointer>())) ->::type> : std::true_type {}; - -template <class _Alloc> -struct __has_max_size - : true_type -{ -}; + typedef void* pointer; + typedef const void* const_pointer; + typedef void value_type; -template <class _Alloc> -struct __has_select_on_container_copy_construction - : false_type -{ + template <class _Up> struct rebind {typedef allocator<_Up> other;}; }; -#endif // _LIBCPP_CXX03_LANG - -template <class _Alloc, class _Ptr, bool = __has_difference_type<_Alloc>::value> -struct __alloc_traits_difference_type +template <> +class _LIBCPP_TEMPLATE_VIS _LIBCPP_DEPRECATED_IN_CXX17 allocator<const void> { - typedef _LIBCPP_NODEBUG_TYPE typename pointer_traits<_Ptr>::difference_type type; -}; +public: + typedef const void* pointer; + typedef const void* const_pointer; + typedef const void value_type; -template <class _Alloc, class _Ptr> -struct __alloc_traits_difference_type<_Alloc, _Ptr, true> -{ - typedef _LIBCPP_NODEBUG_TYPE typename _Alloc::difference_type type; + template <class _Up> struct rebind {typedef allocator<_Up> other;}; }; +#endif -template <class _Tp> -struct __is_default_allocator : false_type {}; +// allocator template <class _Tp> -struct __is_default_allocator<_VSTD::allocator<_Tp> > : true_type {}; - - - -template <class _Alloc, - bool = __has_construct<_Alloc, typename _Alloc::value_type*, typename _Alloc::value_type&&>::value && !__is_default_allocator<_Alloc>::value - > -struct __is_cpp17_move_insertable; -template <class _Alloc> -struct __is_cpp17_move_insertable<_Alloc, true> : std::true_type {}; -template <class _Alloc> -struct __is_cpp17_move_insertable<_Alloc, false> : std::is_move_constructible<typename _Alloc::value_type> {}; - -template <class _Alloc, - bool = __has_construct<_Alloc, typename _Alloc::value_type*, const typename _Alloc::value_type&>::value && !__is_default_allocator<_Alloc>::value - > -struct __is_cpp17_copy_insertable; -template <class _Alloc> -struct __is_cpp17_copy_insertable<_Alloc, true> : __is_cpp17_move_insertable<_Alloc> {}; -template <class _Alloc> -struct __is_cpp17_copy_insertable<_Alloc, false> : integral_constant<bool, - std::is_copy_constructible<typename _Alloc::value_type>::value && - __is_cpp17_move_insertable<_Alloc>::value> - {}; - - - -template <class _Alloc> -struct _LIBCPP_TEMPLATE_VIS allocator_traits +class _LIBCPP_TEMPLATE_VIS allocator { - typedef _Alloc allocator_type; - typedef typename allocator_type::value_type value_type; - - typedef typename __pointer_type<value_type, allocator_type>::type pointer; - typedef typename __const_pointer<value_type, pointer, allocator_type>::type const_pointer; - typedef typename __void_pointer<pointer, allocator_type>::type void_pointer; - typedef typename __const_void_pointer<pointer, allocator_type>::type const_void_pointer; - - typedef typename __alloc_traits_difference_type<allocator_type, pointer>::type difference_type; - typedef typename __size_type<allocator_type, difference_type>::type size_type; - - typedef typename __propagate_on_container_copy_assignment<allocator_type>::type - propagate_on_container_copy_assignment; - typedef typename __propagate_on_container_move_assignment<allocator_type>::type - propagate_on_container_move_assignment; - typedef typename __propagate_on_container_swap<allocator_type>::type - propagate_on_container_swap; - typedef typename __is_always_equal<allocator_type>::type - is_always_equal; - -#ifndef _LIBCPP_CXX03_LANG - template <class _Tp> using rebind_alloc = - typename __allocator_traits_rebind<allocator_type, _Tp>::type; - template <class _Tp> using rebind_traits = allocator_traits<rebind_alloc<_Tp> >; -#else // _LIBCPP_CXX03_LANG - template <class _Tp> struct rebind_alloc - {typedef typename __allocator_traits_rebind<allocator_type, _Tp>::type other;}; - template <class _Tp> struct rebind_traits - {typedef allocator_traits<typename rebind_alloc<_Tp>::other> other;}; -#endif // _LIBCPP_CXX03_LANG - - _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY - static pointer allocate(allocator_type& __a, size_type __n) - {return __a.allocate(__n);} - _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY - static pointer allocate(allocator_type& __a, size_type __n, const_void_pointer __hint) - {return __allocate(__a, __n, __hint, - __has_allocate_hint<allocator_type, size_type, const_void_pointer>());} - - _LIBCPP_INLINE_VISIBILITY - static void deallocate(allocator_type& __a, pointer __p, size_type __n) _NOEXCEPT - {__a.deallocate(__p, __n);} - - template <class _Tp, class... _Args> - _LIBCPP_INLINE_VISIBILITY - static void construct(allocator_type& __a, _Tp* __p, _Args&&... __args) - {__construct(__has_construct<allocator_type, _Tp*, _Args...>(), - __a, __p, _VSTD::forward<_Args>(__args)...);} - - template <class _Tp> - _LIBCPP_INLINE_VISIBILITY - static void destroy(allocator_type& __a, _Tp* __p) - {__destroy(__has_destroy<allocator_type, _Tp*>(), __a, __p);} - - _LIBCPP_INLINE_VISIBILITY - static size_type max_size(const allocator_type& __a) _NOEXCEPT - {return __max_size(__has_max_size<const allocator_type>(), __a);} - - _LIBCPP_INLINE_VISIBILITY - static allocator_type - select_on_container_copy_construction(const allocator_type& __a) - {return __select_on_container_copy_construction( - __has_select_on_container_copy_construction<const allocator_type>(), - __a);} - - template <class _Ptr> - _LIBCPP_INLINE_VISIBILITY - static - void - __construct_forward_with_exception_guarantees(allocator_type& __a, _Ptr __begin1, _Ptr __end1, _Ptr& __begin2) - { - static_assert(__is_cpp17_move_insertable<allocator_type>::value, - "The specified type does not meet the requirements of Cpp17MoveInsertible"); - for (; __begin1 != __end1; ++__begin1, (void) ++__begin2) - construct(__a, _VSTD::__to_address(__begin2), -#ifdef _LIBCPP_NO_EXCEPTIONS - _VSTD::move(*__begin1) -#else - _VSTD::move_if_noexcept(*__begin1) -#endif - ); - } - - template <class _Tp> - _LIBCPP_INLINE_VISIBILITY - static - typename enable_if - < - (__is_default_allocator<allocator_type>::value - || !__has_construct<allocator_type, _Tp*, _Tp>::value) && - is_trivially_move_constructible<_Tp>::value, - void - >::type - __construct_forward_with_exception_guarantees(allocator_type&, _Tp* __begin1, _Tp* __end1, _Tp*& __begin2) - { - ptrdiff_t _Np = __end1 - __begin1; - if (_Np > 0) - { - _VSTD::memcpy(__begin2, __begin1, _Np * sizeof(_Tp)); - __begin2 += _Np; - } - } - - template <class _Iter, class _Ptr> - _LIBCPP_INLINE_VISIBILITY - static - void - __construct_range_forward(allocator_type& __a, _Iter __begin1, _Iter __end1, _Ptr& __begin2) - { - for (; __begin1 != __end1; ++__begin1, (void) ++__begin2) - construct(__a, _VSTD::__to_address(__begin2), *__begin1); - } +public: + typedef size_t size_type; + typedef ptrdiff_t difference_type; + typedef _Tp value_type; + typedef true_type propagate_on_container_move_assignment; + typedef true_type is_always_equal; - template <class _SourceTp, class _DestTp, - class _RawSourceTp = typename remove_const<_SourceTp>::type, - class _RawDestTp = typename remove_const<_DestTp>::type> - _LIBCPP_INLINE_VISIBILITY - static - typename enable_if - < - is_trivially_copy_constructible<_DestTp>::value && - is_same<_RawSourceTp, _RawDestTp>::value && - (__is_default_allocator<allocator_type>::value || - !__has_construct<allocator_type, _DestTp*, _SourceTp&>::value), - void - >::type - __construct_range_forward(allocator_type&, _SourceTp* __begin1, _SourceTp* __end1, _DestTp*& __begin2) - { - ptrdiff_t _Np = __end1 - __begin1; - if (_Np > 0) - { - _VSTD::memcpy(const_cast<_RawDestTp*>(__begin2), __begin1, _Np * sizeof(_DestTp)); - __begin2 += _Np; - } - } + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + allocator() _NOEXCEPT { } - template <class _Ptr> - _LIBCPP_INLINE_VISIBILITY - static - void - __construct_backward_with_exception_guarantees(allocator_type& __a, _Ptr __begin1, _Ptr __end1, _Ptr& __end2) - { - static_assert(__is_cpp17_move_insertable<allocator_type>::value, - "The specified type does not meet the requirements of Cpp17MoveInsertable"); - while (__end1 != __begin1) - { - construct(__a, _VSTD::__to_address(__end2 - 1), -#ifdef _LIBCPP_NO_EXCEPTIONS - _VSTD::move(*--__end1) -#else - _VSTD::move_if_noexcept(*--__end1) -#endif - ); - --__end2; - } - } + template <class _Up> + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + allocator(const allocator<_Up>&) _NOEXCEPT { } - template <class _Tp> - _LIBCPP_INLINE_VISIBILITY - static - typename enable_if - < - (__is_default_allocator<allocator_type>::value - || !__has_construct<allocator_type, _Tp*, _Tp>::value) && - is_trivially_move_constructible<_Tp>::value, - void - >::type - __construct_backward_with_exception_guarantees(allocator_type&, _Tp* __begin1, _Tp* __end1, _Tp*& __end2) - { - ptrdiff_t _Np = __end1 - __begin1; - __end2 -= _Np; - if (_Np > 0) - _VSTD::memcpy(__end2, __begin1, _Np * sizeof(_Tp)); + _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + _Tp* allocate(size_t __n) { + if (__n > allocator_traits<allocator>::max_size(*this)) + __throw_length_error("allocator<T>::allocate(size_t n)" + " 'n' exceeds maximum supported size"); + if (__libcpp_is_constant_evaluated()) { + return static_cast<_Tp*>(::operator new(__n * sizeof(_Tp))); + } else { + return static_cast<_Tp*>(_VSTD::__libcpp_allocate(__n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp))); } + } -private: - - _LIBCPP_INLINE_VISIBILITY - static pointer __allocate(allocator_type& __a, size_type __n, - const_void_pointer __hint, true_type) - { - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - return __a.allocate(__n, __hint); - _LIBCPP_SUPPRESS_DEPRECATED_POP + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + void deallocate(_Tp* __p, size_t __n) _NOEXCEPT { + if (__libcpp_is_constant_evaluated()) { + ::operator delete(__p); + } else { + _VSTD::__libcpp_deallocate((void*)__p, __n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp)); } - _LIBCPP_INLINE_VISIBILITY - static pointer __allocate(allocator_type& __a, size_type __n, - const_void_pointer, false_type) - {return __a.allocate(__n);} + } - template <class _Tp, class... _Args> - _LIBCPP_INLINE_VISIBILITY - static void __construct(true_type, allocator_type& __a, _Tp* __p, _Args&&... __args) - { - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - __a.construct(__p, _VSTD::forward<_Args>(__args)...); - _LIBCPP_SUPPRESS_DEPRECATED_POP - } + // C++20 Removed members +#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) + _LIBCPP_DEPRECATED_IN_CXX17 typedef _Tp* pointer; + _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp* const_pointer; + _LIBCPP_DEPRECATED_IN_CXX17 typedef _Tp& reference; + _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp& const_reference; - template <class _Tp, class... _Args> - _LIBCPP_INLINE_VISIBILITY - static void __construct(false_type, allocator_type&, _Tp* __p, _Args&&... __args) - { - ::new ((void*)__p) _Tp(_VSTD::forward<_Args>(__args)...); - } + template <class _Up> + struct _LIBCPP_DEPRECATED_IN_CXX17 rebind { + typedef allocator<_Up> other; + }; - template <class _Tp> - _LIBCPP_INLINE_VISIBILITY - static void __destroy(true_type, allocator_type& __a, _Tp* __p) - { - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - __a.destroy(__p); - _LIBCPP_SUPPRESS_DEPRECATED_POP - } - template <class _Tp> - _LIBCPP_INLINE_VISIBILITY - static void __destroy(false_type, allocator_type&, _Tp* __p) - { - __p->~_Tp(); - } + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY + pointer address(reference __x) const _NOEXCEPT { + return _VSTD::addressof(__x); + } + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY + const_pointer address(const_reference __x) const _NOEXCEPT { + return _VSTD::addressof(__x); + } - _LIBCPP_INLINE_VISIBILITY - static size_type __max_size(true_type, const allocator_type& __a) _NOEXCEPT - { - _LIBCPP_SUPPRESS_DEPRECATED_PUSH - return __a.max_size(); - _LIBCPP_SUPPRESS_DEPRECATED_POP - } + _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_DEPRECATED_IN_CXX17 + _Tp* allocate(size_t __n, const void*) { + return allocate(__n); + } - _LIBCPP_INLINE_VISIBILITY - static size_type __max_size(false_type, const allocator_type&) _NOEXCEPT - {return numeric_limits<size_type>::max() / sizeof(value_type);} + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT { + return size_type(~0) / sizeof(_Tp); + } - _LIBCPP_INLINE_VISIBILITY - static allocator_type - __select_on_container_copy_construction(true_type, const allocator_type& __a) - {return __a.select_on_container_copy_construction();} - _LIBCPP_INLINE_VISIBILITY - static allocator_type - __select_on_container_copy_construction(false_type, const allocator_type& __a) - {return __a;} -}; + template <class _Up, class... _Args> + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY + void construct(_Up* __p, _Args&&... __args) { + ::new ((void*)__p) _Up(_VSTD::forward<_Args>(__args)...); + } -template <class _Traits, class _Tp> -struct __rebind_alloc_helper -{ -#ifndef _LIBCPP_CXX03_LANG - typedef _LIBCPP_NODEBUG_TYPE typename _Traits::template rebind_alloc<_Tp> type; -#else - typedef typename _Traits::template rebind_alloc<_Tp>::other type; + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY + void destroy(pointer __p) { + __p->~_Tp(); + } #endif }; -// allocator - template <class _Tp> -class _LIBCPP_TEMPLATE_VIS allocator +class _LIBCPP_TEMPLATE_VIS allocator<const _Tp> { public: -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) - _LIBCPP_DEPRECATED_IN_CXX17 typedef size_t size_type; - _LIBCPP_DEPRECATED_IN_CXX17 typedef ptrdiff_t difference_type; - _LIBCPP_DEPRECATED_IN_CXX17 typedef _Tp* pointer; - _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp* const_pointer; - _LIBCPP_DEPRECATED_IN_CXX17 typedef _Tp& reference; - _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp& const_reference; - - template <class _Up> struct _LIBCPP_DEPRECATED_IN_CXX17 rebind {typedef allocator<_Up> other;}; -#endif - - typedef _Tp value_type; - - typedef true_type propagate_on_container_move_assignment; - typedef true_type is_always_equal; + typedef size_t size_type; + typedef ptrdiff_t difference_type; + typedef const _Tp value_type; + typedef true_type propagate_on_container_move_assignment; + typedef true_type is_always_equal; _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 - allocator() _NOEXCEPT {} + allocator() _NOEXCEPT { } template <class _Up> _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 - allocator(const allocator<_Up>&) _NOEXCEPT {} - -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY - pointer address(reference __x) const _NOEXCEPT - {return _VSTD::addressof(__x);} - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY - const_pointer address(const_reference __x) const _NOEXCEPT - {return _VSTD::addressof(__x);} -#endif + allocator(const allocator<_Up>&) _NOEXCEPT { } - _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _Tp* allocate(size_t __n) - { - // TODO(mpark): Replace with `allocator_traits<allocator>::max_size(*this)`. - if (__n > (size_t(~0) / sizeof(_Tp))) - __throw_length_error("allocator<T>::allocate(size_t n)" + _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + const _Tp* allocate(size_t __n) { + if (__n > allocator_traits<allocator>::max_size(*this)) + __throw_length_error("allocator<const T>::allocate(size_t n)" " 'n' exceeds maximum supported size"); - return static_cast<_Tp*>(_VSTD::__libcpp_allocate(__n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp))); + if (__libcpp_is_constant_evaluated()) { + return static_cast<const _Tp*>(::operator new(__n * sizeof(_Tp))); + } else { + return static_cast<const _Tp*>(_VSTD::__libcpp_allocate(__n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp))); } + } -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) - _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_DEPRECATED_IN_CXX17 - _Tp* allocate(size_t __n, const void*) { return allocate(__n); } -#endif - - _LIBCPP_INLINE_VISIBILITY void deallocate(_Tp* __p, size_t __n) _NOEXCEPT - {_VSTD::__libcpp_deallocate((void*)__p, __n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp));} - -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT - {return size_type(~0) / sizeof(_Tp);} - - template <class _Up, class... _Args> - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY - void - construct(_Up* __p, _Args&&... __args) - { - ::new((void*)__p) _Up(_VSTD::forward<_Args>(__args)...); + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + void deallocate(const _Tp* __p, size_t __n) { + if (__libcpp_is_constant_evaluated()) { + ::operator delete(const_cast<_Tp*>(__p)); + } else { + _VSTD::__libcpp_deallocate((void*) const_cast<_Tp *>(__p), __n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp)); } - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY void destroy(pointer __p) {__p->~_Tp();} -#endif -}; + } -template <class _Tp> -class _LIBCPP_TEMPLATE_VIS allocator<const _Tp> -{ -public: + // C++20 Removed members #if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) - _LIBCPP_DEPRECATED_IN_CXX17 typedef size_t size_type; - _LIBCPP_DEPRECATED_IN_CXX17 typedef ptrdiff_t difference_type; _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp* pointer; _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp* const_pointer; _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp& reference; _LIBCPP_DEPRECATED_IN_CXX17 typedef const _Tp& const_reference; - template <class _Up> struct _LIBCPP_DEPRECATED_IN_CXX17 rebind {typedef allocator<_Up> other;}; -#endif - - typedef const _Tp value_type; - - typedef true_type propagate_on_container_move_assignment; - typedef true_type is_always_equal; - - _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 - allocator() _NOEXCEPT {} - template <class _Up> - _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 - allocator(const allocator<_Up>&) _NOEXCEPT {} + struct _LIBCPP_DEPRECATED_IN_CXX17 rebind { + typedef allocator<_Up> other; + }; -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY - const_pointer address(const_reference __x) const _NOEXCEPT - {return _VSTD::addressof(__x);} -#endif - - _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY const _Tp* allocate(size_t __n) - { - // TODO(mpark): Replace with `allocator_traits<allocator>::max_size(*this)`. - if (__n > (size_t(~0) / sizeof(_Tp))) - __throw_length_error("allocator<const T>::allocate(size_t n)" - " 'n' exceeds maximum supported size"); - return static_cast<const _Tp*>(_VSTD::__libcpp_allocate(__n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp))); + const_pointer address(const_reference __x) const _NOEXCEPT { + return _VSTD::addressof(__x); } -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY _LIBCPP_DEPRECATED_IN_CXX17 - const _Tp* allocate(size_t __n, const void*) { return allocate(__n); } -#endif - - _LIBCPP_INLINE_VISIBILITY void deallocate(const _Tp* __p, size_t __n) - {_VSTD::__libcpp_deallocate((void*) const_cast<_Tp *>(__p), __n * sizeof(_Tp), _LIBCPP_ALIGNOF(_Tp));} + const _Tp* allocate(size_t __n, const void*) { + return allocate(__n); + } -#if _LIBCPP_STD_VER <= 17 || defined(_LIBCPP_ENABLE_CXX20_REMOVED_ALLOCATOR_MEMBERS) - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT - {return size_type(~0) / sizeof(_Tp);} + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY size_type max_size() const _NOEXCEPT { + return size_type(~0) / sizeof(_Tp); + } -#if !defined(_LIBCPP_HAS_NO_RVALUE_REFERENCES) && !defined(_LIBCPP_HAS_NO_VARIADICS) template <class _Up, class... _Args> - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY - void - construct(_Up* __p, _Args&&... __args) - { - ::new((void*)__p) _Up(_VSTD::forward<_Args>(__args)...); - } -#else // !defined(_LIBCPP_HAS_NO_RVALUE_REFERENCES) && !defined(_LIBCPP_HAS_NO_VARIADICS) - _LIBCPP_INLINE_VISIBILITY - void - construct(pointer __p) - { - ::new((void*) const_cast<_Tp *>(__p)) _Tp(); - } -# if defined(_LIBCPP_HAS_NO_RVALUE_REFERENCES) + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY + void construct(_Up* __p, _Args&&... __args) { + ::new ((void*)__p) _Up(_VSTD::forward<_Args>(__args)...); + } - template <class _A0> - _LIBCPP_INLINE_VISIBILITY - void - construct(pointer __p, _A0& __a0) - { - ::new((void*) const_cast<_Tp *>(__p)) _Tp(__a0); - } - template <class _A0> - _LIBCPP_INLINE_VISIBILITY - void - construct(pointer __p, const _A0& __a0) - { - ::new((void*) const_cast<_Tp *>(__p)) _Tp(__a0); - } -# endif // defined(_LIBCPP_HAS_NO_RVALUE_REFERENCES) - template <class _A0, class _A1> - _LIBCPP_INLINE_VISIBILITY - void - construct(pointer __p, _A0& __a0, _A1& __a1) - { - ::new((void*) const_cast<_Tp *>(__p)) _Tp(__a0, __a1); - } - template <class _A0, class _A1> - _LIBCPP_INLINE_VISIBILITY - void - construct(pointer __p, const _A0& __a0, _A1& __a1) - { - ::new((void*) const_cast<_Tp *>(__p)) _Tp(__a0, __a1); - } - template <class _A0, class _A1> - _LIBCPP_INLINE_VISIBILITY - void - construct(pointer __p, _A0& __a0, const _A1& __a1) - { - ::new((void*) const_cast<_Tp *>(__p)) _Tp(__a0, __a1); - } - template <class _A0, class _A1> - _LIBCPP_INLINE_VISIBILITY - void - construct(pointer __p, const _A0& __a0, const _A1& __a1) - { - ::new((void*) const_cast<_Tp *>(__p)) _Tp(__a0, __a1); - } -#endif // !defined(_LIBCPP_HAS_NO_RVALUE_REFERENCES) && !defined(_LIBCPP_HAS_NO_VARIADICS) - _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY void destroy(pointer __p) {__p->~_Tp();} + _LIBCPP_DEPRECATED_IN_CXX17 _LIBCPP_INLINE_VISIBILITY + void destroy(pointer __p) { + __p->~_Tp(); + } #endif }; template <class _Tp, class _Up> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 bool operator==(const allocator<_Tp>&, const allocator<_Up>&) _NOEXCEPT {return true;} template <class _Tp, class _Up> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 bool operator!=(const allocator<_Tp>&, const allocator<_Up>&) _NOEXCEPT {return false;} +template <class _Alloc, class _Ptr> +_LIBCPP_INLINE_VISIBILITY +void __construct_forward_with_exception_guarantees(_Alloc& __a, _Ptr __begin1, _Ptr __end1, _Ptr& __begin2) { + static_assert(__is_cpp17_move_insertable<_Alloc>::value, + "The specified type does not meet the requirements of Cpp17MoveInsertable"); + typedef allocator_traits<_Alloc> _Traits; + for (; __begin1 != __end1; ++__begin1, (void)++__begin2) { + _Traits::construct(__a, _VSTD::__to_address(__begin2), +#ifdef _LIBCPP_NO_EXCEPTIONS + _VSTD::move(*__begin1) +#else + _VSTD::move_if_noexcept(*__begin1) +#endif + ); + } +} + +template <class _Alloc, class _Tp, typename enable_if< + (__is_default_allocator<_Alloc>::value || !__has_construct<_Alloc, _Tp*, _Tp>::value) && + is_trivially_move_constructible<_Tp>::value +>::type> +_LIBCPP_INLINE_VISIBILITY +void __construct_forward_with_exception_guarantees(_Alloc&, _Tp* __begin1, _Tp* __end1, _Tp*& __begin2) { + ptrdiff_t _Np = __end1 - __begin1; + if (_Np > 0) { + _VSTD::memcpy(__begin2, __begin1, _Np * sizeof(_Tp)); + __begin2 += _Np; + } +} + +template <class _Alloc, class _Iter, class _Ptr> +_LIBCPP_INLINE_VISIBILITY +void __construct_range_forward(_Alloc& __a, _Iter __begin1, _Iter __end1, _Ptr& __begin2) { + typedef allocator_traits<_Alloc> _Traits; + for (; __begin1 != __end1; ++__begin1, (void) ++__begin2) { + _Traits::construct(__a, _VSTD::__to_address(__begin2), *__begin1); + } +} + +template <class _Alloc, class _Source, class _Dest, + class _RawSource = typename remove_const<_Source>::type, + class _RawDest = typename remove_const<_Dest>::type, + class = + typename enable_if< + is_trivially_copy_constructible<_Dest>::value && + is_same<_RawSource, _RawDest>::value && + (__is_default_allocator<_Alloc>::value || !__has_construct<_Alloc, _Dest*, _Source&>::value) + >::type> +_LIBCPP_INLINE_VISIBILITY +void __construct_range_forward(_Alloc&, _Source* __begin1, _Source* __end1, _Dest*& __begin2) { + ptrdiff_t _Np = __end1 - __begin1; + if (_Np > 0) { + _VSTD::memcpy(const_cast<_RawDest*>(__begin2), __begin1, _Np * sizeof(_Dest)); + __begin2 += _Np; + } +} + +template <class _Alloc, class _Ptr> +_LIBCPP_INLINE_VISIBILITY +void __construct_backward_with_exception_guarantees(_Alloc& __a, _Ptr __begin1, _Ptr __end1, _Ptr& __end2) { + static_assert(__is_cpp17_move_insertable<_Alloc>::value, + "The specified type does not meet the requirements of Cpp17MoveInsertable"); + typedef allocator_traits<_Alloc> _Traits; + while (__end1 != __begin1) { + _Traits::construct(__a, _VSTD::__to_address(__end2 - 1), +#ifdef _LIBCPP_NO_EXCEPTIONS + _VSTD::move(*--__end1) +#else + _VSTD::move_if_noexcept(*--__end1) +#endif + ); + --__end2; + } +} + +template <class _Alloc, class _Tp, class = typename enable_if< + (__is_default_allocator<_Alloc>::value || !__has_construct<_Alloc, _Tp*, _Tp>::value) && + is_trivially_move_constructible<_Tp>::value +>::type> +_LIBCPP_INLINE_VISIBILITY +void __construct_backward_with_exception_guarantees(_Alloc&, _Tp* __begin1, _Tp* __end1, _Tp*& __end2) { + ptrdiff_t _Np = __end1 - __begin1; + __end2 -= _Np; + if (_Np > 0) + _VSTD::memcpy(__end2, __begin1, _Np * sizeof(_Tp)); +} + template <class _OutputIterator, class _Tp> class _LIBCPP_TEMPLATE_VIS raw_storage_iterator : public iterator<output_iterator_tag, @@ -1953,10 +1079,10 @@ public: _LIBCPP_INLINE_VISIBILITY explicit raw_storage_iterator(_OutputIterator __x) : __x_(__x) {} _LIBCPP_INLINE_VISIBILITY raw_storage_iterator& operator*() {return *this;} _LIBCPP_INLINE_VISIBILITY raw_storage_iterator& operator=(const _Tp& __element) - {::new(_VSTD::addressof(*__x_)) _Tp(__element); return *this;} + {::new ((void*)_VSTD::addressof(*__x_)) _Tp(__element); return *this;} #if _LIBCPP_STD_VER >= 14 _LIBCPP_INLINE_VISIBILITY raw_storage_iterator& operator=(_Tp&& __element) - {::new(_VSTD::addressof(*__x_)) _Tp(_VSTD::move(__element)); return *this;} + {::new ((void*)_VSTD::addressof(*__x_)) _Tp(_VSTD::move(__element)); return *this;} #endif _LIBCPP_INLINE_VISIBILITY raw_storage_iterator& operator++() {++__x_; return *this;} _LIBCPP_INLINE_VISIBILITY raw_storage_iterator operator++(int) @@ -1982,8 +1108,8 @@ get_temporary_buffer(ptrdiff_t __n) _NOEXCEPT #if !defined(_LIBCPP_HAS_NO_ALIGNED_ALLOCATION) if (__is_overaligned_for_new(_LIBCPP_ALIGNOF(_Tp))) { - std::align_val_t __al = - std::align_val_t(std::alignment_of<_Tp>::value); + align_val_t __al = + align_val_t(alignment_of<_Tp>::value); __r.first = static_cast<_Tp*>(::operator new( __n * sizeof(_Tp), __al, nothrow)); } else { @@ -2052,7 +1178,7 @@ public: _LIBCPP_INLINE_VISIBILITY _Tp* release() _NOEXCEPT { _Tp* __t = __ptr_; - __ptr_ = 0; + __ptr_ = nullptr; return __t; } _LIBCPP_INLINE_VISIBILITY void reset(_Tp* __p = 0) _NOEXCEPT @@ -2160,19 +1286,19 @@ struct __compressed_pair_elem<_Tp, _Idx, true> : private _Tp { template <class _T1, class _T2> class __compressed_pair : private __compressed_pair_elem<_T1, 0>, private __compressed_pair_elem<_T2, 1> { - typedef _LIBCPP_NODEBUG_TYPE __compressed_pair_elem<_T1, 0> _Base1; - typedef _LIBCPP_NODEBUG_TYPE __compressed_pair_elem<_T2, 1> _Base2; - +public: // NOTE: This static assert should never fire because __compressed_pair // is *almost never* used in a scenario where it's possible for T1 == T2. // (The exception is std::function where it is possible that the function // object and the allocator have the same type). static_assert((!is_same<_T1, _T2>::value), - "__compressed_pair cannot be instantated when T1 and T2 are the same type; " + "__compressed_pair cannot be instantiated when T1 and T2 are the same type; " "The current implementation is NOT ABI-compatible with the previous " "implementation for this configuration"); -public: + typedef _LIBCPP_NODEBUG_TYPE __compressed_pair_elem<_T1, 0> _Base1; + typedef _LIBCPP_NODEBUG_TYPE __compressed_pair_elem<_T2, 1> _Base2; + template <bool _Dummy = true, class = typename enable_if< __dependent_type<is_default_constructible<_T1>, _Dummy>::value && @@ -2185,7 +1311,7 @@ public: template <class _U1, class _U2> _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR __compressed_pair(_U1&& __t1, _U2&& __t2) - : _Base1(std::forward<_U1>(__t1)), _Base2(std::forward<_U2>(__t2)) {} + : _Base1(_VSTD::forward<_U1>(__t1)), _Base2(_VSTD::forward<_U2>(__t2)) {} #ifndef _LIBCPP_CXX03_LANG template <class... _Args1, class... _Args2> @@ -2218,12 +1344,21 @@ public: return static_cast<_Base2 const&>(*this).__get(); } + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR + static _Base1* __get_first_base(__compressed_pair* __pair) _NOEXCEPT { + return static_cast<_Base1*>(__pair); + } + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR + static _Base2* __get_second_base(__compressed_pair* __pair) _NOEXCEPT { + return static_cast<_Base2*>(__pair); + } + _LIBCPP_INLINE_VISIBILITY void swap(__compressed_pair& __x) _NOEXCEPT_(__is_nothrow_swappable<_T1>::value && __is_nothrow_swappable<_T2>::value) { - using std::swap; + using _VSTD::swap; swap(first(), __x.first()); swap(second(), __x.second()); } @@ -2316,12 +1451,18 @@ struct __unique_ptr_deleter_sfinae<_Deleter&> { typedef false_type __enable_rval_overload; }; +#if defined(_LIBCPP_ABI_ENABLE_UNIQUE_PTR_TRIVIAL_ABI) +# define _LIBCPP_UNIQUE_PTR_TRIVIAL_ABI __attribute__((trivial_abi)) +#else +# define _LIBCPP_UNIQUE_PTR_TRIVIAL_ABI +#endif + template <class _Tp, class _Dp = default_delete<_Tp> > -class _LIBCPP_TEMPLATE_VIS unique_ptr { +class _LIBCPP_UNIQUE_PTR_TRIVIAL_ABI _LIBCPP_TEMPLATE_VIS unique_ptr { public: typedef _Tp element_type; typedef _Dp deleter_type; - typedef _LIBCPP_NODEBUG_TYPE typename __pointer_type<_Tp, deleter_type>::type pointer; + typedef _LIBCPP_NODEBUG_TYPE typename __pointer<_Tp, deleter_type>::type pointer; static_assert(!is_rvalue_reference<deleter_type>::value, "the specified deleter type cannot be an rvalue reference"); @@ -2525,11 +1666,11 @@ public: template <class _Tp, class _Dp> -class _LIBCPP_TEMPLATE_VIS unique_ptr<_Tp[], _Dp> { +class _LIBCPP_UNIQUE_PTR_TRIVIAL_ABI _LIBCPP_TEMPLATE_VIS unique_ptr<_Tp[], _Dp> { public: typedef _Tp element_type; typedef _Dp deleter_type; - typedef typename __pointer_type<_Tp, deleter_type>::type pointer; + typedef typename __pointer<_Tp, deleter_type>::type pointer; private: __compressed_pair<pointer, deleter_type> __ptr_; @@ -2987,7 +2128,7 @@ public: : __size_(__s) {} template <class _Tp> - _LIBCPP_INLINE_VISIBILITY void __incr(_Tp*) _NOEXCEPT + _LIBCPP_INLINE_VISIBILITY void __incr() _NOEXCEPT {__incr(integral_constant<bool, is_trivially_destructible<_Tp>::value>());} template <class _Tp> @@ -3029,7 +2170,7 @@ uninitialized_copy(_InputIterator __f, _InputIterator __l, _ForwardIterator __r) { #endif for (; __f != __l; ++__f, (void) ++__r) - ::new (static_cast<void*>(_VSTD::addressof(*__r))) value_type(*__f); + ::new ((void*)_VSTD::addressof(*__r)) value_type(*__f); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -3053,7 +2194,7 @@ uninitialized_copy_n(_InputIterator __f, _Size __n, _ForwardIterator __r) { #endif for (; __n > 0; ++__f, (void) ++__r, (void) --__n) - ::new (static_cast<void*>(_VSTD::addressof(*__r))) value_type(*__f); + ::new ((void*)_VSTD::addressof(*__r)) value_type(*__f); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -3077,7 +2218,7 @@ uninitialized_fill(_ForwardIterator __f, _ForwardIterator __l, const _Tp& __x) { #endif for (; __f != __l; ++__f) - ::new (static_cast<void*>(_VSTD::addressof(*__f))) value_type(__x); + ::new ((void*)_VSTD::addressof(*__f)) value_type(__x); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -3100,7 +2241,7 @@ uninitialized_fill_n(_ForwardIterator __f, _Size __n, const _Tp& __x) { #endif for (; __n > 0; ++__f, (void) --__n) - ::new (static_cast<void*>(_VSTD::addressof(*__f))) value_type(__x); + ::new ((void*)_VSTD::addressof(*__f)) value_type(__x); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -3115,22 +2256,15 @@ uninitialized_fill_n(_ForwardIterator __f, _Size __n, const _Tp& __x) #if _LIBCPP_STD_VER > 14 -template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY -void destroy_at(_Tp* __loc) { - _LIBCPP_ASSERT(__loc, "null pointer given to destroy_at"); - __loc->~_Tp(); -} - template <class _ForwardIterator> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 void destroy(_ForwardIterator __first, _ForwardIterator __last) { for (; __first != __last; ++__first) _VSTD::destroy_at(_VSTD::addressof(*__first)); } template <class _ForwardIterator, class _Size> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _ForwardIterator destroy_n(_ForwardIterator __first, _Size __n) { for (; __n > 0; (void)++__first, --__n) _VSTD::destroy_at(_VSTD::addressof(*__first)); @@ -3146,7 +2280,7 @@ void uninitialized_default_construct(_ForwardIterator __first, _ForwardIterator try { #endif for (; __idx != __last; ++__idx) - ::new((void*)_VSTD::addressof(*__idx)) _Vt; + ::new ((void*)_VSTD::addressof(*__idx)) _Vt; #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) { _VSTD::destroy(__first, __idx); @@ -3164,7 +2298,7 @@ _ForwardIterator uninitialized_default_construct_n(_ForwardIterator __first, _Si try { #endif for (; __n > 0; (void)++__idx, --__n) - ::new((void*)_VSTD::addressof(*__idx)) _Vt; + ::new ((void*)_VSTD::addressof(*__idx)) _Vt; return __idx; #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) { @@ -3184,7 +2318,7 @@ void uninitialized_value_construct(_ForwardIterator __first, _ForwardIterator __ try { #endif for (; __idx != __last; ++__idx) - ::new((void*)_VSTD::addressof(*__idx)) _Vt(); + ::new ((void*)_VSTD::addressof(*__idx)) _Vt(); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) { _VSTD::destroy(__first, __idx); @@ -3202,7 +2336,7 @@ _ForwardIterator uninitialized_value_construct_n(_ForwardIterator __first, _Size try { #endif for (; __n > 0; (void)++__idx, --__n) - ::new((void*)_VSTD::addressof(*__idx)) _Vt(); + ::new ((void*)_VSTD::addressof(*__idx)) _Vt(); return __idx; #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) { @@ -3222,7 +2356,7 @@ _ForwardIt uninitialized_move(_InputIt __first, _InputIt __last, _ForwardIt __fi try { #endif for (; __first != __last; (void)++__idx, ++__first) - ::new((void*)_VSTD::addressof(*__idx)) _Vt(std::move(*__first)); + ::new ((void*)_VSTD::addressof(*__idx)) _Vt(_VSTD::move(*__first)); return __idx; #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) { @@ -3242,7 +2376,7 @@ uninitialized_move_n(_InputIt __first, _Size __n, _ForwardIt __first_res) { try { #endif for (; __n > 0; ++__idx, (void)++__first, --__n) - ::new((void*)_VSTD::addressof(*__idx)) _Vt(std::move(*__first)); + ::new ((void*)_VSTD::addressof(*__idx)) _Vt(_VSTD::move(*__first)); return {__first, __idx}; #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) { @@ -3389,13 +2523,7 @@ public: long use_count() const _NOEXCEPT {return __shared_count::use_count();} __shared_weak_count* lock() _NOEXCEPT; - // Define the function out only if we build static libc++ without RTTI. - // Otherwise we may break clients who need to compile their projects with - // -fno-rtti and yet link against a libc++.dylib compiled - // without -fno-rtti. -#if !defined(_LIBCPP_NO_RTTI) || !defined(_LIBCPP_BUILD_STATIC) virtual const void* __get_deleter(const type_info&) const _NOEXCEPT; -#endif private: virtual void __on_zero_shared_weak() _NOEXCEPT = 0; }; @@ -3452,69 +2580,86 @@ __shared_ptr_pointer<_Tp, _Dp, _Alloc>::__on_zero_shared_weak() _NOEXCEPT } template <class _Tp, class _Alloc> -class __shared_ptr_emplace - : public __shared_weak_count +struct __shared_ptr_emplace + : __shared_weak_count { - __compressed_pair<_Alloc, _Tp> __data_; -public: - - _LIBCPP_INLINE_VISIBILITY - __shared_ptr_emplace(_Alloc __a) - : __data_(_VSTD::move(__a), __value_init_tag()) {} - + template<class ..._Args> + _LIBCPP_HIDE_FROM_ABI + explicit __shared_ptr_emplace(_Alloc __a, _Args&& ...__args) + : __storage_(_VSTD::move(__a)) + { +#if _LIBCPP_STD_VER > 17 + using _TpAlloc = typename __allocator_traits_rebind<_Alloc, _Tp>::type; + _TpAlloc __tmp(*__get_alloc()); + allocator_traits<_TpAlloc>::construct(__tmp, __get_elem(), _VSTD::forward<_Args>(__args)...); +#else + ::new ((void*)__get_elem()) _Tp(_VSTD::forward<_Args>(__args)...); +#endif + } -#ifndef _LIBCPP_HAS_NO_VARIADICS - template <class ..._Args> - _LIBCPP_INLINE_VISIBILITY - __shared_ptr_emplace(_Alloc __a, _Args&& ...__args) - : __data_(piecewise_construct, _VSTD::forward_as_tuple(__a), - _VSTD::forward_as_tuple(_VSTD::forward<_Args>(__args)...)) {} -#else // _LIBCPP_HAS_NO_VARIADICS + _LIBCPP_HIDE_FROM_ABI + _Alloc* __get_alloc() _NOEXCEPT { return __storage_.__get_alloc(); } - template <class _A0> - _LIBCPP_INLINE_VISIBILITY - __shared_ptr_emplace(_Alloc __a, _A0& __a0) - : __data_(__a, _Tp(__a0)) {} + _LIBCPP_HIDE_FROM_ABI + _Tp* __get_elem() _NOEXCEPT { return __storage_.__get_elem(); } - template <class _A0, class _A1> - _LIBCPP_INLINE_VISIBILITY - __shared_ptr_emplace(_Alloc __a, _A0& __a0, _A1& __a1) - : __data_(__a, _Tp(__a0, __a1)) {} +private: + virtual void __on_zero_shared() _NOEXCEPT { +#if _LIBCPP_STD_VER > 17 + using _TpAlloc = typename __allocator_traits_rebind<_Alloc, _Tp>::type; + _TpAlloc __tmp(*__get_alloc()); + allocator_traits<_TpAlloc>::destroy(__tmp, __get_elem()); +#else + __get_elem()->~_Tp(); +#endif + } - template <class _A0, class _A1, class _A2> - _LIBCPP_INLINE_VISIBILITY - __shared_ptr_emplace(_Alloc __a, _A0& __a0, _A1& __a1, _A2& __a2) - : __data_(__a, _Tp(__a0, __a1, __a2)) {} + virtual void __on_zero_shared_weak() _NOEXCEPT { + using _ControlBlockAlloc = typename __allocator_traits_rebind<_Alloc, __shared_ptr_emplace>::type; + using _ControlBlockPointer = typename allocator_traits<_ControlBlockAlloc>::pointer; + _ControlBlockAlloc __tmp(*__get_alloc()); + __storage_.~_Storage(); + allocator_traits<_ControlBlockAlloc>::deallocate(__tmp, + pointer_traits<_ControlBlockPointer>::pointer_to(*this), 1); + } -#endif // _LIBCPP_HAS_NO_VARIADICS + // This class implements the control block for non-array shared pointers created + // through `std::allocate_shared` and `std::make_shared`. + // + // In previous versions of the library, we used a compressed pair to store + // both the _Alloc and the _Tp. This implies using EBO, which is incompatible + // with Allocator construction for _Tp. To allow implementing P0674 in C++20, + // we now use a properly aligned char buffer while making sure that we maintain + // the same layout that we had when we used a compressed pair. + using _CompressedPair = __compressed_pair<_Alloc, _Tp>; + struct _ALIGNAS_TYPE(_CompressedPair) _Storage { + char __blob_[sizeof(_CompressedPair)]; + + _LIBCPP_HIDE_FROM_ABI explicit _Storage(_Alloc&& __a) { + ::new ((void*)__get_alloc()) _Alloc(_VSTD::move(__a)); + } + _LIBCPP_HIDE_FROM_ABI ~_Storage() { + __get_alloc()->~_Alloc(); + } + _Alloc* __get_alloc() _NOEXCEPT { + _CompressedPair *__as_pair = reinterpret_cast<_CompressedPair*>(__blob_); + typename _CompressedPair::_Base1* __first = _CompressedPair::__get_first_base(__as_pair); + _Alloc *__alloc = reinterpret_cast<_Alloc*>(__first); + return __alloc; + } + _LIBCPP_NO_CFI _Tp* __get_elem() _NOEXCEPT { + _CompressedPair *__as_pair = reinterpret_cast<_CompressedPair*>(__blob_); + typename _CompressedPair::_Base2* __second = _CompressedPair::__get_second_base(__as_pair); + _Tp *__elem = reinterpret_cast<_Tp*>(__second); + return __elem; + } + }; -private: - virtual void __on_zero_shared() _NOEXCEPT; - virtual void __on_zero_shared_weak() _NOEXCEPT; -public: - _LIBCPP_INLINE_VISIBILITY - _Tp* get() _NOEXCEPT {return _VSTD::addressof(__data_.second());} + static_assert(_LIBCPP_ALIGNOF(_Storage) == _LIBCPP_ALIGNOF(_CompressedPair), ""); + static_assert(sizeof(_Storage) == sizeof(_CompressedPair), ""); + _Storage __storage_; }; -template <class _Tp, class _Alloc> -void -__shared_ptr_emplace<_Tp, _Alloc>::__on_zero_shared() _NOEXCEPT -{ - __data_.second().~_Tp(); -} - -template <class _Tp, class _Alloc> -void -__shared_ptr_emplace<_Tp, _Alloc>::__on_zero_shared_weak() _NOEXCEPT -{ - typedef typename __allocator_traits_rebind<_Alloc, __shared_ptr_emplace>::type _Al; - typedef allocator_traits<_Al> _ATraits; - typedef pointer_traits<typename _ATraits::pointer> _PTraits; - _Al __a(__data_.first()); - __data_.first().~_Alloc(); - __a.deallocate(_PTraits::pointer_to(*this), 1); -} - struct __shared_ptr_dummy_rebind_allocator_type; template <> class _LIBCPP_TEMPLATE_VIS allocator<__shared_ptr_dummy_rebind_allocator_type> @@ -3537,8 +2682,14 @@ struct __compatible_with : is_convertible<_Tp*, _Up*> {}; #endif // _LIBCPP_STD_VER > 14 +#if defined(_LIBCPP_ABI_ENABLE_SHARED_PTR_TRIVIAL_ABI) +# define _LIBCPP_SHARED_PTR_TRIVIAL_ABI __attribute__((trivial_abi)) +#else +# define _LIBCPP_SHARED_PTR_TRIVIAL_ABI +#endif + template<class _Tp> -class _LIBCPP_TEMPLATE_VIS shared_ptr +class _LIBCPP_SHARED_PTR_TRIVIAL_ABI _LIBCPP_TEMPLATE_VIS shared_ptr { public: #if _LIBCPP_STD_VER > 14 @@ -3577,25 +2728,17 @@ public: shared_ptr(const shared_ptr<_Yp>& __r, typename enable_if<__compatible_with<_Yp, element_type>::value, __nat>::type = __nat()) _NOEXCEPT; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_ptr(shared_ptr&& __r) _NOEXCEPT; template<class _Yp> _LIBCPP_INLINE_VISIBILITY shared_ptr(shared_ptr<_Yp>&& __r, typename enable_if<__compatible_with<_Yp, element_type>::value, __nat>::type = __nat()) _NOEXCEPT; -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES template<class _Yp> explicit shared_ptr(const weak_ptr<_Yp>& __r, typename enable_if<is_convertible<_Yp*, element_type*>::value, __nat>::type= __nat()); #if _LIBCPP_STD_VER <= 14 || defined(_LIBCPP_ENABLE_CXX17_REMOVED_AUTO_PTR) -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES template<class _Yp> shared_ptr(auto_ptr<_Yp>&& __r, typename enable_if<is_convertible<_Yp*, element_type*>::value, __nat>::type = __nat()); -#else - template<class _Yp> - shared_ptr(auto_ptr<_Yp> __r, - typename enable_if<is_convertible<_Yp*, element_type*>::value, __nat>::type = __nat()); -#endif #endif template <class _Yp, class _Dp> shared_ptr(unique_ptr<_Yp, _Dp>&&, @@ -3628,7 +2771,6 @@ public: >::type _LIBCPP_INLINE_VISIBILITY operator=(const shared_ptr<_Yp>& __r) _NOEXCEPT; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY shared_ptr& operator=(shared_ptr&& __r) _NOEXCEPT; template<class _Yp> @@ -3650,19 +2792,6 @@ public: >::type& operator=(auto_ptr<_Yp>&& __r); #endif -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES -#if _LIBCPP_STD_VER <= 14 || defined(_LIBCPP_ENABLE_CXX17_REMOVED_AUTO_PTR) - template<class _Yp> - _LIBCPP_INLINE_VISIBILITY - typename enable_if - < - !is_array<_Yp>::value && - is_convertible<_Yp*, element_type*>::value, - shared_ptr& - >::type - operator=(auto_ptr<_Yp> __r); -#endif -#endif template <class _Yp, class _Dp> typename enable_if < @@ -3670,13 +2799,8 @@ public: is_convertible<typename unique_ptr<_Yp, _Dp>::pointer, element_type*>::value, shared_ptr& >::type -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY operator=(unique_ptr<_Yp, _Dp>&& __r); -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES - _LIBCPP_INLINE_VISIBILITY - operator=(unique_ptr<_Yp, _Dp> __r); -#endif _LIBCPP_INLINE_VISIBILITY void swap(shared_ptr& __r) _NOEXCEPT; @@ -3724,7 +2848,7 @@ public: _LIBCPP_INLINE_VISIBILITY bool unique() const _NOEXCEPT {return use_count() == 1;} _LIBCPP_INLINE_VISIBILITY - _LIBCPP_EXPLICIT operator bool() const _NOEXCEPT {return get() != 0;} + _LIBCPP_EXPLICIT operator bool() const _NOEXCEPT {return get() != nullptr;} template <class _Up> _LIBCPP_INLINE_VISIBILITY bool owner_before(shared_ptr<_Up> const& __p) const _NOEXCEPT @@ -3828,8 +2952,8 @@ template<class _Tp> inline _LIBCPP_CONSTEXPR shared_ptr<_Tp>::shared_ptr() _NOEXCEPT - : __ptr_(0), - __cntrl_(0) + : __ptr_(nullptr), + __cntrl_(nullptr) { } @@ -3837,8 +2961,8 @@ template<class _Tp> inline _LIBCPP_CONSTEXPR shared_ptr<_Tp>::shared_ptr(nullptr_t) _NOEXCEPT - : __ptr_(0), - __cntrl_(0) + : __ptr_(nullptr), + __cntrl_(nullptr) { } @@ -3883,7 +3007,7 @@ shared_ptr<_Tp>::shared_ptr(_Yp* __p, _Dp __d, template<class _Tp> template<class _Dp> shared_ptr<_Tp>::shared_ptr(nullptr_t __p, _Dp __d) - : __ptr_(0) + : __ptr_(nullptr) { #ifndef _LIBCPP_NO_EXCEPTIONS try @@ -3917,8 +3041,7 @@ shared_ptr<_Tp>::shared_ptr(_Yp* __p, _Dp __d, _Alloc __a, typedef __allocator_destructor<_A2> _D2; _A2 __a2(__a); unique_ptr<_CntrlBlk, _D2> __hold2(__a2.allocate(1), _D2(__a2, 1)); - ::new(static_cast<void*>(_VSTD::addressof(*__hold2.get()))) - _CntrlBlk(__p, __d, __a); + ::new ((void*)_VSTD::addressof(*__hold2.get())) _CntrlBlk(__p, __d, __a); __cntrl_ = _VSTD::addressof(*__hold2.release()); __enable_weak_this(__p, __p); #ifndef _LIBCPP_NO_EXCEPTIONS @@ -3934,7 +3057,7 @@ shared_ptr<_Tp>::shared_ptr(_Yp* __p, _Dp __d, _Alloc __a, template<class _Tp> template<class _Dp, class _Alloc> shared_ptr<_Tp>::shared_ptr(nullptr_t __p, _Dp __d, _Alloc __a) - : __ptr_(0) + : __ptr_(nullptr) { #ifndef _LIBCPP_NO_EXCEPTIONS try @@ -3945,8 +3068,7 @@ shared_ptr<_Tp>::shared_ptr(nullptr_t __p, _Dp __d, _Alloc __a) typedef __allocator_destructor<_A2> _D2; _A2 __a2(__a); unique_ptr<_CntrlBlk, _D2> __hold2(__a2.allocate(1), _D2(__a2, 1)); - ::new(static_cast<void*>(_VSTD::addressof(*__hold2.get()))) - _CntrlBlk(__p, __d, __a); + ::new ((void*)_VSTD::addressof(*__hold2.get())) _CntrlBlk(__p, __d, __a); __cntrl_ = _VSTD::addressof(*__hold2.release()); #ifndef _LIBCPP_NO_EXCEPTIONS } @@ -3992,16 +3114,14 @@ shared_ptr<_Tp>::shared_ptr(const shared_ptr<_Yp>& __r, __cntrl_->__add_shared(); } -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Tp> inline shared_ptr<_Tp>::shared_ptr(shared_ptr&& __r) _NOEXCEPT : __ptr_(__r.__ptr_), __cntrl_(__r.__cntrl_) { - __r.__ptr_ = 0; - __r.__cntrl_ = 0; + __r.__ptr_ = nullptr; + __r.__cntrl_ = nullptr; } template<class _Tp> @@ -4013,20 +3133,14 @@ shared_ptr<_Tp>::shared_ptr(shared_ptr<_Yp>&& __r, : __ptr_(__r.__ptr_), __cntrl_(__r.__cntrl_) { - __r.__ptr_ = 0; - __r.__cntrl_ = 0; + __r.__ptr_ = nullptr; + __r.__cntrl_ = nullptr; } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - #if _LIBCPP_STD_VER <= 14 || defined(_LIBCPP_ENABLE_CXX17_REMOVED_AUTO_PTR) template<class _Tp> template<class _Yp> -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES shared_ptr<_Tp>::shared_ptr(auto_ptr<_Yp>&& __r, -#else -shared_ptr<_Tp>::shared_ptr(auto_ptr<_Yp> __r, -#endif typename enable_if<is_convertible<_Yp*, element_type*>::value, __nat>::type) : __ptr_(__r.get()) { @@ -4121,8 +3235,6 @@ shared_ptr<_Tp>::operator=(const shared_ptr<_Yp>& __r) _NOEXCEPT return *this; } -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Tp> inline shared_ptr<_Tp>& @@ -4179,43 +3291,6 @@ shared_ptr<_Tp>::operator=(unique_ptr<_Yp, _Dp>&& __r) return *this; } -#else // _LIBCPP_HAS_NO_RVALUE_REFERENCES - -#if _LIBCPP_STD_VER <= 14 || defined(_LIBCPP_ENABLE_CXX17_REMOVED_AUTO_PTR) -template<class _Tp> -template<class _Yp> -inline _LIBCPP_INLINE_VISIBILITY -typename enable_if -< - !is_array<_Yp>::value && - is_convertible<_Yp*, typename shared_ptr<_Tp>::element_type*>::value, - shared_ptr<_Tp>& ->::type -shared_ptr<_Tp>::operator=(auto_ptr<_Yp> __r) -{ - shared_ptr(__r).swap(*this); - return *this; -} -#endif - -template<class _Tp> -template <class _Yp, class _Dp> -inline _LIBCPP_INLINE_VISIBILITY -typename enable_if -< - !is_array<_Yp>::value && - is_convertible<typename unique_ptr<_Yp, _Dp>::pointer, - typename shared_ptr<_Tp>::element_type*>::value, - shared_ptr<_Tp>& ->::type -shared_ptr<_Tp>::operator=(unique_ptr<_Yp, _Dp> __r) -{ - shared_ptr(_VSTD::move(__r)).swap(*this); - return *this; -} - -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Tp> inline void @@ -4272,50 +3347,26 @@ shared_ptr<_Tp>::reset(_Yp* __p, _Dp __d, _Alloc __a) shared_ptr(__p, __d, __a).swap(*this); } -template<class _Tp, class ..._Args> -inline _LIBCPP_INLINE_VISIBILITY -typename enable_if -< - !is_array<_Tp>::value, - shared_ptr<_Tp> ->::type -make_shared(_Args&& ...__args) +// +// std::allocate_shared and std::make_shared +// +template<class _Tp, class _Alloc, class ..._Args, class = _EnableIf<!is_array<_Tp>::value> > +_LIBCPP_HIDE_FROM_ABI +shared_ptr<_Tp> allocate_shared(const _Alloc& __a, _Args&& ...__args) { - static_assert(is_constructible<_Tp, _Args...>::value, "Can't construct object in make_shared"); - typedef __shared_ptr_emplace<_Tp, allocator<_Tp> > _CntrlBlk; - typedef allocator<_CntrlBlk> _A2; - typedef __allocator_destructor<_A2> _D2; - - _A2 __a2; - unique_ptr<_CntrlBlk, _D2> __hold2(__a2.allocate(1), _D2(__a2, 1)); - ::new(__hold2.get()) _CntrlBlk(__a2, _VSTD::forward<_Args>(__args)...); - - _Tp *__ptr = __hold2.get()->get(); - return shared_ptr<_Tp>::__create_with_control_block(__ptr, __hold2.release()); + using _ControlBlock = __shared_ptr_emplace<_Tp, _Alloc>; + using _ControlBlockAllocator = typename __allocator_traits_rebind<_Alloc, _ControlBlock>::type; + __allocation_guard<_ControlBlockAllocator> __guard(__a, 1); + ::new ((void*)_VSTD::addressof(*__guard.__get())) _ControlBlock(__a, _VSTD::forward<_Args>(__args)...); + auto __control_block = __guard.__release_ptr(); + return shared_ptr<_Tp>::__create_with_control_block((*__control_block).__get_elem(), _VSTD::addressof(*__control_block)); } -template<class _Tp, class _Alloc, class ..._Args> -inline _LIBCPP_INLINE_VISIBILITY -typename enable_if -< - !is_array<_Tp>::value, - shared_ptr<_Tp> ->::type -allocate_shared(const _Alloc& __a, _Args&& ...__args) +template<class _Tp, class ..._Args, class = _EnableIf<!is_array<_Tp>::value> > +_LIBCPP_HIDE_FROM_ABI +shared_ptr<_Tp> make_shared(_Args&& ...__args) { - static_assert( is_constructible<_Tp, _Args...>::value, "Can't construct object in allocate_shared"); - - typedef __shared_ptr_emplace<_Tp, _Alloc> _CntrlBlk; - typedef typename __allocator_traits_rebind<_Alloc, _CntrlBlk>::type _A2; - typedef __allocator_destructor<_A2> _D2; - - _A2 __a2(__a); - unique_ptr<_CntrlBlk, _D2> __hold2(__a2.allocate(1), _D2(__a2, 1)); - ::new(static_cast<void*>(_VSTD::addressof(*__hold2.get()))) - _CntrlBlk(__a, _VSTD::forward<_Args>(__args)...); - - typename shared_ptr<_Tp>::element_type *__p = __hold2.get()->get(); - return shared_ptr<_Tp>::__create_with_control_block(__p, _VSTD::addressof(*__hold2.release())); + return _VSTD::allocate_shared<_Tp>(allocator<_Tp>(), _VSTD::forward<_Args>(__args)...); } template<class _Tp, class _Up> @@ -4526,7 +3577,7 @@ get_deleter(const shared_ptr<_Tp>& __p) _NOEXCEPT #endif // _LIBCPP_NO_RTTI template<class _Tp> -class _LIBCPP_TEMPLATE_VIS weak_ptr +class _LIBCPP_SHARED_PTR_TRIVIAL_ABI _LIBCPP_TEMPLATE_VIS weak_ptr { public: typedef _Tp element_type; @@ -4546,13 +3597,11 @@ public: typename enable_if<is_convertible<_Yp*, _Tp*>::value, __nat*>::type = 0) _NOEXCEPT; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES _LIBCPP_INLINE_VISIBILITY weak_ptr(weak_ptr&& __r) _NOEXCEPT; template<class _Yp> _LIBCPP_INLINE_VISIBILITY weak_ptr(weak_ptr<_Yp>&& __r, typename enable_if<is_convertible<_Yp*, _Tp*>::value, __nat*>::type = 0) _NOEXCEPT; -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES ~weak_ptr(); _LIBCPP_INLINE_VISIBILITY @@ -4566,8 +3615,6 @@ public: _LIBCPP_INLINE_VISIBILITY operator=(weak_ptr<_Yp> const& __r) _NOEXCEPT; -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - _LIBCPP_INLINE_VISIBILITY weak_ptr& operator=(weak_ptr&& __r) _NOEXCEPT; template<class _Yp> @@ -4579,8 +3626,6 @@ public: _LIBCPP_INLINE_VISIBILITY operator=(weak_ptr<_Yp>&& __r) _NOEXCEPT; -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Yp> typename enable_if < @@ -4600,7 +3645,7 @@ public: {return __cntrl_ ? __cntrl_->use_count() : 0;} _LIBCPP_INLINE_VISIBILITY bool expired() const _NOEXCEPT - {return __cntrl_ == 0 || __cntrl_->use_count() == 0;} + {return __cntrl_ == nullptr || __cntrl_->use_count() == 0;} shared_ptr<_Tp> lock() const _NOEXCEPT; template<class _Up> _LIBCPP_INLINE_VISIBILITY @@ -4624,8 +3669,8 @@ template<class _Tp> inline _LIBCPP_CONSTEXPR weak_ptr<_Tp>::weak_ptr() _NOEXCEPT - : __ptr_(0), - __cntrl_(0) + : __ptr_(nullptr), + __cntrl_(nullptr) { } @@ -4665,16 +3710,14 @@ weak_ptr<_Tp>::weak_ptr(weak_ptr<_Yp> const& __r, __cntrl_->__add_weak(); } -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Tp> inline weak_ptr<_Tp>::weak_ptr(weak_ptr&& __r) _NOEXCEPT : __ptr_(__r.__ptr_), __cntrl_(__r.__cntrl_) { - __r.__ptr_ = 0; - __r.__cntrl_ = 0; + __r.__ptr_ = nullptr; + __r.__cntrl_ = nullptr; } template<class _Tp> @@ -4686,12 +3729,10 @@ weak_ptr<_Tp>::weak_ptr(weak_ptr<_Yp>&& __r, : __ptr_(__r.__ptr_), __cntrl_(__r.__cntrl_) { - __r.__ptr_ = 0; - __r.__cntrl_ = 0; + __r.__ptr_ = nullptr; + __r.__cntrl_ = nullptr; } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Tp> weak_ptr<_Tp>::~weak_ptr() { @@ -4722,8 +3763,6 @@ weak_ptr<_Tp>::operator=(weak_ptr<_Yp> const& __r) _NOEXCEPT return *this; } -#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Tp> inline weak_ptr<_Tp>& @@ -4747,8 +3786,6 @@ weak_ptr<_Tp>::operator=(weak_ptr<_Yp>&& __r) _NOEXCEPT return *this; } -#endif // _LIBCPP_HAS_NO_RVALUE_REFERENCES - template<class _Tp> template<class _Yp> inline @@ -4795,7 +3832,7 @@ shared_ptr<_Tp>::shared_ptr(const weak_ptr<_Yp>& __r, : __ptr_(__r.__ptr_), __cntrl_(__r.__cntrl_ ? __r.__cntrl_->lock() : __r.__cntrl_) { - if (__cntrl_ == 0) + if (__cntrl_ == nullptr) __throw_bad_weak_ptr(); } @@ -5130,35 +4167,35 @@ _LIBCPP_FUNC_VIS void* align(size_t __align, size_t __sz, void*& __ptr, size_t& // --- Helper for container swap -- template <typename _Alloc> -inline _LIBCPP_INLINE_VISIBILITY -void __swap_allocator(_Alloc & __a1, _Alloc & __a2) +_LIBCPP_INLINE_VISIBILITY +void __swap_allocator(_Alloc & __a1, _Alloc & __a2, true_type) #if _LIBCPP_STD_VER >= 14 _NOEXCEPT #else _NOEXCEPT_(__is_nothrow_swappable<_Alloc>::value) #endif { - __swap_allocator(__a1, __a2, - integral_constant<bool, _VSTD::allocator_traits<_Alloc>::propagate_on_container_swap::value>()); + using _VSTD::swap; + swap(__a1, __a2); } template <typename _Alloc> -_LIBCPP_INLINE_VISIBILITY -void __swap_allocator(_Alloc & __a1, _Alloc & __a2, true_type) +inline _LIBCPP_INLINE_VISIBILITY +void __swap_allocator(_Alloc &, _Alloc &, false_type) _NOEXCEPT {} + +template <typename _Alloc> +inline _LIBCPP_INLINE_VISIBILITY +void __swap_allocator(_Alloc & __a1, _Alloc & __a2) #if _LIBCPP_STD_VER >= 14 _NOEXCEPT #else _NOEXCEPT_(__is_nothrow_swappable<_Alloc>::value) #endif { - using _VSTD::swap; - swap(__a1, __a2); + _VSTD::__swap_allocator(__a1, __a2, + integral_constant<bool, _VSTD::allocator_traits<_Alloc>::propagate_on_container_swap::value>()); } -template <typename _Alloc> -inline _LIBCPP_INLINE_VISIBILITY -void __swap_allocator(_Alloc &, _Alloc &, false_type) _NOEXCEPT {} - template <typename _Alloc, typename _Traits=allocator_traits<_Alloc> > struct __noexcept_move_assign_container : public integral_constant<bool, _Traits::propagate_on_container_move_assignment::value @@ -5170,7 +4207,6 @@ struct __noexcept_move_assign_container : public integral_constant<bool, > {}; -#ifndef _LIBCPP_HAS_NO_VARIADICS template <class _Tp, class _Alloc> struct __temp_value { typedef allocator_traits<_Alloc> _Traits; @@ -5190,7 +4226,6 @@ struct __temp_value { ~__temp_value() { _Traits::destroy(__a, __addr()); } }; -#endif template<typename _Alloc, typename = void, typename = void> struct __is_allocator : false_type {}; @@ -5214,7 +4249,7 @@ struct __builtin_new_allocator { : __size_(__size), __align_(__align) {} void operator()(void* p) const _NOEXCEPT { - std::__libcpp_deallocate(p, __size_, __align_); + _VSTD::__libcpp_deallocate(p, __size_, __align_); } private: @@ -5225,13 +4260,13 @@ struct __builtin_new_allocator { typedef unique_ptr<void, __builtin_new_deleter> __holder_t; static __holder_t __allocate_bytes(size_t __s, size_t __align) { - return __holder_t(std::__libcpp_allocate(__s, __align), + return __holder_t(_VSTD::__libcpp_allocate(__s, __align), __builtin_new_deleter(__s, __align)); } static void __deallocate_bytes(void* __p, size_t __s, size_t __align) _NOEXCEPT { - std::__libcpp_deallocate(__p, __s, __align); + _VSTD::__libcpp_deallocate(__p, __s, __align); } template <class _Tp> diff --git a/lib/libcxx/include/module.modulemap b/lib/libcxx/include/module.modulemap index b8d2a6669a..750cd38588 100644 --- a/lib/libcxx/include/module.modulemap +++ b/lib/libcxx/include/module.modulemap @@ -522,6 +522,7 @@ module std [system] { } // FIXME: These should be private. + module __bits { header "__bits" export * } module __bit_reference { header "__bit_reference" export * } module __debug { header "__debug" export * } module __errc { header "__errc" export * } diff --git a/lib/libcxx/include/mutex b/lib/libcxx/include/mutex index 62780bd073..f098ccba48 100644 --- a/lib/libcxx/include/mutex +++ b/lib/libcxx/include/mutex @@ -626,7 +626,7 @@ private: _LIBCPP_INLINE_VISIBILITY void __execute(__tuple_indices<_Indices...>) { - __invoke(_VSTD::get<0>(_VSTD::move(__f_)), _VSTD::get<_Indices>(_VSTD::move(__f_))...); + _VSTD::__invoke(_VSTD::get<0>(_VSTD::move(__f_)), _VSTD::get<_Indices>(_VSTD::move(__f_))...); } }; diff --git a/lib/libcxx/include/new b/lib/libcxx/include/new index 40d351e9b7..0562cef458 100644 --- a/lib/libcxx/include/new +++ b/lib/libcxx/include/new @@ -49,11 +49,11 @@ new_handler get_new_handler() noexcept; template <class T> constexpr T* launder(T* p) noexcept; // C++17 } // std -void* operator new(std::size_t size); // replaceable, nodiscard in C++2a -void* operator new(std::size_t size, std::align_val_t alignment); // replaceable, C++17, nodiscard in C++2a -void* operator new(std::size_t size, const std::nothrow_t&) noexcept; // replaceable, nodiscard in C++2a +void* operator new(std::size_t size); // replaceable, nodiscard in C++20 +void* operator new(std::size_t size, std::align_val_t alignment); // replaceable, C++17, nodiscard in C++20 +void* operator new(std::size_t size, const std::nothrow_t&) noexcept; // replaceable, nodiscard in C++20 void* operator new(std::size_t size, std::align_val_t alignment, - const std::nothrow_t&) noexcept; // replaceable, C++17, nodiscard in C++2a + const std::nothrow_t&) noexcept; // replaceable, C++17, nodiscard in C++20 void operator delete(void* ptr) noexcept; // replaceable void operator delete(void* ptr, std::size_t size) noexcept; // replaceable, C++14 void operator delete(void* ptr, std::align_val_t alignment) noexcept; // replaceable, C++17 @@ -63,12 +63,12 @@ void operator delete(void* ptr, const std::nothrow_t&) noexcept; // repla void operator delete(void* ptr, std:align_val_t alignment, const std::nothrow_t&) noexcept; // replaceable, C++17 -void* operator new[](std::size_t size); // replaceable, nodiscard in C++2a +void* operator new[](std::size_t size); // replaceable, nodiscard in C++20 void* operator new[](std::size_t size, - std::align_val_t alignment) noexcept; // replaceable, C++17, nodiscard in C++2a -void* operator new[](std::size_t size, const std::nothrow_t&) noexcept; // replaceable, nodiscard in C++2a + std::align_val_t alignment) noexcept; // replaceable, C++17, nodiscard in C++20 +void* operator new[](std::size_t size, const std::nothrow_t&) noexcept; // replaceable, nodiscard in C++20 void* operator new[](std::size_t size, std::align_val_t alignment, - const std::nothrow_t&) noexcept; // replaceable, C++17, nodiscard in C++2a + const std::nothrow_t&) noexcept; // replaceable, C++17, nodiscard in C++20 void operator delete[](void* ptr) noexcept; // replaceable void operator delete[](void* ptr, std::size_t size) noexcept; // replaceable, C++14 void operator delete[](void* ptr, @@ -79,21 +79,20 @@ void operator delete[](void* ptr, const std::nothrow_t&) noexcept; // repla void operator delete[](void* ptr, std::align_val_t alignment, const std::nothrow_t&) noexcept; // replaceable, C++17 -void* operator new (std::size_t size, void* ptr) noexcept; // nodiscard in C++2a -void* operator new[](std::size_t size, void* ptr) noexcept; // nodiscard in C++2a +void* operator new (std::size_t size, void* ptr) noexcept; // nodiscard in C++20 +void* operator new[](std::size_t size, void* ptr) noexcept; // nodiscard in C++20 void operator delete (void* ptr, void*) noexcept; void operator delete[](void* ptr, void*) noexcept; */ #include <__config> +#include <__availability> +#include <cstddef> +#include <cstdlib> #include <exception> #include <type_traits> -#include <cstddef> #include <version> -#ifdef _LIBCPP_NO_EXCEPTIONS -#include <cstdlib> -#endif #if defined(_LIBCPP_ABI_VCRUNTIME) #include <new.h> @@ -117,11 +116,6 @@ void operator delete[](void* ptr, void*) noexcept; # define _LIBCPP_HAS_NO_SIZED_DEALLOCATION #endif -#if !__has_builtin(__builtin_operator_new) || \ - __has_builtin(__builtin_operator_new) < 201802L -#define _LIBCPP_HAS_NO_BUILTIN_OVERLOADED_OPERATOR_NEW_DELETE -#endif - namespace std // purposefully not using versioning namespace { @@ -234,31 +228,54 @@ _LIBCPP_CONSTEXPR inline _LIBCPP_INLINE_VISIBILITY bool __is_overaligned_for_new #endif } -inline _LIBCPP_INLINE_VISIBILITY void *__libcpp_allocate(size_t __size, size_t __align) { +template <class ..._Args> +_LIBCPP_INLINE_VISIBILITY +void* __libcpp_operator_new(_Args ...__args) { +#if __has_builtin(__builtin_operator_new) && __has_builtin(__builtin_operator_delete) + return __builtin_operator_new(__args...); +#else + return ::operator new(__args...); +#endif +} + +template <class ..._Args> +_LIBCPP_INLINE_VISIBILITY +void __libcpp_operator_delete(_Args ...__args) { +#if __has_builtin(__builtin_operator_new) && __has_builtin(__builtin_operator_delete) + __builtin_operator_delete(__args...); +#else + ::operator delete(__args...); +#endif +} + +inline _LIBCPP_INLINE_VISIBILITY +void *__libcpp_allocate(size_t __size, size_t __align) { #ifndef _LIBCPP_HAS_NO_ALIGNED_ALLOCATION if (__is_overaligned_for_new(__align)) { const align_val_t __align_val = static_cast<align_val_t>(__align); -# ifdef _LIBCPP_HAS_NO_BUILTIN_OVERLOADED_OPERATOR_NEW_DELETE - return ::operator new(__size, __align_val); -# else - return __builtin_operator_new(__size, __align_val); -# endif + return __libcpp_operator_new(__size, __align_val); } -#else - ((void)__align); #endif -#ifdef _LIBCPP_HAS_NO_BUILTIN_OPERATOR_NEW_DELETE - return ::operator new(__size); + + (void)__align; + return __libcpp_operator_new(__size); +} + +template <class ..._Args> +_LIBCPP_INLINE_VISIBILITY +void __do_deallocate_handle_size(void *__ptr, size_t __size, _Args ...__args) { +#ifdef _LIBCPP_HAS_NO_SIZED_DEALLOCATION + (void)__size; + return __libcpp_operator_delete(__ptr, __args...); #else - return __builtin_operator_new(__size); + return __libcpp_operator_delete(__ptr, __size, __args...); #endif } -struct _DeallocateCaller { - static inline _LIBCPP_INLINE_VISIBILITY - void __do_deallocate_handle_size_align(void *__ptr, size_t __size, size_t __align) { +inline _LIBCPP_INLINE_VISIBILITY +void __libcpp_deallocate(void* __ptr, size_t __size, size_t __align) { #if defined(_LIBCPP_HAS_NO_ALIGNED_ALLOCATION) - ((void)__align); + (void)__align; return __do_deallocate_handle_size(__ptr, __size); #else if (__is_overaligned_for_new(__align)) { @@ -268,81 +285,51 @@ struct _DeallocateCaller { return __do_deallocate_handle_size(__ptr, __size); } #endif - } +} - static inline _LIBCPP_INLINE_VISIBILITY - void __do_deallocate_handle_align(void *__ptr, size_t __align) { +inline _LIBCPP_INLINE_VISIBILITY void __libcpp_deallocate_unsized(void* __ptr, size_t __align) { #if defined(_LIBCPP_HAS_NO_ALIGNED_ALLOCATION) - ((void)__align); - return __do_call(__ptr); + (void)__align; + return __libcpp_operator_delete(__ptr); #else if (__is_overaligned_for_new(__align)) { const align_val_t __align_val = static_cast<align_val_t>(__align); - return __do_call(__ptr, __align_val); + return __libcpp_operator_delete(__ptr, __align_val); } else { - return __do_call(__ptr); + return __libcpp_operator_delete(__ptr); } #endif - } - - private: - static inline void __do_deallocate_handle_size(void *__ptr, size_t __size) { -#ifdef _LIBCPP_HAS_NO_SIZED_DEALLOCATION - ((void)__size); - return __do_call(__ptr); -#else - return __do_call(__ptr, __size); -#endif - } - -#ifndef _LIBCPP_HAS_NO_ALIGNED_ALLOCATION - static inline void __do_deallocate_handle_size(void *__ptr, size_t __size, align_val_t __align) { -#ifdef _LIBCPP_HAS_NO_SIZED_DEALLOCATION - ((void)__size); - return __do_call(__ptr, __align); -#else - return __do_call(__ptr, __size, __align); -#endif - } -#endif - -private: - template <class _A1, class _A2> - static inline void __do_call(void *__ptr, _A1 __a1, _A2 __a2) { -#if defined(_LIBCPP_HAS_NO_BUILTIN_OPERATOR_NEW_DELETE) || \ - defined(_LIBCPP_HAS_NO_BUILTIN_OVERLOADED_OPERATOR_NEW_DELETE) - return ::operator delete(__ptr, __a1, __a2); -#else - return __builtin_operator_delete(__ptr, __a1, __a2); -#endif - } +} - template <class _A1> - static inline void __do_call(void *__ptr, _A1 __a1) { -#if defined(_LIBCPP_HAS_NO_BUILTIN_OPERATOR_NEW_DELETE) || \ - defined(_LIBCPP_HAS_NO_BUILTIN_OVERLOADED_OPERATOR_NEW_DELETE) - return ::operator delete(__ptr, __a1); +#if !defined(_LIBCPP_HAS_NO_LIBRARY_ALIGNED_ALLOCATION) +// Low-level helpers to call the aligned allocation and deallocation functions +// on the target platform. This is used to implement libc++'s own memory +// allocation routines -- if you need to allocate memory inside the library, +// chances are that you want to use `__libcpp_allocate` instead. +// +// Returns the allocated memory, or `nullptr` on failure. +inline _LIBCPP_INLINE_VISIBILITY +void* __libcpp_aligned_alloc(std::size_t __alignment, std::size_t __size) { +#if defined(_LIBCPP_MSVCRT_LIKE) + return ::_aligned_malloc(__size, __alignment); #else - return __builtin_operator_delete(__ptr, __a1); + void* __result = nullptr; + ::posix_memalign(&__result, __alignment, __size); + // If posix_memalign fails, __result is unmodified so we still return `nullptr`. + return __result; #endif - } +} - static inline void __do_call(void *__ptr) { -#ifdef _LIBCPP_HAS_NO_BUILTIN_OPERATOR_NEW_DELETE - return ::operator delete(__ptr); +inline _LIBCPP_INLINE_VISIBILITY +void __libcpp_aligned_free(void* __ptr) { +#if defined(_LIBCPP_MSVCRT_LIKE) + ::_aligned_free(__ptr); #else - return __builtin_operator_delete(__ptr); + ::free(__ptr); #endif - } -}; - -inline _LIBCPP_INLINE_VISIBILITY void __libcpp_deallocate(void* __ptr, size_t __size, size_t __align) { - _DeallocateCaller::__do_deallocate_handle_size_align(__ptr, __size, __align); } +#endif // !_LIBCPP_HAS_NO_ALIGNED_ALLOCATION -inline _LIBCPP_INLINE_VISIBILITY void __libcpp_deallocate_unsized(void* __ptr, size_t __align) { - _DeallocateCaller::__do_deallocate_handle_align(__ptr, __align); -} template <class _Tp> _LIBCPP_NODISCARD_AFTER_CXX17 inline diff --git a/lib/libcxx/include/numbers b/lib/libcxx/include/numbers index e7d981be4a..38dad99554 100644 --- a/lib/libcxx/include/numbers +++ b/lib/libcxx/include/numbers @@ -100,7 +100,7 @@ template <class T> inline constexpr T egamma_v = __illformed<T>{}; template <class T> inline constexpr T phi_v = __illformed<T>{}; template <class T> -concept __floating_point = std::is_floating_point_v<T>; +concept __floating_point = is_floating_point_v<T>; template <__floating_point T> inline constexpr T e_v<T> = 2.718281828459045235360287471352662; template <__floating_point T> inline constexpr T log2e_v<T> = 1.442695040888963407359924681001892; diff --git a/lib/libcxx/include/numeric b/lib/libcxx/include/numeric index 5ceadc1775..4f202bb84f 100644 --- a/lib/libcxx/include/numeric +++ b/lib/libcxx/include/numeric @@ -17,115 +17,116 @@ namespace std { template <class InputIterator, class T> - T + constexpr T // constexpr since C++20 accumulate(InputIterator first, InputIterator last, T init); template <class InputIterator, class T, class BinaryOperation> - T + constexpr T // constexpr since C++20 accumulate(InputIterator first, InputIterator last, T init, BinaryOperation binary_op); template<class InputIterator> - typename iterator_traits<InputIterator>::value_type + constexpr typename iterator_traits<InputIterator>::value_type // constexpr since C++20 reduce(InputIterator first, InputIterator last); // C++17 template<class InputIterator, class T> - T + constexpr T // constexpr since C++20 reduce(InputIterator first, InputIterator last, T init); // C++17 template<class InputIterator, class T, class BinaryOperation> - T + constexpr T // constexpr since C++20 reduce(InputIterator first, InputIterator last, T init, BinaryOperation binary_op); // C++17 template <class InputIterator1, class InputIterator2, class T> - T + constexpr T // constexpr since C++20 inner_product(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, T init); template <class InputIterator1, class InputIterator2, class T, class BinaryOperation1, class BinaryOperation2> - T + constexpr T // constexpr since C++20 inner_product(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, T init, BinaryOperation1 binary_op1, BinaryOperation2 binary_op2); template<class InputIterator1, class InputIterator2, class T> - T + constexpr T // constexpr since C++20 transform_reduce(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, T init); // C++17 template<class InputIterator1, class InputIterator2, class T, class BinaryOperation1, class BinaryOperation2> - T + constexpr T // constexpr since C++20 transform_reduce(InputIterator1 first1, InputIterator1 last1, InputIterator2 first2, T init, BinaryOperation1 binary_op1, BinaryOperation2 binary_op2); // C++17 template<class InputIterator, class T, class BinaryOperation, class UnaryOperation> - T + constexpr T // constexpr since C++20 transform_reduce(InputIterator first, InputIterator last, T init, BinaryOperation binary_op, UnaryOperation unary_op); // C++17 template <class InputIterator, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr since C++20 partial_sum(InputIterator first, InputIterator last, OutputIterator result); template <class InputIterator, class OutputIterator, class BinaryOperation> - OutputIterator + constexpr OutputIterator // constexpr since C++20 partial_sum(InputIterator first, InputIterator last, OutputIterator result, BinaryOperation binary_op); template<class InputIterator, class OutputIterator, class T> - OutputIterator + constexpr OutputIterator // constexpr since C++20 exclusive_scan(InputIterator first, InputIterator last, OutputIterator result, T init); // C++17 template<class InputIterator, class OutputIterator, class T, class BinaryOperation> - OutputIterator + constexpr OutputIterator // constexpr since C++20 exclusive_scan(InputIterator first, InputIterator last, OutputIterator result, T init, BinaryOperation binary_op); // C++17 template<class InputIterator, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr since C++20 inclusive_scan(InputIterator first, InputIterator last, OutputIterator result); // C++17 template<class InputIterator, class OutputIterator, class BinaryOperation> - OutputIterator + constexpr OutputIterator // constexpr since C++20 inclusive_scan(InputIterator first, InputIterator last, OutputIterator result, BinaryOperation binary_op); // C++17 template<class InputIterator, class OutputIterator, class BinaryOperation, class T> - OutputIterator + constexpr OutputIterator // constexpr since C++20 inclusive_scan(InputIterator first, InputIterator last, OutputIterator result, BinaryOperation binary_op, T init); // C++17 template<class InputIterator, class OutputIterator, class T, class BinaryOperation, class UnaryOperation> - OutputIterator + constexpr OutputIterator // constexpr since C++20 transform_exclusive_scan(InputIterator first, InputIterator last, OutputIterator result, T init, BinaryOperation binary_op, UnaryOperation unary_op); // C++17 template<class InputIterator, class OutputIterator, class BinaryOperation, class UnaryOperation> - OutputIterator + constexpr OutputIterator // constexpr since C++20 transform_inclusive_scan(InputIterator first, InputIterator last, OutputIterator result, BinaryOperation binary_op, UnaryOperation unary_op); // C++17 template<class InputIterator, class OutputIterator, class BinaryOperation, class UnaryOperation, class T> - OutputIterator + constexpr OutputIterator // constexpr since C++20 transform_inclusive_scan(InputIterator first, InputIterator last, OutputIterator result, BinaryOperation binary_op, UnaryOperation unary_op, T init); // C++17 template <class InputIterator, class OutputIterator> - OutputIterator + constexpr OutputIterator // constexpr since C++20 adjacent_difference(InputIterator first, InputIterator last, OutputIterator result); template <class InputIterator, class OutputIterator, class BinaryOperation> - OutputIterator + constexpr OutputIterator // constexpr since C++20 adjacent_difference(InputIterator first, InputIterator last, OutputIterator result, BinaryOperation binary_op); template <class ForwardIterator, class T> - void iota(ForwardIterator first, ForwardIterator last, T value); + constexpr void // constexpr since C++20 + iota(ForwardIterator first, ForwardIterator last, T value); template <class M, class N> constexpr common_type_t<M,N> gcd(M m, N n); // C++17 @@ -133,9 +134,11 @@ template <class M, class N> template <class M, class N> constexpr common_type_t<M,N> lcm(M m, N n); // C++17 -integer midpoint(integer a, integer b); // C++20 -pointer midpoint(pointer a, pointer b); // C++20 -floating_point midpoint(floating_point a, floating_point b); // C++20 +template<class T> + constexpr T midpoint(T a, T b) noexcept; // C++20 + +template<class T> + constexpr T* midpoint(T* a, T* b); // C++20 } // std @@ -158,28 +161,36 @@ _LIBCPP_PUSH_MACROS _LIBCPP_BEGIN_NAMESPACE_STD template <class _InputIterator, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp accumulate(_InputIterator __first, _InputIterator __last, _Tp __init) { for (; __first != __last; ++__first) +#if _LIBCPP_STD_VER > 17 + __init = _VSTD::move(__init) + *__first; +#else __init = __init + *__first; +#endif return __init; } template <class _InputIterator, class _Tp, class _BinaryOperation> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp accumulate(_InputIterator __first, _InputIterator __last, _Tp __init, _BinaryOperation __binary_op) { for (; __first != __last; ++__first) +#if _LIBCPP_STD_VER > 17 + __init = __binary_op(_VSTD::move(__init), *__first); +#else __init = __binary_op(__init, *__first); +#endif return __init; } #if _LIBCPP_STD_VER > 14 template <class _InputIterator, class _Tp, class _BinaryOp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp reduce(_InputIterator __first, _InputIterator __last, _Tp __init, _BinaryOp __b) { @@ -189,7 +200,7 @@ reduce(_InputIterator __first, _InputIterator __last, _Tp __init, _BinaryOp __b) } template <class _InputIterator, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp reduce(_InputIterator __first, _InputIterator __last, _Tp __init) { @@ -197,7 +208,7 @@ reduce(_InputIterator __first, _InputIterator __last, _Tp __init) } template <class _InputIterator> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename iterator_traits<_InputIterator>::value_type reduce(_InputIterator __first, _InputIterator __last) { @@ -207,29 +218,37 @@ reduce(_InputIterator __first, _InputIterator __last) #endif template <class _InputIterator1, class _InputIterator2, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp inner_product(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _Tp __init) { for (; __first1 != __last1; ++__first1, (void) ++__first2) +#if _LIBCPP_STD_VER > 17 + __init = _VSTD::move(__init) + *__first1 * *__first2; +#else __init = __init + *__first1 * *__first2; +#endif return __init; } template <class _InputIterator1, class _InputIterator2, class _Tp, class _BinaryOperation1, class _BinaryOperation2> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp inner_product(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _Tp __init, _BinaryOperation1 __binary_op1, _BinaryOperation2 __binary_op2) { for (; __first1 != __last1; ++__first1, (void) ++__first2) +#if _LIBCPP_STD_VER > 17 + __init = __binary_op1(_VSTD::move(__init), __binary_op2(*__first1, *__first2)); +#else __init = __binary_op1(__init, __binary_op2(*__first1, *__first2)); +#endif return __init; } #if _LIBCPP_STD_VER > 14 template <class _InputIterator, class _Tp, class _BinaryOp, class _UnaryOp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp transform_reduce(_InputIterator __first, _InputIterator __last, _Tp __init, _BinaryOp __b, _UnaryOp __u) @@ -241,7 +260,7 @@ transform_reduce(_InputIterator __first, _InputIterator __last, template <class _InputIterator1, class _InputIterator2, class _Tp, class _BinaryOp1, class _BinaryOp2> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp transform_reduce(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _Tp __init, _BinaryOp1 __b1, _BinaryOp2 __b2) @@ -252,7 +271,7 @@ transform_reduce(_InputIterator1 __first1, _InputIterator1 __last1, } template <class _InputIterator1, class _InputIterator2, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _Tp transform_reduce(_InputIterator1 __first1, _InputIterator1 __last1, _InputIterator2 __first2, _Tp __init) @@ -263,7 +282,7 @@ transform_reduce(_InputIterator1 __first1, _InputIterator1 __last1, #endif template <class _InputIterator, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator partial_sum(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { @@ -273,7 +292,11 @@ partial_sum(_InputIterator __first, _InputIterator __last, _OutputIterator __res *__result = __t; for (++__first, (void) ++__result; __first != __last; ++__first, (void) ++__result) { +#if _LIBCPP_STD_VER > 17 + __t = _VSTD::move(__t) + *__first; +#else __t = __t + *__first; +#endif *__result = __t; } } @@ -281,7 +304,7 @@ partial_sum(_InputIterator __first, _InputIterator __last, _OutputIterator __res } template <class _InputIterator, class _OutputIterator, class _BinaryOperation> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator partial_sum(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _BinaryOperation __binary_op) @@ -292,7 +315,11 @@ partial_sum(_InputIterator __first, _InputIterator __last, _OutputIterator __res *__result = __t; for (++__first, (void) ++__result; __first != __last; ++__first, (void) ++__result) { +#if _LIBCPP_STD_VER > 17 + __t = __binary_op(_VSTD::move(__t), *__first); +#else __t = __binary_op(__t, *__first); +#endif *__result = __t; } } @@ -301,27 +328,30 @@ partial_sum(_InputIterator __first, _InputIterator __last, _OutputIterator __res #if _LIBCPP_STD_VER > 14 template <class _InputIterator, class _OutputIterator, class _Tp, class _BinaryOp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator exclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _Tp __init, _BinaryOp __b) { if (__first != __last) { - _Tp __saved = __init; - do + _Tp __tmp(__b(__init, *__first)); + while (true) { - __init = __b(__init, *__first); - *__result = __saved; - __saved = __init; + *__result = _VSTD::move(__init); ++__result; - } while (++__first != __last); + ++__first; + if (__first == __last) + break; + __init = _VSTD::move(__tmp); + __tmp = __b(__init, *__first); + } } return __result; } template <class _InputIterator, class _OutputIterator, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator exclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _Tp __init) @@ -330,40 +360,43 @@ exclusive_scan(_InputIterator __first, _InputIterator __last, } template <class _InputIterator, class _OutputIterator, class _Tp, class _BinaryOp> +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator inclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _BinaryOp __b, _Tp __init) { for (; __first != __last; ++__first, (void) ++__result) { __init = __b(__init, *__first); *__result = __init; - } + } return __result; } template <class _InputIterator, class _OutputIterator, class _BinaryOp> +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator inclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _BinaryOp __b) { if (__first != __last) { - typename std::iterator_traits<_InputIterator>::value_type __init = *__first; + typename iterator_traits<_InputIterator>::value_type __init = *__first; *__result++ = __init; if (++__first != __last) return _VSTD::inclusive_scan(__first, __last, __result, __b, __init); - } + } return __result; } template <class _InputIterator, class _OutputIterator> +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator inclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { - return _VSTD::inclusive_scan(__first, __last, __result, std::plus<>()); + return _VSTD::inclusive_scan(__first, __last, __result, _VSTD::plus<>()); } template <class _InputIterator, class _OutputIterator, class _Tp, class _BinaryOp, class _UnaryOp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator transform_exclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _Tp __init, @@ -384,7 +417,9 @@ transform_exclusive_scan(_InputIterator __first, _InputIterator __last, } template <class _InputIterator, class _OutputIterator, class _Tp, class _BinaryOp, class _UnaryOp> -_OutputIterator transform_inclusive_scan(_InputIterator __first, _InputIterator __last, +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 +_OutputIterator +transform_inclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _BinaryOp __b, _UnaryOp __u, _Tp __init) { for (; __first != __last; ++__first, (void) ++__result) { @@ -396,61 +431,71 @@ _OutputIterator transform_inclusive_scan(_InputIterator __first, _InputIterator } template <class _InputIterator, class _OutputIterator, class _BinaryOp, class _UnaryOp> -_OutputIterator transform_inclusive_scan(_InputIterator __first, _InputIterator __last, +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 +_OutputIterator +transform_inclusive_scan(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _BinaryOp __b, _UnaryOp __u) { if (__first != __last) { - typename std::iterator_traits<_InputIterator>::value_type __init = __u(*__first); + typename iterator_traits<_InputIterator>::value_type __init = __u(*__first); *__result++ = __init; if (++__first != __last) return _VSTD::transform_inclusive_scan(__first, __last, __result, __b, __u, __init); - } + } return __result; } #endif template <class _InputIterator, class _OutputIterator> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator adjacent_difference(_InputIterator __first, _InputIterator __last, _OutputIterator __result) { if (__first != __last) { - typename iterator_traits<_InputIterator>::value_type __t1(*__first); - *__result = __t1; + typename iterator_traits<_InputIterator>::value_type __acc(*__first); + *__result = __acc; for (++__first, (void) ++__result; __first != __last; ++__first, (void) ++__result) { - typename iterator_traits<_InputIterator>::value_type __t2(*__first); - *__result = __t2 - __t1; - __t1 = _VSTD::move(__t2); + typename iterator_traits<_InputIterator>::value_type __val(*__first); +#if _LIBCPP_STD_VER > 17 + *__result = __val - _VSTD::move(__acc); +#else + *__result = __val - __acc; +#endif + __acc = _VSTD::move(__val); } } return __result; } template <class _InputIterator, class _OutputIterator, class _BinaryOperation> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 _OutputIterator adjacent_difference(_InputIterator __first, _InputIterator __last, _OutputIterator __result, _BinaryOperation __binary_op) { if (__first != __last) { - typename iterator_traits<_InputIterator>::value_type __t1(*__first); - *__result = __t1; + typename iterator_traits<_InputIterator>::value_type __acc(*__first); + *__result = __acc; for (++__first, (void) ++__result; __first != __last; ++__first, (void) ++__result) { - typename iterator_traits<_InputIterator>::value_type __t2(*__first); - *__result = __binary_op(__t2, __t1); - __t1 = _VSTD::move(__t2); + typename iterator_traits<_InputIterator>::value_type __val(*__first); +#if _LIBCPP_STD_VER > 17 + *__result = __binary_op(__val, _VSTD::move(__acc)); +#else + *__result = __binary_op(__val, __acc); +#endif + __acc = _VSTD::move(__val); } } return __result; } template <class _ForwardIterator, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 void iota(_ForwardIterator __first, _ForwardIterator __last, _Tp __value_) { @@ -467,9 +512,9 @@ struct __ct_abs<_Result, _Source, true> { _LIBCPP_CONSTEXPR _LIBCPP_INLINE_VISIBILITY _Result operator()(_Source __t) const noexcept { - if (__t >= 0) return __t; - if (__t == numeric_limits<_Source>::min()) return -static_cast<_Result>(__t); - return -__t; + if (__t >= 0) return __t; + if (__t == numeric_limits<_Source>::min()) return -static_cast<_Result>(__t); + return -__t; } }; @@ -531,8 +576,8 @@ enable_if_t<is_integral_v<_Tp> && !is_same_v<bool, _Tp> && !is_null_pointer_v<_T midpoint(_Tp __a, _Tp __b) noexcept _LIBCPP_DISABLE_UBSAN_UNSIGNED_INTEGER_CHECK { - using _Up = std::make_unsigned_t<_Tp>; - constexpr _Up __bitshift = std::numeric_limits<_Up>::digits - 1; + using _Up = make_unsigned_t<_Tp>; + constexpr _Up __bitshift = numeric_limits<_Up>::digits - 1; _Up __diff = _Up(__b) - _Up(__a); _Up __sign_bit = __b < __a; diff --git a/lib/libcxx/include/optional b/lib/libcxx/include/optional index a147d69da0..97a0bbe66c 100644 --- a/lib/libcxx/include/optional +++ b/lib/libcxx/include/optional @@ -147,6 +147,7 @@ template<class T> */ #include <__config> +#include <__availability> #include <__debug> #include <__functional_base> #include <functional> @@ -320,7 +321,7 @@ struct __optional_storage_base : __optional_destruct_base<_Tp> void __construct(_Args&&... __args) { _LIBCPP_ASSERT(!has_value(), "__construct called for engaged __optional_storage"); - ::new((void*)_VSTD::addressof(this->__val_)) value_type(_VSTD::forward<_Args>(__args)...); + ::new ((void*)_VSTD::addressof(this->__val_)) value_type(_VSTD::forward<_Args>(__args)...); this->__engaged_ = true; } @@ -656,7 +657,7 @@ private: } template <class _Up, class _QUp = _QualUp> static constexpr bool __enable_assign() { - // Construction and assignability of _Qup to _Tp has already been + // Construction and assignability of _QUp to _Tp has already been // checked. return !__check_constructible_from_opt<_Up>::value && !__check_assignable_from_opt<_Up>::value; diff --git a/lib/libcxx/include/ostream b/lib/libcxx/include/ostream index 697732d54e..f5eb8a8946 100644 --- a/lib/libcxx/include/ostream +++ b/lib/libcxx/include/ostream @@ -126,9 +126,8 @@ template <class charT, class traits> basic_ostream<charT,traits>& flush(basic_ostream<charT,traits>& os); // rvalue stream insertion -template <class charT, class traits, class T> - basic_ostream<charT, traits>& - operator<<(basic_ostream<charT, traits>&& os, const T& x); +template <class Stream, class T> + Stream&& operator<<(Stream&& os, const T& x); } // std @@ -1028,15 +1027,20 @@ flush(basic_ostream<_CharT, _Traits>& __os) #ifndef _LIBCPP_CXX03_LANG +template <class _Stream, class _Tp, class = void> +struct __is_ostreamable : false_type { }; + template <class _Stream, class _Tp> -inline _LIBCPP_INLINE_VISIBILITY -typename enable_if -< - !is_lvalue_reference<_Stream>::value && - is_base_of<ios_base, _Stream>::value, - _Stream&& ->::type -operator<<(_Stream&& __os, const _Tp& __x) +struct __is_ostreamable<_Stream, _Tp, decltype( + _VSTD::declval<_Stream>() << _VSTD::declval<_Tp>(), void() +)> : true_type { }; + +template <class _Stream, class _Tp, class = typename enable_if< + _And<is_base_of<ios_base, _Stream>, + __is_ostreamable<_Stream&, const _Tp&>>::value +>::type> +_LIBCPP_INLINE_VISIBILITY +_Stream&& operator<<(_Stream&& __os, const _Tp& __x) { __os << __x; return _VSTD::move(__os); @@ -1097,10 +1101,8 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const bitset<_Size>& __x) use_facet<ctype<_CharT> >(__os.getloc()).widen('1')); } -#ifndef _LIBCPP_DO_NOT_ASSUME_STREAMS_EXPLICIT_INSTANTIATION_IN_DYLIB _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_ostream<char>) _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_ostream<wchar_t>) -#endif _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/include/queue b/lib/libcxx/include/queue index 33c25e0dfc..a2048c1e22 100644 --- a/lib/libcxx/include/queue +++ b/lib/libcxx/include/queue @@ -112,18 +112,11 @@ protected: Compare comp; public: - priority_queue() = default; - ~priority_queue() = default; - - priority_queue(const priority_queue& q) = default; - priority_queue(priority_queue&& q) = default; - - priority_queue& operator=(const priority_queue& q) = default; - priority_queue& operator=(priority_queue&& q) = default; - - explicit priority_queue(const Compare& comp); - priority_queue(const Compare& comp, const container_type& c); - explicit priority_queue(const Compare& comp, container_type&& c); + priority_queue() : priority_queue(Compare()) {} // C++20 + explicit priority_queue(const Compare& x) : priority_queue(x, Container()) {} + priority_queue(const Compare& x, const Container&); + explicit priority_queue(const Compare& x = Compare(), Container&&= Container()); // before C++20 + priority_queue(const Compare& x, Container&&); // C++20 template <class InputIterator> priority_queue(InputIterator first, InputIterator last, const Compare& comp = Compare()); @@ -474,7 +467,7 @@ public: priority_queue(const value_compare& __comp, const container_type& __c); #ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY - explicit priority_queue(const value_compare& __comp, container_type&& __c); + priority_queue(const value_compare& __comp, container_type&& __c); #endif template <class _InputIter> _LIBCPP_INLINE_VISIBILITY diff --git a/lib/libcxx/include/random b/lib/libcxx/include/random index ce3d135e06..6e0d2ecb47 100644 --- a/lib/libcxx/include/random +++ b/lib/libcxx/include/random @@ -36,7 +36,9 @@ public: static constexpr result_type default_seed = 1u; // constructors and seeding functions - explicit linear_congruential_engine(result_type s = default_seed); + explicit linear_congruential_engine(result_type s = default_seed); // before C++20 + linear_congruential_engine() : linear_congruential_engine(default_seed) {} // C++20 + explicit linear_congruential_engine(result_type s); // C++20 template<class Sseq> explicit linear_congruential_engine(Sseq& q); void seed(result_type s = default_seed); template<class Sseq> void seed(Sseq& q); @@ -96,7 +98,9 @@ public: static constexpr result_type default_seed = 5489u; // constructors and seeding functions - explicit mersenne_twister_engine(result_type value = default_seed); + explicit mersenne_twister_engine(result_type s = default_seed); // before C++20 + mersenne_twister_engine() : mersenne_twister_engine(default_seed) {} // C++20 + explicit mersenne_twister_engine(result_type s); // C++20 template<class Sseq> explicit mersenne_twister_engine(Sseq& q); void seed(result_type value = default_seed); template<class Sseq> void seed(Sseq& q); @@ -154,7 +158,9 @@ public: static constexpr result_type default_seed = 19780503u; // constructors and seeding functions - explicit subtract_with_carry_engine(result_type value = default_seed); + explicit subtract_with_carry_engine(result_type value = default_seed); // before C++20 + subtract_with_carry_engine() : subtract_with_carry_engine(default_seed) {} // C++20 + explicit subtract_with_carry_engine(result_type value); // C++20 template<class Sseq> explicit subtract_with_carry_engine(Sseq& q); void seed(result_type value = default_seed); template<class Sseq> void seed(Sseq& q); @@ -385,7 +391,9 @@ public: static constexpr result_type max() { return numeric_limits<result_type>::max(); } // constructors - explicit random_device(const string& token = "/dev/urandom"); + explicit random_device(const string& token = implementation-defined); // before C++20 + random_device() : random_device(implementation-defined) {} // C++20 + explicit random_device(const string& token); // C++20 // generating functions result_type operator()(); @@ -456,7 +464,10 @@ public: // constructors and reset functions explicit uniform_int_distribution(IntType a = 0, - IntType b = numeric_limits<IntType>::max()); + IntType b = numeric_limits<IntType>::max()); // before C++20 + uniform_int_distribution() : uniform_int_distribution(0) {} // C++20 + explicit uniform_int_distribution(IntType a, + IntType b = numeric_limits<IntType>::max()); // C++20 explicit uniform_int_distribution(const param_type& parm); void reset(); @@ -515,7 +526,9 @@ public: }; // constructors and reset functions - explicit uniform_real_distribution(RealType a = 0.0, RealType b = 1.0); + explicit uniform_real_distribution(RealType a = 0.0, RealType b = 1.0); // before C++20 + uniform_real_distribution() : uniform_real_distribution(0.0) {} // C++20 + explicit uniform_real_distribution(RealType a, RealType b = 1.0); // C++20 explicit uniform_real_distribution(const param_type& parm); void reset(); @@ -571,7 +584,9 @@ public: }; // constructors and reset functions - explicit bernoulli_distribution(double p = 0.5); + explicit bernoulli_distribution(double p = 0.5); // before C++20 + bernoulli_distribution() : bernoulli_distribution(0.5) {} // C++20 + explicit bernoulli_distribution(double p); // C++20 explicit bernoulli_distribution(const param_type& parm); void reset(); @@ -628,7 +643,9 @@ public: }; // constructors and reset functions - explicit binomial_distribution(IntType t = 1, double p = 0.5); + explicit binomial_distribution(IntType t = 1, double p = 0.5); // before C++20 + binomial_distribution() : binomial_distribution(1) {} // C++20 + explicit binomial_distribution(IntType t, double p = 0.5); // C++20 explicit binomial_distribution(const param_type& parm); void reset(); @@ -685,7 +702,9 @@ public: }; // constructors and reset functions - explicit geometric_distribution(double p = 0.5); + explicit geometric_distribution(double p = 0.5); // before C++20 + geometric_distribution() : geometric_distribution(0.5) {} // C++20 + explicit geometric_distribution(double p); // C++20 explicit geometric_distribution(const param_type& parm); void reset(); @@ -742,7 +761,9 @@ public: }; // constructor and reset functions - explicit negative_binomial_distribution(result_type k = 1, double p = 0.5); + explicit negative_binomial_distribution(IntType k = 1, double p = 0.5); // before C++20 + negative_binomial_distribution() : negative_binomial_distribution(1) {} // C++20 + explicit negative_binomial_distribution(IntType k, double p = 0.5); // C++20 explicit negative_binomial_distribution(const param_type& parm); void reset(); @@ -799,7 +820,9 @@ public: }; // constructors and reset functions - explicit poisson_distribution(double mean = 1.0); + explicit poisson_distribution(double mean = 1.0); // before C++20 + poisson_distribution() : poisson_distribution(1.0) {} // C++20 + explicit poisson_distribution(double mean); // C++20 explicit poisson_distribution(const param_type& parm); void reset(); @@ -855,7 +878,9 @@ public: }; // constructors and reset functions - explicit exponential_distribution(result_type lambda = 1.0); + explicit exponential_distribution(RealType lambda = 1.0); // before C++20 + exponential_distribution() : exponential_distribution(1.0) {} // C++20 + explicit exponential_distribution(RealType lambda); // C++20 explicit exponential_distribution(const param_type& parm); void reset(); @@ -912,7 +937,9 @@ public: }; // constructors and reset functions - explicit gamma_distribution(result_type alpha = 1, result_type beta = 1); + explicit gamma_distribution(RealType alpha = 0.0, RealType beta = 1.0); // before C++20 + gamma_distribution() : gamma_distribution(0.0) {} // C++20 + explicit gamma_distribution(RealType alpha, RealType beta = 1.0); // C++20 explicit gamma_distribution(const param_type& parm); void reset(); @@ -970,7 +997,9 @@ public: }; // constructor and reset functions - explicit weibull_distribution(result_type a = 1, result_type b = 1); + explicit weibull_distribution(RealType a = 1.0, RealType b = 1.0); // before C++20 + weibull_distribution() : weibull_distribution(1.0) {} // C++20 + explicit weibull_distribution(RealType a, RealType b = 1.0); // C++20 explicit weibull_distribution(const param_type& parm); void reset(); @@ -1028,7 +1057,9 @@ public: }; // constructor and reset functions - explicit extreme_value_distribution(result_type a = 0, result_type b = 1); + explicit extreme_value_distribution(RealType a = 0.0, RealType b = 1.0); // before C++20 + extreme_value_distribution() : extreme_value_distribution(0.0) {} // C++20 + explicit extreme_value_distribution(RealType a, RealType b = 1.0); // C++20 explicit extreme_value_distribution(const param_type& parm); void reset(); @@ -1086,7 +1117,9 @@ public: }; // constructors and reset functions - explicit normal_distribution(result_type mean = 0, result_type stddev = 1); + explicit normal_distribution(RealType mean = 0.0, RealType stddev = 1.0); // before C++20 + normal_distribution() : normal_distribution(0.0) {} // C++20 + explicit normal_distribution(RealType mean, RealType stddev = 1.0); // C++20 explicit normal_distribution(const param_type& parm); void reset(); @@ -1144,7 +1177,9 @@ public: }; // constructor and reset functions - explicit lognormal_distribution(result_type m = 0, result_type s = 1); + explicit lognormal_distribution(RealType mean = 0.0, RealType stddev = 1.0); // before C++20 + lognormal_distribution() : lognormal_distribution(0.0) {} // C++20 + explicit lognormal_distribution(RealType mean, RealType stddev = 1.0); // C++20 explicit lognormal_distribution(const param_type& parm); void reset(); @@ -1201,7 +1236,9 @@ public: }; // constructor and reset functions - explicit chi_squared_distribution(result_type n = 1); + explicit chi_squared_distribution(RealType n = 1.0); // before C++20 + chi_squared_distribution() : chi_squared_distribution(1.0) {} // C++20 + explicit chi_squared_distribution(RealType n); // C++20 explicit chi_squared_distribution(const param_type& parm); void reset(); @@ -1258,7 +1295,9 @@ public: }; // constructor and reset functions - explicit cauchy_distribution(result_type a = 0, result_type b = 1); + explicit cauchy_distribution(RealType a = 0.0, RealType b = 1.0); // before C++20 + cauchy_distribution() : cauchy_distribution(0.0) {} // C++20 + explicit cauchy_distribution(RealType a, RealType b = 1.0); // C++20 explicit cauchy_distribution(const param_type& parm); void reset(); @@ -1316,7 +1355,9 @@ public: }; // constructor and reset functions - explicit fisher_f_distribution(result_type m = 1, result_type n = 1); + explicit fisher_f_distribution(RealType m = 1.0, RealType n = 1.0); // before C++20 + fisher_f_distribution() : fisher_f_distribution(1.0) {} // C++20 + explicit fisher_f_distribution(RealType m, RealType n = 1.0); // C++20 explicit fisher_f_distribution(const param_type& parm); void reset(); @@ -1373,7 +1414,9 @@ public: }; // constructor and reset functions - explicit student_t_distribution(result_type n = 1); + explicit student_t_distribution(RealType n = 1.0); // before C++20 + student_t_distribution() : student_t_distribution(1.0) {} // C++20 + explicit student_t_distribution(RealType n); // C++20 explicit student_t_distribution(const param_type& parm); void reset(); @@ -1642,8 +1685,7 @@ class piecewise_linear_distribution #include <numeric> #include <vector> #include <string> -#include <istream> -#include <ostream> +#include <iosfwd> #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header @@ -1669,7 +1711,23 @@ struct __is_seed_sequence template <unsigned long long __a, unsigned long long __c, unsigned long long __m, unsigned long long _Mp, - bool _MightOverflow = (__a != 0 && __m != 0 && __m-1 > (_Mp-__c)/__a)> + bool _MightOverflow = (__a != 0 && __m != 0 && __m-1 > (_Mp-__c)/__a), + bool _OverflowOK = ((__m|__m-1) > __m), // m = 2^n + bool _SchrageOK = (__a != 0 && __m != 0 && __m % __a <= __m / __a)> // r <= q +struct __lce_alg_picker +{ + static_assert(__a != 0 || __m != 0 || !_MightOverflow || _OverflowOK || _SchrageOK, + "The current values of a, c, and m cannot generate a number " + "within bounds of linear_congruential_engine."); + + static _LIBCPP_CONSTEXPR const bool __use_schrage = _MightOverflow && + !_OverflowOK && + _SchrageOK; +}; + +template <unsigned long long __a, unsigned long long __c, + unsigned long long __m, unsigned long long _Mp, + bool _UseSchrage = __lce_alg_picker<__a, __c, __m, _Mp>::__use_schrage> struct __lce_ta; // 64 @@ -1843,6 +1901,7 @@ private: static_assert(__m == 0 || __a < __m, "linear_congruential_engine invalid parameters"); static_assert(__m == 0 || __c < __m, "linear_congruential_engine invalid parameters"); + static_assert(_VSTD::is_unsigned<_UIntType>::value, "_UIntType must be unsigned type"); public: static _LIBCPP_CONSTEXPR const result_type _Min = __c == 0u ? 1u: 0u; static _LIBCPP_CONSTEXPR const result_type _Max = __m - 1u; @@ -1859,9 +1918,17 @@ public: static _LIBCPP_CONSTEXPR const result_type default_seed = 1u; // constructors and seeding functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + linear_congruential_engine() : linear_congruential_engine(default_seed) {} + _LIBCPP_INLINE_VISIBILITY + explicit linear_congruential_engine(result_type __s) { seed(__s); } +#else _LIBCPP_INLINE_VISIBILITY - explicit linear_congruential_engine(result_type __s = default_seed) - {seed(__s);} + explicit linear_congruential_engine(result_type __s = default_seed) { + seed(__s); + } +#endif template<class _Sseq> _LIBCPP_INLINE_VISIBILITY explicit linear_congruential_engine(_Sseq& __q, @@ -1982,7 +2049,8 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const linear_congruential_engine<_UIntType, __a, __c, __m>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left); + typedef basic_ostream<_CharT, _Traits> _Ostream; + __os.flags(_Ostream::dec | _Ostream::left); __os.fill(__os.widen(' ')); return __os << __x.__x_; } @@ -1994,7 +2062,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, linear_congruential_engine<_UIntType, __a, __c, __m>& __x) { __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); _UIntType __t; __is >> __t; if (!__is.fail()) @@ -2106,9 +2175,17 @@ public: static _LIBCPP_CONSTEXPR const result_type default_seed = 5489u; // constructors and seeding functions +#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY - explicit mersenne_twister_engine(result_type __sd = default_seed) - {seed(__sd);} + mersenne_twister_engine() : mersenne_twister_engine(default_seed) {} + _LIBCPP_INLINE_VISIBILITY + explicit mersenne_twister_engine(result_type __sd) { seed(__sd); } +#else + _LIBCPP_INLINE_VISIBILITY + explicit mersenne_twister_engine(result_type __sd = default_seed) { + seed(__sd); + } +#endif template<class _Sseq> _LIBCPP_INLINE_VISIBILITY explicit mersenne_twister_engine(_Sseq& __q, @@ -2453,7 +2530,8 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, _Bp, _Tp, _Cp, _Lp, _Fp>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left); + typedef basic_ostream<_CharT, _Traits> _Ostream; + __os.flags(_Ostream::dec | _Ostream::left); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.__x_[__x.__i_]; @@ -2474,7 +2552,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, _Bp, _Tp, _Cp, _Lp, _Fp>& __x) { __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); _UInt __t[_Np]; for (size_t __i = 0; __i < _Np; ++__i) __is >> __t[__i]; @@ -2562,9 +2641,17 @@ public: static _LIBCPP_CONSTEXPR const result_type default_seed = 19780503u; // constructors and seeding functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + subtract_with_carry_engine() : subtract_with_carry_engine(default_seed) {} + _LIBCPP_INLINE_VISIBILITY + explicit subtract_with_carry_engine(result_type __sd) { seed(__sd); } +#else _LIBCPP_INLINE_VISIBILITY - explicit subtract_with_carry_engine(result_type __sd = default_seed) - {seed(__sd);} + explicit subtract_with_carry_engine(result_type __sd = default_seed) { + seed(__sd); + } +#endif template<class _Sseq> _LIBCPP_INLINE_VISIBILITY explicit subtract_with_carry_engine(_Sseq& __q, @@ -2773,7 +2860,8 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const subtract_with_carry_engine<_UInt, _Wp, _Sp, _Rp>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left); + typedef basic_ostream<_CharT, _Traits> _Ostream; + __os.flags(_Ostream::dec | _Ostream::left); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.__x_[__x.__i_]; @@ -2792,7 +2880,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, subtract_with_carry_engine<_UInt, _Wp, _Sp, _Rp>& __x) { __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); _UInt __t[_Rp+1]; for (size_t __i = 0; __i < _Rp+1; ++__i) __is >> __t[__i]; @@ -2955,7 +3044,8 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const discard_block_engine<_Eng, _Pp, _Rp>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left); + typedef basic_ostream<_CharT, _Traits> _Ostream; + __os.flags(_Ostream::dec | _Ostream::left); _CharT __sp = __os.widen(' '); __os.fill(__sp); return __os << __x.__e_ << __sp << __x.__n_; @@ -2968,7 +3058,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, discard_block_engine<_Eng, _Pp, _Rp>& __x) { __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); _Eng __e; int __n; __is >> __e >> __n; @@ -3440,7 +3531,8 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const shuffle_order_engine<_Eng, _Kp>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left); + typedef basic_ostream<_CharT, _Traits> _Ostream; + __os.flags(_Ostream::dec | _Ostream::left); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.__e_ << __sp << __x._V_[0]; @@ -3457,7 +3549,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, { typedef typename shuffle_order_engine<_Eng, _Kp>::result_type result_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); _Eng __e; result_type _Vp[_Kp+1]; __is >> __e; @@ -3477,6 +3570,8 @@ typedef shuffle_order_engine<minstd_rand0, 256> knuth_b; // random_device +#if !defined(_LIBCPP_HAS_NO_RANDOM_DEVICE) + class _LIBCPP_TYPE_VIS random_device { #ifdef _LIBCPP_USING_DEV_RANDOM @@ -3496,7 +3591,12 @@ public: static _LIBCPP_CONSTEXPR result_type max() { return _Max;} // constructors +#ifndef _LIBCPP_CXX03_LANG + random_device() : random_device("/dev/urandom") {} + explicit random_device(const string& __token); +#else explicit random_device(const string& __token = "/dev/urandom"); +#endif ~random_device(); // generating functions @@ -3511,6 +3611,8 @@ private: random_device& operator=(const random_device&); // = delete; }; +#endif // !_LIBCPP_HAS_NO_RANDOM_DEVICE + // seed_seq class _LIBCPP_TEMPLATE_VIS seed_seq @@ -3663,7 +3765,8 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const uniform_int_distribution<_IT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left); + typedef basic_ostream<_CharT, _Traits> _Ostream; + __os.flags(_Ostream::dec | _Ostream::left); _CharT __sp = __os.widen(' '); __os.fill(__sp); return __os << __x.a() << __sp << __x.b(); @@ -3678,7 +3781,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __a; result_type __b; __is >> __a >> __b; @@ -3726,9 +3830,16 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + uniform_real_distribution() : uniform_real_distribution(0) {} + explicit uniform_real_distribution(result_type __a, result_type __b = 1) + : __p_(param_type(__a, __b)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit uniform_real_distribution(result_type __a = 0, result_type __b = 1) : __p_(param_type(__a, __b)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit uniform_real_distribution(const param_type& __p) : __p_(__p) {} _LIBCPP_INLINE_VISIBILITY @@ -3784,8 +3895,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const uniform_real_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); return __os << __x.a() << __sp << __x.b(); @@ -3800,7 +3912,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __a; result_type __b; __is >> __a >> __b; @@ -3842,9 +3955,15 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + bernoulli_distribution() : bernoulli_distribution(0.5) {} _LIBCPP_INLINE_VISIBILITY - explicit bernoulli_distribution(double __p = 0.5) - : __p_(param_type(__p)) {} + explicit bernoulli_distribution(double __p) : __p_(param_type(__p)) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit bernoulli_distribution(double __p = 0.5) : __p_(param_type(__p)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit bernoulli_distribution(const param_type& __p) : __p_(__p) {} _LIBCPP_INLINE_VISIBILITY @@ -3895,8 +4014,9 @@ basic_ostream<_CharT, _Traits>& operator<<(basic_ostream<_CharT, _Traits>& __os, const bernoulli_distribution& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); return __os << __x.p(); @@ -3909,7 +4029,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, bernoulli_distribution& __x) typedef bernoulli_distribution _Eng; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); double __p; __is >> __p; if (!__is.fail()) @@ -3958,9 +4079,17 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + binomial_distribution() : binomial_distribution(1) {} + _LIBCPP_INLINE_VISIBILITY + explicit binomial_distribution(result_type __t, double __p = 0.5) + : __p_(param_type(__t, __p)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit binomial_distribution(result_type __t = 1, double __p = 0.5) : __p_(param_type(__t, __p)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit binomial_distribution(const param_type& __p) : __p_(__p) {} _LIBCPP_INLINE_VISIBILITY @@ -3999,12 +4128,12 @@ public: {return !(__x == __y);} }; -#ifndef _LIBCPP_MSVCRT +#ifndef _LIBCPP_MSVCRT_LIKE extern "C" double lgamma_r(double, int *); #endif inline _LIBCPP_INLINE_VISIBILITY double __libcpp_lgamma(double __d) { -#if defined(_LIBCPP_MSVCRT) +#if defined(_LIBCPP_MSVCRT_LIKE) return lgamma(__d); #else int __sign; @@ -4079,8 +4208,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const binomial_distribution<_IntType>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); return __os << __x.t() << __sp << __x.p(); @@ -4095,7 +4225,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __t; double __p; __is >> __t >> __p; @@ -4138,9 +4269,17 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + exponential_distribution() : exponential_distribution(1) {} + _LIBCPP_INLINE_VISIBILITY + explicit exponential_distribution(result_type __lambda) + : __p_(param_type(__lambda)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit exponential_distribution(result_type __lambda = 1) : __p_(param_type(__lambda)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit exponential_distribution(const param_type& __p) : __p_(__p) {} _LIBCPP_INLINE_VISIBILITY @@ -4197,8 +4336,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const exponential_distribution<_RealType>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); return __os << __x.lambda(); } @@ -4211,7 +4351,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __lambda; __is >> __lambda; if (!__is.fail()) @@ -4259,9 +4400,18 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + normal_distribution() : normal_distribution(0) {} _LIBCPP_INLINE_VISIBILITY - explicit normal_distribution(result_type __mean = 0, result_type __stddev = 1) + explicit normal_distribution(result_type __mean, result_type __stddev = 1) : __p_(param_type(__mean, __stddev)), _V_hot_(false) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit normal_distribution(result_type __mean = 0, + result_type __stddev = 1) + : __p_(param_type(__mean, __stddev)), _V_hot_(false) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit normal_distribution(const param_type& __p) : __p_(__p), _V_hot_(false) {} @@ -4351,8 +4501,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const normal_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.mean() << __sp << __x.stddev() << __sp << __x._V_hot_; @@ -4370,7 +4521,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __mean; result_type __stddev; result_type _Vp = 0; @@ -4437,9 +4589,18 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + lognormal_distribution() : lognormal_distribution(0) {} + _LIBCPP_INLINE_VISIBILITY + explicit lognormal_distribution(result_type __m, result_type __s = 1) + : __p_(param_type(__m, __s)) {} +#else _LIBCPP_INLINE_VISIBILITY - explicit lognormal_distribution(result_type __m = 0, result_type __s = 1) + explicit lognormal_distribution(result_type __m = 0, + result_type __s = 1) : __p_(param_type(__m, __s)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit lognormal_distribution(const param_type& __p) : __p_(__p) {} @@ -4557,8 +4718,17 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + poisson_distribution() : poisson_distribution(1.0) {} + _LIBCPP_INLINE_VISIBILITY + explicit poisson_distribution(double __mean) + : __p_(__mean) {} +#else _LIBCPP_INLINE_VISIBILITY - explicit poisson_distribution(double __mean = 1.0) : __p_(__mean) {} + explicit poisson_distribution(double __mean = 1.0) + : __p_(__mean) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit poisson_distribution(const param_type& __p) : __p_(__p) {} _LIBCPP_INLINE_VISIBILITY @@ -4618,7 +4788,7 @@ poisson_distribution<_IntType>::param_type::param_type(double __mean) { __s_ = _VSTD::sqrt(__mean_); __d_ = 6 * __mean_ * __mean_; - __l_ = std::trunc(__mean_ - 1.1484); + __l_ = _VSTD::trunc(__mean_ - 1.1484); __omega_ = .3989423 / __s_; double __b1_ = .4166667E-1 / __mean_; double __b2_ = .3 * __b1_ * __b1_; @@ -4650,13 +4820,13 @@ poisson_distribution<_IntType>::operator()(_URNG& __urng, const param_type& __pr double __u; if (__g > 0) { - __tx = std::trunc(__g); + __tx = _VSTD::trunc(__g); if (__tx >= __pr.__l_) - return std::__clamp_to_integral<result_type>(__tx); + return _VSTD::__clamp_to_integral<result_type>(__tx); __difmuk = __pr.__mean_ - __tx; __u = __urd(__urng); if (__pr.__d_ * __u >= __difmuk * __difmuk * __difmuk) - return std::__clamp_to_integral<result_type>(__tx); + return _VSTD::__clamp_to_integral<result_type>(__tx); } exponential_distribution<double> __edist; for (bool __using_exp_dist = false; true; __using_exp_dist = true) @@ -4672,7 +4842,7 @@ poisson_distribution<_IntType>::operator()(_URNG& __urng, const param_type& __pr __u += __u - 1; __t = 1.8 + (__u < 0 ? -__e : __e); } while (__t <= -.6744); - __tx = std::trunc(__pr.__mean_ + __pr.__s_ * __t); + __tx = _VSTD::trunc(__pr.__mean_ + __pr.__s_ * __t); __difmuk = __pr.__mean_ - __tx; __using_exp_dist = true; } @@ -4716,7 +4886,7 @@ poisson_distribution<_IntType>::operator()(_URNG& __urng, const param_type& __pr } } } - return std::__clamp_to_integral<result_type>(__tx); + return _VSTD::__clamp_to_integral<result_type>(__tx); } template <class _CharT, class _Traits, class _IntType> @@ -4725,8 +4895,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const poisson_distribution<_IntType>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); return __os << __x.mean(); } @@ -4738,7 +4909,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef poisson_distribution<_IntType> _Eng; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); double __mean; __is >> __mean; if (!__is.fail()) @@ -4784,9 +4956,17 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + weibull_distribution() : weibull_distribution(1) {} + _LIBCPP_INLINE_VISIBILITY + explicit weibull_distribution(result_type __a, result_type __b = 1) + : __p_(param_type(__a, __b)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit weibull_distribution(result_type __a = 1, result_type __b = 1) : __p_(param_type(__a, __b)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit weibull_distribution(const param_type& __p) : __p_(__p) {} @@ -4836,8 +5016,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const weibull_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.a() << __sp << __x.b(); @@ -4853,7 +5034,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __a; result_type __b; __is >> __a >> __b; @@ -4898,9 +5080,18 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + extreme_value_distribution() : extreme_value_distribution(0) {} _LIBCPP_INLINE_VISIBILITY - explicit extreme_value_distribution(result_type __a = 0, result_type __b = 1) + explicit extreme_value_distribution(result_type __a, result_type __b = 1) : __p_(param_type(__a, __b)) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit extreme_value_distribution(result_type __a = 0, + result_type __b = 1) + : __p_(param_type(__a, __b)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit extreme_value_distribution(const param_type& __p) : __p_(__p) {} @@ -4955,8 +5146,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const extreme_value_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.a() << __sp << __x.b(); @@ -4972,7 +5164,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __a; result_type __b; __is >> __a >> __b; @@ -5019,9 +5212,18 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + gamma_distribution() : gamma_distribution(1) {} _LIBCPP_INLINE_VISIBILITY - explicit gamma_distribution(result_type __alpha = 1, result_type __beta = 1) + explicit gamma_distribution(result_type __alpha, result_type __beta = 1) : __p_(param_type(__alpha, __beta)) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit gamma_distribution(result_type __alpha = 1, + result_type __beta = 1) + : __p_(param_type(__alpha, __beta)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit gamma_distribution(const param_type& __p) : __p_(__p) {} @@ -5127,8 +5329,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const gamma_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.alpha() << __sp << __x.beta(); @@ -5144,7 +5347,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __alpha; result_type __beta; __is >> __alpha >> __beta; @@ -5191,9 +5395,18 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + negative_binomial_distribution() : negative_binomial_distribution(1) {} _LIBCPP_INLINE_VISIBILITY - explicit negative_binomial_distribution(result_type __k = 1, double __p = 0.5) + explicit negative_binomial_distribution(result_type __k, double __p = 0.5) : __p_(__k, __p) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit negative_binomial_distribution(result_type __k = 1, + double __p = 0.5) + : __p_(__k, __p) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit negative_binomial_distribution(const param_type& __p) : __p_(__p) {} _LIBCPP_INLINE_VISIBILITY @@ -5263,8 +5476,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const negative_binomial_distribution<_IntType>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); return __os << __x.k() << __sp << __x.p(); @@ -5279,7 +5493,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __k; double __p; __is >> __k >> __p; @@ -5322,8 +5537,17 @@ private: public: // constructors and reset functions +#ifndef _LIBCPP_CXX03_LANG _LIBCPP_INLINE_VISIBILITY - explicit geometric_distribution(double __p = 0.5) : __p_(__p) {} + geometric_distribution() : geometric_distribution(0.5) {} + _LIBCPP_INLINE_VISIBILITY + explicit geometric_distribution(double __p) + : __p_(__p) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit geometric_distribution(double __p = 0.5) + : __p_(__p) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit geometric_distribution(const param_type& __p) : __p_(__p) {} _LIBCPP_INLINE_VISIBILITY @@ -5369,8 +5593,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const geometric_distribution<_IntType>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); return __os << __x.p(); } @@ -5382,7 +5607,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef geometric_distribution<_IntType> _Eng; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); double __p; __is >> __p; if (!__is.fail()) @@ -5424,9 +5650,17 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + chi_squared_distribution() : chi_squared_distribution(1) {} + _LIBCPP_INLINE_VISIBILITY + explicit chi_squared_distribution(result_type __n) + : __p_(param_type(__n)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit chi_squared_distribution(result_type __n = 1) : __p_(param_type(__n)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit chi_squared_distribution(const param_type& __p) : __p_(__p) {} @@ -5473,8 +5707,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const chi_squared_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); __os << __x.n(); return __os; } @@ -5488,7 +5723,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __n; __is >> __n; if (!__is.fail()) @@ -5534,9 +5770,17 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + cauchy_distribution() : cauchy_distribution(0) {} + _LIBCPP_INLINE_VISIBILITY + explicit cauchy_distribution(result_type __a, result_type __b = 1) + : __p_(param_type(__a, __b)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit cauchy_distribution(result_type __a = 0, result_type __b = 1) : __p_(param_type(__a, __b)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit cauchy_distribution(const param_type& __p) : __p_(__p) {} @@ -5593,8 +5837,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const cauchy_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.a() << __sp << __x.b(); @@ -5610,7 +5855,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __a; result_type __b; __is >> __a >> __b; @@ -5657,9 +5903,17 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + fisher_f_distribution() : fisher_f_distribution(1) {} + _LIBCPP_INLINE_VISIBILITY + explicit fisher_f_distribution(result_type __m, result_type __n = 1) + : __p_(param_type(__m, __n)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit fisher_f_distribution(result_type __m = 1, result_type __n = 1) : __p_(param_type(__m, __n)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit fisher_f_distribution(const param_type& __p) : __p_(__p) {} @@ -5715,8 +5969,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const fisher_f_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); __os << __x.m() << __sp << __x.n(); @@ -5732,7 +5987,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __m; result_type __n; __is >> __m >> __n; @@ -5776,9 +6032,17 @@ private: public: // constructor and reset functions +#ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + student_t_distribution() : student_t_distribution(1) {} + _LIBCPP_INLINE_VISIBILITY + explicit student_t_distribution(result_type __n) + : __p_(param_type(__n)) {} +#else _LIBCPP_INLINE_VISIBILITY explicit student_t_distribution(result_type __n = 1) : __p_(param_type(__n)) {} +#endif _LIBCPP_INLINE_VISIBILITY explicit student_t_distribution(const param_type& __p) : __p_(__p) {} @@ -5831,8 +6095,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const student_t_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); __os << __x.n(); return __os; } @@ -5846,7 +6111,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef typename _Eng::result_type result_type; typedef typename _Eng::param_type param_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); result_type __n; __is >> __n; if (!__is.fail()) @@ -6054,8 +6320,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const discrete_distribution<_IT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); size_t __n = __x.__p_.__p_.size(); @@ -6071,7 +6338,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, discrete_distribution<_IT>& __x) { __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); size_t __n; __is >> __n; vector<double> __p(__n); @@ -6356,8 +6624,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const piecewise_constant_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); size_t __n = __x.__p_.__b_.size(); @@ -6383,7 +6652,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef piecewise_constant_distribution<_RT> _Eng; typedef typename _Eng::result_type result_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); size_t __n; __is >> __n; vector<result_type> __b(__n); @@ -6696,8 +6966,9 @@ operator<<(basic_ostream<_CharT, _Traits>& __os, const piecewise_linear_distribution<_RT>& __x) { __save_flags<_CharT, _Traits> __lx(__os); - __os.flags(ios_base::dec | ios_base::left | ios_base::fixed | - ios_base::scientific); + typedef basic_ostream<_CharT, _Traits> _OStream; + __os.flags(_OStream::dec | _OStream::left | _OStream::fixed | + _OStream::scientific); _CharT __sp = __os.widen(' '); __os.fill(__sp); size_t __n = __x.__p_.__b_.size(); @@ -6723,7 +6994,8 @@ operator>>(basic_istream<_CharT, _Traits>& __is, typedef piecewise_linear_distribution<_RT> _Eng; typedef typename _Eng::result_type result_type; __save_flags<_CharT, _Traits> __lx(__is); - __is.flags(ios_base::dec | ios_base::skipws); + typedef basic_istream<_CharT, _Traits> _Istream; + __is.flags(_Istream::dec | _Istream::skipws); size_t __n; __is >> __n; vector<result_type> __b(__n); diff --git a/lib/libcxx/include/regex b/lib/libcxx/include/regex index f42f1ecd16..d78e4888a6 100644 --- a/lib/libcxx/include/regex +++ b/lib/libcxx/include/regex @@ -32,7 +32,8 @@ enum syntax_option_type extended = unspecified, awk = unspecified, grep = unspecified, - egrep = unspecified + egrep = unspecified, + multiline = unspecified }; constexpr syntax_option_type operator~(syntax_option_type f); @@ -142,6 +143,7 @@ public: static constexpr regex_constants::syntax_option_type awk = regex_constants::awk; static constexpr regex_constants::syntax_option_type grep = regex_constants::grep; static constexpr regex_constants::syntax_option_type egrep = regex_constants::egrep; + static constexpr regex_constants::syntax_option_type multiline = regex_constants::multiline; // construct/copy/destroy: basic_regex(); @@ -453,7 +455,9 @@ public: typedef basic_string<char_type> string_type; // construct/copy/destroy: - explicit match_results(const Allocator& a = Allocator()); + explicit match_results(const Allocator& a = Allocator()); // before C++20 + match_results() : match_results(Allocator()) {} // C++20 + explicit match_results(const Allocator& a); // C++20 match_results(const match_results& m); match_results(match_results&& m) noexcept; match_results& operator=(const match_results& m); @@ -802,7 +806,9 @@ enum syntax_option_type extended = 1 << 5, awk = 1 << 6, grep = 1 << 7, - egrep = 1 << 8 + egrep = 1 << 8, + // 1 << 9 may be used by ECMAScript + multiline = 1 << 10 }; inline _LIBCPP_CONSTEXPR @@ -1982,24 +1988,33 @@ __word_boundary<_CharT, _Traits>::__exec(__state& __s) const // __l_anchor template <class _CharT> -class __l_anchor +_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR +bool __is_eol(_CharT c) +{ + return c == '\r' || c == '\n'; +} + +template <class _CharT> +class __l_anchor_multiline : public __owns_one_state<_CharT> { typedef __owns_one_state<_CharT> base; + bool __multiline; + public: typedef _VSTD::__state<_CharT> __state; _LIBCPP_INLINE_VISIBILITY - __l_anchor(__node<_CharT>* __s) - : base(__s) {} + __l_anchor_multiline(bool __multiline, __node<_CharT>* __s) + : base(__s), __multiline(__multiline) {} virtual void __exec(__state&) const; }; template <class _CharT> void -__l_anchor<_CharT>::__exec(__state& __s) const +__l_anchor_multiline<_CharT>::__exec(__state& __s) const { if (__s.__at_first_ && __s.__current_ == __s.__first_ && !(__s.__flags_ & regex_constants::match_not_bol)) @@ -2007,6 +2022,13 @@ __l_anchor<_CharT>::__exec(__state& __s) const __s.__do_ = __state::__accept_but_not_consume; __s.__node_ = this->first(); } + else if (__multiline && + !__s.__at_first_ && + __is_eol(*_VSTD::prev(__s.__current_))) + { + __s.__do_ = __state::__accept_but_not_consume; + __s.__node_ = this->first(); + } else { __s.__do_ = __state::__reject; @@ -2017,24 +2039,26 @@ __l_anchor<_CharT>::__exec(__state& __s) const // __r_anchor template <class _CharT> -class __r_anchor +class __r_anchor_multiline : public __owns_one_state<_CharT> { typedef __owns_one_state<_CharT> base; + bool __multiline; + public: typedef _VSTD::__state<_CharT> __state; _LIBCPP_INLINE_VISIBILITY - __r_anchor(__node<_CharT>* __s) - : base(__s) {} + __r_anchor_multiline(bool __multiline, __node<_CharT>* __s) + : base(__s), __multiline(__multiline) {} virtual void __exec(__state&) const; }; template <class _CharT> void -__r_anchor<_CharT>::__exec(__state& __s) const +__r_anchor_multiline<_CharT>::__exec(__state& __s) const { if (__s.__current_ == __s.__last_ && !(__s.__flags_ & regex_constants::match_not_eol)) @@ -2042,6 +2066,11 @@ __r_anchor<_CharT>::__exec(__state& __s) const __s.__do_ = __state::__accept_but_not_consume; __s.__node_ = this->first(); } + else if (__multiline && __is_eol(*__s.__current_)) + { + __s.__do_ = __state::__accept_but_not_consume; + __s.__node_ = this->first(); + } else { __s.__do_ = __state::__reject; @@ -2448,7 +2477,7 @@ __bracket_expression<_CharT, _Traits>::__exec(__state& __s) const { const bool __in_neg_mask = __traits_.isctype(__ch, __neg_mask_); const bool __in_neg_chars = - std::find(__neg_chars_.begin(), __neg_chars_.end(), __ch) != + _VSTD::find(__neg_chars_.begin(), __neg_chars_.end(), __ch) != __neg_chars_.end(); if (!(__in_neg_mask || __in_neg_chars)) { @@ -2507,7 +2536,17 @@ __exit: template <class _CharT, class _Traits> class __lookahead; template <class _CharT, class _Traits = regex_traits<_CharT> > -class _LIBCPP_TEMPLATE_VIS basic_regex + class _LIBCPP_TEMPLATE_VIS basic_regex; + +typedef basic_regex<char> regex; +typedef basic_regex<wchar_t> wregex; + +template <class _CharT, class _Traits> +class + _LIBCPP_TEMPLATE_VIS + _LIBCPP_PREFERRED_NAME(regex) + _LIBCPP_PREFERRED_NAME(wregex) + basic_regex { public: // types: @@ -2541,17 +2580,18 @@ public: static const regex_constants::syntax_option_type awk = regex_constants::awk; static const regex_constants::syntax_option_type grep = regex_constants::grep; static const regex_constants::syntax_option_type egrep = regex_constants::egrep; + static const regex_constants::syntax_option_type multiline = regex_constants::multiline; // construct/copy/destroy: _LIBCPP_INLINE_VISIBILITY basic_regex() : __flags_(regex_constants::ECMAScript), __marked_count_(0), __loop_count_(0), __open_count_(0), - __end_(0) + __end_(nullptr) {} _LIBCPP_INLINE_VISIBILITY explicit basic_regex(const value_type* __p, flag_type __f = regex_constants::ECMAScript) : __flags_(__f), __marked_count_(0), __loop_count_(0), __open_count_(0), - __end_(0) + __end_(nullptr) { __init(__p, __p + __traits_.length(__p)); } @@ -2559,7 +2599,7 @@ public: _LIBCPP_INLINE_VISIBILITY basic_regex(const value_type* __p, size_t __len, flag_type __f = regex_constants::ECMAScript) : __flags_(__f), __marked_count_(0), __loop_count_(0), __open_count_(0), - __end_(0) + __end_(nullptr) { __init(__p, __p + __len); } @@ -2571,7 +2611,7 @@ public: explicit basic_regex(const basic_string<value_type, _ST, _SA>& __p, flag_type __f = regex_constants::ECMAScript) : __flags_(__f), __marked_count_(0), __loop_count_(0), __open_count_(0), - __end_(0) + __end_(nullptr) { __init(__p.begin(), __p.end()); } @@ -2581,7 +2621,7 @@ public: basic_regex(_ForwardIterator __first, _ForwardIterator __last, flag_type __f = regex_constants::ECMAScript) : __flags_(__f), __marked_count_(0), __loop_count_(0), __open_count_(0), - __end_(0) + __end_(nullptr) { __init(__first, __last); } @@ -2590,7 +2630,7 @@ public: basic_regex(initializer_list<value_type> __il, flag_type __f = regex_constants::ECMAScript) : __flags_(__f), __marked_count_(0), __loop_count_(0), __open_count_(0), - __end_(0) + __end_(nullptr) { __init(__il.begin(), __il.end()); } @@ -2707,6 +2747,12 @@ private: _LIBCPP_INLINE_VISIBILITY unsigned __loop_count() const {return __loop_count_;} + _LIBCPP_INLINE_VISIBILITY + bool __use_multiline() const + { + return __get_grammar(__flags_) == ECMAScript && (__flags_ & multiline); + } + template <class _ForwardIterator> void __init(_ForwardIterator __first, _ForwardIterator __last); @@ -4119,7 +4165,7 @@ basic_regex<_CharT, _Traits>::__parse_DUP_COUNT(_ForwardIterator __first, __first != __last && ( __val = __traits_.value(*__first, 10)) != -1; ++__first) { - if (__c >= std::numeric_limits<int>::max() / 10) + if (__c >= numeric_limits<int>::max() / 10) __throw_regex_error<regex_constants::error_badbrace>(); __c *= 10; __c += __val; @@ -4383,7 +4429,7 @@ basic_regex<_CharT, _Traits>::__parse_decimal_escape(_ForwardIterator __first, for (++__first; __first != __last && '0' <= *__first && *__first <= '9'; ++__first) { - if (__v >= std::numeric_limits<unsigned>::max() / 10) + if (__v >= numeric_limits<unsigned>::max() / 10) __throw_regex_error<regex_constants::error_backref>(); __v = 10 * __v + *__first - '0'; } @@ -4746,7 +4792,7 @@ template <class _CharT, class _Traits> void basic_regex<_CharT, _Traits>::__push_l_anchor() { - __end_->first() = new __l_anchor<_CharT>(__end_->first()); + __end_->first() = new __l_anchor_multiline<_CharT>(__use_multiline(), __end_->first()); __end_ = static_cast<__owns_one_state<_CharT>*>(__end_->first()); } @@ -4754,7 +4800,7 @@ template <class _CharT, class _Traits> void basic_regex<_CharT, _Traits>::__push_r_anchor() { - __end_->first() = new __r_anchor<_CharT>(__end_->first()); + __end_->first() = new __r_anchor_multiline<_CharT>(__use_multiline(), __end_->first()); __end_ = static_cast<__owns_one_state<_CharT>*>(__end_->first()); } @@ -4845,13 +4891,21 @@ basic_regex<_CharT, _Traits>::__push_lookahead(const basic_regex& __exp, __end_ = static_cast<__owns_one_state<_CharT>*>(__end_->first()); } -typedef basic_regex<char> regex; -typedef basic_regex<wchar_t> wregex; - // sub_match +typedef sub_match<const char*> csub_match; +typedef sub_match<const wchar_t*> wcsub_match; +typedef sub_match<string::const_iterator> ssub_match; +typedef sub_match<wstring::const_iterator> wssub_match; + template <class _BidirectionalIterator> -class _LIBCPP_TEMPLATE_VIS sub_match +class + _LIBCPP_TEMPLATE_VIS + _LIBCPP_PREFERRED_NAME(csub_match) + _LIBCPP_PREFERRED_NAME(wcsub_match) + _LIBCPP_PREFERRED_NAME(ssub_match) + _LIBCPP_PREFERRED_NAME(wssub_match) + sub_match : public pair<_BidirectionalIterator, _BidirectionalIterator> { public: @@ -4886,11 +4940,6 @@ public: {return str().compare(__s);} }; -typedef sub_match<const char*> csub_match; -typedef sub_match<const wchar_t*> wcsub_match; -typedef sub_match<string::const_iterator> ssub_match; -typedef sub_match<wstring::const_iterator> wssub_match; - template <class _BiIter> inline _LIBCPP_INLINE_VISIBILITY bool @@ -5273,8 +5322,19 @@ operator<<(basic_ostream<_CharT, _ST>& __os, const sub_match<_BiIter>& __m) return __os << __m.str(); } +typedef match_results<const char*> cmatch; +typedef match_results<const wchar_t*> wcmatch; +typedef match_results<string::const_iterator> smatch; +typedef match_results<wstring::const_iterator> wsmatch; + template <class _BidirectionalIterator, class _Allocator> -class _LIBCPP_TEMPLATE_VIS match_results +class + _LIBCPP_TEMPLATE_VIS + _LIBCPP_PREFERRED_NAME(cmatch) + _LIBCPP_PREFERRED_NAME(wcmatch) + _LIBCPP_PREFERRED_NAME(smatch) + _LIBCPP_PREFERRED_NAME(wsmatch) + match_results { public: typedef _Allocator allocator_type; @@ -5299,7 +5359,13 @@ public: typedef basic_string<char_type> string_type; // construct/copy/destroy: +#ifndef _LIBCPP_CXX03_LANG + match_results() : match_results(allocator_type()) {} + explicit match_results(const allocator_type& __a); +#else explicit match_results(const allocator_type& __a = allocator_type()); +#endif + // match_results(const match_results&) = default; // match_results& operator=(const match_results&) = default; // match_results(match_results&& __m) = default; @@ -5556,7 +5622,7 @@ match_results<_BidirectionalIterator, _Allocator>::format(_OutputIter __output_i '0' <= __fmt_first[1] && __fmt_first[1] <= '9') { ++__fmt_first; - if (__idx >= std::numeric_limits<size_t>::max() / 10) + if (__idx >= numeric_limits<size_t>::max() / 10) __throw_regex_error<regex_constants::error_escape>(); __idx = 10 * __idx + *__fmt_first - '0'; } @@ -5594,11 +5660,6 @@ match_results<_BidirectionalIterator, _Allocator>::swap(match_results& __m) swap(__ready_, __m.__ready_); } -typedef match_results<const char*> cmatch; -typedef match_results<const wchar_t*> wcmatch; -typedef match_results<string::const_iterator> smatch; -typedef match_results<wstring::const_iterator> wsmatch; - template <class _BidirectionalIterator, class _Allocator> bool operator==(const match_results<_BidirectionalIterator, _Allocator>& __x, @@ -6182,7 +6243,21 @@ regex_match(const basic_string<_CharT, _ST, _SA>& __s, template <class _BidirectionalIterator, class _CharT = typename iterator_traits<_BidirectionalIterator>::value_type, class _Traits = regex_traits<_CharT> > -class _LIBCPP_TEMPLATE_VIS regex_iterator + class _LIBCPP_TEMPLATE_VIS regex_iterator; + +typedef regex_iterator<const char*> cregex_iterator; +typedef regex_iterator<const wchar_t*> wcregex_iterator; +typedef regex_iterator<string::const_iterator> sregex_iterator; +typedef regex_iterator<wstring::const_iterator> wsregex_iterator; + +template <class _BidirectionalIterator, class _CharT, class _Traits> +class + _LIBCPP_TEMPLATE_VIS + _LIBCPP_PREFERRED_NAME(cregex_iterator) + _LIBCPP_PREFERRED_NAME(wcregex_iterator) + _LIBCPP_PREFERRED_NAME(sregex_iterator) + _LIBCPP_PREFERRED_NAME(wsregex_iterator) + regex_iterator { public: typedef basic_regex<_CharT, _Traits> regex_type; @@ -6291,17 +6366,26 @@ regex_iterator<_BidirectionalIterator, _CharT, _Traits>::operator++() return *this; } -typedef regex_iterator<const char*> cregex_iterator; -typedef regex_iterator<const wchar_t*> wcregex_iterator; -typedef regex_iterator<string::const_iterator> sregex_iterator; -typedef regex_iterator<wstring::const_iterator> wsregex_iterator; - // regex_token_iterator template <class _BidirectionalIterator, class _CharT = typename iterator_traits<_BidirectionalIterator>::value_type, class _Traits = regex_traits<_CharT> > -class _LIBCPP_TEMPLATE_VIS regex_token_iterator + class _LIBCPP_TEMPLATE_VIS regex_token_iterator; + +typedef regex_token_iterator<const char*> cregex_token_iterator; +typedef regex_token_iterator<const wchar_t*> wcregex_token_iterator; +typedef regex_token_iterator<string::const_iterator> sregex_token_iterator; +typedef regex_token_iterator<wstring::const_iterator> wsregex_token_iterator; + +template <class _BidirectionalIterator, class _CharT, class _Traits> +class + _LIBCPP_TEMPLATE_VIS + _LIBCPP_PREFERRED_NAME(cregex_token_iterator) + _LIBCPP_PREFERRED_NAME(wcregex_token_iterator) + _LIBCPP_PREFERRED_NAME(sregex_token_iterator) + _LIBCPP_PREFERRED_NAME(wsregex_token_iterator) + regex_token_iterator { public: typedef basic_regex<_CharT, _Traits> regex_type; @@ -6367,7 +6451,7 @@ public: regex_constants::match_flag_type __m = regex_constants::match_default); #if _LIBCPP_STD_VER > 11 - template <std::size_t _Np> + template <size_t _Np> regex_token_iterator(_BidirectionalIterator __a, _BidirectionalIterator __b, const regex_type&& __re, @@ -6579,11 +6663,6 @@ regex_token_iterator<_BidirectionalIterator, _CharT, _Traits>::operator++() return *this; } -typedef regex_token_iterator<const char*> cregex_token_iterator; -typedef regex_token_iterator<const wchar_t*> wcregex_token_iterator; -typedef regex_token_iterator<string::const_iterator> sregex_token_iterator; -typedef regex_token_iterator<wstring::const_iterator> wsregex_token_iterator; - // regex_replace template <class _OutputIterator, class _BidirectionalIterator, diff --git a/lib/libcxx/include/semaphore b/lib/libcxx/include/semaphore index 447bc2f385..0e0434f7fb 100644 --- a/lib/libcxx/include/semaphore +++ b/lib/libcxx/include/semaphore @@ -46,9 +46,9 @@ using binary_semaphore = counting_semaphore<1>; */ #include <__config> +#include <__availability> #include <__threading_support> #include <atomic> -#include <cassert> #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header @@ -58,6 +58,9 @@ using binary_semaphore = counting_semaphore<1>; # error <semaphore> is not supported on this single threaded system #endif +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + #if _LIBCPP_STD_VER >= 14 _LIBCPP_BEGIN_NAMESPACE_STD @@ -68,7 +71,7 @@ __atomic_semaphore_base is the general-case implementation, to be used for user-requested least-max values that exceed the OS implementation support (incl. when the OS has no support of its own) and for binary semaphores. -It is a typical Dijsktra semaphore algorithm over atomics, wait and notify +It is a typical Dijkstra semaphore algorithm over atomics, wait and notify functions. It avoids contention against users' own use of those facilities. */ @@ -232,4 +235,6 @@ _LIBCPP_END_NAMESPACE_STD #endif // _LIBCPP_STD_VER >= 14 +_LIBCPP_POP_MACROS + #endif //_LIBCPP_SEMAPHORE diff --git a/lib/libcxx/include/shared_mutex b/lib/libcxx/include/shared_mutex index fcafd8c0f4..5448d8a807 100644 --- a/lib/libcxx/include/shared_mutex +++ b/lib/libcxx/include/shared_mutex @@ -123,6 +123,7 @@ template <class Mutex> */ #include <__config> +#include <__availability> #include <version> _LIBCPP_PUSH_MACROS diff --git a/lib/libcxx/include/span b/lib/libcxx/include/span index b307c98aee..4f63d0ac4e 100644 --- a/lib/libcxx/include/span +++ b/lib/libcxx/include/span @@ -132,11 +132,14 @@ template<class Container> #pragma GCC system_header #endif +_LIBCPP_PUSH_MACROS +#include <__undef_macros> + _LIBCPP_BEGIN_NAMESPACE_STD #if _LIBCPP_STD_VER > 17 -inline constexpr size_t dynamic_extent = (numeric_limits<size_t>::max)(); +inline constexpr size_t dynamic_extent = numeric_limits<size_t>::max(); template <typename _Tp, size_t _Extent = dynamic_extent> class span; @@ -546,4 +549,6 @@ template<class _Container> _LIBCPP_END_NAMESPACE_STD +_LIBCPP_POP_MACROS + #endif // _LIBCPP_SPAN diff --git a/lib/libcxx/include/sstream b/lib/libcxx/include/sstream index 14c91971c2..7ce85be6ac 100644 --- a/lib/libcxx/include/sstream +++ b/lib/libcxx/include/sstream @@ -25,8 +25,10 @@ public: typedef typename traits_type::off_type off_type; typedef Allocator allocator_type; - // 27.8.1.1 Constructors: - explicit basic_stringbuf(ios_base::openmode which = ios_base::in | ios_base::out); + // 27.8.1.1 [stringbuf.cons], constructors: + explicit basic_stringbuf(ios_base::openmode which = ios_base::in | ios_base::out); // before C++20 + basic_stringbuf() : basic_stringbuf(ios_base::in | ios_base::out) {} // C++20 + explicit basic_stringbuf(ios_base::openmode which); // C++20 explicit basic_stringbuf(const basic_string<char_type, traits_type, allocator_type>& str, ios_base::openmode which = ios_base::in | ios_base::out); basic_stringbuf(basic_stringbuf&& rhs); @@ -71,7 +73,10 @@ public: typedef Allocator allocator_type; // 27.8.2.1 Constructors: - explicit basic_istringstream(ios_base::openmode which = ios_base::in); + explicit basic_istringstream(ios_base::openmode which = ios_base::in); // before C++20 + basic_istringstream() : basic_istringstream(ios_base::in) {} // C++20 + explicit basic_istringstream(ios_base::openmode which); // C++20 + explicit basic_istringstream(const basic_string<char_type, traits_type,allocator_type>& str, ios_base::openmode which = ios_base::in); basic_istringstream(basic_istringstream&& rhs); @@ -107,7 +112,10 @@ public: typedef Allocator allocator_type; // 27.8.3.1 Constructors/destructor: - explicit basic_ostringstream(ios_base::openmode which = ios_base::out); + explicit basic_ostringstream(ios_base::openmode which = ios_base::out); // before C++20 + basic_ostringstream() : basic_ostringstream(ios_base::out) {} // C++20 + explicit basic_ostringstream(ios_base::openmode which); // C++20 + explicit basic_ostringstream(const basic_string<char_type, traits_type, allocator_type>& str, ios_base::openmode which = ios_base::out); basic_ostringstream(basic_ostringstream&& rhs); @@ -143,7 +151,10 @@ public: typedef Allocator allocator_type; // constructors/destructor - explicit basic_stringstream(ios_base::openmode which = ios_base::out|ios_base::in); + explicit basic_stringstream(ios_base::openmode which = ios_base::out | ios_base::in); // before C++20 + basic_stringstream() : basic_stringstream(ios_base::out | ios_base::in) {} // C++20 + explicit basic_stringstream(ios_base::openmode which); // C++20 + explicit basic_stringstream(const basic_string<char_type, traits_type, allocator_type>& str, ios_base::openmode which = ios_base::out|ios_base::in); basic_stringstream(basic_stringstream&& rhs); @@ -207,18 +218,33 @@ private: ios_base::openmode __mode_; public: - // 27.8.1.1 Constructors: - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_stringbuf(ios_base::openmode __wch = ios_base::in | ios_base::out); - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_stringbuf(const string_type& __s, - ios_base::openmode __wch = ios_base::in | ios_base::out); + // 30.8.2.1 [stringbuf.cons], constructors #ifndef _LIBCPP_CXX03_LANG + _LIBCPP_INLINE_VISIBILITY + basic_stringbuf() : basic_stringbuf(ios_base::in | ios_base::out) {} + + _LIBCPP_INLINE_VISIBILITY + explicit basic_stringbuf(ios_base::openmode __wch) + : __hm_(nullptr), __mode_(__wch) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit basic_stringbuf(ios_base::openmode __wch = ios_base::in | + ios_base::out) + : __hm_(nullptr), __mode_(__wch) {} +#endif + + _LIBCPP_INLINE_VISIBILITY + explicit basic_stringbuf(const string_type& __s, + ios_base::openmode __wch = ios_base::in | ios_base::out) + : __str_(__s.get_allocator()), __hm_(nullptr), __mode_(__wch) + { + str(__s); + } + basic_stringbuf(basic_stringbuf&& __rhs); // 27.8.1.2 Assign and swap: basic_stringbuf& operator=(basic_stringbuf&& __rhs); -#endif void swap(basic_stringbuf& __rhs); // 27.8.1.3 Get and set: @@ -232,31 +258,14 @@ protected: virtual int_type overflow (int_type __c = traits_type::eof()); virtual pos_type seekoff(off_type __off, ios_base::seekdir __way, ios_base::openmode __wch = ios_base::in | ios_base::out); - inline _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY virtual pos_type seekpos(pos_type __sp, - ios_base::openmode __wch = ios_base::in | ios_base::out); + ios_base::openmode __wch = ios_base::in | ios_base::out) { + return seekoff(__sp, ios_base::beg, __wch); + } }; template <class _CharT, class _Traits, class _Allocator> -basic_stringbuf<_CharT, _Traits, _Allocator>::basic_stringbuf(ios_base::openmode __wch) - : __hm_(0), - __mode_(__wch) -{ -} - -template <class _CharT, class _Traits, class _Allocator> -basic_stringbuf<_CharT, _Traits, _Allocator>::basic_stringbuf(const string_type& __s, - ios_base::openmode __wch) - : __str_(__s.get_allocator()), - __hm_(0), - __mode_(__wch) -{ - str(__s); -} - -#ifndef _LIBCPP_CXX03_LANG - -template <class _CharT, class _Traits, class _Allocator> basic_stringbuf<_CharT, _Traits, _Allocator>::basic_stringbuf(basic_stringbuf&& __rhs) : __mode_(__rhs.__mode_) { @@ -345,8 +354,6 @@ basic_stringbuf<_CharT, _Traits, _Allocator>::operator=(basic_stringbuf&& __rhs) return *this; } -#endif // _LIBCPP_CXX03_LANG - template <class _CharT, class _Traits, class _Allocator> void basic_stringbuf<_CharT, _Traits, _Allocator>::swap(basic_stringbuf& __rhs) @@ -453,7 +460,7 @@ void basic_stringbuf<_CharT, _Traits, _Allocator>::str(const string_type& __s) { __str_ = __s; - __hm_ = 0; + __hm_ = nullptr; if (__mode_ & ios_base::in) { __hm_ = const_cast<char_type*>(__str_.data()) + __str_.size(); @@ -600,9 +607,9 @@ basic_stringbuf<_CharT, _Traits, _Allocator>::seekoff(off_type __off, return pos_type(-1); if (__noff != 0) { - if ((__wch & ios_base::in) && this->gptr() == 0) + if ((__wch & ios_base::in) && this->gptr() == nullptr) return pos_type(-1); - if ((__wch & ios_base::out) && this->pptr() == 0) + if ((__wch & ios_base::out) && this->pptr() == nullptr) return pos_type(-1); } if (__wch & ios_base::in) @@ -615,14 +622,6 @@ basic_stringbuf<_CharT, _Traits, _Allocator>::seekoff(off_type __off, return pos_type(__noff); } -template <class _CharT, class _Traits, class _Allocator> -typename basic_stringbuf<_CharT, _Traits, _Allocator>::pos_type -basic_stringbuf<_CharT, _Traits, _Allocator>::seekpos(pos_type __sp, - ios_base::openmode __wch) -{ - return seekoff(__sp, ios_base::beg, __wch); -} - // basic_istringstream template <class _CharT, class _Traits, class _Allocator> @@ -643,75 +642,63 @@ private: basic_stringbuf<char_type, traits_type, allocator_type> __sb_; public: - // 27.8.2.1 Constructors: - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_istringstream(ios_base::openmode __wch = ios_base::in); - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_istringstream(const string_type& __s, - ios_base::openmode __wch = ios_base::in); + // 30.8.3.1 [istringstream.cons], constructors #ifndef _LIBCPP_CXX03_LANG - inline _LIBCPP_INLINE_VISIBILITY - basic_istringstream(basic_istringstream&& __rhs); + _LIBCPP_INLINE_VISIBILITY + basic_istringstream() : basic_istringstream(ios_base::in) {} + + _LIBCPP_INLINE_VISIBILITY + explicit basic_istringstream(ios_base::openmode __wch) + : basic_istream<_CharT, _Traits>(&__sb_), __sb_(__wch | ios_base::in) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit basic_istringstream(ios_base::openmode __wch = ios_base::in) + : basic_istream<_CharT, _Traits>(&__sb_), __sb_(__wch | ios_base::in) {} +#endif + + _LIBCPP_INLINE_VISIBILITY + explicit basic_istringstream(const string_type& __s, + ios_base::openmode __wch = ios_base::in) + : basic_istream<_CharT, _Traits>(&__sb_) + , __sb_(__s, __wch | ios_base::in) + { } + + _LIBCPP_INLINE_VISIBILITY + basic_istringstream(basic_istringstream&& __rhs) + : basic_istream<_CharT, _Traits>(_VSTD::move(__rhs)) + , __sb_(_VSTD::move(__rhs.__sb_)) + { + basic_istream<_CharT, _Traits>::set_rdbuf(&__sb_); + } // 27.8.2.2 Assign and swap: - basic_istringstream& operator=(basic_istringstream&& __rhs); -#endif // _LIBCPP_CXX03_LANG - inline _LIBCPP_INLINE_VISIBILITY - void swap(basic_istringstream& __rhs); + basic_istringstream& operator=(basic_istringstream&& __rhs) { + basic_istream<char_type, traits_type>::operator=(_VSTD::move(__rhs)); + __sb_ = _VSTD::move(__rhs.__sb_); + return *this; + } + _LIBCPP_INLINE_VISIBILITY + void swap(basic_istringstream& __rhs) { + basic_istream<char_type, traits_type>::swap(__rhs); + __sb_.swap(__rhs.__sb_); + } // 27.8.2.3 Members: - inline _LIBCPP_INLINE_VISIBILITY - basic_stringbuf<char_type, traits_type, allocator_type>* rdbuf() const; - inline _LIBCPP_INLINE_VISIBILITY - string_type str() const; - inline _LIBCPP_INLINE_VISIBILITY - void str(const string_type& __s); + _LIBCPP_INLINE_VISIBILITY + basic_stringbuf<char_type, traits_type, allocator_type>* rdbuf() const { + return const_cast<basic_stringbuf<char_type, traits_type, allocator_type>*>(&__sb_); + } + _LIBCPP_INLINE_VISIBILITY + string_type str() const { + return __sb_.str(); + } + _LIBCPP_INLINE_VISIBILITY + void str(const string_type& __s) { + __sb_.str(__s); + } }; template <class _CharT, class _Traits, class _Allocator> -basic_istringstream<_CharT, _Traits, _Allocator>::basic_istringstream(ios_base::openmode __wch) - : basic_istream<_CharT, _Traits>(&__sb_), - __sb_(__wch | ios_base::in) -{ -} - -template <class _CharT, class _Traits, class _Allocator> -basic_istringstream<_CharT, _Traits, _Allocator>::basic_istringstream(const string_type& __s, - ios_base::openmode __wch) - : basic_istream<_CharT, _Traits>(&__sb_), - __sb_(__s, __wch | ios_base::in) -{ -} - -#ifndef _LIBCPP_CXX03_LANG - -template <class _CharT, class _Traits, class _Allocator> -basic_istringstream<_CharT, _Traits, _Allocator>::basic_istringstream(basic_istringstream&& __rhs) - : basic_istream<_CharT, _Traits>(_VSTD::move(__rhs)), - __sb_(_VSTD::move(__rhs.__sb_)) -{ - basic_istream<_CharT, _Traits>::set_rdbuf(&__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> -basic_istringstream<_CharT, _Traits, _Allocator>& -basic_istringstream<_CharT, _Traits, _Allocator>::operator=(basic_istringstream&& __rhs) -{ - basic_istream<char_type, traits_type>::operator=(_VSTD::move(__rhs)); - __sb_ = _VSTD::move(__rhs.__sb_); - return *this; -} - -#endif // _LIBCPP_CXX03_LANG - -template <class _CharT, class _Traits, class _Allocator> -void basic_istringstream<_CharT, _Traits, _Allocator>::swap(basic_istringstream& __rhs) -{ - basic_istream<char_type, traits_type>::swap(__rhs); - __sb_.swap(__rhs.__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY void swap(basic_istringstream<_CharT, _Traits, _Allocator>& __x, @@ -720,26 +707,6 @@ swap(basic_istringstream<_CharT, _Traits, _Allocator>& __x, __x.swap(__y); } -template <class _CharT, class _Traits, class _Allocator> -basic_stringbuf<_CharT, _Traits, _Allocator>* -basic_istringstream<_CharT, _Traits, _Allocator>::rdbuf() const -{ - return const_cast<basic_stringbuf<char_type, traits_type, allocator_type>*>(&__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> -basic_string<_CharT, _Traits, _Allocator> -basic_istringstream<_CharT, _Traits, _Allocator>::str() const -{ - return __sb_.str(); -} - -template <class _CharT, class _Traits, class _Allocator> -void basic_istringstream<_CharT, _Traits, _Allocator>::str(const string_type& __s) -{ - __sb_.str(__s); -} - // basic_ostringstream template <class _CharT, class _Traits, class _Allocator> @@ -760,76 +727,66 @@ private: basic_stringbuf<char_type, traits_type, allocator_type> __sb_; public: - // 27.8.2.1 Constructors: - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_ostringstream(ios_base::openmode __wch = ios_base::out); - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_ostringstream(const string_type& __s, - ios_base::openmode __wch = ios_base::out); + // 30.8.4.1 [ostringstream.cons], constructors #ifndef _LIBCPP_CXX03_LANG - inline _LIBCPP_INLINE_VISIBILITY - basic_ostringstream(basic_ostringstream&& __rhs); + _LIBCPP_INLINE_VISIBILITY + basic_ostringstream() : basic_ostringstream(ios_base::out) {} + + _LIBCPP_INLINE_VISIBILITY + explicit basic_ostringstream(ios_base::openmode __wch) + : basic_ostream<_CharT, _Traits>(&__sb_), + __sb_(__wch | ios_base::out) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit basic_ostringstream(ios_base::openmode __wch = ios_base::out) + : basic_ostream<_CharT, _Traits>(&__sb_), + __sb_(__wch | ios_base::out) {} +#endif + + _LIBCPP_INLINE_VISIBILITY + explicit basic_ostringstream(const string_type& __s, + ios_base::openmode __wch = ios_base::out) + : basic_ostream<_CharT, _Traits>(&__sb_) + , __sb_(__s, __wch | ios_base::out) + { } + + _LIBCPP_INLINE_VISIBILITY + basic_ostringstream(basic_ostringstream&& __rhs) + : basic_ostream<_CharT, _Traits>(_VSTD::move(__rhs)) + , __sb_(_VSTD::move(__rhs.__sb_)) + { + basic_ostream<_CharT, _Traits>::set_rdbuf(&__sb_); + } // 27.8.2.2 Assign and swap: - basic_ostringstream& operator=(basic_ostringstream&& __rhs); -#endif // _LIBCPP_CXX03_LANG - inline _LIBCPP_INLINE_VISIBILITY - void swap(basic_ostringstream& __rhs); + basic_ostringstream& operator=(basic_ostringstream&& __rhs) { + basic_ostream<char_type, traits_type>::operator=(_VSTD::move(__rhs)); + __sb_ = _VSTD::move(__rhs.__sb_); + return *this; + } + + _LIBCPP_INLINE_VISIBILITY + void swap(basic_ostringstream& __rhs) { + basic_ostream<char_type, traits_type>::swap(__rhs); + __sb_.swap(__rhs.__sb_); + } // 27.8.2.3 Members: - inline _LIBCPP_INLINE_VISIBILITY - basic_stringbuf<char_type, traits_type, allocator_type>* rdbuf() const; - inline _LIBCPP_INLINE_VISIBILITY - string_type str() const; - inline _LIBCPP_INLINE_VISIBILITY - void str(const string_type& __s); + _LIBCPP_INLINE_VISIBILITY + basic_stringbuf<char_type, traits_type, allocator_type>* rdbuf() const { + return const_cast<basic_stringbuf<char_type, traits_type, allocator_type>*>(&__sb_); + } + _LIBCPP_INLINE_VISIBILITY + string_type str() const { + return __sb_.str(); + } + _LIBCPP_INLINE_VISIBILITY + void str(const string_type& __s) { + __sb_.str(__s); + } }; template <class _CharT, class _Traits, class _Allocator> -basic_ostringstream<_CharT, _Traits, _Allocator>::basic_ostringstream(ios_base::openmode __wch) - : basic_ostream<_CharT, _Traits>(&__sb_), - __sb_(__wch | ios_base::out) -{ -} - -template <class _CharT, class _Traits, class _Allocator> -basic_ostringstream<_CharT, _Traits, _Allocator>::basic_ostringstream(const string_type& __s, - ios_base::openmode __wch) - : basic_ostream<_CharT, _Traits>(&__sb_), - __sb_(__s, __wch | ios_base::out) -{ -} - -#ifndef _LIBCPP_CXX03_LANG - -template <class _CharT, class _Traits, class _Allocator> -basic_ostringstream<_CharT, _Traits, _Allocator>::basic_ostringstream(basic_ostringstream&& __rhs) - : basic_ostream<_CharT, _Traits>(_VSTD::move(__rhs)), - __sb_(_VSTD::move(__rhs.__sb_)) -{ - basic_ostream<_CharT, _Traits>::set_rdbuf(&__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> -basic_ostringstream<_CharT, _Traits, _Allocator>& -basic_ostringstream<_CharT, _Traits, _Allocator>::operator=(basic_ostringstream&& __rhs) -{ - basic_ostream<char_type, traits_type>::operator=(_VSTD::move(__rhs)); - __sb_ = _VSTD::move(__rhs.__sb_); - return *this; -} - -#endif // _LIBCPP_CXX03_LANG - -template <class _CharT, class _Traits, class _Allocator> -void -basic_ostringstream<_CharT, _Traits, _Allocator>::swap(basic_ostringstream& __rhs) -{ - basic_ostream<char_type, traits_type>::swap(__rhs); - __sb_.swap(__rhs.__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY void swap(basic_ostringstream<_CharT, _Traits, _Allocator>& __x, @@ -838,27 +795,6 @@ swap(basic_ostringstream<_CharT, _Traits, _Allocator>& __x, __x.swap(__y); } -template <class _CharT, class _Traits, class _Allocator> -basic_stringbuf<_CharT, _Traits, _Allocator>* -basic_ostringstream<_CharT, _Traits, _Allocator>::rdbuf() const -{ - return const_cast<basic_stringbuf<char_type, traits_type, allocator_type>*>(&__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> -basic_string<_CharT, _Traits, _Allocator> -basic_ostringstream<_CharT, _Traits, _Allocator>::str() const -{ - return __sb_.str(); -} - -template <class _CharT, class _Traits, class _Allocator> -void -basic_ostringstream<_CharT, _Traits, _Allocator>::str(const string_type& __s) -{ - __sb_.str(__s); -} - // basic_stringstream template <class _CharT, class _Traits, class _Allocator> @@ -879,76 +815,64 @@ private: basic_stringbuf<char_type, traits_type, allocator_type> __sb_; public: - // 27.8.2.1 Constructors: - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_stringstream(ios_base::openmode __wch = ios_base::in | ios_base::out); - inline _LIBCPP_INLINE_VISIBILITY - explicit basic_stringstream(const string_type& __s, - ios_base::openmode __wch = ios_base::in | ios_base::out); + // 30.8.5.1 [stringstream.cons], constructors #ifndef _LIBCPP_CXX03_LANG - inline _LIBCPP_INLINE_VISIBILITY - basic_stringstream(basic_stringstream&& __rhs); + _LIBCPP_INLINE_VISIBILITY + basic_stringstream() : basic_stringstream(ios_base::in | ios_base::out) {} + + _LIBCPP_INLINE_VISIBILITY + explicit basic_stringstream(ios_base::openmode __wch) + : basic_iostream<_CharT, _Traits>(&__sb_), __sb_(__wch) {} +#else + _LIBCPP_INLINE_VISIBILITY + explicit basic_stringstream(ios_base::openmode __wch = ios_base::in | + ios_base::out) + : basic_iostream<_CharT, _Traits>(&__sb_), __sb_(__wch) {} +#endif + + _LIBCPP_INLINE_VISIBILITY + explicit basic_stringstream(const string_type& __s, + ios_base::openmode __wch = ios_base::in | ios_base::out) + : basic_iostream<_CharT, _Traits>(&__sb_) + , __sb_(__s, __wch) + { } + + _LIBCPP_INLINE_VISIBILITY + basic_stringstream(basic_stringstream&& __rhs) + : basic_iostream<_CharT, _Traits>(_VSTD::move(__rhs)) + , __sb_(_VSTD::move(__rhs.__sb_)) + { + basic_istream<_CharT, _Traits>::set_rdbuf(&__sb_); + } // 27.8.2.2 Assign and swap: - basic_stringstream& operator=(basic_stringstream&& __rhs); -#endif // _LIBCPP_CXX03_LANG - inline _LIBCPP_INLINE_VISIBILITY - void swap(basic_stringstream& __rhs); + basic_stringstream& operator=(basic_stringstream&& __rhs) { + basic_iostream<char_type, traits_type>::operator=(_VSTD::move(__rhs)); + __sb_ = _VSTD::move(__rhs.__sb_); + return *this; + } + _LIBCPP_INLINE_VISIBILITY + void swap(basic_stringstream& __rhs) { + basic_iostream<char_type, traits_type>::swap(__rhs); + __sb_.swap(__rhs.__sb_); + } // 27.8.2.3 Members: - inline _LIBCPP_INLINE_VISIBILITY - basic_stringbuf<char_type, traits_type, allocator_type>* rdbuf() const; - inline _LIBCPP_INLINE_VISIBILITY - string_type str() const; - inline _LIBCPP_INLINE_VISIBILITY - void str(const string_type& __s); + _LIBCPP_INLINE_VISIBILITY + basic_stringbuf<char_type, traits_type, allocator_type>* rdbuf() const { + return const_cast<basic_stringbuf<char_type, traits_type, allocator_type>*>(&__sb_); + } + _LIBCPP_INLINE_VISIBILITY + string_type str() const { + return __sb_.str(); + } + _LIBCPP_INLINE_VISIBILITY + void str(const string_type& __s) { + __sb_.str(__s); + } }; template <class _CharT, class _Traits, class _Allocator> -basic_stringstream<_CharT, _Traits, _Allocator>::basic_stringstream(ios_base::openmode __wch) - : basic_iostream<_CharT, _Traits>(&__sb_), - __sb_(__wch) -{ -} - -template <class _CharT, class _Traits, class _Allocator> -basic_stringstream<_CharT, _Traits, _Allocator>::basic_stringstream(const string_type& __s, - ios_base::openmode __wch) - : basic_iostream<_CharT, _Traits>(&__sb_), - __sb_(__s, __wch) -{ -} - -#ifndef _LIBCPP_CXX03_LANG - -template <class _CharT, class _Traits, class _Allocator> -basic_stringstream<_CharT, _Traits, _Allocator>::basic_stringstream(basic_stringstream&& __rhs) - : basic_iostream<_CharT, _Traits>(_VSTD::move(__rhs)), - __sb_(_VSTD::move(__rhs.__sb_)) -{ - basic_istream<_CharT, _Traits>::set_rdbuf(&__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> -basic_stringstream<_CharT, _Traits, _Allocator>& -basic_stringstream<_CharT, _Traits, _Allocator>::operator=(basic_stringstream&& __rhs) -{ - basic_iostream<char_type, traits_type>::operator=(_VSTD::move(__rhs)); - __sb_ = _VSTD::move(__rhs.__sb_); - return *this; -} - -#endif // _LIBCPP_CXX03_LANG - -template <class _CharT, class _Traits, class _Allocator> -void -basic_stringstream<_CharT, _Traits, _Allocator>::swap(basic_stringstream& __rhs) -{ - basic_iostream<char_type, traits_type>::swap(__rhs); - __sb_.swap(__rhs.__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY void swap(basic_stringstream<_CharT, _Traits, _Allocator>& __x, @@ -957,26 +881,12 @@ swap(basic_stringstream<_CharT, _Traits, _Allocator>& __x, __x.swap(__y); } -template <class _CharT, class _Traits, class _Allocator> -basic_stringbuf<_CharT, _Traits, _Allocator>* -basic_stringstream<_CharT, _Traits, _Allocator>::rdbuf() const -{ - return const_cast<basic_stringbuf<char_type, traits_type, allocator_type>*>(&__sb_); -} - -template <class _CharT, class _Traits, class _Allocator> -basic_string<_CharT, _Traits, _Allocator> -basic_stringstream<_CharT, _Traits, _Allocator>::str() const -{ - return __sb_.str(); -} - -template <class _CharT, class _Traits, class _Allocator> -void -basic_stringstream<_CharT, _Traits, _Allocator>::str(const string_type& __s) -{ - __sb_.str(__s); -} +#if defined(_LIBCPP_ABI_ENABLE_ADDITIONAL_IOSTREAM_EXPLICIT_INSTANTIATIONS_1) +_LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_stringbuf<char>) +_LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_stringstream<char>) +_LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_ostringstream<char>) +_LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_istringstream<char>) +#endif _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/include/stdexcept b/lib/libcxx/include/stdexcept index 7a7f367940..7b5de7ea36 100644 --- a/lib/libcxx/include/stdexcept +++ b/lib/libcxx/include/stdexcept @@ -42,11 +42,9 @@ public: */ #include <__config> +#include <cstdlib> #include <exception> #include <iosfwd> // for string forward decl -#ifdef _LIBCPP_NO_EXCEPTIONS -#include <cstdlib> -#endif #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) #pragma GCC system_header diff --git a/lib/libcxx/include/streambuf b/lib/libcxx/include/streambuf index 48c07d5e1f..0dfa31449b 100644 --- a/lib/libcxx/include/streambuf +++ b/lib/libcxx/include/streambuf @@ -308,12 +308,12 @@ basic_streambuf<_CharT, _Traits>::~basic_streambuf() template <class _CharT, class _Traits> basic_streambuf<_CharT, _Traits>::basic_streambuf() - : __binp_(0), - __ninp_(0), - __einp_(0), - __bout_(0), - __nout_(0), - __eout_(0) + : __binp_(nullptr), + __ninp_(nullptr), + __einp_(nullptr), + __bout_(nullptr), + __nout_(nullptr), + __eout_(nullptr) { } @@ -485,13 +485,11 @@ basic_streambuf<_CharT, _Traits>::overflow(int_type) return traits_type::eof(); } -#ifndef _LIBCPP_DO_NOT_ASSUME_STREAMS_EXPLICIT_INSTANTIATION_IN_DYLIB _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_streambuf<char>) _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_streambuf<wchar_t>) _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_ios<char>) _LIBCPP_EXTERN_TEMPLATE(class _LIBCPP_EXTERN_TEMPLATE_TYPE_VIS basic_ios<wchar_t>) -#endif _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/include/string b/lib/libcxx/include/string index 2f846eda06..687795c79b 100644 --- a/lib/libcxx/include/string +++ b/lib/libcxx/include/string @@ -153,7 +153,8 @@ public: void resize(size_type n, value_type c); void resize(size_type n); - void reserve(size_type res_arg = 0); + void reserve(size_type res_arg); + void reserve(); // deprecated in C++20 void shrink_to_fit(); void clear() noexcept; bool empty() const noexcept; @@ -316,12 +317,16 @@ public: int compare(size_type pos1, size_type n1, const value_type* s) const; int compare(size_type pos1, size_type n1, const value_type* s, size_type n2) const; - bool starts_with(basic_string_view<charT, traits> sv) const noexcept; // C++2a - bool starts_with(charT c) const noexcept; // C++2a - bool starts_with(const charT* s) const; // C++2a - bool ends_with(basic_string_view<charT, traits> sv) const noexcept; // C++2a - bool ends_with(charT c) const noexcept; // C++2a - bool ends_with(const charT* s) const; // C++2a + bool starts_with(basic_string_view<charT, traits> sv) const noexcept; // C++20 + bool starts_with(charT c) const noexcept; // C++20 + bool starts_with(const charT* s) const; // C++20 + bool ends_with(basic_string_view<charT, traits> sv) const noexcept; // C++20 + bool ends_with(charT c) const noexcept; // C++20 + bool ends_with(const charT* s) const; // C++20 + + constexpr bool contains(basic_string_view<charT, traits> sv) const noexcept; // C++2b + constexpr bool contains(charT c) const noexcept; // C++2b + constexpr bool contains(const charT* s) const; // C++2b bool __invariants() const; }; @@ -448,15 +453,15 @@ typedef basic_string<wchar_t> wstring; typedef basic_string<char16_t> u16string; typedef basic_string<char32_t> u32string; -int stoi (const string& str, size_t* idx = 0, int base = 10); -long stol (const string& str, size_t* idx = 0, int base = 10); -unsigned long stoul (const string& str, size_t* idx = 0, int base = 10); -long long stoll (const string& str, size_t* idx = 0, int base = 10); -unsigned long long stoull(const string& str, size_t* idx = 0, int base = 10); +int stoi (const string& str, size_t* idx = nullptr, int base = 10); +long stol (const string& str, size_t* idx = nullptr, int base = 10); +unsigned long stoul (const string& str, size_t* idx = nullptr, int base = 10); +long long stoll (const string& str, size_t* idx = nullptr, int base = 10); +unsigned long long stoull(const string& str, size_t* idx = nullptr, int base = 10); -float stof (const string& str, size_t* idx = 0); -double stod (const string& str, size_t* idx = 0); -long double stold(const string& str, size_t* idx = 0); +float stof (const string& str, size_t* idx = nullptr); +double stod (const string& str, size_t* idx = nullptr); +long double stold(const string& str, size_t* idx = nullptr); string to_string(int val); string to_string(unsigned val); @@ -468,15 +473,15 @@ string to_string(float val); string to_string(double val); string to_string(long double val); -int stoi (const wstring& str, size_t* idx = 0, int base = 10); -long stol (const wstring& str, size_t* idx = 0, int base = 10); -unsigned long stoul (const wstring& str, size_t* idx = 0, int base = 10); -long long stoll (const wstring& str, size_t* idx = 0, int base = 10); -unsigned long long stoull(const wstring& str, size_t* idx = 0, int base = 10); +int stoi (const wstring& str, size_t* idx = nullptr, int base = 10); +long stol (const wstring& str, size_t* idx = nullptr, int base = 10); +unsigned long stoul (const wstring& str, size_t* idx = nullptr, int base = 10); +long long stoll (const wstring& str, size_t* idx = nullptr, int base = 10); +unsigned long long stoull(const wstring& str, size_t* idx = nullptr, int base = 10); -float stof (const wstring& str, size_t* idx = 0); -double stod (const wstring& str, size_t* idx = 0); -long double stold(const wstring& str, size_t* idx = 0); +float stof (const wstring& str, size_t* idx = nullptr); +double stod (const wstring& str, size_t* idx = nullptr); +long double stold(const wstring& str, size_t* idx = nullptr); wstring to_wstring(int val); wstring to_wstring(unsigned val); @@ -665,8 +670,26 @@ struct __padding<_CharT, 1> #endif // _LIBCPP_ABI_ALTERNATE_STRING_LAYOUT +#ifndef _LIBCPP_NO_HAS_CHAR8_T +typedef basic_string<char8_t> u8string; +#endif + +#ifndef _LIBCPP_HAS_NO_UNICODE_CHARS +typedef basic_string<char16_t> u16string; +typedef basic_string<char32_t> u32string; +#endif // _LIBCPP_HAS_NO_UNICODE_CHARS + template<class _CharT, class _Traits, class _Allocator> -class _LIBCPP_TEMPLATE_VIS basic_string +class + _LIBCPP_TEMPLATE_VIS +#ifndef _LIBCPP_NO_HAS_CHAR8_T + _LIBCPP_PREFERRED_NAME(u8string) +#endif +#ifndef _LIBCPP_HAS_NO_UNICODE_CHARS + _LIBCPP_PREFERRED_NAME(u16string) + _LIBCPP_PREFERRED_NAME(u32string) +#endif + basic_string : private __basic_string_common<true> { public: @@ -816,7 +839,7 @@ public: basic_string(const _CharT* __s) : __r_(__default_init_tag(), __default_init_tag()) { _LIBCPP_ASSERT(__s != nullptr, "basic_string(const char*) detected nullptr"); __init(__s, traits_type::length(__s)); -# if _LIBCPP_DEBUG_LEVEL >= 2 +# if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); # endif } @@ -890,7 +913,7 @@ public: _LIBCPP_INLINE_VISIBILITY basic_string& operator=(const value_type* __s) {return assign(__s);} basic_string& operator=(value_type __c); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY iterator begin() _NOEXCEPT {return iterator(this, __get_pointer());} @@ -916,7 +939,7 @@ public: _LIBCPP_INLINE_VISIBILITY const_iterator end() const _NOEXCEPT {return const_iterator(__get_pointer() + size());} -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_INLINE_VISIBILITY reverse_iterator rbegin() _NOEXCEPT {return reverse_iterator(end());} @@ -954,13 +977,13 @@ public: void resize(size_type __n, value_type __c); _LIBCPP_INLINE_VISIBILITY void resize(size_type __n) {resize(__n, value_type());} - void reserve(size_type __res_arg); + void reserve(size_type __requested_capacity); _LIBCPP_INLINE_VISIBILITY void __resize_default_init(size_type __n); + _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_INLINE_VISIBILITY + void reserve() _NOEXCEPT {shrink_to_fit();} _LIBCPP_INLINE_VISIBILITY - void reserve() _NOEXCEPT {reserve(0);} - _LIBCPP_INLINE_VISIBILITY - void shrink_to_fit() _NOEXCEPT {reserve();} + void shrink_to_fit() _NOEXCEPT; _LIBCPP_INLINE_VISIBILITY void clear() _NOEXCEPT; _LIBCPP_NODISCARD_AFTER_CXX17 _LIBCPP_INLINE_VISIBILITY @@ -1414,22 +1437,38 @@ public: { return ends_with(__self_view(__s)); } #endif +#if _LIBCPP_STD_VER > 20 + constexpr _LIBCPP_INLINE_VISIBILITY + bool contains(__self_view __sv) const noexcept + { return __self_view(data(), size()).contains(__sv); } + + constexpr _LIBCPP_INLINE_VISIBILITY + bool contains(value_type __c) const noexcept + { return __self_view(data(), size()).contains(__c); } + + constexpr _LIBCPP_INLINE_VISIBILITY + bool contains(const value_type* __s) const + { return __self_view(data(), size()).contains(__s); } +#endif + _LIBCPP_INLINE_VISIBILITY bool __invariants() const; _LIBCPP_INLINE_VISIBILITY void __clear_and_shrink() _NOEXCEPT; + _LIBCPP_INLINE_VISIBILITY void __shrink_or_extend(size_type __target_capacity); + _LIBCPP_INLINE_VISIBILITY bool __is_long() const _NOEXCEPT {return bool(__r_.first().__s.__size_ & __short_mask);} -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const; bool __decrementable(const const_iterator* __i) const; bool __addable(const const_iterator* __i, ptrdiff_t __n) const; bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const; -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 private: _LIBCPP_INLINE_VISIBILITY @@ -1726,21 +1765,21 @@ inline void basic_string<_CharT, _Traits, _Allocator>::__invalidate_all_iterators() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__invalidate_all(this); -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif } template <class _CharT, class _Traits, class _Allocator> inline void basic_string<_CharT, _Traits, _Allocator>::__invalidate_iterators_past(size_type -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __pos #endif ) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); if (__c) { @@ -1753,12 +1792,12 @@ basic_string<_CharT, _Traits, _Allocator>::__invalidate_iterators_past(size_type { (*__p)->__c_ = nullptr; if (--__c->end_ != __p) - memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); } } __get_db()->unlock(); } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 } template <class _CharT, class _Traits, class _Allocator> @@ -1767,7 +1806,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string() _NOEXCEPT_(is_nothrow_default_constructible<allocator_type>::value) : __r_(__default_init_tag(), __default_init_tag()) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __zero(); @@ -1783,7 +1822,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const allocator_type& __ #endif : __r_(__default_init_tag(), __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __zero(); @@ -1845,7 +1884,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const _CharT* __s, const { _LIBCPP_ASSERT(__s != nullptr, "basic_string(const char*, allocator) detected nullptr"); __init(__s, traits_type::length(__s)); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1857,7 +1896,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const _CharT* __s, size_ { _LIBCPP_ASSERT(__n == 0 || __s != nullptr, "basic_string(const char*, n) detected nullptr"); __init(__s, __n); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1869,7 +1908,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const _CharT* __s, size_ { _LIBCPP_ASSERT(__n == 0 || __s != nullptr, "basic_string(const char*, n, allocator) detected nullptr"); __init(__s, __n); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1884,7 +1923,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const basic_string& __st __init_copy_ctor_external(_VSTD::__to_address(__str.__get_long_pointer()), __str.__get_long_size()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1899,7 +1938,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string( else __init_copy_ctor_external(_VSTD::__to_address(__str.__get_long_pointer()), __str.__get_long_size()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1936,7 +1975,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(basic_string&& __str) : __r_(_VSTD::move(__str.__r_)) { __str.__zero(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); if (__is_long()) __get_db()->swap(this, &__str); @@ -1955,7 +1994,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(basic_string&& __str, co __r_.first().__r = __str.__r_.first().__r; __str.__zero(); } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); if (__is_long()) __get_db()->swap(this, &__str); @@ -1994,7 +2033,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(size_type __n, _CharT __ : __r_(__default_init_tag(), __default_init_tag()) { __init(__n, __c); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2005,7 +2044,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(size_type __n, _CharT __ : __r_(__default_init_tag(), __a) { __init(__n, __c); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2020,7 +2059,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const basic_string& __st if (__pos > __str_sz) this->__throw_out_of_range(); __init(__str.data() + __pos, _VSTD::min(__n, __str_sz - __pos)); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2035,7 +2074,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const basic_string& __st if (__pos > __str_sz) this->__throw_out_of_range(); __init(__str.data() + __pos, __str_sz - __pos); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2049,7 +2088,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string( __self_view __sv0 = __t; __self_view __sv = __sv0.substr(__pos, __n); __init(__sv.data(), __sv.size()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2061,7 +2100,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const _Tp & __t) { __self_view __sv = __t; __init(__sv.data(), __sv.size()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2073,7 +2112,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(const _Tp & __t, const _ { __self_view __sv = __t; __init(__sv.data(), __sv.size()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2141,7 +2180,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(_InputIterator __first, : __r_(__default_init_tag(), __default_init_tag()) { __init(__first, __last); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2154,7 +2193,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string(_InputIterator __first, : __r_(__default_init_tag(), __a) { __init(__first, __last); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2168,7 +2207,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string( : __r_(__default_init_tag(), __default_init_tag()) { __init(__il.begin(), __il.end()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2181,7 +2220,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string( : __r_(__default_init_tag(), __a) { __init(__il.begin(), __il.end()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2191,7 +2230,7 @@ basic_string<_CharT, _Traits, _Allocator>::basic_string( template <class _CharT, class _Traits, class _Allocator> basic_string<_CharT, _Traits, _Allocator>::~basic_string() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__erase_c(this); #endif if (__is_long()) @@ -2768,7 +2807,7 @@ _EnableIf > basic_string<_CharT, _Traits, _Allocator>::insert(const_iterator __pos, _InputIterator __first, _InputIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__pos) == this, "string::insert(iterator, range) called with an iterator not" " referring to this string"); @@ -2787,7 +2826,7 @@ _EnableIf > basic_string<_CharT, _Traits, _Allocator>::insert(const_iterator __pos, _ForwardIterator __first, _ForwardIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__pos) == this, "string::insert(iterator, range) called with an iterator not" " referring to this string"); @@ -2903,7 +2942,7 @@ inline typename basic_string<_CharT, _Traits, _Allocator>::iterator basic_string<_CharT, _Traits, _Allocator>::insert(const_iterator __pos, size_type __n, value_type __c) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__pos) == this, "string::insert(iterator, n, value) called with an iterator not" " referring to this string"); @@ -3137,7 +3176,7 @@ inline typename basic_string<_CharT, _Traits, _Allocator>::iterator basic_string<_CharT, _Traits, _Allocator>::erase(const_iterator __pos) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__pos) == this, "string::erase(iterator) called with an iterator not" " referring to this string"); @@ -3155,7 +3194,7 @@ inline typename basic_string<_CharT, _Traits, _Allocator>::iterator basic_string<_CharT, _Traits, _Allocator>::erase(const_iterator __first, const_iterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__first) == this, "string::erase(iterator, iterator) called with an iterator not" " referring to this string"); @@ -3262,65 +3301,88 @@ basic_string<_CharT, _Traits, _Allocator>::max_size() const _NOEXCEPT template <class _CharT, class _Traits, class _Allocator> void -basic_string<_CharT, _Traits, _Allocator>::reserve(size_type __res_arg) +basic_string<_CharT, _Traits, _Allocator>::reserve(size_type __requested_capacity) { - if (__res_arg > max_size()) + if (__requested_capacity > max_size()) this->__throw_length_error(); + +#if _LIBCPP_STD_VER > 17 + // Reserve never shrinks as of C++20. + if (__requested_capacity <= capacity()) return; +#endif + + size_type __target_capacity = _VSTD::max(__requested_capacity, size()); + __target_capacity = __recommend(__target_capacity); + if (__target_capacity == capacity()) return; + + __shrink_or_extend(__target_capacity); +} + +template <class _CharT, class _Traits, class _Allocator> +void +basic_string<_CharT, _Traits, _Allocator>::shrink_to_fit() _NOEXCEPT +{ + size_type __target_capacity = __recommend(size()); + if (__target_capacity == capacity()) return; + + __shrink_or_extend(__target_capacity); +} + +template <class _CharT, class _Traits, class _Allocator> +void +basic_string<_CharT, _Traits, _Allocator>::__shrink_or_extend(size_type __target_capacity) +{ size_type __cap = capacity(); size_type __sz = size(); - __res_arg = _VSTD::max(__res_arg, __sz); - __res_arg = __recommend(__res_arg); - if (__res_arg != __cap) + + pointer __new_data, __p; + bool __was_long, __now_long; + if (__target_capacity == __min_cap - 1) { - pointer __new_data, __p; - bool __was_long, __now_long; - if (__res_arg == __min_cap - 1) - { - __was_long = true; - __now_long = false; - __new_data = __get_short_pointer(); - __p = __get_long_pointer(); - } + __was_long = true; + __now_long = false; + __new_data = __get_short_pointer(); + __p = __get_long_pointer(); + } + else + { + if (__target_capacity > __cap) + __new_data = __alloc_traits::allocate(__alloc(), __target_capacity+1); else { - if (__res_arg > __cap) - __new_data = __alloc_traits::allocate(__alloc(), __res_arg+1); - else + #ifndef _LIBCPP_NO_EXCEPTIONS + try { - #ifndef _LIBCPP_NO_EXCEPTIONS - try - { - #endif // _LIBCPP_NO_EXCEPTIONS - __new_data = __alloc_traits::allocate(__alloc(), __res_arg+1); - #ifndef _LIBCPP_NO_EXCEPTIONS - } - catch (...) - { - return; - } - #else // _LIBCPP_NO_EXCEPTIONS - if (__new_data == nullptr) - return; - #endif // _LIBCPP_NO_EXCEPTIONS + #endif // _LIBCPP_NO_EXCEPTIONS + __new_data = __alloc_traits::allocate(__alloc(), __target_capacity+1); + #ifndef _LIBCPP_NO_EXCEPTIONS } - __now_long = true; - __was_long = __is_long(); - __p = __get_pointer(); - } - traits_type::copy(_VSTD::__to_address(__new_data), - _VSTD::__to_address(__p), size()+1); - if (__was_long) - __alloc_traits::deallocate(__alloc(), __p, __cap+1); - if (__now_long) - { - __set_long_cap(__res_arg+1); - __set_long_size(__sz); - __set_long_pointer(__new_data); + catch (...) + { + return; + } + #else // _LIBCPP_NO_EXCEPTIONS + if (__new_data == nullptr) + return; + #endif // _LIBCPP_NO_EXCEPTIONS } - else - __set_short_size(__sz); - __invalidate_all_iterators(); + __now_long = true; + __was_long = __is_long(); + __p = __get_pointer(); } + traits_type::copy(_VSTD::__to_address(__new_data), + _VSTD::__to_address(__p), size()+1); + if (__was_long) + __alloc_traits::deallocate(__alloc(), __p, __cap+1); + if (__now_long) + { + __set_long_cap(__target_capacity+1); + __set_long_size(__sz); + __set_long_pointer(__new_data); + } + else + __set_short_size(__sz); + __invalidate_all_iterators(); } template <class _CharT, class _Traits, class _Allocator> @@ -3426,7 +3488,7 @@ basic_string<_CharT, _Traits, _Allocator>::swap(basic_string& __str) __is_nothrow_swappable<allocator_type>::value) #endif { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 if (!__is_long()) __get_db()->__invalidate_all(this); if (!__str.__is_long()) @@ -3438,7 +3500,7 @@ basic_string<_CharT, _Traits, _Allocator>::swap(basic_string& __str) __alloc_traits::is_always_equal::value || __alloc() == __str.__alloc(), "swapping non-equal allocators"); _VSTD::swap(__r_.first(), __str.__r_.first()); - __swap_allocator(__alloc(), __str.__alloc()); + _VSTD::__swap_allocator(__alloc(), __str.__alloc()); } // find @@ -3939,9 +4001,9 @@ basic_string<_CharT, _Traits, _Allocator>::__invariants() const return false; if (capacity() < __min_cap - 1) return false; - if (data() == 0) + if (data() == nullptr) return false; - if (data()[size()] != value_type(0)) + if (data()[size()] != value_type()) return false; return true; } @@ -3959,6 +4021,7 @@ basic_string<_CharT, _Traits, _Allocator>::__clear_and_shrink() _NOEXCEPT __alloc_traits::deallocate(__alloc(), __get_long_pointer(), capacity() + 1); __set_long_cap(0); __set_short_size(0); + traits_type::assign(*__get_short_pointer(), value_type()); } } @@ -4300,24 +4363,15 @@ swap(basic_string<_CharT, _Traits, _Allocator>& __lhs, __lhs.swap(__rhs); } -#ifndef _LIBCPP_NO_HAS_CHAR8_T -typedef basic_string<char8_t> u8string; -#endif - -#ifndef _LIBCPP_HAS_NO_UNICODE_CHARS -typedef basic_string<char16_t> u16string; -typedef basic_string<char32_t> u32string; -#endif // _LIBCPP_HAS_NO_UNICODE_CHARS - -_LIBCPP_FUNC_VIS int stoi (const string& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS long stol (const string& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS unsigned long stoul (const string& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS long long stoll (const string& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS unsigned long long stoull(const string& __str, size_t* __idx = 0, int __base = 10); +_LIBCPP_FUNC_VIS int stoi (const string& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS long stol (const string& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS unsigned long stoul (const string& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS long long stoll (const string& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS unsigned long long stoull(const string& __str, size_t* __idx = nullptr, int __base = 10); -_LIBCPP_FUNC_VIS float stof (const string& __str, size_t* __idx = 0); -_LIBCPP_FUNC_VIS double stod (const string& __str, size_t* __idx = 0); -_LIBCPP_FUNC_VIS long double stold(const string& __str, size_t* __idx = 0); +_LIBCPP_FUNC_VIS float stof (const string& __str, size_t* __idx = nullptr); +_LIBCPP_FUNC_VIS double stod (const string& __str, size_t* __idx = nullptr); +_LIBCPP_FUNC_VIS long double stold(const string& __str, size_t* __idx = nullptr); _LIBCPP_FUNC_VIS string to_string(int __val); _LIBCPP_FUNC_VIS string to_string(unsigned __val); @@ -4329,15 +4383,15 @@ _LIBCPP_FUNC_VIS string to_string(float __val); _LIBCPP_FUNC_VIS string to_string(double __val); _LIBCPP_FUNC_VIS string to_string(long double __val); -_LIBCPP_FUNC_VIS int stoi (const wstring& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS long stol (const wstring& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS unsigned long stoul (const wstring& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS long long stoll (const wstring& __str, size_t* __idx = 0, int __base = 10); -_LIBCPP_FUNC_VIS unsigned long long stoull(const wstring& __str, size_t* __idx = 0, int __base = 10); +_LIBCPP_FUNC_VIS int stoi (const wstring& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS long stol (const wstring& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS unsigned long stoul (const wstring& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS long long stoll (const wstring& __str, size_t* __idx = nullptr, int __base = 10); +_LIBCPP_FUNC_VIS unsigned long long stoull(const wstring& __str, size_t* __idx = nullptr, int __base = 10); -_LIBCPP_FUNC_VIS float stof (const wstring& __str, size_t* __idx = 0); -_LIBCPP_FUNC_VIS double stod (const wstring& __str, size_t* __idx = 0); -_LIBCPP_FUNC_VIS long double stold(const wstring& __str, size_t* __idx = 0); +_LIBCPP_FUNC_VIS float stof (const wstring& __str, size_t* __idx = nullptr); +_LIBCPP_FUNC_VIS double stod (const wstring& __str, size_t* __idx = nullptr); +_LIBCPP_FUNC_VIS long double stold(const wstring& __str, size_t* __idx = nullptr); _LIBCPP_FUNC_VIS wstring to_wstring(int __val); _LIBCPP_FUNC_VIS wstring to_wstring(unsigned __val); @@ -4425,7 +4479,7 @@ inline _LIBCPP_INLINE_VISIBILITY } #endif -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 template<class _CharT, class _Traits, class _Allocator> bool @@ -4459,7 +4513,7 @@ basic_string<_CharT, _Traits, _Allocator>::__subscriptable(const const_iterator* return this->data() <= __p && __p < this->data() + this->size(); } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 #if _LIBCPP_STD_VER > 11 // Literal suffixes for basic_string [basic.string.literals] diff --git a/lib/libcxx/include/string_view b/lib/libcxx/include/string_view index 8a684a8f96..bc0245cf2b 100644 --- a/lib/libcxx/include/string_view +++ b/lib/libcxx/include/string_view @@ -142,12 +142,16 @@ namespace std { constexpr size_type find_last_not_of(const charT* s, size_type pos, size_type n) const; constexpr size_type find_last_not_of(const charT* s, size_type pos = npos) const; - constexpr bool starts_with(basic_string_view s) const noexcept; // C++2a - constexpr bool starts_with(charT c) const noexcept; // C++2a - constexpr bool starts_with(const charT* s) const; // C++2a - constexpr bool ends_with(basic_string_view s) const noexcept; // C++2a - constexpr bool ends_with(charT c) const noexcept; // C++2a - constexpr bool ends_with(const charT* s) const; // C++2a + constexpr bool starts_with(basic_string_view s) const noexcept; // C++20 + constexpr bool starts_with(charT c) const noexcept; // C++20 + constexpr bool starts_with(const charT* s) const; // C++20 + constexpr bool ends_with(basic_string_view s) const noexcept; // C++20 + constexpr bool ends_with(charT c) const noexcept; // C++20 + constexpr bool ends_with(const charT* s) const; // C++20 + + constexpr bool contains(basic_string_view s) const noexcept; // C++2b + constexpr bool contains(charT c) const noexcept; // C++2b + constexpr bool contains(const charT* s) const; // C++2b private: const_pointer data_; // exposition only @@ -192,7 +196,26 @@ _LIBCPP_PUSH_MACROS _LIBCPP_BEGIN_NAMESPACE_STD template<class _CharT, class _Traits = char_traits<_CharT> > -class _LIBCPP_TEMPLATE_VIS basic_string_view { + class _LIBCPP_TEMPLATE_VIS basic_string_view; + +typedef basic_string_view<char> string_view; +#ifndef _LIBCPP_NO_HAS_CHAR8_T +typedef basic_string_view<char8_t> u8string_view; +#endif +typedef basic_string_view<char16_t> u16string_view; +typedef basic_string_view<char32_t> u32string_view; +typedef basic_string_view<wchar_t> wstring_view; + +template<class _CharT, class _Traits> +class + _LIBCPP_PREFERRED_NAME(string_view) +#ifndef _LIBCPP_NO_HAS_CHAR8_T + _LIBCPP_PREFERRED_NAME(u8string_view) +#endif + _LIBCPP_PREFERRED_NAME(u16string_view) + _LIBCPP_PREFERRED_NAME(u32string_view) + _LIBCPP_PREFERRED_NAME(wstring_view) + basic_string_view { public: // types typedef _Traits traits_type; @@ -236,7 +259,7 @@ public: _LIBCPP_CONSTEXPR _LIBCPP_INLINE_VISIBILITY basic_string_view(const _CharT* __s) - : __data(__s), __size(std::__char_traits_length_checked<_Traits>(__s)) {} + : __data(__s), __size(_VSTD::__char_traits_length_checked<_Traits>(__s)) {} // [string.view.iterators], iterators _LIBCPP_CONSTEXPR _LIBCPP_INLINE_VISIBILITY @@ -278,7 +301,9 @@ public: // [string.view.access], element access _LIBCPP_CONSTEXPR _LIBCPP_INLINE_VISIBILITY - const_reference operator[](size_type __pos) const _NOEXCEPT { return __data[__pos]; } + const_reference operator[](size_type __pos) const _NOEXCEPT { + return _LIBCPP_ASSERT(__pos < size(), "string_view[] index out of bounds"), __data[__pos]; + } _LIBCPP_CONSTEXPR _LIBCPP_INLINE_VISIBILITY const_reference at(size_type __pos) const @@ -601,6 +626,20 @@ public: { return ends_with(basic_string_view(__s)); } #endif +#if _LIBCPP_STD_VER > 20 + constexpr _LIBCPP_INLINE_VISIBILITY + bool contains(basic_string_view __sv) const noexcept + { return find(__sv) != npos; } + + constexpr _LIBCPP_INLINE_VISIBILITY + bool contains(value_type __c) const noexcept + { return find(__c) != npos; } + + constexpr _LIBCPP_INLINE_VISIBILITY + bool contains(const value_type* __s) const + { return find(__s) != npos; } +#endif + private: const value_type* __data; size_type __size; @@ -774,14 +813,6 @@ basic_ostream<_CharT, _Traits>& operator<<(basic_ostream<_CharT, _Traits>& __os, basic_string_view<_CharT, _Traits> __str); -typedef basic_string_view<char> string_view; -#ifndef _LIBCPP_NO_HAS_CHAR8_T -typedef basic_string_view<char8_t> u8string_view; -#endif -typedef basic_string_view<char16_t> u16string_view; -typedef basic_string_view<char32_t> u32string_view; -typedef basic_string_view<wchar_t> wstring_view; - // [string.view.hash] template<class _CharT> struct _LIBCPP_TEMPLATE_VIS hash<basic_string_view<_CharT, char_traits<_CharT> > > diff --git a/lib/libcxx/include/strstream b/lib/libcxx/include/strstream index 31999bbae1..0062777cd4 100644 --- a/lib/libcxx/include/strstream +++ b/lib/libcxx/include/strstream @@ -17,14 +17,17 @@ class strstreambuf : public basic_streambuf<char> { public: - explicit strstreambuf(streamsize alsize_arg = 0); + explicit strstreambuf(streamsize alsize_arg = 0); // before C++20 + strstreambuf() : strstreambuf(0) {} // C++20 + explicit strstreambuf(streamsize alsize_arg); // C++20 + strstreambuf(void* (*palloc_arg)(size_t), void (*pfree_arg)(void*)); - strstreambuf(char* gnext_arg, streamsize n, char* pbeg_arg = 0); + strstreambuf(char* gnext_arg, streamsize n, char* pbeg_arg = nullptr); strstreambuf(const char* gnext_arg, streamsize n); - strstreambuf(signed char* gnext_arg, streamsize n, signed char* pbeg_arg = 0); + strstreambuf(signed char* gnext_arg, streamsize n, signed char* pbeg_arg = nullptr); strstreambuf(const signed char* gnext_arg, streamsize n); - strstreambuf(unsigned char* gnext_arg, streamsize n, unsigned char* pbeg_arg = 0); + strstreambuf(unsigned char* gnext_arg, streamsize n, unsigned char* pbeg_arg = nullptr); strstreambuf(const unsigned char* gnext_arg, streamsize n); strstreambuf(strstreambuf&& rhs); @@ -140,14 +143,19 @@ class _LIBCPP_TYPE_VIS strstreambuf : public streambuf { public: +#ifndef _LIBCPP_CXX03_LANG + strstreambuf() : strstreambuf(0) {} + explicit strstreambuf(streamsize __alsize); +#else explicit strstreambuf(streamsize __alsize = 0); +#endif strstreambuf(void* (*__palloc)(size_t), void (*__pfree)(void*)); - strstreambuf(char* __gnext, streamsize __n, char* __pbeg = 0); + strstreambuf(char* __gnext, streamsize __n, char* __pbeg = nullptr); strstreambuf(const char* __gnext, streamsize __n); - strstreambuf(signed char* __gnext, streamsize __n, signed char* __pbeg = 0); + strstreambuf(signed char* __gnext, streamsize __n, signed char* __pbeg = nullptr); strstreambuf(const signed char* __gnext, streamsize __n); - strstreambuf(unsigned char* __gnext, streamsize __n, unsigned char* __pbeg = 0); + strstreambuf(unsigned char* __gnext, streamsize __n, unsigned char* __pbeg = nullptr); strstreambuf(const unsigned char* __gnext, streamsize __n); #ifndef _LIBCPP_CXX03_LANG @@ -290,7 +298,7 @@ public: _LIBCPP_INLINE_VISIBILITY ostrstream(char* __s, int __n, ios_base::openmode __mode = ios_base::out) : ostream(&__sb_), - __sb_(__s, __n, __s + (__mode & ios::app ? strlen(__s) : 0)) + __sb_(__s, __n, __s + (__mode & ios::app ? _VSTD::strlen(__s) : 0)) {} #ifndef _LIBCPP_CXX03_LANG @@ -350,7 +358,7 @@ public: _LIBCPP_INLINE_VISIBILITY strstream(char* __s, int __n, ios_base::openmode __mode = ios_base::in | ios_base::out) : iostream(&__sb_), - __sb_(__s, __n, __s + (__mode & ios::app ? strlen(__s) : 0)) + __sb_(__s, __n, __s + (__mode & ios::app ? _VSTD::strlen(__s) : 0)) {} #ifndef _LIBCPP_CXX03_LANG diff --git a/lib/libcxx/include/system_error b/lib/libcxx/include/system_error index 74e889aa9a..b714e1d426 100644 --- a/lib/libcxx/include/system_error +++ b/lib/libcxx/include/system_error @@ -253,7 +253,7 @@ public: template <class _Ep> _LIBCPP_INLINE_VISIBILITY error_condition(_Ep __e, - typename enable_if<is_error_condition_enum<_Ep>::value>::type* = 0 + typename enable_if<is_error_condition_enum<_Ep>::value>::type* = nullptr ) _NOEXCEPT {*this = make_error_condition(__e);} @@ -325,7 +325,7 @@ public: template <class _Ep> _LIBCPP_INLINE_VISIBILITY error_code(_Ep __e, - typename enable_if<is_error_code_enum<_Ep>::value>::type* = 0 + typename enable_if<is_error_code_enum<_Ep>::value>::type* = nullptr ) _NOEXCEPT {*this = make_error_code(__e);} diff --git a/lib/libcxx/include/thread b/lib/libcxx/include/thread index 6eff1800ac..34e0c2a239 100644 --- a/lib/libcxx/include/thread +++ b/lib/libcxx/include/thread @@ -277,18 +277,18 @@ inline _LIBCPP_INLINE_VISIBILITY void __thread_execute(tuple<_TSp, _Fp, _Args...>& __t, __tuple_indices<_Indices...>) { - __invoke(_VSTD::move(_VSTD::get<1>(__t)), _VSTD::move(_VSTD::get<_Indices>(__t))...); + _VSTD::__invoke(_VSTD::move(_VSTD::get<1>(__t)), _VSTD::move(_VSTD::get<_Indices>(__t))...); } template <class _Fp> _LIBCPP_INLINE_VISIBILITY void* __thread_proxy(void* __vp) { - // _Fp = std::tuple< unique_ptr<__thread_struct>, Functor, Args...> - std::unique_ptr<_Fp> __p(static_cast<_Fp*>(__vp)); - __thread_local_data().set_pointer(_VSTD::get<0>(*__p).release()); + // _Fp = tuple< unique_ptr<__thread_struct>, Functor, Args...> + unique_ptr<_Fp> __p(static_cast<_Fp*>(__vp)); + __thread_local_data().set_pointer(_VSTD::get<0>(*__p.get()).release()); typedef typename __make_tuple_indices<tuple_size<_Fp>::value, 2>::type _Index; - __thread_execute(*__p, _Index()); + _VSTD::__thread_execute(*__p.get(), _Index()); return nullptr; } @@ -300,11 +300,11 @@ thread::thread(_Fp&& __f, _Args&&... __args) typedef unique_ptr<__thread_struct> _TSPtr; _TSPtr __tsp(new __thread_struct); typedef tuple<_TSPtr, typename decay<_Fp>::type, typename decay<_Args>::type...> _Gp; - _VSTD::unique_ptr<_Gp> __p( - new _Gp(std::move(__tsp), - __decay_copy(_VSTD::forward<_Fp>(__f)), - __decay_copy(_VSTD::forward<_Args>(__args))...)); - int __ec = __libcpp_thread_create(&__t_, &__thread_proxy<_Gp>, __p.get()); + unique_ptr<_Gp> __p( + new _Gp(_VSTD::move(__tsp), + _VSTD::__decay_copy(_VSTD::forward<_Fp>(__f)), + _VSTD::__decay_copy(_VSTD::forward<_Args>(__args))...)); + int __ec = _VSTD::__libcpp_thread_create(&__t_, &__thread_proxy<_Gp>, __p.get()); if (__ec == 0) __p.release(); else @@ -326,7 +326,7 @@ struct __thread_invoke_pair { template <class _Fp> void* __thread_proxy_cxx03(void* __vp) { - std::unique_ptr<_Fp> __p(static_cast<_Fp*>(__vp)); + unique_ptr<_Fp> __p(static_cast<_Fp*>(__vp)); __thread_local_data().set_pointer(__p->__tsp_.release()); (__p->__fn_)(); return nullptr; @@ -337,9 +337,9 @@ thread::thread(_Fp __f) { typedef __thread_invoke_pair<_Fp> _InvokePair; - typedef std::unique_ptr<_InvokePair> _PairPtr; + typedef unique_ptr<_InvokePair> _PairPtr; _PairPtr __pp(new _InvokePair(__f)); - int __ec = __libcpp_thread_create(&__t_, &__thread_proxy_cxx03<_InvokePair>, __pp.get()); + int __ec = _VSTD::__libcpp_thread_create(&__t_, &__thread_proxy_cxx03<_InvokePair>, __pp.get()); if (__ec == 0) __pp.release(); else @@ -360,25 +360,24 @@ template <class _Rep, class _Period> void sleep_for(const chrono::duration<_Rep, _Period>& __d) { - using namespace chrono; - if (__d > duration<_Rep, _Period>::zero()) + if (__d > chrono::duration<_Rep, _Period>::zero()) { #if defined(_LIBCPP_COMPILER_GCC) && (__powerpc__ || __POWERPC__) // GCC's long double const folding is incomplete for IBM128 long doubles. - _LIBCPP_CONSTEXPR duration<long double> _Max = duration<long double>(ULLONG_MAX/1000000000ULL) ; + _LIBCPP_CONSTEXPR chrono::duration<long double> _Max = chrono::duration<long double>(ULLONG_MAX/1000000000ULL) ; #else - _LIBCPP_CONSTEXPR duration<long double> _Max = nanoseconds::max(); + _LIBCPP_CONSTEXPR chrono::duration<long double> _Max = chrono::nanoseconds::max(); #endif - nanoseconds __ns; + chrono::nanoseconds __ns; if (__d < _Max) { - __ns = duration_cast<nanoseconds>(__d); + __ns = chrono::duration_cast<chrono::nanoseconds>(__d); if (__ns < __d) ++__ns; } else - __ns = nanoseconds::max(); - sleep_for(__ns); + __ns = chrono::nanoseconds::max(); + this_thread::sleep_for(__ns); } } @@ -386,7 +385,6 @@ template <class _Clock, class _Duration> void sleep_until(const chrono::time_point<_Clock, _Duration>& __t) { - using namespace chrono; mutex __mut; condition_variable __cv; unique_lock<mutex> __lk(__mut); @@ -399,8 +397,7 @@ inline _LIBCPP_INLINE_VISIBILITY void sleep_until(const chrono::time_point<chrono::steady_clock, _Duration>& __t) { - using namespace chrono; - sleep_for(__t - steady_clock::now()); + this_thread::sleep_for(__t - chrono::steady_clock::now()); } inline _LIBCPP_INLINE_VISIBILITY diff --git a/lib/libcxx/include/tuple b/lib/libcxx/include/tuple index 1f80b70759..c3c7db5ff1 100644 --- a/lib/libcxx/include/tuple +++ b/lib/libcxx/include/tuple @@ -1393,7 +1393,7 @@ struct _LIBCPP_TEMPLATE_VIS uses_allocator<tuple<_Tp...>, _Alloc> template <class _T1, class _T2> template <class... _Args1, class... _Args2, size_t ..._I1, size_t ..._I2> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 pair<_T1, _T2>::pair(piecewise_construct_t, tuple<_Args1...>& __first_args, tuple<_Args2...>& __second_args, __tuple_indices<_I1...>, __tuple_indices<_I2...>) diff --git a/lib/libcxx/include/type_traits b/lib/libcxx/include/type_traits index be031037d6..59dfd1e9ba 100644 --- a/lib/libcxx/include/type_traits +++ b/lib/libcxx/include/type_traits @@ -51,6 +51,7 @@ namespace std template <class T> struct is_arithmetic; template <class T> struct is_fundamental; template <class T> struct is_member_pointer; + template <class T> struct is_scoped_enum; // C++2b template <class T> struct is_scalar; template <class T> struct is_object; template <class T> struct is_compound; @@ -284,6 +285,8 @@ namespace std = is_compound<T>::value; // C++17 template <class T> inline constexpr bool is_member_pointer_v = is_member_pointer<T>::value; // C++17 + template <class T> inline constexpr bool is_scoped_enum_v + = is_scoped_enum<T>::value; // C++2b // See C++14 20.10.4.3, type properties template <class T> inline constexpr bool is_const_v @@ -514,7 +517,7 @@ template <template <class...> class, class ...> false_type __sfinae_test_impl(...); template <template <class ...> class _Templ, class ..._Args> -using _IsValidExpansion _LIBCPP_NODEBUG_TYPE = decltype(std::__sfinae_test_impl<_Templ, _Args...>(0)); +using _IsValidExpansion _LIBCPP_NODEBUG_TYPE = decltype(__sfinae_test_impl<_Templ, _Args...>(0)); template <class> struct __void_t { typedef void type; }; @@ -595,75 +598,6 @@ using __is_primary_template = _IsValidExpansion< __test_for_primary_template, _Tp >; -// addressof -#ifndef _LIBCPP_HAS_NO_BUILTIN_ADDRESSOF - -template <class _Tp> -inline _LIBCPP_CONSTEXPR_AFTER_CXX14 -_LIBCPP_NO_CFI _LIBCPP_INLINE_VISIBILITY -_Tp* -addressof(_Tp& __x) _NOEXCEPT -{ - return __builtin_addressof(__x); -} - -#else - -template <class _Tp> -inline _LIBCPP_NO_CFI _LIBCPP_INLINE_VISIBILITY -_Tp* -addressof(_Tp& __x) _NOEXCEPT -{ - return reinterpret_cast<_Tp *>( - const_cast<char *>(&reinterpret_cast<const volatile char &>(__x))); -} - -#endif // _LIBCPP_HAS_NO_BUILTIN_ADDRESSOF - -#if defined(_LIBCPP_HAS_OBJC_ARC) && !defined(_LIBCPP_PREDEFINED_OBJC_ARC_ADDRESSOF) -// Objective-C++ Automatic Reference Counting uses qualified pointers -// that require special addressof() signatures. When -// _LIBCPP_PREDEFINED_OBJC_ARC_ADDRESSOF is defined, the compiler -// itself is providing these definitions. Otherwise, we provide them. -template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY -__strong _Tp* -addressof(__strong _Tp& __x) _NOEXCEPT -{ - return &__x; -} - -#ifdef _LIBCPP_HAS_OBJC_ARC_WEAK -template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY -__weak _Tp* -addressof(__weak _Tp& __x) _NOEXCEPT -{ - return &__x; -} -#endif - -template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY -__autoreleasing _Tp* -addressof(__autoreleasing _Tp& __x) _NOEXCEPT -{ - return &__x; -} - -template <class _Tp> -inline _LIBCPP_INLINE_VISIBILITY -__unsafe_unretained _Tp* -addressof(__unsafe_unretained _Tp& __x) _NOEXCEPT -{ - return &__x; -} -#endif - -#if !defined(_LIBCPP_CXX03_LANG) -template <class _Tp> _Tp* addressof(const _Tp&&) noexcept = delete; -#endif - struct __two {char __lx[2];}; // helper class: @@ -1731,6 +1665,21 @@ _LIBCPP_INLINE_VAR _LIBCPP_CONSTEXPR bool is_base_of_v = is_base_of<_Bp, _Dp>::value; #endif +// __is_core_convertible + +// [conv.general]/3 says "E is convertible to T" whenever "T t=E;" is well-formed. +// We can't test for that, but we can test implicit convertibility by passing it +// to a function. Notice that __is_core_convertible<void,void> is false, +// and __is_core_convertible<immovable-type,immovable-type> is true in C++17 and later. + +template <class _Tp, class _Up, class = void> +struct __is_core_convertible : public false_type {}; + +template <class _Tp, class _Up> +struct __is_core_convertible<_Tp, _Up, decltype( + static_cast<void(*)(_Up)>(0) ( static_cast<_Tp(*)()>(0)() ) +)> : public true_type {}; + // is_convertible #if __has_feature(is_convertible_to) && !defined(_LIBCPP_USE_IS_CONVERTIBLE_FALLBACK) @@ -1821,7 +1770,7 @@ template <typename _Tp> static void __test_noexcept(_Tp) noexcept; template<typename _Fm, typename _To> -static bool_constant<noexcept(__test_noexcept<_To>(declval<_Fm>()))> +static bool_constant<noexcept(_VSTD::__test_noexcept<_To>(declval<_Fm>()))> __is_nothrow_convertible_test(); template <typename _Fm, typename _To> @@ -2550,7 +2499,7 @@ _LIBCPP_INLINE_VAR _LIBCPP_CONSTEXPR bool is_destructible_v = __is_destructible( // if it's a function, return false // if it's void, return false // if it's an array of unknown bound, return false -// Otherwise, return "std::declval<_Up&>().~_Up()" is well-formed +// Otherwise, return "declval<_Up&>().~_Up()" is well-formed // where _Up is remove_all_extents<_Tp>::type template <class> @@ -2876,167 +2825,13 @@ struct __member_pointer_class_type<_Ret _ClassType::*> { typedef _ClassType type; }; -// result_of - -template <class _Callable> class result_of; - -#ifdef _LIBCPP_HAS_NO_VARIADICS - -template <class _Fn, bool, bool> -class __result_of -{ -}; - -template <class _Fn> -class __result_of<_Fn(), true, false> -{ -public: - typedef decltype(declval<_Fn>()()) type; -}; - -template <class _Fn, class _A0> -class __result_of<_Fn(_A0), true, false> -{ -public: - typedef decltype(declval<_Fn>()(declval<_A0>())) type; -}; - -template <class _Fn, class _A0, class _A1> -class __result_of<_Fn(_A0, _A1), true, false> -{ -public: - typedef decltype(declval<_Fn>()(declval<_A0>(), declval<_A1>())) type; -}; - -template <class _Fn, class _A0, class _A1, class _A2> -class __result_of<_Fn(_A0, _A1, _A2), true, false> -{ -public: - typedef decltype(declval<_Fn>()(declval<_A0>(), declval<_A1>(), declval<_A2>())) type; -}; - -template <class _MP, class _Tp, bool _IsMemberFunctionPtr> -struct __result_of_mp; - -// member function pointer - -template <class _MP, class _Tp> -struct __result_of_mp<_MP, _Tp, true> - : public __identity<typename __member_pointer_traits<_MP>::_ReturnType> -{ -}; - -// member data pointer - -template <class _MP, class _Tp, bool> -struct __result_of_mdp; - -template <class _Rp, class _Class, class _Tp> -struct __result_of_mdp<_Rp _Class::*, _Tp, false> -{ - typedef typename __apply_cv<decltype(*_VSTD::declval<_Tp>()), _Rp>::type& type; -}; - -template <class _Rp, class _Class, class _Tp> -struct __result_of_mdp<_Rp _Class::*, _Tp, true> -{ - typedef typename __apply_cv<_Tp, _Rp>::type& type; -}; - -template <class _Rp, class _Class, class _Tp> -struct __result_of_mp<_Rp _Class::*, _Tp, false> - : public __result_of_mdp<_Rp _Class::*, _Tp, - is_base_of<_Class, typename remove_reference<_Tp>::type>::value> -{ -}; - - - -template <class _Fn, class _Tp> -class __result_of<_Fn(_Tp), false, true> // _Fn must be member pointer - : public __result_of_mp<typename remove_reference<_Fn>::type, - _Tp, - is_member_function_pointer<typename remove_reference<_Fn>::type>::value> -{ -}; - -template <class _Fn, class _Tp, class _A0> -class __result_of<_Fn(_Tp, _A0), false, true> // _Fn must be member pointer - : public __result_of_mp<typename remove_reference<_Fn>::type, - _Tp, - is_member_function_pointer<typename remove_reference<_Fn>::type>::value> -{ -}; - -template <class _Fn, class _Tp, class _A0, class _A1> -class __result_of<_Fn(_Tp, _A0, _A1), false, true> // _Fn must be member pointer - : public __result_of_mp<typename remove_reference<_Fn>::type, - _Tp, - is_member_function_pointer<typename remove_reference<_Fn>::type>::value> -{ -}; - -template <class _Fn, class _Tp, class _A0, class _A1, class _A2> -class __result_of<_Fn(_Tp, _A0, _A1, _A2), false, true> // _Fn must be member pointer - : public __result_of_mp<typename remove_reference<_Fn>::type, - _Tp, - is_member_function_pointer<typename remove_reference<_Fn>::type>::value> -{ -}; - -// result_of - -template <class _Fn> -class _LIBCPP_TEMPLATE_VIS result_of<_Fn()> - : public __result_of<_Fn(), - is_class<typename remove_reference<_Fn>::type>::value || - is_function<typename remove_pointer<typename remove_reference<_Fn>::type>::type>::value, - is_member_pointer<typename remove_reference<_Fn>::type>::value - > -{ -}; - -template <class _Fn, class _A0> -class _LIBCPP_TEMPLATE_VIS result_of<_Fn(_A0)> - : public __result_of<_Fn(_A0), - is_class<typename remove_reference<_Fn>::type>::value || - is_function<typename remove_pointer<typename remove_reference<_Fn>::type>::type>::value, - is_member_pointer<typename remove_reference<_Fn>::type>::value - > -{ -}; - -template <class _Fn, class _A0, class _A1> -class _LIBCPP_TEMPLATE_VIS result_of<_Fn(_A0, _A1)> - : public __result_of<_Fn(_A0, _A1), - is_class<typename remove_reference<_Fn>::type>::value || - is_function<typename remove_pointer<typename remove_reference<_Fn>::type>::type>::value, - is_member_pointer<typename remove_reference<_Fn>::type>::value - > -{ -}; - -template <class _Fn, class _A0, class _A1, class _A2> -class _LIBCPP_TEMPLATE_VIS result_of<_Fn(_A0, _A1, _A2)> - : public __result_of<_Fn(_A0, _A1, _A2), - is_class<typename remove_reference<_Fn>::type>::value || - is_function<typename remove_pointer<typename remove_reference<_Fn>::type>::type>::value, - is_member_pointer<typename remove_reference<_Fn>::type>::value - > -{ -}; - -#endif // _LIBCPP_HAS_NO_VARIADICS - // template <class T, class... Args> struct is_constructible; -namespace __is_construct -{ -struct __nat {}; -} +#if defined(_LIBCPP_COMPILER_GCC) && _GNUC_VER_NEW >= 10000 +# define _LIBCPP_GCC_SUPPORTS_IS_CONSTRUCTIBLE +#endif -#if !defined(_LIBCPP_CXX03_LANG) && (!__has_feature(is_constructible) || \ - defined(_LIBCPP_TESTING_FALLBACK_IS_CONSTRUCTIBLE)) +#if !defined(_LIBCPP_CXX03_LANG) && !__has_feature(is_constructible) && !defined(_LIBCPP_GCC_SUPPORTS_IS_CONSTRUCTIBLE) template <class _Tp, class... _Args> struct __libcpp_is_constructible; @@ -3151,7 +2946,7 @@ struct __libcpp_is_constructible<_Tp&&, _A0> #endif -#if __has_feature(is_constructible) +#if __has_feature(is_constructible) || defined(_LIBCPP_GCC_SUPPORTS_IS_CONSTRUCTIBLE) template <class _Tp, class ..._Args> struct _LIBCPP_TEMPLATE_VIS is_constructible : public integral_constant<bool, __is_constructible(_Tp, _Args...)> @@ -3442,7 +3237,7 @@ void __implicit_conversion_to(_Tp) noexcept { } template <class _Tp, class _Arg> struct __libcpp_is_nothrow_constructible</*is constructible*/true, /*is reference*/true, _Tp, _Arg> - : public integral_constant<bool, noexcept(__implicit_conversion_to<_Tp>(declval<_Arg>()))> + : public integral_constant<bool, noexcept(_VSTD::__implicit_conversion_to<_Tp>(declval<_Arg>()))> { }; @@ -3807,7 +3602,7 @@ auto __invoke_constexpr(__any, _Args&& ...__args) -> __nat; template <class _Fp, class _A0, class ..._Args, class = __enable_if_bullet1<_Fp, _A0>> inline _LIBCPP_INLINE_VISIBILITY -auto +_LIBCPP_CONSTEXPR_AFTER_CXX17 auto __invoke(_Fp&& __f, _A0&& __a0, _Args&& ...__args) _LIBCPP_INVOKE_RETURN((_VSTD::forward<_A0>(__a0).*__f)(_VSTD::forward<_Args>(__args)...)) @@ -3821,7 +3616,7 @@ _LIBCPP_INVOKE_RETURN((_VSTD::forward<_A0>(__a0).*__f)(_VSTD::forward<_Args>(__a template <class _Fp, class _A0, class ..._Args, class = __enable_if_bullet2<_Fp, _A0>> inline _LIBCPP_INLINE_VISIBILITY -auto +_LIBCPP_CONSTEXPR_AFTER_CXX17 auto __invoke(_Fp&& __f, _A0&& __a0, _Args&& ...__args) _LIBCPP_INVOKE_RETURN((__a0.get().*__f)(_VSTD::forward<_Args>(__args)...)) @@ -3835,7 +3630,7 @@ _LIBCPP_INVOKE_RETURN((__a0.get().*__f)(_VSTD::forward<_Args>(__args)...)) template <class _Fp, class _A0, class ..._Args, class = __enable_if_bullet3<_Fp, _A0>> inline _LIBCPP_INLINE_VISIBILITY -auto +_LIBCPP_CONSTEXPR_AFTER_CXX17 auto __invoke(_Fp&& __f, _A0&& __a0, _Args&& ...__args) _LIBCPP_INVOKE_RETURN(((*_VSTD::forward<_A0>(__a0)).*__f)(_VSTD::forward<_Args>(__args)...)) @@ -3851,7 +3646,7 @@ _LIBCPP_INVOKE_RETURN(((*_VSTD::forward<_A0>(__a0)).*__f)(_VSTD::forward<_Args>( template <class _Fp, class _A0, class = __enable_if_bullet4<_Fp, _A0>> inline _LIBCPP_INLINE_VISIBILITY -auto +_LIBCPP_CONSTEXPR_AFTER_CXX17 auto __invoke(_Fp&& __f, _A0&& __a0) _LIBCPP_INVOKE_RETURN(_VSTD::forward<_A0>(__a0).*__f) @@ -3865,7 +3660,7 @@ _LIBCPP_INVOKE_RETURN(_VSTD::forward<_A0>(__a0).*__f) template <class _Fp, class _A0, class = __enable_if_bullet5<_Fp, _A0>> inline _LIBCPP_INLINE_VISIBILITY -auto +_LIBCPP_CONSTEXPR_AFTER_CXX17 auto __invoke(_Fp&& __f, _A0&& __a0) _LIBCPP_INVOKE_RETURN(__a0.get().*__f) @@ -3879,7 +3674,7 @@ _LIBCPP_INVOKE_RETURN(__a0.get().*__f) template <class _Fp, class _A0, class = __enable_if_bullet6<_Fp, _A0>> inline _LIBCPP_INLINE_VISIBILITY -auto +_LIBCPP_CONSTEXPR_AFTER_CXX17 auto __invoke(_Fp&& __f, _A0&& __a0) _LIBCPP_INVOKE_RETURN((*_VSTD::forward<_A0>(__a0)).*__f) @@ -3894,7 +3689,7 @@ _LIBCPP_INVOKE_RETURN((*_VSTD::forward<_A0>(__a0)).*__f) template <class _Fp, class ..._Args> inline _LIBCPP_INLINE_VISIBILITY -auto +_LIBCPP_CONSTEXPR_AFTER_CXX17 auto __invoke(_Fp&& __f, _Args&& ...__args) _LIBCPP_INVOKE_RETURN(_VSTD::forward<_Fp>(__f)(_VSTD::forward<_Args>(__args)...)) @@ -3982,14 +3777,97 @@ struct __invoke_of { }; +#endif // _LIBCPP_CXX03_LANG + // result_of +template <class _Callable> class result_of; + +#ifndef _LIBCPP_CXX03_LANG + template <class _Fp, class ..._Args> class _LIBCPP_TEMPLATE_VIS result_of<_Fp(_Args...)> : public __invoke_of<_Fp, _Args...> { }; +#else // C++03 + +template <class _Fn, bool, bool> +class __result_of +{ +}; + +template <class _Fn, class ..._Args> +class __result_of<_Fn(_Args...), true, false> +{ +public: + typedef decltype(declval<_Fn>()(declval<_Args>()...)) type; +}; + +template <class _MP, class _Tp, bool _IsMemberFunctionPtr> +struct __result_of_mp; + +// member function pointer + +template <class _MP, class _Tp> +struct __result_of_mp<_MP, _Tp, true> + : public __identity<typename __member_pointer_traits<_MP>::_ReturnType> +{ +}; + +// member data pointer + +template <class _MP, class _Tp, bool> +struct __result_of_mdp; + +template <class _Rp, class _Class, class _Tp> +struct __result_of_mdp<_Rp _Class::*, _Tp, false> +{ + typedef typename __apply_cv<decltype(*_VSTD::declval<_Tp>()), _Rp>::type& type; +}; + +template <class _Rp, class _Class, class _Tp> +struct __result_of_mdp<_Rp _Class::*, _Tp, true> +{ + typedef typename __apply_cv<_Tp, _Rp>::type& type; +}; + +template <class _Rp, class _Class, class _Tp> +struct __result_of_mp<_Rp _Class::*, _Tp, false> + : public __result_of_mdp<_Rp _Class::*, _Tp, + is_base_of<_Class, typename remove_reference<_Tp>::type>::value> +{ +}; + +template <class _Fn, class _Tp> +class __result_of<_Fn(_Tp), false, true> // _Fn must be member pointer + : public __result_of_mp<typename remove_reference<_Fn>::type, + _Tp, + is_member_function_pointer<typename remove_reference<_Fn>::type>::value> +{ +}; + +template <class _Fn, class _Tp, class ..._Args> +class __result_of<_Fn(_Tp, _Args...), false, true> // _Fn must be member pointer + : public __result_of_mp<typename remove_reference<_Fn>::type, + _Tp, + is_member_function_pointer<typename remove_reference<_Fn>::type>::value> +{ +}; + +template <class _Fn, class ..._Args> +class _LIBCPP_TEMPLATE_VIS result_of<_Fn(_Args...)> + : public __result_of<_Fn(_Args...), + is_class<typename remove_reference<_Fn>::type>::value || + is_function<typename remove_pointer<typename remove_reference<_Fn>::type>::type>::value, + is_member_pointer<typename remove_reference<_Fn>::type>::value + > +{ +}; + +#endif // C++03 + #if _LIBCPP_STD_VER > 11 template <class _Tp> using result_of_t = typename result_of<_Tp>::type; #endif @@ -4045,8 +3923,6 @@ _LIBCPP_INLINE_VAR constexpr bool is_nothrow_invocable_r_v #endif // _LIBCPP_STD_VER > 14 -#endif // !defined(_LIBCPP_CXX03_LANG) - template <class _Tp> struct __is_swappable; template <class _Tp> struct __is_nothrow_swappable; @@ -4319,6 +4195,25 @@ struct __has_operator_addressof #endif // _LIBCPP_CXX03_LANG +// is_scoped_enum [meta.unary.prop] + +#if _LIBCPP_STD_VER > 20 +template <class _Tp, bool = is_enum_v<_Tp> > +struct __is_scoped_enum_helper : false_type {}; + +template <class _Tp> +struct __is_scoped_enum_helper<_Tp, true> + : public bool_constant<!is_convertible_v<_Tp, underlying_type_t<_Tp> > > {}; + +template <class _Tp> +struct _LIBCPP_TEMPLATE_VIS is_scoped_enum + : public __is_scoped_enum_helper<_Tp> {}; + +template <class _Tp> +_LIBCPP_INLINE_VAR _LIBCPP_CONSTEXPR bool is_scoped_enum_v = + is_scoped_enum<_Tp>::value; +#endif + #if _LIBCPP_STD_VER > 14 template <class... _Args> @@ -4341,7 +4236,6 @@ _LIBCPP_INLINE_VAR constexpr bool negation_v #endif // _LIBCPP_STD_VER > 14 // These traits are used in __tree and __hash_table -#ifndef _LIBCPP_CXX03_LANG struct __extract_key_fail_tag {}; struct __extract_key_self_tag {}; struct __extract_key_first_tag {}; @@ -4353,7 +4247,7 @@ struct __can_extract_key __extract_key_fail_tag>::type {}; template <class _Pair, class _Key, class _First, class _Second> -struct __can_extract_key<_Pair, _Key, pair<_First, _Second>> +struct __can_extract_key<_Pair, _Key, pair<_First, _Second> > : conditional<_IsSame<typename remove_const<_First>::type, _Key>::value, __extract_key_first_tag, __extract_key_fail_tag>::type {}; @@ -4371,8 +4265,6 @@ template <class _ValTy, class _Key, class _RawValTy> struct __can_extract_map_key<_ValTy, _Key, _Key, _RawValTy> : false_type {}; -#endif - #ifndef _LIBCPP_HAS_NO_BUILTIN_IS_CONSTANT_EVALUATED #if _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY diff --git a/lib/libcxx/include/typeinfo b/lib/libcxx/include/typeinfo index 7e76da5387..048e57614a 100644 --- a/lib/libcxx/include/typeinfo +++ b/lib/libcxx/include/typeinfo @@ -57,6 +57,7 @@ public: */ #include <__config> +#include <__availability> #include <exception> #include <cstddef> #include <cstdint> @@ -141,7 +142,7 @@ public: // comparison is equal. // -------------------------------------------------------------------------- // // NonUniqueARMRTTIBit -// (selected on ARM64 regardless of _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION) +// (_LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION = 3) // -------------------------------------------------------------------------- // // This implementation of type_info does not assume always a unique copy of // the RTTI for a given type inside a program. It packs the pointer to the @@ -160,6 +161,24 @@ public: // the pointer when it constructs the type_info, depending on whether it can // guarantee uniqueness for that specific type_info. +// This value can be overriden in the __config_site. When it's not overriden, +// we pick a default implementation based on the platform here. +#ifndef _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION + + // Windows binaries can't merge typeinfos, so use the NonUnique implementation. +# ifdef _LIBCPP_OBJECT_FORMAT_COFF +# define _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION 2 + + // On arm64 on Apple platforms, use the special NonUniqueARMRTTIBit implementation. +# elif defined(__APPLE__) && defined(__LP64__) && !defined(__x86_64__) +# define _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION 3 + + // On all other platforms, assume the Itanium C++ ABI and use the Unique implementation. +# else +# define _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION 1 +# endif +#endif + struct __type_info_implementations { struct __string_impl_base { typedef const char* __type_name_t; @@ -243,7 +262,7 @@ struct __type_info_implementations { private: // The unique bit is the top bit. It is expected that __type_name_t is 64 bits when // this implementation is actually used. - typedef std::integral_constant<__type_name_t, + typedef integral_constant<__type_name_t, (1ULL << ((__CHAR_BIT__ * sizeof(__type_name_t)) - 1))> __non_unique_rtti_bit; _LIBCPP_INLINE_VISIBILITY @@ -257,12 +276,12 @@ struct __type_info_implementations { }; typedef -#if defined(__APPLE__) && defined(__LP64__) && !defined(__x86_64__) - __non_unique_arm_rtti_bit_impl -#elif _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION == 1 +#if _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION == 1 __unique_impl #elif _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION == 2 __non_unique_impl +#elif _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION == 3 + __non_unique_arm_rtti_bit_impl #else # error invalid configuration for _LIBCPP_TYPEINFO_COMPARISON_IMPLEMENTATION #endif diff --git a/lib/libcxx/include/unordered_map b/lib/libcxx/include/unordered_map index 6156cfddd7..d595302e63 100644 --- a/lib/libcxx/include/unordered_map +++ b/lib/libcxx/include/unordered_map @@ -173,10 +173,22 @@ public: iterator find(const key_type& k); const_iterator find(const key_type& k) const; + template<typename K> + iterator find(const K& x); // C++20 + template<typename K> + const_iterator find(const K& x) const; // C++20 size_type count(const key_type& k) const; + template<typename K> + size_type count(const K& k) const; // C++20 bool contains(const key_type& k) const; // C++20 + template<typename K> + bool contains(const K& k) const; // C++20 pair<iterator, iterator> equal_range(const key_type& k); pair<const_iterator, const_iterator> equal_range(const key_type& k) const; + template<typename K> + pair<iterator, iterator> equal_range(const K& k); // C++20 + template<typename K> + pair<const_iterator, const_iterator> equal_range(const K& k) const; // C++20 mapped_type& operator[](const key_type& k); mapped_type& operator[](key_type&& k); @@ -355,10 +367,22 @@ public: iterator find(const key_type& k); const_iterator find(const key_type& k) const; + template<typename K> + iterator find(const K& x); // C++20 + template<typename K> + const_iterator find(const K& x) const; // C++20 size_type count(const key_type& k) const; + template<typename K> + size_type count(const K& k) const; // C++20 bool contains(const key_type& k) const; // C++20 + template<typename K> + bool contains(const K& k) const; // C++20 pair<iterator, iterator> equal_range(const key_type& k); pair<const_iterator, const_iterator> equal_range(const key_type& k) const; + template<typename K> + pair<iterator, iterator> equal_range(const K& k); // C++20 + template<typename K> + pair<const_iterator, const_iterator> equal_range(const K& k) const; // C++20 size_type bucket_count() const noexcept; size_type max_bucket_count() const noexcept; @@ -423,7 +447,7 @@ template <class Key, class T, class Hash, class Pred, class Alloc> _LIBCPP_BEGIN_NAMESPACE_STD -template <class _Key, class _Cp, class _Hash, +template <class _Key, class _Cp, class _Hash, class _Pred, bool = is_empty<_Hash>::value && !__libcpp_is_final<_Hash>::value> class __unordered_map_hasher : private _Hash @@ -445,6 +469,12 @@ public: _LIBCPP_INLINE_VISIBILITY size_t operator()(const _Key& __x) const {return static_cast<const _Hash&>(*this)(__x);} +#if _LIBCPP_STD_VER > 17 + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + size_t operator()(const _K2& __x) const + {return static_cast<const _Hash&>(*this)(__x);} +#endif void swap(__unordered_map_hasher&__y) _NOEXCEPT_(__is_nothrow_swappable<_Hash>::value) { @@ -453,8 +483,8 @@ public: } }; -template <class _Key, class _Cp, class _Hash> -class __unordered_map_hasher<_Key, _Cp, _Hash, false> +template <class _Key, class _Cp, class _Hash, class _Pred> +class __unordered_map_hasher<_Key, _Cp, _Hash, _Pred, false> { _Hash __hash_; public: @@ -474,6 +504,12 @@ public: _LIBCPP_INLINE_VISIBILITY size_t operator()(const _Key& __x) const {return __hash_(__x);} +#if _LIBCPP_STD_VER > 17 + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + size_t operator()(const _K2& __x) const + {return __hash_(__x);} +#endif void swap(__unordered_map_hasher&__y) _NOEXCEPT_(__is_nothrow_swappable<_Hash>::value) { @@ -482,17 +518,17 @@ public: } }; -template <class _Key, class _Cp, class _Hash, bool __b> +template <class _Key, class _Cp, class _Hash, class _Pred, bool __b> inline _LIBCPP_INLINE_VISIBILITY void -swap(__unordered_map_hasher<_Key, _Cp, _Hash, __b>& __x, - __unordered_map_hasher<_Key, _Cp, _Hash, __b>& __y) +swap(__unordered_map_hasher<_Key, _Cp, _Hash, _Pred, __b>& __x, + __unordered_map_hasher<_Key, _Cp, _Hash, _Pred, __b>& __y) _NOEXCEPT_(_NOEXCEPT_(__x.swap(__y))) { __x.swap(__y); } -template <class _Key, class _Cp, class _Pred, +template <class _Key, class _Cp, class _Pred, class _Hash, bool = is_empty<_Pred>::value && !__libcpp_is_final<_Pred>::value> class __unordered_map_equal : private _Pred @@ -517,6 +553,24 @@ public: _LIBCPP_INLINE_VISIBILITY bool operator()(const _Key& __x, const _Cp& __y) const {return static_cast<const _Pred&>(*this)(__x, __y.__get_value().first);} +#if _LIBCPP_STD_VER > 17 + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _Cp& __x, const _K2& __y) const + {return static_cast<const _Pred&>(*this)(__x.__get_value().first, __y);} + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _K2& __x, const _Cp& __y) const + {return static_cast<const _Pred&>(*this)(__x, __y.__get_value().first);} + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _Key& __x, const _K2& __y) const + {return static_cast<const _Pred&>(*this)(__x, __y);} + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _K2& __x, const _Key& __y) const + {return static_cast<const _Pred&>(*this)(__x, __y);} +#endif void swap(__unordered_map_equal&__y) _NOEXCEPT_(__is_nothrow_swappable<_Pred>::value) { @@ -525,8 +579,8 @@ public: } }; -template <class _Key, class _Cp, class _Pred> -class __unordered_map_equal<_Key, _Cp, _Pred, false> +template <class _Key, class _Cp, class _Pred, class _Hash> +class __unordered_map_equal<_Key, _Cp, _Pred, _Hash, false> { _Pred __pred_; public: @@ -549,6 +603,24 @@ public: _LIBCPP_INLINE_VISIBILITY bool operator()(const _Key& __x, const _Cp& __y) const {return __pred_(__x, __y.__get_value().first);} +#if _LIBCPP_STD_VER > 17 + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _Cp& __x, const _K2& __y) const + {return __pred_(__x.__get_value().first, __y);} + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _K2& __x, const _Cp& __y) const + {return __pred_(__x, __y.__get_value().first);} + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _Key& __x, const _K2& __y) const + {return __pred_(__x, __y);} + template <typename _K2, typename = _EnableIf<__is_transparent<_Hash, _K2>::value && __is_transparent<_Pred, _K2>::value>> + _LIBCPP_INLINE_VISIBILITY + bool operator()(const _K2& __x, const _Key& __y) const + {return __pred_(__x, __y);} +#endif void swap(__unordered_map_equal&__y) _NOEXCEPT_(__is_nothrow_swappable<_Pred>::value) { @@ -557,11 +629,11 @@ public: } }; -template <class _Key, class _Cp, class _Pred, bool __b> +template <class _Key, class _Cp, class _Pred, class _Hash, bool __b> inline _LIBCPP_INLINE_VISIBILITY void -swap(__unordered_map_equal<_Key, _Cp, _Pred, __b>& __x, - __unordered_map_equal<_Key, _Cp, _Pred, __b>& __y) +swap(__unordered_map_equal<_Key, _Cp, _Pred, _Hash, __b>& __x, + __unordered_map_equal<_Key, _Cp, _Pred, _Hash, __b>& __y) _NOEXCEPT_(_NOEXCEPT_(__x.swap(__y))) { __x.swap(__y); @@ -858,11 +930,11 @@ public: "Invalid allocator::value_type"); private: - typedef __hash_value_type<key_type, mapped_type> __value_type; - typedef __unordered_map_hasher<key_type, __value_type, hasher> __hasher; - typedef __unordered_map_equal<key_type, __value_type, key_equal> __key_equal; + typedef __hash_value_type<key_type, mapped_type> __value_type; + typedef __unordered_map_hasher<key_type, __value_type, hasher, key_equal> __hasher; + typedef __unordered_map_equal<key_type, __value_type, key_equal, hasher> __key_equal; typedef typename __rebind_alloc_helper<allocator_traits<allocator_type>, - __value_type>::type __allocator_type; + __value_type>::type __allocator_type; typedef __hash_table<__value_type, __hasher, __key_equal, __allocator_type> __table; @@ -906,7 +978,7 @@ public: unordered_map() _NOEXCEPT_(is_nothrow_default_constructible<__table>::value) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1025,7 +1097,7 @@ public: {return __table_.__insert_unique(__x);} iterator insert(const_iterator __p, const value_type& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "unordered_map::insert(const_iterator, const value_type&) called with an iterator not" " referring to this unordered_map"); @@ -1049,7 +1121,7 @@ public: {return __table_.__insert_unique(_VSTD::move(__x));} iterator insert(const_iterator __p, value_type&& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "unordered_map::insert(const_iterator, const value_type&) called with an iterator not" " referring to this unordered_map"); @@ -1070,7 +1142,7 @@ public: _LIBCPP_INLINE_VISIBILITY iterator insert(const_iterator __p, _Pp&& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "unordered_map::insert(const_iterator, value_type&&) called with an iterator not" " referring to this unordered_map"); @@ -1089,7 +1161,7 @@ public: template <class... _Args> _LIBCPP_INLINE_VISIBILITY iterator emplace_hint(const_iterator __p, _Args&&... __args) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, "unordered_map::emplace_hint(const_iterator, args...) called with an iterator not" " referring to this unordered_map"); @@ -1106,7 +1178,7 @@ public: _LIBCPP_INLINE_VISIBILITY pair<iterator, bool> try_emplace(const key_type& __k, _Args&&... __args) { - return __table_.__emplace_unique_key_args(__k, _VSTD::piecewise_construct, + return __table_.__emplace_unique_key_args(__k, piecewise_construct, _VSTD::forward_as_tuple(__k), _VSTD::forward_as_tuple(_VSTD::forward<_Args>(__args)...)); } @@ -1115,7 +1187,7 @@ public: _LIBCPP_INLINE_VISIBILITY pair<iterator, bool> try_emplace(key_type&& __k, _Args&&... __args) { - return __table_.__emplace_unique_key_args(__k, _VSTD::piecewise_construct, + return __table_.__emplace_unique_key_args(__k, piecewise_construct, _VSTD::forward_as_tuple(_VSTD::move(__k)), _VSTD::forward_as_tuple(_VSTD::forward<_Args>(__args)...)); } @@ -1124,7 +1196,7 @@ public: _LIBCPP_INLINE_VISIBILITY iterator try_emplace(const_iterator __h, const key_type& __k, _Args&&... __args) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__h) == this, "unordered_map::try_emplace(const_iterator, key, args...) called with an iterator not" " referring to this unordered_map"); @@ -1138,7 +1210,7 @@ public: _LIBCPP_INLINE_VISIBILITY iterator try_emplace(const_iterator __h, key_type&& __k, _Args&&... __args) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__h) == this, "unordered_map::try_emplace(const_iterator, key, args...) called with an iterator not" " referring to this unordered_map"); @@ -1280,11 +1352,34 @@ public: iterator find(const key_type& __k) {return __table_.find(__k);} _LIBCPP_INLINE_VISIBILITY const_iterator find(const key_type& __k) const {return __table_.find(__k);} + + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, iterator> + find(const _K2& __k) {return __table_.find(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, const_iterator> + find(const _K2& __k) const {return __table_.find(__k);} + #endif // _LIBCPP_STD_VER > 17 + _LIBCPP_INLINE_VISIBILITY size_type count(const key_type& __k) const {return __table_.__count_unique(__k);} #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, size_type> + count(const _K2& __k) const {return __table_.__count_unique(__k);} + #endif // _LIBCPP_STD_VER > 17 + #if _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY bool contains(const key_type& __k) const {return find(__k) != end();} + + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, bool> + contains(const _K2& __k) const {return find(__k) != end();} #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY pair<iterator, iterator> equal_range(const key_type& __k) @@ -1292,6 +1387,16 @@ public: _LIBCPP_INLINE_VISIBILITY pair<const_iterator, const_iterator> equal_range(const key_type& __k) const {return __table_.__equal_range_unique(__k);} + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<iterator, iterator>> + equal_range(const _K2& __k) {return __table_.__equal_range_unique(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<const_iterator, const_iterator>> + equal_range(const _K2& __k) const {return __table_.__equal_range_unique(__k);} + #endif // _LIBCPP_STD_VER > 17 mapped_type& operator[](const key_type& __k); #ifndef _LIBCPP_CXX03_LANG @@ -1336,7 +1441,7 @@ public: _LIBCPP_INLINE_VISIBILITY void reserve(size_type __n) {__table_.reserve(__n);} -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const {return __table_.__dereferenceable(&__i->__i_);} @@ -1347,7 +1452,7 @@ public: bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const {return __table_.__addable(&__i->__i_, __n);} -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 private: @@ -1428,7 +1533,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( size_type __n, const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1440,7 +1545,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const allocator_type& __a) : __table_(__hf, __eql, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1452,7 +1557,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const allocator_type& __a) : __table_(typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1462,7 +1567,7 @@ template <class _InputIterator> unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( _InputIterator __first, _InputIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__first, __last); @@ -1475,7 +1580,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1489,7 +1594,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const hasher& __hf, const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1501,7 +1606,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const unordered_map& __u) : __table_(__u.__table_) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -1513,7 +1618,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const unordered_map& __u, const allocator_type& __a) : __table_(__u.__table_, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -1529,7 +1634,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( _NOEXCEPT_(is_nothrow_move_constructible<__table>::value) : __table_(_VSTD::move(__u.__table_)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); __get_db()->swap(this, &__u); #endif @@ -1540,7 +1645,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( unordered_map&& __u, const allocator_type& __a) : __table_(_VSTD::move(__u.__table_), typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__a != __u.get_allocator()) @@ -1551,7 +1656,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( __u.__table_.remove((__i++).__i_)->__value_.__move()); } } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 else __get_db()->swap(this, &__u); #endif @@ -1561,7 +1666,7 @@ template <class _Key, class _Tp, class _Hash, class _Pred, class _Alloc> unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( initializer_list<value_type> __il) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__il.begin(), __il.end()); @@ -1573,7 +1678,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1586,7 +1691,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_map( const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1633,8 +1738,8 @@ _Tp& unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::operator[](const key_type& __k) { return __table_.__emplace_unique_key_args(__k, - std::piecewise_construct, std::forward_as_tuple(__k), - std::forward_as_tuple()).first->__get_value().second; + piecewise_construct, _VSTD::forward_as_tuple(__k), + _VSTD::forward_as_tuple()).first->__get_value().second; } template <class _Key, class _Tp, class _Hash, class _Pred, class _Alloc> @@ -1642,8 +1747,8 @@ _Tp& unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::operator[](key_type&& __k) { return __table_.__emplace_unique_key_args(__k, - std::piecewise_construct, std::forward_as_tuple(std::move(__k)), - std::forward_as_tuple()).first->__get_value().second; + piecewise_construct, _VSTD::forward_as_tuple(_VSTD::move(__k)), + _VSTD::forward_as_tuple()).first->__get_value().second; } #else // _LIBCPP_CXX03_LANG @@ -1657,7 +1762,7 @@ unordered_map<_Key, _Tp, _Hash, _Pred, _Alloc>::__construct_node_with_key(const __h.get_deleter().__first_constructed = true; __node_traits::construct(__na, _VSTD::addressof(__h->__value_.__get_value().second)); __h.get_deleter().__second_constructed = true; - return _LIBCPP_EXPLICIT_MOVE(__h); // explicitly moved for C++03 + return __h; } template <class _Key, class _Tp, class _Hash, class _Pred, class _Alloc> @@ -1762,11 +1867,11 @@ public: "Invalid allocator::value_type"); private: - typedef __hash_value_type<key_type, mapped_type> __value_type; - typedef __unordered_map_hasher<key_type, __value_type, hasher> __hasher; - typedef __unordered_map_equal<key_type, __value_type, key_equal> __key_equal; + typedef __hash_value_type<key_type, mapped_type> __value_type; + typedef __unordered_map_hasher<key_type, __value_type, hasher, key_equal> __hasher; + typedef __unordered_map_equal<key_type, __value_type, key_equal, hasher> __key_equal; typedef typename __rebind_alloc_helper<allocator_traits<allocator_type>, - __value_type>::type __allocator_type; + __value_type>::type __allocator_type; typedef __hash_table<__value_type, __hasher, __key_equal, __allocator_type> __table; @@ -1807,7 +1912,7 @@ public: unordered_multimap() _NOEXCEPT_(is_nothrow_default_constructible<__table>::value) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2059,11 +2164,32 @@ public: iterator find(const key_type& __k) {return __table_.find(__k);} _LIBCPP_INLINE_VISIBILITY const_iterator find(const key_type& __k) const {return __table_.find(__k);} + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, iterator> + find(const _K2& __k) {return __table_.find(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, const_iterator> + find(const _K2& __k) const {return __table_.find(__k);} + #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY size_type count(const key_type& __k) const {return __table_.__count_multi(__k);} #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, size_type> + count(const _K2& __k) const {return __table_.__count_multi(__k);} + #endif // _LIBCPP_STD_VER > 17 + #if _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY bool contains(const key_type& __k) const {return find(__k) != end();} + + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, bool> + contains(const _K2& __k) const {return find(__k) != end();} #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY pair<iterator, iterator> equal_range(const key_type& __k) @@ -2071,6 +2197,16 @@ public: _LIBCPP_INLINE_VISIBILITY pair<const_iterator, const_iterator> equal_range(const key_type& __k) const {return __table_.__equal_range_multi(__k);} + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<iterator, iterator>> + equal_range(const _K2& __k) {return __table_.__equal_range_multi(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<const_iterator, const_iterator>> + equal_range(const _K2& __k) const {return __table_.__equal_range_multi(__k);} + #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY size_type bucket_count() const _NOEXCEPT {return __table_.bucket_count();} @@ -2108,7 +2244,7 @@ public: _LIBCPP_INLINE_VISIBILITY void reserve(size_type __n) {__table_.reserve(__n);} -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const {return __table_.__dereferenceable(&__i->__i_);} @@ -2119,7 +2255,7 @@ public: bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const {return __table_.__addable(&__i->__i_, __n);} -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 }; @@ -2196,7 +2332,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( size_type __n, const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -2208,7 +2344,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const allocator_type& __a) : __table_(__hf, __eql, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -2219,7 +2355,7 @@ template <class _InputIterator> unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( _InputIterator __first, _InputIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__first, __last); @@ -2232,7 +2368,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -2246,7 +2382,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const hasher& __hf, const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -2259,7 +2395,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const allocator_type& __a) : __table_(typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -2269,7 +2405,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const unordered_multimap& __u) : __table_(__u.__table_) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -2281,7 +2417,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const unordered_multimap& __u, const allocator_type& __a) : __table_(__u.__table_, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -2297,7 +2433,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( _NOEXCEPT_(is_nothrow_move_constructible<__table>::value) : __table_(_VSTD::move(__u.__table_)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); __get_db()->swap(this, &__u); #endif @@ -2308,7 +2444,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( unordered_multimap&& __u, const allocator_type& __a) : __table_(_VSTD::move(__u.__table_), typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__a != __u.get_allocator()) @@ -2320,7 +2456,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( __u.__table_.remove((__i++).__i_)->__value_.__move()); } } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 else __get_db()->swap(this, &__u); #endif @@ -2330,7 +2466,7 @@ template <class _Key, class _Tp, class _Hash, class _Pred, class _Alloc> unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( initializer_list<value_type> __il) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__il.begin(), __il.end()); @@ -2342,7 +2478,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -2355,7 +2491,7 @@ unordered_multimap<_Key, _Tp, _Hash, _Pred, _Alloc>::unordered_multimap( const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, typename __table::allocator_type(__a)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); diff --git a/lib/libcxx/include/unordered_set b/lib/libcxx/include/unordered_set index 6c4ad93800..80f460d7ed 100644 --- a/lib/libcxx/include/unordered_set +++ b/lib/libcxx/include/unordered_set @@ -145,10 +145,22 @@ public: iterator find(const key_type& k); const_iterator find(const key_type& k) const; + template<typename K> + iterator find(const K& x); // C++20 + template<typename K> + const_iterator find(const K& x) const; // C++20 size_type count(const key_type& k) const; + template<typename K> + size_type count(const K& k) const; // C++20 bool contains(const key_type& k) const; // C++20 + template<typename K> + bool contains(const K& k) const; // C++20 pair<iterator, iterator> equal_range(const key_type& k); pair<const_iterator, const_iterator> equal_range(const key_type& k) const; + template<typename K> + pair<iterator, iterator> equal_range(const K& k); // C++20 + template<typename K> + pair<const_iterator, const_iterator> equal_range(const K& k) const; // C++20 size_type bucket_count() const noexcept; size_type max_bucket_count() const noexcept; @@ -310,10 +322,22 @@ public: iterator find(const key_type& k); const_iterator find(const key_type& k) const; + template<typename K> + iterator find(const K& x); // C++20 + template<typename K> + const_iterator find(const K& x) const; // C++20 size_type count(const key_type& k) const; + template<typename K> + size_type count(const K& k) const; // C++20 bool contains(const key_type& k) const; // C++20 + template<typename K> + bool contains(const K& k) const; // C++20 pair<iterator, iterator> equal_range(const key_type& k); pair<const_iterator, const_iterator> equal_range(const key_type& k) const; + template<typename K> + pair<iterator, iterator> equal_range(const K& k); // C++20 + template<typename K> + pair<const_iterator, const_iterator> equal_range(const K& k) const; // C++20 size_type bucket_count() const noexcept; size_type max_bucket_count() const noexcept; @@ -425,7 +449,7 @@ public: unordered_set() _NOEXCEPT_(is_nothrow_default_constructible<__table>::value) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -539,7 +563,7 @@ public: {return __table_.__emplace_unique(_VSTD::forward<_Args>(__args)...);} template <class... _Args> _LIBCPP_INLINE_VISIBILITY -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 iterator emplace_hint(const_iterator __p, _Args&&... __args) { _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, @@ -556,7 +580,7 @@ public: pair<iterator, bool> insert(value_type&& __x) {return __table_.__insert_unique(_VSTD::move(__x));} _LIBCPP_INLINE_VISIBILITY -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 iterator insert(const_iterator __p, value_type&& __x) { _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, @@ -577,7 +601,7 @@ public: {return __table_.__insert_unique(__x);} _LIBCPP_INLINE_VISIBILITY -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 iterator insert(const_iterator __p, const value_type& __x) { _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__p) == this, @@ -679,11 +703,32 @@ public: iterator find(const key_type& __k) {return __table_.find(__k);} _LIBCPP_INLINE_VISIBILITY const_iterator find(const key_type& __k) const {return __table_.find(__k);} + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, iterator> + find(const _K2& __k) {return __table_.find(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, const_iterator> + find(const _K2& __k) const {return __table_.find(__k);} + #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY size_type count(const key_type& __k) const {return __table_.__count_unique(__k);} #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, size_type> + count(const _K2& __k) const {return __table_.__count_unique(__k);} + #endif // _LIBCPP_STD_VER > 17 + #if _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY bool contains(const key_type& __k) const {return find(__k) != end();} + + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, bool> + contains(const _K2& __k) const {return find(__k) != end();} #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY pair<iterator, iterator> equal_range(const key_type& __k) @@ -691,6 +736,16 @@ public: _LIBCPP_INLINE_VISIBILITY pair<const_iterator, const_iterator> equal_range(const key_type& __k) const {return __table_.__equal_range_unique(__k);} + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<iterator, iterator>> + equal_range(const _K2& __k) {return __table_.__equal_range_unique(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<const_iterator, const_iterator>> + equal_range(const _K2& __k) const {return __table_.__equal_range_unique(__k);} + #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY size_type bucket_count() const _NOEXCEPT {return __table_.bucket_count();} @@ -726,7 +781,7 @@ public: _LIBCPP_INLINE_VISIBILITY void reserve(size_type __n) {__table_.reserve(__n);} -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const {return __table_.__dereferenceable(__i);} @@ -737,7 +792,7 @@ public: bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const {return __table_.__addable(__i, __n);} -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 }; @@ -802,7 +857,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set(size_type __n, const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -813,7 +868,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set(size_type __n, const hasher& __hf, const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -824,7 +879,7 @@ template <class _InputIterator> unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( _InputIterator __first, _InputIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__first, __last); @@ -837,7 +892,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -851,7 +906,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( const hasher& __hf, const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -864,7 +919,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( const allocator_type& __a) : __table_(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -874,7 +929,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( const unordered_set& __u) : __table_(__u.__table_) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -886,7 +941,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( const unordered_set& __u, const allocator_type& __a) : __table_(__u.__table_, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -902,7 +957,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( _NOEXCEPT_(is_nothrow_move_constructible<__table>::value) : __table_(_VSTD::move(__u.__table_)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); __get_db()->swap(this, &__u); #endif @@ -913,7 +968,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( unordered_set&& __u, const allocator_type& __a) : __table_(_VSTD::move(__u.__table_), __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__a != __u.get_allocator()) @@ -922,7 +977,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( while (__u.size() != 0) __table_.__insert_unique(_VSTD::move(__u.__table_.remove(__i++)->__value_)); } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 else __get_db()->swap(this, &__u); #endif @@ -932,7 +987,7 @@ template <class _Value, class _Hash, class _Pred, class _Alloc> unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( initializer_list<value_type> __il) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__il.begin(), __il.end()); @@ -944,7 +999,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -957,7 +1012,7 @@ unordered_set<_Value, _Hash, _Pred, _Alloc>::unordered_set( const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1091,7 +1146,7 @@ public: unordered_multiset() _NOEXCEPT_(is_nothrow_default_constructible<__table>::value) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1314,11 +1369,32 @@ public: iterator find(const key_type& __k) {return __table_.find(__k);} _LIBCPP_INLINE_VISIBILITY const_iterator find(const key_type& __k) const {return __table_.find(__k);} + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, iterator> + find(const _K2& __k) {return __table_.find(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, const_iterator> + find(const _K2& __k) const {return __table_.find(__k);} + #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY size_type count(const key_type& __k) const {return __table_.__count_multi(__k);} #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, size_type> + count(const _K2& __k) const {return __table_.__count_multi(__k);} + #endif // _LIBCPP_STD_VER > 17 + #if _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY bool contains(const key_type& __k) const {return find(__k) != end();} + + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, bool> + contains(const _K2& __k) const {return find(__k) != end();} #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY pair<iterator, iterator> equal_range(const key_type& __k) @@ -1326,6 +1402,16 @@ public: _LIBCPP_INLINE_VISIBILITY pair<const_iterator, const_iterator> equal_range(const key_type& __k) const {return __table_.__equal_range_multi(__k);} + #if _LIBCPP_STD_VER > 17 + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<iterator, iterator>> + equal_range(const _K2& __k) {return __table_.__equal_range_multi(__k);} + template <typename _K2> + _LIBCPP_INLINE_VISIBILITY + _EnableIf<__is_transparent<hasher, _K2>::value && __is_transparent<key_equal, _K2>::value, pair<const_iterator, const_iterator>> + equal_range(const _K2& __k) const {return __table_.__equal_range_multi(__k);} + #endif // _LIBCPP_STD_VER > 17 _LIBCPP_INLINE_VISIBILITY size_type bucket_count() const _NOEXCEPT {return __table_.bucket_count();} @@ -1361,7 +1447,7 @@ public: _LIBCPP_INLINE_VISIBILITY void reserve(size_type __n) {__table_.reserve(__n);} -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const {return __table_.__dereferenceable(__i);} @@ -1372,7 +1458,7 @@ public: bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const {return __table_.__addable(__i, __n);} -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 }; @@ -1435,7 +1521,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( size_type __n, const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1447,7 +1533,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const allocator_type& __a) : __table_(__hf, __eql, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1458,7 +1544,7 @@ template <class _InputIterator> unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( _InputIterator __first, _InputIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__first, __last); @@ -1471,7 +1557,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const hasher& __hf, const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1485,7 +1571,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const hasher& __hf, const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1498,7 +1584,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const allocator_type& __a) : __table_(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -1508,7 +1594,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const unordered_multiset& __u) : __table_(__u.__table_) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -1520,7 +1606,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const unordered_multiset& __u, const allocator_type& __a) : __table_(__u.__table_, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__u.bucket_count()); @@ -1536,7 +1622,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( _NOEXCEPT_(is_nothrow_move_constructible<__table>::value) : __table_(_VSTD::move(__u.__table_)) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); __get_db()->swap(this, &__u); #endif @@ -1547,7 +1633,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( unordered_multiset&& __u, const allocator_type& __a) : __table_(_VSTD::move(__u.__table_), __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__a != __u.get_allocator()) @@ -1556,7 +1642,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( while (__u.size() != 0) __table_.__insert_multi(_VSTD::move(__u.__table_.remove(__i++)->__value_)); } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 else __get_db()->swap(this, &__u); #endif @@ -1566,7 +1652,7 @@ template <class _Value, class _Hash, class _Pred, class _Alloc> unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( initializer_list<value_type> __il) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif insert(__il.begin(), __il.end()); @@ -1578,7 +1664,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const key_equal& __eql) : __table_(__hf, __eql) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); @@ -1591,7 +1677,7 @@ unordered_multiset<_Value, _Hash, _Pred, _Alloc>::unordered_multiset( const key_equal& __eql, const allocator_type& __a) : __table_(__hf, __eql, __a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif __table_.rehash(__n); diff --git a/lib/libcxx/include/utility b/lib/libcxx/include/utility index 7ac322bfe7..6f27af7064 100644 --- a/lib/libcxx/include/utility +++ b/lib/libcxx/include/utility @@ -499,7 +499,7 @@ struct _LIBCPP_TEMPLATE_VIS pair second(_VSTD::get<1>(_VSTD::forward<_Tuple>(__p))) {} template <class... _Args1, class... _Args2> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 pair(piecewise_construct_t __pc, tuple<_Args1...> __first_args, tuple<_Args2...> __second_args) _NOEXCEPT_((is_nothrow_constructible<first_type, _Args1...>::value && @@ -508,7 +508,7 @@ struct _LIBCPP_TEMPLATE_VIS pair typename __make_tuple_indices<sizeof...(_Args1)>::type(), typename __make_tuple_indices<sizeof...(_Args2) >::type()) {} - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 pair& operator=(typename conditional< is_copy_assignable<first_type>::value && is_copy_assignable<second_type>::value, @@ -521,7 +521,7 @@ struct _LIBCPP_TEMPLATE_VIS pair return *this; } - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 pair& operator=(typename conditional< is_move_assignable<first_type>::value && is_move_assignable<second_type>::value, @@ -537,7 +537,7 @@ struct _LIBCPP_TEMPLATE_VIS pair template <class _Tuple, _EnableB< _CheckTLC<_Tuple>::template __enable_assign<_Tuple>() > = false> - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 pair& operator=(_Tuple&& __p) { first = _VSTD::get<0>(_VSTD::forward<_Tuple>(__p)); second = _VSTD::get<1>(_VSTD::forward<_Tuple>(__p)); @@ -545,7 +545,7 @@ struct _LIBCPP_TEMPLATE_VIS pair } #endif - _LIBCPP_INLINE_VISIBILITY + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 void swap(pair& __p) _NOEXCEPT_(__is_nothrow_swappable<first_type>::value && __is_nothrow_swappable<second_type>::value) @@ -558,10 +558,10 @@ private: #ifndef _LIBCPP_CXX03_LANG template <class... _Args1, class... _Args2, size_t... _I1, size_t... _I2> - _LIBCPP_INLINE_VISIBILITY - pair(piecewise_construct_t, - tuple<_Args1...>& __first_args, tuple<_Args2...>& __second_args, - __tuple_indices<_I1...>, __tuple_indices<_I2...>); + _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 + pair(piecewise_construct_t, + tuple<_Args1...>& __first_args, tuple<_Args2...>& __second_args, + __tuple_indices<_I1...>, __tuple_indices<_I2...>); #endif }; @@ -619,7 +619,7 @@ operator<=(const pair<_T1,_T2>& __x, const pair<_T1,_T2>& __y) } template <class _T1, class _T2> -inline _LIBCPP_INLINE_VISIBILITY +inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR_AFTER_CXX17 typename enable_if < __is_swappable<_T1>::value && @@ -970,7 +970,7 @@ _Size __loadword(const void* __p) { _Size __r; - std::memcpy(&__r, __p, sizeof(__r)); + _VSTD::memcpy(&__r, __p, sizeof(__r)); return __r; } @@ -1189,7 +1189,7 @@ __murmur2_or_cityhash<_Size, 64>::operator()(const void* __key, _Size __len) __v = __weak_hash_len_32_with_seeds(__s, __v.second * __k1, __x + __w.first); __w = __weak_hash_len_32_with_seeds(__s + 32, __z + __w.second, __y + __loadword<_Size>(__s + 16)); - std::swap(__z, __x); + _VSTD::swap(__z, __x); __s += 64; __len -= 64; } while (__len != 0); @@ -1364,6 +1364,16 @@ struct _LIBCPP_TEMPLATE_VIS hash<unsigned char> size_t operator()(unsigned char __v) const _NOEXCEPT {return static_cast<size_t>(__v);} }; +#ifndef _LIBCPP_NO_HAS_CHAR8_T +template <> +struct _LIBCPP_TEMPLATE_VIS hash<char8_t> + : public unary_function<char8_t, size_t> +{ + _LIBCPP_INLINE_VISIBILITY + size_t operator()(char8_t __v) const _NOEXCEPT {return static_cast<size_t>(__v);} +}; +#endif // !_LIBCPP_NO_HAS_CHAR8_T + #ifndef _LIBCPP_HAS_NO_UNICODE_CHARS template <> @@ -1506,7 +1516,7 @@ struct _LIBCPP_TEMPLATE_VIS hash<long double> // -0.0 and 0.0 should return same hash if (__v == 0.0L) return 0; -#if defined(__i386__) +#if defined(__i386__) || (defined(__x86_64__) && defined(__ILP32__)) // Zero out padding bits union { @@ -1593,7 +1603,7 @@ using __check_hash_requirements _LIBCPP_NODEBUG_TYPE = integral_constant<bool, __invokable_r<size_t, _Hash, _Key const&>::value >; -template <class _Key, class _Hash = std::hash<_Key> > +template <class _Key, class _Hash = hash<_Key> > using __has_enabled_hash _LIBCPP_NODEBUG_TYPE = integral_constant<bool, __check_hash_requirements<_Key, _Hash>::value && is_default_constructible<_Hash>::value diff --git a/lib/libcxx/include/valarray b/lib/libcxx/include/valarray index c048a6d7e4..787d8aca2f 100644 --- a/lib/libcxx/include/valarray +++ b/lib/libcxx/include/valarray @@ -136,6 +136,7 @@ public: void operator>>=(const valarray<value_type>& v) const; void operator=(const value_type& x) const; + void operator=(const valarray<T>& val_arr) const; slice_array() = delete; }; @@ -802,7 +803,7 @@ private: public: // construct/destroy: _LIBCPP_INLINE_VISIBILITY - valarray() : __begin_(0), __end_(0) {} + valarray() : __begin_(nullptr), __end_(nullptr) {} inline _LIBCPP_HIDE_FROM_ABI_AFTER_V1 explicit valarray(size_t __n); _LIBCPP_INLINE_VISIBILITY @@ -1264,6 +1265,9 @@ public: _LIBCPP_INLINE_VISIBILITY void operator=(const value_type& __x) const; + _LIBCPP_INLINE_VISIBILITY + void operator=(const valarray<value_type>& __va) const; + private: _LIBCPP_INLINE_VISIBILITY slice_array(const slice& __sl, const valarray<value_type>& __v) @@ -1304,6 +1308,15 @@ slice_array<_Tp>::operator=(const _Expr& __v) const } template <class _Tp> +inline void +slice_array<_Tp>::operator=(const valarray<value_type>& __va) const +{ + value_type* __t = __vp_; + for (size_t __i = 0; __i < __va.size(); ++__i, __t += __stride_) + *__t = __va[__i]; +} + +template <class _Tp> template <class _Expr> inline typename enable_if @@ -2738,11 +2751,9 @@ __val_expr<_ValExpr>::operator valarray<__val_expr::result_type>() const if (__n) { __r.__begin_ = - __r.__end_ = - static_cast<result_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(result_type), _LIBCPP_ALIGNOF(result_type))); + __r.__end_ = allocator<result_type>().allocate(__n); for (size_t __i = 0; __i != __n; ++__r.__end_, ++__i) - ::new (__r.__end_) result_type(__expr_[__i]); + ::new ((void*)__r.__end_) result_type(__expr_[__i]); } return __r; } @@ -2752,19 +2763,18 @@ __val_expr<_ValExpr>::operator valarray<__val_expr::result_type>() const template <class _Tp> inline valarray<_Tp>::valarray(size_t __n) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { if (__n) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { #endif // _LIBCPP_NO_EXCEPTIONS for (size_t __n_left = __n; __n_left; --__n_left, ++__end_) - ::new (__end_) value_type(); + ::new ((void*)__end_) value_type(); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -2779,27 +2789,26 @@ valarray<_Tp>::valarray(size_t __n) template <class _Tp> inline valarray<_Tp>::valarray(const value_type& __x, size_t __n) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { resize(__n, __x); } template <class _Tp> valarray<_Tp>::valarray(const value_type* __p, size_t __n) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { if (__n) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { #endif // _LIBCPP_NO_EXCEPTIONS for (size_t __n_left = __n; __n_left; ++__end_, ++__p, --__n_left) - ::new (__end_) value_type(*__p); + ::new ((void*)__end_) value_type(*__p); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -2813,19 +2822,18 @@ valarray<_Tp>::valarray(const value_type* __p, size_t __n) template <class _Tp> valarray<_Tp>::valarray(const valarray& __v) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { if (__v.size()) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__v.size() * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__v.size()); #ifndef _LIBCPP_NO_EXCEPTIONS try { #endif // _LIBCPP_NO_EXCEPTIONS for (value_type* __p = __v.__begin_; __p != __v.__end_; ++__end_, ++__p) - ::new (__end_) value_type(*__p); + ::new ((void*)__end_) value_type(*__p); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -2850,21 +2858,20 @@ valarray<_Tp>::valarray(valarray&& __v) _NOEXCEPT template <class _Tp> valarray<_Tp>::valarray(initializer_list<value_type> __il) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { const size_t __n = __il.size(); if (__n) { - __begin_ = __end_ = static_cast<value_type*>( -_VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { #endif // _LIBCPP_NO_EXCEPTIONS size_t __n_left = __n; for (const value_type* __p = __il.begin(); __n_left; ++__end_, ++__p, --__n_left) - ::new (__end_) value_type(*__p); + ::new ((void*)__end_) value_type(*__p); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -2880,21 +2887,20 @@ _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))) template <class _Tp> valarray<_Tp>::valarray(const slice_array<value_type>& __sa) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { const size_t __n = __sa.__size_; if (__n) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { #endif // _LIBCPP_NO_EXCEPTIONS size_t __n_left = __n; for (const value_type* __p = __sa.__vp_; __n_left; ++__end_, __p += __sa.__stride_, --__n_left) - ::new (__end_) value_type(*__p); + ::new ((void*)__end_) value_type(*__p); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -2908,14 +2914,13 @@ valarray<_Tp>::valarray(const slice_array<value_type>& __sa) template <class _Tp> valarray<_Tp>::valarray(const gslice_array<value_type>& __ga) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { const size_t __n = __ga.__1d_.size(); if (__n) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { @@ -2924,7 +2929,7 @@ valarray<_Tp>::valarray(const gslice_array<value_type>& __ga) const value_type* __s = __ga.__vp_; for (_Ip __i = __ga.__1d_.__begin_, __e = __ga.__1d_.__end_; __i != __e; ++__i, ++__end_) - ::new (__end_) value_type(__s[*__i]); + ::new ((void*)__end_) value_type(__s[*__i]); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -2938,14 +2943,13 @@ valarray<_Tp>::valarray(const gslice_array<value_type>& __ga) template <class _Tp> valarray<_Tp>::valarray(const mask_array<value_type>& __ma) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { const size_t __n = __ma.__1d_.size(); if (__n) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { @@ -2954,7 +2958,7 @@ valarray<_Tp>::valarray(const mask_array<value_type>& __ma) const value_type* __s = __ma.__vp_; for (_Ip __i = __ma.__1d_.__begin_, __e = __ma.__1d_.__end_; __i != __e; ++__i, ++__end_) - ::new (__end_) value_type(__s[*__i]); + ::new ((void*)__end_) value_type(__s[*__i]); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -2968,14 +2972,13 @@ valarray<_Tp>::valarray(const mask_array<value_type>& __ma) template <class _Tp> valarray<_Tp>::valarray(const indirect_array<value_type>& __ia) - : __begin_(0), - __end_(0) + : __begin_(nullptr), + __end_(nullptr) { const size_t __n = __ia.__1d_.size(); if (__n) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { @@ -2984,7 +2987,7 @@ valarray<_Tp>::valarray(const indirect_array<value_type>& __ia) const value_type* __s = __ia.__vp_; for (_Ip __i = __ia.__1d_.__begin_, __e = __ia.__1d_.__end_; __i != __e; ++__i, ++__end_) - ::new (__end_) value_type(__s[*__i]); + ::new ((void*)__end_) value_type(__s[*__i]); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) @@ -3011,8 +3014,7 @@ valarray<_Tp>::__assign_range(const value_type* __f, const value_type* __l) if (size() != __n) { __clear(size()); - __begin_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = allocator<value_type>().allocate(__n); __end_ = __begin_ + __n; _VSTD::uninitialized_copy(__f, __l, __begin_); } else { @@ -3265,12 +3267,9 @@ valarray<_Tp>::operator+() const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __r.__begin_ = __r.__end_ = allocator<value_type>().allocate(__n); for (const value_type* __p = __begin_; __n; ++__r.__end_, ++__p, --__n) - ::new (__r.__end_) value_type(+*__p); + ::new ((void*)__r.__end_) value_type(+*__p); } return __r; } @@ -3283,12 +3282,9 @@ valarray<_Tp>::operator-() const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __r.__begin_ = __r.__end_ = allocator<value_type>().allocate(__n); for (const value_type* __p = __begin_; __n; ++__r.__end_, ++__p, --__n) - ::new (__r.__end_) value_type(-*__p); + ::new ((void*)__r.__end_) value_type(-*__p); } return __r; } @@ -3301,12 +3297,9 @@ valarray<_Tp>::operator~() const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __r.__begin_ = __r.__end_ = allocator<value_type>().allocate(__n); for (const value_type* __p = __begin_; __n; ++__r.__end_, ++__p, --__n) - ::new (__r.__end_) value_type(~*__p); + ::new ((void*)__r.__end_) value_type(~*__p); } return __r; } @@ -3319,11 +3312,9 @@ valarray<_Tp>::operator!() const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<bool*>(_VSTD::__libcpp_allocate(__n * sizeof(bool), _LIBCPP_ALIGNOF(bool))); + __r.__begin_ = __r.__end_ = allocator<bool>().allocate(__n); for (const value_type* __p = __begin_; __n; ++__r.__end_, ++__p, --__n) - ::new (__r.__end_) bool(!*__p); + ::new ((void*)__r.__end_) bool(!*__p); } return __r; } @@ -3639,10 +3630,7 @@ valarray<_Tp>::shift(int __i) const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __r.__begin_ = __r.__end_ = allocator<value_type>().allocate(__n); const value_type* __sb; value_type* __tb; value_type* __te; @@ -3661,11 +3649,11 @@ valarray<_Tp>::shift(int __i) const __te = __r.__begin_ + __n; } for (; __r.__end_ != __tb; ++__r.__end_) - ::new (__r.__end_) value_type(); + ::new ((void*)__r.__end_) value_type(); for (; __r.__end_ != __te; ++__r.__end_, ++__sb) - ::new (__r.__end_) value_type(*__sb); + ::new ((void*)__r.__end_) value_type(*__sb); for (__te = __r.__begin_ + __n; __r.__end_ != __te; ++__r.__end_) - ::new (__r.__end_) value_type(); + ::new ((void*)__r.__end_) value_type(); } return __r; } @@ -3678,16 +3666,13 @@ valarray<_Tp>::cshift(int __i) const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __r.__begin_ = __r.__end_ = allocator<value_type>().allocate(__n); __i %= static_cast<int>(__n); const value_type* __m = __i >= 0 ? __begin_ + __i : __end_ + __i; for (const value_type* __s = __m; __s != __end_; ++__r.__end_, ++__s) - ::new (__r.__end_) value_type(*__s); + ::new ((void*)__r.__end_) value_type(*__s); for (const value_type* __s = __begin_; __s != __m; ++__r.__end_, ++__s) - ::new (__r.__end_) value_type(*__s); + ::new ((void*)__r.__end_) value_type(*__s); } return __r; } @@ -3700,12 +3685,9 @@ valarray<_Tp>::apply(value_type __f(value_type)) const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __r.__begin_ = __r.__end_ = allocator<value_type>().allocate(__n); for (const value_type* __p = __begin_; __n; ++__r.__end_, ++__p, --__n) - ::new (__r.__end_) value_type(__f(*__p)); + ::new ((void*)__r.__end_) value_type(__f(*__p)); } return __r; } @@ -3718,12 +3700,9 @@ valarray<_Tp>::apply(value_type __f(const value_type&)) const size_t __n = size(); if (__n) { - __r.__begin_ = - __r.__end_ = - static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __r.__begin_ = __r.__end_ = allocator<value_type>().allocate(__n); for (const value_type* __p = __begin_; __n; ++__r.__end_, ++__p, --__n) - ::new (__r.__end_) value_type(__f(*__p)); + ::new ((void*)__r.__end_) value_type(__f(*__p)); } return __r; } @@ -3736,7 +3715,7 @@ void valarray<_Tp>::__clear(size_t __capacity) { while (__end_ != __begin_) (--__end_)->~value_type(); - _VSTD::__libcpp_deallocate(__begin_, __capacity * sizeof(value_type), _LIBCPP_ALIGNOF(value_type)); + allocator<value_type>().deallocate(__begin_, __capacity); __begin_ = __end_ = nullptr; } } @@ -3748,14 +3727,13 @@ valarray<_Tp>::resize(size_t __n, value_type __x) __clear(size()); if (__n) { - __begin_ = __end_ = static_cast<value_type*>( - _VSTD::__libcpp_allocate(__n * sizeof(value_type), _LIBCPP_ALIGNOF(value_type))); + __begin_ = __end_ = allocator<value_type>().allocate(__n); #ifndef _LIBCPP_NO_EXCEPTIONS try { #endif // _LIBCPP_NO_EXCEPTIONS for (size_t __n_left = __n; __n_left; --__n_left, ++__end_) - ::new (__end_) value_type(__x); + ::new ((void*)__end_) value_type(__x); #ifndef _LIBCPP_NO_EXCEPTIONS } catch (...) diff --git a/lib/libcxx/include/variant b/lib/libcxx/include/variant index 03557239a6..770dd335ba 100644 --- a/lib/libcxx/include/variant +++ b/lib/libcxx/include/variant @@ -169,6 +169,9 @@ namespace std { template <class Visitor, class... Variants> constexpr see below visit(Visitor&&, Variants&&...); + template <class R, class Visitor, class... Variants> + constexpr R visit(Visitor&&, Variants&&...); // since C++20 + // 20.7.7, class monostate struct monostate; @@ -197,6 +200,7 @@ namespace std { */ #include <__config> +#include <__availability> #include <__tuple> #include <array> #include <exception> @@ -227,7 +231,10 @@ public: _LIBCPP_BEGIN_NAMESPACE_STD -#if _LIBCPP_STD_VER > 14 +// TODO: GCC 5 lies about its support for C++17 (it says it supports it but it +// really doesn't). That breaks variant, which uses some C++17 features. +// Remove this once we drop support for GCC 5. +#if _LIBCPP_STD_VER > 14 && !(defined(_LIBCPP_COMPILER_GCC) && _GNUC_VER_NEW < 6000) _LIBCPP_NORETURN inline _LIBCPP_INLINE_VISIBILITY @@ -290,9 +297,9 @@ struct _LIBCPP_TEMPLATE_VIS variant_alternative<_Ip, variant<_Types...>> { _LIBCPP_INLINE_VAR constexpr size_t variant_npos = static_cast<size_t>(-1); constexpr int __choose_index_type(unsigned int __num_elem) { - if (__num_elem < std::numeric_limits<unsigned char>::max()) + if (__num_elem < numeric_limits<unsigned char>::max()) return 0; - if (__num_elem < std::numeric_limits<unsigned short>::max()) + if (__num_elem < numeric_limits<unsigned short>::max()) return 1; return 2; } @@ -482,12 +489,12 @@ private: return __result{{_VSTD::forward<_Fs>(__fs)...}}; } - template <std::size_t... _Is> + template <size_t... _Is> struct __dispatcher { template <class _Fp, class... _Vs> inline _LIBCPP_INLINE_VISIBILITY static constexpr decltype(auto) __dispatch(_Fp __f, _Vs... __vs) { - return __invoke_constexpr( + return _VSTD::__invoke_constexpr( static_cast<_Fp>(__f), __access::__base::__get_alt<_Is>(static_cast<_Vs>(__vs))...); } @@ -579,6 +586,16 @@ struct __variant { __make_value_visitor(_VSTD::forward<_Visitor>(__visitor)), _VSTD::forward<_Vs>(__vs)...); } +#if _LIBCPP_STD_VER > 17 + template <class _Rp, class _Visitor, class... _Vs> + inline _LIBCPP_INLINE_VISIBILITY + static constexpr _Rp __visit_value(_Visitor&& __visitor, + _Vs&&... __vs) { + return __visit_alt( + __make_value_visitor<_Rp>(_VSTD::forward<_Visitor>(__visitor)), + _VSTD::forward<_Vs>(__vs)...); + } +#endif private: template <class _Visitor, class... _Values> @@ -595,17 +612,48 @@ private: __std_visit_exhaustive_visitor_check< _Visitor, decltype((_VSTD::forward<_Alts>(__alts).__value))...>(); - return __invoke_constexpr(_VSTD::forward<_Visitor>(__visitor), + return _VSTD::__invoke_constexpr(_VSTD::forward<_Visitor>(__visitor), _VSTD::forward<_Alts>(__alts).__value...); } _Visitor&& __visitor; }; +#if _LIBCPP_STD_VER > 17 + template <class _Rp, class _Visitor> + struct __value_visitor_return_type { + template <class... _Alts> + inline _LIBCPP_INLINE_VISIBILITY + constexpr _Rp operator()(_Alts&&... __alts) const { + __std_visit_exhaustive_visitor_check< + _Visitor, + decltype((_VSTD::forward<_Alts>(__alts).__value))...>(); + if constexpr (is_void_v<_Rp>) { + _VSTD::__invoke_constexpr(_VSTD::forward<_Visitor>(__visitor), + _VSTD::forward<_Alts>(__alts).__value...); + } + else { + return _VSTD::__invoke_constexpr(_VSTD::forward<_Visitor>(__visitor), + _VSTD::forward<_Alts>(__alts).__value...); + } + } + + _Visitor&& __visitor; + }; +#endif + template <class _Visitor> inline _LIBCPP_INLINE_VISIBILITY static constexpr auto __make_value_visitor(_Visitor&& __visitor) { return __value_visitor<_Visitor>{_VSTD::forward<_Visitor>(__visitor)}; } + +#if _LIBCPP_STD_VER > 17 + template <class _Rp, class _Visitor> + inline _LIBCPP_INLINE_VISIBILITY + static constexpr auto __make_value_visitor(_Visitor&& __visitor) { + return __value_visitor_return_type<_Rp, _Visitor>{_VSTD::forward<_Visitor>(__visitor)}; + } +#endif }; } // namespace __visitation @@ -720,12 +768,12 @@ protected: }; template <class _Traits, _Trait = _Traits::__destructible_trait> -class _LIBCPP_TEMPLATE_VIS __destructor; +class _LIBCPP_TEMPLATE_VIS __dtor; #define _LIBCPP_VARIANT_DESTRUCTOR(destructible_trait, destructor, destroy) \ template <class... _Types> \ - class _LIBCPP_TEMPLATE_VIS __destructor<__traits<_Types...>, \ - destructible_trait> \ + class _LIBCPP_TEMPLATE_VIS __dtor<__traits<_Types...>, \ + destructible_trait> \ : public __base<destructible_trait, _Types...> { \ using __base_type = __base<destructible_trait, _Types...>; \ using __index_t = typename __base_type::__index_t; \ @@ -734,11 +782,11 @@ class _LIBCPP_TEMPLATE_VIS __destructor; using __base_type::__base_type; \ using __base_type::operator=; \ \ - __destructor(const __destructor&) = default; \ - __destructor(__destructor&&) = default; \ + __dtor(const __dtor&) = default; \ + __dtor(__dtor&&) = default; \ destructor \ - __destructor& operator=(const __destructor&) = default; \ - __destructor& operator=(__destructor&&) = default; \ + __dtor& operator=(const __dtor&) = default; \ + __dtor& operator=(__dtor&&) = default; \ \ protected: \ inline _LIBCPP_INLINE_VISIBILITY \ @@ -747,12 +795,12 @@ class _LIBCPP_TEMPLATE_VIS __destructor; _LIBCPP_VARIANT_DESTRUCTOR( _Trait::_TriviallyAvailable, - ~__destructor() = default;, + ~__dtor() = default;, void __destroy() noexcept { this->__index = __variant_npos<__index_t>; }); _LIBCPP_VARIANT_DESTRUCTOR( _Trait::_Available, - ~__destructor() { __destroy(); }, + ~__dtor() { __destroy(); }, void __destroy() noexcept { if (!this->valueless_by_exception()) { __visitation::__base::__visit_alt( @@ -767,14 +815,14 @@ _LIBCPP_VARIANT_DESTRUCTOR( _LIBCPP_VARIANT_DESTRUCTOR( _Trait::_Unavailable, - ~__destructor() = delete;, + ~__dtor() = delete;, void __destroy() noexcept = delete;); #undef _LIBCPP_VARIANT_DESTRUCTOR template <class _Traits> -class _LIBCPP_TEMPLATE_VIS __constructor : public __destructor<_Traits> { - using __base_type = __destructor<_Traits>; +class _LIBCPP_TEMPLATE_VIS __ctor : public __dtor<_Traits> { + using __base_type = __dtor<_Traits>; public: using __base_type::__base_type; @@ -791,7 +839,7 @@ protected: template <class _Rhs> inline _LIBCPP_INLINE_VISIBILITY - static void __generic_construct(__constructor& __lhs, _Rhs&& __rhs) { + static void __generic_construct(__ctor& __lhs, _Rhs&& __rhs) { __lhs.__destroy(); if (!__rhs.valueless_by_exception()) { __visitation::__base::__visit_alt_at( @@ -813,10 +861,10 @@ class _LIBCPP_TEMPLATE_VIS __move_constructor; #define _LIBCPP_VARIANT_MOVE_CONSTRUCTOR(move_constructible_trait, \ move_constructor) \ template <class... _Types> \ - class _LIBCPP_TEMPLATE_VIS __move_constructor<__traits<_Types...>, \ - move_constructible_trait> \ - : public __constructor<__traits<_Types...>> { \ - using __base_type = __constructor<__traits<_Types...>>; \ + class _LIBCPP_TEMPLATE_VIS __move_constructor<__traits<_Types...>, \ + move_constructible_trait> \ + : public __ctor<__traits<_Types...>> { \ + using __base_type = __ctor<__traits<_Types...>>; \ \ public: \ using __base_type::__base_type; \ @@ -1104,7 +1152,7 @@ struct __narrowing_check { template <class _Dest> static auto __test_impl(_Dest (&&)[1]) -> __identity<_Dest>; template <class _Dest, class _Source> - using _Apply _LIBCPP_NODEBUG_TYPE = decltype(__test_impl<_Dest>({std::declval<_Source>()})); + using _Apply _LIBCPP_NODEBUG_TYPE = decltype(__test_impl<_Dest>({_VSTD::declval<_Source>()})); }; template <class _Dest, class _Source> @@ -1510,7 +1558,7 @@ template <class _Operator> struct __convert_to_bool { template <class _T1, class _T2> _LIBCPP_INLINE_VISIBILITY constexpr bool operator()(_T1 && __t1, _T2&& __t2) const { - static_assert(std::is_convertible<decltype(_Operator{}(_VSTD::forward<_T1>(__t1), _VSTD::forward<_T2>(__t2))), bool>::value, + static_assert(is_convertible<decltype(_Operator{}(_VSTD::forward<_T1>(__t1), _VSTD::forward<_T2>(__t2))), bool>::value, "the relational operator does not return a type which is implicitly convertible to bool"); return _Operator{}(_VSTD::forward<_T1>(__t1), _VSTD::forward<_T2>(__t2)); } @@ -1587,21 +1635,38 @@ constexpr bool operator>=(const variant<_Types...>& __lhs, __lhs.index(), __convert_to_bool<greater_equal<>>{}, __lhs, __rhs); } +template <class... _Vs> +inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_AVAILABILITY_THROW_BAD_VARIANT_ACCESS +constexpr void __throw_if_valueless(_Vs&&... __vs) { + const bool __valueless = (... || __vs.valueless_by_exception()); + if (__valueless) { + __throw_bad_variant_access(); + } +} + template <class _Visitor, class... _Vs> inline _LIBCPP_INLINE_VISIBILITY _LIBCPP_AVAILABILITY_THROW_BAD_VARIANT_ACCESS constexpr decltype(auto) visit(_Visitor&& __visitor, _Vs&&... __vs) { using __variant_detail::__visitation::__variant; - bool __results[] = {__vs.valueless_by_exception()...}; - for (bool __result : __results) { - if (__result) { - __throw_bad_variant_access(); - } - } + _VSTD::__throw_if_valueless(_VSTD::forward<_Vs>(__vs)...); return __variant::__visit_value(_VSTD::forward<_Visitor>(__visitor), _VSTD::forward<_Vs>(__vs)...); } +#if _LIBCPP_STD_VER > 17 +template <class _Rp, class _Visitor, class... _Vs> +inline _LIBCPP_INLINE_VISIBILITY +_LIBCPP_AVAILABILITY_THROW_BAD_VARIANT_ACCESS +constexpr _Rp visit(_Visitor&& __visitor, _Vs&&... __vs) { + using __variant_detail::__visitation::__variant; + _VSTD::__throw_if_valueless(_VSTD::forward<_Vs>(__vs)...); + return __variant::__visit_value<_Rp>(_VSTD::forward<_Visitor>(__visitor), + _VSTD::forward<_Vs>(__vs)...); +} +#endif + struct _LIBCPP_TEMPLATE_VIS monostate {}; inline _LIBCPP_INLINE_VISIBILITY diff --git a/lib/libcxx/include/vector b/lib/libcxx/include/vector index 1007beeaaf..8e2df79f92 100644 --- a/lib/libcxx/include/vector +++ b/lib/libcxx/include/vector @@ -454,7 +454,7 @@ inline _LIBCPP_INLINE_VISIBILITY __vector_base<_Tp, _Allocator>::__vector_base(allocator_type&& __a) _NOEXCEPT : __begin_(nullptr), __end_(nullptr), - __end_cap_(nullptr, std::move(__a)) {} + __end_cap_(nullptr, _VSTD::move(__a)) {} #endif template <class _Tp, class _Allocator> @@ -496,7 +496,7 @@ public: _LIBCPP_INLINE_VISIBILITY vector() _NOEXCEPT_(is_nothrow_default_constructible<allocator_type>::value) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -508,7 +508,7 @@ public: #endif : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif } @@ -551,7 +551,7 @@ public: ~vector() { __annotate_delete(); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__erase_c(this); #endif } @@ -789,14 +789,14 @@ public: bool __invariants() const; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 bool __dereferenceable(const const_iterator* __i) const; bool __decrementable(const const_iterator* __i) const; bool __addable(const const_iterator* __i, ptrdiff_t __n) const; bool __subscriptable(const const_iterator* __i, ptrdiff_t __n) const; -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 private: _LIBCPP_INLINE_VISIBILITY void __invalidate_all_iterators(); @@ -931,7 +931,7 @@ private: #ifndef _LIBCPP_HAS_NO_DEDUCTION_GUIDES template<class _InputIterator, - class _Alloc = typename std::allocator<typename iterator_traits<_InputIterator>::value_type>, + class _Alloc = allocator<typename iterator_traits<_InputIterator>::value_type>, class = typename enable_if<__is_allocator<_Alloc>::value, void>::type > vector(_InputIterator, _InputIterator) @@ -951,8 +951,7 @@ vector<_Tp, _Allocator>::__swap_out_circular_buffer(__split_buffer<value_type, a { __annotate_delete(); - __alloc_traits::__construct_backward_with_exception_guarantees( - this->__alloc(), this->__begin_, this->__end_, __v.__begin_); + _VSTD::__construct_backward_with_exception_guarantees(this->__alloc(), this->__begin_, this->__end_, __v.__begin_); _VSTD::swap(this->__begin_, __v.__begin_); _VSTD::swap(this->__end_, __v.__end_); _VSTD::swap(this->__end_cap(), __v.__end_cap()); @@ -967,10 +966,8 @@ vector<_Tp, _Allocator>::__swap_out_circular_buffer(__split_buffer<value_type, a { __annotate_delete(); pointer __r = __v.__begin_; - __alloc_traits::__construct_backward_with_exception_guarantees( - this->__alloc(), this->__begin_, __p, __v.__begin_); - __alloc_traits::__construct_forward_with_exception_guarantees( - this->__alloc(), __p, this->__end_, __v.__end_); + _VSTD::__construct_backward_with_exception_guarantees(this->__alloc(), this->__begin_, __p, __v.__begin_); + _VSTD::__construct_forward_with_exception_guarantees(this->__alloc(), __p, this->__end_, __v.__end_); _VSTD::swap(this->__begin_, __v.__begin_); _VSTD::swap(this->__end_, __v.__end_); _VSTD::swap(this->__end_cap(), __v.__end_cap()); @@ -1077,7 +1074,7 @@ typename enable_if vector<_Tp, _Allocator>::__construct_at_end(_ForwardIterator __first, _ForwardIterator __last, size_type __n) { _ConstructTransaction __tx(*this, __n); - __alloc_traits::__construct_range_forward(this->__alloc(), __first, __last, __tx.__pos_); + _VSTD::__construct_range_forward(this->__alloc(), __first, __last, __tx.__pos_); } // Default constructs __n objects starting at __end_ @@ -1121,7 +1118,7 @@ vector<_Tp, _Allocator>::__append(size_type __n, const_reference __x) template <class _Tp, class _Allocator> vector<_Tp, _Allocator>::vector(size_type __n) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__n > 0) @@ -1136,7 +1133,7 @@ template <class _Tp, class _Allocator> vector<_Tp, _Allocator>::vector(size_type __n, const allocator_type& __a) : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__n > 0) @@ -1150,7 +1147,7 @@ vector<_Tp, _Allocator>::vector(size_type __n, const allocator_type& __a) template <class _Tp, class _Allocator> vector<_Tp, _Allocator>::vector(size_type __n, const value_type& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__n > 0) @@ -1164,7 +1161,7 @@ template <class _Tp, class _Allocator> vector<_Tp, _Allocator>::vector(size_type __n, const value_type& __x, const allocator_type& __a) : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__n > 0) @@ -1184,7 +1181,7 @@ vector<_Tp, _Allocator>::vector(_InputIterator __first, typename iterator_traits<_InputIterator>::reference>::value, _InputIterator>::type __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __first != __last; ++__first) @@ -1201,7 +1198,7 @@ vector<_Tp, _Allocator>::vector(_InputIterator __first, _InputIterator __last, c typename iterator_traits<_InputIterator>::reference>::value>::type*) : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif for (; __first != __last; ++__first) @@ -1217,7 +1214,7 @@ vector<_Tp, _Allocator>::vector(_ForwardIterator __first, typename iterator_traits<_ForwardIterator>::reference>::value, _ForwardIterator>::type __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif size_type __n = static_cast<size_type>(_VSTD::distance(__first, __last)); @@ -1237,7 +1234,7 @@ vector<_Tp, _Allocator>::vector(_ForwardIterator __first, _ForwardIterator __las typename iterator_traits<_ForwardIterator>::reference>::value>::type*) : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif size_type __n = static_cast<size_type>(_VSTD::distance(__first, __last)); @@ -1252,7 +1249,7 @@ template <class _Tp, class _Allocator> vector<_Tp, _Allocator>::vector(const vector& __x) : __base(__alloc_traits::select_on_container_copy_construction(__x.__alloc())) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif size_type __n = __x.size(); @@ -1267,7 +1264,7 @@ template <class _Tp, class _Allocator> vector<_Tp, _Allocator>::vector(const vector& __x, const allocator_type& __a) : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif size_type __n = __x.size(); @@ -1290,7 +1287,7 @@ vector<_Tp, _Allocator>::vector(vector&& __x) #endif : __base(_VSTD::move(__x.__alloc())) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); __get_db()->swap(this, &__x); #endif @@ -1305,7 +1302,7 @@ inline _LIBCPP_INLINE_VISIBILITY vector<_Tp, _Allocator>::vector(vector&& __x, const allocator_type& __a) : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__a == __x.__alloc()) @@ -1314,7 +1311,7 @@ vector<_Tp, _Allocator>::vector(vector&& __x, const allocator_type& __a) this->__end_ = __x.__end_; this->__end_cap() = __x.__end_cap(); __x.__begin_ = __x.__end_ = __x.__end_cap() = nullptr; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->swap(this, &__x); #endif } @@ -1329,7 +1326,7 @@ template <class _Tp, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY vector<_Tp, _Allocator>::vector(initializer_list<value_type> __il) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__il.size() > 0) @@ -1344,7 +1341,7 @@ inline _LIBCPP_INLINE_VISIBILITY vector<_Tp, _Allocator>::vector(initializer_list<value_type> __il, const allocator_type& __a) : __base(__a) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__insert_c(this); #endif if (__il.size() > 0) @@ -1390,7 +1387,7 @@ vector<_Tp, _Allocator>::__move_assign(vector& __c, true_type) this->__end_ = __c.__end_; this->__end_cap() = __c.__end_cap(); __c.__begin_ = __c.__end_ = __c.__end_cap() = nullptr; -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->swap(this, &__c); #endif } @@ -1493,7 +1490,7 @@ inline _LIBCPP_INLINE_VISIBILITY typename vector<_Tp, _Allocator>::iterator vector<_Tp, _Allocator>::__make_iter(pointer __p) _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return iterator(this, __p); #else return iterator(__p); @@ -1505,7 +1502,7 @@ inline _LIBCPP_INLINE_VISIBILITY typename vector<_Tp, _Allocator>::const_iterator vector<_Tp, _Allocator>::__make_iter(const_pointer __p) const _NOEXCEPT { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 return const_iterator(this, __p); #else return const_iterator(__p); @@ -1709,7 +1706,7 @@ inline _LIBCPP_INLINE_VISIBILITY typename vector<_Tp, _Allocator>::iterator vector<_Tp, _Allocator>::erase(const_iterator __position) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__position) == this, "vector::erase(iterator) called with an iterator not" " referring to this vector"); @@ -1728,7 +1725,7 @@ template <class _Tp, class _Allocator> typename vector<_Tp, _Allocator>::iterator vector<_Tp, _Allocator>::erase(const_iterator __first, const_iterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__first) == this, "vector::erase(iterator, iterator) called with an iterator not" " referring to this vector"); @@ -1769,7 +1766,7 @@ template <class _Tp, class _Allocator> typename vector<_Tp, _Allocator>::iterator vector<_Tp, _Allocator>::insert(const_iterator __position, const_reference __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__position) == this, "vector::insert(iterator, x) called with an iterator not" " referring to this vector"); @@ -1806,7 +1803,7 @@ template <class _Tp, class _Allocator> typename vector<_Tp, _Allocator>::iterator vector<_Tp, _Allocator>::insert(const_iterator __position, value_type&& __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__position) == this, "vector::insert(iterator, x) called with an iterator not" " referring to this vector"); @@ -1839,7 +1836,7 @@ template <class... _Args> typename vector<_Tp, _Allocator>::iterator vector<_Tp, _Allocator>::emplace(const_iterator __position, _Args&&... __args) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__position) == this, "vector::emplace(iterator, x) called with an iterator not" " referring to this vector"); @@ -1874,7 +1871,7 @@ template <class _Tp, class _Allocator> typename vector<_Tp, _Allocator>::iterator vector<_Tp, _Allocator>::insert(const_iterator __position, size_type __n, const_reference __x) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__position) == this, "vector::insert(iterator, n, x) called with an iterator not" " referring to this vector"); @@ -1925,7 +1922,7 @@ typename enable_if >::type vector<_Tp, _Allocator>::insert(const_iterator __position, _InputIterator __first, _InputIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__position) == this, "vector::insert(iterator, range) called with an iterator not" " referring to this vector"); @@ -1978,7 +1975,7 @@ typename enable_if >::type vector<_Tp, _Allocator>::insert(const_iterator __position, _ForwardIterator __first, _ForwardIterator __last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 _LIBCPP_ASSERT(__get_const_db()->__find_c_from_i(&__position) == this, "vector::insert(iterator, range) called with an iterator not" " referring to this vector"); @@ -2057,11 +2054,11 @@ vector<_Tp, _Allocator>::swap(vector& __x) _VSTD::swap(this->__begin_, __x.__begin_); _VSTD::swap(this->__end_, __x.__end_); _VSTD::swap(this->__end_cap(), __x.__end_cap()); - __swap_allocator(this->__alloc(), __x.__alloc(), + _VSTD::__swap_allocator(this->__alloc(), __x.__alloc(), integral_constant<bool,__alloc_traits::propagate_on_container_swap::value>()); -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->swap(this, &__x); -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif } template <class _Tp, class _Allocator> @@ -2085,7 +2082,7 @@ vector<_Tp, _Allocator>::__invariants() const return true; } -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 template <class _Tp, class _Allocator> bool @@ -2117,16 +2114,16 @@ vector<_Tp, _Allocator>::__subscriptable(const const_iterator* __i, ptrdiff_t __ return this->__begin_ <= __p && __p < this->__end_; } -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif // _LIBCPP_DEBUG_LEVEL == 2 template <class _Tp, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY void vector<_Tp, _Allocator>::__invalidate_all_iterators() { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __get_db()->__invalidate_all(this); -#endif // _LIBCPP_DEBUG_LEVEL >= 2 +#endif } @@ -2134,7 +2131,7 @@ template <class _Tp, class _Allocator> inline _LIBCPP_INLINE_VISIBILITY void vector<_Tp, _Allocator>::__invalidate_iterators_past(pointer __new_last) { -#if _LIBCPP_DEBUG_LEVEL >= 2 +#if _LIBCPP_DEBUG_LEVEL == 2 __c_node* __c = __get_db()->__find_c_and_lock(this); for (__i_node** __p = __c->end_; __p != __c->beg_; ) { --__p; @@ -2142,7 +2139,7 @@ vector<_Tp, _Allocator>::__invalidate_iterators_past(pointer __new_last) { if (__i->base() > __new_last) { (*__p)->__c_ = nullptr; if (--__c->end_ != __p) - memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); + _VSTD::memmove(__p, __p+1, (__c->end_ - __p)*sizeof(__i_node*)); } } __get_db()->unlock(); @@ -2883,7 +2880,7 @@ inline _LIBCPP_INLINE_VISIBILITY vector<bool, _Allocator>::vector(vector&& __v) #endif : __begin_(__v.__begin_), __size_(__v.__size_), - __cap_alloc_(std::move(__v.__cap_alloc_)) { + __cap_alloc_(_VSTD::move(__v.__cap_alloc_)) { __v.__begin_ = nullptr; __v.__size_ = 0; __v.__cap() = 0; @@ -3235,7 +3232,7 @@ vector<bool, _Allocator>::swap(vector& __x) _VSTD::swap(this->__begin_, __x.__begin_); _VSTD::swap(this->__size_, __x.__size_); _VSTD::swap(this->__cap(), __x.__cap()); - __swap_allocator(this->__alloc(), __x.__alloc(), + _VSTD::__swap_allocator(this->__alloc(), __x.__alloc(), integral_constant<bool, __alloc_traits::propagate_on_container_swap::value>()); } diff --git a/lib/libcxx/include/version b/lib/libcxx/include/version index acedd03073..813bc1ab9e 100644 --- a/lib/libcxx/include/version +++ b/lib/libcxx/include/version @@ -15,20 +15,30 @@ Macro name Value Headers __cpp_lib_addressof_constexpr 201603L <memory> -__cpp_lib_allocator_traits_is_always_equal 201411L <memory> <scoped_allocator> <string> - <deque> <forward_list> <list> - <vector> <map> <set> - <unordered_map> <unordered_set> +__cpp_lib_allocator_traits_is_always_equal 201411L <deque> <forward_list> <list> + <map> <memory> <scoped_allocator> + <set> <string> <unordered_map> + <unordered_set> <vector> __cpp_lib_any 201606L <any> __cpp_lib_apply 201603L <tuple> -__cpp_lib_array_constexpr 201811L <iterator> <array> +__cpp_lib_array_constexpr 201811L <array> <iterator> 201603L // C++17 __cpp_lib_as_const 201510L <utility> +__cpp_lib_assume_aligned 201811L <memory> +__cpp_lib_atomic_flag_test 201907L <atomic> +__cpp_lib_atomic_float 201711L <atomic> __cpp_lib_atomic_is_always_lock_free 201603L <atomic> +__cpp_lib_atomic_lock_free_type_aliases 201907L <atomic> __cpp_lib_atomic_ref 201806L <atomic> -__cpp_lib_bind_front 201811L <functional> +__cpp_lib_atomic_shared_ptr 201711L <atomic> +__cpp_lib_atomic_value_initialization 201911L <atomic> <memory> +__cpp_lib_atomic_wait 201907L <atomic> +__cpp_lib_barrier 201907L <barrier> +__cpp_lib_bind_front 201907L <functional> __cpp_lib_bit_cast 201806L <bit> +__cpp_lib_bitops 201907L <bit> __cpp_lib_bool_constant 201505L <type_traits> +__cpp_lib_bounded_array_traits 201902L <type_traits> __cpp_lib_boyer_moore_searcher 201603L <functional> __cpp_lib_byte 201603L <cstddef> __cpp_lib_char8_t 201811L <atomic> <filesystem> <istream> @@ -38,18 +48,29 @@ __cpp_lib_chrono 201611L <chrono> __cpp_lib_chrono_udls 201304L <chrono> __cpp_lib_clamp 201603L <algorithm> __cpp_lib_complex_udls 201309L <complex> -__cpp_lib_concepts 201806L <concepts> -__cpp_lib_constexpr_misc 201811L <array> <functional> <iterator> - <string_view> <tuple> <utility> -__cpp_lib_constexpr_swap_algorithms 201806L <algorithm> +__cpp_lib_concepts 202002L <concepts> +__cpp_lib_constexpr_algorithms 201806L <algorithm> +__cpp_lib_constexpr_complex 201711L <complex> +__cpp_lib_constexpr_dynamic_alloc 201907L <memory> +__cpp_lib_constexpr_functional 201907L <functional> +__cpp_lib_constexpr_iterator 201811L <iterator> +__cpp_lib_constexpr_memory 201811L <memory> +__cpp_lib_constexpr_numeric 201911L <numeric> +__cpp_lib_constexpr_string 201907L <string> +__cpp_lib_constexpr_string_view 201811L <string_view> +__cpp_lib_constexpr_tuple 201811L <tuple> +__cpp_lib_constexpr_utility 201811L <utility> +__cpp_lib_constexpr_vector 201907L <vector> +__cpp_lib_coroutine 201902L <coroutine> __cpp_lib_destroying_delete 201806L <new> __cpp_lib_enable_shared_from_this 201603L <memory> __cpp_lib_endian 201907L <bit> -__cpp_lib_erase_if 202002L <string> <deque> <forward_list> - <list> <vector> <map> - <set> <unordered_map> <unordered_set> +__cpp_lib_erase_if 202002L <deque> <forward_list> <list> + <map> <set> <string> + <unordered_map> <unordered_set> <vector> __cpp_lib_exchange_function 201304L <utility> -__cpp_lib_execution 201603L <execution> +__cpp_lib_execution 201902L <execution> + 201603L // C++17 __cpp_lib_filesystem 201703L <filesystem> __cpp_lib_gcd_lcm 201606L <numeric> __cpp_lib_generic_associative_lookup 201304L <map> <set> @@ -58,16 +79,24 @@ __cpp_lib_hardware_interference_size 201703L <new> __cpp_lib_has_unique_object_representations 201606L <type_traits> __cpp_lib_hypot 201603L <cmath> __cpp_lib_incomplete_container_elements 201505L <forward_list> <list> <vector> +__cpp_lib_int_pow2 202002L <bit> +__cpp_lib_integer_comparison_functions 202002L <utility> __cpp_lib_integer_sequence 201304L <utility> __cpp_lib_integral_constant_callable 201304L <type_traits> -__cpp_lib_interpolate 201902L <numeric> +__cpp_lib_interpolate 201902L <cmath> <numeric> __cpp_lib_invoke 201411L <functional> __cpp_lib_is_aggregate 201703L <type_traits> __cpp_lib_is_constant_evaluated 201811L <type_traits> __cpp_lib_is_final 201402L <type_traits> __cpp_lib_is_invocable 201703L <type_traits> +__cpp_lib_is_layout_compatible 201907L <type_traits> +__cpp_lib_is_nothrow_convertible 201806L <type_traits> __cpp_lib_is_null_pointer 201309L <type_traits> +__cpp_lib_is_pointer_interconvertible 201907L <type_traits> +__cpp_lib_is_scoped_enum 202011L <type_traits> __cpp_lib_is_swappable 201603L <type_traits> +__cpp_lib_jthread 201911L <stop_token> <thread> +__cpp_lib_latch 201907L <latch> __cpp_lib_launder 201606L <new> __cpp_lib_list_remove_return_type 201806L <forward_list> <list> __cpp_lib_logical_traits 201510L <type_traits> @@ -80,40 +109,55 @@ __cpp_lib_math_special_functions 201603L <cmath> __cpp_lib_memory_resource 201603L <memory_resource> __cpp_lib_node_extract 201606L <map> <set> <unordered_map> <unordered_set> -__cpp_lib_nonmember_container_access 201411L <iterator> <array> <deque> - <forward_list> <list> <map> +__cpp_lib_nonmember_container_access 201411L <array> <deque> <forward_list> + <iterator> <list> <map> <regex> <set> <string> <unordered_map> <unordered_set> <vector> __cpp_lib_not_fn 201603L <functional> __cpp_lib_null_iterators 201304L <iterator> __cpp_lib_optional 201606L <optional> __cpp_lib_parallel_algorithm 201603L <algorithm> <numeric> +__cpp_lib_polymorphic_allocator 201902L <memory> __cpp_lib_quoted_string_io 201304L <iomanip> __cpp_lib_ranges 201811L <algorithm> <functional> <iterator> <memory> <ranges> __cpp_lib_raw_memory_algorithms 201606L <memory> +__cpp_lib_remove_cvref 201711L <type_traits> __cpp_lib_result_of_sfinae 201210L <functional> <type_traits> __cpp_lib_robust_nonmodifying_seq_ops 201304L <algorithm> __cpp_lib_sample 201603L <algorithm> __cpp_lib_scoped_lock 201703L <mutex> +__cpp_lib_semaphore 201907L <semaphore> __cpp_lib_shared_mutex 201505L <shared_mutex> __cpp_lib_shared_ptr_arrays 201611L <memory> __cpp_lib_shared_ptr_weak_type 201606L <memory> __cpp_lib_shared_timed_mutex 201402L <shared_mutex> +__cpp_lib_shift 201806L <algorithm> +__cpp_lib_smart_ptr_for_overwrite 202002L <memory> +__cpp_lib_source_location 201907L <source_location> __cpp_lib_span 202002L <span> +__cpp_lib_ssize 201902L <iterator> +__cpp_lib_stacktrace 202011L <stacktrace> +__cpp_lib_starts_ends_with 201711L <string> <string_view> +__cpp_lib_stdatomic_h 202011L <stdatomic.h> +__cpp_lib_string_contains 202011L <string> <string_view> __cpp_lib_string_udls 201304L <string> -__cpp_lib_string_view 201606L <string> <string_view> -__cpp_lib_three_way_comparison 201711L <compare> +__cpp_lib_string_view 201803L <string> <string_view> + 201606L // C++17 +__cpp_lib_syncbuf 201803L <syncstream> +__cpp_lib_three_way_comparison 201907L <compare> +__cpp_lib_to_address 201711L <memory> __cpp_lib_to_array 201907L <array> __cpp_lib_to_chars 201611L <utility> __cpp_lib_transformation_trait_aliases 201304L <type_traits> -__cpp_lib_transparent_operators 201510L <functional> +__cpp_lib_transparent_operators 201510L <functional> <memory> 201210L // C++14 __cpp_lib_tuple_element_t 201402L <tuple> -__cpp_lib_tuples_by_type 201304L <utility> <tuple> +__cpp_lib_tuples_by_type 201304L <tuple> <utility> __cpp_lib_type_trait_variable_templates 201510L <type_traits> __cpp_lib_uncaught_exceptions 201411L <exception> __cpp_lib_unordered_map_try_emplace 201411L <unordered_map> +__cpp_lib_unwrap_ref 201811L <functional> __cpp_lib_variant 201606L <variant> __cpp_lib_void_t 201411L <type_traits> @@ -200,7 +244,7 @@ __cpp_lib_void_t 201411L <type_traits> # if !defined(_LIBCPP_HAS_NO_THREADS) # define __cpp_lib_shared_mutex 201505L # endif -// # define __cpp_lib_shared_ptr_arrays 201611L +# define __cpp_lib_shared_ptr_arrays 201611L # define __cpp_lib_shared_ptr_weak_type 201606L # define __cpp_lib_string_view 201606L // # define __cpp_lib_to_chars 201611L @@ -216,35 +260,105 @@ __cpp_lib_void_t 201411L <type_traits> #if _LIBCPP_STD_VER > 17 # undef __cpp_lib_array_constexpr # define __cpp_lib_array_constexpr 201811L +// # define __cpp_lib_assume_aligned 201811L +# if !defined(_LIBCPP_HAS_NO_THREADS) +# define __cpp_lib_atomic_flag_test 201907L +# endif +# if !defined(_LIBCPP_HAS_NO_THREADS) +// # define __cpp_lib_atomic_float 201711L +# endif +# if !defined(_LIBCPP_HAS_NO_THREADS) +# define __cpp_lib_atomic_lock_free_type_aliases 201907L +# endif # if !defined(_LIBCPP_HAS_NO_THREADS) // # define __cpp_lib_atomic_ref 201806L # endif -// # define __cpp_lib_bind_front 201811L +# if !defined(_LIBCPP_HAS_NO_THREADS) +// # define __cpp_lib_atomic_shared_ptr 201711L +# endif +# if !defined(_LIBCPP_HAS_NO_THREADS) +// # define __cpp_lib_atomic_value_initialization 201911L +# endif +# if !defined(_LIBCPP_HAS_NO_THREADS) +# define __cpp_lib_atomic_wait 201907L +# endif +# if !defined(_LIBCPP_HAS_NO_THREADS) +# define __cpp_lib_barrier 201907L +# endif +// # define __cpp_lib_bind_front 201907L // # define __cpp_lib_bit_cast 201806L +// # define __cpp_lib_bitops 201907L +# define __cpp_lib_bounded_array_traits 201902L # if !defined(_LIBCPP_NO_HAS_CHAR8_T) # define __cpp_lib_char8_t 201811L # endif -// # define __cpp_lib_concepts 201806L -// # define __cpp_lib_constexpr_misc 201811L -// # define __cpp_lib_constexpr_swap_algorithms 201806L +// # define __cpp_lib_concepts 202002L +// # define __cpp_lib_constexpr_algorithms 201806L +// # define __cpp_lib_constexpr_complex 201711L +# define __cpp_lib_constexpr_dynamic_alloc 201907L +# define __cpp_lib_constexpr_functional 201907L +// # define __cpp_lib_constexpr_iterator 201811L +// # define __cpp_lib_constexpr_memory 201811L +# define __cpp_lib_constexpr_numeric 201911L +// # define __cpp_lib_constexpr_string 201907L +// # define __cpp_lib_constexpr_string_view 201811L +// # define __cpp_lib_constexpr_tuple 201811L +# define __cpp_lib_constexpr_utility 201811L +// # define __cpp_lib_constexpr_vector 201907L +// # define __cpp_lib_coroutine 201902L # if _LIBCPP_STD_VER > 17 && defined(__cpp_impl_destroying_delete) && __cpp_impl_destroying_delete >= 201806L # define __cpp_lib_destroying_delete 201806L # endif # define __cpp_lib_endian 201907L # define __cpp_lib_erase_if 202002L -// # define __cpp_lib_generic_unordered_lookup 201811L +# undef __cpp_lib_execution +// # define __cpp_lib_execution 201902L +# define __cpp_lib_generic_unordered_lookup 201811L +# define __cpp_lib_int_pow2 202002L +// # define __cpp_lib_integer_comparison_functions 202002L # define __cpp_lib_interpolate 201902L # if !defined(_LIBCPP_HAS_NO_BUILTIN_IS_CONSTANT_EVALUATED) # define __cpp_lib_is_constant_evaluated 201811L # endif +// # define __cpp_lib_is_layout_compatible 201907L +# define __cpp_lib_is_nothrow_convertible 201806L +// # define __cpp_lib_is_pointer_interconvertible 201907L +# if !defined(_LIBCPP_HAS_NO_THREADS) +// # define __cpp_lib_jthread 201911L +# endif +# if !defined(_LIBCPP_HAS_NO_THREADS) +# define __cpp_lib_latch 201907L +# endif # define __cpp_lib_list_remove_return_type 201806L # if defined(__cpp_concepts) && __cpp_concepts >= 201811L # define __cpp_lib_math_constants 201907L # endif +// # define __cpp_lib_polymorphic_allocator 201902L // # define __cpp_lib_ranges 201811L +# define __cpp_lib_remove_cvref 201711L +# if !defined(_LIBCPP_HAS_NO_THREADS) +# define __cpp_lib_semaphore 201907L +# endif +# define __cpp_lib_shift 201806L +// # define __cpp_lib_smart_ptr_for_overwrite 202002L +// # define __cpp_lib_source_location 201907L # define __cpp_lib_span 202002L -// # define __cpp_lib_three_way_comparison 201711L +# define __cpp_lib_ssize 201902L +# define __cpp_lib_starts_ends_with 201711L +# undef __cpp_lib_string_view +# define __cpp_lib_string_view 201803L +// # define __cpp_lib_syncbuf 201803L +// # define __cpp_lib_three_way_comparison 201907L +# define __cpp_lib_to_address 201711L # define __cpp_lib_to_array 201907L +# define __cpp_lib_unwrap_ref 201811L +#endif + +#if _LIBCPP_STD_VER > 20 +# define __cpp_lib_is_scoped_enum 202011L +// # define __cpp_lib_stacktrace 202011L +// # define __cpp_lib_stdatomic_h 202011L +# define __cpp_lib_string_contains 202011L #endif #endif // _LIBCPP_VERSIONH diff --git a/lib/libcxx/include/wctype.h b/lib/libcxx/include/wctype.h index bdcf37234c..2d4c36d992 100644 --- a/lib/libcxx/include/wctype.h +++ b/lib/libcxx/include/wctype.h @@ -50,7 +50,9 @@ wctrans_t wctrans(const char* property); #pragma GCC system_header #endif -#include_next <wctype.h> +#if __has_include_next(<wctype.h>) +# include_next <wctype.h> +#endif #ifdef __cplusplus diff --git a/lib/libcxx/src/atomic.cpp b/lib/libcxx/src/atomic.cpp index 65d4837bb4..9ae1fb5199 100644 --- a/lib/libcxx/src/atomic.cpp +++ b/lib/libcxx/src/atomic.cpp @@ -13,14 +13,18 @@ #include <atomic> #include <functional> -#include <iostream> - #ifdef __linux__ #include <unistd.h> #include <linux/futex.h> #include <sys/syscall.h> +// libc++ uses SYS_futex as a universal syscall name. However, on 32 bit architectures +// with a 64 bit time_t, we need to specify SYS_futex_time64. +#if !defined(SYS_futex) && defined(SYS_futex_time64) +# define SYS_futex SYS_futex_time64 +#endif + #else // <- Add other operating systems here // Baseline needs no new headers diff --git a/lib/libcxx/src/barrier.cpp b/lib/libcxx/src/barrier.cpp index c5e33cbba3..9ee476993b 100644 --- a/lib/libcxx/src/barrier.cpp +++ b/lib/libcxx/src/barrier.cpp @@ -26,21 +26,15 @@ public: } __tickets[64]; }; - ptrdiff_t& __expected; - unique_ptr<char[]> __state_allocation; - __state_t* __state; + ptrdiff_t& __expected; + unique_ptr<__state_t[]> __state; _LIBCPP_HIDDEN __barrier_algorithm_base(ptrdiff_t& __expected) : __expected(__expected) { size_t const __count = (__expected + 1) >> 1; - size_t const __size = sizeof(__state_t) * __count; - size_t __allocation_size = __size + alignof(__state_t); - __state_allocation = unique_ptr<char[]>(new char[__allocation_size]); - void* __allocation = __state_allocation.get(); - void* const __state_ = align(alignof(__state_t), __size, __allocation, __allocation_size); - __state = new (__state_) __barrier_algorithm_base::__state_t[__count]; + __state = unique_ptr<__state_t[]>(new __state_t[__count]); } _LIBCPP_HIDDEN bool __arrive(__barrier_phase_t __old_phase) diff --git a/lib/libcxx/src/chrono.cpp b/lib/libcxx/src/chrono.cpp index f0a5d50ddf..085fbfde26 100644 --- a/lib/libcxx/src/chrono.cpp +++ b/lib/libcxx/src/chrono.cpp @@ -13,11 +13,15 @@ #include "include/apple_availability.h" #if __has_include(<unistd.h>) -#include <unistd.h> +# include <unistd.h> +#endif + +#if __has_include(<sys/time.h>) +# include <sys/time.h> // for gettimeofday and timeval #endif #if !defined(__APPLE__) && _POSIX_TIMERS > 0 -#define _LIBCPP_USE_CLOCK_GETTIME +# define _LIBCPP_USE_CLOCK_GETTIME #endif #if defined(_LIBCPP_WIN32API) @@ -27,12 +31,12 @@ # if _WIN32_WINNT >= _WIN32_WINNT_WIN8 # include <winapifamily.h> # endif -#else -# if !defined(CLOCK_REALTIME) -# include <sys/time.h> // for gettimeofday and timeval -# endif // !defined(CLOCK_REALTIME) #endif // defined(_LIBCPP_WIN32API) +#if __has_include(<mach/mach_time.h>) +# include <mach/mach_time.h> +#endif + #if defined(__ELF__) && defined(_LIBCPP_LINK_RT_LIB) # pragma comment(lib, "rt") #endif @@ -42,14 +46,13 @@ _LIBCPP_BEGIN_NAMESPACE_STD namespace chrono { +// // system_clock +// -const bool system_clock::is_steady; - -system_clock::time_point -system_clock::now() _NOEXCEPT -{ #if defined(_LIBCPP_WIN32API) + +static system_clock::time_point __libcpp_system_clock_now() { // FILETIME is in 100ns units using filetime_duration = _VSTD::chrono::duration<__int64, @@ -60,31 +63,42 @@ system_clock::now() _NOEXCEPT static _LIBCPP_CONSTEXPR const seconds nt_to_unix_epoch{11644473600}; FILETIME ft; -#if _WIN32_WINNT >= _WIN32_WINNT_WIN8 -#if WINAPI_FAMILY_PARTITION(WINAPI_PARTITION_DESKTOP) +#if _WIN32_WINNT >= _WIN32_WINNT_WIN8 && WINAPI_FAMILY_PARTITION(WINAPI_PARTITION_DESKTOP) GetSystemTimePreciseAsFileTime(&ft); #else GetSystemTimeAsFileTime(&ft); #endif -#else - GetSystemTimeAsFileTime(&ft); -#endif filetime_duration d{(static_cast<__int64>(ft.dwHighDateTime) << 32) | static_cast<__int64>(ft.dwLowDateTime)}; - return time_point(duration_cast<duration>(d - nt_to_unix_epoch)); -#else -#if defined(CLOCK_REALTIME) + return system_clock::time_point(duration_cast<system_clock::duration>(d - nt_to_unix_epoch)); +} + +#elif defined(CLOCK_REALTIME) && defined(_LIBCPP_USE_CLOCK_GETTIME) + +static system_clock::time_point __libcpp_system_clock_now() { struct timespec tp; if (0 != clock_gettime(CLOCK_REALTIME, &tp)) __throw_system_error(errno, "clock_gettime(CLOCK_REALTIME) failed"); - return time_point(seconds(tp.tv_sec) + microseconds(tp.tv_nsec / 1000)); + return system_clock::time_point(seconds(tp.tv_sec) + microseconds(tp.tv_nsec / 1000)); +} + #else + +static system_clock::time_point __libcpp_system_clock_now() { timeval tv; gettimeofday(&tv, 0); - return time_point(seconds(tv.tv_sec) + microseconds(tv.tv_usec)); -#endif // CLOCK_REALTIME + return system_clock::time_point(seconds(tv.tv_sec) + microseconds(tv.tv_usec)); +} + #endif + +const bool system_clock::is_steady; + +system_clock::time_point +system_clock::now() _NOEXCEPT +{ + return __libcpp_system_clock_now(); } time_t @@ -99,35 +113,85 @@ system_clock::from_time_t(time_t t) _NOEXCEPT return system_clock::time_point(seconds(t)); } -#ifndef _LIBCPP_HAS_NO_MONOTONIC_CLOCK +// // steady_clock // // Warning: If this is not truly steady, then it is non-conforming. It is // better for it to not exist and have the rest of libc++ use system_clock // instead. +// -const bool steady_clock::is_steady; +#ifndef _LIBCPP_HAS_NO_MONOTONIC_CLOCK #if defined(__APPLE__) -#if !defined(CLOCK_MONOTONIC_RAW) -# error "Building libc++ on Apple platforms requires CLOCK_MONOTONIC_RAW" +// TODO(ldionne): +// This old implementation of steady_clock is retained until Chrome drops supports +// for macOS < 10.12. The issue is that they link libc++ statically into their +// application, which means that libc++ must support being built for such deployment +// targets. See https://llvm.org/D74489 for details. +#if (defined(__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__) && __ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ < 101200) || \ + (defined(__ENVIRONMENT_IPHONE_OS_VERSION_MIN_REQUIRED__) && __ENVIRONMENT_IPHONE_OS_VERSION_MIN_REQUIRED__ < 100000) || \ + (defined(__ENVIRONMENT_TV_OS_VERSION_MIN_REQUIRED__) && __ENVIRONMENT_TV_OS_VERSION_MIN_REQUIRED__ < 100000) || \ + (defined(__ENVIRONMENT_WATCH_OS_VERSION_MIN_REQUIRED__) && __ENVIRONMENT_WATCH_OS_VERSION_MIN_REQUIRED__ < 30000) +# define _LIBCPP_USE_OLD_MACH_ABSOLUTE_TIME #endif +#if defined(_LIBCPP_USE_OLD_MACH_ABSOLUTE_TIME) + +// mach_absolute_time() * MachInfo.numer / MachInfo.denom is the number of +// nanoseconds since the computer booted up. MachInfo.numer and MachInfo.denom +// are run time constants supplied by the OS. This clock has no relationship +// to the Gregorian calendar. It's main use is as a high resolution timer. + +// MachInfo.numer / MachInfo.denom is often 1 on the latest equipment. Specialize +// for that case as an optimization. + +static steady_clock::rep steady_simplified() { + return static_cast<steady_clock::rep>(mach_absolute_time()); +} +static double compute_steady_factor() { + mach_timebase_info_data_t MachInfo; + mach_timebase_info(&MachInfo); + return static_cast<double>(MachInfo.numer) / MachInfo.denom; +} + +static steady_clock::rep steady_full() { + static const double factor = compute_steady_factor(); + return static_cast<steady_clock::rep>(mach_absolute_time() * factor); +} + +typedef steady_clock::rep (*FP)(); + +static FP init_steady_clock() { + mach_timebase_info_data_t MachInfo; + mach_timebase_info(&MachInfo); + if (MachInfo.numer == MachInfo.denom) + return &steady_simplified; + return &steady_full; +} + +static steady_clock::time_point __libcpp_steady_clock_now() { + static FP fp = init_steady_clock(); + return steady_clock::time_point(steady_clock::duration(fp())); +} + +#else // vvvvv default behavior for Apple platforms vvvvv + // On Apple platforms, only CLOCK_UPTIME_RAW, CLOCK_MONOTONIC_RAW or // mach_absolute_time are able to time functions in the nanosecond range. // Furthermore, only CLOCK_MONOTONIC_RAW is truly monotonic, because it // also counts cycles when the system is asleep. Thus, it is the only // acceptable implementation of steady_clock. -steady_clock::time_point -steady_clock::now() _NOEXCEPT -{ +static steady_clock::time_point __libcpp_steady_clock_now() { struct timespec tp; if (0 != clock_gettime(CLOCK_MONOTONIC_RAW, &tp)) __throw_system_error(errno, "clock_gettime(CLOCK_MONOTONIC_RAW) failed"); - return time_point(seconds(tp.tv_sec) + nanoseconds(tp.tv_nsec)); + return steady_clock::time_point(seconds(tp.tv_sec) + nanoseconds(tp.tv_nsec)); } +#endif + #elif defined(_LIBCPP_WIN32API) // https://msdn.microsoft.com/en-us/library/windows/desktop/ms644905(v=vs.85).aspx says: @@ -138,36 +202,43 @@ steady_clock::now() _NOEXCEPT static LARGE_INTEGER __QueryPerformanceFrequency() { - LARGE_INTEGER val; - (void) QueryPerformanceFrequency(&val); - return val; + LARGE_INTEGER val; + (void) QueryPerformanceFrequency(&val); + return val; } -steady_clock::time_point -steady_clock::now() _NOEXCEPT -{ +static steady_clock::time_point __libcpp_steady_clock_now() { static const LARGE_INTEGER freq = __QueryPerformanceFrequency(); LARGE_INTEGER counter; (void) QueryPerformanceCounter(&counter); - return time_point(duration(counter.QuadPart * nano::den / freq.QuadPart)); + auto seconds = counter.QuadPart / freq.QuadPart; + auto fractions = counter.QuadPart % freq.QuadPart; + auto dur = seconds * nano::den + fractions * nano::den / freq.QuadPart; + return steady_clock::time_point(steady_clock::duration(dur)); } #elif defined(CLOCK_MONOTONIC) -steady_clock::time_point -steady_clock::now() _NOEXCEPT -{ +static steady_clock::time_point __libcpp_steady_clock_now() { struct timespec tp; if (0 != clock_gettime(CLOCK_MONOTONIC, &tp)) __throw_system_error(errno, "clock_gettime(CLOCK_MONOTONIC) failed"); - return time_point(seconds(tp.tv_sec) + nanoseconds(tp.tv_nsec)); + return steady_clock::time_point(seconds(tp.tv_sec) + nanoseconds(tp.tv_nsec)); } #else -# error "Monotonic clock not implemented" +# error "Monotonic clock not implemented on this platform" #endif +const bool steady_clock::is_steady; + +steady_clock::time_point +steady_clock::now() _NOEXCEPT +{ + return __libcpp_steady_clock_now(); +} + #endif // !_LIBCPP_HAS_NO_MONOTONIC_CLOCK } diff --git a/lib/libcxx/src/experimental/memory_resource.cpp b/lib/libcxx/src/experimental/memory_resource.cpp index 68c5bc99cc..1304ef3df7 100644 --- a/lib/libcxx/src/experimental/memory_resource.cpp +++ b/lib/libcxx/src/experimental/memory_resource.cpp @@ -76,16 +76,6 @@ union ResourceInitHelper { ~ResourceInitHelper() {} }; -// Detect if the init_priority attribute is supported. -#if (defined(_LIBCPP_COMPILER_GCC) && defined(__APPLE__)) \ - || defined(_LIBCPP_COMPILER_MSVC) -// GCC on Apple doesn't support the init priority attribute, -// and MSVC doesn't support any GCC attributes. -# define _LIBCPP_INIT_PRIORITY_MAX -#else -# define _LIBCPP_INIT_PRIORITY_MAX __attribute__((init_priority(101))) -#endif - // When compiled in C++14 this initialization should be a constant expression. // Only in C++11 is "init_priority" needed to ensure initialization order. #if _LIBCPP_STD_VER > 11 diff --git a/lib/libcxx/src/filesystem/directory_iterator.cpp b/lib/libcxx/src/filesystem/directory_iterator.cpp index e8941b3494..2721dea5c9 100644 --- a/lib/libcxx/src/filesystem/directory_iterator.cpp +++ b/lib/libcxx/src/filesystem/directory_iterator.cpp @@ -10,6 +10,7 @@ #include "__config" #if defined(_LIBCPP_WIN32API) #define WIN32_LEAN_AND_MEAN +#define NOMINMAX #include <windows.h> #else #include <dirent.h> @@ -72,16 +73,20 @@ static pair<string_view, file_type> posix_readdir(DIR* dir_stream, } } #else +// defined(_LIBCPP_WIN32API) -static file_type get_file_type(const WIN32_FIND_DATA& data) { - //auto attrs = data.dwFileAttributes; - // FIXME(EricWF) - return file_type::unknown; +static file_type get_file_type(const WIN32_FIND_DATAW& data) { + if (data.dwFileAttributes & FILE_ATTRIBUTE_REPARSE_POINT && + data.dwReserved0 == IO_REPARSE_TAG_SYMLINK) + return file_type::symlink; + if (data.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) + return file_type::directory; + return file_type::regular; } -static uintmax_t get_file_size(const WIN32_FIND_DATA& data) { - return (data.nFileSizeHigh * (MAXDWORD + 1)) + data.nFileSizeLow; +static uintmax_t get_file_size(const WIN32_FIND_DATAW& data) { + return (static_cast<uint64_t>(data.nFileSizeHigh) << 32) + data.nFileSizeLow; } -static file_time_type get_write_time(const WIN32_FIND_DATA& data) { +static file_time_type get_write_time(const WIN32_FIND_DATAW& data) { ULARGE_INTEGER tmp; const FILETIME& time = data.ftLastWriteTime; tmp.u.LowPart = time.dwLowDateTime; @@ -110,15 +115,21 @@ public: __dir_stream(const path& root, directory_options opts, error_code& ec) : __stream_(INVALID_HANDLE_VALUE), __root_(root) { - __stream_ = ::FindFirstFile(root.c_str(), &__data_); + if (root.native().empty()) { + ec = make_error_code(errc::no_such_file_or_directory); + return; + } + __stream_ = ::FindFirstFileW((root / "*").c_str(), &__data_); if (__stream_ == INVALID_HANDLE_VALUE) { - ec = error_code(::GetLastError(), generic_category()); + ec = detail::make_windows_error(GetLastError()); const bool ignore_permission_denied = bool(opts & directory_options::skip_permission_denied); if (ignore_permission_denied && ec.value() == ERROR_ACCESS_DENIED) ec.clear(); return; } + if (!assign()) + advance(ec); } ~__dir_stream() noexcept { @@ -130,35 +141,39 @@ public: bool good() const noexcept { return __stream_ != INVALID_HANDLE_VALUE; } bool advance(error_code& ec) { - while (::FindNextFile(__stream_, &__data_)) { - if (!strcmp(__data_.cFileName, ".") || strcmp(__data_.cFileName, "..")) - continue; - // FIXME: Cache more of this - //directory_entry::__cached_data cdata; - //cdata.__type_ = get_file_type(__data_); - //cdata.__size_ = get_file_size(__data_); - //cdata.__write_time_ = get_write_time(__data_); - __entry_.__assign_iter_entry( - __root_ / __data_.cFileName, - directory_entry::__create_iter_result(detail::get_file_type(__data))); - return true; + while (::FindNextFileW(__stream_, &__data_)) { + if (assign()) + return true; } - ec = error_code(::GetLastError(), generic_category()); close(); return false; } + bool assign() { + if (!wcscmp(__data_.cFileName, L".") || !wcscmp(__data_.cFileName, L"..")) + return false; + // FIXME: Cache more of this + //directory_entry::__cached_data cdata; + //cdata.__type_ = get_file_type(__data_); + //cdata.__size_ = get_file_size(__data_); + //cdata.__write_time_ = get_write_time(__data_); + __entry_.__assign_iter_entry( + __root_ / __data_.cFileName, + directory_entry::__create_iter_result(detail::get_file_type(__data_))); + return true; + } + private: error_code close() noexcept { error_code ec; if (!::FindClose(__stream_)) - ec = error_code(::GetLastError(), generic_category()); + ec = detail::make_windows_error(GetLastError()); __stream_ = INVALID_HANDLE_VALUE; return ec; } HANDLE __stream_{INVALID_HANDLE_VALUE}; - WIN32_FIND_DATA __data_; + WIN32_FIND_DATAW __data_; public: path __root_; diff --git a/lib/libcxx/src/filesystem/filesystem_common.h b/lib/libcxx/src/filesystem/filesystem_common.h index fe5c42f5e6..e0fdbccf96 100644 --- a/lib/libcxx/src/filesystem/filesystem_common.h +++ b/lib/libcxx/src/filesystem/filesystem_common.h @@ -13,14 +13,17 @@ #include "filesystem" #include "array" #include "chrono" -#include "cstdlib" #include "climits" - -#include <unistd.h> -#include <sys/stat.h> -#include <sys/statvfs.h> -#include <sys/time.h> // for ::utimes as used in __last_write_time -#include <fcntl.h> /* values for fchmodat */ +#include "cstdlib" +#include "ctime" + +#if !defined(_LIBCPP_WIN32API) +# include <unistd.h> +# include <sys/stat.h> +# include <sys/statvfs.h> +# include <sys/time.h> // for ::utimes as used in __last_write_time +# include <fcntl.h> /* values for fchmodat */ +#endif #include "../include/apple_availability.h" @@ -37,9 +40,21 @@ #pragma GCC diagnostic ignored "-Wunused-function" #endif +#if defined(_LIBCPP_WIN32API) +#define PS(x) (L##x) +#else +#define PS(x) (x) +#endif + _LIBCPP_BEGIN_NAMESPACE_FILESYSTEM namespace detail { + +#if defined(_LIBCPP_WIN32API) +// Non anonymous, to allow access from two translation units. +errc __win_err_to_errc(int err); +#endif + namespace { static string format_string_imp(const char* msg, ...) { @@ -47,7 +62,7 @@ static string format_string_imp(const char* msg, ...) { struct GuardVAList { va_list& target; bool active = true; - GuardVAList(va_list& target) : target(target), active(true) {} + GuardVAList(va_list& tgt) : target(tgt), active(true) {} void clear() { if (active) va_end(target); @@ -93,8 +108,8 @@ static string format_string_imp(const char* msg, ...) { return result; } -const char* unwrap(string const& s) { return s.c_str(); } -const char* unwrap(path const& p) { return p.native().c_str(); } +const path::value_type* unwrap(path::string_type const& s) { return s.c_str(); } +const path::value_type* unwrap(path const& p) { return p.native().c_str(); } template <class Arg> Arg const& unwrap(Arg const& a) { static_assert(!is_class<Arg>::value, "cannot pass class here"); @@ -111,6 +126,12 @@ error_code capture_errno() { return error_code(errno, generic_category()); } +#if defined(_LIBCPP_WIN32API) +error_code make_windows_error(int err) { + return make_error_code(__win_err_to_errc(err)); +} +#endif + template <class T> T error_value(); template <> @@ -119,6 +140,12 @@ template <> bool error_value<bool>() { return false; } +#if __SIZEOF_SIZE_T__ != __SIZEOF_LONG_LONG__ +template <> +size_t error_value<size_t>() { + return size_t(-1); +} +#endif template <> uintmax_t error_value<uintmax_t>() { return uintmax_t(-1); @@ -134,50 +161,50 @@ path error_value<path>() { template <class T> struct ErrorHandler { - const char* func_name; - error_code* ec = nullptr; - const path* p1 = nullptr; - const path* p2 = nullptr; + const char* func_name_; + error_code* ec_ = nullptr; + const path* p1_ = nullptr; + const path* p2_ = nullptr; ErrorHandler(const char* fname, error_code* ec, const path* p1 = nullptr, const path* p2 = nullptr) - : func_name(fname), ec(ec), p1(p1), p2(p2) { - if (ec) - ec->clear(); + : func_name_(fname), ec_(ec), p1_(p1), p2_(p2) { + if (ec_) + ec_->clear(); } - T report(const error_code& m_ec) const { - if (ec) { - *ec = m_ec; + T report(const error_code& ec) const { + if (ec_) { + *ec_ = ec; return error_value<T>(); } - string what = string("in ") + func_name; - switch (bool(p1) + bool(p2)) { + string what = string("in ") + func_name_; + switch (bool(p1_) + bool(p2_)) { case 0: - __throw_filesystem_error(what, m_ec); + __throw_filesystem_error(what, ec); case 1: - __throw_filesystem_error(what, *p1, m_ec); + __throw_filesystem_error(what, *p1_, ec); case 2: - __throw_filesystem_error(what, *p1, *p2, m_ec); + __throw_filesystem_error(what, *p1_, *p2_, ec); } _LIBCPP_UNREACHABLE(); } template <class... Args> - T report(const error_code& m_ec, const char* msg, Args const&... args) const { - if (ec) { - *ec = m_ec; + T report(const error_code& ec, const char* msg, Args const&... args) const { + if (ec_) { + *ec_ = ec; return error_value<T>(); } string what = - string("in ") + func_name + ": " + format_string(msg, args...); - switch (bool(p1) + bool(p2)) { + string("in ") + func_name_ + ": " + format_string(msg, args...); + switch (bool(p1_) + bool(p2_)) { case 0: - __throw_filesystem_error(what, m_ec); + __throw_filesystem_error(what, ec); case 1: - __throw_filesystem_error(what, *p1, m_ec); + __throw_filesystem_error(what, *p1_, ec); case 2: - __throw_filesystem_error(what, *p1, *p2, m_ec); + __throw_filesystem_error(what, *p1_, *p2_, ec); } _LIBCPP_UNREACHABLE(); } @@ -197,8 +224,9 @@ private: using chrono::duration; using chrono::duration_cast; -using TimeSpec = struct ::timespec; -using StatT = struct ::stat; +using TimeSpec = struct timespec; +using TimeVal = struct timeval; +using StatT = struct stat; template <class FileTimeT, class TimeT, bool IsFloat = is_floating_point<typename FileTimeT::rep>::value> @@ -380,26 +408,38 @@ public: using fs_time = time_util<file_time_type, time_t, TimeSpec>; #if defined(__APPLE__) -TimeSpec extract_mtime(StatT const& st) { return st.st_mtimespec; } -TimeSpec extract_atime(StatT const& st) { return st.st_atimespec; } +inline TimeSpec extract_mtime(StatT const& st) { return st.st_mtimespec; } +inline TimeSpec extract_atime(StatT const& st) { return st.st_atimespec; } +#elif defined(__MVS__) +inline TimeSpec extract_mtime(StatT const& st) { + TimeSpec TS = {st.st_mtime, 0}; + return TS; +} +inline TimeSpec extract_atime(StatT const& st) { + TimeSpec TS = {st.st_atime, 0}; + return TS; +} #else -TimeSpec extract_mtime(StatT const& st) { return st.st_mtim; } -TimeSpec extract_atime(StatT const& st) { return st.st_atim; } +inline TimeSpec extract_mtime(StatT const& st) { return st.st_mtim; } +inline TimeSpec extract_atime(StatT const& st) { return st.st_atim; } #endif -// allow the utimes implementation to compile even it we're not going -// to use it. - -bool posix_utimes(const path& p, std::array<TimeSpec, 2> const& TS, - error_code& ec) { +inline TimeVal make_timeval(TimeSpec const& ts) { using namespace chrono; auto Convert = [](long nsec) { - using int_type = decltype(std::declval< ::timeval>().tv_usec); + using int_type = decltype(std::declval<TimeVal>().tv_usec); auto dur = duration_cast<microseconds>(nanoseconds(nsec)).count(); return static_cast<int_type>(dur); }; - struct ::timeval ConvertedTS[2] = {{TS[0].tv_sec, Convert(TS[0].tv_nsec)}, - {TS[1].tv_sec, Convert(TS[1].tv_nsec)}}; + TimeVal TV = {}; + TV.tv_sec = ts.tv_sec; + TV.tv_usec = Convert(ts.tv_nsec); + return TV; +} + +inline bool posix_utimes(const path& p, std::array<TimeSpec, 2> const& TS, + error_code& ec) { + TimeVal ConvertedTS[2] = {make_timeval(TS[0]), make_timeval(TS[1])}; if (::utimes(p.c_str(), ConvertedTS) == -1) { ec = capture_errno(); return true; diff --git a/lib/libcxx/src/filesystem/operations.cpp b/lib/libcxx/src/filesystem/operations.cpp index 5c98671fee..50a895dc2f 100644 --- a/lib/libcxx/src/filesystem/operations.cpp +++ b/lib/libcxx/src/filesystem/operations.cpp @@ -9,8 +9,6 @@ #include "filesystem" #include "array" #include "iterator" -#include "fstream" -#include "random" /* for unique_path */ #include "string_view" #include "type_traits" #include "vector" @@ -19,40 +17,51 @@ #include "filesystem_common.h" -#include <unistd.h> -#include <sys/stat.h> -#include <sys/statvfs.h> +#if defined(_LIBCPP_WIN32API) +# define WIN32_LEAN_AND_MEAN +# define NOMINMAX +# include <windows.h> +#else +# include <unistd.h> +# include <sys/stat.h> +# include <sys/statvfs.h> +#endif #include <time.h> #include <fcntl.h> /* values for fchmodat */ -#if defined(__linux__) -#include <linux/version.h> -#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 33) -#include <sys/sendfile.h> -#define _LIBCPP_USE_SENDFILE -#endif +#if __has_include(<sys/sendfile.h>) +# include <sys/sendfile.h> +# define _LIBCPP_FILESYSTEM_USE_SENDFILE #elif defined(__APPLE__) || __has_include(<copyfile.h>) -#include <copyfile.h> -#define _LIBCPP_USE_COPYFILE +# include <copyfile.h> +# define _LIBCPP_FILESYSTEM_USE_COPYFILE +#else +# include "fstream" +# define _LIBCPP_FILESYSTEM_USE_FSTREAM #endif #if !defined(CLOCK_REALTIME) -#include <sys/time.h> // for gettimeofday and timeval -#endif // !defined(CLOCK_REALTIME) - -#if defined(__ELF__) && defined(_LIBCPP_LINK_RT_LIB) -#pragma comment(lib, "rt") +# include <sys/time.h> // for gettimeofday and timeval #endif -#if defined(_LIBCPP_COMPILER_GCC) -#if _GNUC_VER < 500 -#pragma GCC diagnostic ignored "-Wmissing-field-initializers" -#endif +#if defined(__ELF__) && defined(_LIBCPP_LINK_RT_LIB) +# pragma comment(lib, "rt") #endif _LIBCPP_BEGIN_NAMESPACE_FILESYSTEM namespace { + +bool isSeparator(path::value_type C) { + if (C == '/') + return true; +#if defined(_LIBCPP_WIN32API) + if (C == '\\') + return true; +#endif + return false; +} + namespace parser { using string_view_t = path::__string_view; @@ -179,11 +188,14 @@ public: switch (State) { case PS_BeforeBegin: case PS_AtEnd: - return ""; + return PS(""); case PS_InRootDir: - return "/"; + if (RawEntry[0] == '\\') + return PS("\\"); + else + return PS("/"); case PS_InTrailingSep: - return ""; + return PS(""); case PS_InRootName: case PS_InFilenames: return RawEntry; @@ -270,29 +282,29 @@ private: } PosPtr consumeSeparator(PosPtr P, PosPtr End) const noexcept { - if (P == End || *P != '/') + if (P == End || !isSeparator(*P)) return nullptr; const int Inc = P < End ? 1 : -1; P += Inc; - while (P != End && *P == '/') + while (P != End && isSeparator(*P)) P += Inc; return P; } PosPtr consumeName(PosPtr P, PosPtr End) const noexcept { - if (P == End || *P == '/') + if (P == End || isSeparator(*P)) return nullptr; const int Inc = P < End ? 1 : -1; P += Inc; - while (P != End && *P != '/') + while (P != End && !isSeparator(*P)) P += Inc; return P; } }; string_view_pair separate_filename(string_view_t const& s) { - if (s == "." || s == ".." || s.empty()) - return string_view_pair{s, ""}; + if (s == PS(".") || s == PS("..") || s.empty()) + return string_view_pair{s, PS("")}; auto pos = s.find_last_of('.'); if (pos == string_view_t::npos || pos == 0) return string_view_pair{s, string_view_t{}}; @@ -308,6 +320,73 @@ string_view_t createView(PosPtr S, PosPtr E) noexcept { // POSIX HELPERS +#if defined(_LIBCPP_WIN32API) +namespace detail { + +errc __win_err_to_errc(int err) { + constexpr struct { + DWORD win; + errc errc; + } win_error_mapping[] = { + {ERROR_ACCESS_DENIED, errc::permission_denied}, + {ERROR_ALREADY_EXISTS, errc::file_exists}, + {ERROR_BAD_NETPATH, errc::no_such_file_or_directory}, + {ERROR_BAD_UNIT, errc::no_such_device}, + {ERROR_BROKEN_PIPE, errc::broken_pipe}, + {ERROR_BUFFER_OVERFLOW, errc::filename_too_long}, + {ERROR_BUSY, errc::device_or_resource_busy}, + {ERROR_BUSY_DRIVE, errc::device_or_resource_busy}, + {ERROR_CANNOT_MAKE, errc::permission_denied}, + {ERROR_CANTOPEN, errc::io_error}, + {ERROR_CANTREAD, errc::io_error}, + {ERROR_CANTWRITE, errc::io_error}, + {ERROR_CURRENT_DIRECTORY, errc::permission_denied}, + {ERROR_DEV_NOT_EXIST, errc::no_such_device}, + {ERROR_DEVICE_IN_USE, errc::device_or_resource_busy}, + {ERROR_DIR_NOT_EMPTY, errc::directory_not_empty}, + {ERROR_DIRECTORY, errc::invalid_argument}, + {ERROR_DISK_FULL, errc::no_space_on_device}, + {ERROR_FILE_EXISTS, errc::file_exists}, + {ERROR_FILE_NOT_FOUND, errc::no_such_file_or_directory}, + {ERROR_HANDLE_DISK_FULL, errc::no_space_on_device}, + {ERROR_INVALID_ACCESS, errc::permission_denied}, + {ERROR_INVALID_DRIVE, errc::no_such_device}, + {ERROR_INVALID_FUNCTION, errc::function_not_supported}, + {ERROR_INVALID_HANDLE, errc::invalid_argument}, + {ERROR_INVALID_NAME, errc::no_such_file_or_directory}, + {ERROR_INVALID_PARAMETER, errc::invalid_argument}, + {ERROR_LOCK_VIOLATION, errc::no_lock_available}, + {ERROR_LOCKED, errc::no_lock_available}, + {ERROR_NEGATIVE_SEEK, errc::invalid_argument}, + {ERROR_NOACCESS, errc::permission_denied}, + {ERROR_NOT_ENOUGH_MEMORY, errc::not_enough_memory}, + {ERROR_NOT_READY, errc::resource_unavailable_try_again}, + {ERROR_NOT_SAME_DEVICE, errc::cross_device_link}, + {ERROR_NOT_SUPPORTED, errc::not_supported}, + {ERROR_OPEN_FAILED, errc::io_error}, + {ERROR_OPEN_FILES, errc::device_or_resource_busy}, + {ERROR_OPERATION_ABORTED, errc::operation_canceled}, + {ERROR_OUTOFMEMORY, errc::not_enough_memory}, + {ERROR_PATH_NOT_FOUND, errc::no_such_file_or_directory}, + {ERROR_READ_FAULT, errc::io_error}, + {ERROR_REPARSE_TAG_INVALID, errc::invalid_argument}, + {ERROR_RETRY, errc::resource_unavailable_try_again}, + {ERROR_SEEK, errc::io_error}, + {ERROR_SHARING_VIOLATION, errc::permission_denied}, + {ERROR_TOO_MANY_OPEN_FILES, errc::too_many_files_open}, + {ERROR_WRITE_FAULT, errc::io_error}, + {ERROR_WRITE_PROTECT, errc::permission_denied}, + }; + + for (const auto &pair : win_error_mapping) + if (pair.win == static_cast<DWORD>(err)) + return pair.errc; + return errc::invalid_argument; +} + +} // namespace detail +#endif + namespace detail { namespace { @@ -503,19 +582,25 @@ _FilesystemClock::time_point _FilesystemClock::now() noexcept { filesystem_error::~filesystem_error() {} +#if defined(_LIBCPP_WIN32API) +#define PS_FMT "%ls" +#else +#define PS_FMT "%s" +#endif + void filesystem_error::__create_what(int __num_paths) { const char* derived_what = system_error::what(); __storage_->__what_ = [&]() -> string { - const char* p1 = path1().native().empty() ? "\"\"" : path1().c_str(); - const char* p2 = path2().native().empty() ? "\"\"" : path2().c_str(); + const path::value_type* p1 = path1().native().empty() ? PS("\"\"") : path1().c_str(); + const path::value_type* p2 = path2().native().empty() ? PS("\"\"") : path2().c_str(); switch (__num_paths) { default: return detail::format_string("filesystem error: %s", derived_what); case 1: - return detail::format_string("filesystem error: %s [%s]", derived_what, + return detail::format_string("filesystem error: %s [" PS_FMT "]", derived_what, p1); case 2: - return detail::format_string("filesystem error: %s [%s] [%s]", + return detail::format_string("filesystem error: %s [" PS_FMT "] [" PS_FMT "]", derived_what, p1, p2); } }(); @@ -542,14 +627,18 @@ path __canonical(path const& orig_p, error_code* ec) { ErrorHandler<path> err("canonical", ec, &orig_p, &cwd); path p = __do_absolute(orig_p, &cwd, ec); -#if _POSIX_VERSION >= 200112 +#if defined(_POSIX_VERSION) && _POSIX_VERSION >= 200112 std::unique_ptr<char, decltype(&::free)> hold(::realpath(p.c_str(), nullptr), &::free); if (hold.get() == nullptr) return err.report(capture_errno()); return {hold.get()}; #else - char buff[PATH_MAX + 1]; + #if defined(__MVS__) && !defined(PATH_MAX) + char buff[ _XOPEN_PATH_MAX + 1 ]; + #else + char buff[PATH_MAX + 1]; + #endif char* ret; if ((ret = ::realpath(p.c_str(), buff)) == nullptr) return err.report(capture_errno()); @@ -646,96 +735,83 @@ void __copy(const path& from, const path& to, copy_options options, namespace detail { namespace { -#ifdef _LIBCPP_USE_SENDFILE -bool copy_file_impl_sendfile(FileDescriptor& read_fd, FileDescriptor& write_fd, - error_code& ec) { - - size_t count = read_fd.get_stat().st_size; - do { - ssize_t res; - if ((res = ::sendfile(write_fd.fd, read_fd.fd, nullptr, count)) == -1) { - ec = capture_errno(); - return false; - } - count -= res; - } while (count > 0); - - ec.clear(); +#if defined(_LIBCPP_FILESYSTEM_USE_SENDFILE) + bool copy_file_impl(FileDescriptor& read_fd, FileDescriptor& write_fd, error_code& ec) { + size_t count = read_fd.get_stat().st_size; + do { + ssize_t res; + if ((res = ::sendfile(write_fd.fd, read_fd.fd, nullptr, count)) == -1) { + ec = capture_errno(); + return false; + } + count -= res; + } while (count > 0); - return true; -} -#elif defined(_LIBCPP_USE_COPYFILE) -bool copy_file_impl_copyfile(FileDescriptor& read_fd, FileDescriptor& write_fd, - error_code& ec) { - struct CopyFileState { - copyfile_state_t state; - CopyFileState() { state = copyfile_state_alloc(); } - ~CopyFileState() { copyfile_state_free(state); } - - private: - CopyFileState(CopyFileState const&) = delete; - CopyFileState& operator=(CopyFileState const&) = delete; - }; + ec.clear(); - CopyFileState cfs; - if (fcopyfile(read_fd.fd, write_fd.fd, cfs.state, COPYFILE_DATA) < 0) { - ec = capture_errno(); - return false; + return true; } +#elif defined(_LIBCPP_FILESYSTEM_USE_COPYFILE) + bool copy_file_impl(FileDescriptor& read_fd, FileDescriptor& write_fd, error_code& ec) { + struct CopyFileState { + copyfile_state_t state; + CopyFileState() { state = copyfile_state_alloc(); } + ~CopyFileState() { copyfile_state_free(state); } - ec.clear(); - return true; -} -#endif + private: + CopyFileState(CopyFileState const&) = delete; + CopyFileState& operator=(CopyFileState const&) = delete; + }; -// Note: This function isn't guarded by ifdef's even though it may be unused -// in order to assure it still compiles. -__attribute__((unused)) bool copy_file_impl_default(FileDescriptor& read_fd, - FileDescriptor& write_fd, - error_code& ec) { - ifstream in; - in.__open(read_fd.fd, ios::binary); - if (!in.is_open()) { - // This assumes that __open didn't reset the error code. - ec = capture_errno(); - return false; - } - ofstream out; - out.__open(write_fd.fd, ios::binary); - if (!out.is_open()) { - ec = capture_errno(); - return false; - } + CopyFileState cfs; + if (fcopyfile(read_fd.fd, write_fd.fd, cfs.state, COPYFILE_DATA) < 0) { + ec = capture_errno(); + return false; + } - if (in.good() && out.good()) { - using InIt = istreambuf_iterator<char>; - using OutIt = ostreambuf_iterator<char>; - InIt bin(in); - InIt ein; - OutIt bout(out); - copy(bin, ein, bout); - } - if (out.fail() || in.fail()) { - ec = make_error_code(errc::io_error); - return false; + ec.clear(); + return true; } +#elif defined(_LIBCPP_FILESYSTEM_USE_FSTREAM) + bool copy_file_impl(FileDescriptor& read_fd, FileDescriptor& write_fd, error_code& ec) { + ifstream in; + in.__open(read_fd.fd, ios::binary); + if (!in.is_open()) { + // This assumes that __open didn't reset the error code. + ec = capture_errno(); + return false; + } + read_fd.fd = -1; + ofstream out; + out.__open(write_fd.fd, ios::binary); + if (!out.is_open()) { + ec = capture_errno(); + return false; + } + write_fd.fd = -1; + + if (in.good() && out.good()) { + using InIt = istreambuf_iterator<char>; + using OutIt = ostreambuf_iterator<char>; + InIt bin(in); + InIt ein; + OutIt bout(out); + copy(bin, ein, bout); + } + if (out.fail() || in.fail()) { + ec = make_error_code(errc::io_error); + return false; + } - ec.clear(); - return true; -} - -bool copy_file_impl(FileDescriptor& from, FileDescriptor& to, error_code& ec) { -#if defined(_LIBCPP_USE_SENDFILE) - return copy_file_impl_sendfile(from, to, ec); -#elif defined(_LIBCPP_USE_COPYFILE) - return copy_file_impl_copyfile(from, to, ec); + ec.clear(); + return true; + } #else - return copy_file_impl_default(from, to, ec); -#endif -} +# error "Unknown implementation for copy_file_impl" +#endif // copy_file_impl implementation -} // namespace -} // namespace detail +} // end anonymous namespace +} // end namespace detail bool __copy_file(const path& from, const path& to, copy_options options, error_code* ec) { @@ -870,8 +946,17 @@ bool __create_directory(const path& p, error_code* ec) { if (::mkdir(p.c_str(), static_cast<int>(perms::all)) == 0) return true; - if (errno != EEXIST) + + if (errno == EEXIST) { + error_code mec = capture_errno(); + error_code ignored_ec; + const file_status st = status(p, ignored_ec); + if (!is_directory(st)) { + err.report(mec); + } + } else { err.report(capture_errno()); + } return false; } @@ -889,8 +974,17 @@ bool __create_directory(path const& p, path const& attributes, error_code* ec) { if (::mkdir(p.c_str(), attr_stat.st_mode) == 0) return true; - if (errno != EEXIST) + + if (errno == EEXIST) { + error_code mec = capture_errno(); + error_code ignored_ec; + const file_status st = status(p, ignored_ec); + if (!is_directory(st)) { + err.report(mec); + } + } else { err.report(capture_errno()); + } return false; } @@ -1225,10 +1319,10 @@ path __temp_directory_path(error_code* ec) { error_code m_ec; file_status st = detail::posix_stat(p, &m_ec); if (!status_known(st)) - return err.report(m_ec, "cannot access path \"%s\"", p); + return err.report(m_ec, "cannot access path \"" PS_FMT "\"", p); if (!exists(st) || !is_directory(st)) - return err.report(errc::not_a_directory, "path \"%s\" is not a directory", + return err.report(errc::not_a_directory, "path \"" PS_FMT "\" is not a directory", p); return p; @@ -1284,7 +1378,7 @@ path& path::replace_extension(path const& replacement) { } if (!replacement.empty()) { if (replacement.native()[0] != '.') { - __pn_ += "."; + __pn_ += PS("."); } __pn_.append(replacement.__pn_); } @@ -1314,7 +1408,7 @@ string_view_t path::__root_path_raw() const { auto PP = PathParser::CreateBegin(__pn_); if (PP.State == PathParser::PS_InRootName) { auto NextCh = PP.peek(); - if (NextCh && *NextCh == '/') { + if (NextCh && isSeparator(*NextCh)) { ++PP; return createView(__pn_.data(), &PP.RawEntry.back()); } @@ -1406,12 +1500,16 @@ enum PathPartKind : unsigned char { static PathPartKind ClassifyPathPart(string_view_t Part) { if (Part.empty()) return PK_TrailingSep; - if (Part == ".") + if (Part == PS(".")) return PK_Dot; - if (Part == "..") + if (Part == PS("..")) return PK_DotDot; - if (Part == "/") + if (Part == PS("/")) + return PK_RootSep; +#if defined(_LIBCPP_WIN32API) + if (Part == PS("\\")) return PK_RootSep; +#endif return PK_Filename; } @@ -1459,7 +1557,7 @@ path path::lexically_normal() const { NewPathSize -= Parts.back().first.size(); Parts.pop_back(); } else if (LastKind != PK_RootSep) - AddPart(PK_DotDot, ".."); + AddPart(PK_DotDot, PS("..")); MaybeNeedTrailingSep = LastKind == PK_Filename; break; } @@ -1474,7 +1572,7 @@ path path::lexically_normal() const { } // [fs.path.generic]p6.8: If the path is empty, add a dot. if (Parts.empty()) - return "."; + return PS("."); // [fs.path.generic]p6.7: If the last filename is dot-dot, remove any // trailing directory-separator. @@ -1486,7 +1584,7 @@ path path::lexically_normal() const { Result /= PK.first; if (NeedTrailingSep) - Result /= ""; + Result /= PS(""); return Result; } @@ -1495,9 +1593,9 @@ static int DetermineLexicalElementCount(PathParser PP) { int Count = 0; for (; PP; ++PP) { auto Elem = *PP; - if (Elem == "..") + if (Elem == PS("..")) --Count; - else if (Elem != "." && Elem != "") + else if (Elem != PS(".") && Elem != PS("")) ++Count; } return Count; @@ -1544,15 +1642,15 @@ path path::lexically_relative(const path& base) const { return {}; // if n == 0 and (a == end() || a->empty()), returns path("."); otherwise - if (ElemCount == 0 && (PP.atEnd() || *PP == "")) - return "."; + if (ElemCount == 0 && (PP.atEnd() || *PP == PS(""))) + return PS("."); // return a path constructed with 'n' dot-dot elements, followed by the the // elements of '*this' after the mismatch. path Result; // FIXME: Reserve enough room in Result that it won't have to re-allocate. while (ElemCount--) - Result /= ".."; + Result /= PS(".."); for (; PP; ++PP) Result /= *PP; return Result; @@ -1565,7 +1663,7 @@ static int CompareRootName(PathParser *LHS, PathParser *RHS) { return 0; auto GetRootName = [](PathParser *Parser) -> string_view_t { - return Parser->inRootName() ? **Parser : ""; + return Parser->inRootName() ? **Parser : PS(""); }; int res = GetRootName(LHS).compare(GetRootName(RHS)); ConsumeRootName(LHS); @@ -1674,6 +1772,36 @@ path::iterator& path::iterator::__decrement() { return *this; } +#if defined(_LIBCPP_WIN32API) +//////////////////////////////////////////////////////////////////////////// +// Windows path conversions +size_t __wide_to_char(const wstring &str, char *out, size_t outlen) { + if (str.empty()) + return 0; + ErrorHandler<size_t> err("__wide_to_char", nullptr); + UINT codepage = AreFileApisANSI() ? CP_ACP : CP_OEMCP; + BOOL used_default = FALSE; + int ret = WideCharToMultiByte(codepage, 0, str.data(), str.size(), out, + outlen, nullptr, &used_default); + if (ret <= 0 || used_default) + return err.report(errc::illegal_byte_sequence); + return ret; +} + +size_t __char_to_wide(const string &str, wchar_t *out, size_t outlen) { + if (str.empty()) + return 0; + ErrorHandler<size_t> err("__char_to_wide", nullptr); + UINT codepage = AreFileApisANSI() ? CP_ACP : CP_OEMCP; + int ret = MultiByteToWideChar(codepage, MB_ERR_INVALID_CHARS, str.data(), + str.size(), out, outlen); + if (ret <= 0) + return err.report(errc::illegal_byte_sequence); + return ret; +} +#endif + + /////////////////////////////////////////////////////////////////////////////// // directory entry definitions /////////////////////////////////////////////////////////////////////////////// diff --git a/lib/libcxx/src/include/config_elast.h b/lib/libcxx/src/include/config_elast.h index 501cbc4ffe..3113f9fb5c 100644 --- a/lib/libcxx/src/include/config_elast.h +++ b/lib/libcxx/src/include/config_elast.h @@ -17,10 +17,14 @@ #include <errno.h> #endif +// Note: _LIBCPP_ELAST needs to be defined only on platforms +// where strerror/strerror_r can't handle out-of-range errno values. #if defined(ELAST) #define _LIBCPP_ELAST ELAST #elif defined(_NEWLIB_VERSION) #define _LIBCPP_ELAST __ELASTERROR +#elif defined(__NuttX__) +// No _LIBCPP_ELAST needed on NuttX #elif defined(__Fuchsia__) // No _LIBCPP_ELAST needed on Fuchsia #elif defined(__wasi__) diff --git a/lib/libcxx/src/include/refstring.h b/lib/libcxx/src/include/refstring.h index e464b79ba8..cefd7caf0f 100644 --- a/lib/libcxx/src/include/refstring.h +++ b/lib/libcxx/src/include/refstring.h @@ -13,12 +13,25 @@ #include <stdexcept> #include <cstddef> #include <cstring> -#ifdef __APPLE__ -#include <dlfcn.h> -#include <mach-o/dyld.h> -#endif #include "atomic_support.h" +// MacOS and iOS used to ship with libstdc++, and still support old applications +// linking against libstdc++. The libc++ and libstdc++ exceptions are supposed +// to be ABI compatible, such that they can be thrown from one library and caught +// in the other. +// +// For that reason, we must look for libstdc++ in the same process and if found, +// check the string stored in the exception object to see if it is the GCC empty +// string singleton before manipulating the reference count. This is done so that +// if an exception is created with a zero-length string in libstdc++, libc++abi +// won't try to delete the memory. +#if defined(__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__) || \ + defined(__ENVIRONMENT_IPHONE_OS_VERSION_MIN_REQUIRED__) +# define _LIBCPP_CHECK_FOR_GCC_EMPTY_STRING_STORAGE +# include <dlfcn.h> +# include <mach-o/dyld.h> +#endif + _LIBCPP_BEGIN_NAMESPACE_STD namespace __refstring_imp { namespace { @@ -40,7 +53,7 @@ inline char * data_from_rep(_Rep_base *rep) noexcept { return data + sizeof(*rep); } -#if defined(__APPLE__) +#if defined(_LIBCPP_CHECK_FOR_GCC_EMPTY_STRING_STORAGE) inline const char* compute_gcc_empty_string_storage() _NOEXCEPT { @@ -115,7 +128,7 @@ __libcpp_refstring::~__libcpp_refstring() { inline bool __libcpp_refstring::__uses_refcount() const { -#ifdef __APPLE__ +#if defined(_LIBCPP_CHECK_FOR_GCC_EMPTY_STRING_STORAGE) return __imp_ != get_gcc_empty_string_storage(); #else return true; diff --git a/lib/libcxx/src/ios.cpp b/lib/libcxx/src/ios.cpp index 2dc84be828..3a92964657 100644 --- a/lib/libcxx/src/ios.cpp +++ b/lib/libcxx/src/ios.cpp @@ -15,30 +15,14 @@ #include "__locale" #include "algorithm" #include "include/config_elast.h" -#include "istream" #include "limits" #include "memory" #include "new" -#include "streambuf" #include "string" #include "__undef_macros" _LIBCPP_BEGIN_NAMESPACE_STD -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ios<char>; -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ios<wchar_t>; - -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_streambuf<char>; -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_streambuf<wchar_t>; - -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_istream<char>; -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_istream<wchar_t>; - -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ostream<char>; -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ostream<wchar_t>; - -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_iostream<char>; - class _LIBCPP_HIDDEN __iostream_category : public __do_message { diff --git a/lib/libcxx/src/ios.instantiations.cpp b/lib/libcxx/src/ios.instantiations.cpp new file mode 100644 index 0000000000..1a23687d12 --- /dev/null +++ b/lib/libcxx/src/ios.instantiations.cpp @@ -0,0 +1,43 @@ +//===----------------------------------------------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#include "__config" +#include "fstream" +#include "ios" +#include "istream" +#include "ostream" +#include "sstream" +#include "streambuf" + + +_LIBCPP_BEGIN_NAMESPACE_STD + +// Original explicit instantiations provided in the library +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ios<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ios<wchar_t>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_streambuf<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_streambuf<wchar_t>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_istream<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_istream<wchar_t>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ostream<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ostream<wchar_t>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_iostream<char>; + +// Additional instantiations added later. Whether programs rely on these being +// available is protected by _LIBCPP_ABI_ENABLE_ADDITIONAL_IOSTREAM_EXPLICIT_INSTANTIATIONS_1. +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_stringbuf<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_stringstream<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ostringstream<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_istringstream<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ifstream<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_ofstream<char>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS basic_filebuf<char>; + +// Add more here if needed... + +_LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/src/iostream.cpp b/lib/libcxx/src/iostream.cpp index ad1920abc6..ea95534c76 100644 --- a/lib/libcxx/src/iostream.cpp +++ b/lib/libcxx/src/iostream.cpp @@ -77,7 +77,7 @@ __asm__("?wclog@" _LIBCPP_ABI_NAMESPACE_STR "@std@@3V?$basic_ostream@_WU?$char_t #endif ; -_LIBCPP_HIDDEN ios_base::Init __start_std_streams; +_LIBCPP_HIDDEN ios_base::Init __start_std_streams _LIBCPP_INIT_PRIORITY_MAX; // On Windows the TLS storage for locales needs to be initialized before we create // the standard streams, otherwise it may not be alive during program termination diff --git a/lib/libcxx/src/locale.cpp b/lib/libcxx/src/locale.cpp index b9180880e4..a0209d0ce8 100644 --- a/lib/libcxx/src/locale.cpp +++ b/lib/libcxx/src/locale.cpp @@ -29,8 +29,8 @@ #include "cwctype" #include "__sso_allocator" #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) -#include "support/win32/locale_win32.h" -#elif !defined(__BIONIC__) +#include "__support/win32/locale_win32.h" +#elif !defined(__BIONIC__) && !defined(__NuttX__) #include <langinfo.h> #endif #include <stdlib.h> @@ -158,7 +158,7 @@ const locale::category locale::all; class _LIBCPP_HIDDEN locale::__imp : public facet { - enum {N = 28}; + enum {N = 30}; #if defined(_LIBCPP_COMPILER_MSVC) // FIXME: MSVC doesn't support aligned parameters by value. // I can't get the __sso_allocator to work here @@ -202,8 +202,14 @@ locale::__imp::__imp(size_t refs) install(&make<_VSTD::ctype<wchar_t> >(1u)); install(&make<codecvt<char, char, mbstate_t> >(1u)); install(&make<codecvt<wchar_t, char, mbstate_t> >(1u)); +_LIBCPP_SUPPRESS_DEPRECATED_PUSH install(&make<codecvt<char16_t, char, mbstate_t> >(1u)); install(&make<codecvt<char32_t, char, mbstate_t> >(1u)); +_LIBCPP_SUPPRESS_DEPRECATED_POP +#ifndef _LIBCPP_NO_HAS_CHAR8_T + install(&make<codecvt<char16_t, char8_t, mbstate_t> >(1u)); + install(&make<codecvt<char32_t, char8_t, mbstate_t> >(1u)); +#endif install(&make<numpunct<char> >(1u)); install(&make<numpunct<wchar_t> >(1u)); install(&make<num_get<char> >(1u)); @@ -245,8 +251,14 @@ locale::__imp::__imp(const string& name, size_t refs) install(new ctype_byname<wchar_t>(name_)); install(new codecvt_byname<char, char, mbstate_t>(name_)); install(new codecvt_byname<wchar_t, char, mbstate_t>(name_)); +_LIBCPP_SUPPRESS_DEPRECATED_PUSH install(new codecvt_byname<char16_t, char, mbstate_t>(name_)); install(new codecvt_byname<char32_t, char, mbstate_t>(name_)); +_LIBCPP_SUPPRESS_DEPRECATED_POP +#ifndef _LIBCPP_NO_HAS_CHAR8_T + install(new codecvt_byname<char16_t, char8_t, mbstate_t>(name_)); + install(new codecvt_byname<char32_t, char8_t, mbstate_t>(name_)); +#endif install(new numpunct_byname<char>(name_)); install(new numpunct_byname<wchar_t>(name_)); install(new moneypunct_byname<char, false>(name_)); @@ -315,8 +327,14 @@ locale::__imp::__imp(const __imp& other, const string& name, locale::category c) install(new ctype_byname<wchar_t>(name)); install(new codecvt_byname<char, char, mbstate_t>(name)); install(new codecvt_byname<wchar_t, char, mbstate_t>(name)); +_LIBCPP_SUPPRESS_DEPRECATED_PUSH install(new codecvt_byname<char16_t, char, mbstate_t>(name)); install(new codecvt_byname<char32_t, char, mbstate_t>(name)); +_LIBCPP_SUPPRESS_DEPRECATED_POP +#ifndef _LIBCPP_NO_HAS_CHAR8_T + install(new codecvt_byname<char16_t, char8_t, mbstate_t>(name)); + install(new codecvt_byname<char32_t, char8_t, mbstate_t>(name)); +#endif } if (c & locale::monetary) { @@ -385,8 +403,14 @@ locale::__imp::__imp(const __imp& other, const __imp& one, locale::category c) install_from<_VSTD::ctype<char> >(one); install_from<_VSTD::ctype<wchar_t> >(one); install_from<_VSTD::codecvt<char, char, mbstate_t> >(one); +_LIBCPP_SUPPRESS_DEPRECATED_PUSH install_from<_VSTD::codecvt<char16_t, char, mbstate_t> >(one); install_from<_VSTD::codecvt<char32_t, char, mbstate_t> >(one); +_LIBCPP_SUPPRESS_DEPRECATED_POP +#ifndef _LIBCPP_NO_HAS_CHAR8_T + install_from<_VSTD::codecvt<char16_t, char8_t, mbstate_t> >(one); + install_from<_VSTD::codecvt<char32_t, char8_t, mbstate_t> >(one); +#endif install_from<_VSTD::codecvt<wchar_t, char, mbstate_t> >(one); } if (c & locale::monetary) @@ -1149,7 +1173,7 @@ ctype<char>::__classic_upper_table() _NOEXCEPT { return _LIBCPP_GET_C_LOCALE->__ctype_toupper; } -#elif __NetBSD__ +#elif defined(__NetBSD__) const short* ctype<char>::__classic_lower_table() _NOEXCEPT { @@ -3171,6 +3195,87 @@ codecvt<char16_t, char, mbstate_t>::do_max_length() const _NOEXCEPT return 4; } +#ifndef _LIBCPP_NO_HAS_CHAR8_T + +// template <> class codecvt<char16_t, char8_t, mbstate_t> + +locale::id codecvt<char16_t, char8_t, mbstate_t>::id; + +codecvt<char16_t, char8_t, mbstate_t>::~codecvt() +{ +} + +codecvt<char16_t, char8_t, mbstate_t>::result +codecvt<char16_t, char8_t, mbstate_t>::do_out(state_type&, + const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, + extern_type* to, extern_type* to_end, extern_type*& to_nxt) const +{ + const uint16_t* _frm = reinterpret_cast<const uint16_t*>(frm); + const uint16_t* _frm_end = reinterpret_cast<const uint16_t*>(frm_end); + const uint16_t* _frm_nxt = _frm; + uint8_t* _to = reinterpret_cast<uint8_t*>(to); + uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); + uint8_t* _to_nxt = _to; + result r = utf16_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); + frm_nxt = frm + (_frm_nxt - _frm); + to_nxt = to + (_to_nxt - _to); + return r; +} + +codecvt<char16_t, char8_t, mbstate_t>::result +codecvt<char16_t, char8_t, mbstate_t>::do_in(state_type&, + const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, + intern_type* to, intern_type* to_end, intern_type*& to_nxt) const +{ + const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); + const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); + const uint8_t* _frm_nxt = _frm; + uint16_t* _to = reinterpret_cast<uint16_t*>(to); + uint16_t* _to_end = reinterpret_cast<uint16_t*>(to_end); + uint16_t* _to_nxt = _to; + result r = utf8_to_utf16(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); + frm_nxt = frm + (_frm_nxt - _frm); + to_nxt = to + (_to_nxt - _to); + return r; +} + +codecvt<char16_t, char8_t, mbstate_t>::result +codecvt<char16_t, char8_t, mbstate_t>::do_unshift(state_type&, + extern_type* to, extern_type*, extern_type*& to_nxt) const +{ + to_nxt = to; + return noconv; +} + +int +codecvt<char16_t, char8_t, mbstate_t>::do_encoding() const _NOEXCEPT +{ + return 0; +} + +bool +codecvt<char16_t, char8_t, mbstate_t>::do_always_noconv() const _NOEXCEPT +{ + return false; +} + +int +codecvt<char16_t, char8_t, mbstate_t>::do_length(state_type&, + const extern_type* frm, const extern_type* frm_end, size_t mx) const +{ + const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); + const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); + return utf8_to_utf16_length(_frm, _frm_end, mx); +} + +int +codecvt<char16_t, char8_t, mbstate_t>::do_max_length() const _NOEXCEPT +{ + return 4; +} + +#endif + // template <> class codecvt<char32_t, char, mbstate_t> locale::id codecvt<char32_t, char, mbstate_t>::id; @@ -3248,6 +3353,87 @@ codecvt<char32_t, char, mbstate_t>::do_max_length() const _NOEXCEPT return 4; } +#ifndef _LIBCPP_NO_HAS_CHAR8_T + +// template <> class codecvt<char32_t, char8_t, mbstate_t> + +locale::id codecvt<char32_t, char8_t, mbstate_t>::id; + +codecvt<char32_t, char8_t, mbstate_t>::~codecvt() +{ +} + +codecvt<char32_t, char8_t, mbstate_t>::result +codecvt<char32_t, char8_t, mbstate_t>::do_out(state_type&, + const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, + extern_type* to, extern_type* to_end, extern_type*& to_nxt) const +{ + const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); + const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); + const uint32_t* _frm_nxt = _frm; + uint8_t* _to = reinterpret_cast<uint8_t*>(to); + uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); + uint8_t* _to_nxt = _to; + result r = ucs4_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); + frm_nxt = frm + (_frm_nxt - _frm); + to_nxt = to + (_to_nxt - _to); + return r; +} + +codecvt<char32_t, char8_t, mbstate_t>::result +codecvt<char32_t, char8_t, mbstate_t>::do_in(state_type&, + const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, + intern_type* to, intern_type* to_end, intern_type*& to_nxt) const +{ + const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); + const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); + const uint8_t* _frm_nxt = _frm; + uint32_t* _to = reinterpret_cast<uint32_t*>(to); + uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); + uint32_t* _to_nxt = _to; + result r = utf8_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); + frm_nxt = frm + (_frm_nxt - _frm); + to_nxt = to + (_to_nxt - _to); + return r; +} + +codecvt<char32_t, char8_t, mbstate_t>::result +codecvt<char32_t, char8_t, mbstate_t>::do_unshift(state_type&, + extern_type* to, extern_type*, extern_type*& to_nxt) const +{ + to_nxt = to; + return noconv; +} + +int +codecvt<char32_t, char8_t, mbstate_t>::do_encoding() const _NOEXCEPT +{ + return 0; +} + +bool +codecvt<char32_t, char8_t, mbstate_t>::do_always_noconv() const _NOEXCEPT +{ + return false; +} + +int +codecvt<char32_t, char8_t, mbstate_t>::do_length(state_type&, + const extern_type* frm, const extern_type* frm_end, size_t mx) const +{ + const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); + const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); + return utf8_to_ucs4_length(_frm, _frm_end, mx); +} + +int +codecvt<char32_t, char8_t, mbstate_t>::do_max_length() const _NOEXCEPT +{ + return 4; +} + +#endif + // __codecvt_utf8<wchar_t> __codecvt_utf8<wchar_t>::result @@ -5128,7 +5314,7 @@ __time_get_storage<wchar_t>::init(const ctype<wchar_t>& ct) mb = mbstate_t(); const char* bb = buf; size_t j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); - if (j == size_t(-1)) + if (j == size_t(-1) || j == 0) __throw_runtime_error("locale not supported"); wbe = wbuf + j; __weeks_[i].assign(wbuf, wbe); @@ -5136,7 +5322,7 @@ __time_get_storage<wchar_t>::init(const ctype<wchar_t>& ct) mb = mbstate_t(); bb = buf; j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); - if (j == size_t(-1)) + if (j == size_t(-1) || j == 0) __throw_runtime_error("locale not supported"); wbe = wbuf + j; __weeks_[i+7].assign(wbuf, wbe); @@ -5149,7 +5335,7 @@ __time_get_storage<wchar_t>::init(const ctype<wchar_t>& ct) mb = mbstate_t(); const char* bb = buf; size_t j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); - if (j == size_t(-1)) + if (j == size_t(-1) || j == 0) __throw_runtime_error("locale not supported"); wbe = wbuf + j; __months_[i].assign(wbuf, wbe); @@ -5157,7 +5343,7 @@ __time_get_storage<wchar_t>::init(const ctype<wchar_t>& ct) mb = mbstate_t(); bb = buf; j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); - if (j == size_t(-1)) + if (j == size_t(-1) || j == 0) __throw_runtime_error("locale not supported"); wbe = wbuf + j; __months_[i+12].assign(wbuf, wbe); @@ -6148,7 +6334,11 @@ template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS messages_byname<wchar_t> template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char, char, mbstate_t>; template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<wchar_t, char, mbstate_t>; -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char16_t, char, mbstate_t>; -template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char32_t, char, mbstate_t>; +template class _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char16_t, char, mbstate_t>; +template class _LIBCPP_DEPRECATED_IN_CXX20 _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char32_t, char, mbstate_t>; +#ifndef _LIBCPP_NO_HAS_CHAR8_T +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char16_t, char8_t, mbstate_t>; +template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char32_t, char8_t, mbstate_t>; +#endif _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/src/memory.cpp b/lib/libcxx/src/memory.cpp index 633c9a6f56..5a5894fd94 100644 --- a/lib/libcxx/src/memory.cpp +++ b/lib/libcxx/src/memory.cpp @@ -124,16 +124,12 @@ __shared_weak_count::lock() _NOEXCEPT return nullptr; } -#if !defined(_LIBCPP_NO_RTTI) || !defined(_LIBCPP_BUILD_STATIC) - const void* __shared_weak_count::__get_deleter(const type_info&) const _NOEXCEPT { return nullptr; } -#endif // _LIBCPP_NO_RTTI - #if !defined(_LIBCPP_HAS_NO_ATOMIC_HEADER) _LIBCPP_SAFE_STATIC static const std::size_t __sp_mut_count = 16; diff --git a/lib/libcxx/src/new.cpp b/lib/libcxx/src/new.cpp index 901e785658..9d01330ba7 100644 --- a/lib/libcxx/src/new.cpp +++ b/lib/libcxx/src/new.cpp @@ -64,7 +64,7 @@ operator new(std::size_t size) _THROW_BAD_ALLOC if (size == 0) size = 1; void* p; - while ((p = ::malloc(size)) == 0) + while ((p = ::malloc(size)) == nullptr) { // If malloc fails and there is a new_handler, // call it to try free up memory. @@ -85,7 +85,7 @@ _LIBCPP_WEAK void* operator new(size_t size, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCPP_NO_EXCEPTIONS try { @@ -111,7 +111,7 @@ _LIBCPP_WEAK void* operator new[](size_t size, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCPP_NO_EXCEPTIONS try { @@ -178,15 +178,16 @@ operator new(std::size_t size, std::align_val_t alignment) _THROW_BAD_ALLOC size = 1; if (static_cast<size_t>(alignment) < sizeof(void*)) alignment = std::align_val_t(sizeof(void*)); + + // Try allocating memory. If allocation fails and there is a new_handler, + // call it to try free up memory, and try again until it succeeds, or until + // the new_handler decides to terminate. + // + // If allocation fails and there is no new_handler, we throw bad_alloc + // (or return nullptr if exceptions are disabled). void* p; -#if defined(_LIBCPP_MSVCRT_LIKE) - while ((p = _aligned_malloc(size, static_cast<size_t>(alignment))) == nullptr) -#else - while (::posix_memalign(&p, static_cast<size_t>(alignment), size) != 0) -#endif + while ((p = std::__libcpp_aligned_alloc(static_cast<std::size_t>(alignment), size)) == nullptr) { - // If posix_memalign fails and there is a new_handler, - // call it to try free up memory. std::new_handler nh = std::get_new_handler(); if (nh) nh(); @@ -194,7 +195,6 @@ operator new(std::size_t size, std::align_val_t alignment) _THROW_BAD_ALLOC #ifndef _LIBCPP_NO_EXCEPTIONS throw std::bad_alloc(); #else - p = nullptr; // posix_memalign doesn't initialize 'p' on failure break; #endif } @@ -206,7 +206,7 @@ _LIBCPP_WEAK void* operator new(size_t size, std::align_val_t alignment, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCPP_NO_EXCEPTIONS try { @@ -232,7 +232,7 @@ _LIBCPP_WEAK void* operator new[](size_t size, std::align_val_t alignment, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCPP_NO_EXCEPTIONS try { @@ -251,11 +251,7 @@ _LIBCPP_WEAK void operator delete(void* ptr, std::align_val_t) _NOEXCEPT { -#if defined(_LIBCPP_MSVCRT_LIKE) - ::_aligned_free(ptr); -#else - ::free(ptr); -#endif + std::__libcpp_aligned_free(ptr); } _LIBCPP_WEAK diff --git a/lib/libcxx/src/optional.cpp b/lib/libcxx/src/optional.cpp index 3aaf5ea553..86d013b350 100644 --- a/lib/libcxx/src/optional.cpp +++ b/lib/libcxx/src/optional.cpp @@ -7,6 +7,7 @@ //===----------------------------------------------------------------------===// #include "optional" +#include "__availability" namespace std { diff --git a/lib/libcxx/src/random.cpp b/lib/libcxx/src/random.cpp index 04adc59f9b..29aa43b1e1 100644 --- a/lib/libcxx/src/random.cpp +++ b/lib/libcxx/src/random.cpp @@ -13,6 +13,7 @@ #define _CRT_RAND_S #endif // defined(_LIBCPP_USING_WIN32_RANDOM) +#include "limits" #include "random" #include "system_error" @@ -29,6 +30,10 @@ #elif defined(_LIBCPP_USING_DEV_RANDOM) #include <fcntl.h> #include <unistd.h> +#if __has_include(<sys/ioctl.h>) && __has_include(<linux/random.h>) +#include <sys/ioctl.h> +#include <linux/random.h> +#endif #elif defined(_LIBCPP_USING_NACL_RANDOM) #include <nacl/nacl_random.h> #endif @@ -172,7 +177,23 @@ random_device::operator()() double random_device::entropy() const _NOEXCEPT { +#if defined(_LIBCPP_USING_DEV_RANDOM) && defined(RNDGETENTCNT) + int ent; + if (::ioctl(__f_, RNDGETENTCNT, &ent) < 0) + return 0; + + if (ent < 0) return 0; + + if (ent > std::numeric_limits<result_type>::digits) + return std::numeric_limits<result_type>::digits; + + return ent; +#elif defined(__OpenBSD__) + return std::numeric_limits<result_type>::digits; +#else + return 0; +#endif } _LIBCPP_END_NAMESPACE_STD diff --git a/lib/libcxx/src/support/runtime/exception_fallback.ipp b/lib/libcxx/src/support/runtime/exception_fallback.ipp index 376a0381f5..faa112f7f3 100644 --- a/lib/libcxx/src/support/runtime/exception_fallback.ipp +++ b/lib/libcxx/src/support/runtime/exception_fallback.ipp @@ -49,7 +49,6 @@ get_terminate() _NOEXCEPT return __libcpp_atomic_load(&__terminate_handler); } -#ifndef __EMSCRIPTEN__ // We provide this in JS _LIBCPP_NORETURN void terminate() _NOEXCEPT @@ -72,9 +71,7 @@ terminate() _NOEXCEPT } #endif // _LIBCPP_NO_EXCEPTIONS } -#endif // !__EMSCRIPTEN__ -#if !defined(__EMSCRIPTEN__) bool uncaught_exception() _NOEXCEPT { return uncaught_exceptions() > 0; } int uncaught_exceptions() _NOEXCEPT @@ -83,7 +80,6 @@ int uncaught_exceptions() _NOEXCEPT fprintf(stderr, "uncaught_exceptions not yet implemented\n"); ::abort(); } -#endif // !__EMSCRIPTEN__ exception::~exception() _NOEXCEPT diff --git a/lib/libcxx/src/support/solaris/xlocale.cpp b/lib/libcxx/src/support/solaris/xlocale.cpp index d68a39f4df..d25adcd21d 100644 --- a/lib/libcxx/src/support/solaris/xlocale.cpp +++ b/lib/libcxx/src/support/solaris/xlocale.cpp @@ -8,7 +8,7 @@ #ifdef __sun__ -#include "support/solaris/xlocale.h" +#include "__support/solaris/xlocale.h" #include <stdarg.h> #include <stdio.h> #include <sys/localedef.h> diff --git a/lib/libcxx/src/support/win32/locale_win32.cpp b/lib/libcxx/src/support/win32/locale_win32.cpp index b7062db352..e7c6005fc1 100644 --- a/lib/libcxx/src/support/win32/locale_win32.cpp +++ b/lib/libcxx/src/support/win32/locale_win32.cpp @@ -1,5 +1,5 @@ // -*- C++ -*- -//===-------------------- support/win32/locale_win32.cpp ------------------===// +//===----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/src/support/win32/support.cpp b/lib/libcxx/src/support/win32/support.cpp index d156e02e3e..52453f5479 100644 --- a/lib/libcxx/src/support/win32/support.cpp +++ b/lib/libcxx/src/support/win32/support.cpp @@ -1,5 +1,5 @@ // -*- C++ -*- -//===----------------------- support/win32/support.h ----------------------===// +//===----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/src/support/win32/thread_win32.cpp b/lib/libcxx/src/support/win32/thread_win32.cpp index 83e7e9f6ce..35c4c87145 100644 --- a/lib/libcxx/src/support/win32/thread_win32.cpp +++ b/lib/libcxx/src/support/win32/thread_win32.cpp @@ -1,5 +1,5 @@ // -*- C++ -*- -//===-------------------- support/win32/thread_win32.cpp ------------------===// +//===----------------------------------------------------------------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. diff --git a/lib/libcxx/src/thread.cpp b/lib/libcxx/src/thread.cpp index 5f44e9e40f..5959d8b711 100644 --- a/lib/libcxx/src/thread.cpp +++ b/lib/libcxx/src/thread.cpp @@ -14,17 +14,9 @@ #include "vector" #include "future" #include "limits" -#include <sys/types.h> - -#if defined(__unix__) || (defined(__APPLE__) && defined(__MACH__)) -# include <sys/param.h> -# if defined(__FreeBSD__) || defined(__NetBSD__) || defined(__OpenBSD__) || defined(__DragonFly__) || defined(__APPLE__) -# include <sys/sysctl.h> -# endif -#endif // defined(__unix__) || (defined(__APPLE__) && defined(__MACH__)) #if __has_include(<unistd.h>) -#include <unistd.h> +# include <unistd.h> // for sysconf #endif #if defined(__NetBSD__) @@ -80,13 +72,7 @@ thread::detach() unsigned thread::hardware_concurrency() _NOEXCEPT { -#if defined(CTL_HW) && defined(HW_NCPU) - unsigned n; - int mib[2] = {CTL_HW, HW_NCPU}; - std::size_t s = sizeof(n); - sysctl(mib, 2, &n, &s, 0, 0); - return n; -#elif defined(_SC_NPROCESSORS_ONLN) +#if defined(_SC_NPROCESSORS_ONLN) long result = sysconf(_SC_NPROCESSORS_ONLN); // sysconf returns -1 if the name is invalid, the option does not exist or // does not have a definite limit. diff --git a/lib/libcxxabi/include/__cxxabi_config.h b/lib/libcxxabi/include/__cxxabi_config.h index b5444d67ee..cffedb88df 100644 --- a/lib/libcxxabi/include/__cxxabi_config.h +++ b/lib/libcxxabi/include/__cxxabi_config.h @@ -18,6 +18,19 @@ #define __has_attribute(_attribute_) 0 #endif +#if defined(__clang__) +# define _LIBCXXABI_COMPILER_CLANG +# ifndef __apple_build_version__ +# define _LIBCXXABI_CLANG_VER (__clang_major__ * 100 + __clang_minor__) +# endif +#elif defined(__GNUC__) +# define _LIBCXXABI_COMPILER_GCC +#elif defined(_MSC_VER) +# define _LIBCXXABI_COMPILER_MSVC +#elif defined(__IBMCPP__) +# define _LIBCXXABI_COMPILER_IBM +#endif + #if defined(_WIN32) #if defined(_LIBCXXABI_DISABLE_VISIBILITY_ANNOTATIONS) #define _LIBCXXABI_HIDDEN @@ -53,7 +66,7 @@ #endif #endif -#if defined(_WIN32) +#if defined(_LIBCXXABI_COMPILER_MSVC) #define _LIBCXXABI_WEAK #else #define _LIBCXXABI_WEAK __attribute__((__weak__)) @@ -72,7 +85,7 @@ #endif // wasm32 follows the arm32 ABI convention of using 32-bit guard. -#if defined(__arm__) || defined(__wasm32__) +#if defined(__arm__) || defined(__wasm32__) || defined(__ARM64_ARCH_8_32__) # define _LIBCXXABI_GUARD_ABI_ARM #endif diff --git a/lib/libcxxabi/include/cxxabi.h b/lib/libcxxabi/include/cxxabi.h index d21d3e1e23..43ce6f5f74 100644 --- a/lib/libcxxabi/include/cxxabi.h +++ b/lib/libcxxabi/include/cxxabi.h @@ -21,6 +21,7 @@ #define _LIBCPPABI_VERSION 1002 #define _LIBCXXABI_NORETURN __attribute__((noreturn)) +#define _LIBCXXABI_ALWAYS_COLD __attribute__((cold)) #ifdef __cplusplus @@ -78,13 +79,13 @@ extern _LIBCXXABI_FUNC_VIS _LIBCXXABI_NORETURN void __cxa_deleted_virtual(void); // 3.3.2 One-time Construction API #if defined(_LIBCXXABI_GUARD_ABI_ARM) -extern _LIBCXXABI_FUNC_VIS int __cxa_guard_acquire(uint32_t *); -extern _LIBCXXABI_FUNC_VIS void __cxa_guard_release(uint32_t *); -extern _LIBCXXABI_FUNC_VIS void __cxa_guard_abort(uint32_t *); +extern _LIBCXXABI_FUNC_VIS _LIBCXXABI_ALWAYS_COLD int __cxa_guard_acquire(uint32_t *); +extern _LIBCXXABI_FUNC_VIS _LIBCXXABI_ALWAYS_COLD void __cxa_guard_release(uint32_t *); +extern _LIBCXXABI_FUNC_VIS _LIBCXXABI_ALWAYS_COLD void __cxa_guard_abort(uint32_t *); #else -extern _LIBCXXABI_FUNC_VIS int __cxa_guard_acquire(uint64_t *); -extern _LIBCXXABI_FUNC_VIS void __cxa_guard_release(uint64_t *); -extern _LIBCXXABI_FUNC_VIS void __cxa_guard_abort(uint64_t *); +extern _LIBCXXABI_FUNC_VIS _LIBCXXABI_ALWAYS_COLD int __cxa_guard_acquire(uint64_t *); +extern _LIBCXXABI_FUNC_VIS _LIBCXXABI_ALWAYS_COLD void __cxa_guard_release(uint64_t *); +extern _LIBCXXABI_FUNC_VIS _LIBCXXABI_ALWAYS_COLD void __cxa_guard_abort(uint64_t *); #endif // 3.3.3 Array Construction and Destruction API @@ -136,9 +137,9 @@ __cxa_vec_cctor(void *dest_array, void *src_array, size_t element_count, void (*destructor)(void *)); // 3.3.5.3 Runtime API -extern _LIBCXXABI_FUNC_VIS int __cxa_atexit(void (*f)(void *), void *p, - void *d); -extern _LIBCXXABI_FUNC_VIS int __cxa_finalize(void *); +// These functions are part of the C++ ABI, but they are not defined in libc++abi: +// int __cxa_atexit(void (*)(void *), void *, void *); +// void __cxa_finalize(void *); // 3.4 Demangler API extern _LIBCXXABI_FUNC_VIS char *__cxa_demangle(const char *mangled_name, diff --git a/lib/libcxxabi/src/cxa_default_handlers.cpp b/lib/libcxxabi/src/cxa_default_handlers.cpp index d2f823d2b7..a24ee01453 100644 --- a/lib/libcxxabi/src/cxa_default_handlers.cpp +++ b/lib/libcxxabi/src/cxa_default_handlers.cpp @@ -45,6 +45,7 @@ static void demangling_terminate_handler() exception_header + 1; const __shim_type_info* thrown_type = static_cast<const __shim_type_info*>(exception_header->exceptionType); +#if !defined(LIBCXXABI_NON_DEMANGLING_TERMINATE) // Try to get demangled name of thrown_type int status; char buf[1024]; @@ -52,6 +53,9 @@ static void demangling_terminate_handler() const char* name = __cxa_demangle(thrown_type->name(), buf, &len, &status); if (status != 0) name = thrown_type->name(); +#else + const char* name = thrown_type->name(); +#endif // If the uncaught exception can be caught with std::exception& const __shim_type_info* catch_type = static_cast<const __shim_type_info*>(&typeid(std::exception)); diff --git a/lib/libcxxabi/src/cxa_guard_impl.h b/lib/libcxxabi/src/cxa_guard_impl.h index f6a698e23a..6f873f241f 100644 --- a/lib/libcxxabi/src/cxa_guard_impl.h +++ b/lib/libcxxabi/src/cxa_guard_impl.h @@ -54,6 +54,14 @@ #endif #endif +#if defined(__clang__) +# pragma clang diagnostic push +# pragma clang diagnostic ignored "-Wtautological-pointer-compare" +#elif defined(__GNUC__) +# pragma GCC diagnostic push +# pragma GCC diagnostic ignored "-Waddress" +#endif + // To make testing possible, this header is included from both cxa_guard.cpp // and a number of tests. // @@ -112,25 +120,25 @@ class AtomicInt { public: using MemoryOrder = std::__libcpp_atomic_order; - explicit AtomicInt(IntType *b) : b(b) {} + explicit AtomicInt(IntType *b) : b_(b) {} AtomicInt(AtomicInt const&) = delete; AtomicInt& operator=(AtomicInt const&) = delete; IntType load(MemoryOrder ord) { - return std::__libcpp_atomic_load(b, ord); + return std::__libcpp_atomic_load(b_, ord); } void store(IntType val, MemoryOrder ord) { - std::__libcpp_atomic_store(b, val, ord); + std::__libcpp_atomic_store(b_, val, ord); } IntType exchange(IntType new_val, MemoryOrder ord) { - return std::__libcpp_atomic_exchange(b, new_val, ord); + return std::__libcpp_atomic_exchange(b_, new_val, ord); } bool compare_exchange(IntType *expected, IntType desired, MemoryOrder ord_success, MemoryOrder ord_failure) { - return std::__libcpp_atomic_compare_exchange(b, expected, desired, ord_success, ord_failure); + return std::__libcpp_atomic_compare_exchange(b_, expected, desired, ord_success, ord_failure); } private: - IntType *b; + IntType *b_; }; //===----------------------------------------------------------------------===// @@ -154,14 +162,7 @@ constexpr uint32_t (*PlatformThreadID)() = nullptr; constexpr bool PlatformSupportsThreadID() { -#ifdef __clang__ -#pragma clang diagnostic push -#pragma clang diagnostic ignored "-Wtautological-pointer-compare" -#endif return +PlatformThreadID != nullptr; -#ifdef __clang__ -#pragma clang diagnostic pop -#endif } //===----------------------------------------------------------------------===// @@ -375,18 +376,18 @@ private: LockGuard& operator=(LockGuard const&) = delete; explicit LockGuard(const char* calling_func) - : calling_func(calling_func) { + : calling_func_(calling_func) { if (global_mutex.lock()) - ABORT_WITH_MESSAGE("%s failed to acquire mutex", calling_func); + ABORT_WITH_MESSAGE("%s failed to acquire mutex", calling_func_); } ~LockGuard() { if (global_mutex.unlock()) - ABORT_WITH_MESSAGE("%s failed to release mutex", calling_func); + ABORT_WITH_MESSAGE("%s failed to release mutex", calling_func_); } private: - const char* const calling_func; + const char* const calling_func_; }; }; @@ -411,14 +412,7 @@ constexpr void (*PlatformFutexWake)(int*) = nullptr; #endif constexpr bool PlatformSupportsFutex() { -#ifdef __clang__ -#pragma clang diagnostic push -#pragma clang diagnostic ignored "-Wtautological-pointer-compare" -#endif return +PlatformFutexWait != nullptr; -#ifdef __clang__ -#pragma clang diagnostic pop -#endif } /// InitByteFutex - Manages initialization using atomics and the futex syscall @@ -589,4 +583,10 @@ using SelectedImplementation = } // end namespace } // end namespace __cxxabiv1 +#if defined(__clang__) +# pragma clang diagnostic pop +#elif defined(__GNUC__) +# pragma GCC diagnostic pop +#endif + #endif // LIBCXXABI_SRC_INCLUDE_CXA_GUARD_IMPL_H diff --git a/lib/libcxxabi/src/cxa_personality.cpp b/lib/libcxxabi/src/cxa_personality.cpp index f276257761..81aa85165d 100644 --- a/lib/libcxxabi/src/cxa_personality.cpp +++ b/lib/libcxxabi/src/cxa_personality.cpp @@ -684,27 +684,21 @@ static void scan_eh_tab(scan_results &results, _Unwind_Action actions, return; } landingPad = (uintptr_t)lpStart + landingPad; + results.landingPad = landingPad; #else // __USING_SJLJ_EXCEPTIONS__ ++landingPad; #endif // __USING_SJLJ_EXCEPTIONS__ if (actionEntry == 0) { // Found a cleanup - // If this is a type 1 or type 2 search, there are no handlers - // If this is a type 3 search, you want to install the cleanup. - if ((actions & _UA_CLEANUP_PHASE) && !(actions & _UA_HANDLER_FRAME)) - { - results.ttypeIndex = 0; // Redundant but clarifying - results.landingPad = landingPad; - results.reason = _URC_HANDLER_FOUND; - return; - } - // No handler here - results.reason = _URC_CONTINUE_UNWIND; + results.reason = actions & _UA_SEARCH_PHASE + ? _URC_CONTINUE_UNWIND + : _URC_HANDLER_FOUND; return; } // Convert 1-based byte offset into const uint8_t* action = actionTableStart + (actionEntry - 1); + bool hasCleanup = false; // Scan action entries until you find a matching handler, cleanup, or the end of action list while (true) { @@ -720,27 +714,17 @@ static void scan_eh_tab(scan_results &results, _Unwind_Action actions, native_exception, unwind_exception); if (catchType == 0) { - // Found catch (...) catches everything, including foreign exceptions - // If this is a type 1 search save state and return _URC_HANDLER_FOUND - // If this is a type 2 search save state and return _URC_HANDLER_FOUND - // If this is a type 3 search !_UA_FORCE_UNWIND, we should have found this in phase 1! - // If this is a type 3 search _UA_FORCE_UNWIND, ignore handler and continue scan - if ((actions & _UA_SEARCH_PHASE) || (actions & _UA_HANDLER_FRAME)) - { - // Save state and return _URC_HANDLER_FOUND - results.ttypeIndex = ttypeIndex; - results.actionRecord = actionRecord; - results.landingPad = landingPad; - results.adjustedPtr = get_thrown_object_ptr(unwind_exception); - results.reason = _URC_HANDLER_FOUND; - return; - } - else if (!(actions & _UA_FORCE_UNWIND)) - { - // It looks like the exception table has changed - // on us. Likely stack corruption! - call_terminate(native_exception, unwind_exception); - } + // Found catch (...) catches everything, including + // foreign exceptions. This is search phase, cleanup + // phase with foreign exception, or forced unwinding. + assert(actions & (_UA_SEARCH_PHASE | _UA_HANDLER_FRAME | + _UA_FORCE_UNWIND)); + results.ttypeIndex = ttypeIndex; + results.actionRecord = actionRecord; + results.adjustedPtr = + get_thrown_object_ptr(unwind_exception); + results.reason = _URC_HANDLER_FOUND; + return; } // Else this is a catch (T) clause and will never // catch a foreign exception @@ -757,36 +741,25 @@ static void scan_eh_tab(scan_results &results, _Unwind_Action actions, } if (catchType->can_catch(excpType, adjustedPtr)) { - // Found a matching handler - // If this is a type 1 search save state and return _URC_HANDLER_FOUND - // If this is a type 3 search and !_UA_FORCE_UNWIND, we should have found this in phase 1! - // If this is a type 3 search and _UA_FORCE_UNWIND, ignore handler and continue scan - if (actions & _UA_SEARCH_PHASE) - { - // Save state and return _URC_HANDLER_FOUND - results.ttypeIndex = ttypeIndex; - results.actionRecord = actionRecord; - results.landingPad = landingPad; - results.adjustedPtr = adjustedPtr; - results.reason = _URC_HANDLER_FOUND; - return; - } - else if (!(actions & _UA_FORCE_UNWIND)) - { - // It looks like the exception table has changed - // on us. Likely stack corruption! - call_terminate(native_exception, unwind_exception); - } + // Found a matching handler. This is either search + // phase or forced unwinding. + assert(actions & + (_UA_SEARCH_PHASE | _UA_FORCE_UNWIND)); + results.ttypeIndex = ttypeIndex; + results.actionRecord = actionRecord; + results.adjustedPtr = adjustedPtr; + results.reason = _URC_HANDLER_FOUND; + return; } } // Scan next action ... } else if (ttypeIndex < 0) { - // Found an exception spec. If this is a foreign exception, - // it is always caught. - if (native_exception) - { + // Found an exception specification. + if (actions & _UA_FORCE_UNWIND) { + // Skip if forced unwinding. + } else if (native_exception) { // Does the exception spec catch this native exception? __cxa_exception* exception_header = (__cxa_exception*)(unwind_exception+1) - 1; void* adjustedPtr = get_thrown_object_ptr(unwind_exception); @@ -801,77 +774,38 @@ static void scan_eh_tab(scan_results &results, _Unwind_Action actions, ttypeEncoding, excpType, adjustedPtr, unwind_exception)) { - // native exception caught by exception spec - // If this is a type 1 search, save state and return _URC_HANDLER_FOUND - // If this is a type 3 search !_UA_FORCE_UNWIND, we should have found this in phase 1! - // If this is a type 3 search _UA_FORCE_UNWIND, ignore handler and continue scan - if (actions & _UA_SEARCH_PHASE) - { - // Save state and return _URC_HANDLER_FOUND - results.ttypeIndex = ttypeIndex; - results.actionRecord = actionRecord; - results.landingPad = landingPad; - results.adjustedPtr = adjustedPtr; - results.reason = _URC_HANDLER_FOUND; - return; - } - else if (!(actions & _UA_FORCE_UNWIND)) - { - // It looks like the exception table has changed - // on us. Likely stack corruption! - call_terminate(native_exception, unwind_exception); - } - } - } - else - { - // foreign exception caught by exception spec - // If this is a type 1 search, save state and return _URC_HANDLER_FOUND - // If this is a type 2 search, save state and return _URC_HANDLER_FOUND - // If this is a type 3 search !_UA_FORCE_UNWIND, we should have found this in phase 1! - // If this is a type 3 search _UA_FORCE_UNWIND, ignore handler and continue scan - if ((actions & _UA_SEARCH_PHASE) || (actions & _UA_HANDLER_FRAME)) - { - // Save state and return _URC_HANDLER_FOUND + // Native exception caught by exception + // specification. + assert(actions & _UA_SEARCH_PHASE); results.ttypeIndex = ttypeIndex; results.actionRecord = actionRecord; - results.landingPad = landingPad; - results.adjustedPtr = get_thrown_object_ptr(unwind_exception); + results.adjustedPtr = adjustedPtr; results.reason = _URC_HANDLER_FOUND; return; } - else if (!(actions & _UA_FORCE_UNWIND)) - { - // It looks like the exception table has changed - // on us. Likely stack corruption! - call_terminate(native_exception, unwind_exception); - } - } - // Scan next action ... - } - else // ttypeIndex == 0 - { - // Found a cleanup - // If this is a type 1 search, ignore it and continue scan - // If this is a type 2 search, ignore it and continue scan - // If this is a type 3 search, save state and return _URC_HANDLER_FOUND - if ((actions & _UA_CLEANUP_PHASE) && !(actions & _UA_HANDLER_FRAME)) - { - // Save state and return _URC_HANDLER_FOUND + } else { + // foreign exception caught by exception spec results.ttypeIndex = ttypeIndex; results.actionRecord = actionRecord; - results.landingPad = landingPad; - results.adjustedPtr = get_thrown_object_ptr(unwind_exception); + results.adjustedPtr = + get_thrown_object_ptr(unwind_exception); results.reason = _URC_HANDLER_FOUND; return; } + // Scan next action ... + } else { + hasCleanup = true; } const uint8_t* temp = action; int64_t actionOffset = readSLEB128(&temp); if (actionOffset == 0) { - // End of action list, no matching handler or cleanup found - results.reason = _URC_CONTINUE_UNWIND; + // End of action list. If this is phase 2 and we have found + // a cleanup (ttypeIndex=0), return _URC_HANDLER_FOUND; + // otherwise return _URC_CONTINUE_UNWIND. + results.reason = hasCleanup && actions & _UA_CLEANUP_PHASE + ? _URC_HANDLER_FOUND + : _URC_CONTINUE_UNWIND; return; } // Go to next action @@ -962,78 +896,51 @@ __gxx_personality_v0 bool native_exception = (exceptionClass & get_vendor_and_language) == (kOurExceptionClass & get_vendor_and_language); scan_results results; + // Process a catch handler for a native exception first. + if (actions == (_UA_CLEANUP_PHASE | _UA_HANDLER_FRAME) && + native_exception) { + // Reload the results from the phase 1 cache. + __cxa_exception* exception_header = + (__cxa_exception*)(unwind_exception + 1) - 1; + results.ttypeIndex = exception_header->handlerSwitchValue; + results.actionRecord = exception_header->actionRecord; + results.languageSpecificData = exception_header->languageSpecificData; + results.landingPad = + reinterpret_cast<uintptr_t>(exception_header->catchTemp); + results.adjustedPtr = exception_header->adjustedPtr; + + // Jump to the handler. + set_registers(unwind_exception, context, results); + return _URC_INSTALL_CONTEXT; + } + + // In other cases we need to scan LSDA. + scan_eh_tab(results, actions, native_exception, unwind_exception, context); + if (results.reason == _URC_CONTINUE_UNWIND || + results.reason == _URC_FATAL_PHASE1_ERROR) + return results.reason; + if (actions & _UA_SEARCH_PHASE) { // Phase 1 search: All we're looking for in phase 1 is a handler that // halts unwinding - scan_eh_tab(results, actions, native_exception, unwind_exception, context); - if (results.reason == _URC_HANDLER_FOUND) - { - // Found one. Can we cache the results somewhere to optimize phase 2? - if (native_exception) - { - __cxa_exception* exception_header = (__cxa_exception*)(unwind_exception+1) - 1; - exception_header->handlerSwitchValue = static_cast<int>(results.ttypeIndex); - exception_header->actionRecord = results.actionRecord; - exception_header->languageSpecificData = results.languageSpecificData; - exception_header->catchTemp = reinterpret_cast<void*>(results.landingPad); - exception_header->adjustedPtr = results.adjustedPtr; - } - return _URC_HANDLER_FOUND; - } - // Did not find a catching-handler. Return the results of the scan - // (normally _URC_CONTINUE_UNWIND, but could have been _URC_FATAL_PHASE1_ERROR - // if we were called improperly). - return results.reason; - } - if (actions & _UA_CLEANUP_PHASE) - { - // Phase 2 search: - // Did we find a catching handler in phase 1? - if (actions & _UA_HANDLER_FRAME) - { - // Yes, phase 1 said we have a catching handler here. - // Did we cache the results of the scan? - if (native_exception) - { - // Yes, reload the results from the cache. - __cxa_exception* exception_header = (__cxa_exception*)(unwind_exception+1) - 1; - results.ttypeIndex = exception_header->handlerSwitchValue; - results.actionRecord = exception_header->actionRecord; - results.languageSpecificData = exception_header->languageSpecificData; - results.landingPad = reinterpret_cast<uintptr_t>(exception_header->catchTemp); - results.adjustedPtr = exception_header->adjustedPtr; - } - else - { - // No, do the scan again to reload the results. - scan_eh_tab(results, actions, native_exception, unwind_exception, context); - // Phase 1 told us we would find a handler. Now in Phase 2 we - // didn't find a handler. The eh table should not be changing! - if (results.reason != _URC_HANDLER_FOUND) - call_terminate(native_exception, unwind_exception); - } - // Jump to the handler - set_registers(unwind_exception, context, results); - return _URC_INSTALL_CONTEXT; - } - // Either we didn't do a phase 1 search (due to forced unwinding), or - // phase 1 reported no catching-handlers. - // Search for a (non-catching) cleanup - scan_eh_tab(results, actions, native_exception, unwind_exception, context); - if (results.reason == _URC_HANDLER_FOUND) - { - // Found a non-catching handler. Jump to it: - set_registers(unwind_exception, context, results); - return _URC_INSTALL_CONTEXT; + assert(results.reason == _URC_HANDLER_FOUND); + if (native_exception) { + // For a native exception, cache the LSDA result. + __cxa_exception* exc = (__cxa_exception*)(unwind_exception + 1) - 1; + exc->handlerSwitchValue = static_cast<int>(results.ttypeIndex); + exc->actionRecord = results.actionRecord; + exc->languageSpecificData = results.languageSpecificData; + exc->catchTemp = reinterpret_cast<void*>(results.landingPad); + exc->adjustedPtr = results.adjustedPtr; } - // Did not find a cleanup. Return the results of the scan - // (normally _URC_CONTINUE_UNWIND, but could have been _URC_FATAL_PHASE2_ERROR - // if we were called improperly). - return results.reason; + return _URC_HANDLER_FOUND; } - // We were called improperly: neither a phase 1 or phase 2 search - return _URC_FATAL_PHASE1_ERROR; + + assert(actions & _UA_CLEANUP_PHASE); + assert(results.reason == _URC_HANDLER_FOUND); + set_registers(unwind_exception, context, results); + return _URC_INSTALL_CONTEXT; } #if defined(__SEH__) && !defined(__USING_SJLJ_EXCEPTIONS__) diff --git a/lib/libcxxabi/src/demangle/ItaniumDemangle.h b/lib/libcxxabi/src/demangle/ItaniumDemangle.h index 6ab8732183..e5fca98f92 100644 --- a/lib/libcxxabi/src/demangle/ItaniumDemangle.h +++ b/lib/libcxxabi/src/demangle/ItaniumDemangle.h @@ -82,6 +82,7 @@ X(PostfixExpr) \ X(ConditionalExpr) \ X(MemberExpr) \ + X(SubobjectExpr) \ X(EnclosingExpr) \ X(CastExpr) \ X(SizeofParamPackExpr) \ @@ -91,10 +92,10 @@ X(PrefixExpr) \ X(FunctionParam) \ X(ConversionExpr) \ + X(PointerToMemberConversionExpr) \ X(InitListExpr) \ X(FoldExpr) \ X(ThrowExpr) \ - X(UUIDOfExpr) \ X(BoolExpr) \ X(StringLiteral) \ X(LambdaExpr) \ @@ -1656,6 +1657,40 @@ public: } }; +class SubobjectExpr : public Node { + const Node *Type; + const Node *SubExpr; + StringView Offset; + NodeArray UnionSelectors; + bool OnePastTheEnd; + +public: + SubobjectExpr(const Node *Type_, const Node *SubExpr_, StringView Offset_, + NodeArray UnionSelectors_, bool OnePastTheEnd_) + : Node(KSubobjectExpr), Type(Type_), SubExpr(SubExpr_), Offset(Offset_), + UnionSelectors(UnionSelectors_), OnePastTheEnd(OnePastTheEnd_) {} + + template<typename Fn> void match(Fn F) const { + F(Type, SubExpr, Offset, UnionSelectors, OnePastTheEnd); + } + + void printLeft(OutputStream &S) const override { + SubExpr->print(S); + S += ".<"; + Type->print(S); + S += " at offset "; + if (Offset.empty()) { + S += "0"; + } else if (Offset[0] == 'n') { + S += "-"; + S += Offset.dropFront(); + } else { + S += Offset; + } + S += ">"; + } +}; + class EnclosingExpr : public Node { const StringView Prefix; const Node *Infix; @@ -1843,6 +1878,28 @@ public: } }; +class PointerToMemberConversionExpr : public Node { + const Node *Type; + const Node *SubExpr; + StringView Offset; + +public: + PointerToMemberConversionExpr(const Node *Type_, const Node *SubExpr_, + StringView Offset_) + : Node(KPointerToMemberConversionExpr), Type(Type_), SubExpr(SubExpr_), + Offset(Offset_) {} + + template<typename Fn> void match(Fn F) const { F(Type, SubExpr, Offset); } + + void printLeft(OutputStream &S) const override { + S += "("; + Type->print(S); + S += ")("; + SubExpr->print(S); + S += ")"; + } +}; + class InitListExpr : public Node { const Node *Ty; NodeArray Inits; @@ -1977,21 +2034,6 @@ public: } }; -// MSVC __uuidof extension, generated by clang in -fms-extensions mode. -class UUIDOfExpr : public Node { - Node *Operand; -public: - UUIDOfExpr(Node *Operand_) : Node(KUUIDOfExpr), Operand(Operand_) {} - - template<typename Fn> void match(Fn F) const { F(Operand); } - - void printLeft(OutputStream &S) const override { - S << "__uuidof("; - Operand->print(S); - S << ")"; - } -}; - class BoolExpr : public Node { bool Value; @@ -2313,9 +2355,9 @@ template <typename Derived, typename Alloc> struct AbstractManglingParser { TemplateParamList Params; public: - ScopedTemplateParamList(AbstractManglingParser *Parser) - : Parser(Parser), - OldNumTemplateParamLists(Parser->TemplateParams.size()) { + ScopedTemplateParamList(AbstractManglingParser *TheParser) + : Parser(TheParser), + OldNumTemplateParamLists(TheParser->TemplateParams.size()) { Parser->TemplateParams.push_back(&Params); } ~ScopedTemplateParamList() { @@ -2437,6 +2479,8 @@ template <typename Derived, typename Alloc> struct AbstractManglingParser { Node *parseConversionExpr(); Node *parseBracedExpr(); Node *parseFoldExpr(); + Node *parsePointerToMemberConversionExpr(); + Node *parseSubobjectExpr(); /// Parse the <type> production. Node *parseType(); @@ -4404,6 +4448,50 @@ Node *AbstractManglingParser<Derived, Alloc>::parseFoldExpr() { return make<FoldExpr>(IsLeftFold, OperatorName, Pack, Init); } +// <expression> ::= mc <parameter type> <expr> [<offset number>] E +// +// Not yet in the spec: https://github.com/itanium-cxx-abi/cxx-abi/issues/47 +template <typename Derived, typename Alloc> +Node *AbstractManglingParser<Derived, Alloc>::parsePointerToMemberConversionExpr() { + Node *Ty = getDerived().parseType(); + if (!Ty) + return nullptr; + Node *Expr = getDerived().parseExpr(); + if (!Expr) + return nullptr; + StringView Offset = getDerived().parseNumber(true); + if (!consumeIf('E')) + return nullptr; + return make<PointerToMemberConversionExpr>(Ty, Expr, Offset); +} + +// <expression> ::= so <referent type> <expr> [<offset number>] <union-selector>* [p] E +// <union-selector> ::= _ [<number>] +// +// Not yet in the spec: https://github.com/itanium-cxx-abi/cxx-abi/issues/47 +template <typename Derived, typename Alloc> +Node *AbstractManglingParser<Derived, Alloc>::parseSubobjectExpr() { + Node *Ty = getDerived().parseType(); + if (!Ty) + return nullptr; + Node *Expr = getDerived().parseExpr(); + if (!Expr) + return nullptr; + StringView Offset = getDerived().parseNumber(true); + size_t SelectorsBegin = Names.size(); + while (consumeIf('_')) { + Node *Selector = make<NameType>(parseNumber()); + if (!Selector) + return nullptr; + Names.push_back(Selector); + } + bool OnePastTheEnd = consumeIf('p'); + if (!consumeIf('E')) + return nullptr; + return make<SubobjectExpr>( + Ty, Expr, Offset, popTrailingNodeArray(SelectorsBegin), OnePastTheEnd); +} + // <expression> ::= <unary operator-name> <expression> // ::= <binary operator-name> <expression> <expression> // ::= <ternary operator-name> <expression> <expression> <expression> @@ -4661,6 +4749,9 @@ Node *AbstractManglingParser<Derived, Alloc>::parseExpr() { return nullptr; case 'm': switch (First[1]) { + case 'c': + First += 2; + return parsePointerToMemberConversionExpr(); case 'i': First += 2; return getDerived().parseBinaryExpr("-"); @@ -4808,6 +4899,9 @@ Node *AbstractManglingParser<Derived, Alloc>::parseExpr() { return Ex; return make<CastExpr>("static_cast", T, Ex); } + case 'o': + First += 2; + return parseSubobjectExpr(); case 'p': { First += 2; Node *Child = getDerived().parseExpr(); @@ -4903,6 +4997,43 @@ Node *AbstractManglingParser<Derived, Alloc>::parseExpr() { } } return nullptr; + case 'u': { + ++First; + Node *Name = getDerived().parseSourceName(/*NameState=*/nullptr); + if (!Name) + return nullptr; + // Special case legacy __uuidof mangling. The 't' and 'z' appear where the + // standard encoding expects a <template-arg>, and would be otherwise be + // interpreted as <type> node 'short' or 'ellipsis'. However, neither + // __uuidof(short) nor __uuidof(...) can actually appear, so there is no + // actual conflict here. + if (Name->getBaseName() == "__uuidof") { + if (numLeft() < 2) + return nullptr; + if (*First == 't') { + ++First; + Node *Ty = getDerived().parseType(); + if (!Ty) + return nullptr; + return make<CallExpr>(Name, makeNodeArray(&Ty, &Ty + 1)); + } + if (*First == 'z') { + ++First; + Node *Ex = getDerived().parseExpr(); + if (!Ex) + return nullptr; + return make<CallExpr>(Name, makeNodeArray(&Ex, &Ex + 1)); + } + } + size_t ExprsBegin = Names.size(); + while (!consumeIf('E')) { + Node *E = getDerived().parseTemplateArg(); + if (E == nullptr) + return E; + Names.push_back(E); + } + return make<CallExpr>(Name, popTrailingNodeArray(ExprsBegin)); + } case '1': case '2': case '3': @@ -4914,21 +5045,6 @@ Node *AbstractManglingParser<Derived, Alloc>::parseExpr() { case '9': return getDerived().parseUnresolvedName(); } - - if (consumeIf("u8__uuidoft")) { - Node *Ty = getDerived().parseType(); - if (!Ty) - return nullptr; - return make<UUIDOfExpr>(Ty); - } - - if (consumeIf("u8__uuidofz")) { - Node *Ex = getDerived().parseExpr(); - if (!Ex) - return nullptr; - return make<UUIDOfExpr>(Ex); - } - return nullptr; } @@ -4975,6 +5091,16 @@ Node *AbstractManglingParser<Derived, Alloc>::parseSpecialName() { switch (look()) { case 'T': switch (look(1)) { + // TA <template-arg> # template parameter object + // + // Not yet in the spec: https://github.com/itanium-cxx-abi/cxx-abi/issues/63 + case 'A': { + First += 2; + Node *Arg = getDerived().parseTemplateArg(); + if (Arg == nullptr) + return nullptr; + return make<SpecialName>("template parameter object for ", Arg); + } // TV <type> # virtual table case 'V': { First += 2; @@ -5103,7 +5229,7 @@ Node *AbstractManglingParser<Derived, Alloc>::parseEncoding() { decltype(TemplateParams) OldParams; public: - SaveTemplateParams(AbstractManglingParser *Parser) : Parser(Parser) { + SaveTemplateParams(AbstractManglingParser *TheParser) : Parser(TheParser) { OldParams = std::move(Parser->TemplateParams); Parser->TemplateParams.clear(); } @@ -5203,7 +5329,12 @@ struct FloatData<long double> #else static const size_t mangled_size = 20; // May need to be adjusted to 16 or 24 on other platforms #endif - static const size_t max_demangled_size = 40; + // `-0x1.ffffffffffffffffffffffffffffp+16383` + 'L' + '\0' == 42 bytes. + // 28 'f's * 4 bits == 112 bits, which is the number of mantissa bits. + // Negatives are one character longer than positives. + // `0x1.` and `p` are constant, and exponents `+16383` and `-16382` are the + // same length. 1 sign bit, 112 mantissa bits, and 15 exponent bits == 128. + static const size_t max_demangled_size = 42; static constexpr const char *spec = "%LaL"; }; diff --git a/lib/libcxxabi/src/demangle/Utility.h b/lib/libcxxabi/src/demangle/Utility.h index 04e1936ebb..846a5f0818 100644 --- a/lib/libcxxabi/src/demangle/Utility.h +++ b/lib/libcxxabi/src/demangle/Utility.h @@ -52,7 +52,7 @@ class OutputStream { char *TempPtr = std::end(Temp); while (N) { - *--TempPtr = '0' + char(N % 10); + *--TempPtr = char('0' + N % 10); N /= 10; } diff --git a/lib/libcxxabi/src/fallback_malloc.cpp b/lib/libcxxabi/src/fallback_malloc.cpp index fdae40764a..f3d7937793 100644 --- a/lib/libcxxabi/src/fallback_malloc.cpp +++ b/lib/libcxxabi/src/fallback_malloc.cpp @@ -6,9 +6,6 @@ // //===----------------------------------------------------------------------===// -// Define _LIBCPP_BUILDING_LIBRARY to ensure _LIBCPP_HAS_NO_LIBRARY_ALIGNED_ALLOCATION -// is only defined when libc aligned allocation is not available. -#define _LIBCPP_BUILDING_LIBRARY #include "fallback_malloc.h" #include <__threading_support> @@ -20,6 +17,7 @@ #include <stdlib.h> // for malloc, calloc, free #include <string.h> // for memset +#include <new> // for std::__libcpp_aligned_{alloc,free} // A small, simple heap manager based (loosely) on // the startup heap manager from FreeBSD, optimized for space. @@ -144,29 +142,26 @@ void fallback_free(void* ptr) { mutexor mtx(&heap_mutex); #ifdef DEBUG_FALLBACK_MALLOC - std::cout << "Freeing item at " << offset_from_node(cp) << " of size " - << cp->len << std::endl; + std::printf("Freeing item at %d of size %d\n", offset_from_node(cp), cp->len); #endif for (p = freelist, prev = 0; p && p != list_end; prev = p, p = node_from_offset(p->next_node)) { #ifdef DEBUG_FALLBACK_MALLOC - std::cout << " p, cp, after (p), after(cp) " << offset_from_node(p) << ' ' - << offset_from_node(cp) << ' ' << offset_from_node(after(p)) - << ' ' << offset_from_node(after(cp)) << std::endl; + std::printf(" p=%d, cp=%d, after(p)=%d, after(cp)=%d\n", + offset_from_node(p), offset_from_node(cp), + offset_from_node(after(p)), offset_from_node(after(cp))); #endif if (after(p) == cp) { #ifdef DEBUG_FALLBACK_MALLOC - std::cout << " Appending onto chunk at " << offset_from_node(p) - << std::endl; + std::printf(" Appending onto chunk at %d\n", offset_from_node(p)); #endif p->len = static_cast<heap_size>( p->len + cp->len); // make the free heap_node larger return; } else if (after(cp) == p) { // there's a free heap_node right after #ifdef DEBUG_FALLBACK_MALLOC - std::cout << " Appending free chunk at " << offset_from_node(p) - << std::endl; + std::printf(" Appending free chunk at %d\n", offset_from_node(p)); #endif cp->len = static_cast<heap_size>(cp->len + p->len); if (prev == 0) { @@ -179,8 +174,7 @@ void fallback_free(void* ptr) { } // Nothing to merge with, add it to the start of the free list #ifdef DEBUG_FALLBACK_MALLOC - std::cout << " Making new free list entry " << offset_from_node(cp) - << std::endl; + std::printf(" Making new free list entry %d\n", offset_from_node(cp)); #endif cp->next_node = offset_from_node(freelist); freelist = cp; @@ -195,11 +189,11 @@ size_t print_free_list() { for (p = freelist, prev = 0; p && p != list_end; prev = p, p = node_from_offset(p->next_node)) { - std::cout << (prev == 0 ? "" : " ") << "Offset: " << offset_from_node(p) - << "\tsize: " << p->len << " Next: " << p->next_node << std::endl; + std::printf("%sOffset: %d\tsize: %d Next: %d\n", + (prev == 0 ? "" : " "), offset_from_node(p), p->len, p->next_node); total_free += p->len; } - std::cout << "Total Free space: " << total_free << std::endl; + std::printf("Total Free space: %d\n", total_free); return total_free; } #endif @@ -211,7 +205,7 @@ struct __attribute__((aligned)) __aligned_type {}; void* __aligned_malloc_with_fallback(size_t size) { #if defined(_WIN32) - if (void* dest = _aligned_malloc(size, alignof(__aligned_type))) + if (void* dest = std::__libcpp_aligned_alloc(alignof(__aligned_type), size)) return dest; #elif defined(_LIBCPP_HAS_NO_LIBRARY_ALIGNED_ALLOCATION) if (void* dest = ::malloc(size)) @@ -219,8 +213,7 @@ void* __aligned_malloc_with_fallback(size_t size) { #else if (size == 0) size = 1; - void* dest; - if (::posix_memalign(&dest, __alignof(__aligned_type), size) == 0) + if (void* dest = std::__libcpp_aligned_alloc(__alignof(__aligned_type), size)) return dest; #endif return fallback_malloc(size); @@ -241,10 +234,10 @@ void __aligned_free_with_fallback(void* ptr) { if (is_fallback_ptr(ptr)) fallback_free(ptr); else { -#if defined(_WIN32) - ::_aligned_free(ptr); -#else +#if defined(_LIBCPP_HAS_NO_LIBRARY_ALIGNED_ALLOCATION) ::free(ptr); +#else + std::__libcpp_aligned_free(ptr); #endif } } diff --git a/lib/libcxxabi/src/include/refstring.h b/lib/libcxxabi/src/include/refstring.h deleted file mode 100644 index 5905b5c932..0000000000 --- a/lib/libcxxabi/src/include/refstring.h +++ /dev/null @@ -1,131 +0,0 @@ -//===------------------------ __refstring ---------------------------------===// -// -// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. -// See https://llvm.org/LICENSE.txt for license information. -// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception -// -//===----------------------------------------------------------------------===// - -// FIXME: This file is copied from libcxx/src/include/refstring.h. Instead of -// duplicating the file in libc++abi we should require that the libc++ sources -// are available when building libc++abi. - -#ifndef _LIBCPPABI_REFSTRING_H -#define _LIBCPPABI_REFSTRING_H - -#include <__config> -#include <stdexcept> -#include <cstddef> -#include <cstring> -#ifdef __APPLE__ -#include <dlfcn.h> -#include <mach-o/dyld.h> -#endif -#include "atomic_support.h" - -_LIBCPP_BEGIN_NAMESPACE_STD - -namespace __refstring_imp { namespace { -typedef int count_t; - -struct _Rep_base { - std::size_t len; - std::size_t cap; - count_t count; -}; - -inline _Rep_base* rep_from_data(const char *data_) noexcept { - char *data = const_cast<char *>(data_); - return reinterpret_cast<_Rep_base *>(data - sizeof(_Rep_base)); -} - -inline char * data_from_rep(_Rep_base *rep) noexcept { - char *data = reinterpret_cast<char *>(rep); - return data + sizeof(*rep); -} - -#if defined(__APPLE__) -inline -const char* compute_gcc_empty_string_storage() _NOEXCEPT -{ - void* handle = dlopen("/usr/lib/libstdc++.6.dylib", RTLD_NOLOAD); - if (handle == nullptr) - return nullptr; - void* sym = dlsym(handle, "_ZNSs4_Rep20_S_empty_rep_storageE"); - if (sym == nullptr) - return nullptr; - return data_from_rep(reinterpret_cast<_Rep_base *>(sym)); -} - -inline -const char* -get_gcc_empty_string_storage() _NOEXCEPT -{ - static const char* p = compute_gcc_empty_string_storage(); - return p; -} -#endif - -}} // namespace __refstring_imp - -using namespace __refstring_imp; - -inline -__libcpp_refstring::__libcpp_refstring(const char* msg) { - std::size_t len = strlen(msg); - _Rep_base* rep = static_cast<_Rep_base *>(::operator new(sizeof(*rep) + len + 1)); - rep->len = len; - rep->cap = len; - rep->count = 0; - char *data = data_from_rep(rep); - std::memcpy(data, msg, len + 1); - __imp_ = data; -} - -inline -__libcpp_refstring::__libcpp_refstring(const __libcpp_refstring &s) _NOEXCEPT - : __imp_(s.__imp_) -{ - if (__uses_refcount()) - __libcpp_atomic_add(&rep_from_data(__imp_)->count, 1); -} - -inline -__libcpp_refstring& __libcpp_refstring::operator=(__libcpp_refstring const& s) _NOEXCEPT { - bool adjust_old_count = __uses_refcount(); - struct _Rep_base *old_rep = rep_from_data(__imp_); - __imp_ = s.__imp_; - if (__uses_refcount()) - __libcpp_atomic_add(&rep_from_data(__imp_)->count, 1); - if (adjust_old_count) - { - if (__libcpp_atomic_add(&old_rep->count, count_t(-1)) < 0) - { - ::operator delete(old_rep); - } - } - return *this; -} - -inline -__libcpp_refstring::~__libcpp_refstring() { - if (__uses_refcount()) { - _Rep_base* rep = rep_from_data(__imp_); - if (__libcpp_atomic_add(&rep->count, count_t(-1)) < 0) { - ::operator delete(rep); - } - } -} - -inline -bool __libcpp_refstring::__uses_refcount() const { -#ifdef __APPLE__ - return __imp_ != get_gcc_empty_string_storage(); -#else - return true; -#endif -} - -_LIBCPP_END_NAMESPACE_STD - -#endif //_LIBCPPABI_REFSTRING_H diff --git a/lib/libcxxabi/src/private_typeinfo.cpp b/lib/libcxxabi/src/private_typeinfo.cpp index 55a90b3ae1..c77ad669c4 100644 --- a/lib/libcxxabi/src/private_typeinfo.cpp +++ b/lib/libcxxabi/src/private_typeinfo.cpp @@ -61,6 +61,16 @@ is_equal(const std::type_info* x, const std::type_info* y, bool use_strcmp) return x == y || strcmp(x->name(), y->name()) == 0; } +static inline ptrdiff_t update_offset_to_base(const char* vtable, + ptrdiff_t offset_to_base) { +#if __has_feature(cxx_abi_relative_vtable) + // VTable components are 32 bits in the relative vtables ABI. + return *reinterpret_cast<const int32_t*>(vtable + offset_to_base); +#else + return *reinterpret_cast<const ptrdiff_t*>(vtable + offset_to_base); +#endif +} + namespace __cxxabiv1 { @@ -297,7 +307,7 @@ __base_class_type_info::has_unambiguous_public_base(__dynamic_cast_info* info, if (__offset_flags & __virtual_mask) { const char* vtable = *static_cast<const char*const*>(adjustedPtr); - offset_to_base = *reinterpret_cast<const ptrdiff_t*>(vtable + offset_to_base); + offset_to_base = update_offset_to_base(vtable, offset_to_base); } } __base_type->has_unambiguous_public_base( @@ -615,10 +625,26 @@ __dynamic_cast(const void *static_ptr, const __class_type_info *static_type, // Possible future optimization: Take advantage of src2dst_offset // Get (dynamic_ptr, dynamic_type) from static_ptr +#if __has_feature(cxx_abi_relative_vtable) + // The vtable address will point to the first virtual function, which is 8 + // bytes after the start of the vtable (4 for the offset from top + 4 for the typeinfo component). + const int32_t* vtable = + *reinterpret_cast<const int32_t* const*>(static_ptr); + int32_t offset_to_derived = vtable[-2]; + const void* dynamic_ptr = static_cast<const char*>(static_ptr) + offset_to_derived; + + // The typeinfo component is now a relative offset to a proxy. + int32_t offset_to_ti_proxy = vtable[-1]; + const uint8_t* ptr_to_ti_proxy = + reinterpret_cast<const uint8_t*>(vtable) + offset_to_ti_proxy; + const __class_type_info* dynamic_type = + *(reinterpret_cast<const __class_type_info* const*>(ptr_to_ti_proxy)); +#else void **vtable = *static_cast<void ** const *>(static_ptr); ptrdiff_t offset_to_derived = reinterpret_cast<ptrdiff_t>(vtable[-2]); const void* dynamic_ptr = static_cast<const char*>(static_ptr) + offset_to_derived; const __class_type_info* dynamic_type = static_cast<const __class_type_info*>(vtable[-1]); +#endif // Initialize answer to nullptr. This will be changed from the search // results if a non-null answer is found. Regardless, this is what will @@ -641,6 +667,7 @@ __dynamic_cast(const void *static_ptr, const __class_type_info *static_type, { // We get here only if there is some kind of visibility problem // in client code. + static_assert(std::atomic<size_t>::is_always_lock_free, ""); static std::atomic<size_t> error_count(0); size_t error_count_snapshot = error_count.fetch_add(1, std::memory_order_relaxed); if ((error_count_snapshot & (error_count_snapshot-1)) == 0) @@ -667,6 +694,7 @@ __dynamic_cast(const void *static_ptr, const __class_type_info *static_type, if (info.path_dst_ptr_to_static_ptr == unknown && info.path_dynamic_ptr_to_static_ptr == unknown) { + static_assert(std::atomic<size_t>::is_always_lock_free, ""); static std::atomic<size_t> error_count(0); size_t error_count_snapshot = error_count.fetch_add(1, std::memory_order_relaxed); if ((error_count_snapshot & (error_count_snapshot-1)) == 0) @@ -1265,7 +1293,7 @@ __base_class_type_info::search_above_dst(__dynamic_cast_info* info, if (__offset_flags & __virtual_mask) { const char* vtable = *static_cast<const char*const*>(current_ptr); - offset_to_base = *reinterpret_cast<const ptrdiff_t*>(vtable + offset_to_base); + offset_to_base = update_offset_to_base(vtable, offset_to_base); } __base_type->search_above_dst(info, dst_ptr, static_cast<const char*>(current_ptr) + offset_to_base, @@ -1285,7 +1313,7 @@ __base_class_type_info::search_below_dst(__dynamic_cast_info* info, if (__offset_flags & __virtual_mask) { const char* vtable = *static_cast<const char*const*>(current_ptr); - offset_to_base = *reinterpret_cast<const ptrdiff_t*>(vtable + offset_to_base); + offset_to_base = update_offset_to_base(vtable, offset_to_base); } __base_type->search_below_dst(info, static_cast<const char*>(current_ptr) + offset_to_base, diff --git a/lib/libcxxabi/src/stdlib_exception.cpp b/lib/libcxxabi/src/stdlib_exception.cpp index 88d911d96c..b0cc431f24 100644 --- a/lib/libcxxabi/src/stdlib_exception.cpp +++ b/lib/libcxxabi/src/stdlib_exception.cpp @@ -6,7 +6,6 @@ // //===----------------------------------------------------------------------===// -#define _LIBCPP_BUILDING_LIBRARY #include <new> #include <exception> diff --git a/lib/libcxxabi/src/stdlib_new_delete.cpp b/lib/libcxxabi/src/stdlib_new_delete.cpp index 698c5f7c29..8ef3057dd4 100644 --- a/lib/libcxxabi/src/stdlib_new_delete.cpp +++ b/lib/libcxxabi/src/stdlib_new_delete.cpp @@ -8,7 +8,6 @@ // This file implements the new and delete operators. //===----------------------------------------------------------------------===// -#define _LIBCPP_BUILDING_LIBRARY #include "__cxxabi_config.h" #include <new> #include <cstdlib> @@ -28,7 +27,7 @@ operator new(std::size_t size) _THROW_BAD_ALLOC if (size == 0) size = 1; void* p; - while ((p = ::malloc(size)) == 0) + while ((p = ::malloc(size)) == nullptr) { // If malloc fails and there is a new_handler, // call it to try free up memory. @@ -49,7 +48,7 @@ _LIBCXXABI_WEAK void* operator new(size_t size, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCXXABI_NO_EXCEPTIONS try { @@ -75,7 +74,7 @@ _LIBCXXABI_WEAK void* operator new[](size_t size, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCXXABI_NO_EXCEPTIONS try { @@ -94,8 +93,7 @@ _LIBCXXABI_WEAK void operator delete(void* ptr) _NOEXCEPT { - if (ptr) - ::free(ptr); + ::free(ptr); } _LIBCXXABI_WEAK @@ -143,15 +141,16 @@ operator new(std::size_t size, std::align_val_t alignment) _THROW_BAD_ALLOC size = 1; if (static_cast<size_t>(alignment) < sizeof(void*)) alignment = std::align_val_t(sizeof(void*)); + + // Try allocating memory. If allocation fails and there is a new_handler, + // call it to try free up memory, and try again until it succeeds, or until + // the new_handler decides to terminate. + // + // If allocation fails and there is no new_handler, we throw bad_alloc + // (or return nullptr if exceptions are disabled). void* p; -#if defined(_LIBCPP_WIN32API) - while ((p = _aligned_malloc(size, static_cast<size_t>(alignment))) == nullptr) -#else - while (::posix_memalign(&p, static_cast<size_t>(alignment), size) != 0) -#endif + while ((p = std::__libcpp_aligned_alloc(static_cast<std::size_t>(alignment), size)) == nullptr) { - // If posix_memalign fails and there is a new_handler, - // call it to try free up memory. std::new_handler nh = std::get_new_handler(); if (nh) nh(); @@ -159,7 +158,6 @@ operator new(std::size_t size, std::align_val_t alignment) _THROW_BAD_ALLOC #ifndef _LIBCXXABI_NO_EXCEPTIONS throw std::bad_alloc(); #else - p = nullptr; // posix_memalign doesn't initialize 'p' on failure break; #endif } @@ -171,7 +169,7 @@ _LIBCXXABI_WEAK void* operator new(size_t size, std::align_val_t alignment, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCXXABI_NO_EXCEPTIONS try { @@ -197,7 +195,7 @@ _LIBCXXABI_WEAK void* operator new[](size_t size, std::align_val_t alignment, const std::nothrow_t&) _NOEXCEPT { - void* p = 0; + void* p = nullptr; #ifndef _LIBCXXABI_NO_EXCEPTIONS try { @@ -216,12 +214,7 @@ _LIBCXXABI_WEAK void operator delete(void* ptr, std::align_val_t) _NOEXCEPT { - if (ptr) -#if defined(_LIBCPP_WIN32API) - ::_aligned_free(ptr); -#else - ::free(ptr); -#endif + std::__libcpp_aligned_free(ptr); } _LIBCXXABI_WEAK diff --git a/lib/libcxxabi/src/stdlib_stdexcept.cpp b/lib/libcxxabi/src/stdlib_stdexcept.cpp index 2cff8ee2f0..4a464e4889 100644 --- a/lib/libcxxabi/src/stdlib_stdexcept.cpp +++ b/lib/libcxxabi/src/stdlib_stdexcept.cpp @@ -6,7 +6,7 @@ // //===----------------------------------------------------------------------===// -#include "include/refstring.h" +#include "../../libcxx/src/include/refstring.h" #include "stdexcept" #include "new" #include <cstdlib> diff --git a/lib/libunwind/include/__libunwind_config.h b/lib/libunwind/include/__libunwind_config.h index 71d77ca651..34ac6f717d 100644 --- a/lib/libunwind/include/__libunwind_config.h +++ b/lib/libunwind/include/__libunwind_config.h @@ -25,8 +25,12 @@ #define _LIBUNWIND_HIGHEST_DWARF_REGISTER_SPARC 31 #define _LIBUNWIND_HIGHEST_DWARF_REGISTER_HEXAGON 34 #define _LIBUNWIND_HIGHEST_DWARF_REGISTER_RISCV 64 +#define _LIBUNWIND_HIGHEST_DWARF_REGISTER_VE 143 #if defined(_LIBUNWIND_IS_NATIVE_ONLY) +# if defined(__linux__) +# define _LIBUNWIND_TARGET_LINUX 1 +# endif # if defined(__i386__) # define _LIBUNWIND_TARGET_I386 # define _LIBUNWIND_CONTEXT_SIZE 8 @@ -135,6 +139,11 @@ # error "Unsupported RISC-V ABI" # endif # define _LIBUNWIND_HIGHEST_DWARF_REGISTER _LIBUNWIND_HIGHEST_DWARF_REGISTER_RISCV +# elif defined(__ve__) +# define _LIBUNWIND_TARGET_VE 1 +# define _LIBUNWIND_CONTEXT_SIZE 67 +# define _LIBUNWIND_CURSOR_SIZE 79 +# define _LIBUNWIND_HIGHEST_DWARF_REGISTER _LIBUNWIND_HIGHEST_DWARF_REGISTER_VE # else # error "Unsupported architecture." # endif @@ -151,6 +160,7 @@ # define _LIBUNWIND_TARGET_SPARC 1 # define _LIBUNWIND_TARGET_HEXAGON 1 # define _LIBUNWIND_TARGET_RISCV 1 +# define _LIBUNWIND_TARGET_VE 1 # define _LIBUNWIND_CONTEXT_SIZE 167 # define _LIBUNWIND_CURSOR_SIZE 179 # define _LIBUNWIND_HIGHEST_DWARF_REGISTER 287 diff --git a/lib/libunwind/include/libunwind.h b/lib/libunwind/include/libunwind.h index 23ef47f4ac..5bae8d02f7 100644 --- a/lib/libunwind/include/libunwind.h +++ b/lib/libunwind/include/libunwind.h @@ -43,6 +43,12 @@ #define LIBUNWIND_AVAIL #endif +#if defined(_WIN32) && defined(__SEH__) + #define LIBUNWIND_CURSOR_ALIGNMENT_ATTR __attribute__((__aligned__(16))) +#else + #define LIBUNWIND_CURSOR_ALIGNMENT_ATTR +#endif + /* error codes */ enum { UNW_ESUCCESS = 0, /* no error */ @@ -68,7 +74,7 @@ typedef struct unw_context_t unw_context_t; struct unw_cursor_t { uint64_t data[_LIBUNWIND_CURSOR_SIZE]; -}; +} LIBUNWIND_CURSOR_ALIGNMENT_ATTR; typedef struct unw_cursor_t unw_cursor_t; typedef struct unw_addr_space *unw_addr_space_t; @@ -941,4 +947,156 @@ enum { UNW_RISCV_F31 = 63, }; +// VE register numbers +enum { + UNW_VE_S0 = 0, + UNW_VE_S1 = 1, + UNW_VE_S2 = 2, + UNW_VE_S3 = 3, + UNW_VE_S4 = 4, + UNW_VE_S5 = 5, + UNW_VE_S6 = 6, + UNW_VE_S7 = 7, + UNW_VE_S8 = 8, + UNW_VE_S9 = 9, + UNW_VE_S10 = 10, + UNW_VE_S11 = 11, + UNW_VE_S12 = 12, + UNW_VE_S13 = 13, + UNW_VE_S14 = 14, + UNW_VE_S15 = 15, + UNW_VE_S16 = 16, + UNW_VE_S17 = 17, + UNW_VE_S18 = 18, + UNW_VE_S19 = 19, + UNW_VE_S20 = 20, + UNW_VE_S21 = 21, + UNW_VE_S22 = 22, + UNW_VE_S23 = 23, + UNW_VE_S24 = 24, + UNW_VE_S25 = 25, + UNW_VE_S26 = 26, + UNW_VE_S27 = 27, + UNW_VE_S28 = 28, + UNW_VE_S29 = 29, + UNW_VE_S30 = 30, + UNW_VE_S31 = 31, + UNW_VE_S32 = 32, + UNW_VE_S33 = 33, + UNW_VE_S34 = 34, + UNW_VE_S35 = 35, + UNW_VE_S36 = 36, + UNW_VE_S37 = 37, + UNW_VE_S38 = 38, + UNW_VE_S39 = 39, + UNW_VE_S40 = 40, + UNW_VE_S41 = 41, + UNW_VE_S42 = 42, + UNW_VE_S43 = 43, + UNW_VE_S44 = 44, + UNW_VE_S45 = 45, + UNW_VE_S46 = 46, + UNW_VE_S47 = 47, + UNW_VE_S48 = 48, + UNW_VE_S49 = 49, + UNW_VE_S50 = 50, + UNW_VE_S51 = 51, + UNW_VE_S52 = 52, + UNW_VE_S53 = 53, + UNW_VE_S54 = 54, + UNW_VE_S55 = 55, + UNW_VE_S56 = 56, + UNW_VE_S57 = 57, + UNW_VE_S58 = 58, + UNW_VE_S59 = 59, + UNW_VE_S60 = 60, + UNW_VE_S61 = 61, + UNW_VE_S62 = 62, + UNW_VE_S63 = 63, + UNW_VE_V0 = 64 + 0, + UNW_VE_V1 = 64 + 1, + UNW_VE_V2 = 64 + 2, + UNW_VE_V3 = 64 + 3, + UNW_VE_V4 = 64 + 4, + UNW_VE_V5 = 64 + 5, + UNW_VE_V6 = 64 + 6, + UNW_VE_V7 = 64 + 7, + UNW_VE_V8 = 64 + 8, + UNW_VE_V9 = 64 + 9, + UNW_VE_V10 = 64 + 10, + UNW_VE_V11 = 64 + 11, + UNW_VE_V12 = 64 + 12, + UNW_VE_V13 = 64 + 13, + UNW_VE_V14 = 64 + 14, + UNW_VE_V15 = 64 + 15, + UNW_VE_V16 = 64 + 16, + UNW_VE_V17 = 64 + 17, + UNW_VE_V18 = 64 + 18, + UNW_VE_V19 = 64 + 19, + UNW_VE_V20 = 64 + 20, + UNW_VE_V21 = 64 + 21, + UNW_VE_V22 = 64 + 22, + UNW_VE_V23 = 64 + 23, + UNW_VE_V24 = 64 + 24, + UNW_VE_V25 = 64 + 25, + UNW_VE_V26 = 64 + 26, + UNW_VE_V27 = 64 + 27, + UNW_VE_V28 = 64 + 28, + UNW_VE_V29 = 64 + 29, + UNW_VE_V30 = 64 + 30, + UNW_VE_V31 = 64 + 31, + UNW_VE_V32 = 64 + 32, + UNW_VE_V33 = 64 + 33, + UNW_VE_V34 = 64 + 34, + UNW_VE_V35 = 64 + 35, + UNW_VE_V36 = 64 + 36, + UNW_VE_V37 = 64 + 37, + UNW_VE_V38 = 64 + 38, + UNW_VE_V39 = 64 + 39, + UNW_VE_V40 = 64 + 40, + UNW_VE_V41 = 64 + 41, + UNW_VE_V42 = 64 + 42, + UNW_VE_V43 = 64 + 43, + UNW_VE_V44 = 64 + 44, + UNW_VE_V45 = 64 + 45, + UNW_VE_V46 = 64 + 46, + UNW_VE_V47 = 64 + 47, + UNW_VE_V48 = 64 + 48, + UNW_VE_V49 = 64 + 49, + UNW_VE_V50 = 64 + 50, + UNW_VE_V51 = 64 + 51, + UNW_VE_V52 = 64 + 52, + UNW_VE_V53 = 64 + 53, + UNW_VE_V54 = 64 + 54, + UNW_VE_V55 = 64 + 55, + UNW_VE_V56 = 64 + 56, + UNW_VE_V57 = 64 + 57, + UNW_VE_V58 = 64 + 58, + UNW_VE_V59 = 64 + 59, + UNW_VE_V60 = 64 + 60, + UNW_VE_V61 = 64 + 61, + UNW_VE_V62 = 64 + 62, + UNW_VE_V63 = 64 + 63, + UNW_VE_VM0 = 128 + 0, + UNW_VE_VM1 = 128 + 1, + UNW_VE_VM2 = 128 + 2, + UNW_VE_VM3 = 128 + 3, + UNW_VE_VM4 = 128 + 4, + UNW_VE_VM5 = 128 + 5, + UNW_VE_VM6 = 128 + 6, + UNW_VE_VM7 = 128 + 7, + UNW_VE_VM8 = 128 + 8, + UNW_VE_VM9 = 128 + 9, + UNW_VE_VM10 = 128 + 10, + UNW_VE_VM11 = 128 + 11, + UNW_VE_VM12 = 128 + 12, + UNW_VE_VM13 = 128 + 13, + UNW_VE_VM14 = 128 + 14, + UNW_VE_VM15 = 128 + 15, // = 143 + + // Following registers don't have DWARF register numbers. + UNW_VE_VIXR = 144, + UNW_VE_VL = 145, +}; + #endif diff --git a/lib/libunwind/src/AddressSpace.hpp b/lib/libunwind/src/AddressSpace.hpp index 93395ffb3b..171318ff63 100644 --- a/lib/libunwind/src/AddressSpace.hpp +++ b/lib/libunwind/src/AddressSpace.hpp @@ -17,6 +17,12 @@ #include <stdlib.h> #include <string.h> +#include "libunwind.h" +#include "config.h" +#include "dwarf2.h" +#include "EHHeaderParser.hpp" +#include "Registers.hpp" + #ifndef _LIBUNWIND_USE_DLADDR #if !defined(_LIBUNWIND_IS_BAREMETAL) && !defined(_WIN32) #define _LIBUNWIND_USE_DLADDR 1 @@ -40,19 +46,6 @@ struct EHABIIndexEntry { #endif #ifdef __APPLE__ -#include <mach-o/getsect.h> -namespace libunwind { - bool checkKeyMgrRegisteredFDEs(uintptr_t targetAddr, void *&fde); -} -#endif - -#include "libunwind.h" -#include "config.h" -#include "dwarf2.h" -#include "EHHeaderParser.hpp" -#include "Registers.hpp" - -#ifdef __APPLE__ struct dyld_unwind_sections { @@ -62,43 +55,9 @@ namespace libunwind { const void* compact_unwind_section; uintptr_t compact_unwind_section_length; }; - #if (defined(__MAC_OS_X_VERSION_MIN_REQUIRED) \ - && (__MAC_OS_X_VERSION_MIN_REQUIRED >= 1070)) \ - || defined(__IPHONE_OS_VERSION_MIN_REQUIRED) - // In 10.7.0 or later, libSystem.dylib implements this function. - extern "C" bool _dyld_find_unwind_sections(void *, dyld_unwind_sections *); - #else - // In 10.6.x and earlier, we need to implement this functionality. Note - // that this requires a newer version of libmacho (from cctools) than is - // present in libSystem on 10.6.x (for getsectiondata). - static inline bool _dyld_find_unwind_sections(void* addr, - dyld_unwind_sections* info) { - // Find mach-o image containing address. - Dl_info dlinfo; - if (!dladdr(addr, &dlinfo)) - return false; -#if __LP64__ - const struct mach_header_64 *mh = (const struct mach_header_64 *)dlinfo.dli_fbase; -#else - const struct mach_header *mh = (const struct mach_header *)dlinfo.dli_fbase; -#endif - // Initialize the return struct - info->mh = (const struct mach_header *)mh; - info->dwarf_section = getsectiondata(mh, "__TEXT", "__eh_frame", &info->dwarf_section_length); - info->compact_unwind_section = getsectiondata(mh, "__TEXT", "__unwind_info", &info->compact_unwind_section_length); - - if (!info->dwarf_section) { - info->dwarf_section_length = 0; - } - - if (!info->compact_unwind_section) { - info->compact_unwind_section_length = 0; - } - - return true; - } - #endif + // In 10.7.0 or later, libSystem.dylib implements this function. + extern "C" bool _dyld_find_unwind_sections(void *, dyld_unwind_sections *); #elif defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) && defined(_LIBUNWIND_IS_BAREMETAL) @@ -139,22 +98,15 @@ extern char __eh_frame_hdr_end; extern char __exidx_start; extern char __exidx_end; -#elif defined(_LIBUNWIND_ARM_EHABI) || defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) +#elif defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) && defined(_WIN32) -// ELF-based systems may use dl_iterate_phdr() to access sections -// containing unwinding information. The ElfW() macro for pointer-size -// independent ELF header traversal is not provided by <link.h> on some -// systems (e.g., FreeBSD). On these systems the data structures are -// just called Elf_XXX. Define ElfW() locally. -#ifndef _WIN32 -#include <link.h> -#else #include <windows.h> #include <psapi.h> -#endif -#if !defined(ElfW) -#define ElfW(type) Elf_##type -#endif + +#elif defined(_LIBUNWIND_USE_DL_ITERATE_PHDR) || \ + defined(_LIBUNWIND_USE_DL_UNWIND_FIND_EXIDX) + +#include <link.h> #endif @@ -162,11 +114,15 @@ namespace libunwind { /// Used by findUnwindSections() to return info about needed sections. struct UnwindInfoSections { -#if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) || defined(_LIBUNWIND_SUPPORT_DWARF_INDEX) || \ - defined(_LIBUNWIND_SUPPORT_COMPACT_UNWIND) - // No dso_base for SEH or ARM EHABI. +#if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) || \ + defined(_LIBUNWIND_SUPPORT_COMPACT_UNWIND) || \ + defined(_LIBUNWIND_USE_DL_ITERATE_PHDR) + // No dso_base for SEH. uintptr_t dso_base; #endif +#if defined(_LIBUNWIND_USE_DL_ITERATE_PHDR) + uintptr_t text_segment_length; +#endif #if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) uintptr_t dwarf_section; uintptr_t dwarf_section_length; @@ -290,11 +246,11 @@ inline int64_t LocalAddressSpace::getSLEB128(pint_t &addr, pint_t end) { if (p == pend) _LIBUNWIND_ABORT("truncated sleb128 expression"); byte = *p++; - result |= ((byte & 0x7f) << bit); + result |= (uint64_t)(byte & 0x7f) << bit; bit += 7; } while (byte & 0x80); // sign extend negative numbers - if ((byte & 0x40) != 0) + if ((byte & 0x40) != 0 && bit < 64) result |= (-1ULL) << bit; addr = (pint_t) p; return result; @@ -392,23 +348,14 @@ LocalAddressSpace::getEncodedP(pint_t &addr, pint_t end, uint8_t encoding, return result; } -#ifdef __APPLE__ -#elif defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) && defined(_LIBUNWIND_IS_BAREMETAL) -#elif defined(_LIBUNWIND_ARM_EHABI) && defined(_LIBUNWIND_IS_BAREMETAL) -#elif defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) && defined(_WIN32) -#elif defined(_LIBUNWIND_SUPPORT_SEH_UNWIND) && defined(_WIN32) -#elif defined(_LIBUNWIND_ARM_EHABI) && defined(__BIONIC__) -// Code inside findUnwindSections handles all these cases. -// -// Although the above ifdef chain is ugly, there doesn't seem to be a cleaner -// way to handle it. The generalized boolean expression is: -// -// A OR (B AND C) OR (D AND C) OR (B AND E) OR (F AND E) OR (D AND G) -// -// Running it through various boolean expression simplifiers gives expressions -// that don't help at all. -#elif defined(_LIBUNWIND_ARM_EHABI) || defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) +#if defined(_LIBUNWIND_USE_DL_ITERATE_PHDR) +// The ElfW() macro for pointer-size independent ELF header traversal is not +// provided by <link.h> on some systems (e.g., FreeBSD). On these systems the +// data structures are just called Elf_XXX. Define ElfW() locally. +#if !defined(ElfW) + #define ElfW(type) Elf_##type +#endif #if !defined(Elf_Half) typedef ElfW(Half) Elf_Half; #endif @@ -447,16 +394,12 @@ struct _LIBUNWIND_HIDDEN dl_iterate_cb_data { uintptr_t targetAddr; }; -#if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) - #if !defined(_LIBUNWIND_SUPPORT_DWARF_INDEX) - #error "_LIBUNWIND_SUPPORT_DWARF_UNWIND requires _LIBUNWIND_SUPPORT_DWARF_INDEX on this platform." - #endif - #if defined(_LIBUNWIND_USE_FRAME_HEADER_CACHE) #include "FrameHeaderCache.hpp" -// There should be just one of these per process. -static FrameHeaderCache ProcessFrameHeaderCache; +// Typically there is one cache per process, but when libunwind is built as a +// hermetic static library, then each shared object may have its own cache. +static FrameHeaderCache TheFrameHeaderCache; #endif static bool checkAddrInSegment(const Elf_Phdr *phdr, size_t image_base, @@ -466,95 +409,93 @@ static bool checkAddrInSegment(const Elf_Phdr *phdr, size_t image_base, uintptr_t end = begin + phdr->p_memsz; if (cbdata->targetAddr >= begin && cbdata->targetAddr < end) { cbdata->sects->dso_base = begin; - cbdata->sects->dwarf_section_length = phdr->p_memsz; + cbdata->sects->text_segment_length = phdr->p_memsz; return true; } } return false; } +static bool checkForUnwindInfoSegment(const Elf_Phdr *phdr, size_t image_base, + dl_iterate_cb_data *cbdata) { +#if defined(_LIBUNWIND_SUPPORT_DWARF_INDEX) + if (phdr->p_type == PT_GNU_EH_FRAME) { + EHHeaderParser<LocalAddressSpace>::EHHeaderInfo hdrInfo; + uintptr_t eh_frame_hdr_start = image_base + phdr->p_vaddr; + cbdata->sects->dwarf_index_section = eh_frame_hdr_start; + cbdata->sects->dwarf_index_section_length = phdr->p_memsz; + if (EHHeaderParser<LocalAddressSpace>::decodeEHHdr( + *cbdata->addressSpace, eh_frame_hdr_start, phdr->p_memsz, + hdrInfo)) { + // .eh_frame_hdr records the start of .eh_frame, but not its size. + // Rely on a zero terminator to find the end of the section. + cbdata->sects->dwarf_section = hdrInfo.eh_frame_ptr; + cbdata->sects->dwarf_section_length = UINTPTR_MAX; + return true; + } + } + return false; +#elif defined(_LIBUNWIND_ARM_EHABI) + if (phdr->p_type == PT_ARM_EXIDX) { + uintptr_t exidx_start = image_base + phdr->p_vaddr; + cbdata->sects->arm_section = exidx_start; + cbdata->sects->arm_section_length = phdr->p_memsz; + return true; + } + return false; +#else +#error Need one of _LIBUNWIND_SUPPORT_DWARF_INDEX or _LIBUNWIND_ARM_EHABI +#endif +} + static int findUnwindSectionsByPhdr(struct dl_phdr_info *pinfo, size_t pinfo_size, void *data) { auto cbdata = static_cast<dl_iterate_cb_data *>(data); if (pinfo->dlpi_phnum == 0 || cbdata->targetAddr < pinfo->dlpi_addr) return 0; #if defined(_LIBUNWIND_USE_FRAME_HEADER_CACHE) - if (ProcessFrameHeaderCache.find(pinfo, pinfo_size, data)) + if (TheFrameHeaderCache.find(pinfo, pinfo_size, data)) return 1; +#else + // Avoid warning about unused variable. + (void)pinfo_size; #endif Elf_Addr image_base = calculateImageBase(pinfo); - bool found_obj = false; - bool found_hdr = false; - // Third phdr is usually the executable phdr. - if (pinfo->dlpi_phnum > 2) - found_obj = checkAddrInSegment(&pinfo->dlpi_phdr[2], image_base, cbdata); + // Most shared objects seen in this callback function likely don't contain the + // target address, so optimize for that. Scan for a matching PT_LOAD segment + // first and bail when it isn't found. + bool found_text = false; + for (Elf_Half i = 0; i < pinfo->dlpi_phnum; ++i) { + if (checkAddrInSegment(&pinfo->dlpi_phdr[i], image_base, cbdata)) { + found_text = true; + break; + } + } + if (!found_text) + return 0; - // PT_GNU_EH_FRAME is usually near the end. Iterate backward. We already know - // that there is one or more phdrs. + // PT_GNU_EH_FRAME and PT_ARM_EXIDX are usually near the end. Iterate + // backward. + bool found_unwind = false; for (Elf_Half i = pinfo->dlpi_phnum; i > 0; i--) { const Elf_Phdr *phdr = &pinfo->dlpi_phdr[i - 1]; - if (!found_hdr && phdr->p_type == PT_GNU_EH_FRAME) { - EHHeaderParser<LocalAddressSpace>::EHHeaderInfo hdrInfo; - uintptr_t eh_frame_hdr_start = image_base + phdr->p_vaddr; - cbdata->sects->dwarf_index_section = eh_frame_hdr_start; - cbdata->sects->dwarf_index_section_length = phdr->p_memsz; - found_hdr = EHHeaderParser<LocalAddressSpace>::decodeEHHdr( - *cbdata->addressSpace, eh_frame_hdr_start, phdr->p_memsz, - hdrInfo); - if (found_hdr) - cbdata->sects->dwarf_section = hdrInfo.eh_frame_ptr; - } else if (!found_obj) { - found_obj = checkAddrInSegment(phdr, image_base, cbdata); - } - if (found_obj && found_hdr) { -#if defined(_LIBUNWIND_USE_FRAME_HEADER_CACHE) - ProcessFrameHeaderCache.add(cbdata->sects); -#endif - return 1; + if (checkForUnwindInfoSegment(phdr, image_base, cbdata)) { + found_unwind = true; + break; } } - cbdata->sects->dwarf_section_length = 0; - return 0; -} - -#else // defined(LIBUNWIND_SUPPORT_DWARF_UNWIND) -// Given all the #ifdef's above, the code here is for -// defined(LIBUNWIND_ARM_EHABI) - -static int findUnwindSectionsByPhdr(struct dl_phdr_info *pinfo, size_t, - void *data) { - auto *cbdata = static_cast<dl_iterate_cb_data *>(data); - bool found_obj = false; - bool found_hdr = false; - - assert(cbdata); - assert(cbdata->sects); - - if (cbdata->targetAddr < pinfo->dlpi_addr) + if (!found_unwind) return 0; - Elf_Addr image_base = calculateImageBase(pinfo); - - for (Elf_Half i = 0; i < pinfo->dlpi_phnum; i++) { - const Elf_Phdr *phdr = &pinfo->dlpi_phdr[i]; - if (phdr->p_type == PT_LOAD) { - uintptr_t begin = image_base + phdr->p_vaddr; - uintptr_t end = begin + phdr->p_memsz; - if (cbdata->targetAddr >= begin && cbdata->targetAddr < end) - found_obj = true; - } else if (phdr->p_type == PT_ARM_EXIDX) { - uintptr_t exidx_start = image_base + phdr->p_vaddr; - cbdata->sects->arm_section = exidx_start; - cbdata->sects->arm_section_length = phdr->p_memsz; - found_hdr = true; - } - } - return found_obj && found_hdr; +#if defined(_LIBUNWIND_USE_FRAME_HEADER_CACHE) + TheFrameHeaderCache.add(cbdata->sects); +#endif + return 1; } -#endif // defined(LIBUNWIND_SUPPORT_DWARF_UNWIND) -#endif // defined(_LIBUNWIND_ARM_EHABI) || defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) + +#endif // defined(_LIBUNWIND_USE_DL_ITERATE_PHDR) inline bool LocalAddressSpace::findUnwindSections(pint_t targetAddr, @@ -572,6 +513,7 @@ inline bool LocalAddressSpace::findUnwindSections(pint_t targetAddr, return true; } #elif defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) && defined(_LIBUNWIND_IS_BAREMETAL) + info.dso_base = 0; // Bare metal is statically linked, so no need to ask the dynamic loader info.dwarf_section_length = (uintptr_t)(&__eh_frame_end - &__eh_frame_start); info.dwarf_section = (uintptr_t)(&__eh_frame_start); @@ -638,16 +580,14 @@ inline bool LocalAddressSpace::findUnwindSections(pint_t targetAddr, (void)targetAddr; (void)info; return true; -#elif defined(_LIBUNWIND_ARM_EHABI) && defined(__BIONIC__) - // For ARM EHABI, Bionic didn't implement dl_iterate_phdr until API 21. After - // API 21, dl_iterate_phdr exists, but dl_unwind_find_exidx is much faster. +#elif defined(_LIBUNWIND_USE_DL_UNWIND_FIND_EXIDX) int length = 0; info.arm_section = (uintptr_t)dl_unwind_find_exidx((_Unwind_Ptr)targetAddr, &length); info.arm_section_length = (uintptr_t)length * sizeof(EHABIIndexEntry); if (info.arm_section && info.arm_section_length) return true; -#elif defined(_LIBUNWIND_ARM_EHABI) || defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) +#elif defined(_LIBUNWIND_USE_DL_ITERATE_PHDR) dl_iterate_cb_data cb_data = {this, &info, targetAddr}; int found = dl_iterate_phdr(findUnwindSectionsByPhdr, &cb_data); return static_cast<bool>(found); @@ -658,14 +598,10 @@ inline bool LocalAddressSpace::findUnwindSections(pint_t targetAddr, inline bool LocalAddressSpace::findOtherFDE(pint_t targetAddr, pint_t &fde) { -#ifdef __APPLE__ - return checkKeyMgrRegisteredFDEs(targetAddr, *((void**)&fde)); -#else // TO DO: if OS has way to dynamically register FDEs, check that. (void)targetAddr; (void)fde; return false; -#endif } inline bool LocalAddressSpace::findFunctionName(pint_t addr, char *buf, diff --git a/lib/libunwind/src/DwarfInstructions.hpp b/lib/libunwind/src/DwarfInstructions.hpp index ee98f538d4..c39cabe1f7 100644 --- a/lib/libunwind/src/DwarfInstructions.hpp +++ b/lib/libunwind/src/DwarfInstructions.hpp @@ -93,7 +93,8 @@ typename A::pint_t DwarfInstructions<A, R>::getSavedRegister( case CFI_Parser<A>::kRegisterInRegister: return registers.getRegister((int)savedReg.value); - + case CFI_Parser<A>::kRegisterUndefined: + return 0; case CFI_Parser<A>::kRegisterUnused: case CFI_Parser<A>::kRegisterOffsetFromCFA: // FIX ME @@ -117,6 +118,7 @@ double DwarfInstructions<A, R>::getSavedFloatRegister( case CFI_Parser<A>::kRegisterIsExpression: case CFI_Parser<A>::kRegisterUnused: + case CFI_Parser<A>::kRegisterUndefined: case CFI_Parser<A>::kRegisterOffsetFromCFA: case CFI_Parser<A>::kRegisterInRegister: // FIX ME @@ -140,6 +142,7 @@ v128 DwarfInstructions<A, R>::getSavedVectorRegister( case CFI_Parser<A>::kRegisterIsExpression: case CFI_Parser<A>::kRegisterUnused: + case CFI_Parser<A>::kRegisterUndefined: case CFI_Parser<A>::kRegisterOffsetFromCFA: case CFI_Parser<A>::kRegisterInRegister: // FIX ME @@ -190,6 +193,10 @@ int DwarfInstructions<A, R>::stepWithDwarf(A &addressSpace, pint_t pc, prolog.savedRegisters[i])); else return UNW_EBADREG; + } else if (i == (int)cieInfo.returnAddressRegister) { + // Leaf function keeps the return address in register and there is no + // explicit intructions how to restore it. + returnAddress = registers.getRegister(cieInfo.returnAddressRegister); } } diff --git a/lib/libunwind/src/DwarfParser.hpp b/lib/libunwind/src/DwarfParser.hpp index d05ac46836..de0eb6de9d 100644 --- a/lib/libunwind/src/DwarfParser.hpp +++ b/lib/libunwind/src/DwarfParser.hpp @@ -69,6 +69,7 @@ public: }; enum RegisterSavedWhere { kRegisterUnused, + kRegisterUndefined, kRegisterInCFA, kRegisterOffsetFromCFA, kRegisterInRegister, @@ -87,9 +88,6 @@ public: int32_t cfaRegisterOffset; // CFA = (cfaRegister)+cfaRegisterOffset int64_t cfaExpression; // CFA = expression uint32_t spExtraArgSize; - uint32_t codeOffsetAtStackDecrement; - bool registersInOtherRegisters; - bool sameValueUsed; RegisterLocation savedRegisters[kMaxRegisterNumber + 1]; enum class InitializeTime { kLazy, kNormal }; @@ -134,8 +132,26 @@ public: PrologInfo info; }; + struct RememberStack { + PrologInfoStackEntry *entry; + RememberStack() : entry(nullptr) {} + ~RememberStack() { +#if defined(_LIBUNWIND_REMEMBER_CLEANUP_NEEDED) + // Clean up rememberStack. Even in the case where every + // DW_CFA_remember_state is paired with a DW_CFA_restore_state, + // parseInstructions can skip restore opcodes if it reaches the target PC + // and stops interpreting, so we have to make sure we don't leak memory. + while (entry) { + PrologInfoStackEntry *next = entry->next; + _LIBUNWIND_REMEMBER_FREE(entry); + entry = next; + } +#endif + } + }; + static bool findFDE(A &addressSpace, pint_t pc, pint_t ehSectionStart, - uint32_t sectionLength, pint_t fdeHint, FDE_Info *fdeInfo, + uintptr_t sectionLength, pint_t fdeHint, FDE_Info *fdeInfo, CIE_Info *cieInfo); static const char *decodeFDE(A &addressSpace, pint_t fdeStart, FDE_Info *fdeInfo, CIE_Info *cieInfo); @@ -144,13 +160,6 @@ public: int arch, PrologInfo *results); static const char *parseCIE(A &addressSpace, pint_t cie, CIE_Info *cieInfo); - -private: - static bool parseInstructions(A &addressSpace, pint_t instructions, - pint_t instructionsEnd, const CIE_Info &cieInfo, - pint_t pcoffset, - PrologInfoStackEntry *&rememberStack, int arch, - PrologInfo *results); }; /// Parse a FDE into a CIE_Info and an FDE_Info @@ -166,7 +175,7 @@ const char *CFI_Parser<A>::decodeFDE(A &addressSpace, pint_t fdeStart, p += 8; } if (cfiLength == 0) - return "FDE has zero length"; // end marker + return "FDE has zero length"; // zero terminator uint32_t ciePointer = addressSpace.get32(p); if (ciePointer == 0) return "FDE is really a CIE"; // this is a CIE not an FDE @@ -211,11 +220,13 @@ const char *CFI_Parser<A>::decodeFDE(A &addressSpace, pint_t fdeStart, /// Scan an eh_frame section to find an FDE for a pc template <typename A> bool CFI_Parser<A>::findFDE(A &addressSpace, pint_t pc, pint_t ehSectionStart, - uint32_t sectionLength, pint_t fdeHint, + uintptr_t sectionLength, pint_t fdeHint, FDE_Info *fdeInfo, CIE_Info *cieInfo) { //fprintf(stderr, "findFDE(0x%llX)\n", (long long)pc); pint_t p = (fdeHint != 0) ? fdeHint : ehSectionStart; - const pint_t ehSectionEnd = p + sectionLength; + const pint_t ehSectionEnd = (sectionLength == UINTPTR_MAX) + ? static_cast<pint_t>(-1) + : (ehSectionStart + sectionLength); while (p < ehSectionEnd) { pint_t currentCFI = p; //fprintf(stderr, "findFDE() CFI at 0x%llX\n", (long long)p); @@ -227,7 +238,7 @@ bool CFI_Parser<A>::findFDE(A &addressSpace, pint_t pc, pint_t ehSectionStart, p += 8; } if (cfiLength == 0) - return false; // end marker + return false; // zero terminator uint32_t id = addressSpace.get32(p); if (id == 0) { // Skip over CIEs. @@ -336,7 +347,8 @@ const char *CFI_Parser<A>::parseCIE(A &addressSpace, pint_t cie, // parse data alignment factor cieInfo->dataAlignFactor = (int)addressSpace.getSLEB128(p, cieContentEnd); // parse return address register - uint64_t raReg = addressSpace.getULEB128(p, cieContentEnd); + uint64_t raReg = (version == 1) ? addressSpace.get8(p++) + : addressSpace.getULEB128(p, cieContentEnd); assert(raReg < 255 && "return address register too large"); cieInfo->returnAddressRegister = (uint8_t)raReg; // parse augmentation data based on augmentation string @@ -390,418 +402,409 @@ bool CFI_Parser<A>::parseFDEInstructions(A &addressSpace, const FDE_Info &fdeInfo, const CIE_Info &cieInfo, pint_t upToPC, int arch, PrologInfo *results) { - PrologInfoStackEntry *rememberStack = NULL; - - // parse CIE then FDE instructions - bool returnValue = - parseInstructions(addressSpace, cieInfo.cieInstructions, - cieInfo.cieStart + cieInfo.cieLength, cieInfo, - (pint_t)(-1), rememberStack, arch, results) && - parseInstructions(addressSpace, fdeInfo.fdeInstructions, - fdeInfo.fdeStart + fdeInfo.fdeLength, cieInfo, - upToPC - fdeInfo.pcStart, rememberStack, arch, results); - -#if !defined(_LIBUNWIND_NO_HEAP) - // Clean up rememberStack. Even in the case where every DW_CFA_remember_state - // is paired with a DW_CFA_restore_state, parseInstructions can skip restore - // opcodes if it reaches the target PC and stops interpreting, so we have to - // make sure we don't leak memory. - while (rememberStack) { - PrologInfoStackEntry *next = rememberStack->next; - free(rememberStack); - rememberStack = next; - } -#endif - - return returnValue; -} + // Alloca is used for the allocation of the rememberStack entries. It removes + // the dependency on new/malloc but the below for loop can not be refactored + // into functions. Entry could be saved during the processing of a CIE and + // restored by an FDE. + RememberStack rememberStack; + + struct ParseInfo { + pint_t instructions; + pint_t instructionsEnd; + pint_t pcoffset; + }; -/// "run" the DWARF instructions -template <typename A> -bool CFI_Parser<A>::parseInstructions(A &addressSpace, pint_t instructions, - pint_t instructionsEnd, - const CIE_Info &cieInfo, pint_t pcoffset, - PrologInfoStackEntry *&rememberStack, - int arch, PrologInfo *results) { - pint_t p = instructions; - pint_t codeOffset = 0; - // initialState initialized as registers in results are modified. Use - // PrologInfo accessor functions to avoid reading uninitialized data. - PrologInfo initialState(PrologInfo::InitializeTime::kLazy); - - _LIBUNWIND_TRACE_DWARF("parseInstructions(instructions=0x%0" PRIx64 ")\n", - static_cast<uint64_t>(instructionsEnd)); - - // see DWARF Spec, section 6.4.2 for details on unwind opcodes - while ((p < instructionsEnd) && (codeOffset < pcoffset)) { - uint64_t reg; - uint64_t reg2; - int64_t offset; - uint64_t length; - uint8_t opcode = addressSpace.get8(p); - uint8_t operand; -#if !defined(_LIBUNWIND_NO_HEAP) - PrologInfoStackEntry *entry; -#endif - ++p; - switch (opcode) { - case DW_CFA_nop: - _LIBUNWIND_TRACE_DWARF("DW_CFA_nop\n"); - break; - case DW_CFA_set_loc: - codeOffset = - addressSpace.getEncodedP(p, instructionsEnd, cieInfo.pointerEncoding); - _LIBUNWIND_TRACE_DWARF("DW_CFA_set_loc\n"); - break; - case DW_CFA_advance_loc1: - codeOffset += (addressSpace.get8(p) * cieInfo.codeAlignFactor); - p += 1; - _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc1: new offset=%" PRIu64 "\n", - static_cast<uint64_t>(codeOffset)); - break; - case DW_CFA_advance_loc2: - codeOffset += (addressSpace.get16(p) * cieInfo.codeAlignFactor); - p += 2; - _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc2: new offset=%" PRIu64 "\n", - static_cast<uint64_t>(codeOffset)); - break; - case DW_CFA_advance_loc4: - codeOffset += (addressSpace.get32(p) * cieInfo.codeAlignFactor); - p += 4; - _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc4: new offset=%" PRIu64 "\n", - static_cast<uint64_t>(codeOffset)); - break; - case DW_CFA_offset_extended: - reg = addressSpace.getULEB128(p, instructionsEnd); - offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) - * cieInfo.dataAlignFactor; - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_offset_extended DWARF unwind, reg too big"); - return false; - } - results->setRegister(reg, kRegisterInCFA, offset, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_offset_extended(reg=%" PRIu64 ", " - "offset=%" PRId64 ")\n", - reg, offset); - break; - case DW_CFA_restore_extended: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_restore_extended DWARF unwind, reg too big"); - return false; - } - results->restoreRegisterToInitialState(reg, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_restore_extended(reg=%" PRIu64 ")\n", reg); - break; - case DW_CFA_undefined: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_undefined DWARF unwind, reg too big"); - return false; - } - results->setRegisterLocation(reg, kRegisterUnused, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_undefined(reg=%" PRIu64 ")\n", reg); - break; - case DW_CFA_same_value: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_same_value DWARF unwind, reg too big"); - return false; - } - // <rdar://problem/8456377> DW_CFA_same_value unsupported - // "same value" means register was stored in frame, but its current - // value has not changed, so no need to restore from frame. - // We model this as if the register was never saved. - results->setRegisterLocation(reg, kRegisterUnused, initialState); - // set flag to disable conversion to compact unwind - results->sameValueUsed = true; - _LIBUNWIND_TRACE_DWARF("DW_CFA_same_value(reg=%" PRIu64 ")\n", reg); - break; - case DW_CFA_register: - reg = addressSpace.getULEB128(p, instructionsEnd); - reg2 = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_register DWARF unwind, reg too big"); - return false; - } - if (reg2 > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_register DWARF unwind, reg2 too big"); - return false; - } - results->setRegister(reg, kRegisterInRegister, (int64_t)reg2, - initialState); - // set flag to disable conversion to compact unwind - results->registersInOtherRegisters = true; - _LIBUNWIND_TRACE_DWARF( - "DW_CFA_register(reg=%" PRIu64 ", reg2=%" PRIu64 ")\n", reg, reg2); - break; -#if !defined(_LIBUNWIND_NO_HEAP) - case DW_CFA_remember_state: - // avoid operator new, because that would be an upward dependency - entry = (PrologInfoStackEntry *)malloc(sizeof(PrologInfoStackEntry)); - if (entry != NULL) { - entry->next = rememberStack; - entry->info = *results; - rememberStack = entry; - } else { - return false; - } - _LIBUNWIND_TRACE_DWARF("DW_CFA_remember_state\n"); - break; - case DW_CFA_restore_state: - if (rememberStack != NULL) { - PrologInfoStackEntry *top = rememberStack; - *results = top->info; - rememberStack = top->next; - free((char *)top); - } else { - return false; - } - _LIBUNWIND_TRACE_DWARF("DW_CFA_restore_state\n"); - break; -#endif - case DW_CFA_def_cfa: - reg = addressSpace.getULEB128(p, instructionsEnd); - offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0("malformed DW_CFA_def_cfa DWARF unwind, reg too big"); - return false; - } - results->cfaRegister = (uint32_t)reg; - results->cfaRegisterOffset = (int32_t)offset; - _LIBUNWIND_TRACE_DWARF( - "DW_CFA_def_cfa(reg=%" PRIu64 ", offset=%" PRIu64 ")\n", reg, offset); - break; - case DW_CFA_def_cfa_register: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_def_cfa_register DWARF unwind, reg too big"); - return false; - } - results->cfaRegister = (uint32_t)reg; - _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_register(%" PRIu64 ")\n", reg); - break; - case DW_CFA_def_cfa_offset: - results->cfaRegisterOffset = (int32_t) - addressSpace.getULEB128(p, instructionsEnd); - results->codeOffsetAtStackDecrement = (uint32_t)codeOffset; - _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_offset(%d)\n", - results->cfaRegisterOffset); - break; - case DW_CFA_def_cfa_expression: - results->cfaRegister = 0; - results->cfaExpression = (int64_t)p; - length = addressSpace.getULEB128(p, instructionsEnd); - assert(length < static_cast<pint_t>(~0) && "pointer overflow"); - p += static_cast<pint_t>(length); - _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_expression(expression=0x%" PRIx64 - ", length=%" PRIu64 ")\n", - results->cfaExpression, length); - break; - case DW_CFA_expression: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_expression DWARF unwind, reg too big"); - return false; - } - results->setRegister(reg, kRegisterAtExpression, (int64_t)p, - initialState); - length = addressSpace.getULEB128(p, instructionsEnd); - assert(length < static_cast<pint_t>(~0) && "pointer overflow"); - p += static_cast<pint_t>(length); - _LIBUNWIND_TRACE_DWARF("DW_CFA_expression(reg=%" PRIu64 ", " - "expression=0x%" PRIx64 ", " - "length=%" PRIu64 ")\n", - reg, results->savedRegisters[reg].value, length); - break; - case DW_CFA_offset_extended_sf: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_offset_extended_sf DWARF unwind, reg too big"); - return false; - } - offset = - addressSpace.getSLEB128(p, instructionsEnd) * cieInfo.dataAlignFactor; - results->setRegister(reg, kRegisterInCFA, offset, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_offset_extended_sf(reg=%" PRIu64 ", " - "offset=%" PRId64 ")\n", - reg, offset); - break; - case DW_CFA_def_cfa_sf: - reg = addressSpace.getULEB128(p, instructionsEnd); - offset = - addressSpace.getSLEB128(p, instructionsEnd) * cieInfo.dataAlignFactor; - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_def_cfa_sf DWARF unwind, reg too big"); - return false; - } - results->cfaRegister = (uint32_t)reg; - results->cfaRegisterOffset = (int32_t)offset; - _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_sf(reg=%" PRIu64 ", " - "offset=%" PRId64 ")\n", - reg, offset); - break; - case DW_CFA_def_cfa_offset_sf: - results->cfaRegisterOffset = (int32_t) - (addressSpace.getSLEB128(p, instructionsEnd) * cieInfo.dataAlignFactor); - results->codeOffsetAtStackDecrement = (uint32_t)codeOffset; - _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_offset_sf(%d)\n", - results->cfaRegisterOffset); - break; - case DW_CFA_val_offset: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG( - "malformed DW_CFA_val_offset DWARF unwind, reg (%" PRIu64 - ") out of range\n", - reg); - return false; - } - offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) - * cieInfo.dataAlignFactor; - results->setRegister(reg, kRegisterOffsetFromCFA, offset, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_val_offset(reg=%" PRIu64 ", " - "offset=%" PRId64 "\n", - reg, offset); - break; - case DW_CFA_val_offset_sf: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_val_offset_sf DWARF unwind, reg too big"); - return false; - } - offset = - addressSpace.getSLEB128(p, instructionsEnd) * cieInfo.dataAlignFactor; - results->setRegister(reg, kRegisterOffsetFromCFA, offset, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_val_offset_sf(reg=%" PRIu64 ", " - "offset=%" PRId64 "\n", - reg, offset); - break; - case DW_CFA_val_expression: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0( - "malformed DW_CFA_val_expression DWARF unwind, reg too big"); - return false; - } - results->setRegister(reg, kRegisterIsExpression, (int64_t)p, - initialState); - length = addressSpace.getULEB128(p, instructionsEnd); - assert(length < static_cast<pint_t>(~0) && "pointer overflow"); - p += static_cast<pint_t>(length); - _LIBUNWIND_TRACE_DWARF("DW_CFA_val_expression(reg=%" PRIu64 ", " - "expression=0x%" PRIx64 ", length=%" PRIu64 ")\n", - reg, results->savedRegisters[reg].value, length); - break; - case DW_CFA_GNU_args_size: - length = addressSpace.getULEB128(p, instructionsEnd); - results->spExtraArgSize = (uint32_t)length; - _LIBUNWIND_TRACE_DWARF("DW_CFA_GNU_args_size(%" PRIu64 ")\n", length); - break; - case DW_CFA_GNU_negative_offset_extended: - reg = addressSpace.getULEB128(p, instructionsEnd); - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG0("malformed DW_CFA_GNU_negative_offset_extended DWARF " - "unwind, reg too big"); - return false; + ParseInfo parseInfoArray[] = { + {cieInfo.cieInstructions, cieInfo.cieStart + cieInfo.cieLength, + (pint_t)(-1)}, + {fdeInfo.fdeInstructions, fdeInfo.fdeStart + fdeInfo.fdeLength, + upToPC - fdeInfo.pcStart}}; + + for (const auto &info : parseInfoArray) { + pint_t p = info.instructions; + pint_t instructionsEnd = info.instructionsEnd; + pint_t pcoffset = info.pcoffset; + pint_t codeOffset = 0; + + // initialState initialized as registers in results are modified. Use + // PrologInfo accessor functions to avoid reading uninitialized data. + PrologInfo initialState(PrologInfo::InitializeTime::kLazy); + + _LIBUNWIND_TRACE_DWARF("parseFDEInstructions(instructions=0x%0" PRIx64 + ")\n", + static_cast<uint64_t>(instructionsEnd)); + + // see DWARF Spec, section 6.4.2 for details on unwind opcodes + while ((p < instructionsEnd) && (codeOffset < pcoffset)) { + uint64_t reg; + uint64_t reg2; + int64_t offset; + uint64_t length; + uint8_t opcode = addressSpace.get8(p); + uint8_t operand; + + ++p; + switch (opcode) { + case DW_CFA_nop: + _LIBUNWIND_TRACE_DWARF("DW_CFA_nop\n"); + break; + case DW_CFA_set_loc: + codeOffset = addressSpace.getEncodedP(p, instructionsEnd, + cieInfo.pointerEncoding); + _LIBUNWIND_TRACE_DWARF("DW_CFA_set_loc\n"); + break; + case DW_CFA_advance_loc1: + codeOffset += (addressSpace.get8(p) * cieInfo.codeAlignFactor); + p += 1; + _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc1: new offset=%" PRIu64 "\n", + static_cast<uint64_t>(codeOffset)); + break; + case DW_CFA_advance_loc2: + codeOffset += (addressSpace.get16(p) * cieInfo.codeAlignFactor); + p += 2; + _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc2: new offset=%" PRIu64 "\n", + static_cast<uint64_t>(codeOffset)); + break; + case DW_CFA_advance_loc4: + codeOffset += (addressSpace.get32(p) * cieInfo.codeAlignFactor); + p += 4; + _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc4: new offset=%" PRIu64 "\n", + static_cast<uint64_t>(codeOffset)); + break; + case DW_CFA_offset_extended: + reg = addressSpace.getULEB128(p, instructionsEnd); + offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor; + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_offset_extended DWARF unwind, reg too big"); + return false; + } + results->setRegister(reg, kRegisterInCFA, offset, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_offset_extended(reg=%" PRIu64 ", " + "offset=%" PRId64 ")\n", + reg, offset); + break; + case DW_CFA_restore_extended: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_restore_extended DWARF unwind, reg too big"); + return false; + } + results->restoreRegisterToInitialState(reg, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_restore_extended(reg=%" PRIu64 ")\n", + reg); + break; + case DW_CFA_undefined: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_undefined DWARF unwind, reg too big"); + return false; + } + results->setRegisterLocation(reg, kRegisterUndefined, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_undefined(reg=%" PRIu64 ")\n", reg); + break; + case DW_CFA_same_value: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_same_value DWARF unwind, reg too big"); + return false; + } + // <rdar://problem/8456377> DW_CFA_same_value unsupported + // "same value" means register was stored in frame, but its current + // value has not changed, so no need to restore from frame. + // We model this as if the register was never saved. + results->setRegisterLocation(reg, kRegisterUnused, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_same_value(reg=%" PRIu64 ")\n", reg); + break; + case DW_CFA_register: + reg = addressSpace.getULEB128(p, instructionsEnd); + reg2 = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_register DWARF unwind, reg too big"); + return false; + } + if (reg2 > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_register DWARF unwind, reg2 too big"); + return false; + } + results->setRegister(reg, kRegisterInRegister, (int64_t)reg2, + initialState); + _LIBUNWIND_TRACE_DWARF( + "DW_CFA_register(reg=%" PRIu64 ", reg2=%" PRIu64 ")\n", reg, reg2); + break; + case DW_CFA_remember_state: { + // Avoid operator new because that would be an upward dependency. + // Avoid malloc because it needs heap allocation. + PrologInfoStackEntry *entry = + (PrologInfoStackEntry *)_LIBUNWIND_REMEMBER_ALLOC( + sizeof(PrologInfoStackEntry)); + if (entry != NULL) { + entry->next = rememberStack.entry; + entry->info = *results; + rememberStack.entry = entry; + } else { + return false; + } + _LIBUNWIND_TRACE_DWARF("DW_CFA_remember_state\n"); + break; } - offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) - * cieInfo.dataAlignFactor; - results->setRegister(reg, kRegisterInCFA, -offset, initialState); - _LIBUNWIND_TRACE_DWARF( - "DW_CFA_GNU_negative_offset_extended(%" PRId64 ")\n", offset); - break; + case DW_CFA_restore_state: + if (rememberStack.entry != NULL) { + PrologInfoStackEntry *top = rememberStack.entry; + *results = top->info; + rememberStack.entry = top->next; + _LIBUNWIND_REMEMBER_FREE(top); + } else { + return false; + } + _LIBUNWIND_TRACE_DWARF("DW_CFA_restore_state\n"); + break; + case DW_CFA_def_cfa: + reg = addressSpace.getULEB128(p, instructionsEnd); + offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0("malformed DW_CFA_def_cfa DWARF unwind, reg too big"); + return false; + } + results->cfaRegister = (uint32_t)reg; + results->cfaRegisterOffset = (int32_t)offset; + _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa(reg=%" PRIu64 ", offset=%" PRIu64 + ")\n", + reg, offset); + break; + case DW_CFA_def_cfa_register: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_def_cfa_register DWARF unwind, reg too big"); + return false; + } + results->cfaRegister = (uint32_t)reg; + _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_register(%" PRIu64 ")\n", reg); + break; + case DW_CFA_def_cfa_offset: + results->cfaRegisterOffset = + (int32_t)addressSpace.getULEB128(p, instructionsEnd); + _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_offset(%d)\n", + results->cfaRegisterOffset); + break; + case DW_CFA_def_cfa_expression: + results->cfaRegister = 0; + results->cfaExpression = (int64_t)p; + length = addressSpace.getULEB128(p, instructionsEnd); + assert(length < static_cast<pint_t>(~0) && "pointer overflow"); + p += static_cast<pint_t>(length); + _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_expression(expression=0x%" PRIx64 + ", length=%" PRIu64 ")\n", + results->cfaExpression, length); + break; + case DW_CFA_expression: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_expression DWARF unwind, reg too big"); + return false; + } + results->setRegister(reg, kRegisterAtExpression, (int64_t)p, + initialState); + length = addressSpace.getULEB128(p, instructionsEnd); + assert(length < static_cast<pint_t>(~0) && "pointer overflow"); + p += static_cast<pint_t>(length); + _LIBUNWIND_TRACE_DWARF("DW_CFA_expression(reg=%" PRIu64 ", " + "expression=0x%" PRIx64 ", " + "length=%" PRIu64 ")\n", + reg, results->savedRegisters[reg].value, length); + break; + case DW_CFA_offset_extended_sf: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_offset_extended_sf DWARF unwind, reg too big"); + return false; + } + offset = addressSpace.getSLEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor; + results->setRegister(reg, kRegisterInCFA, offset, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_offset_extended_sf(reg=%" PRIu64 ", " + "offset=%" PRId64 ")\n", + reg, offset); + break; + case DW_CFA_def_cfa_sf: + reg = addressSpace.getULEB128(p, instructionsEnd); + offset = addressSpace.getSLEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor; + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_def_cfa_sf DWARF unwind, reg too big"); + return false; + } + results->cfaRegister = (uint32_t)reg; + results->cfaRegisterOffset = (int32_t)offset; + _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_sf(reg=%" PRIu64 ", " + "offset=%" PRId64 ")\n", + reg, offset); + break; + case DW_CFA_def_cfa_offset_sf: + results->cfaRegisterOffset = + (int32_t)(addressSpace.getSLEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor); + _LIBUNWIND_TRACE_DWARF("DW_CFA_def_cfa_offset_sf(%d)\n", + results->cfaRegisterOffset); + break; + case DW_CFA_val_offset: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG( + "malformed DW_CFA_val_offset DWARF unwind, reg (%" PRIu64 + ") out of range\n", + reg); + return false; + } + offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor; + results->setRegister(reg, kRegisterOffsetFromCFA, offset, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_val_offset(reg=%" PRIu64 ", " + "offset=%" PRId64 "\n", + reg, offset); + break; + case DW_CFA_val_offset_sf: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_val_offset_sf DWARF unwind, reg too big"); + return false; + } + offset = addressSpace.getSLEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor; + results->setRegister(reg, kRegisterOffsetFromCFA, offset, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_val_offset_sf(reg=%" PRIu64 ", " + "offset=%" PRId64 "\n", + reg, offset); + break; + case DW_CFA_val_expression: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0( + "malformed DW_CFA_val_expression DWARF unwind, reg too big"); + return false; + } + results->setRegister(reg, kRegisterIsExpression, (int64_t)p, + initialState); + length = addressSpace.getULEB128(p, instructionsEnd); + assert(length < static_cast<pint_t>(~0) && "pointer overflow"); + p += static_cast<pint_t>(length); + _LIBUNWIND_TRACE_DWARF("DW_CFA_val_expression(reg=%" PRIu64 ", " + "expression=0x%" PRIx64 ", length=%" PRIu64 + ")\n", + reg, results->savedRegisters[reg].value, length); + break; + case DW_CFA_GNU_args_size: + length = addressSpace.getULEB128(p, instructionsEnd); + results->spExtraArgSize = (uint32_t)length; + _LIBUNWIND_TRACE_DWARF("DW_CFA_GNU_args_size(%" PRIu64 ")\n", length); + break; + case DW_CFA_GNU_negative_offset_extended: + reg = addressSpace.getULEB128(p, instructionsEnd); + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG0("malformed DW_CFA_GNU_negative_offset_extended DWARF " + "unwind, reg too big"); + return false; + } + offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor; + results->setRegister(reg, kRegisterInCFA, -offset, initialState); + _LIBUNWIND_TRACE_DWARF( + "DW_CFA_GNU_negative_offset_extended(%" PRId64 ")\n", offset); + break; #if defined(_LIBUNWIND_TARGET_AARCH64) || defined(_LIBUNWIND_TARGET_SPARC) - // The same constant is used to represent different instructions on - // AArch64 (negate_ra_state) and SPARC (window_save). - static_assert(DW_CFA_AARCH64_negate_ra_state == DW_CFA_GNU_window_save, - "uses the same constant"); - case DW_CFA_AARCH64_negate_ra_state: - switch (arch) { + // The same constant is used to represent different instructions on + // AArch64 (negate_ra_state) and SPARC (window_save). + static_assert(DW_CFA_AARCH64_negate_ra_state == DW_CFA_GNU_window_save, + "uses the same constant"); + case DW_CFA_AARCH64_negate_ra_state: + switch (arch) { #if defined(_LIBUNWIND_TARGET_AARCH64) case REGISTERS_ARM64: { int64_t value = results->savedRegisters[UNW_ARM64_RA_SIGN_STATE].value ^ 0x1; - results->setRegisterValue(UNW_ARM64_RA_SIGN_STATE, value, initialState); + results->setRegisterValue(UNW_ARM64_RA_SIGN_STATE, value, + initialState); _LIBUNWIND_TRACE_DWARF("DW_CFA_AARCH64_negate_ra_state\n"); } break; #endif #if defined(_LIBUNWIND_TARGET_SPARC) - // case DW_CFA_GNU_window_save: - case REGISTERS_SPARC: - _LIBUNWIND_TRACE_DWARF("DW_CFA_GNU_window_save()\n"); - for (reg = UNW_SPARC_O0; reg <= UNW_SPARC_O7; reg++) { - results->setRegister(reg, kRegisterInRegister, - ((int64_t)reg - UNW_SPARC_O0) + UNW_SPARC_I0, - initialState); - } + // case DW_CFA_GNU_window_save: + case REGISTERS_SPARC: + _LIBUNWIND_TRACE_DWARF("DW_CFA_GNU_window_save()\n"); + for (reg = UNW_SPARC_O0; reg <= UNW_SPARC_O7; reg++) { + results->setRegister(reg, kRegisterInRegister, + ((int64_t)reg - UNW_SPARC_O0) + UNW_SPARC_I0, + initialState); + } - for (reg = UNW_SPARC_L0; reg <= UNW_SPARC_I7; reg++) { - results->setRegister(reg, kRegisterInCFA, - ((int64_t)reg - UNW_SPARC_L0) * 4, initialState); + for (reg = UNW_SPARC_L0; reg <= UNW_SPARC_I7; reg++) { + results->setRegister(reg, kRegisterInCFA, + ((int64_t)reg - UNW_SPARC_L0) * 4, + initialState); + } + break; +#endif } break; -#endif - } - break; #else - (void)arch; + (void)arch; #endif - default: - operand = opcode & 0x3F; - switch (opcode & 0xC0) { - case DW_CFA_offset: - reg = operand; - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG("malformed DW_CFA_offset DWARF unwind, reg (%" PRIu64 - ") out of range", - reg); - return false; - } - offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) - * cieInfo.dataAlignFactor; - results->setRegister(reg, kRegisterInCFA, offset, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_offset(reg=%d, offset=%" PRId64 ")\n", - operand, offset); - break; - case DW_CFA_advance_loc: - codeOffset += operand * cieInfo.codeAlignFactor; - _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc: new offset=%" PRIu64 "\n", - static_cast<uint64_t>(codeOffset)); - break; - case DW_CFA_restore: - reg = operand; - if (reg > kMaxRegisterNumber) { - _LIBUNWIND_LOG("malformed DW_CFA_restore DWARF unwind, reg (%" PRIu64 - ") out of range", - reg); + default: + operand = opcode & 0x3F; + switch (opcode & 0xC0) { + case DW_CFA_offset: + reg = operand; + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG("malformed DW_CFA_offset DWARF unwind, reg (%" PRIu64 + ") out of range", + reg); + return false; + } + offset = (int64_t)addressSpace.getULEB128(p, instructionsEnd) * + cieInfo.dataAlignFactor; + results->setRegister(reg, kRegisterInCFA, offset, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_offset(reg=%d, offset=%" PRId64 ")\n", + operand, offset); + break; + case DW_CFA_advance_loc: + codeOffset += operand * cieInfo.codeAlignFactor; + _LIBUNWIND_TRACE_DWARF("DW_CFA_advance_loc: new offset=%" PRIu64 "\n", + static_cast<uint64_t>(codeOffset)); + break; + case DW_CFA_restore: + reg = operand; + if (reg > kMaxRegisterNumber) { + _LIBUNWIND_LOG( + "malformed DW_CFA_restore DWARF unwind, reg (%" PRIu64 + ") out of range", + reg); + return false; + } + results->restoreRegisterToInitialState(reg, initialState); + _LIBUNWIND_TRACE_DWARF("DW_CFA_restore(reg=%" PRIu64 ")\n", + static_cast<uint64_t>(operand)); + break; + default: + _LIBUNWIND_TRACE_DWARF("unknown CFA opcode 0x%02X\n", opcode); return false; } - results->restoreRegisterToInitialState(reg, initialState); - _LIBUNWIND_TRACE_DWARF("DW_CFA_restore(reg=%" PRIu64 ")\n", - static_cast<uint64_t>(operand)); - break; - default: - _LIBUNWIND_TRACE_DWARF("unknown CFA opcode 0x%02X\n", opcode); - return false; } } } - return true; } diff --git a/lib/libunwind/src/FrameHeaderCache.hpp b/lib/libunwind/src/FrameHeaderCache.hpp index 813fcd408b..54d5d33c3c 100644 --- a/lib/libunwind/src/FrameHeaderCache.hpp +++ b/lib/libunwind/src/FrameHeaderCache.hpp @@ -32,7 +32,7 @@ class _LIBUNWIND_HIDDEN FrameHeaderCache { struct CacheEntry { uintptr_t LowPC() { return Info.dso_base; }; - uintptr_t HighPC() { return Info.dso_base + Info.dwarf_section_length; }; + uintptr_t HighPC() { return Info.dso_base + Info.text_segment_length; }; UnwindInfoSections Info; CacheEntry *Next; }; diff --git a/lib/libunwind/src/Registers.hpp b/lib/libunwind/src/Registers.hpp index c76b05bf31..efeaf43559 100644 --- a/lib/libunwind/src/Registers.hpp +++ b/lib/libunwind/src/Registers.hpp @@ -36,9 +36,12 @@ enum { REGISTERS_SPARC, REGISTERS_HEXAGON, REGISTERS_RISCV, + REGISTERS_VE, }; #if defined(_LIBUNWIND_TARGET_I386) +class _LIBUNWIND_HIDDEN Registers_x86; +extern "C" void __libunwind_Registers_x86_jumpto(Registers_x86 *); /// Registers_x86 holds the register state of a thread in a 32-bit intel /// process. class _LIBUNWIND_HIDDEN Registers_x86 { @@ -56,7 +59,7 @@ public: v128 getVectorRegister(int num) const; void setVectorRegister(int num, v128 value); static const char *getRegisterName(int num); - void jumpto(); + void jumpto() { __libunwind_Registers_x86_jumpto(this); } static int lastDwarfRegNum() { return _LIBUNWIND_HIGHEST_DWARF_REGISTER_X86; } static int getArch() { return REGISTERS_X86; } @@ -248,6 +251,8 @@ inline void Registers_x86::setVectorRegister(int, v128) { #if defined(_LIBUNWIND_TARGET_X86_64) /// Registers_x86_64 holds the register state of a thread in a 64-bit intel /// process. +class _LIBUNWIND_HIDDEN Registers_x86_64; +extern "C" void __libunwind_Registers_x86_64_jumpto(Registers_x86_64 *); class _LIBUNWIND_HIDDEN Registers_x86_64 { public: Registers_x86_64(); @@ -263,7 +268,7 @@ public: v128 getVectorRegister(int num) const; void setVectorRegister(int num, v128 value); static const char *getRegisterName(int num); - void jumpto(); + void jumpto() { __libunwind_Registers_x86_64_jumpto(this); } static int lastDwarfRegNum() { return _LIBUNWIND_HIGHEST_DWARF_REGISTER_X86_64; } static int getArch() { return REGISTERS_X86_64; } @@ -1510,12 +1515,12 @@ inline void Registers_ppc64::setFloatRegister(int regNum, double value) { } inline bool Registers_ppc64::validVectorRegister(int regNum) const { -#ifdef PPC64_HAS_VMX +#if defined(__VSX__) if (regNum >= UNW_PPC64_VS0 && regNum <= UNW_PPC64_VS31) return true; if (regNum >= UNW_PPC64_VS32 && regNum <= UNW_PPC64_VS63) return true; -#else +#elif defined(__ALTIVEC__) if (regNum >= UNW_PPC64_V0 && regNum <= UNW_PPC64_V31) return true; #endif @@ -1771,6 +1776,8 @@ inline const char *Registers_ppc64::getRegisterName(int regNum) { #if defined(_LIBUNWIND_TARGET_AARCH64) /// Registers_arm64 holds the register state of a thread in a 64-bit arm /// process. +class _LIBUNWIND_HIDDEN Registers_arm64; +extern "C" void __libunwind_Registers_arm64_jumpto(Registers_arm64 *); class _LIBUNWIND_HIDDEN Registers_arm64 { public: Registers_arm64(); @@ -1786,7 +1793,7 @@ public: v128 getVectorRegister(int num) const; void setVectorRegister(int num, v128 value); static const char *getRegisterName(int num); - void jumpto(); + void jumpto() { __libunwind_Registers_arm64_jumpto(this); } static int lastDwarfRegNum() { return _LIBUNWIND_HIGHEST_DWARF_REGISTER_ARM64; } static int getArch() { return REGISTERS_ARM64; } @@ -3977,6 +3984,447 @@ inline void Registers_riscv::setVectorRegister(int, v128) { _LIBUNWIND_ABORT("no riscv vector register support yet"); } #endif // _LIBUNWIND_TARGET_RISCV + +#if defined(_LIBUNWIND_TARGET_VE) +/// Registers_ve holds the register state of a thread in a VE process. +class _LIBUNWIND_HIDDEN Registers_ve { +public: + Registers_ve(); + Registers_ve(const void *registers); + + bool validRegister(int num) const; + uint64_t getRegister(int num) const; + void setRegister(int num, uint64_t value); + bool validFloatRegister(int num) const; + double getFloatRegister(int num) const; + void setFloatRegister(int num, double value); + bool validVectorRegister(int num) const; + v128 getVectorRegister(int num) const; + void setVectorRegister(int num, v128 value); + static const char *getRegisterName(int num); + void jumpto(); + static int lastDwarfRegNum() { return _LIBUNWIND_HIGHEST_DWARF_REGISTER_VE; } + static int getArch() { return REGISTERS_VE; } + + uint64_t getSP() const { return _registers.__s[11]; } + void setSP(uint64_t value) { _registers.__s[11] = value; } + uint64_t getIP() const { return _registers.__ic; } + void setIP(uint64_t value) { _registers.__ic = value; } + +private: + // FIXME: Need to store not only scalar registers but also vector and vector + // mask registers. VEOS uses mcontext_t defined in ucontext.h. It takes + // 524288 bytes (65536*8 bytes), though. Currently, we use libunwind for + // SjLj exception support only, so Registers_ve is not implemented completely. + struct ve_thread_state_t { + uint64_t __s[64]; // s0-s64 + uint64_t __ic; // Instruction counter (IC) + uint64_t __vixr; // Vector Index Register + uint64_t __vl; // Vector Length Register + }; + + ve_thread_state_t _registers; // total 67 registers + + // Currently no vector register is preserved. +}; + +inline Registers_ve::Registers_ve(const void *registers) { + static_assert((check_fit<Registers_ve, unw_context_t>::does_fit), + "ve registers do not fit into unw_context_t"); + memcpy(&_registers, static_cast<const uint8_t *>(registers), + sizeof(_registers)); + static_assert(sizeof(_registers) == 536, + "expected vector register offset to be 536"); +} + +inline Registers_ve::Registers_ve() { + memset(&_registers, 0, sizeof(_registers)); +} + +inline bool Registers_ve::validRegister(int regNum) const { + if (regNum >= UNW_VE_S0 && regNum <= UNW_VE_S63) + return true; + + switch (regNum) { + case UNW_REG_IP: + case UNW_REG_SP: + case UNW_VE_VIXR: + case UNW_VE_VL: + return true; + default: + return false; + } +} + +inline uint64_t Registers_ve::getRegister(int regNum) const { + if (regNum >= UNW_VE_S0 && regNum <= UNW_VE_S63) + return _registers.__s[regNum - UNW_VE_S0]; + + switch (regNum) { + case UNW_REG_IP: + return _registers.__ic; + case UNW_REG_SP: + return _registers.__s[11]; + case UNW_VE_VIXR: + return _registers.__vixr; + case UNW_VE_VL: + return _registers.__vl; + } + _LIBUNWIND_ABORT("unsupported ve register"); +} + +inline void Registers_ve::setRegister(int regNum, uint64_t value) { + if (regNum >= UNW_VE_S0 && regNum <= UNW_VE_S63) { + _registers.__s[regNum - UNW_VE_S0] = value; + return; + } + + switch (regNum) { + case UNW_REG_IP: + _registers.__ic = value; + return; + case UNW_REG_SP: + _registers.__s[11] = value; + return; + case UNW_VE_VIXR: + _registers.__vixr = value; + return; + case UNW_VE_VL: + _registers.__vl = value; + return; + } + _LIBUNWIND_ABORT("unsupported ve register"); +} + +inline bool Registers_ve::validFloatRegister(int /* regNum */) const { + return false; +} + +inline double Registers_ve::getFloatRegister(int /* regNum */) const { + _LIBUNWIND_ABORT("VE doesn't have float registers"); +} + +inline void Registers_ve::setFloatRegister(int /* regNum */, + double /* value */) { + _LIBUNWIND_ABORT("VE doesn't have float registers"); +} + +inline bool Registers_ve::validVectorRegister(int /* regNum */) const { + return false; +} + +inline v128 Registers_ve::getVectorRegister(int /* regNum */) const { + _LIBUNWIND_ABORT("VE vector support not implemented"); +} + +inline void Registers_ve::setVectorRegister(int /* regNum */, + v128 /* value */) { + _LIBUNWIND_ABORT("VE vector support not implemented"); +} + +inline const char *Registers_ve::getRegisterName(int regNum) { + switch (regNum) { + case UNW_REG_IP: + return "ip"; + case UNW_REG_SP: + return "sp"; + case UNW_VE_VIXR: + return "vixr"; + case UNW_VE_VL: + return "vl"; + case UNW_VE_S0: + return "s0"; + case UNW_VE_S1: + return "s1"; + case UNW_VE_S2: + return "s2"; + case UNW_VE_S3: + return "s3"; + case UNW_VE_S4: + return "s4"; + case UNW_VE_S5: + return "s5"; + case UNW_VE_S6: + return "s6"; + case UNW_VE_S7: + return "s7"; + case UNW_VE_S8: + return "s8"; + case UNW_VE_S9: + return "s9"; + case UNW_VE_S10: + return "s10"; + case UNW_VE_S11: + return "s11"; + case UNW_VE_S12: + return "s12"; + case UNW_VE_S13: + return "s13"; + case UNW_VE_S14: + return "s14"; + case UNW_VE_S15: + return "s15"; + case UNW_VE_S16: + return "s16"; + case UNW_VE_S17: + return "s17"; + case UNW_VE_S18: + return "s18"; + case UNW_VE_S19: + return "s19"; + case UNW_VE_S20: + return "s20"; + case UNW_VE_S21: + return "s21"; + case UNW_VE_S22: + return "s22"; + case UNW_VE_S23: + return "s23"; + case UNW_VE_S24: + return "s24"; + case UNW_VE_S25: + return "s25"; + case UNW_VE_S26: + return "s26"; + case UNW_VE_S27: + return "s27"; + case UNW_VE_S28: + return "s28"; + case UNW_VE_S29: + return "s29"; + case UNW_VE_S30: + return "s30"; + case UNW_VE_S31: + return "s31"; + case UNW_VE_S32: + return "s32"; + case UNW_VE_S33: + return "s33"; + case UNW_VE_S34: + return "s34"; + case UNW_VE_S35: + return "s35"; + case UNW_VE_S36: + return "s36"; + case UNW_VE_S37: + return "s37"; + case UNW_VE_S38: + return "s38"; + case UNW_VE_S39: + return "s39"; + case UNW_VE_S40: + return "s40"; + case UNW_VE_S41: + return "s41"; + case UNW_VE_S42: + return "s42"; + case UNW_VE_S43: + return "s43"; + case UNW_VE_S44: + return "s44"; + case UNW_VE_S45: + return "s45"; + case UNW_VE_S46: + return "s46"; + case UNW_VE_S47: + return "s47"; + case UNW_VE_S48: + return "s48"; + case UNW_VE_S49: + return "s49"; + case UNW_VE_S50: + return "s50"; + case UNW_VE_S51: + return "s51"; + case UNW_VE_S52: + return "s52"; + case UNW_VE_S53: + return "s53"; + case UNW_VE_S54: + return "s54"; + case UNW_VE_S55: + return "s55"; + case UNW_VE_S56: + return "s56"; + case UNW_VE_S57: + return "s57"; + case UNW_VE_S58: + return "s58"; + case UNW_VE_S59: + return "s59"; + case UNW_VE_S60: + return "s60"; + case UNW_VE_S61: + return "s61"; + case UNW_VE_S62: + return "s62"; + case UNW_VE_S63: + return "s63"; + case UNW_VE_V0: + return "v0"; + case UNW_VE_V1: + return "v1"; + case UNW_VE_V2: + return "v2"; + case UNW_VE_V3: + return "v3"; + case UNW_VE_V4: + return "v4"; + case UNW_VE_V5: + return "v5"; + case UNW_VE_V6: + return "v6"; + case UNW_VE_V7: + return "v7"; + case UNW_VE_V8: + return "v8"; + case UNW_VE_V9: + return "v9"; + case UNW_VE_V10: + return "v10"; + case UNW_VE_V11: + return "v11"; + case UNW_VE_V12: + return "v12"; + case UNW_VE_V13: + return "v13"; + case UNW_VE_V14: + return "v14"; + case UNW_VE_V15: + return "v15"; + case UNW_VE_V16: + return "v16"; + case UNW_VE_V17: + return "v17"; + case UNW_VE_V18: + return "v18"; + case UNW_VE_V19: + return "v19"; + case UNW_VE_V20: + return "v20"; + case UNW_VE_V21: + return "v21"; + case UNW_VE_V22: + return "v22"; + case UNW_VE_V23: + return "v23"; + case UNW_VE_V24: + return "v24"; + case UNW_VE_V25: + return "v25"; + case UNW_VE_V26: + return "v26"; + case UNW_VE_V27: + return "v27"; + case UNW_VE_V28: + return "v28"; + case UNW_VE_V29: + return "v29"; + case UNW_VE_V30: + return "v30"; + case UNW_VE_V31: + return "v31"; + case UNW_VE_V32: + return "v32"; + case UNW_VE_V33: + return "v33"; + case UNW_VE_V34: + return "v34"; + case UNW_VE_V35: + return "v35"; + case UNW_VE_V36: + return "v36"; + case UNW_VE_V37: + return "v37"; + case UNW_VE_V38: + return "v38"; + case UNW_VE_V39: + return "v39"; + case UNW_VE_V40: + return "v40"; + case UNW_VE_V41: + return "v41"; + case UNW_VE_V42: + return "v42"; + case UNW_VE_V43: + return "v43"; + case UNW_VE_V44: + return "v44"; + case UNW_VE_V45: + return "v45"; + case UNW_VE_V46: + return "v46"; + case UNW_VE_V47: + return "v47"; + case UNW_VE_V48: + return "v48"; + case UNW_VE_V49: + return "v49"; + case UNW_VE_V50: + return "v50"; + case UNW_VE_V51: + return "v51"; + case UNW_VE_V52: + return "v52"; + case UNW_VE_V53: + return "v53"; + case UNW_VE_V54: + return "v54"; + case UNW_VE_V55: + return "v55"; + case UNW_VE_V56: + return "v56"; + case UNW_VE_V57: + return "v57"; + case UNW_VE_V58: + return "v58"; + case UNW_VE_V59: + return "v59"; + case UNW_VE_V60: + return "v60"; + case UNW_VE_V61: + return "v61"; + case UNW_VE_V62: + return "v62"; + case UNW_VE_V63: + return "v63"; + case UNW_VE_VM0: + return "vm0"; + case UNW_VE_VM1: + return "vm1"; + case UNW_VE_VM2: + return "vm2"; + case UNW_VE_VM3: + return "vm3"; + case UNW_VE_VM4: + return "vm4"; + case UNW_VE_VM5: + return "vm5"; + case UNW_VE_VM6: + return "vm6"; + case UNW_VE_VM7: + return "vm7"; + case UNW_VE_VM8: + return "vm8"; + case UNW_VE_VM9: + return "vm9"; + case UNW_VE_VM10: + return "vm10"; + case UNW_VE_VM11: + return "vm11"; + case UNW_VE_VM12: + return "vm12"; + case UNW_VE_VM13: + return "vm13"; + case UNW_VE_VM14: + return "vm14"; + case UNW_VE_VM15: + return "vm15"; + } + return "unknown register"; +} +#endif // _LIBUNWIND_TARGET_VE + } // namespace libunwind #endif // __REGISTERS_HPP__ diff --git a/lib/libunwind/src/Unwind-seh.cpp b/lib/libunwind/src/Unwind-seh.cpp index 403ab2d771..6e2b4e73e4 100644 --- a/lib/libunwind/src/Unwind-seh.cpp +++ b/lib/libunwind/src/Unwind-seh.cpp @@ -46,18 +46,6 @@ using namespace libunwind; /// handling. #define STATUS_GCC_UNWIND MAKE_GCC_EXCEPTION(1) // 0x21474343 -/// Class of foreign exceptions based on unrecognized SEH exceptions. -static const uint64_t kSEHExceptionClass = 0x434C4E4753454800; // CLNGSEH\0 - -/// Exception cleanup routine used by \c _GCC_specific_handler to -/// free foreign exceptions. -static void seh_exc_cleanup(_Unwind_Reason_Code urc, _Unwind_Exception *exc) { - (void)urc; - if (exc->exception_class != kSEHExceptionClass) - _LIBUNWIND_ABORT("SEH cleanup called on non-SEH exception"); - free(exc); -} - static int __unw_init_seh(unw_cursor_t *cursor, CONTEXT *ctx); static DISPATCHER_CONTEXT *__unw_seh_get_disp_ctx(unw_cursor_t *cursor); static void __unw_seh_set_disp_ctx(unw_cursor_t *cursor, @@ -108,10 +96,10 @@ _GCC_specific_handler(PEXCEPTION_RECORD ms_exc, PVOID frame, PCONTEXT ms_ctx, } } else { // Foreign exception. - exc = (_Unwind_Exception *)malloc(sizeof(_Unwind_Exception)); - exc->exception_class = kSEHExceptionClass; - exc->exception_cleanup = seh_exc_cleanup; - memset(exc->private_, 0, sizeof(exc->private_)); + // We can't interact with them (we don't know the original target frame + // that we should pass on to RtlUnwindEx in _Unwind_Resume), so just + // pass without calling our destructors here. + return ExceptionContinueSearch; } if (!ctx) { __unw_init_seh(&cursor, disp->ContextRecord); diff --git a/lib/libunwind/src/Unwind-sjlj.c b/lib/libunwind/src/Unwind-sjlj.c index 85a4cc3c69..fd2a95b74c 100644 --- a/lib/libunwind/src/Unwind-sjlj.c +++ b/lib/libunwind/src/Unwind-sjlj.c @@ -32,11 +32,23 @@ struct _Unwind_FunctionContext { // next function in stack of handlers struct _Unwind_FunctionContext *prev; +#if defined(__ve__) + // VE requires to store 64 bit pointers in the buffer for SjLj execption. + // We expand the size of values defined here. This size must be matched + // to the size returned by TargetMachine::getSjLjDataSize(). + + // set by calling function before registering to be the landing pad + uint64_t resumeLocation; + + // set by personality handler to be parameters passed to landing pad function + uint64_t resumeParameters[4]; +#else // set by calling function before registering to be the landing pad uint32_t resumeLocation; // set by personality handler to be parameters passed to landing pad function uint32_t resumeParameters[4]; +#endif // set by calling function before registering _Unwind_Personality_Fn personality; // arm offset=24 diff --git a/lib/libunwind/src/UnwindCursor.hpp b/lib/libunwind/src/UnwindCursor.hpp index f346c720d2..e537ed84dd 100644 --- a/lib/libunwind/src/UnwindCursor.hpp +++ b/lib/libunwind/src/UnwindCursor.hpp @@ -81,6 +81,7 @@ template <typename A> class _LIBUNWIND_HIDDEN DwarfFDECache { typedef typename A::pint_t pint_t; public: + static constexpr pint_t kSearchAll = static_cast<pint_t>(-1); static pint_t findFDE(pint_t mh, pint_t pc); static void add(pint_t mh, pint_t ip_start, pint_t ip_end, pint_t fde); static void removeAllIn(pint_t mh); @@ -138,7 +139,7 @@ typename A::pint_t DwarfFDECache<A>::findFDE(pint_t mh, pint_t pc) { pint_t result = 0; _LIBUNWIND_LOG_IF_FALSE(_lock.lock_shared()); for (entry *p = _buffer; p < _bufferUsed; ++p) { - if ((mh == p->mh) || (mh == 0)) { + if ((mh == p->mh) || (mh == kSearchAll)) { if ((p->ip_start <= pc) && (pc < p->ip_end)) { result = p->fde; break; @@ -530,6 +531,8 @@ UnwindCursor<A, R>::UnwindCursor(unw_context_t *context, A &as) : _addressSpace(as), _unwindInfoMissing(false) { static_assert((check_fit<UnwindCursor<A, R>, unw_cursor_t>::does_fit), "UnwindCursor<> does not fit in unw_cursor_t"); + static_assert((alignof(UnwindCursor<A, R>) <= alignof(unw_cursor_t)), + "UnwindCursor<> requires more alignment than unw_cursor_t"); memset(&_info, 0, sizeof(_info)); memset(&_histTable, 0, sizeof(_histTable)); _dispContext.ContextRecord = &_msContext; @@ -922,7 +925,29 @@ private: } #endif +#if defined(_LIBUNWIND_TARGET_LINUX) && defined(_LIBUNWIND_TARGET_AARCH64) + bool setInfoForSigReturn() { + R dummy; + return setInfoForSigReturn(dummy); + } + int stepThroughSigReturn() { + R dummy; + return stepThroughSigReturn(dummy); + } + bool setInfoForSigReturn(Registers_arm64 &); + int stepThroughSigReturn(Registers_arm64 &); + template <typename Registers> bool setInfoForSigReturn(Registers &) { + return false; + } + template <typename Registers> int stepThroughSigReturn(Registers &) { + return UNW_STEP_END; + } +#endif + #if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) + bool getInfoFromFdeCie(const typename CFI_Parser<A>::FDE_Info &fdeInfo, + const typename CFI_Parser<A>::CIE_Info &cieInfo, + pint_t pc, uintptr_t dso_base); bool getInfoFromDwarfSection(pint_t pc, const UnwindInfoSections §s, uint32_t fdeSectionOffsetHint=0); int stepWithDwarfFDE() { @@ -1173,6 +1198,9 @@ private: unw_proc_info_t _info; bool _unwindInfoMissing; bool _isSignalFrame; +#if defined(_LIBUNWIND_TARGET_LINUX) && defined(_LIBUNWIND_TARGET_AARCH64) + bool _isSigReturn = false; +#endif }; @@ -1182,6 +1210,8 @@ UnwindCursor<A, R>::UnwindCursor(unw_context_t *context, A &as) _isSignalFrame(false) { static_assert((check_fit<UnwindCursor<A, R>, unw_cursor_t>::does_fit), "UnwindCursor<> does not fit in unw_cursor_t"); + static_assert((alignof(UnwindCursor<A, R>) <= alignof(unw_cursor_t)), + "UnwindCursor<> requires more alignment than unw_cursor_t"); memset(&_info, 0, sizeof(_info)); } @@ -1473,6 +1503,32 @@ bool UnwindCursor<A, R>::getInfoFromEHABISection( #if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) template <typename A, typename R> +bool UnwindCursor<A, R>::getInfoFromFdeCie( + const typename CFI_Parser<A>::FDE_Info &fdeInfo, + const typename CFI_Parser<A>::CIE_Info &cieInfo, pint_t pc, + uintptr_t dso_base) { + typename CFI_Parser<A>::PrologInfo prolog; + if (CFI_Parser<A>::parseFDEInstructions(_addressSpace, fdeInfo, cieInfo, pc, + R::getArch(), &prolog)) { + // Save off parsed FDE info + _info.start_ip = fdeInfo.pcStart; + _info.end_ip = fdeInfo.pcEnd; + _info.lsda = fdeInfo.lsda; + _info.handler = cieInfo.personality; + // Some frameless functions need SP altered when resuming in function, so + // propagate spExtraArgSize. + _info.gp = prolog.spExtraArgSize; + _info.flags = 0; + _info.format = dwarfEncoding(); + _info.unwind_info = fdeInfo.fdeStart; + _info.unwind_info_size = static_cast<uint32_t>(fdeInfo.fdeLength); + _info.extra = static_cast<unw_word_t>(dso_base); + return true; + } + return false; +} + +template <typename A, typename R> bool UnwindCursor<A, R>::getInfoFromDwarfSection(pint_t pc, const UnwindInfoSections §s, uint32_t fdeSectionOffsetHint) { @@ -1483,7 +1539,7 @@ bool UnwindCursor<A, R>::getInfoFromDwarfSection(pint_t pc, // If compact encoding table gave offset into dwarf section, go directly there if (fdeSectionOffsetHint != 0) { foundFDE = CFI_Parser<A>::findFDE(_addressSpace, pc, sects.dwarf_section, - (uint32_t)sects.dwarf_section_length, + sects.dwarf_section_length, sects.dwarf_section + fdeSectionOffsetHint, &fdeInfo, &cieInfo); } @@ -1500,7 +1556,7 @@ bool UnwindCursor<A, R>::getInfoFromDwarfSection(pint_t pc, if (cachedFDE != 0) { foundFDE = CFI_Parser<A>::findFDE(_addressSpace, pc, sects.dwarf_section, - (uint32_t)sects.dwarf_section_length, + sects.dwarf_section_length, cachedFDE, &fdeInfo, &cieInfo); foundInCache = foundFDE; } @@ -1508,25 +1564,11 @@ bool UnwindCursor<A, R>::getInfoFromDwarfSection(pint_t pc, if (!foundFDE) { // Still not found, do full scan of __eh_frame section. foundFDE = CFI_Parser<A>::findFDE(_addressSpace, pc, sects.dwarf_section, - (uint32_t)sects.dwarf_section_length, 0, + sects.dwarf_section_length, 0, &fdeInfo, &cieInfo); } if (foundFDE) { - typename CFI_Parser<A>::PrologInfo prolog; - if (CFI_Parser<A>::parseFDEInstructions(_addressSpace, fdeInfo, cieInfo, pc, - R::getArch(), &prolog)) { - // Save off parsed FDE info - _info.start_ip = fdeInfo.pcStart; - _info.end_ip = fdeInfo.pcEnd; - _info.lsda = fdeInfo.lsda; - _info.handler = cieInfo.personality; - _info.gp = prolog.spExtraArgSize; - _info.flags = 0; - _info.format = dwarfEncoding(); - _info.unwind_info = fdeInfo.fdeStart; - _info.unwind_info_size = (uint32_t)fdeInfo.fdeLength; - _info.extra = (unw_word_t) sects.dso_base; - + if (getInfoFromFdeCie(fdeInfo, cieInfo, pc, sects.dso_base)) { // Add to cache (to make next lookup faster) if we had no hint // and there was no index. if (!foundInCache && (fdeSectionOffsetHint == 0)) { @@ -1759,12 +1801,12 @@ bool UnwindCursor<A, R>::getInfoFromCompactEncodingSection(pint_t pc, } } - // extact personality routine, if encoding says function has one + // extract personality routine, if encoding says function has one uint32_t personalityIndex = (encoding & UNWIND_PERSONALITY_MASK) >> (__builtin_ctz(UNWIND_PERSONALITY_MASK)); if (personalityIndex != 0) { --personalityIndex; // change 1-based to zero-based index - if (personalityIndex > sectionHeader.personalityArrayCount()) { + if (personalityIndex >= sectionHeader.personalityArrayCount()) { _LIBUNWIND_DEBUG_LOG("found encoding 0x%08X with personality index %d, " "but personality table has only %d entries", encoding, personalityIndex, @@ -1853,7 +1895,11 @@ bool UnwindCursor<A, R>::getInfoFromSEH(pint_t pc) { template <typename A, typename R> void UnwindCursor<A, R>::setInfoBasedOnIPRegister(bool isReturnAddress) { - pint_t pc = (pint_t)this->getReg(UNW_REG_IP); +#if defined(_LIBUNWIND_TARGET_LINUX) && defined(_LIBUNWIND_TARGET_AARCH64) + _isSigReturn = false; +#endif + + pint_t pc = static_cast<pint_t>(this->getReg(UNW_REG_IP)); #if defined(_LIBUNWIND_ARM_EHABI) // Remove the thumb bit so the IP represents the actual instruction address. // This matches the behaviour of _Unwind_GetIP on arm. @@ -1926,68 +1972,102 @@ void UnwindCursor<A, R>::setInfoBasedOnIPRegister(bool isReturnAddress) { #if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) // There is no static unwind info for this pc. Look to see if an FDE was // dynamically registered for it. - pint_t cachedFDE = DwarfFDECache<A>::findFDE(0, pc); + pint_t cachedFDE = DwarfFDECache<A>::findFDE(DwarfFDECache<A>::kSearchAll, + pc); if (cachedFDE != 0) { - CFI_Parser<LocalAddressSpace>::FDE_Info fdeInfo; - CFI_Parser<LocalAddressSpace>::CIE_Info cieInfo; - const char *msg = CFI_Parser<A>::decodeFDE(_addressSpace, - cachedFDE, &fdeInfo, &cieInfo); - if (msg == NULL) { - typename CFI_Parser<A>::PrologInfo prolog; - if (CFI_Parser<A>::parseFDEInstructions(_addressSpace, fdeInfo, cieInfo, - pc, R::getArch(), &prolog)) { - // save off parsed FDE info - _info.start_ip = fdeInfo.pcStart; - _info.end_ip = fdeInfo.pcEnd; - _info.lsda = fdeInfo.lsda; - _info.handler = cieInfo.personality; - _info.gp = prolog.spExtraArgSize; - // Some frameless functions need SP - // altered when resuming in function. - _info.flags = 0; - _info.format = dwarfEncoding(); - _info.unwind_info = fdeInfo.fdeStart; - _info.unwind_info_size = (uint32_t)fdeInfo.fdeLength; - _info.extra = 0; + typename CFI_Parser<A>::FDE_Info fdeInfo; + typename CFI_Parser<A>::CIE_Info cieInfo; + if (!CFI_Parser<A>::decodeFDE(_addressSpace, cachedFDE, &fdeInfo, &cieInfo)) + if (getInfoFromFdeCie(fdeInfo, cieInfo, pc, 0)) return; - } - } } // Lastly, ask AddressSpace object about platform specific ways to locate // other FDEs. pint_t fde; if (_addressSpace.findOtherFDE(pc, fde)) { - CFI_Parser<LocalAddressSpace>::FDE_Info fdeInfo; - CFI_Parser<LocalAddressSpace>::CIE_Info cieInfo; + typename CFI_Parser<A>::FDE_Info fdeInfo; + typename CFI_Parser<A>::CIE_Info cieInfo; if (!CFI_Parser<A>::decodeFDE(_addressSpace, fde, &fdeInfo, &cieInfo)) { // Double check this FDE is for a function that includes the pc. - if ((fdeInfo.pcStart <= pc) && (pc < fdeInfo.pcEnd)) { - typename CFI_Parser<A>::PrologInfo prolog; - if (CFI_Parser<A>::parseFDEInstructions(_addressSpace, fdeInfo, cieInfo, - pc, R::getArch(), &prolog)) { - // save off parsed FDE info - _info.start_ip = fdeInfo.pcStart; - _info.end_ip = fdeInfo.pcEnd; - _info.lsda = fdeInfo.lsda; - _info.handler = cieInfo.personality; - _info.gp = prolog.spExtraArgSize; - _info.flags = 0; - _info.format = dwarfEncoding(); - _info.unwind_info = fdeInfo.fdeStart; - _info.unwind_info_size = (uint32_t)fdeInfo.fdeLength; - _info.extra = 0; + if ((fdeInfo.pcStart <= pc) && (pc < fdeInfo.pcEnd)) + if (getInfoFromFdeCie(fdeInfo, cieInfo, pc, 0)) return; - } - } } } #endif // #if defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) +#if defined(_LIBUNWIND_TARGET_LINUX) && defined(_LIBUNWIND_TARGET_AARCH64) + if (setInfoForSigReturn()) + return; +#endif + // no unwind info, flag that we can't reliably unwind _unwindInfoMissing = true; } +#if defined(_LIBUNWIND_TARGET_LINUX) && defined(_LIBUNWIND_TARGET_AARCH64) +template <typename A, typename R> +bool UnwindCursor<A, R>::setInfoForSigReturn(Registers_arm64 &) { + // Look for the sigreturn trampoline. The trampoline's body is two + // specific instructions (see below). Typically the trampoline comes from the + // vDSO[1] (i.e. the __kernel_rt_sigreturn function). A libc might provide its + // own restorer function, though, or user-mode QEMU might write a trampoline + // onto the stack. + // + // This special code path is a fallback that is only used if the trampoline + // lacks proper (e.g. DWARF) unwind info. On AArch64, a new DWARF register + // constant for the PC needs to be defined before DWARF can handle a signal + // trampoline. This code may segfault if the target PC is unreadable, e.g.: + // - The PC points at a function compiled without unwind info, and which is + // part of an execute-only mapping (e.g. using -Wl,--execute-only). + // - The PC is invalid and happens to point to unreadable or unmapped memory. + // + // [1] https://github.com/torvalds/linux/blob/master/arch/arm64/kernel/vdso/sigreturn.S + const pint_t pc = static_cast<pint_t>(this->getReg(UNW_REG_IP)); + // Look for instructions: mov x8, #0x8b; svc #0x0 + if (_addressSpace.get32(pc) == 0xd2801168 && + _addressSpace.get32(pc + 4) == 0xd4000001) { + _info = {}; + _isSigReturn = true; + return true; + } + return false; +} + +template <typename A, typename R> +int UnwindCursor<A, R>::stepThroughSigReturn(Registers_arm64 &) { + // In the signal trampoline frame, sp points to an rt_sigframe[1], which is: + // - 128-byte siginfo struct + // - ucontext struct: + // - 8-byte long (uc_flags) + // - 8-byte pointer (uc_link) + // - 24-byte stack_t + // - 128-byte signal set + // - 8 bytes of padding because sigcontext has 16-byte alignment + // - sigcontext/mcontext_t + // [1] https://github.com/torvalds/linux/blob/master/arch/arm64/kernel/signal.c + const pint_t kOffsetSpToSigcontext = (128 + 8 + 8 + 24 + 128 + 8); // 304 + + // Offsets from sigcontext to each register. + const pint_t kOffsetGprs = 8; // offset to "__u64 regs[31]" field + const pint_t kOffsetSp = 256; // offset to "__u64 sp" field + const pint_t kOffsetPc = 264; // offset to "__u64 pc" field + + pint_t sigctx = _registers.getSP() + kOffsetSpToSigcontext; + + for (int i = 0; i <= 30; ++i) { + uint64_t value = _addressSpace.get64(sigctx + kOffsetGprs + + static_cast<pint_t>(i * 8)); + _registers.setRegister(UNW_ARM64_X0 + i, value); + } + _registers.setSP(_addressSpace.get64(sigctx + kOffsetSp)); + _registers.setIP(_addressSpace.get64(sigctx + kOffsetPc)); + _isSignalFrame = true; + return UNW_STEP_SUCCESS; +} +#endif // defined(_LIBUNWIND_TARGET_LINUX) && defined(_LIBUNWIND_TARGET_AARCH64) + template <typename A, typename R> int UnwindCursor<A, R>::step() { // Bottom of stack is defined is when unwind info cannot be found. @@ -1996,20 +2076,27 @@ int UnwindCursor<A, R>::step() { // Use unwinding info to modify register set as if function returned. int result; +#if defined(_LIBUNWIND_TARGET_LINUX) && defined(_LIBUNWIND_TARGET_AARCH64) + if (_isSigReturn) { + result = this->stepThroughSigReturn(); + } else +#endif + { #if defined(_LIBUNWIND_SUPPORT_COMPACT_UNWIND) - result = this->stepWithCompactEncoding(); + result = this->stepWithCompactEncoding(); #elif defined(_LIBUNWIND_SUPPORT_SEH_UNWIND) - result = this->stepWithSEHData(); + result = this->stepWithSEHData(); #elif defined(_LIBUNWIND_SUPPORT_DWARF_UNWIND) - result = this->stepWithDwarfFDE(); + result = this->stepWithDwarfFDE(); #elif defined(_LIBUNWIND_ARM_EHABI) - result = this->stepWithEHABI(); + result = this->stepWithEHABI(); #else #error Need _LIBUNWIND_SUPPORT_COMPACT_UNWIND or \ _LIBUNWIND_SUPPORT_SEH_UNWIND or \ _LIBUNWIND_SUPPORT_DWARF_UNWIND or \ _LIBUNWIND_ARM_EHABI #endif + } // update info based on new PC if (result == UNW_STEP_SUCCESS) { diff --git a/lib/libunwind/src/UnwindLevel1.c b/lib/libunwind/src/UnwindLevel1.c index 3e75b5f13c..68e5e48b8c 100644 --- a/lib/libunwind/src/UnwindLevel1.c +++ b/lib/libunwind/src/UnwindLevel1.c @@ -39,8 +39,7 @@ unwind_phase1(unw_context_t *uc, unw_cursor_t *cursor, _Unwind_Exception *except __unw_init_local(cursor, uc); // Walk each frame looking for a place to stop. - bool handlerNotFound = true; - while (handlerNotFound) { + while (true) { // Ask libunwind to get next frame (skip over first which is // _Unwind_RaiseException). int stepResult = __unw_step(cursor); @@ -102,7 +101,6 @@ unwind_phase1(unw_context_t *uc, unw_cursor_t *cursor, _Unwind_Exception *except case _URC_HANDLER_FOUND: // found a catch clause or locals that need destructing in this frame // stop search and remember stack pointer at the frame - handlerNotFound = false; __unw_get_reg(cursor, UNW_REG_SP, &sp); exception_object->private_2 = (uintptr_t)sp; _LIBUNWIND_TRACE_UNWINDING( diff --git a/lib/libunwind/src/UnwindRegistersRestore.S b/lib/libunwind/src/UnwindRegistersRestore.S index 5d54432152..289afe98b0 100644 --- a/lib/libunwind/src/UnwindRegistersRestore.S +++ b/lib/libunwind/src/UnwindRegistersRestore.S @@ -13,14 +13,10 @@ #if !defined(__USING_SJLJ_EXCEPTIONS__) #if defined(__i386__) -DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind13Registers_x866jumptoEv) +DEFINE_LIBUNWIND_FUNCTION(__libunwind_Registers_x86_jumpto) # -# void libunwind::Registers_x86::jumpto() +# extern "C" void __libunwind_Registers_x86_jumpto(Registers_x86 *); # -#if defined(_WIN32) -# On windows, the 'this' pointer is passed in ecx instead of on the stack - movl %ecx, %eax -#else # On entry: # + + # +-----------------------+ @@ -30,7 +26,6 @@ DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind13Registers_x866jumptoEv) # +-----------------------+ <-- SP # + + movl 4(%esp), %eax -#endif # set up eax and ret on new stack location movl 28(%eax), %edx # edx holds new stack pointer subl $8,%edx @@ -60,9 +55,9 @@ DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind13Registers_x866jumptoEv) #elif defined(__x86_64__) -DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind16Registers_x86_646jumptoEv) +DEFINE_LIBUNWIND_FUNCTION(__libunwind_Registers_x86_64_jumpto) # -# void libunwind::Registers_x86_64::jumpto() +# extern "C" void __libunwind_Registers_x86_64_jumpto(Registers_x86_64 *); # #if defined(_WIN64) # On entry, thread_state pointer is in rcx; move it into rdi @@ -175,7 +170,7 @@ DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind15Registers_ppc646jumptoEv) PPC64_LR(30) PPC64_LR(31) -#ifdef PPC64_HAS_VMX +#if defined(__VSX__) // restore VS registers // (note that this also restores floating point registers and V registers, @@ -317,6 +312,7 @@ PPC64_CLVS_BOTTOM(n) PPC64_LF(30) PPC64_LF(31) +#if defined(__ALTIVEC__) // restore vector registers if any are in use ld %r5, PPC64_OFFS_VRSAVE(%r3) // test VRsave cmpwi %r5, 0 @@ -378,6 +374,7 @@ PPC64_CLV_UNALIGNED_BOTTOM(n) PPC64_CLV_UNALIGNEDh(31) #endif +#endif Lnovec: ld %r0, PPC64_OFFS_CR(%r3) @@ -436,6 +433,7 @@ DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind13Registers_ppc6jumptoEv) lwz %r30,128(%r3) lwz %r31,132(%r3) +#ifndef __NO_FPRS__ // restore float registers lfd %f0, 160(%r3) lfd %f1, 168(%r3) @@ -469,7 +467,9 @@ DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind13Registers_ppc6jumptoEv) lfd %f29,392(%r3) lfd %f30,400(%r3) lfd %f31,408(%r3) +#endif +#if defined(__ALTIVEC__) // restore vector registers if any are in use lwz %r5, 156(%r3) // test VRsave cmpwi %r5, 0 @@ -542,6 +542,7 @@ DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind13Registers_ppc6jumptoEv) LOAD_VECTOR_UNALIGNEDh(29) LOAD_VECTOR_UNALIGNEDh(30) LOAD_VECTOR_UNALIGNEDh(31) +#endif Lnovec: lwz %r0, 136(%r3) // __cr @@ -560,13 +561,13 @@ Lnovec: #elif defined(__aarch64__) // -// void libunwind::Registers_arm64::jumpto() +// extern "C" void __libunwind_Registers_arm64_jumpto(Registers_arm64 *); // // On entry: // thread_state pointer is in x0 // .p2align 2 -DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind15Registers_arm646jumptoEv) +DEFINE_LIBUNWIND_FUNCTION(__libunwind_Registers_arm64_jumpto) // skip restore of x0,x1 for now ldp x2, x3, [x0, #0x010] ldp x4, x5, [x0, #0x020] diff --git a/lib/libunwind/src/UnwindRegistersSave.S b/lib/libunwind/src/UnwindRegistersSave.S index 51bb9b0688..94fc836545 100644 --- a/lib/libunwind/src/UnwindRegistersSave.S +++ b/lib/libunwind/src/UnwindRegistersSave.S @@ -384,7 +384,7 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext) mfvrsave %r0 std %r0, PPC64_OFFS_VRSAVE(%r3) -#ifdef PPC64_HAS_VMX +#if defined(__VSX__) // save VS registers // (note that this also saves floating point registers and V registers, // because part of VS is mapped to these registers) @@ -501,6 +501,7 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext) PPC64_STF(30) PPC64_STF(31) +#if defined(__ALTIVEC__) // save vector registers // Use 16-bytes below the stack pointer as an @@ -549,6 +550,7 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext) PPC64_STV_UNALIGNED(31) #endif +#endif li %r3, 0 // return UNW_ESUCCESS blr @@ -608,6 +610,7 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext) mfctr %r0 stw %r0, 148(%r3) +#if !defined(__NO_FPRS__) // save float registers stfd %f0, 160(%r3) stfd %f1, 168(%r3) @@ -641,8 +644,9 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext) stfd %f29,392(%r3) stfd %f30,400(%r3) stfd %f31,408(%r3) +#endif - +#if defined(__ALTIVEC__) // save vector registers subi %r4, %r1, 16 @@ -692,6 +696,7 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext) SAVE_VECTOR_UNALIGNED(%v29, 424+0x1D0) SAVE_VECTOR_UNALIGNED(%v30, 424+0x1E0) SAVE_VECTOR_UNALIGNED(%v31, 424+0x1F0) +#endif li %r3, 0 // return UNW_ESUCCESS blr diff --git a/lib/libunwind/src/Unwind_AppleExtras.cpp b/lib/libunwind/src/Unwind_AppleExtras.cpp index 536303993e..e3d41ca2b4 100644 --- a/lib/libunwind/src/Unwind_AppleExtras.cpp +++ b/lib/libunwind/src/Unwind_AppleExtras.cpp @@ -8,35 +8,6 @@ //===----------------------------------------------------------------------===// #include "config.h" -#include "AddressSpace.hpp" -#include "DwarfParser.hpp" - - -// private keymgr stuff -#define KEYMGR_GCC3_DW2_OBJ_LIST 302 -extern "C" { - extern void _keymgr_set_and_unlock_processwide_ptr(int key, void *ptr); - extern void *_keymgr_get_and_lock_processwide_ptr(int key); -} - -// undocumented libgcc "struct object" -struct libgcc_object { - void *start; - void *unused1; - void *unused2; - void *fde; - unsigned long encoding; - void *fde_end; - libgcc_object *next; -}; - -// undocumented libgcc "struct km_object_info" referenced by -// KEYMGR_GCC3_DW2_OBJ_LIST -struct libgcc_object_info { - libgcc_object *seen_objects; - libgcc_object *unseen_objects; - unsigned spare[2]; -}; // static linker symbols to prevent wrong two level namespace for _Unwind symbols @@ -140,44 +111,3 @@ NOT_HERE_BEFORE_5_0(_Unwind_SjLj_Resume_or_Rethrow) NOT_HERE_BEFORE_5_0(_Unwind_SjLj_Unregister) #endif // defined(_LIBUNWIND_BUILD_SJLJ_APIS) - - -namespace libunwind { - -_LIBUNWIND_HIDDEN -bool checkKeyMgrRegisteredFDEs(uintptr_t pc, void *&fde) { -#if __MAC_OS_X_VERSION_MIN_REQUIRED - // lastly check for old style keymgr registration of dynamically generated - // FDEs acquire exclusive access to libgcc_object_info - libgcc_object_info *head = (libgcc_object_info *) - _keymgr_get_and_lock_processwide_ptr(KEYMGR_GCC3_DW2_OBJ_LIST); - if (head != NULL) { - // look at each FDE in keymgr - for (libgcc_object *ob = head->unseen_objects; ob != NULL; ob = ob->next) { - CFI_Parser<LocalAddressSpace>::FDE_Info fdeInfo; - CFI_Parser<LocalAddressSpace>::CIE_Info cieInfo; - const char *msg = CFI_Parser<LocalAddressSpace>::decodeFDE( - LocalAddressSpace::sThisAddressSpace, - (uintptr_t)ob->fde, &fdeInfo, &cieInfo); - if (msg == NULL) { - // Check if this FDE is for a function that includes the pc - if ((fdeInfo.pcStart <= pc) && (pc < fdeInfo.pcEnd)) { - fde = (void*)fdeInfo.pcStart; - _keymgr_set_and_unlock_processwide_ptr(KEYMGR_GCC3_DW2_OBJ_LIST, - head); - return true; - } - } - } - } - // release libgcc_object_info - _keymgr_set_and_unlock_processwide_ptr(KEYMGR_GCC3_DW2_OBJ_LIST, head); -#else - (void)pc; - (void)fde; -#endif - return false; -} - -} - diff --git a/lib/libunwind/src/assembly.h b/lib/libunwind/src/assembly.h index 4cf179e13e..f2f7c84830 100644 --- a/lib/libunwind/src/assembly.h +++ b/lib/libunwind/src/assembly.h @@ -25,9 +25,6 @@ #define PPC64_OFFS_VRSAVE 304 #define PPC64_OFFS_FP 312 #define PPC64_OFFS_V 824 -#ifdef _ARCH_PWR8 -#define PPC64_HAS_VMX -#endif #elif defined(__APPLE__) && defined(__aarch64__) #define SEPARATOR %% #else @@ -48,6 +45,24 @@ #define PPC64_OPD2 #endif +#if defined(__ARM_FEATURE_BTI_DEFAULT) + .pushsection ".note.gnu.property", "a" SEPARATOR \ + .balign 8 SEPARATOR \ + .long 4 SEPARATOR \ + .long 0x10 SEPARATOR \ + .long 0x5 SEPARATOR \ + .asciz "GNU" SEPARATOR \ + .long 0xc0000000 SEPARATOR /* GNU_PROPERTY_AARCH64_FEATURE_1_AND */ \ + .long 4 SEPARATOR \ + .long 3 SEPARATOR /* GNU_PROPERTY_AARCH64_FEATURE_1_BTI AND */ \ + /* GNU_PROPERTY_AARCH64_FEATURE_1_PAC */ \ + .long 0 SEPARATOR \ + .popsection SEPARATOR +#define AARCH64_BTI bti c +#else +#define AARCH64_BTI +#endif + #define GLUE2(a, b) a ## b #define GLUE(a, b) GLUE2(a, b) #define SYMBOL_NAME(name) GLUE(__USER_LABEL_PREFIX__, name) @@ -144,7 +159,8 @@ SYMBOL_IS_FUNC(SYMBOL_NAME(name)) SEPARATOR \ PPC64_OPD1 \ SYMBOL_NAME(name): \ - PPC64_OPD2 + PPC64_OPD2 \ + AARCH64_BTI #if defined(__arm__) #if !defined(__ARM_ARCH) diff --git a/lib/libunwind/src/config.h b/lib/libunwind/src/config.h index 842fd829af..9efed05405 100644 --- a/lib/libunwind/src/config.h +++ b/lib/libunwind/src/config.h @@ -18,23 +18,15 @@ #include <stdint.h> #include <stdlib.h> -// Define static_assert() unless already defined by compiler. -#ifndef __has_feature - #define __has_feature(__x) 0 -#endif -#if !(__has_feature(cxx_static_assert)) && !defined(static_assert) - #define static_assert(__b, __m) \ - extern int compile_time_assert_failed[ ( __b ) ? 1 : -1 ] \ - __attribute__( ( unused ) ); -#endif +#include <__libunwind_config.h> // Platform specific configuration defines. #ifdef __APPLE__ #if defined(FOR_DYLD) - #define _LIBUNWIND_SUPPORT_COMPACT_UNWIND + #define _LIBUNWIND_SUPPORT_COMPACT_UNWIND 1 #else - #define _LIBUNWIND_SUPPORT_COMPACT_UNWIND - #define _LIBUNWIND_SUPPORT_DWARF_UNWIND 1 + #define _LIBUNWIND_SUPPORT_COMPACT_UNWIND 1 + #define _LIBUNWIND_SUPPORT_DWARF_UNWIND 1 #endif #elif defined(_WIN32) #ifdef __SEH__ @@ -42,8 +34,19 @@ #else #define _LIBUNWIND_SUPPORT_DWARF_UNWIND 1 #endif +#elif defined(_LIBUNWIND_IS_BAREMETAL) + #if !defined(_LIBUNWIND_ARM_EHABI) + #define _LIBUNWIND_SUPPORT_DWARF_UNWIND 1 + #define _LIBUNWIND_SUPPORT_DWARF_INDEX 1 + #endif +#elif defined(__BIONIC__) && defined(_LIBUNWIND_ARM_EHABI) + // For ARM EHABI, Bionic didn't implement dl_iterate_phdr until API 21. After + // API 21, dl_iterate_phdr exists, but dl_unwind_find_exidx is much faster. + #define _LIBUNWIND_USE_DL_UNWIND_FIND_EXIDX 1 #else - #if defined(__ARM_DWARF_EH__) || !defined(__arm__) + // Assume an ELF system with a dl_iterate_phdr function. + #define _LIBUNWIND_USE_DL_ITERATE_PHDR 1 + #if !defined(_LIBUNWIND_ARM_EHABI) #define _LIBUNWIND_SUPPORT_DWARF_UNWIND 1 #define _LIBUNWIND_SUPPORT_DWARF_INDEX 1 #endif @@ -91,6 +94,8 @@ #error Unsupported target #endif +// Apple/armv7k defaults to DWARF/Compact unwinding, but its libunwind also +// needs to include the SJLJ APIs. #if (defined(__APPLE__) && defined(__arm__)) || defined(__USING_SJLJ_EXCEPTIONS__) #define _LIBUNWIND_BUILD_SJLJ_APIS #endif @@ -111,8 +116,27 @@ #endif #endif -#if defined(__powerpc64__) && defined(_ARCH_PWR8) -#define PPC64_HAS_VMX +#ifndef _LIBUNWIND_REMEMBER_HEAP_ALLOC +#if defined(_LIBUNWIND_REMEMBER_STACK_ALLOC) || defined(__APPLE__) || \ + defined(__linux__) || defined(__ANDROID__) || defined(__MINGW32__) || \ + defined(_LIBUNWIND_IS_BAREMETAL) +#define _LIBUNWIND_REMEMBER_ALLOC(_size) alloca(_size) +#define _LIBUNWIND_REMEMBER_FREE(_ptr) \ + do { \ + } while (0) +#elif defined(_WIN32) +#define _LIBUNWIND_REMEMBER_ALLOC(_size) _malloca(_size) +#define _LIBUNWIND_REMEMBER_FREE(_ptr) _freea(_ptr) +#define _LIBUNWIND_REMEMBER_CLEANUP_NEEDED +#else +#define _LIBUNWIND_REMEMBER_ALLOC(_size) malloc(_size) +#define _LIBUNWIND_REMEMBER_FREE(_ptr) free(_ptr) +#define _LIBUNWIND_REMEMBER_CLEANUP_NEEDED +#endif +#else /* _LIBUNWIND_REMEMBER_HEAP_ALLOC */ +#define _LIBUNWIND_REMEMBER_ALLOC(_size) malloc(_size) +#define _LIBUNWIND_REMEMBER_FREE(_ptr) free(_ptr) +#define _LIBUNWIND_REMEMBER_CLEANUP_NEEDED #endif #if defined(NDEBUG) && defined(_LIBUNWIND_IS_BAREMETAL) diff --git a/lib/libunwind/src/libunwind.cpp b/lib/libunwind/src/libunwind.cpp index fd079da308..c21461b1f4 100644 --- a/lib/libunwind/src/libunwind.cpp +++ b/lib/libunwind/src/libunwind.cpp @@ -62,6 +62,8 @@ _LIBUNWIND_HIDDEN int __unw_init_local(unw_cursor_t *cursor, # define REGISTER_KIND Registers_sparc #elif defined(__riscv) && __riscv_xlen == 64 # define REGISTER_KIND Registers_riscv +#elif defined(__ve__) +# define REGISTER_KIND Registers_ve #else # error Architecture not supported #endif diff --git a/lib/std/target.zig b/lib/std/target.zig index 7e05f35932..0c94b9407e 100644 --- a/lib/std/target.zig +++ b/lib/std/target.zig @@ -36,11 +36,11 @@ pub const Target = struct { openbsd, solaris, windows, + zos, haiku, minix, rtems, nacl, - cnk, aix, cuda, nvcl, @@ -232,11 +232,11 @@ pub const Target = struct { .kfreebsd, .lv2, .solaris, + .zos, .haiku, .minix, .rtems, .nacl, - .cnk, .aix, .cuda, .nvcl, @@ -391,10 +391,10 @@ pub const Target = struct { .kfreebsd, .lv2, .solaris, + .zos, .minix, .rtems, .nacl, - .cnk, .aix, .cuda, .nvcl, @@ -430,8 +430,8 @@ pub const Target = struct { pub const powerpc = @import("target/powerpc.zig"); pub const riscv = @import("target/riscv.zig"); pub const sparc = @import("target/sparc.zig"); - pub const spirv = @import("target/spirv.zig"); pub const systemz = @import("target/systemz.zig"); + pub const ve = @import("target/ve.zig"); pub const wasm = @import("target/wasm.zig"); pub const x86 = @import("target/x86.zig"); @@ -443,6 +443,7 @@ pub const Target = struct { gnueabi, gnueabihf, gnux32, + gnuilp32, code16, eabi, eabihf, @@ -468,10 +469,10 @@ pub const Target = struct { .dragonfly, .lv2, .solaris, + .zos, .minix, .rtems, .nacl, - .cnk, .aix, .cuda, .nvcl, @@ -592,7 +593,7 @@ pub const Target = struct { pub const Set = struct { ints: [usize_count]usize, - pub const needed_bit_count = 168; + pub const needed_bit_count = 172; pub const byte_count = (needed_bit_count + 7) / 8; pub const usize_count = (byte_count + (@sizeOf(usize) - 1)) / @sizeOf(usize); pub const Index = std.math.Log2Int(std.meta.Int(.unsigned, usize_count * @bitSizeOf(usize))); @@ -714,6 +715,7 @@ pub const Target = struct { avr, bpfel, bpfeb, + csky, hexagon, mips, mipsel, @@ -721,6 +723,7 @@ pub const Target = struct { mips64el, msp430, powerpc, + powerpcle, powerpc64, powerpc64le, r600, @@ -831,7 +834,7 @@ pub const Target = struct { .le32 => ._NONE, .mips => ._MIPS, .mipsel => ._MIPS_RS3_LE, - .powerpc => ._PPC, + .powerpc, .powerpcle => ._PPC, .r600 => ._NONE, .riscv32 => ._RISCV, .sparc => ._SPARC, @@ -870,6 +873,7 @@ pub const Target = struct { .amdgcn => ._NONE, .bpfel => ._BPF, .bpfeb => ._BPF, + .csky => ._NONE, .sparcv9 => ._SPARCV9, .s390x => ._S390, .ve => ._NONE, @@ -890,7 +894,7 @@ pub const Target = struct { .le32 => .Unknown, .mips => .Unknown, .mipsel => .Unknown, - .powerpc => .POWERPC, + .powerpc, .powerpcle => .POWERPC, .r600 => .Unknown, .riscv32 => .RISCV32, .sparc => .Unknown, @@ -929,6 +933,7 @@ pub const Target = struct { .amdgcn => .Unknown, .bpfel => .Unknown, .bpfeb => .Unknown, + .csky => .Unknown, .sparcv9 => .Unknown, .s390x => .Unknown, .ve => .Unknown, @@ -948,6 +953,7 @@ pub const Target = struct { .amdil, .amdil64, .bpfel, + .csky, .hexagon, .hsail, .hsail64, @@ -961,6 +967,7 @@ pub const Target = struct { .nvptx64, .sparcel, .tcele, + .powerpcle, .powerpc64le, .r600, .riscv32, @@ -1011,11 +1018,13 @@ pub const Target = struct { .arc, .arm, .armeb, + .csky, .hexagon, .le32, .mips, .mipsel, .powerpc, + .powerpcle, .r600, .riscv32, .sparc, @@ -1065,17 +1074,14 @@ pub const Target = struct { } } - /// Returns a name that matches the lib/std/target/* directory name. + /// Returns a name that matches the lib/std/target/* source file name. pub fn genericName(arch: Arch) []const u8 { return switch (arch) { .arm, .armeb, .thumb, .thumbeb => "arm", .aarch64, .aarch64_be, .aarch64_32 => "aarch64", - .avr => "avr", .bpfel, .bpfeb => "bpf", - .hexagon => "hexagon", .mips, .mipsel, .mips64, .mips64el => "mips", - .msp430 => "msp430", - .powerpc, .powerpc64, .powerpc64le => "powerpc", + .powerpc, .powerpcle, .powerpc64, .powerpc64le => "powerpc", .amdgcn => "amdgpu", .riscv32, .riscv64 => "riscv", .sparc, .sparcv9, .sparcel => "sparc", @@ -1098,13 +1104,14 @@ pub const Target = struct { .hexagon => &hexagon.all_features, .mips, .mipsel, .mips64, .mips64el => &mips.all_features, .msp430 => &msp430.all_features, - .powerpc, .powerpc64, .powerpc64le => &powerpc.all_features, + .powerpc, .powerpcle, .powerpc64, .powerpc64le => &powerpc.all_features, .amdgcn => &amdgpu.all_features, .riscv32, .riscv64 => &riscv.all_features, .sparc, .sparcv9, .sparcel => &sparc.all_features, .s390x => &systemz.all_features, .i386, .x86_64 => &x86.all_features, .nvptx, .nvptx64 => &nvptx.all_features, + .ve => &ve.all_features, .wasm32, .wasm64 => &wasm.all_features, else => &[0]Cpu.Feature{}, @@ -1121,13 +1128,14 @@ pub const Target = struct { .hexagon => comptime allCpusFromDecls(hexagon.cpu), .mips, .mipsel, .mips64, .mips64el => comptime allCpusFromDecls(mips.cpu), .msp430 => comptime allCpusFromDecls(msp430.cpu), - .powerpc, .powerpc64, .powerpc64le => comptime allCpusFromDecls(powerpc.cpu), + .powerpc, .powerpcle, .powerpc64, .powerpc64le => comptime allCpusFromDecls(powerpc.cpu), .amdgcn => comptime allCpusFromDecls(amdgpu.cpu), .riscv32, .riscv64 => comptime allCpusFromDecls(riscv.cpu), .sparc, .sparcv9, .sparcel => comptime allCpusFromDecls(sparc.cpu), .s390x => comptime allCpusFromDecls(systemz.cpu), .i386, .x86_64 => comptime allCpusFromDecls(x86.cpu), .nvptx, .nvptx64 => comptime allCpusFromDecls(nvptx.cpu), + .ve => comptime allCpusFromDecls(ve.cpu), .wasm32, .wasm64 => comptime allCpusFromDecls(wasm.cpu), else => &[0]*const Model{}, @@ -1177,17 +1185,19 @@ pub const Target = struct { .mips64, .mips64el => &mips.cpu.mips64, .msp430 => &msp430.cpu.generic, .powerpc => &powerpc.cpu.ppc32, + .powerpcle => &powerpc.cpu.ppc32, .powerpc64 => &powerpc.cpu.ppc64, .powerpc64le => &powerpc.cpu.ppc64le, .amdgcn => &amdgpu.cpu.generic, .riscv32 => &riscv.cpu.generic_rv32, .riscv64 => &riscv.cpu.generic_rv64, - .sparc, .sparcel => &sparc.cpu.v8, + .sparc, .sparcel => &sparc.cpu.generic, .sparcv9 => &sparc.cpu.v9, .s390x => &systemz.cpu.generic, .i386 => &x86.cpu._i386, .x86_64 => &x86.cpu.x86_64, .nvptx, .nvptx64 => &nvptx.cpu.sm_20, + .ve => &ve.cpu.generic, .wasm32, .wasm64 => &wasm.cpu.generic, else => &S.generic_model, @@ -1201,6 +1211,7 @@ pub const Target = struct { .riscv64 => &riscv.cpu.baseline_rv64, .i386 => &x86.cpu.pentium4, .nvptx, .nvptx64 => &nvptx.cpu.sm_20, + .sparc, .sparcel => &sparc.cpu.v8, else => generic(arch), }; @@ -1490,7 +1501,7 @@ pub const Target = struct { return print(&result, "/lib{s}/{s}", .{ lib_suffix, loader }); }, - .powerpc => return copy(&result, "/lib/ld.so.1"), + .powerpc, .powerpcle => return copy(&result, "/lib/ld.so.1"), .powerpc64, .powerpc64le => return copy(&result, "/lib64/ld64.so.2"), .s390x => return copy(&result, "/lib64/ld64.so.1"), .sparcv9 => return copy(&result, "/lib64/ld-linux.so.2"), @@ -1519,6 +1530,7 @@ pub const Target = struct { // TODO go over each item in this list and either move it to the above list, or // implement the standard dynamic linker path code for it. .arc, + .csky, .hexagon, .msp430, .r600, @@ -1573,10 +1585,10 @@ pub const Target = struct { .kfreebsd, .lv2, .solaris, + .zos, .minix, .rtems, .nacl, - .cnk, .aix, .cuda, .nvcl, diff --git a/lib/std/target/aarch64.zig b/lib/std/target/aarch64.zig index 9a2b3e7322..03c2031207 100644 --- a/lib/std/target/aarch64.zig +++ b/lib/std/target/aarch64.zig @@ -1,14 +1,10 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; pub const Feature = enum { - a34, a65, a76, aes, @@ -17,8 +13,6 @@ pub const Feature = enum { altnzcv, am, amvs, - apple_a10, - apple_a11, apple_a12, apple_a13, apple_a7, @@ -26,6 +20,7 @@ pub const Feature = enum { arith_cbz_fusion, balance_fp_ops, bf16, + brbe, bti, call_saved_x10, call_saved_x11, @@ -39,7 +34,11 @@ pub const Feature = enum { ccdp, ccidx, ccpp, + cmp_bcc_fusion, complxnum, + contextidr_el2, + cortex_a78c, + cortex_r82, crc, crypto, custom_cheap_as_move, @@ -49,14 +48,14 @@ pub const Feature = enum { ecv, ete, exynos_cheap_as_move, - exynosm4, + exynos_m4, f32mm, f64mm, fgt, - fmi, + flagm, force_32bit_jump_tables, - fp_armv8, fp16fml, + fp_armv8, fptoint, fullfp16, fuse_address, @@ -67,28 +66,32 @@ pub const Feature = enum { fuse_literals, harden_sls_blr, harden_sls_retbr, + hcx, i8mm, jsconv, lor, + ls64, lse, lsl_fast, mpam, mte, neon, - neoversee1, - neoversen1, + neoverse_e1, + neoverse_n1, + neoverse_n2, + neoverse_v1, no_neg_immediates, nv, - pa, + outline_atomics, pan, pan_rwv, + pauth, perfmon, pmu, predictable_select_expensive, predres, rand, ras, - rasv8_4, rcpc, rcpc_immo, rdm, @@ -126,6 +129,7 @@ pub const Feature = enum { slow_strqro_store, sm4, spe, + spe_eef, specrestrict, ssbs, strict_align, @@ -148,14 +152,18 @@ pub const Feature = enum { use_experimental_zeroing_pseudos, use_postra_scheduler, use_reciprocal_square_root, - v8a, v8_1a, v8_2a, v8_3a, v8_4a, v8_5a, v8_6a, + v8_7a, + v8a, + v8r, vh, + wfxt, + xs, zcm, zcz, zcz_fp, @@ -170,26 +178,16 @@ pub const all_features = blk: { const len = @typeInfo(Feature).Enum.fields.len; std.debug.assert(len <= CpuFeature.Set.needed_bit_count); var result: [len]CpuFeature = undefined; - result[@enumToInt(Feature.a34)] = .{ - .llvm_name = "a35", - .description = "Cortex-A34 ARM processors", - .dependencies = featureSet(&[_]Feature{ - .crc, - .crypto, - .perfmon, - .v8a, - }), - }; result[@enumToInt(Feature.a65)] = .{ .llvm_name = "a65", .description = "Cortex-A65 ARM processors", .dependencies = featureSet(&[_]Feature{ .crypto, .dotprod, - .fp_armv8, .fullfp16, - .neon, - .ras, + .fuse_address, + .fuse_aes, + .fuse_literals, .rcpc, .ssbs, .v8_2a, @@ -202,6 +200,7 @@ pub const all_features = blk: { .crypto, .dotprod, .fullfp16, + .fuse_aes, .rcpc, .ssbs, .v8_2a, @@ -241,49 +240,6 @@ pub const all_features = blk: { .am, }), }; - result[@enumToInt(Feature.apple_a10)] = .{ - .llvm_name = "apple-a10", - .description = "Apple A10", - .dependencies = featureSet(&[_]Feature{ - .alternate_sextload_cvt_f32_pattern, - .arith_bcc_fusion, - .arith_cbz_fusion, - .crc, - .crypto, - .disable_latency_sched_heuristic, - .fp_armv8, - .fuse_aes, - .fuse_crypto_eor, - .lor, - .neon, - .pan, - .perfmon, - .rdm, - .vh, - .zcm, - .zcz, - }), - }; - result[@enumToInt(Feature.apple_a11)] = .{ - .llvm_name = "apple-a11", - .description = "Apple A11", - .dependencies = featureSet(&[_]Feature{ - .alternate_sextload_cvt_f32_pattern, - .arith_bcc_fusion, - .arith_cbz_fusion, - .crypto, - .disable_latency_sched_heuristic, - .fp_armv8, - .fullfp16, - .fuse_aes, - .fuse_crypto_eor, - .neon, - .perfmon, - .v8_2a, - .zcm, - .zcz, - }), - }; result[@enumToInt(Feature.apple_a12)] = .{ .llvm_name = "apple-a12", .description = "Apple A12", @@ -293,11 +249,9 @@ pub const all_features = blk: { .arith_cbz_fusion, .crypto, .disable_latency_sched_heuristic, - .fp_armv8, .fullfp16, .fuse_aes, .fuse_crypto_eor, - .neon, .perfmon, .v8_3a, .zcm, @@ -313,12 +267,9 @@ pub const all_features = blk: { .arith_cbz_fusion, .crypto, .disable_latency_sched_heuristic, - .fp_armv8, .fp16fml, - .fullfp16, .fuse_aes, .fuse_crypto_eor, - .neon, .perfmon, .sha3, .v8_4a, @@ -335,10 +286,8 @@ pub const all_features = blk: { .arith_cbz_fusion, .crypto, .disable_latency_sched_heuristic, - .fp_armv8, .fuse_aes, .fuse_crypto_eor, - .neon, .perfmon, .zcm, .zcz, @@ -365,6 +314,11 @@ pub const all_features = blk: { .description = "Enable BFloat16 Extension", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.brbe)] = .{ + .llvm_name = "brbe", + .description = "Enable Branch Record Buffer Extension", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.bti)] = .{ .llvm_name = "bti", .description = "Enable Branch Target Identification", @@ -430,6 +384,11 @@ pub const all_features = blk: { .description = "Enable v8.2 data Cache Clean to Point of Persistence", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.cmp_bcc_fusion)] = .{ + .llvm_name = "cmp-bcc-fusion", + .description = "CPU fuses cmp+bcc operations", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.complxnum)] = .{ .llvm_name = "complxnum", .description = "Enable v8.3-A Floating-point complex number support", @@ -437,6 +396,38 @@ pub const all_features = blk: { .neon, }), }; + result[@enumToInt(Feature.contextidr_el2)] = .{ + .llvm_name = "CONTEXTIDREL2", + .description = "Enable RW operand Context ID Register (EL2)", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.cortex_a78c)] = .{ + .llvm_name = "cortex-a78c", + .description = "Cortex-A78C ARM processors", + .dependencies = featureSet(&[_]Feature{ + .cmp_bcc_fusion, + .crypto, + .dotprod, + .flagm, + .fp16fml, + .fuse_aes, + .pauth, + .perfmon, + .rcpc, + .spe, + .ssbs, + .use_postra_scheduler, + .v8_2a, + }), + }; + result[@enumToInt(Feature.cortex_r82)] = .{ + .llvm_name = "cortex-r82", + .description = "Cortex-R82 ARM Processors", + .dependencies = featureSet(&[_]Feature{ + .use_postra_scheduler, + .v8r, + }), + }; result[@enumToInt(Feature.crc)] = .{ .llvm_name = "crc", .description = "Enable ARMv8 CRC-32 checksum instructions", @@ -447,7 +438,6 @@ pub const all_features = blk: { .description = "Enable cryptographic instructions", .dependencies = featureSet(&[_]Feature{ .aes, - .neon, .sha2, }), }; @@ -490,7 +480,7 @@ pub const all_features = blk: { .custom_cheap_as_move, }), }; - result[@enumToInt(Feature.exynosm4)] = .{ + result[@enumToInt(Feature.exynos_m4)] = .{ .llvm_name = "exynosm4", .description = "Samsung Exynos-M4 processors", .dependencies = featureSet(&[_]Feature{ @@ -532,8 +522,8 @@ pub const all_features = blk: { .description = "Enable fine grained virtualization traps extension", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.fmi)] = .{ - .llvm_name = "fmi", + result[@enumToInt(Feature.flagm)] = .{ + .llvm_name = "flagm", .description = "Enable v8.4-A Flag Manipulation Instructions", .dependencies = featureSet(&[_]Feature{}), }; @@ -542,11 +532,6 @@ pub const all_features = blk: { .description = "Force jump table entries to be 32-bits wide except at MinSize", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.fp_armv8)] = .{ - .llvm_name = "fp-armv8", - .description = "Enable ARMv8 FP", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.fp16fml)] = .{ .llvm_name = "fp16fml", .description = "Enable FP16 FML instructions", @@ -554,6 +539,11 @@ pub const all_features = blk: { .fullfp16, }), }; + result[@enumToInt(Feature.fp_armv8)] = .{ + .llvm_name = "fp-armv8", + .description = "Enable ARMv8 FP", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.fptoint)] = .{ .llvm_name = "fptoint", .description = "Enable FRInt[32|64][Z|X] instructions that round a floating-point number to an integer (in FP format) forcing it to fit into a 32- or 64-bit int", @@ -606,6 +596,11 @@ pub const all_features = blk: { .description = "Harden against straight line speculation across RET and BR instructions", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.hcx)] = .{ + .llvm_name = "hcx", + .description = "Enable Armv8.7-A HCRX_EL2 system register", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.i8mm)] = .{ .llvm_name = "i8mm", .description = "Enable Matrix Multiply Int8 Extension", @@ -623,6 +618,11 @@ pub const all_features = blk: { .description = "Enables ARM v8.1 Limited Ordering Regions extension", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.ls64)] = .{ + .llvm_name = "ls64", + .description = "Enable Armv8.7-A LD64B/ST64B Accelerator Extension", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.lse)] = .{ .llvm_name = "lse", .description = "Enable ARMv8.1 Large System Extension (LSE) atomic instructions", @@ -650,35 +650,62 @@ pub const all_features = blk: { .fp_armv8, }), }; - result[@enumToInt(Feature.neoversee1)] = .{ + result[@enumToInt(Feature.neoverse_e1)] = .{ .llvm_name = "neoversee1", .description = "Neoverse E1 ARM processors", .dependencies = featureSet(&[_]Feature{ .crypto, .dotprod, - .fp_armv8, .fullfp16, - .neon, .rcpc, .ssbs, .v8_2a, }), }; - result[@enumToInt(Feature.neoversen1)] = .{ + result[@enumToInt(Feature.neoverse_n1)] = .{ .llvm_name = "neoversen1", .description = "Neoverse N1 ARM processors", .dependencies = featureSet(&[_]Feature{ .crypto, .dotprod, - .fp_armv8, .fullfp16, - .neon, .rcpc, .spe, .ssbs, .v8_2a, }), }; + result[@enumToInt(Feature.neoverse_n2)] = .{ + .llvm_name = "neoversen2", + .description = "Neoverse N2 ARM processors", + .dependencies = featureSet(&[_]Feature{ + .bf16, + .ete, + .i8mm, + .mte, + .sve2_bitperm, + .v8_5a, + }), + }; + result[@enumToInt(Feature.neoverse_v1)] = .{ + .llvm_name = "neoversev1", + .description = "Neoverse V1 ARM processors", + .dependencies = featureSet(&[_]Feature{ + .bf16, + .ccdp, + .crypto, + .fp16fml, + .fuse_aes, + .i8mm, + .perfmon, + .rand, + .spe, + .ssbs, + .sve, + .use_postra_scheduler, + .v8_4a, + }), + }; result[@enumToInt(Feature.no_neg_immediates)] = .{ .llvm_name = "no-neg-immediates", .description = "Convert immediates and instructions to their negated or complemented equivalent when the immediate does not fit in the encoding.", @@ -686,12 +713,12 @@ pub const all_features = blk: { }; result[@enumToInt(Feature.nv)] = .{ .llvm_name = "nv", - .description = "Enable v8.4-A Nested Virtualization extension", + .description = "Enable v8.4-A Nested Virtualization Enchancement", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.pa)] = .{ - .llvm_name = "pa", - .description = "Enable v8.3-A Pointer Authentication extension", + result[@enumToInt(Feature.outline_atomics)] = .{ + .llvm_name = "outline-atomics", + .description = "Enable out of line atomics to support LSE instructions", .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.pan)] = .{ @@ -706,6 +733,11 @@ pub const all_features = blk: { .pan, }), }; + result[@enumToInt(Feature.pauth)] = .{ + .llvm_name = "pauth", + .description = "Enable v8.3-A Pointer Authentication extension", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.perfmon)] = .{ .llvm_name = "perfmon", .description = "Enable ARMv8 PMUv3 Performance Monitors extension", @@ -736,13 +768,6 @@ pub const all_features = blk: { .description = "Enable ARMv8 Reliability, Availability and Serviceability Extensions", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.rasv8_4)] = .{ - .llvm_name = "rasv8_4", - .description = "Enable v8.4-A Reliability, Availability and Serviceability extension", - .dependencies = featureSet(&[_]Feature{ - .ras, - }), - }; result[@enumToInt(Feature.rcpc)] = .{ .llvm_name = "rcpc", .description = "Enable support for RCPC extension", @@ -906,7 +931,6 @@ pub const all_features = blk: { .llvm_name = "sha3", .description = "Enable SHA512 and SHA3 support", .dependencies = featureSet(&[_]Feature{ - .neon, .sha2, }), }; @@ -937,6 +961,11 @@ pub const all_features = blk: { .description = "Enable Statistical Profiling extension", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.spe_eef)] = .{ + .llvm_name = "spe-eef", + .description = "Enable extra register in the Statistical Profiling Extension", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.specrestrict)] = .{ .llvm_name = "specrestrict", .description = "Enable architectural speculation restriction", @@ -955,7 +984,9 @@ pub const all_features = blk: { result[@enumToInt(Feature.sve)] = .{ .llvm_name = "sve", .description = "Enable Scalable Vector Extension (SVE) instructions", - .dependencies = featureSet(&[_]Feature{}), + .dependencies = featureSet(&[_]Feature{ + .fullfp16, + }), }; result[@enumToInt(Feature.sve2)] = .{ .llvm_name = "sve2", @@ -1060,14 +1091,6 @@ pub const all_features = blk: { .description = "Use the reciprocal square root approximation", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.v8a)] = .{ - .llvm_name = null, - .description = "Support ARM v8a instructions", - .dependencies = featureSet(&[_]Feature{ - .fp_armv8, - .neon, - }), - }; result[@enumToInt(Feature.v8_1a)] = .{ .llvm_name = "v8.1a", .description = "Support ARM v8.1a instructions", @@ -1077,8 +1100,8 @@ pub const all_features = blk: { .lse, .pan, .rdm, - .vh, .v8a, + .vh, }), }; result[@enumToInt(Feature.v8_2a)] = .{ @@ -1099,7 +1122,7 @@ pub const all_features = blk: { .ccidx, .complxnum, .jsconv, - .pa, + .pauth, .rcpc, .v8_2a, }), @@ -1111,11 +1134,10 @@ pub const all_features = blk: { .am, .dit, .dotprod, - .fmi, + .flagm, .mpam, .nv, .pmu, - .rasv8_4, .rcpc_immo, .sel2, .tlb_rmi, @@ -1150,9 +1172,71 @@ pub const all_features = blk: { .v8_5a, }), }; + result[@enumToInt(Feature.v8_7a)] = .{ + .llvm_name = "v8.7a", + .description = "Support ARM v8.7a instructions", + .dependencies = featureSet(&[_]Feature{ + .hcx, + .v8_6a, + .wfxt, + .xs, + }), + }; + result[@enumToInt(Feature.v8a)] = .{ + .llvm_name = null, + .description = "Support ARM v8a instructions", + .dependencies = featureSet(&[_]Feature{ + .neon, + }), + }; + result[@enumToInt(Feature.v8r)] = .{ + .llvm_name = "v8r", + .description = "Support ARM v8r instructions", + .dependencies = featureSet(&[_]Feature{ + .ccidx, + .ccpp, + .complxnum, + .contextidr_el2, + .crc, + .dit, + .dotprod, + .flagm, + .fp16fml, + .jsconv, + .lse, + .pan_rwv, + .pauth, + .perfmon, + .predres, + .ras, + .rcpc_immo, + .rdm, + .sb, + .sel2, + .sha3, + .sm4, + .specrestrict, + .ssbs, + .tlb_rmi, + .tracev8_4, + .uaops, + }), + }; result[@enumToInt(Feature.vh)] = .{ .llvm_name = "vh", .description = "Enables ARM v8.1 Virtual Host extension", + .dependencies = featureSet(&[_]Feature{ + .contextidr_el2, + }), + }; + result[@enumToInt(Feature.wfxt)] = .{ + .llvm_name = "wfxt", + .description = "Enable Armv8.7-A WFET and WFIT instruction", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.xs)] = .{ + .llvm_name = "xs", + .description = "Enable Armv8.7-A limited-TLB-maintenance instruction", .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.zcm)] = .{ @@ -1196,11 +1280,11 @@ pub const cpu = struct { .name = "a64fx", .llvm_name = "a64fx", .features = featureSet(&[_]Feature{ + .aggressive_fma, + .arith_bcc_fusion, .complxnum, - .fp_armv8, - .fullfp16, - .neon, .perfmon, + .predictable_select_expensive, .sha2, .sve, .use_postra_scheduler, @@ -1211,14 +1295,39 @@ pub const cpu = struct { .name = "apple_a10", .llvm_name = "apple-a10", .features = featureSet(&[_]Feature{ - .apple_a10, + .alternate_sextload_cvt_f32_pattern, + .arith_bcc_fusion, + .arith_cbz_fusion, + .crc, + .crypto, + .disable_latency_sched_heuristic, + .fuse_aes, + .fuse_crypto_eor, + .lor, + .pan, + .perfmon, + .rdm, + .vh, + .zcm, + .zcz, }), }; pub const apple_a11 = CpuModel{ .name = "apple_a11", .llvm_name = "apple-a11", .features = featureSet(&[_]Feature{ - .apple_a11, + .alternate_sextload_cvt_f32_pattern, + .arith_bcc_fusion, + .arith_cbz_fusion, + .crypto, + .disable_latency_sched_heuristic, + .fullfp16, + .fuse_aes, + .fuse_crypto_eor, + .perfmon, + .v8_2a, + .zcm, + .zcz, }), }; pub const apple_a12 = CpuModel{ @@ -1235,6 +1344,37 @@ pub const cpu = struct { .apple_a13, }), }; + pub const apple_a14 = CpuModel{ + .name = "apple_a14", + .llvm_name = "apple-a14", + .features = featureSet(&[_]Feature{ + .aggressive_fma, + .alternate_sextload_cvt_f32_pattern, + .altnzcv, + .arith_bcc_fusion, + .arith_cbz_fusion, + .ccdp, + .crypto, + .disable_latency_sched_heuristic, + .fp16fml, + .fptoint, + .fuse_address, + .fuse_aes, + .fuse_arith_logic, + .fuse_crypto_eor, + .fuse_csel, + .fuse_literals, + .perfmon, + .predres, + .sb, + .sha3, + .specrestrict, + .ssbs, + .v8_4a, + .zcm, + .zcz, + }), + }; pub const apple_a7 = CpuModel{ .name = "apple_a7", .llvm_name = "apple-a7", @@ -1283,7 +1423,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .crypto, .fullfp16, - .neon, .v8_2a, }), }; @@ -1291,14 +1430,20 @@ pub const cpu = struct { .name = "cortex_a34", .llvm_name = "cortex-a34", .features = featureSet(&[_]Feature{ - .a34, + .crc, + .crypto, + .perfmon, + .v8a, }), }; pub const cortex_a35 = CpuModel{ .name = "cortex_a35", .llvm_name = "cortex-a35", .features = featureSet(&[_]Feature{ - .a34, + .crc, + .crypto, + .perfmon, + .v8a, }), }; pub const cortex_a53 = CpuModel{ @@ -1366,6 +1511,7 @@ pub const cpu = struct { .crc, .crypto, .fuse_aes, + .fuse_literals, .perfmon, .v8a, }), @@ -1412,11 +1558,11 @@ pub const cpu = struct { .name = "cortex_a77", .llvm_name = "cortex-a77", .features = featureSet(&[_]Feature{ + .cmp_bcc_fusion, .crypto, .dotprod, - .fp_armv8, .fullfp16, - .neon, + .fuse_aes, .rcpc, .v8_2a, }), @@ -1425,12 +1571,11 @@ pub const cpu = struct { .name = "cortex_a78", .llvm_name = "cortex-a78", .features = featureSet(&[_]Feature{ + .cmp_bcc_fusion, .crypto, .dotprod, - .fp_armv8, .fullfp16, .fuse_aes, - .neon, .perfmon, .rcpc, .spe, @@ -1439,16 +1584,29 @@ pub const cpu = struct { .v8_2a, }), }; + pub const cortex_a78c = CpuModel{ + .name = "cortex_a78c", + .llvm_name = "cortex-a78c", + .features = featureSet(&[_]Feature{ + .cortex_a78c, + }), + }; + pub const cortex_r82 = CpuModel{ + .name = "cortex_r82", + .llvm_name = "cortex-r82", + .features = featureSet(&[_]Feature{ + .cortex_r82, + }), + }; pub const cortex_x1 = CpuModel{ .name = "cortex_x1", .llvm_name = "cortex-x1", .features = featureSet(&[_]Feature{ + .cmp_bcc_fusion, .crypto, .dotprod, - .fp_armv8, .fullfp16, .fuse_aes, - .neon, .perfmon, .rcpc, .spe, @@ -1522,14 +1680,14 @@ pub const cpu = struct { .name = "exynos_m4", .llvm_name = "exynos-m4", .features = featureSet(&[_]Feature{ - .exynosm4, + .exynos_m4, }), }; pub const exynos_m5 = CpuModel{ .name = "exynos_m5", .llvm_name = "exynos-m5", .features = featureSet(&[_]Feature{ - .exynosm4, + .exynos_m4, }), }; pub const falkor = CpuModel{ @@ -1555,9 +1713,9 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .ete, .fuse_aes, + .neon, .perfmon, .use_postra_scheduler, - .v8a, }), }; pub const kryo = CpuModel{ @@ -1571,22 +1729,36 @@ pub const cpu = struct { .perfmon, .predictable_select_expensive, .use_postra_scheduler, - .zcz, .v8a, + .zcz, }), }; pub const neoverse_e1 = CpuModel{ .name = "neoverse_e1", .llvm_name = "neoverse-e1", .features = featureSet(&[_]Feature{ - .neoversee1, + .neoverse_e1, }), }; pub const neoverse_n1 = CpuModel{ .name = "neoverse_n1", .llvm_name = "neoverse-n1", .features = featureSet(&[_]Feature{ - .neoversen1, + .neoverse_n1, + }), + }; + pub const neoverse_n2 = CpuModel{ + .name = "neoverse_n2", + .llvm_name = "neoverse-n2", + .features = featureSet(&[_]Feature{ + .neoverse_n2, + }), + }; + pub const neoverse_v1 = CpuModel{ + .name = "neoverse_v1", + .llvm_name = "neoverse-v1", + .features = featureSet(&[_]Feature{ + .neoverse_v1, }), }; pub const saphira = CpuModel{ @@ -1622,9 +1794,7 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .aggressive_fma, .arith_bcc_fusion, - .crc, .crypto, - .lse, .predictable_select_expensive, .use_postra_scheduler, .v8_1a, @@ -1637,12 +1807,7 @@ pub const cpu = struct { .aggressive_fma, .arith_bcc_fusion, .balance_fp_ops, - .crc, .crypto, - .fp_armv8, - .lse, - .neon, - .pa, .perfmon, .predictable_select_expensive, .strict_align, @@ -1695,7 +1860,6 @@ pub const cpu = struct { .custom_cheap_as_move, .dotprod, .fp16fml, - .fullfp16, .fuse_aes, .perfmon, .spe, diff --git a/lib/std/target/amdgpu.zig b/lib/std/target/amdgpu.zig index 09be754964..6e36a34a5a 100644 --- a/lib/std/target/amdgpu.zig +++ b/lib/std/target/amdgpu.zig @@ -1,22 +1,17 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; pub const Feature = enum { @"16_bit_insts", - DumpCode, a16, add_no_carry_insts, aperture_regs, atomic_fadd_insts, auto_waitcnt_before_barrier, ci_insts, - code_object_v3, cumode, dl_insts, dot1_insts, @@ -28,7 +23,6 @@ pub const Feature = enum { dpp, dpp8, ds_src2_insts, - dumpcode, enable_ds128, enable_prt_strict_null, fast_denormal_f32, @@ -47,13 +41,15 @@ pub const Feature = enum { get_wave_id_inst, gfx10, gfx10_3_insts, - gfx10_insts, gfx10_b_encoding, + gfx10_insts, gfx7_gfx8_gfx9_insts, gfx8_insts, gfx9, gfx9_insts, half_rate_64_ops, + image_gather4_d16_bug, + image_store_d16_bug, inst_fwd_prefetch_bug, int_clamp_insts, inv_2pi_inline_imm, @@ -76,8 +72,6 @@ pub const Feature = enum { movrel, no_data_dep_hazard, no_sdst_cmpx, - no_sram_ecc_support, - no_xnack_support, nsa_encoding, nsa_to_vmem_bug, offset_3f_bug, @@ -101,10 +95,13 @@ pub const Feature = enum { si_scheduler, smem_to_vector_write_hazard, southern_islands, - sram_ecc, + sramecc, + sramecc_support, trap_handler, trig_reduced_range, + unaligned_access_mode, unaligned_buffer_access, + unaligned_ds_access, unaligned_scratch_access, unpacked_d16_vmem, unsafe_ds_offset_folding, @@ -120,6 +117,7 @@ pub const Feature = enum { wavefrontsize32, wavefrontsize64, xnack, + xnack_support, }; pub usingnamespace CpuFeature.feature_set_fns(Feature); @@ -133,11 +131,6 @@ pub const all_features = blk: { .description = "Has i16/f16 instructions", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.DumpCode)] = .{ - .llvm_name = "DumpCode", - .description = "Dump MachineInstrs in the CodeEmitter", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.a16)] = .{ .llvm_name = "a16", .description = "Support gfx10-style A16 for 16-bit coordinates/gradients/lod/clamp/mip image operands", @@ -170,11 +163,6 @@ pub const all_features = blk: { .description = "Additional instructions for CI+", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.code_object_v3)] = .{ - .llvm_name = "code-object-v3", - .description = "Generate code object version 3", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.cumode)] = .{ .llvm_name = "cumode", .description = "Enable CU wavefront execution mode", @@ -230,14 +218,9 @@ pub const all_features = blk: { .description = "Has ds_*_src2 instructions", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.dumpcode)] = .{ - .llvm_name = "dumpcode", - .description = "Dump MachineInstrs in the CodeEmitter", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.enable_ds128)] = .{ .llvm_name = "enable-ds128", - .description = "Use ds_read|write_b128", + .description = "Use ds_{read|write}_b128", .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.enable_prt_strict_null)] = .{ @@ -345,7 +328,6 @@ pub const all_features = blk: { .movrel, .no_data_dep_hazard, .no_sdst_cmpx, - .no_sram_ecc_support, .pk_fmac_f16_inst, .register_banking, .s_memrealtime, @@ -353,6 +335,8 @@ pub const all_features = blk: { .sdwa_omod, .sdwa_scalar, .sdwa_sdst, + .unaligned_buffer_access, + .unaligned_ds_access, .vop3_literal, .vop3p, .vscnt, @@ -363,16 +347,16 @@ pub const all_features = blk: { .description = "Additional instructions for GFX10.3", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.gfx10_insts)] = .{ - .llvm_name = "gfx10-insts", - .description = "Additional instructions for GFX10+", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.gfx10_b_encoding)] = .{ .llvm_name = "gfx10_b-encoding", .description = "Encoding format GFX10_B", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.gfx10_insts)] = .{ + .llvm_name = "gfx10-insts", + .description = "Additional instructions for GFX10+", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.gfx7_gfx8_gfx9_insts)] = .{ .llvm_name = "gfx7-gfx8-gfx9-insts", .description = "Instructions shared in GFX7, GFX8, GFX9", @@ -418,9 +402,12 @@ pub const all_features = blk: { .sdwa_omod, .sdwa_scalar, .sdwa_sdst, + .unaligned_buffer_access, + .unaligned_ds_access, .vgpr_index_mode, .vop3p, .wavefrontsize64, + .xnack_support, }), }; result[@enumToInt(Feature.gfx9_insts)] = .{ @@ -433,6 +420,16 @@ pub const all_features = blk: { .description = "Most fp64 instructions are half rate instead of quarter", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.image_gather4_d16_bug)] = .{ + .llvm_name = "image-gather4-d16-bug", + .description = "Image Gather4 D16 hardware bug", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.image_store_d16_bug)] = .{ + .llvm_name = "image-store-d16-bug", + .description = "Image Store D16 hardware bug", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.inst_fwd_prefetch_bug)] = .{ .llvm_name = "inst-fwd-prefetch-bug", .description = "S_INST_PREFETCH instruction causes shader to hang", @@ -455,7 +452,7 @@ pub const all_features = blk: { }; result[@enumToInt(Feature.lds_misaligned_bug)] = .{ .llvm_name = "lds-misaligned-bug", - .description = "Some GFX10 bug with misaligned multi-dword LDS access in WGP mode", + .description = "Some GFX10 bug with multi-dword LDS and flat access that is not naturally aligned in WGP mode", .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.ldsbankcount16)] = .{ @@ -543,16 +540,6 @@ pub const all_features = blk: { .description = "V_CMPX does not write VCC/SGPR in addition to EXEC", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.no_sram_ecc_support)] = .{ - .llvm_name = "no-sram-ecc-support", - .description = "Hardware does not support SRAM ECC", - .dependencies = featureSet(&[_]Feature{}), - }; - result[@enumToInt(Feature.no_xnack_support)] = .{ - .llvm_name = "no-xnack-support", - .description = "Hardware does not support XNACK", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.nsa_encoding)] = .{ .llvm_name = "nsa-encoding", .description = "Support NSA encoding for image instructions", @@ -656,9 +643,9 @@ pub const all_features = blk: { .mad_mac_f32_insts, .mimg_r128, .movrel, - .no_sram_ecc_support, .s_memtime_inst, .trig_reduced_range, + .unaligned_buffer_access, .wavefrontsize64, }), }; @@ -688,16 +675,19 @@ pub const all_features = blk: { .mad_mac_f32_insts, .mimg_r128, .movrel, - .no_sram_ecc_support, - .no_xnack_support, .s_memtime_inst, .trig_reduced_range, .wavefrontsize64, }), }; - result[@enumToInt(Feature.sram_ecc)] = .{ - .llvm_name = "sram-ecc", - .description = "Enable SRAM ECC", + result[@enumToInt(Feature.sramecc)] = .{ + .llvm_name = "sramecc", + .description = "Enable SRAMECC", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.sramecc_support)] = .{ + .llvm_name = "sramecc-support", + .description = "Hardware supports SRAMECC", .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.trap_handler)] = .{ @@ -710,9 +700,19 @@ pub const all_features = blk: { .description = "Requires use of fract on arguments to trig instructions", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.unaligned_access_mode)] = .{ + .llvm_name = "unaligned-access-mode", + .description = "Enable unaligned global, local and region loads and stores if the hardware supports it", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.unaligned_buffer_access)] = .{ .llvm_name = "unaligned-buffer-access", - .description = "Support unaligned global loads and stores", + .description = "Hardware supports unaligned global loads and stores", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.unaligned_ds_access)] = .{ + .llvm_name = "unaligned-ds-access", + .description = "Hardware supports unaligned local and region loads and stores", .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.unaligned_scratch_access)] = .{ @@ -770,7 +770,6 @@ pub const all_features = blk: { .mad_mac_f32_insts, .mimg_r128, .movrel, - .no_sram_ecc_support, .s_memrealtime, .s_memtime_inst, .scalar_stores, @@ -778,6 +777,7 @@ pub const all_features = blk: { .sdwa_mav, .sdwa_out_mods_vopc, .trig_reduced_range, + .unaligned_buffer_access, .vgpr_index_mode, .wavefrontsize64, }), @@ -817,6 +817,11 @@ pub const all_features = blk: { .description = "Enable XNACK support", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.xnack_support)] = .{ + .llvm_name = "xnack-support", + .description = "Hardware supports XNACK", + .dependencies = featureSet(&[_]Feature{}), + }; const ti = @typeInfo(Feature); for (result) |*elem, i| { elem.index = i; @@ -830,9 +835,7 @@ pub const cpu = struct { .name = "bonaire", .llvm_name = "bonaire", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .sea_islands, }), }; @@ -840,22 +843,19 @@ pub const cpu = struct { .name = "carrizo", .llvm_name = "carrizo", .features = featureSet(&[_]Feature{ - .code_object_v3, .fast_fmaf, .half_rate_64_ops, .ldsbankcount32, .unpacked_d16_vmem, .volcanic_islands, - .xnack, + .xnack_support, }), }; pub const fiji = CpuModel{ .name = "fiji", .llvm_name = "fiji", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .unpacked_d16_vmem, .volcanic_islands, }), @@ -879,7 +879,6 @@ pub const cpu = struct { .name = "gfx1010", .llvm_name = "gfx1010", .features = featureSet(&[_]Feature{ - .code_object_v3, .dl_insts, .ds_src2_insts, .flat_segment_offset_bug, @@ -890,7 +889,6 @@ pub const cpu = struct { .lds_misaligned_bug, .ldsbankcount32, .mad_mac_f32_insts, - .no_xnack_support, .nsa_encoding, .nsa_to_vmem_bug, .offset_3f_bug, @@ -903,13 +901,13 @@ pub const cpu = struct { .vcmpx_permlane_hazard, .vmem_to_scalar_write_hazard, .wavefrontsize32, + .xnack_support, }), }; pub const gfx1011 = CpuModel{ .name = "gfx1011", .llvm_name = "gfx1011", .features = featureSet(&[_]Feature{ - .code_object_v3, .dl_insts, .dot1_insts, .dot2_insts, @@ -921,9 +919,9 @@ pub const cpu = struct { .gfx10, .inst_fwd_prefetch_bug, .lds_branch_vmem_war_hazard, + .lds_misaligned_bug, .ldsbankcount32, .mad_mac_f32_insts, - .no_xnack_support, .nsa_encoding, .nsa_to_vmem_bug, .offset_3f_bug, @@ -936,13 +934,13 @@ pub const cpu = struct { .vcmpx_permlane_hazard, .vmem_to_scalar_write_hazard, .wavefrontsize32, + .xnack_support, }), }; pub const gfx1012 = CpuModel{ .name = "gfx1012", .llvm_name = "gfx1012", .features = featureSet(&[_]Feature{ - .code_object_v3, .dl_insts, .dot1_insts, .dot2_insts, @@ -957,7 +955,6 @@ pub const cpu = struct { .lds_misaligned_bug, .ldsbankcount32, .mad_mac_f32_insts, - .no_xnack_support, .nsa_encoding, .nsa_to_vmem_bug, .offset_3f_bug, @@ -970,13 +967,13 @@ pub const cpu = struct { .vcmpx_permlane_hazard, .vmem_to_scalar_write_hazard, .wavefrontsize32, + .xnack_support, }), }; pub const gfx1030 = CpuModel{ .name = "gfx1030", .llvm_name = "gfx1030", .features = featureSet(&[_]Feature{ - .code_object_v3, .dl_insts, .dot1_insts, .dot2_insts, @@ -986,7 +983,57 @@ pub const cpu = struct { .gfx10_3_insts, .gfx10_b_encoding, .ldsbankcount32, - .no_xnack_support, + .nsa_encoding, + .wavefrontsize32, + }), + }; + pub const gfx1031 = CpuModel{ + .name = "gfx1031", + .llvm_name = "gfx1031", + .features = featureSet(&[_]Feature{ + .dl_insts, + .dot1_insts, + .dot2_insts, + .dot5_insts, + .dot6_insts, + .gfx10, + .gfx10_3_insts, + .gfx10_b_encoding, + .ldsbankcount32, + .nsa_encoding, + .wavefrontsize32, + }), + }; + pub const gfx1032 = CpuModel{ + .name = "gfx1032", + .llvm_name = "gfx1032", + .features = featureSet(&[_]Feature{ + .dl_insts, + .dot1_insts, + .dot2_insts, + .dot5_insts, + .dot6_insts, + .gfx10, + .gfx10_3_insts, + .gfx10_b_encoding, + .ldsbankcount32, + .nsa_encoding, + .wavefrontsize32, + }), + }; + pub const gfx1033 = CpuModel{ + .name = "gfx1033", + .llvm_name = "gfx1033", + .features = featureSet(&[_]Feature{ + .dl_insts, + .dot1_insts, + .dot2_insts, + .dot5_insts, + .dot6_insts, + .gfx10, + .gfx10_3_insts, + .gfx10_b_encoding, + .ldsbankcount32, .nsa_encoding, .wavefrontsize32, }), @@ -995,11 +1042,8 @@ pub const cpu = struct { .name = "gfx600", .llvm_name = "gfx600", .features = featureSet(&[_]Feature{ - .code_object_v3, .fast_fmaf, .half_rate_64_ops, - .ldsbankcount32, - .no_xnack_support, .southern_islands, }), }; @@ -1007,9 +1051,13 @@ pub const cpu = struct { .name = "gfx601", .llvm_name = "gfx601", .features = featureSet(&[_]Feature{ - .code_object_v3, - .ldsbankcount32, - .no_xnack_support, + .southern_islands, + }), + }; + pub const gfx602 = CpuModel{ + .name = "gfx602", + .llvm_name = "gfx602", + .features = featureSet(&[_]Feature{ .southern_islands, }), }; @@ -1017,9 +1065,7 @@ pub const cpu = struct { .name = "gfx700", .llvm_name = "gfx700", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .sea_islands, }), }; @@ -1027,11 +1073,9 @@ pub const cpu = struct { .name = "gfx701", .llvm_name = "gfx701", .features = featureSet(&[_]Feature{ - .code_object_v3, .fast_fmaf, .half_rate_64_ops, .ldsbankcount32, - .no_xnack_support, .sea_islands, }), }; @@ -1039,10 +1083,8 @@ pub const cpu = struct { .name = "gfx702", .llvm_name = "gfx702", .features = featureSet(&[_]Feature{ - .code_object_v3, .fast_fmaf, .ldsbankcount16, - .no_xnack_support, .sea_islands, }), }; @@ -1050,9 +1092,7 @@ pub const cpu = struct { .name = "gfx703", .llvm_name = "gfx703", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount16, - .no_xnack_support, .sea_islands, }), }; @@ -1060,9 +1100,15 @@ pub const cpu = struct { .name = "gfx704", .llvm_name = "gfx704", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, + .sea_islands, + }), + }; + pub const gfx705 = CpuModel{ + .name = "gfx705", + .llvm_name = "gfx705", + .features = featureSet(&[_]Feature{ + .ldsbankcount16, .sea_islands, }), }; @@ -1070,22 +1116,19 @@ pub const cpu = struct { .name = "gfx801", .llvm_name = "gfx801", .features = featureSet(&[_]Feature{ - .code_object_v3, .fast_fmaf, .half_rate_64_ops, .ldsbankcount32, .unpacked_d16_vmem, .volcanic_islands, - .xnack, + .xnack_support, }), }; pub const gfx802 = CpuModel{ .name = "gfx802", .llvm_name = "gfx802", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .sgpr_init_bug, .unpacked_d16_vmem, .volcanic_islands, @@ -1095,9 +1138,17 @@ pub const cpu = struct { .name = "gfx803", .llvm_name = "gfx803", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, + .unpacked_d16_vmem, + .volcanic_islands, + }), + }; + pub const gfx805 = CpuModel{ + .name = "gfx805", + .llvm_name = "gfx805", + .features = featureSet(&[_]Feature{ + .ldsbankcount32, + .sgpr_init_bug, .unpacked_d16_vmem, .volcanic_islands, }), @@ -1106,61 +1157,56 @@ pub const cpu = struct { .name = "gfx810", .llvm_name = "gfx810", .features = featureSet(&[_]Feature{ - .code_object_v3, + .image_gather4_d16_bug, + .image_store_d16_bug, .ldsbankcount16, .volcanic_islands, - .xnack, + .xnack_support, }), }; pub const gfx900 = CpuModel{ .name = "gfx900", .llvm_name = "gfx900", .features = featureSet(&[_]Feature{ - .code_object_v3, .gfx9, + .image_gather4_d16_bug, .ldsbankcount32, .mad_mix_insts, - .no_sram_ecc_support, - .no_xnack_support, }), }; pub const gfx902 = CpuModel{ .name = "gfx902", .llvm_name = "gfx902", .features = featureSet(&[_]Feature{ - .code_object_v3, .gfx9, + .image_gather4_d16_bug, .ldsbankcount32, .mad_mix_insts, - .no_sram_ecc_support, - .xnack, }), }; pub const gfx904 = CpuModel{ .name = "gfx904", .llvm_name = "gfx904", .features = featureSet(&[_]Feature{ - .code_object_v3, .fma_mix_insts, .gfx9, + .image_gather4_d16_bug, .ldsbankcount32, - .no_sram_ecc_support, - .no_xnack_support, }), }; pub const gfx906 = CpuModel{ .name = "gfx906", .llvm_name = "gfx906", .features = featureSet(&[_]Feature{ - .code_object_v3, .dl_insts, .dot1_insts, .dot2_insts, .fma_mix_insts, .gfx9, .half_rate_64_ops, + .image_gather4_d16_bug, .ldsbankcount32, - .no_xnack_support, + .sramecc_support, }), }; pub const gfx908 = CpuModel{ @@ -1168,7 +1214,6 @@ pub const cpu = struct { .llvm_name = "gfx908", .features = featureSet(&[_]Feature{ .atomic_fadd_insts, - .code_object_v3, .dl_insts, .dot1_insts, .dot2_insts, @@ -1179,19 +1224,30 @@ pub const cpu = struct { .fma_mix_insts, .gfx9, .half_rate_64_ops, + .image_gather4_d16_bug, .ldsbankcount32, .mai_insts, .mfma_inline_literal_bug, .pk_fmac_f16_inst, - .sram_ecc, + .sramecc_support, }), }; pub const gfx909 = CpuModel{ .name = "gfx909", .llvm_name = "gfx909", .features = featureSet(&[_]Feature{ - .code_object_v3, .gfx9, + .image_gather4_d16_bug, + .ldsbankcount32, + .mad_mix_insts, + }), + }; + pub const gfx90c = CpuModel{ + .name = "gfx90c", + .llvm_name = "gfx90c", + .features = featureSet(&[_]Feature{ + .gfx9, + .image_gather4_d16_bug, .ldsbankcount32, .mad_mix_insts, .xnack, @@ -1201,9 +1257,6 @@ pub const cpu = struct { .name = "hainan", .llvm_name = "hainan", .features = featureSet(&[_]Feature{ - .code_object_v3, - .ldsbankcount32, - .no_xnack_support, .southern_islands, }), }; @@ -1211,11 +1264,9 @@ pub const cpu = struct { .name = "hawaii", .llvm_name = "hawaii", .features = featureSet(&[_]Feature{ - .code_object_v3, .fast_fmaf, .half_rate_64_ops, .ldsbankcount32, - .no_xnack_support, .sea_islands, }), }; @@ -1223,9 +1274,7 @@ pub const cpu = struct { .name = "iceland", .llvm_name = "iceland", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .sgpr_init_bug, .unpacked_d16_vmem, .volcanic_islands, @@ -1235,9 +1284,7 @@ pub const cpu = struct { .name = "kabini", .llvm_name = "kabini", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount16, - .no_xnack_support, .sea_islands, }), }; @@ -1245,9 +1292,7 @@ pub const cpu = struct { .name = "kaveri", .llvm_name = "kaveri", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .sea_islands, }), }; @@ -1255,9 +1300,7 @@ pub const cpu = struct { .name = "mullins", .llvm_name = "mullins", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount16, - .no_xnack_support, .sea_islands, }), }; @@ -1265,9 +1308,6 @@ pub const cpu = struct { .name = "oland", .llvm_name = "oland", .features = featureSet(&[_]Feature{ - .code_object_v3, - .ldsbankcount32, - .no_xnack_support, .southern_islands, }), }; @@ -1275,9 +1315,6 @@ pub const cpu = struct { .name = "pitcairn", .llvm_name = "pitcairn", .features = featureSet(&[_]Feature{ - .code_object_v3, - .ldsbankcount32, - .no_xnack_support, .southern_islands, }), }; @@ -1285,9 +1322,7 @@ pub const cpu = struct { .name = "polaris10", .llvm_name = "polaris10", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .unpacked_d16_vmem, .volcanic_islands, }), @@ -1296,9 +1331,7 @@ pub const cpu = struct { .name = "polaris11", .llvm_name = "polaris11", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, .unpacked_d16_vmem, .volcanic_islands, }), @@ -1307,21 +1340,19 @@ pub const cpu = struct { .name = "stoney", .llvm_name = "stoney", .features = featureSet(&[_]Feature{ - .code_object_v3, + .image_gather4_d16_bug, + .image_store_d16_bug, .ldsbankcount16, .volcanic_islands, - .xnack, + .xnack_support, }), }; pub const tahiti = CpuModel{ .name = "tahiti", .llvm_name = "tahiti", .features = featureSet(&[_]Feature{ - .code_object_v3, .fast_fmaf, .half_rate_64_ops, - .ldsbankcount32, - .no_xnack_support, .southern_islands, }), }; @@ -1329,9 +1360,17 @@ pub const cpu = struct { .name = "tonga", .llvm_name = "tonga", .features = featureSet(&[_]Feature{ - .code_object_v3, .ldsbankcount32, - .no_xnack_support, + .sgpr_init_bug, + .unpacked_d16_vmem, + .volcanic_islands, + }), + }; + pub const tongapro = CpuModel{ + .name = "tongapro", + .llvm_name = "tongapro", + .features = featureSet(&[_]Feature{ + .ldsbankcount32, .sgpr_init_bug, .unpacked_d16_vmem, .volcanic_islands, @@ -1341,9 +1380,6 @@ pub const cpu = struct { .name = "verde", .llvm_name = "verde", .features = featureSet(&[_]Feature{ - .code_object_v3, - .ldsbankcount32, - .no_xnack_support, .southern_islands, }), }; diff --git a/lib/std/target/arm.zig b/lib/std/target/arm.zig index 96365b3e04..ecd08f2861 100644 --- a/lib/std/target/arm.zig +++ b/lib/std/target/arm.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; @@ -53,6 +50,8 @@ pub const Feature = enum { fullfp16, fuse_aes, fuse_literals, + harden_sls_blr, + harden_sls_retbr, has_v4t, has_v5t, has_v5te, @@ -62,6 +61,7 @@ pub const Feature = enum { has_v6t2, has_v7, has_v7clrex, + has_v8, has_v8_1a, has_v8_1m_main, has_v8_2a, @@ -69,7 +69,7 @@ pub const Feature = enum { has_v8_4a, has_v8_5a, has_v8_6a, - has_v8, + has_v8_7a, has_v8m, has_v8m_main, hwdiv, @@ -85,10 +85,10 @@ pub const Feature = enum { mp, muxed_units, mve, - mve_fp, mve1beat, mve2beat, mve4beat, + mve_fp, nacl_trap, neon, neon_fpmovs, @@ -148,10 +148,6 @@ pub const Feature = enum { v7r, v7s, v7ve, - v8a, - v8m, - v8m_main, - v8r, v8_1a, v8_1m_main, v8_2a, @@ -159,6 +155,11 @@ pub const Feature = enum { v8_4a, v8_5a, v8_6a, + v8_7a, + v8a, + v8m, + v8m_main, + v8r, vfp2, vfp2sp, vfp3, @@ -238,7 +239,7 @@ pub const all_features = blk: { .llvm_name = "cde", .description = "Support CDE instructions", .dependencies = featureSet(&[_]Feature{ - .v8m_main, + .has_v8m_main, }), }; result[@enumToInt(Feature.cdecp0)] = .{ @@ -312,7 +313,6 @@ pub const all_features = blk: { .description = "Enable support for Cryptography extensions", .dependencies = featureSet(&[_]Feature{ .aes, - .neon, .sha2, }), }; @@ -419,7 +419,6 @@ pub const all_features = blk: { .description = "Enable ARMv8 FP with only 16 d-registers", .dependencies = featureSet(&[_]Feature{ .fp_armv8d16sp, - .fp64, .vfp4d16, }), }; @@ -434,7 +433,6 @@ pub const all_features = blk: { .llvm_name = "fp-armv8sp", .description = "Enable ARMv8 FP with no double precision", .dependencies = featureSet(&[_]Feature{ - .d32, .fp_armv8d16sp, .vfp4sp, }), @@ -481,6 +479,16 @@ pub const all_features = blk: { .description = "CPU fuses literal generation operations", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.harden_sls_blr)] = .{ + .llvm_name = "harden-sls-blr", + .description = "Harden against straight line speculation across indirect calls", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.harden_sls_retbr)] = .{ + .llvm_name = "harden-sls-retbr", + .description = "Harden against straight line speculation across RETurn and BranchRegister instructions", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.has_v4t)] = .{ .llvm_name = "v4t", .description = "Support ARM v4T instructions", @@ -525,18 +533,18 @@ pub const all_features = blk: { .llvm_name = "v6t2", .description = "Support ARM v6t2 instructions", .dependencies = featureSet(&[_]Feature{ - .thumb2, .has_v6k, .has_v8m, + .thumb2, }), }; result[@enumToInt(Feature.has_v7)] = .{ .llvm_name = "v7", .description = "Support ARM v7 instructions", .dependencies = featureSet(&[_]Feature{ - .perfmon, .has_v6t2, .has_v7clrex, + .perfmon, }), }; result[@enumToInt(Feature.has_v7clrex)] = .{ @@ -592,8 +600,8 @@ pub const all_features = blk: { .llvm_name = "v8.5a", .description = "Support ARM v8.5a instructions", .dependencies = featureSet(&[_]Feature{ - .sb, .has_v8_4a, + .sb, }), }; result[@enumToInt(Feature.has_v8_6a)] = .{ @@ -601,8 +609,15 @@ pub const all_features = blk: { .description = "Support ARM v8.6a instructions", .dependencies = featureSet(&[_]Feature{ .bf16, - .i8mm, .has_v8_5a, + .i8mm, + }), + }; + result[@enumToInt(Feature.has_v8_7a)] = .{ + .llvm_name = "v8.7a", + .description = "Support ARM v8.7a instructions", + .dependencies = featureSet(&[_]Feature{ + .has_v8_6a, }), }; result[@enumToInt(Feature.has_v8m)] = .{ @@ -640,14 +655,14 @@ pub const all_features = blk: { .llvm_name = "iwmmxt", .description = "ARMv5te architecture", .dependencies = featureSet(&[_]Feature{ - .has_v5te, + .v5te, }), }; result[@enumToInt(Feature.iwmmxt2)] = .{ .llvm_name = "iwmmxt2", .description = "ARMv5te architecture", .dependencies = featureSet(&[_]Feature{ - .has_v5te, + .v5te, }), }; result[@enumToInt(Feature.lob)] = .{ @@ -695,15 +710,6 @@ pub const all_features = blk: { .has_v8_1m_main, }), }; - result[@enumToInt(Feature.mve_fp)] = .{ - .llvm_name = "mve.fp", - .description = "Support M-Class Vector Extension with integer and floating ops", - .dependencies = featureSet(&[_]Feature{ - .fp_armv8d16sp, - .fullfp16, - .mve, - }), - }; result[@enumToInt(Feature.mve1beat)] = .{ .llvm_name = "mve1beat", .description = "Model MVE instructions as a 1 beat per tick architecture", @@ -719,6 +725,14 @@ pub const all_features = blk: { .description = "Model MVE instructions as a 4 beats per tick architecture", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.mve_fp)] = .{ + .llvm_name = "mve.fp", + .description = "Support M-Class Vector Extension with integer and floating ops", + .dependencies = featureSet(&[_]Feature{ + .fullfp16, + .mve, + }), + }; result[@enumToInt(Feature.nacl_trap)] = .{ .llvm_name = "nacl-trap", .description = "NaCl trap", @@ -944,32 +958,32 @@ pub const all_features = blk: { .llvm_name = "armv4t", .description = "ARMv4t architecture", .dependencies = featureSet(&[_]Feature{ - .strict_align, .has_v4t, + .strict_align, }), }; result[@enumToInt(Feature.v5t)] = .{ .llvm_name = "armv5t", .description = "ARMv5t architecture", .dependencies = featureSet(&[_]Feature{ - .strict_align, .has_v5t, + .strict_align, }), }; result[@enumToInt(Feature.v5te)] = .{ .llvm_name = "armv5te", .description = "ARMv5te architecture", .dependencies = featureSet(&[_]Feature{ - .strict_align, .has_v5te, + .strict_align, }), }; result[@enumToInt(Feature.v5tej)] = .{ .llvm_name = "armv5tej", .description = "ARMv5tej architecture", .dependencies = featureSet(&[_]Feature{ - .strict_align, .has_v5te, + .strict_align, }), }; result[@enumToInt(Feature.v6)] = .{ @@ -980,18 +994,6 @@ pub const all_features = blk: { .has_v6, }), }; - result[@enumToInt(Feature.v6m)] = .{ - .llvm_name = "armv6-m", - .description = "ARMv6m architecture", - .dependencies = featureSet(&[_]Feature{ - .db, - .mclass, - .noarm, - .strict_align, - .thumb_mode, - .has_v6m, - }), - }; result[@enumToInt(Feature.v6j)] = .{ .llvm_name = "armv6j", .description = "ARMv7a architecture", @@ -1010,8 +1012,20 @@ pub const all_features = blk: { .llvm_name = "armv6kz", .description = "ARMv6kz architecture", .dependencies = featureSet(&[_]Feature{ - .trustzone, .has_v6k, + .trustzone, + }), + }; + result[@enumToInt(Feature.v6m)] = .{ + .llvm_name = "armv6-m", + .description = "ARMv6m architecture", + .dependencies = featureSet(&[_]Feature{ + .db, + .has_v6m, + .mclass, + .noarm, + .strict_align, + .thumb_mode, }), }; result[@enumToInt(Feature.v6sm)] = .{ @@ -1019,11 +1033,11 @@ pub const all_features = blk: { .description = "ARMv6sm architecture", .dependencies = featureSet(&[_]Feature{ .db, + .has_v6m, .mclass, .noarm, .strict_align, .thumb_mode, - .has_v6m, }), }; result[@enumToInt(Feature.v6t2)] = .{ @@ -1041,53 +1055,51 @@ pub const all_features = blk: { .aclass, .db, .dsp, - .neon, .has_v7, + .neon, }), }; - result[@enumToInt(Feature.v7m)] = .{ - .llvm_name = "armv7-m", - .description = "ARMv7m architecture", + result[@enumToInt(Feature.v7em)] = .{ + .llvm_name = "armv7e-m", + .description = "ARMv7em architecture", .dependencies = featureSet(&[_]Feature{ .db, + .dsp, + .has_v7, .hwdiv, .mclass, .noarm, .thumb_mode, - .thumb2, - .has_v7, }), }; - result[@enumToInt(Feature.v7r)] = .{ - .llvm_name = "armv7-r", - .description = "ARMv7r architecture", + result[@enumToInt(Feature.v7k)] = .{ + .llvm_name = "armv7k", + .description = "ARMv7a architecture", .dependencies = featureSet(&[_]Feature{ - .db, - .dsp, - .hwdiv, - .rclass, - .has_v7, + .v7a, }), }; - result[@enumToInt(Feature.v7em)] = .{ - .llvm_name = "armv7e-m", - .description = "ARMv7em architecture", + result[@enumToInt(Feature.v7m)] = .{ + .llvm_name = "armv7-m", + .description = "ARMv7m architecture", .dependencies = featureSet(&[_]Feature{ .db, - .dsp, + .has_v7, .hwdiv, .mclass, .noarm, .thumb_mode, - .thumb2, - .has_v7, }), }; - result[@enumToInt(Feature.v7k)] = .{ - .llvm_name = "armv7k", - .description = "ARMv7a architecture", + result[@enumToInt(Feature.v7r)] = .{ + .llvm_name = "armv7-r", + .description = "ARMv7r architecture", .dependencies = featureSet(&[_]Feature{ - .v7a, + .db, + .dsp, + .has_v7, + .hwdiv, + .rclass, }), }; result[@enumToInt(Feature.v7s)] = .{ @@ -1104,73 +1116,10 @@ pub const all_features = blk: { .aclass, .db, .dsp, - .mp, - .neon, - .trustzone, .has_v7, - .virtualization, - }), - }; - result[@enumToInt(Feature.v8a)] = .{ - .llvm_name = "armv8-a", - .description = "ARMv8a architecture", - .dependencies = featureSet(&[_]Feature{ - .aclass, - .crc, - .crypto, - .db, - .dsp, - .fp_armv8, .mp, .neon, .trustzone, - .has_v8, - .virtualization, - }), - }; - result[@enumToInt(Feature.v8m)] = .{ - .llvm_name = "armv8-m.base", - .description = "ARMv8mBaseline architecture", - .dependencies = featureSet(&[_]Feature{ - .@"8msecext", - .acquire_release, - .db, - .hwdiv, - .mclass, - .noarm, - .strict_align, - .thumb_mode, - .has_v7clrex, - .has_v8m, - }), - }; - result[@enumToInt(Feature.v8m_main)] = .{ - .llvm_name = "armv8-m.main", - .description = "ARMv8mMainline architecture", - .dependencies = featureSet(&[_]Feature{ - .@"8msecext", - .acquire_release, - .db, - .hwdiv, - .mclass, - .noarm, - .thumb_mode, - .has_v8m_main, - }), - }; - result[@enumToInt(Feature.v8r)] = .{ - .llvm_name = "armv8-r", - .description = "ARMv8r architecture", - .dependencies = featureSet(&[_]Feature{ - .crc, - .db, - .dfb, - .dsp, - .fp_armv8, - .mp, - .neon, - .rclass, - .has_v8, .virtualization, }), }; @@ -1184,10 +1133,9 @@ pub const all_features = blk: { .db, .dsp, .fp_armv8, + .has_v8_1a, .mp, - .neon, .trustzone, - .has_v8_1a, .virtualization, }), }; @@ -1198,13 +1146,13 @@ pub const all_features = blk: { .@"8msecext", .acquire_release, .db, + .has_v8_1m_main, .hwdiv, .lob, .mclass, .noarm, .ras, .thumb_mode, - .has_v8_1m_main, }), }; result[@enumToInt(Feature.v8_2a)] = .{ @@ -1217,11 +1165,10 @@ pub const all_features = blk: { .db, .dsp, .fp_armv8, + .has_v8_2a, .mp, - .neon, .ras, .trustzone, - .has_v8_2a, .virtualization, }), }; @@ -1235,11 +1182,10 @@ pub const all_features = blk: { .db, .dsp, .fp_armv8, + .has_v8_3a, .mp, - .neon, .ras, .trustzone, - .has_v8_3a, .virtualization, }), }; @@ -1251,14 +1197,12 @@ pub const all_features = blk: { .crc, .crypto, .db, - .dotprod, .dsp, .fp_armv8, + .has_v8_4a, .mp, - .neon, .ras, .trustzone, - .has_v8_4a, .virtualization, }), }; @@ -1270,14 +1214,12 @@ pub const all_features = blk: { .crc, .crypto, .db, - .dotprod, .dsp, .fp_armv8, + .has_v8_5a, .mp, - .neon, .ras, .trustzone, - .has_v8_5a, .virtualization, }), }; @@ -1289,14 +1231,91 @@ pub const all_features = blk: { .crc, .crypto, .db, - .dotprod, .dsp, .fp_armv8, + .has_v8_6a, .mp, - .neon, .ras, .trustzone, - .has_v8_6a, + .virtualization, + }), + }; + result[@enumToInt(Feature.v8_7a)] = .{ + .llvm_name = "armv8.7-a", + .description = "ARMv86a architecture", + .dependencies = featureSet(&[_]Feature{ + .aclass, + .crc, + .crypto, + .db, + .dsp, + .fp_armv8, + .has_v8_7a, + .mp, + .ras, + .trustzone, + .virtualization, + }), + }; + result[@enumToInt(Feature.v8a)] = .{ + .llvm_name = "armv8-a", + .description = "ARMv8a architecture", + .dependencies = featureSet(&[_]Feature{ + .aclass, + .crc, + .crypto, + .db, + .dsp, + .fp_armv8, + .has_v8, + .mp, + .trustzone, + .virtualization, + }), + }; + result[@enumToInt(Feature.v8m)] = .{ + .llvm_name = "armv8-m.base", + .description = "ARMv8mBaseline architecture", + .dependencies = featureSet(&[_]Feature{ + .@"8msecext", + .acquire_release, + .db, + .has_v7clrex, + .has_v8m, + .hwdiv, + .mclass, + .noarm, + .strict_align, + .thumb_mode, + }), + }; + result[@enumToInt(Feature.v8m_main)] = .{ + .llvm_name = "armv8-m.main", + .description = "ARMv8mMainline architecture", + .dependencies = featureSet(&[_]Feature{ + .@"8msecext", + .acquire_release, + .db, + .has_v8m_main, + .hwdiv, + .mclass, + .noarm, + .thumb_mode, + }), + }; + result[@enumToInt(Feature.v8r)] = .{ + .llvm_name = "armv8-r", + .description = "ARMv8r architecture", + .dependencies = featureSet(&[_]Feature{ + .crc, + .db, + .dfb, + .dsp, + .fp_armv8, + .has_v8, + .mp, + .neon, + .rclass, .virtualization, }), }; @@ -1327,7 +1346,6 @@ pub const all_features = blk: { .llvm_name = "vfp3d16", .description = "Enable VFP3 instructions with only 16 d-registers", .dependencies = featureSet(&[_]Feature{ - .fp64, .vfp2, .vfp3d16sp, }), @@ -1351,7 +1369,6 @@ pub const all_features = blk: { .llvm_name = "vfp4", .description = "Enable VFP4 instructions", .dependencies = featureSet(&[_]Feature{ - .fp16, .vfp3, .vfp4d16, .vfp4sp, @@ -1361,8 +1378,6 @@ pub const all_features = blk: { .llvm_name = "vfp4d16", .description = "Enable VFP4 instructions with only 16 d-registers", .dependencies = featureSet(&[_]Feature{ - .fp16, - .fp64, .vfp3d16, .vfp4d16sp, }), @@ -1379,8 +1394,6 @@ pub const all_features = blk: { .llvm_name = "vfp4sp", .description = "Enable VFP4 instructions with no double precision", .dependencies = featureSet(&[_]Feature{ - .d32, - .fp16, .vfp3sp, .vfp4d16sp, }), @@ -1417,7 +1430,7 @@ pub const all_features = blk: { .llvm_name = "xscale", .description = "ARMv5te architecture", .dependencies = featureSet(&[_]Feature{ - .has_v5te, + .v5te, }), }; result[@enumToInt(Feature.zcz)] = .{ @@ -1480,8 +1493,8 @@ pub const cpu = struct { .name = "arm1136jf_s", .llvm_name = "arm1136jf-s", .features = featureSet(&[_]Feature{ - .v6, .slowfpvmlx, + .v6, .vfp2, }), }; @@ -1496,8 +1509,8 @@ pub const cpu = struct { .name = "arm1156t2f_s", .llvm_name = "arm1156t2f-s", .features = featureSet(&[_]Feature{ - .v6t2, .slowfpvmlx, + .v6t2, .vfp2, }), }; @@ -1519,8 +1532,8 @@ pub const cpu = struct { .name = "arm1176jzf_s", .llvm_name = "arm1176jzf-s", .features = featureSet(&[_]Feature{ - .v6kz, .slowfpvmlx, + .v6kz, .vfp2, }), }; @@ -1654,11 +1667,11 @@ pub const cpu = struct { .name = "cortex_a12", .llvm_name = "cortex-a12", .features = featureSet(&[_]Feature{ - .v7a, .avoid_partial_cpsr, .mp, .ret_addr_stack, .trustzone, + .v7a, .vfp4, .virtualization, .vmlx_forwarding, @@ -1668,14 +1681,13 @@ pub const cpu = struct { .name = "cortex_a15", .llvm_name = "cortex-a15", .features = featureSet(&[_]Feature{ - .v7a, .avoid_partial_cpsr, - .dont_widen_vmovs, .mp, .muxed_units, .ret_addr_stack, .splat_vfp_neon, .trustzone, + .v7a, .vfp4, .virtualization, .vldn_align, @@ -1685,11 +1697,11 @@ pub const cpu = struct { .name = "cortex_a17", .llvm_name = "cortex-a17", .features = featureSet(&[_]Feature{ - .v7a, .avoid_partial_cpsr, .mp, .ret_addr_stack, .trustzone, + .v7a, .vfp4, .virtualization, .vmlx_forwarding, @@ -1699,10 +1711,6 @@ pub const cpu = struct { .name = "cortex_a32", .llvm_name = "cortex-a32", .features = featureSet(&[_]Feature{ - .crc, - .crypto, - .hwdiv, - .hwdiv_arm, .v8a, }), }; @@ -1710,10 +1718,6 @@ pub const cpu = struct { .name = "cortex_a35", .llvm_name = "cortex-a35", .features = featureSet(&[_]Feature{ - .crc, - .crypto, - .hwdiv, - .hwdiv_arm, .v8a, }), }; @@ -1721,13 +1725,13 @@ pub const cpu = struct { .name = "cortex_a5", .llvm_name = "cortex-a5", .features = featureSet(&[_]Feature{ - .v7a, .mp, .ret_addr_stack, .slow_fp_brcc, .slowfpvfmx, .slowfpvmlx, .trustzone, + .v7a, .vfp4, .vmlx_forwarding, }), @@ -1736,49 +1740,39 @@ pub const cpu = struct { .name = "cortex_a53", .llvm_name = "cortex-a53", .features = featureSet(&[_]Feature{ - .v8a, - .crc, - .crypto, .fpao, - .hwdiv, - .hwdiv_arm, + .v8a, }), }; pub const cortex_a55 = CpuModel{ .name = "cortex_a55", .llvm_name = "cortex-a55", .features = featureSet(&[_]Feature{ - .v8_2a, .dotprod, - .hwdiv, - .hwdiv_arm, + .v8_2a, }), }; pub const cortex_a57 = CpuModel{ .name = "cortex_a57", .llvm_name = "cortex-a57", .features = featureSet(&[_]Feature{ - .v8a, .avoid_partial_cpsr, .cheap_predicable_cpsr, - .crc, - .crypto, .fpao, - .hwdiv, - .hwdiv_arm, + .v8a, }), }; pub const cortex_a7 = CpuModel{ .name = "cortex_a7", .llvm_name = "cortex-a7", .features = featureSet(&[_]Feature{ - .v7a, .mp, .ret_addr_stack, .slow_fp_brcc, .slowfpvfmx, .slowfpvmlx, .trustzone, + .v7a, .vfp4, .virtualization, .vmlx_forwarding, @@ -1790,10 +1784,6 @@ pub const cpu = struct { .llvm_name = "cortex-a72", .features = featureSet(&[_]Feature{ .v8a, - .crc, - .crypto, - .hwdiv, - .hwdiv_arm, }), }; pub const cortex_a73 = CpuModel{ @@ -1801,20 +1791,14 @@ pub const cpu = struct { .llvm_name = "cortex-a73", .features = featureSet(&[_]Feature{ .v8a, - .crc, - .crypto, - .hwdiv, - .hwdiv_arm, }), }; pub const cortex_a75 = CpuModel{ .name = "cortex_a75", .llvm_name = "cortex-a75", .features = featureSet(&[_]Feature{ - .v8_2a, .dotprod, - .hwdiv, - .hwdiv_arm, + .v8_2a, }), }; pub const cortex_a76 = CpuModel{ @@ -1822,13 +1806,9 @@ pub const cpu = struct { .llvm_name = "cortex-a76", .features = featureSet(&[_]Feature{ .a76, - .v8_2a, - .crc, - .crypto, .dotprod, .fullfp16, - .hwdiv, - .hwdiv_arm, + .v8_2a, }), }; pub const cortex_a76ae = CpuModel{ @@ -1836,52 +1816,49 @@ pub const cpu = struct { .llvm_name = "cortex-a76ae", .features = featureSet(&[_]Feature{ .a76, - .v8_2a, - .crc, - .crypto, .dotprod, .fullfp16, - .hwdiv, - .hwdiv_arm, + .v8_2a, }), }; pub const cortex_a77 = CpuModel{ .name = "cortex_a77", .llvm_name = "cortex-a77", .features = featureSet(&[_]Feature{ - .v8_2a, - .crc, - .crypto, .dotprod, .fullfp16, - .hwdiv, - .hwdiv_arm, + .v8_2a, }), }; pub const cortex_a78 = CpuModel{ .name = "cortex_a78", .llvm_name = "cortex-a78", .features = featureSet(&[_]Feature{ + .dotprod, + .fullfp16, .v8_2a, - .crc, - .crypto, + }), + }; + pub const cortex_a78c = CpuModel{ + .name = "cortex_a78c", + .llvm_name = "cortex-a78c", + .features = featureSet(&[_]Feature{ .dotprod, .fullfp16, - .hwdiv, - .hwdiv_arm, + .v8_2a, }), }; pub const cortex_a8 = CpuModel{ .name = "cortex_a8", .llvm_name = "cortex-a8", .features = featureSet(&[_]Feature{ - .v7a, .nonpipelined_vfp, .ret_addr_stack, .slow_fp_brcc, .slowfpvfmx, .slowfpvmlx, .trustzone, + .v7a, .vmlx_forwarding, .vmlx_hazards, }), @@ -1890,7 +1867,6 @@ pub const cpu = struct { .name = "cortex_a9", .llvm_name = "cortex-a9", .features = featureSet(&[_]Feature{ - .v7a, .avoid_partial_cpsr, .expand_fp_mlx, .fp16, @@ -1900,6 +1876,7 @@ pub const cpu = struct { .prefer_vmovsr, .ret_addr_stack, .trustzone, + .v7a, .vldn_align, .vmlx_forwarding, .vmlx_hazards, @@ -1930,26 +1907,25 @@ pub const cpu = struct { .name = "cortex_m23", .llvm_name = "cortex-m23", .features = featureSet(&[_]Feature{ - .v8m, .no_movt, + .v8m, }), }; pub const cortex_m3 = CpuModel{ .name = "cortex_m3", .llvm_name = "cortex-m3", .features = featureSet(&[_]Feature{ - .v7m, .loop_align, .m3, .no_branch_predictor, .use_misched, + .v7m, }), }; pub const cortex_m33 = CpuModel{ .name = "cortex_m33", .llvm_name = "cortex-m33", .features = featureSet(&[_]Feature{ - .v8m_main, .dsp, .fp_armv8d16sp, .loop_align, @@ -1957,13 +1933,13 @@ pub const cpu = struct { .slowfpvfmx, .slowfpvmlx, .use_misched, + .v8m_main, }), }; pub const cortex_m35p = CpuModel{ .name = "cortex_m35p", .llvm_name = "cortex-m35p", .features = featureSet(&[_]Feature{ - .v8m_main, .dsp, .fp_armv8d16sp, .loop_align, @@ -1971,18 +1947,19 @@ pub const cpu = struct { .slowfpvfmx, .slowfpvmlx, .use_misched, + .v8m_main, }), }; pub const cortex_m4 = CpuModel{ .name = "cortex_m4", .llvm_name = "cortex-m4", .features = featureSet(&[_]Feature{ - .v7em, .loop_align, .no_branch_predictor, .slowfpvfmx, .slowfpvmlx, .use_misched, + .v7em, .vfp4d16sp, }), }; @@ -1990,45 +1967,45 @@ pub const cpu = struct { .name = "cortex_m55", .llvm_name = "cortex-m55", .features = featureSet(&[_]Feature{ - .v8_1m_main, - .dsp, .fp_armv8d16, .loop_align, .mve_fp, .no_branch_predictor, .slowfpvmlx, .use_misched, + .v8_1m_main, }), }; pub const cortex_m7 = CpuModel{ .name = "cortex_m7", .llvm_name = "cortex-m7", .features = featureSet(&[_]Feature{ - .v7em, .fp_armv8d16, + .use_misched, + .v7em, }), }; pub const cortex_r4 = CpuModel{ .name = "cortex_r4", .llvm_name = "cortex-r4", .features = featureSet(&[_]Feature{ - .v7r, .avoid_partial_cpsr, .r4, .ret_addr_stack, + .v7r, }), }; pub const cortex_r4f = CpuModel{ .name = "cortex_r4f", .llvm_name = "cortex-r4f", .features = featureSet(&[_]Feature{ - .v7r, .avoid_partial_cpsr, .r4, .ret_addr_stack, .slow_fp_brcc, .slowfpvfmx, .slowfpvmlx, + .v7r, .vfp3d16, }), }; @@ -2036,13 +2013,13 @@ pub const cpu = struct { .name = "cortex_r5", .llvm_name = "cortex-r5", .features = featureSet(&[_]Feature{ - .v7r, .avoid_partial_cpsr, .hwdiv_arm, .ret_addr_stack, .slow_fp_brcc, .slowfpvfmx, .slowfpvmlx, + .v7r, .vfp3d16, }), }; @@ -2050,16 +2027,15 @@ pub const cpu = struct { .name = "cortex_r52", .llvm_name = "cortex-r52", .features = featureSet(&[_]Feature{ - .v8r, .fpao, .use_misched, + .v8r, }), }; pub const cortex_r7 = CpuModel{ .name = "cortex_r7", .llvm_name = "cortex-r7", .features = featureSet(&[_]Feature{ - .v7r, .avoid_partial_cpsr, .fp16, .hwdiv_arm, @@ -2068,6 +2044,7 @@ pub const cpu = struct { .slow_fp_brcc, .slowfpvfmx, .slowfpvmlx, + .v7r, .vfp3d16, }), }; @@ -2075,7 +2052,6 @@ pub const cpu = struct { .name = "cortex_r8", .llvm_name = "cortex-r8", .features = featureSet(&[_]Feature{ - .v7r, .avoid_partial_cpsr, .fp16, .hwdiv_arm, @@ -2084,6 +2060,7 @@ pub const cpu = struct { .slow_fp_brcc, .slowfpvfmx, .slowfpvmlx, + .v7r, .vfp3d16, }), }; @@ -2091,34 +2068,25 @@ pub const cpu = struct { .name = "cortex_x1", .llvm_name = "cortex-x1", .features = featureSet(&[_]Feature{ - .v8_2a, - .crc, - .crypto, .dotprod, .fullfp16, - .hwdiv, - .hwdiv_arm, + .v8_2a, }), }; pub const cyclone = CpuModel{ .name = "cyclone", .llvm_name = "cyclone", .features = featureSet(&[_]Feature{ - .v8a, .avoid_movs_shop, .avoid_partial_cpsr, - .crypto, .disable_postra_scheduler, - .hwdiv, - .hwdiv_arm, - .mp, .neonfp, .ret_addr_stack, .slowfpvfmx, .slowfpvmlx, .swift, .use_misched, - .vfp4, + .v8a, .zcz, }), }; @@ -2133,34 +2101,34 @@ pub const cpu = struct { .name = "exynos_m1", .llvm_name = null, .features = featureSet(&[_]Feature{ - .v8a, .exynos, + .v8a, }), }; pub const exynos_m2 = CpuModel{ .name = "exynos_m2", .llvm_name = null, .features = featureSet(&[_]Feature{ - .v8a, .exynos, + .v8a, }), }; pub const exynos_m3 = CpuModel{ .name = "exynos_m3", .llvm_name = "exynos-m3", .features = featureSet(&[_]Feature{ - .v8_2a, .exynos, + .v8a, }), }; pub const exynos_m4 = CpuModel{ .name = "exynos_m4", .llvm_name = "exynos-m4", .features = featureSet(&[_]Feature{ - .v8_2a, .dotprod, .exynos, .fullfp16, + .v8_2a, }), }; pub const exynos_m5 = CpuModel{ @@ -2190,7 +2158,6 @@ pub const cpu = struct { .llvm_name = "krait", .features = featureSet(&[_]Feature{ .avoid_partial_cpsr, - .fp16, .hwdiv, .hwdiv_arm, .muxed_units, @@ -2205,10 +2172,6 @@ pub const cpu = struct { .name = "kryo", .llvm_name = "kryo", .features = featureSet(&[_]Feature{ - .crc, - .crypto, - .hwdiv, - .hwdiv_arm, .v8a, }), }; @@ -2216,8 +2179,8 @@ pub const cpu = struct { .name = "mpcore", .llvm_name = "mpcore", .features = featureSet(&[_]Feature{ - .v6k, .slowfpvmlx, + .v6k, .vfp2, }), }; @@ -2232,12 +2195,27 @@ pub const cpu = struct { .name = "neoverse_n1", .llvm_name = "neoverse-n1", .features = featureSet(&[_]Feature{ - .v8_2a, - .crc, - .crypto, .dotprod, - .hwdiv, - .hwdiv_arm, + .v8_2a, + }), + }; + pub const neoverse_n2 = CpuModel{ + .name = "neoverse_n2", + .llvm_name = "neoverse-n2", + .features = featureSet(&[_]Feature{ + .bf16, + .i8mm, + .v8_5a, + }), + }; + pub const neoverse_v1 = CpuModel{ + .name = "neoverse_v1", + .llvm_name = "neoverse-v1", + .features = featureSet(&[_]Feature{ + .bf16, + .fullfp16, + .i8mm, + .v8_4a, }), }; pub const sc000 = CpuModel{ @@ -2251,10 +2229,10 @@ pub const cpu = struct { .name = "sc300", .llvm_name = "sc300", .features = featureSet(&[_]Feature{ - .v7m, .m3, .no_branch_predictor, .use_misched, + .v7m, }), }; pub const strongarm = CpuModel{ @@ -2289,7 +2267,6 @@ pub const cpu = struct { .name = "swift", .llvm_name = "swift", .features = featureSet(&[_]Feature{ - .v7a, .avoid_movs_shop, .avoid_partial_cpsr, .disable_postra_scheduler, @@ -2308,6 +2285,7 @@ pub const cpu = struct { .slowfpvmlx, .swift, .use_misched, + .v7a, .vfp4, .vmlx_hazards, .wide_stride_vfp, diff --git a/lib/std/target/avr.zig b/lib/std/target/avr.zig index f85867444a..079d30cf92 100644 --- a/lib/std/target/avr.zig +++ b/lib/std/target/avr.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; @@ -163,7 +160,6 @@ pub const all_features = blk: { .dependencies = featureSet(&[_]Feature{ .avr0, .@"break", - .memmappedregs, .sram, .tinyencoding, }), @@ -802,6 +798,13 @@ pub const cpu = struct { .avr5, }), }; + pub const atmega168pb = CpuModel{ + .name = "atmega168pb", + .llvm_name = "atmega168pb", + .features = featureSet(&[_]Feature{ + .avr5, + }), + }; pub const atmega169 = CpuModel{ .name = "atmega169", .llvm_name = "atmega169", @@ -949,6 +952,13 @@ pub const cpu = struct { .avr5, }), }; + pub const atmega324pb = CpuModel{ + .name = "atmega324pb", + .llvm_name = "atmega324pb", + .features = featureSet(&[_]Feature{ + .avr5, + }), + }; pub const atmega325 = CpuModel{ .name = "atmega325", .llvm_name = "atmega325", @@ -1019,6 +1029,13 @@ pub const cpu = struct { .avr5, }), }; + pub const atmega328pb = CpuModel{ + .name = "atmega328pb", + .llvm_name = "atmega328pb", + .features = featureSet(&[_]Feature{ + .avr5, + }), + }; pub const atmega329 = CpuModel{ .name = "atmega329", .llvm_name = "atmega329", @@ -1166,6 +1183,13 @@ pub const cpu = struct { .avr4, }), }; + pub const atmega48pb = CpuModel{ + .name = "atmega48pb", + .llvm_name = "atmega48pb", + .features = featureSet(&[_]Feature{ + .avr4, + }), + }; pub const atmega64 = CpuModel{ .name = "atmega64", .llvm_name = "atmega64", @@ -1338,7 +1362,11 @@ pub const cpu = struct { .name = "atmega8", .llvm_name = "atmega8", .features = featureSet(&[_]Feature{ - .avr4, + .avr2, + .lpmx, + .movw, + .mul, + .spm, }), }; pub const atmega8515 = CpuModel{ @@ -1391,11 +1419,22 @@ pub const cpu = struct { .avr4, }), }; + pub const atmega88pb = CpuModel{ + .name = "atmega88pb", + .llvm_name = "atmega88pb", + .features = featureSet(&[_]Feature{ + .avr4, + }), + }; pub const atmega8a = CpuModel{ .name = "atmega8a", .llvm_name = "atmega8a", .features = featureSet(&[_]Feature{ - .avr4, + .avr2, + .lpmx, + .movw, + .mul, + .spm, }), }; pub const atmega8hva = CpuModel{ @@ -1595,6 +1634,13 @@ pub const cpu = struct { .avr25, }), }; + pub const attiny441 = CpuModel{ + .name = "attiny441", + .llvm_name = "attiny441", + .features = featureSet(&[_]Feature{ + .avr25, + }), + }; pub const attiny44a = CpuModel{ .name = "attiny44a", .llvm_name = "attiny44a", @@ -1651,6 +1697,13 @@ pub const cpu = struct { .avr25, }), }; + pub const attiny841 = CpuModel{ + .name = "attiny841", + .llvm_name = "attiny841", + .features = featureSet(&[_]Feature{ + .avr25, + }), + }; pub const attiny84a = CpuModel{ .name = "attiny84a", .llvm_name = "attiny84a", @@ -1802,7 +1855,7 @@ pub const cpu = struct { .name = "atxmega16e5", .llvm_name = "atxmega16e5", .features = featureSet(&[_]Feature{ - .xmega, + .xmegau, }), }; pub const atxmega192a3 = CpuModel{ @@ -1907,7 +1960,7 @@ pub const cpu = struct { .name = "atxmega32e5", .llvm_name = "atxmega32e5", .features = featureSet(&[_]Feature{ - .xmega, + .xmegau, }), }; pub const atxmega32x1 = CpuModel{ @@ -2005,7 +2058,7 @@ pub const cpu = struct { .name = "atxmega8e5", .llvm_name = "atxmega8e5", .features = featureSet(&[_]Feature{ - .xmega, + .xmegau, }), }; pub const avr1 = CpuModel{ diff --git a/lib/std/target/bpf.zig b/lib/std/target/bpf.zig index 73287ec6a8..3d3032689d 100644 --- a/lib/std/target/bpf.zig +++ b/lib/std/target/bpf.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; diff --git a/lib/std/target/hexagon.zig b/lib/std/target/hexagon.zig index b1f565f52d..aa8d8f6efd 100644 --- a/lib/std/target/hexagon.zig +++ b/lib/std/target/hexagon.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; @@ -92,7 +89,6 @@ pub const all_features = blk: { .llvm_name = "hvxv62", .description = "Hexagon HVX instructions", .dependencies = featureSet(&[_]Feature{ - .hvx, .hvxv60, }), }; @@ -100,8 +96,6 @@ pub const all_features = blk: { .llvm_name = "hvxv65", .description = "Hexagon HVX instructions", .dependencies = featureSet(&[_]Feature{ - .hvx, - .hvxv60, .hvxv62, }), }; @@ -109,9 +103,6 @@ pub const all_features = blk: { .llvm_name = "hvxv66", .description = "Hexagon HVX instructions", .dependencies = featureSet(&[_]Feature{ - .hvx, - .hvxv60, - .hvxv62, .hvxv65, .zreg, }), @@ -120,9 +111,6 @@ pub const all_features = blk: { .llvm_name = "hvxv67", .description = "Hexagon HVX instructions", .dependencies = featureSet(&[_]Feature{ - .hvxv60, - .hvxv62, - .hvxv65, .hvxv66, }), }; @@ -248,7 +236,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .prev65, .small_data, .v5, @@ -265,7 +252,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .prev65, .small_data, .v5, @@ -280,7 +266,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .prev65, .small_data, .v5, @@ -296,7 +281,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .prev65, .small_data, .v5, @@ -313,7 +297,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .prev65, .small_data, .v5, @@ -332,7 +315,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .small_data, .v5, .v55, @@ -351,7 +333,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .small_data, .v5, .v55, @@ -371,7 +352,6 @@ pub const cpu = struct { .memops, .nvj, .nvs, - .packets, .small_data, .v5, .v55, @@ -391,7 +371,6 @@ pub const cpu = struct { .mem_noshuf, .memops, .nvs, - .packets, .small_data, .tinycore, .v5, diff --git a/lib/std/target/mips.zig b/lib/std/target/mips.zig index 59da13ac39..c88d2ec8fb 100644 --- a/lib/std/target/mips.zig +++ b/lib/std/target/mips.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; @@ -108,7 +105,6 @@ pub const all_features = blk: { .llvm_name = "dspr3", .description = "Mips DSP-R3 ASE", .dependencies = featureSet(&[_]Feature{ - .dsp, .dspr2, }), }; @@ -301,10 +297,8 @@ pub const all_features = blk: { .llvm_name = "mips64r6", .description = "Mips64r6 ISA Support [experimental]", .dependencies = featureSet(&[_]Feature{ - .abs2008, .mips32r6, .mips64r5, - .nan2008, }), }; result[@enumToInt(Feature.msa)] = .{ @@ -515,16 +509,13 @@ pub const cpu = struct { .llvm_name = "octeon", .features = featureSet(&[_]Feature{ .cnmips, - .mips64r2, }), }; pub const @"octeon+" = CpuModel{ .name = "octeon+", .llvm_name = "octeon+", .features = featureSet(&[_]Feature{ - .cnmips, .cnmipsp, - .mips64r2, }), }; pub const p5600 = CpuModel{ diff --git a/lib/std/target/msp430.zig b/lib/std/target/msp430.zig index c1005a1d81..be59b09706 100644 --- a/lib/std/target/msp430.zig +++ b/lib/std/target/msp430.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; diff --git a/lib/std/target/nvptx.zig b/lib/std/target/nvptx.zig index b025fbfcf7..6b9b9bf777 100644 --- a/lib/std/target/nvptx.zig +++ b/lib/std/target/nvptx.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; diff --git a/lib/std/target/powerpc.zig b/lib/std/target/powerpc.zig index 2db7d30e8d..4e2200a47f 100644 --- a/lib/std/target/powerpc.zig +++ b/lib/std/target/powerpc.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; @@ -10,6 +7,7 @@ const CpuModel = std.Target.Cpu.Model; pub const Feature = enum { @"64bit", @"64bitregs", + aix, allow_unaligned_fp_access, altivec, booke, @@ -19,6 +17,7 @@ pub const Feature = enum { crypto, direct_move, e500, + efpu2, extdiv, fcpsgn, float128, @@ -32,6 +31,7 @@ pub const Feature = enum { fsqrt, fuse_addi_load, fuse_addis_load, + fuse_store, fusion, hard_float, htm, @@ -44,7 +44,10 @@ pub const Feature = enum { lfiwax, longcall, mfocrf, + mma, + modern_aix_as, msync, + paired_vector_memops, partword_atomics, pcrelative_memops, popcntd, @@ -53,13 +56,12 @@ pub const Feature = enum { power8_vector, power9_altivec, power9_vector, - ppc_postra_sched, - ppc_prera_sched, ppc4xx, ppc6xx, + ppc_postra_sched, + ppc_prera_sched, predictable_select_expensive, prefix_instrs, - qpx, recipprec, secure_plt, slow_popcntd, @@ -86,6 +88,11 @@ pub const all_features = blk: { .description = "Enable 64-bit registers usage for ppc32 [beta]", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.aix)] = .{ + .llvm_name = "aix", + .description = "AIX OS", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.allow_unaligned_fp_access)] = .{ .llvm_name = "allow-unaligned-fp-access", .description = "CPU does not trap on unaligned FP access", @@ -139,6 +146,13 @@ pub const all_features = blk: { .description = "Enable E500/E500mc instructions", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.efpu2)] = .{ + .llvm_name = "efpu2", + .description = "Enable Embedded Floating-Point APU 2 instructions", + .dependencies = featureSet(&[_]Feature{ + .spe, + }), + }; result[@enumToInt(Feature.extdiv)] = .{ .llvm_name = "extdiv", .description = "Enable extended divide instructions", @@ -228,6 +242,13 @@ pub const all_features = blk: { .fusion, }), }; + result[@enumToInt(Feature.fuse_store)] = .{ + .llvm_name = "fuse-store", + .description = "Target supports store clustering", + .dependencies = featureSet(&[_]Feature{ + .fusion, + }), + }; result[@enumToInt(Feature.fusion)] = .{ .llvm_name = "fusion", .description = "Target supports instruction fusion", @@ -292,6 +313,20 @@ pub const all_features = blk: { .description = "Enable the MFOCRF instruction", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.mma)] = .{ + .llvm_name = "mma", + .description = "Enable MMA instructions", + .dependencies = featureSet(&[_]Feature{ + .paired_vector_memops, + .power8_vector, + .power9_altivec, + }), + }; + result[@enumToInt(Feature.modern_aix_as)] = .{ + .llvm_name = "modern-aix-as", + .description = "AIX system assembler is modern enough to support new mnes", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.msync)] = .{ .llvm_name = "msync", .description = "Has only the msync instruction instead of sync", @@ -299,6 +334,13 @@ pub const all_features = blk: { .booke, }), }; + result[@enumToInt(Feature.paired_vector_memops)] = .{ + .llvm_name = "paired-vector-memops", + .description = "32Byte load and store instructions", + .dependencies = featureSet(&[_]Feature{ + .isa_v30_instructions, + }), + }; result[@enumToInt(Feature.partword_atomics)] = .{ .llvm_name = "partword-atomics", .description = "Enable l[bh]arx and st[bh]cx.", @@ -308,7 +350,7 @@ pub const all_features = blk: { .llvm_name = "pcrelative-memops", .description = "Enable PC relative Memory Ops", .dependencies = featureSet(&[_]Feature{ - .isa_v30_instructions, + .prefix_instrs, }), }; result[@enumToInt(Feature.popcntd)] = .{ @@ -351,21 +393,10 @@ pub const all_features = blk: { .llvm_name = "power9-vector", .description = "Enable POWER9 vector instructions", .dependencies = featureSet(&[_]Feature{ - .isa_v30_instructions, .power8_vector, .power9_altivec, }), }; - result[@enumToInt(Feature.ppc_postra_sched)] = .{ - .llvm_name = "ppc-postra-sched", - .description = "Use PowerPC post-RA scheduling strategy", - .dependencies = featureSet(&[_]Feature{}), - }; - result[@enumToInt(Feature.ppc_prera_sched)] = .{ - .llvm_name = "ppc-prera-sched", - .description = "Use PowerPC pre-RA scheduling strategy", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.ppc4xx)] = .{ .llvm_name = "ppc4xx", .description = "Enable PPC 4xx instructions", @@ -376,6 +407,16 @@ pub const all_features = blk: { .description = "Enable PPC 6xx instructions", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.ppc_postra_sched)] = .{ + .llvm_name = "ppc-postra-sched", + .description = "Use PowerPC post-RA scheduling strategy", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.ppc_prera_sched)] = .{ + .llvm_name = "ppc-prera-sched", + .description = "Use PowerPC pre-RA scheduling strategy", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.predictable_select_expensive)] = .{ .llvm_name = "predictable-select-expensive", .description = "Prefer likely predicted branches over selects", @@ -385,18 +426,10 @@ pub const all_features = blk: { .llvm_name = "prefix-instrs", .description = "Enable prefixed instructions", .dependencies = featureSet(&[_]Feature{ - .isa_v30_instructions, .power8_vector, .power9_altivec, }), }; - result[@enumToInt(Feature.qpx)] = .{ - .llvm_name = "qpx", - .description = "Enable QPX instructions", - .dependencies = featureSet(&[_]Feature{ - .fpu, - }), - }; result[@enumToInt(Feature.recipprec)] = .{ .llvm_name = "recipprec", .description = "Assume higher precision reciprocal estimates", @@ -452,94 +485,90 @@ pub const all_features = blk: { }; pub const cpu = struct { - pub const @"ppc440" = CpuModel{ - .name = "ppc440", + pub const @"440" = CpuModel{ + .name = "440", .llvm_name = "440", .features = featureSet(&[_]Feature{ - .booke, .fres, .frsqrte, - .icbt, .isel, .msync, }), }; - pub const @"ppc450" = CpuModel{ - .name = "ppc450", + pub const @"450" = CpuModel{ + .name = "450", .llvm_name = "450", .features = featureSet(&[_]Feature{ - .booke, .fres, .frsqrte, - .icbt, .isel, .msync, }), }; - pub const @"ppc601" = CpuModel{ - .name = "ppc601", + pub const @"601" = CpuModel{ + .name = "601", .llvm_name = "601", .features = featureSet(&[_]Feature{ .fpu, }), }; - pub const @"ppc602" = CpuModel{ - .name = "ppc602", + pub const @"602" = CpuModel{ + .name = "602", .llvm_name = "602", .features = featureSet(&[_]Feature{ .fpu, }), }; - pub const @"ppc603" = CpuModel{ - .name = "ppc603", + pub const @"603" = CpuModel{ + .name = "603", .llvm_name = "603", .features = featureSet(&[_]Feature{ .fres, .frsqrte, }), }; - pub const @"ppc603e" = CpuModel{ - .name = "ppc603e", + pub const @"603e" = CpuModel{ + .name = "603e", .llvm_name = "603e", .features = featureSet(&[_]Feature{ .fres, .frsqrte, }), }; - pub const @"ppc603ev" = CpuModel{ - .name = "ppc603ev", + pub const @"603ev" = CpuModel{ + .name = "603ev", .llvm_name = "603ev", .features = featureSet(&[_]Feature{ .fres, .frsqrte, }), }; - pub const @"ppc604" = CpuModel{ - .name = "ppc604", + pub const @"604" = CpuModel{ + .name = "604", .llvm_name = "604", .features = featureSet(&[_]Feature{ .fres, .frsqrte, }), }; - pub const @"ppc604e" = CpuModel{ - .name = "ppc604e", + pub const @"604e" = CpuModel{ + .name = "604e", .llvm_name = "604e", .features = featureSet(&[_]Feature{ .fres, .frsqrte, }), }; - pub const @"ppc620" = CpuModel{ - .name = "ppc620", + pub const @"620" = CpuModel{ + .name = "620", .llvm_name = "620", .features = featureSet(&[_]Feature{ .fres, .frsqrte, }), }; - pub const @"ppc7400" = CpuModel{ - .name = "ppc7400", + pub const @"7400" = CpuModel{ + .name = "7400", .llvm_name = "7400", .features = featureSet(&[_]Feature{ .altivec, @@ -547,8 +576,8 @@ pub const cpu = struct { .frsqrte, }), }; - pub const @"ppc7450" = CpuModel{ - .name = "ppc7450", + pub const @"7450" = CpuModel{ + .name = "7450", .llvm_name = "7450", .features = featureSet(&[_]Feature{ .altivec, @@ -556,16 +585,16 @@ pub const cpu = struct { .frsqrte, }), }; - pub const @"ppc750" = CpuModel{ - .name = "ppc750", + pub const @"750" = CpuModel{ + .name = "750", .llvm_name = "750", .features = featureSet(&[_]Feature{ .fres, .frsqrte, }), }; - pub const @"ppc970" = CpuModel{ - .name = "ppc970", + pub const @"970" = CpuModel{ + .name = "970", .llvm_name = "970", .features = featureSet(&[_]Feature{ .@"64bit", @@ -592,7 +621,6 @@ pub const cpu = struct { .frsqrte, .frsqrtes, .fsqrt, - .icbt, .isel, .ldbrx, .lfiwax, @@ -602,38 +630,10 @@ pub const cpu = struct { .stfiwx, }), }; - pub const a2q = CpuModel{ - .name = "a2q", - .llvm_name = "a2q", - .features = featureSet(&[_]Feature{ - .@"64bit", - .booke, - .cmpb, - .fcpsgn, - .fpcvt, - .fprnd, - .fre, - .fres, - .frsqrte, - .frsqrtes, - .fsqrt, - .icbt, - .isel, - .ldbrx, - .lfiwax, - .mfocrf, - .qpx, - .recipprec, - .slow_popcntd, - .stfiwx, - }), - }; pub const e500 = CpuModel{ .name = "e500", .llvm_name = "e500", .features = featureSet(&[_]Feature{ - .booke, - .icbt, .isel, .msync, .spe, @@ -644,7 +644,6 @@ pub const cpu = struct { .llvm_name = "e500mc", .features = featureSet(&[_]Feature{ .booke, - .icbt, .isel, .stfiwx, }), @@ -655,7 +654,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .booke, - .icbt, .isel, .mfocrf, .stfiwx, @@ -667,7 +665,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .allow_unaligned_fp_access, - .altivec, .bpermd, .cmpb, .crypto, @@ -681,28 +678,24 @@ pub const cpu = struct { .frsqrte, .frsqrtes, .fsqrt, + .fuse_store, .htm, .icbt, - .isa_v30_instructions, - .isa_v31_instructions, .isel, .ldbrx, .lfiwax, .mfocrf, + .mma, .partword_atomics, .pcrelative_memops, .popcntd, .power10_vector, - .power8_altivec, - .power8_vector, - .power9_altivec, - .power9_vector, + .ppc_postra_sched, + .ppc_prera_sched, .predictable_select_expensive, - .prefix_instrs, .recipprec, .stfiwx, .two_const_nr, - .vsx, }), }; pub const g3 = CpuModel{ @@ -760,7 +753,7 @@ pub const cpu = struct { }; pub const ppc32 = CpuModel{ .name = "ppc32", - .llvm_name = "ppc", + .llvm_name = "ppc32", .features = featureSet(&[_]Feature{ .hard_float, }), @@ -784,7 +777,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .allow_unaligned_fp_access, - .altivec, .bpermd, .cmpb, .crypto, @@ -808,13 +800,11 @@ pub const cpu = struct { .mfocrf, .partword_atomics, .popcntd, - .power8_altivec, .power8_vector, .predictable_select_expensive, .recipprec, .stfiwx, .two_const_nr, - .vsx, }), }; pub const pwr10 = CpuModel{ @@ -823,7 +813,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .allow_unaligned_fp_access, - .altivec, .bpermd, .cmpb, .crypto, @@ -837,28 +826,24 @@ pub const cpu = struct { .frsqrte, .frsqrtes, .fsqrt, + .fuse_store, .htm, .icbt, - .isa_v30_instructions, - .isa_v31_instructions, .isel, .ldbrx, .lfiwax, .mfocrf, + .mma, .partword_atomics, .pcrelative_memops, .popcntd, .power10_vector, - .power8_altivec, - .power8_vector, - .power9_altivec, - .power9_vector, + .ppc_postra_sched, + .ppc_prera_sched, .predictable_select_expensive, - .prefix_instrs, .recipprec, .stfiwx, .two_const_nr, - .vsx, }), }; pub const pwr3 = CpuModel{ @@ -963,7 +948,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .allow_unaligned_fp_access, - .altivec, .bpermd, .cmpb, .extdiv, @@ -992,7 +976,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .allow_unaligned_fp_access, - .altivec, .bpermd, .cmpb, .crypto, @@ -1016,13 +999,11 @@ pub const cpu = struct { .mfocrf, .partword_atomics, .popcntd, - .power8_altivec, .power8_vector, .predictable_select_expensive, .recipprec, .stfiwx, .two_const_nr, - .vsx, }), }; pub const pwr9 = CpuModel{ @@ -1031,7 +1012,6 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .allow_unaligned_fp_access, - .altivec, .bpermd, .cmpb, .crypto, @@ -1047,16 +1027,12 @@ pub const cpu = struct { .fsqrt, .htm, .icbt, - .isa_v30_instructions, .isel, .ldbrx, .lfiwax, .mfocrf, .partword_atomics, .popcntd, - .power8_altivec, - .power8_vector, - .power9_altivec, .power9_vector, .ppc_postra_sched, .ppc_prera_sched, @@ -1065,7 +1041,6 @@ pub const cpu = struct { .stfiwx, .two_const_nr, .vectors_use_two_units, - .vsx, }), }; }; diff --git a/lib/std/target/riscv.zig b/lib/std/target/riscv.zig index b9eea13c87..5b19a936cf 100644 --- a/lib/std/target/riscv.zig +++ b/lib/std/target/riscv.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; @@ -15,6 +12,7 @@ pub const Feature = enum { e, experimental_b, experimental_v, + experimental_zba, experimental_zbb, experimental_zbc, experimental_zbe, @@ -25,6 +23,9 @@ pub const Feature = enum { experimental_zbr, experimental_zbs, experimental_zbt, + experimental_zfh, + experimental_zvamo, + experimental_zvlsseg, f, m, no_rvc_hints, @@ -100,6 +101,7 @@ pub const all_features = blk: { .llvm_name = "experimental-b", .description = "'B' (Bit Manipulation Instructions)", .dependencies = featureSet(&[_]Feature{ + .experimental_zba, .experimental_zbb, .experimental_zbc, .experimental_zbe, @@ -114,9 +116,12 @@ pub const all_features = blk: { result[@enumToInt(Feature.experimental_v)] = .{ .llvm_name = "experimental-v", .description = "'V' (Vector Instructions)", - .dependencies = featureSet(&[_]Feature{ - .f, - }), + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.experimental_zba)] = .{ + .llvm_name = "experimental-zba", + .description = "'Zba' (Address calculation 'B' Instructions)", + .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.experimental_zbb)] = .{ .llvm_name = "experimental-zbb", @@ -168,6 +173,27 @@ pub const all_features = blk: { .description = "'Zbt' (Ternary 'B' Instructions)", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.experimental_zfh)] = .{ + .llvm_name = "experimental-zfh", + .description = "'Zfh' (Half-Precision Floating-Point)", + .dependencies = featureSet(&[_]Feature{ + .f, + }), + }; + result[@enumToInt(Feature.experimental_zvamo)] = .{ + .llvm_name = "experimental-zvamo", + .description = "'Zvamo'(Vector AMO Operations)", + .dependencies = featureSet(&[_]Feature{ + .experimental_v, + }), + }; + result[@enumToInt(Feature.experimental_zvlsseg)] = .{ + .llvm_name = "experimental-zvlsseg", + .description = "'Zvlsseg' (Vector segment load/store instructions)", + .dependencies = featureSet(&[_]Feature{ + .experimental_v, + }), + }; result[@enumToInt(Feature.f)] = .{ .llvm_name = "f", .description = "'F' (Single-Precision Floating-Point)", @@ -364,7 +390,6 @@ pub const cpu = struct { .a, .c, .d, - .f, .m, }), }; @@ -376,7 +401,6 @@ pub const cpu = struct { .a, .c, .d, - .f, .m, }), }; @@ -404,6 +428,18 @@ pub const cpu = struct { .@"64bit", }), }; + pub const sifive_7_rv32 = CpuModel{ + .name = "sifive_7_rv32", + .llvm_name = "sifive-7-rv32", + .features = featureSet(&[_]Feature{}), + }; + pub const sifive_7_rv64 = CpuModel{ + .name = "sifive_7_rv64", + .llvm_name = "sifive-7-rv64", + .features = featureSet(&[_]Feature{ + .@"64bit", + }), + }; pub const sifive_e31 = CpuModel{ .name = "sifive_e31", .llvm_name = "sifive-e31", @@ -413,6 +449,16 @@ pub const cpu = struct { .m, }), }; + pub const sifive_e76 = CpuModel{ + .name = "sifive_e76", + .llvm_name = "sifive-e76", + .features = featureSet(&[_]Feature{ + .a, + .c, + .f, + .m, + }), + }; pub const sifive_u54 = CpuModel{ .name = "sifive_u54", .llvm_name = "sifive-u54", @@ -421,7 +467,17 @@ pub const cpu = struct { .a, .c, .d, - .f, + .m, + }), + }; + pub const sifive_u74 = CpuModel{ + .name = "sifive_u74", + .llvm_name = "sifive-u74", + .features = featureSet(&[_]Feature{ + .@"64bit", + .a, + .c, + .d, .m, }), }; diff --git a/lib/std/target/sparc.zig b/lib/std/target/sparc.zig index a075160d59..5423739be6 100644 --- a/lib/std/target/sparc.zig +++ b/lib/std/target/sparc.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; @@ -160,6 +157,11 @@ pub const cpu = struct { .llvm_name = "f934", .features = featureSet(&[_]Feature{}), }; + pub const generic = CpuModel{ + .name = "generic", + .llvm_name = "generic", + .features = featureSet(&[_]Feature{}), + }; pub const gr712rc = CpuModel{ .name = "gr712rc", .llvm_name = "gr712rc", diff --git a/lib/std/target/systemz.zig b/lib/std/target/systemz.zig index 8a78167e69..65c53984eb 100644 --- a/lib/std/target/systemz.zig +++ b/lib/std/target/systemz.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; diff --git a/lib/std/target/ve.zig b/lib/std/target/ve.zig new file mode 100644 index 0000000000..ff3eea698d --- /dev/null +++ b/lib/std/target/ve.zig @@ -0,0 +1,36 @@ +//! This file is auto-generated by tools/update_cpu_features.zig. + +const std = @import("../std.zig"); +const CpuFeature = std.Target.Cpu.Feature; +const CpuModel = std.Target.Cpu.Model; + +pub const Feature = enum { + vpu, +}; + +pub usingnamespace CpuFeature.feature_set_fns(Feature); + +pub const all_features = blk: { + const len = @typeInfo(Feature).Enum.fields.len; + std.debug.assert(len <= CpuFeature.Set.needed_bit_count); + var result: [len]CpuFeature = undefined; + result[@enumToInt(Feature.vpu)] = .{ + .llvm_name = "vpu", + .description = "Enable the VPU", + .dependencies = featureSet(&[_]Feature{}), + }; + const ti = @typeInfo(Feature); + for (result) |*elem, i| { + elem.index = i; + elem.name = ti.Enum.fields[i].name; + } + break :blk result; +}; + +pub const cpu = struct { + pub const generic = CpuModel{ + .name = "generic", + .llvm_name = "generic", + .features = featureSet(&[_]Feature{}), + }; +}; diff --git a/lib/std/target/wasm.zig b/lib/std/target/wasm.zig index 0a3281c692..4714125f30 100644 --- a/lib/std/target/wasm.zig +++ b/lib/std/target/wasm.zig @@ -1,8 +1,5 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; diff --git a/lib/std/target/x86.zig b/lib/std/target/x86.zig index abe154d509..eeb773b034 100644 --- a/lib/std/target/x86.zig +++ b/lib/std/target/x86.zig @@ -1,13 +1,12 @@ -// SPDX-License-Identifier: MIT -// Copyright (c) 2015-2021 Zig Contributors -// This file is part of [zig](https://ziglang.org/), which is MIT licensed. -// The MIT license requires this copyright notice to be included in all copies -// and substantial portions of the software. +//! This file is auto-generated by tools/update_cpu_features.zig. + const std = @import("../std.zig"); const CpuFeature = std.Target.Cpu.Feature; const CpuModel = std.Target.Cpu.Model; pub const Feature = enum { + @"16bit_mode", + @"32bit_mode", @"3dnow", @"3dnowa", @"64bit", @@ -33,6 +32,7 @@ pub const Feature = enum { avx512vnni, avx512vp2intersect, avx512vpopcntdq, + avxvnni, bmi, bmi2, branchfusion, @@ -64,11 +64,14 @@ pub const Feature = enum { fma, fma4, fsgsbase, + fsrm, fxsr, gfni, + hreset, idivl_to_divb, idivq_to_divl, invpcid, + kl, lea_sp, lea_uses_ag, lvi_cfi, @@ -76,12 +79,10 @@ pub const Feature = enum { lwp, lzcnt, macrofusion, - merge_to_threeway_branch, mmx, movbe, movdir64b, movdiri, - mpx, mwaitx, nopl, pad_short_functions, @@ -120,15 +121,16 @@ pub const Feature = enum { slow_unaligned_mem_32, soft_float, sse, - sse_unaligned_mem, sse2, sse3, sse4_1, sse4_2, sse4a, + sse_unaligned_mem, ssse3, tbm, tsxldtrk, + uintr, use_aa, use_glm_div_sqrt_costs, vaes, @@ -136,6 +138,7 @@ pub const Feature = enum { vzeroupper, waitpkg, wbnoinvd, + widekl, x87, xop, xsave, @@ -150,6 +153,16 @@ pub const all_features = blk: { const len = @typeInfo(Feature).Enum.fields.len; std.debug.assert(len <= CpuFeature.Set.needed_bit_count); var result: [len]CpuFeature = undefined; + result[@enumToInt(Feature.@"16bit_mode")] = .{ + .llvm_name = "16bit-mode", + .description = "16-bit mode (i8086)", + .dependencies = featureSet(&[_]Feature{}), + }; + result[@enumToInt(Feature.@"32bit_mode")] = .{ + .llvm_name = "32bit-mode", + .description = "32-bit mode (80386)", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.@"3dnow")] = .{ .llvm_name = "3dnow", .description = "Enable 3DNow! instructions", @@ -321,6 +334,13 @@ pub const all_features = blk: { .avx512f, }), }; + result[@enumToInt(Feature.avxvnni)] = .{ + .llvm_name = "avxvnni", + .description = "Support AVX_VNNI encoding", + .dependencies = featureSet(&[_]Feature{ + .avx2, + }), + }; result[@enumToInt(Feature.bmi)] = .{ .llvm_name = "bmi", .description = "Support BMI instructions", @@ -485,6 +505,11 @@ pub const all_features = blk: { .description = "Support FS/GS Base instructions", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.fsrm)] = .{ + .llvm_name = "fsrm", + .description = "REP MOVSB of short lengths is faster", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.fxsr)] = .{ .llvm_name = "fxsr", .description = "Support fxsave/fxrestore instructions", @@ -497,6 +522,11 @@ pub const all_features = blk: { .sse2, }), }; + result[@enumToInt(Feature.hreset)] = .{ + .llvm_name = "hreset", + .description = "Has hreset instruction", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.idivl_to_divb)] = .{ .llvm_name = "idivl-to-divb", .description = "Use 8-bit divide for positive values less than 256", @@ -512,6 +542,13 @@ pub const all_features = blk: { .description = "Invalidate Process-Context Identifier", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.kl)] = .{ + .llvm_name = "kl", + .description = "Support Key Locker kl Instructions", + .dependencies = featureSet(&[_]Feature{ + .sse2, + }), + }; result[@enumToInt(Feature.lea_sp)] = .{ .llvm_name = "lea-sp", .description = "Use LEA for adjusting the stack pointer", @@ -547,11 +584,6 @@ pub const all_features = blk: { .description = "Various instructions can be fused with conditional branches", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.merge_to_threeway_branch)] = .{ - .llvm_name = "merge-to-threeway-branch", - .description = "Merge branches to a three-way conditional branch", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.mmx)] = .{ .llvm_name = "mmx", .description = "Enable MMX instructions", @@ -572,11 +604,6 @@ pub const all_features = blk: { .description = "Support movdiri instruction", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.mpx)] = .{ - .llvm_name = "mpx", - .description = "Deprecated. Support MPX instructions", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.mwaitx)] = .{ .llvm_name = "mwaitx", .description = "Enable MONITORX/MWAITX timer functionality", @@ -691,7 +718,7 @@ pub const all_features = blk: { }; result[@enumToInt(Feature.sahf)] = .{ .llvm_name = "sahf", - .description = "Support LAHF and SAHF instructions", + .description = "Support LAHF and SAHF instructions in 64-bit mode", .dependencies = featureSet(&[_]Feature{}), }; result[@enumToInt(Feature.serialize)] = .{ @@ -778,11 +805,6 @@ pub const all_features = blk: { .description = "Enable SSE instructions", .dependencies = featureSet(&[_]Feature{}), }; - result[@enumToInt(Feature.sse_unaligned_mem)] = .{ - .llvm_name = "sse-unaligned-mem", - .description = "Allow unaligned memory operands with SSE instructions", - .dependencies = featureSet(&[_]Feature{}), - }; result[@enumToInt(Feature.sse2)] = .{ .llvm_name = "sse2", .description = "Enable SSE2 instructions", @@ -818,6 +840,11 @@ pub const all_features = blk: { .sse3, }), }; + result[@enumToInt(Feature.sse_unaligned_mem)] = .{ + .llvm_name = "sse-unaligned-mem", + .description = "Allow unaligned memory operands with SSE instructions", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.ssse3)] = .{ .llvm_name = "ssse3", .description = "Enable SSSE3 instructions", @@ -835,6 +862,11 @@ pub const all_features = blk: { .description = "Support TSXLDTRK instructions", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.uintr)] = .{ + .llvm_name = "uintr", + .description = "Has UINTR Instructions", + .dependencies = featureSet(&[_]Feature{}), + }; result[@enumToInt(Feature.use_aa)] = .{ .llvm_name = "use-aa", .description = "Use alias analysis during codegen", @@ -876,6 +908,13 @@ pub const all_features = blk: { .description = "Write Back No Invalidate", .dependencies = featureSet(&[_]Feature{}), }; + result[@enumToInt(Feature.widekl)] = .{ + .llvm_name = "widekl", + .description = "Support Key Locker wide Instructions", + .dependencies = featureSet(&[_]Feature{ + .kl, + }), + }; result[@enumToInt(Feature.x87)] = .{ .llvm_name = "x87", .description = "Enable X87 float instructions", @@ -923,6 +962,97 @@ pub const all_features = blk: { }; pub const cpu = struct { + pub const _i386 = CpuModel{ + .name = "_i386", + .llvm_name = "i386", + .features = featureSet(&[_]Feature{ + .slow_unaligned_mem_16, + .vzeroupper, + .x87, + }), + }; + pub const _i486 = CpuModel{ + .name = "_i486", + .llvm_name = "i486", + .features = featureSet(&[_]Feature{ + .slow_unaligned_mem_16, + .vzeroupper, + .x87, + }), + }; + pub const _i586 = CpuModel{ + .name = "_i586", + .llvm_name = "i586", + .features = featureSet(&[_]Feature{ + .cx8, + .slow_unaligned_mem_16, + .vzeroupper, + .x87, + }), + }; + pub const _i686 = CpuModel{ + .name = "_i686", + .llvm_name = "i686", + .features = featureSet(&[_]Feature{ + .cmov, + .cx8, + .slow_unaligned_mem_16, + .vzeroupper, + .x87, + }), + }; + pub const alderlake = CpuModel{ + .name = "alderlake", + .llvm_name = "alderlake", + .features = featureSet(&[_]Feature{ + .@"64bit", + .adx, + .aes, + .avxvnni, + .bmi, + .bmi2, + .cldemote, + .clflushopt, + .cmov, + .cx16, + .ermsb, + .f16c, + .false_deps_popcnt, + .fast_15bytenop, + .fast_gather, + .fast_scalar_fsqrt, + .fast_shld_rotate, + .fast_variable_shuffle, + .fast_vector_fsqrt, + .fma, + .fsgsbase, + .fxsr, + .hreset, + .idivq_to_divl, + .invpcid, + .lzcnt, + .macrofusion, + .mmx, + .movbe, + .nopl, + .pclmul, + .popcnt, + .prfchw, + .ptwrite, + .rdrnd, + .rdseed, + .sahf, + .serialize, + .sgx, + .slow_3ops_lea, + .vzeroupper, + .waitpkg, + .x87, + .xsavec, + .xsaveopt, + .xsaves, + }), + }; pub const amdfam10 = CpuModel{ .name = "amdfam10", .llvm_name = "amdfam10", @@ -931,7 +1061,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fast_scalar_shift_masks, .fxsr, .lzcnt, @@ -959,43 +1088,45 @@ pub const cpu = struct { .x87, }), }; - pub const athlon_4 = CpuModel{ - .name = "athlon_4", - .llvm_name = "athlon-4", + pub const athlon64 = CpuModel{ + .name = "athlon64", + .llvm_name = "athlon64", .features = featureSet(&[_]Feature{ .@"3dnowa", + .@"64bit", .cmov, .cx8, + .fast_scalar_shift_masks, .fxsr, .nopl, .slow_shld, .slow_unaligned_mem_16, - .sse, + .sse2, .vzeroupper, .x87, }), }; - pub const athlon_fx = CpuModel{ - .name = "athlon_fx", - .llvm_name = "athlon-fx", + pub const athlon64_sse3 = CpuModel{ + .name = "athlon64_sse3", + .llvm_name = "athlon64-sse3", .features = featureSet(&[_]Feature{ .@"3dnowa", .@"64bit", .cmov, - .cx8, + .cx16, .fast_scalar_shift_masks, .fxsr, .nopl, .slow_shld, .slow_unaligned_mem_16, - .sse2, + .sse3, .vzeroupper, .x87, }), }; - pub const athlon_mp = CpuModel{ - .name = "athlon_mp", - .llvm_name = "athlon-mp", + pub const athlon_4 = CpuModel{ + .name = "athlon_4", + .llvm_name = "athlon-4", .features = featureSet(&[_]Feature{ .@"3dnowa", .cmov, @@ -1009,23 +1140,27 @@ pub const cpu = struct { .x87, }), }; - pub const athlon_tbird = CpuModel{ - .name = "athlon_tbird", - .llvm_name = "athlon-tbird", + pub const athlon_fx = CpuModel{ + .name = "athlon_fx", + .llvm_name = "athlon-fx", .features = featureSet(&[_]Feature{ .@"3dnowa", + .@"64bit", .cmov, .cx8, + .fast_scalar_shift_masks, + .fxsr, .nopl, .slow_shld, .slow_unaligned_mem_16, + .sse2, .vzeroupper, .x87, }), }; - pub const athlon_xp = CpuModel{ - .name = "athlon_xp", - .llvm_name = "athlon-xp", + pub const athlon_mp = CpuModel{ + .name = "athlon_mp", + .llvm_name = "athlon-mp", .features = featureSet(&[_]Feature{ .@"3dnowa", .cmov, @@ -1039,39 +1174,32 @@ pub const cpu = struct { .x87, }), }; - pub const athlon64 = CpuModel{ - .name = "athlon64", - .llvm_name = "athlon64", + pub const athlon_tbird = CpuModel{ + .name = "athlon_tbird", + .llvm_name = "athlon-tbird", .features = featureSet(&[_]Feature{ .@"3dnowa", - .@"64bit", .cmov, .cx8, - .fast_scalar_shift_masks, - .fxsr, .nopl, .slow_shld, .slow_unaligned_mem_16, - .sse2, .vzeroupper, .x87, }), }; - pub const athlon64_sse3 = CpuModel{ - .name = "athlon64_sse3", - .llvm_name = "athlon64-sse3", + pub const athlon_xp = CpuModel{ + .name = "athlon_xp", + .llvm_name = "athlon-xp", .features = featureSet(&[_]Feature{ .@"3dnowa", - .@"64bit", .cmov, - .cx16, .cx8, - .fast_scalar_shift_masks, .fxsr, .nopl, .slow_shld, .slow_unaligned_mem_16, - .sse3, + .sse, .vzeroupper, .x87, }), @@ -1083,7 +1211,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fxsr, .idivl_to_divb, .idivq_to_divl, @@ -1109,7 +1236,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fast_scalar_shift_masks, .fxsr, .lzcnt, @@ -1132,7 +1258,6 @@ pub const cpu = struct { .branchfusion, .cmov, .cx16, - .cx8, .fast_11bytenop, .fast_scalar_shift_masks, .fxsr, @@ -1161,7 +1286,6 @@ pub const cpu = struct { .branchfusion, .cmov, .cx16, - .cx8, .f16c, .fast_11bytenop, .fast_bextr, @@ -1194,7 +1318,6 @@ pub const cpu = struct { .branchfusion, .cmov, .cx16, - .cx8, .f16c, .fast_11bytenop, .fast_bextr, @@ -1215,7 +1338,6 @@ pub const cpu = struct { .vzeroupper, .x87, .xop, - .xsave, .xsaveopt, }), }; @@ -1231,7 +1353,6 @@ pub const cpu = struct { .branchfusion, .cmov, .cx16, - .cx8, .f16c, .fast_11bytenop, .fast_bextr, @@ -1255,7 +1376,6 @@ pub const cpu = struct { .vzeroupper, .x87, .xop, - .xsave, .xsaveopt, }), }; @@ -1266,7 +1386,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fxsr, .idivl_to_divb, .idivq_to_divl, @@ -1290,13 +1409,11 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .adx, - .avx, .avx2, .bmi, .bmi2, .cmov, .cx16, - .cx8, .ermsb, .f16c, .false_deps_lzcnt_tzcnt, @@ -1312,7 +1429,6 @@ pub const cpu = struct { .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -1323,10 +1439,8 @@ pub const cpu = struct { .rdseed, .sahf, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsaveopt, }), }; @@ -1337,7 +1451,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fast_15bytenop, .fast_scalar_shift_masks, .fast_vector_shift_masks, @@ -1361,11 +1474,9 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .aes, - .avx, .bmi, .cmov, .cx16, - .cx8, .f16c, .fast_15bytenop, .fast_bextr, @@ -1384,9 +1495,7 @@ pub const cpu = struct { .sahf, .slow_shld, .sse4a, - .ssse3, .x87, - .xsave, .xsaveopt, }), }; @@ -1421,12 +1530,8 @@ pub const cpu = struct { .@"64bit", .adx, .aes, - .avx, - .avx2, - .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512ifma, .avx512vbmi, .avx512vl, @@ -1435,23 +1540,19 @@ pub const cpu = struct { .clflushopt, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .fast_15bytenop, .fast_gather, .fast_scalar_fsqrt, .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, .fxsr, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -1466,10 +1567,8 @@ pub const cpu = struct { .sgx, .sha, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -1482,12 +1581,9 @@ pub const cpu = struct { .@"64bit", .adx, .aes, - .avx, - .avx2, .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512vl, .avx512vnni, .bmi, @@ -1496,9 +1592,7 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .false_deps_popcnt, .fast_15bytenop, .fast_gather, @@ -1506,14 +1600,12 @@ pub const cpu = struct { .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, .fxsr, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -1526,10 +1618,8 @@ pub const cpu = struct { .rdseed, .sahf, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -1542,13 +1632,9 @@ pub const cpu = struct { .@"64bit", .adx, .aes, - .avx, - .avx2, .avx512bf16, - .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512vl, .avx512vnni, .bmi, @@ -1557,9 +1643,7 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .false_deps_popcnt, .fast_15bytenop, .fast_gather, @@ -1567,14 +1651,12 @@ pub const cpu = struct { .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, .fxsr, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -1587,47 +1669,29 @@ pub const cpu = struct { .rdseed, .sahf, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, }), }; - pub const core_avx_i = CpuModel{ - .name = "core_avx_i", - .llvm_name = "core-avx-i", + pub const core2 = CpuModel{ + .name = "core2", + .llvm_name = "core2", .features = featureSet(&[_]Feature{ .@"64bit", - .avx, .cmov, .cx16, - .cx8, - .f16c, - .false_deps_popcnt, - .fast_15bytenop, - .fast_scalar_fsqrt, - .fast_shld_rotate, - .fsgsbase, .fxsr, - .idivq_to_divl, .macrofusion, - .merge_to_threeway_branch, .mmx, .nopl, - .pclmul, - .popcnt, - .rdrnd, .sahf, - .slow_3ops_lea, - .slow_unaligned_mem_32, - .sse4_2, + .slow_unaligned_mem_16, + .ssse3, .vzeroupper, .x87, - .xsave, - .xsaveopt, }), }; pub const core_avx2 = CpuModel{ @@ -1635,13 +1699,11 @@ pub const cpu = struct { .llvm_name = "core-avx2", .features = featureSet(&[_]Feature{ .@"64bit", - .avx, .avx2, .bmi, .bmi2, .cmov, .cx16, - .cx8, .ermsb, .f16c, .false_deps_lzcnt_tzcnt, @@ -1657,7 +1719,6 @@ pub const cpu = struct { .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -1666,30 +1727,38 @@ pub const cpu = struct { .rdrnd, .sahf, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsaveopt, }), }; - pub const core2 = CpuModel{ - .name = "core2", - .llvm_name = "core2", + pub const core_avx_i = CpuModel{ + .name = "core_avx_i", + .llvm_name = "core-avx-i", .features = featureSet(&[_]Feature{ .@"64bit", .cmov, .cx16, - .cx8, + .f16c, + .false_deps_popcnt, + .fast_15bytenop, + .fast_scalar_fsqrt, + .fast_shld_rotate, + .fsgsbase, .fxsr, + .idivq_to_divl, .macrofusion, .mmx, .nopl, + .pclmul, + .popcnt, + .rdrnd, .sahf, - .slow_unaligned_mem_16, - .ssse3, + .slow_3ops_lea, + .slow_unaligned_mem_32, .vzeroupper, .x87, + .xsaveopt, }), }; pub const corei7 = CpuModel{ @@ -1699,7 +1768,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fxsr, .macrofusion, .mmx, @@ -1719,7 +1787,6 @@ pub const cpu = struct { .avx, .cmov, .cx16, - .cx8, .false_deps_popcnt, .fast_15bytenop, .fast_scalar_fsqrt, @@ -1727,7 +1794,6 @@ pub const cpu = struct { .fxsr, .idivq_to_divl, .macrofusion, - .merge_to_threeway_branch, .mmx, .nopl, .pclmul, @@ -1735,10 +1801,8 @@ pub const cpu = struct { .sahf, .slow_3ops_lea, .slow_unaligned_mem_32, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsaveopt, }), }; @@ -1746,8 +1810,12 @@ pub const cpu = struct { .name = "generic", .llvm_name = "generic", .features = featureSet(&[_]Feature{ + .@"64bit", .cx8, - .slow_unaligned_mem_16, + .idivq_to_divl, + .macrofusion, + .slow_3ops_lea, + .slow_incdec, .vzeroupper, .x87, }), @@ -1772,7 +1840,6 @@ pub const cpu = struct { .clflushopt, .cmov, .cx16, - .cx8, .false_deps_popcnt, .fsgsbase, .fxsr, @@ -1790,11 +1857,9 @@ pub const cpu = struct { .slow_lea, .slow_two_mem_ops, .sse4_2, - .ssse3, .use_glm_div_sqrt_costs, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -1809,7 +1874,6 @@ pub const cpu = struct { .clflushopt, .cmov, .cx16, - .cx8, .fsgsbase, .fxsr, .mmx, @@ -1829,11 +1893,9 @@ pub const cpu = struct { .slow_lea, .slow_two_mem_ops, .sse4_2, - .ssse3, .use_glm_div_sqrt_costs, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -1844,13 +1906,11 @@ pub const cpu = struct { .llvm_name = "haswell", .features = featureSet(&[_]Feature{ .@"64bit", - .avx, .avx2, .bmi, .bmi2, .cmov, .cx16, - .cx8, .ermsb, .f16c, .false_deps_lzcnt_tzcnt, @@ -1866,7 +1926,6 @@ pub const cpu = struct { .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -1875,66 +1934,20 @@ pub const cpu = struct { .rdrnd, .sahf, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsaveopt, }), }; - pub const _i386 = CpuModel{ - .name = "_i386", - .llvm_name = "i386", - .features = featureSet(&[_]Feature{ - .slow_unaligned_mem_16, - .vzeroupper, - .x87, - }), - }; - pub const _i486 = CpuModel{ - .name = "_i486", - .llvm_name = "i486", - .features = featureSet(&[_]Feature{ - .slow_unaligned_mem_16, - .vzeroupper, - .x87, - }), - }; - pub const _i586 = CpuModel{ - .name = "_i586", - .llvm_name = "i586", - .features = featureSet(&[_]Feature{ - .cx8, - .slow_unaligned_mem_16, - .vzeroupper, - .x87, - }), - }; - pub const _i686 = CpuModel{ - .name = "_i686", - .llvm_name = "i686", - .features = featureSet(&[_]Feature{ - .cmov, - .cx8, - .slow_unaligned_mem_16, - .vzeroupper, - .x87, - }), - }; pub const icelake_client = CpuModel{ .name = "icelake_client", .llvm_name = "icelake-client", .features = featureSet(&[_]Feature{ .@"64bit", .adx, - .aes, - .avx, - .avx2, .avx512bitalg, - .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512ifma, .avx512vbmi, .avx512vbmi2, @@ -1947,28 +1960,24 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .fast_15bytenop, .fast_gather, .fast_scalar_fsqrt, .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, + .fsrm, .fxsr, .gfni, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, - .pclmul, .pku, .popcnt, .prefer_256_bit, @@ -1980,12 +1989,10 @@ pub const cpu = struct { .sgx, .sha, .slow_3ops_lea, - .sse4_2, .vaes, .vpclmulqdq, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -1997,14 +2004,9 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .adx, - .aes, - .avx, - .avx2, .avx512bitalg, - .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512ifma, .avx512vbmi, .avx512vbmi2, @@ -2017,28 +2019,24 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .fast_15bytenop, .fast_gather, .fast_scalar_fsqrt, .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, + .fsrm, .fxsr, .gfni, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, - .pclmul, .pconfig, .pku, .popcnt, @@ -2051,13 +2049,11 @@ pub const cpu = struct { .sgx, .sha, .slow_3ops_lea, - .sse4_2, .vaes, .vpclmulqdq, .vzeroupper, .wbnoinvd, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -2068,10 +2064,8 @@ pub const cpu = struct { .llvm_name = "ivybridge", .features = featureSet(&[_]Feature{ .@"64bit", - .avx, .cmov, .cx16, - .cx8, .f16c, .false_deps_popcnt, .fast_15bytenop, @@ -2081,7 +2075,6 @@ pub const cpu = struct { .fxsr, .idivq_to_divl, .macrofusion, - .merge_to_threeway_branch, .mmx, .nopl, .pclmul, @@ -2090,10 +2083,8 @@ pub const cpu = struct { .sahf, .slow_3ops_lea, .slow_unaligned_mem_32, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsaveopt, }), }; @@ -2156,7 +2147,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fast_scalar_shift_masks, .fxsr, .nopl, @@ -2176,16 +2166,12 @@ pub const cpu = struct { .aes, .avx512cd, .avx512er, - .avx512f, .avx512pf, .bmi, .bmi2, .cmov, .cx16, - .cx8, - .f16c, .fast_gather, - .fma, .fsgsbase, .fxsr, .idivq_to_divl, @@ -2206,7 +2192,6 @@ pub const cpu = struct { .slow_pmaddwd, .slow_two_mem_ops, .x87, - .xsave, .xsaveopt, }), }; @@ -2219,17 +2204,13 @@ pub const cpu = struct { .aes, .avx512cd, .avx512er, - .avx512f, .avx512pf, .avx512vpopcntdq, .bmi, .bmi2, .cmov, .cx16, - .cx8, - .f16c, .fast_gather, - .fma, .fsgsbase, .fxsr, .idivq_to_divl, @@ -2250,7 +2231,6 @@ pub const cpu = struct { .slow_pmaddwd, .slow_two_mem_ops, .x87, - .xsave, .xsaveopt, }), }; @@ -2258,6 +2238,8 @@ pub const cpu = struct { .name = "lakemont", .llvm_name = "lakemont", .features = featureSet(&[_]Feature{ + .cx8, + .slow_unaligned_mem_16, .vzeroupper, }), }; @@ -2268,7 +2250,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fxsr, .macrofusion, .mmx, @@ -2287,7 +2268,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fxsr, .mmx, .nopl, @@ -2323,7 +2303,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fast_scalar_shift_masks, .fxsr, .nopl, @@ -2341,7 +2320,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fxsr, .macrofusion, .mmx, @@ -2363,32 +2341,6 @@ pub const cpu = struct { .x87, }), }; - pub const pentium_m = CpuModel{ - .name = "pentium_m", - .llvm_name = "pentium-m", - .features = featureSet(&[_]Feature{ - .cmov, - .cx8, - .fxsr, - .mmx, - .nopl, - .slow_unaligned_mem_16, - .sse2, - .vzeroupper, - .x87, - }), - }; - pub const pentium_mmx = CpuModel{ - .name = "pentium_mmx", - .llvm_name = "pentium-mmx", - .features = featureSet(&[_]Feature{ - .cx8, - .mmx, - .slow_unaligned_mem_16, - .vzeroupper, - .x87, - }), - }; pub const pentium2 = CpuModel{ .name = "pentium2", .llvm_name = "pentium2", @@ -2463,6 +2415,32 @@ pub const cpu = struct { .x87, }), }; + pub const pentium_m = CpuModel{ + .name = "pentium_m", + .llvm_name = "pentium-m", + .features = featureSet(&[_]Feature{ + .cmov, + .cx8, + .fxsr, + .mmx, + .nopl, + .slow_unaligned_mem_16, + .sse2, + .vzeroupper, + .x87, + }), + }; + pub const pentium_mmx = CpuModel{ + .name = "pentium_mmx", + .llvm_name = "pentium-mmx", + .features = featureSet(&[_]Feature{ + .cx8, + .mmx, + .slow_unaligned_mem_16, + .vzeroupper, + .x87, + }), + }; pub const pentiumpro = CpuModel{ .name = "pentiumpro", .llvm_name = "pentiumpro", @@ -2498,7 +2476,6 @@ pub const cpu = struct { .avx, .cmov, .cx16, - .cx8, .false_deps_popcnt, .fast_15bytenop, .fast_scalar_fsqrt, @@ -2506,7 +2483,6 @@ pub const cpu = struct { .fxsr, .idivq_to_divl, .macrofusion, - .merge_to_threeway_branch, .mmx, .nopl, .pclmul, @@ -2514,13 +2490,87 @@ pub const cpu = struct { .sahf, .slow_3ops_lea, .slow_unaligned_mem_32, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsaveopt, }), }; + pub const sapphirerapids = CpuModel{ + .name = "sapphirerapids", + .llvm_name = "sapphirerapids", + .features = featureSet(&[_]Feature{ + .@"64bit", + .adx, + .amx_bf16, + .amx_int8, + .avx512bf16, + .avx512bitalg, + .avx512cd, + .avx512dq, + .avx512ifma, + .avx512vbmi, + .avx512vbmi2, + .avx512vl, + .avx512vnni, + .avx512vp2intersect, + .avx512vpopcntdq, + .avxvnni, + .bmi, + .bmi2, + .cldemote, + .clflushopt, + .clwb, + .cmov, + .cx16, + .enqcmd, + .ermsb, + .fast_15bytenop, + .fast_gather, + .fast_scalar_fsqrt, + .fast_shld_rotate, + .fast_variable_shuffle, + .fast_vector_fsqrt, + .fsgsbase, + .fsrm, + .fxsr, + .gfni, + .idivq_to_divl, + .invpcid, + .lzcnt, + .macrofusion, + .mmx, + .movbe, + .movdir64b, + .movdiri, + .nopl, + .pconfig, + .pku, + .popcnt, + .prefer_256_bit, + .prfchw, + .ptwrite, + .rdpid, + .rdrnd, + .rdseed, + .sahf, + .serialize, + .sgx, + .sha, + .shstk, + .slow_3ops_lea, + .tsxldtrk, + .uintr, + .vaes, + .vpclmulqdq, + .vzeroupper, + .waitpkg, + .wbnoinvd, + .x87, + .xsavec, + .xsaveopt, + .xsaves, + }), + }; pub const silvermont = CpuModel{ .name = "silvermont", .llvm_name = "silvermont", @@ -2528,7 +2578,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .false_deps_popcnt, .fast_7bytenop, .fxsr, @@ -2546,7 +2595,6 @@ pub const cpu = struct { .slow_pmulld, .slow_two_mem_ops, .sse4_2, - .ssse3, .vzeroupper, .x87, }), @@ -2558,12 +2606,9 @@ pub const cpu = struct { .@"64bit", .adx, .aes, - .avx, - .avx2, .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512vl, .bmi, .bmi2, @@ -2571,9 +2616,7 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .false_deps_popcnt, .fast_15bytenop, .fast_gather, @@ -2581,14 +2624,12 @@ pub const cpu = struct { .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, .fxsr, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -2601,10 +2642,8 @@ pub const cpu = struct { .rdseed, .sahf, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -2617,14 +2656,12 @@ pub const cpu = struct { .@"64bit", .adx, .aes, - .avx, .avx2, .bmi, .bmi2, .clflushopt, .cmov, .cx16, - .cx8, .ermsb, .f16c, .false_deps_popcnt, @@ -2641,7 +2678,6 @@ pub const cpu = struct { .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -2653,10 +2689,8 @@ pub const cpu = struct { .sahf, .sgx, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -2669,12 +2703,9 @@ pub const cpu = struct { .@"64bit", .adx, .aes, - .avx, - .avx2, .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512vl, .bmi, .bmi2, @@ -2682,9 +2713,7 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .false_deps_popcnt, .fast_15bytenop, .fast_gather, @@ -2692,14 +2721,12 @@ pub const cpu = struct { .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, .fxsr, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .nopl, @@ -2712,10 +2739,8 @@ pub const cpu = struct { .rdseed, .sahf, .slow_3ops_lea, - .sse4_2, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -2728,7 +2753,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .false_deps_popcnt, .fast_7bytenop, .fxsr, @@ -2746,7 +2770,6 @@ pub const cpu = struct { .slow_pmulld, .slow_two_mem_ops, .sse4_2, - .ssse3, .vzeroupper, .x87, }), @@ -2757,14 +2780,9 @@ pub const cpu = struct { .features = featureSet(&[_]Feature{ .@"64bit", .adx, - .aes, - .avx, - .avx2, .avx512bitalg, - .avx512bw, .avx512cd, .avx512dq, - .avx512f, .avx512ifma, .avx512vbmi, .avx512vbmi2, @@ -2778,30 +2796,26 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .ermsb, - .f16c, .fast_15bytenop, .fast_gather, .fast_scalar_fsqrt, .fast_shld_rotate, .fast_variable_shuffle, .fast_vector_fsqrt, - .fma, .fsgsbase, + .fsrm, .fxsr, .gfni, .idivq_to_divl, .invpcid, .lzcnt, .macrofusion, - .merge_to_threeway_branch, .mmx, .movbe, .movdir64b, .movdiri, .nopl, - .pclmul, .pku, .popcnt, .prefer_256_bit, @@ -2814,12 +2828,10 @@ pub const cpu = struct { .sha, .shstk, .slow_3ops_lea, - .sse4_2, .vaes, .vpclmulqdq, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -2835,7 +2847,6 @@ pub const cpu = struct { .clwb, .cmov, .cx16, - .cx8, .fsgsbase, .fxsr, .gfni, @@ -2856,11 +2867,9 @@ pub const cpu = struct { .slow_lea, .slow_two_mem_ops, .sse4_2, - .ssse3, .use_glm_div_sqrt_costs, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -2873,7 +2882,6 @@ pub const cpu = struct { .@"64bit", .cmov, .cx16, - .cx8, .fxsr, .macrofusion, .mmx, @@ -2886,21 +2894,21 @@ pub const cpu = struct { .x87, }), }; - pub const winchip_c6 = CpuModel{ - .name = "winchip_c6", - .llvm_name = "winchip-c6", + pub const winchip2 = CpuModel{ + .name = "winchip2", + .llvm_name = "winchip2", .features = featureSet(&[_]Feature{ - .mmx, + .@"3dnow", .slow_unaligned_mem_16, .vzeroupper, .x87, }), }; - pub const winchip2 = CpuModel{ - .name = "winchip2", - .llvm_name = "winchip2", + pub const winchip_c6 = CpuModel{ + .name = "winchip_c6", + .llvm_name = "winchip-c6", .features = featureSet(&[_]Feature{ - .@"3dnow", + .mmx, .slow_unaligned_mem_16, .vzeroupper, .x87, @@ -2925,6 +2933,100 @@ pub const cpu = struct { .x87, }), }; + pub const x86_64_v2 = CpuModel{ + .name = "x86_64_v2", + .llvm_name = "x86-64-v2", + .features = featureSet(&[_]Feature{ + .@"64bit", + .cmov, + .cx16, + .false_deps_popcnt, + .fast_15bytenop, + .fast_scalar_fsqrt, + .fast_shld_rotate, + .fxsr, + .idivq_to_divl, + .macrofusion, + .mmx, + .nopl, + .popcnt, + .sahf, + .slow_3ops_lea, + .slow_unaligned_mem_32, + .sse4_2, + .vzeroupper, + .x87, + }), + }; + pub const x86_64_v3 = CpuModel{ + .name = "x86_64_v3", + .llvm_name = "x86-64-v3", + .features = featureSet(&[_]Feature{ + .@"64bit", + .avx2, + .bmi, + .bmi2, + .cmov, + .cx16, + .f16c, + .false_deps_lzcnt_tzcnt, + .false_deps_popcnt, + .fast_15bytenop, + .fast_scalar_fsqrt, + .fast_shld_rotate, + .fast_variable_shuffle, + .fma, + .fxsr, + .idivq_to_divl, + .lzcnt, + .macrofusion, + .mmx, + .movbe, + .nopl, + .popcnt, + .sahf, + .slow_3ops_lea, + .vzeroupper, + .x87, + .xsave, + }), + }; + pub const x86_64_v4 = CpuModel{ + .name = "x86_64_v4", + .llvm_name = "x86-64-v4", + .features = featureSet(&[_]Feature{ + .@"64bit", + .avx512bw, + .avx512cd, + .avx512dq, + .avx512vl, + .bmi, + .bmi2, + .cmov, + .cx16, + .false_deps_popcnt, + .fast_15bytenop, + .fast_gather, + .fast_scalar_fsqrt, + .fast_shld_rotate, + .fast_variable_shuffle, + .fast_vector_fsqrt, + .fxsr, + .idivq_to_divl, + .lzcnt, + .macrofusion, + .mmx, + .movbe, + .nopl, + .popcnt, + .prefer_256_bit, + .sahf, + .slow_3ops_lea, + .vzeroupper, + .x87, + .xsave, + }), + }; pub const yonah = CpuModel{ .name = "yonah", .llvm_name = "yonah", @@ -2979,7 +3081,6 @@ pub const cpu = struct { .sse4a, .vzeroupper, .x87, - .xsave, .xsavec, .xsaveopt, .xsaves, @@ -3027,7 +3128,56 @@ pub const cpu = struct { .vzeroupper, .wbnoinvd, .x87, - .xsave, + .xsavec, + .xsaveopt, + .xsaves, + }), + }; + pub const znver3 = CpuModel{ + .name = "znver3", + .llvm_name = "znver3", + .features = featureSet(&[_]Feature{ + .@"64bit", + .adx, + .avx2, + .bmi, + .bmi2, + .branchfusion, + .clflushopt, + .clwb, + .clzero, + .cmov, + .cx16, + .f16c, + .fast_15bytenop, + .fast_bextr, + .fast_lzcnt, + .fast_scalar_shift_masks, + .fma, + .fsgsbase, + .fsrm, + .fxsr, + .invpcid, + .lzcnt, + .mmx, + .movbe, + .mwaitx, + .nopl, + .pku, + .popcnt, + .prfchw, + .rdpid, + .rdrnd, + .rdseed, + .sahf, + .sha, + .slow_shld, + .sse4a, + .vaes, + .vpclmulqdq, + .vzeroupper, + .wbnoinvd, + .x87, .xsavec, .xsaveopt, .xsaves, diff --git a/lib/std/zig/cross_target.zig b/lib/std/zig/cross_target.zig index c34dcc2bd3..8d6f63f5e3 100644 --- a/lib/std/zig/cross_target.zig +++ b/lib/std/zig/cross_target.zig @@ -111,11 +111,11 @@ pub const CrossTarget = struct { .kfreebsd, .lv2, .solaris, + .zos, .haiku, .minix, .rtems, .nacl, - .cnk, .aix, .cuda, .nvcl, @@ -714,11 +714,11 @@ pub const CrossTarget = struct { .kfreebsd, .lv2, .solaris, + .zos, .haiku, .minix, .rtems, .nacl, - .cnk, .aix, .cuda, .nvcl, diff --git a/lib/std/zig/system.zig b/lib/std/zig/system.zig index 2d9f286dd6..485bd479bf 100644 --- a/lib/std/zig/system.zig +++ b/lib/std/zig/system.zig @@ -326,11 +326,33 @@ pub const NativeTargetInfo = struct { cpu_detection_unimplemented = true; break :backup_cpu_detection Target.Cpu.baseline(cpu_arch); }; - cross_target.updateCpuFeatures(&cpu.features); - - var target = try detectAbiAndDynamicLinker(allocator, cpu, os, cross_target); - target.cpu_detection_unimplemented = cpu_detection_unimplemented; - return target; + var result = try detectAbiAndDynamicLinker(allocator, cpu, os, cross_target); + // For x86, we need to populate some CPU feature flags depending on architecture + // and mode: + // * 16bit_mode => if the abi is code16 + // * 32bit_mode => if the arch is i386 + // However, the "mode" flags can be used as overrides, so if the user explicitly + // sets one of them, that takes precedence. + switch (cpu_arch) { + .i386 => { + if (!std.Target.x86.featureSetHasAny(cross_target.cpu_features_add, .{ + .@"16bit_mode", .@"32bit_mode", + })) { + switch (result.target.abi) { + .code16 => result.target.cpu.features.addFeature( + @enumToInt(std.Target.x86.Feature.@"16bit_mode"), + ), + else => result.target.cpu.features.addFeature( + @enumToInt(std.Target.x86.Feature.@"32bit_mode"), + ), + } + } + }, + else => {}, + } + cross_target.updateCpuFeatures(&result.target.cpu.features); + result.cpu_detection_unimplemented = cpu_detection_unimplemented; + return result; } /// First we attempt to use the executable's own binary. If it is dynamically diff --git a/src/Compilation.zig b/src/Compilation.zig index 39e10becec..bbe36d8fce 100644 --- a/src/Compilation.zig +++ b/src/Compilation.zig @@ -2264,6 +2264,7 @@ pub fn addCCArgs( } try argv.append("-D_LIBCPP_DISABLE_VISIBILITY_ANNOTATIONS"); try argv.append("-D_LIBCXXABI_DISABLE_VISIBILITY_ANNOTATIONS"); + try argv.append("-D_LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS"); } const llvm_triple = try @import("codegen/llvm.zig").targetTriple(arena, target); @@ -2847,11 +2848,7 @@ pub fn generateBuiltinZigSource(comp: *Compilation, allocator: *Allocator) ![]u8 const index = @intCast(std.Target.Cpu.Feature.Set.Index, index_usize); const is_enabled = target.cpu.features.isEnabled(index); if (is_enabled) { - // TODO some kind of "zig identifier escape" function rather than - // unconditionally using @"" syntax - try buffer.appendSlice(" .@\""); - try buffer.appendSlice(feature.name); - try buffer.appendSlice("\",\n"); + try buffer.writer().print(" .{},\n", .{std.zig.fmtId(feature.name)}); } } diff --git a/src/clang.zig b/src/clang.zig index 854ee7dc7e..5d00df0357 100644 --- a/src/clang.zig +++ b/src/clang.zig @@ -411,7 +411,7 @@ pub const Expr = opaque { extern fn ZigClangExpr_getBeginLoc(*const Expr) SourceLocation; pub const evaluateAsConstantExpr = ZigClangExpr_EvaluateAsConstantExpr; - extern fn ZigClangExpr_EvaluateAsConstantExpr(*const Expr, *ExprEvalResult, Expr_ConstExprUsage, *const ASTContext) bool; + extern fn ZigClangExpr_EvaluateAsConstantExpr(*const Expr, *ExprEvalResult, Expr_ConstantExprKind, *const ASTContext) bool; }; pub const FieldDecl = opaque { @@ -1304,6 +1304,8 @@ pub const CK = extern enum { IntegralCast, IntegralToBoolean, IntegralToFloating, + FloatingToFixedPoint, + FixedPointToFloating, FixedPointCast, FixedPointToIntegral, IntegralToFixedPoint, @@ -1410,6 +1412,7 @@ pub const DeclKind = extern enum { MSGuid, OMPDeclareMapper, OMPDeclareReduction, + TemplateParamObject, UnresolvedUsingValue, OMPAllocate, OMPRequires, @@ -1520,6 +1523,8 @@ pub const BuiltinTypeKind = extern enum { SveFloat64x4, SveBFloat16x4, SveBool, + VectorQuad, + VectorPair, Void, Bool, Char_U, @@ -1673,9 +1678,11 @@ pub const PreprocessedEntity_EntityKind = extern enum { InclusionDirectiveKind, }; -pub const Expr_ConstExprUsage = extern enum { - EvaluateForCodeGen, - EvaluateForMangling, +pub const Expr_ConstantExprKind = extern enum { + Normal, + NonClassTemplateArgument, + ClassTemplateArgument, + ImmediateInvocation, }; pub const UnaryExprOrTypeTrait_Kind = extern enum { diff --git a/src/clang_options_data.zig b/src/clang_options_data.zig index b1cb6c258c..8e4ab92894 100644 --- a/src/clang_options_data.zig +++ b/src/clang_options_data.zig @@ -15,7 +15,7 @@ flagpd1("CC"), flagpd1("EB"), flagpd1("EL"), flagpd1("Eonly"), -flagpd1("fAAPCSBitfieldLoad"), +flagpd1("faapcs-bitfield-load"), flagpd1("H"), .{ .name = "<input>", @@ -1028,14 +1028,6 @@ flagpsl("MT"), .psl = true, }, .{ - .name = "Zd", - .syntax = .flag, - .zig_equivalent = .other, - .pd1 = true, - .pd2 = false, - .psl = true, -}, -.{ .name = "Ze", .syntax = .flag, .zig_equivalent = .other, @@ -2073,49 +2065,7 @@ flagpd1("analyzer-viz-egraph-graphviz"), flagpd1("analyzer-werror"), flagpd1("fslp-vectorize-aggressive"), flagpd1("fno-slp-vectorize-aggressive"), -flagpd1("fno-record-gcc-switches"), -flagpd1("fexpensive-optimizations"), -flagpd1("fno-expensive-optimizations"), -flagpd1("fdefer-pop"), -flagpd1("fno-defer-pop"), -flagpd1("fextended-identifiers"), -flagpd1("fno-extended-identifiers"), -flagpd1("fhonor-infinites"), -flagpd1("fno-honor-infinites"), -flagpd1("findirect-virtual-calls"), -sepd1("fnew-alignment"), -flagpd1("faligned-new"), -flagpd1("fno-aligned-new"), -flagpd1("fsched-interblock"), -flagpd1("ftree-vectorize"), -flagpd1("fno-tree-vectorize"), -flagpd1("ftree-slp-vectorize"), -flagpd1("fno-tree-slp-vectorize"), -flagpd1("fident"), -flagpd1("fterminated-vtables"), -flagpd1("grecord-gcc-switches"), -flagpd1("gno-record-gcc-switches"), -flagpd1("nocudainc"), -flagpd1("nocudalib"), -.{ - .name = "system-header-prefix", - .syntax = .separate, - .zig_equivalent = .other, - .pd1 = false, - .pd2 = true, - .psl = false, -}, -.{ - .name = "no-system-header-prefix", - .syntax = .separate, - .zig_equivalent = .other, - .pd1 = false, - .pd2 = true, - .psl = false, -}, -flagpd1("integrated-as"), -flagpd1("no-integrated-as"), -flagpd1("fno-ident"), +flagpd1("shared-libasan"), .{ .name = "Gs", .syntax = .flag, @@ -2196,7 +2146,6 @@ flagpd1("fno-ident"), .pd2 = false, .psl = true, }, -flagpd1("fcuda-rdc"), .{ .name = "Os", .syntax = .flag, @@ -2237,9 +2186,52 @@ flagpd1("fcuda-rdc"), .pd2 = false, .psl = true, }, -flagpd1("fno-cuda-rdc"), -flagpd1("shared-libasan"), flagpd1("frecord-gcc-switches"), +flagpd1("fno-record-gcc-switches"), +flagpd1("fident"), +flagpd1("fno-ident"), +flagpd1("fexpensive-optimizations"), +flagpd1("fno-expensive-optimizations"), +flagpd1("fdefer-pop"), +flagpd1("fno-defer-pop"), +flagpd1("fextended-identifiers"), +flagpd1("fno-extended-identifiers"), +flagpd1("fhonor-infinites"), +flagpd1("fno-honor-infinites"), +flagpd1("fcuda-rdc"), +flagpd1("fno-cuda-rdc"), +flagpd1("findirect-virtual-calls"), +sepd1("fnew-alignment"), +flagpd1("faligned-new"), +flagpd1("fno-aligned-new"), +flagpd1("fsched-interblock"), +flagpd1("ftree-vectorize"), +flagpd1("fno-tree-vectorize"), +flagpd1("ftree-slp-vectorize"), +flagpd1("fno-tree-slp-vectorize"), +flagpd1("fterminated-vtables"), +flagpd1("grecord-gcc-switches"), +flagpd1("gno-record-gcc-switches"), +flagpd1("nocudainc"), +flagpd1("nocudalib"), +.{ + .name = "system-header-prefix", + .syntax = .separate, + .zig_equivalent = .other, + .pd1 = false, + .pd2 = true, + .psl = false, +}, +.{ + .name = "no-system-header-prefix", + .syntax = .separate, + .zig_equivalent = .other, + .pd1 = false, + .pd2 = true, + .psl = false, +}, +flagpd1("integrated-as"), +flagpd1("no-integrated-as"), .{ .name = "ansi", .syntax = .flag, @@ -2251,11 +2243,8 @@ flagpd1("frecord-gcc-switches"), sepd1("arch"), flagpd1("arch_errors_fatal"), sepd1("arch_only"), -flagpd1("arcmt-check"), -flagpd1("arcmt-migrate"), flagpd1("arcmt-migrate-emit-errors"), sepd1("arcmt-migrate-report-output"), -flagpd1("arcmt-modify"), flagpd1("ast-dump"), flagpd1("ast-dump-all"), flagpd1("ast-dump-decl-types"), @@ -2451,6 +2440,7 @@ flagpd1("emit-pch"), }, flagpd1("enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang"), sepd1("error-on-deserialized-decl"), +sepd1("exception-model"), sepd1("exported_symbols_list"), .{ .name = "fPIC", @@ -2468,7 +2458,8 @@ sepd1("exported_symbols_list"), .pd2 = false, .psl = false, }, -flagpd1("fno-access-control"), +flagpd1("faapcs-bitfield-width"), +flagpd1("faccess-control"), flagpd1("faddrsig"), flagpd1("faggressive-function-elimination"), flagpd1("falign-commons"), @@ -2481,6 +2472,7 @@ flagpd1("fall-intrinsics"), flagpd1("fallow-editor-placeholders"), flagpd1("fallow-half-arguments-and-returns"), flagpd1("fallow-pch-with-compiler-errors"), +flagpd1("fallow-pcm-with-compiler-errors"), flagpd1("fallow-unsupported"), flagpd1("faltivec"), flagpd1("fkeep-inline-functions"), @@ -2490,6 +2482,7 @@ flagpd1("fapple-link-rtlib"), flagpd1("fapple-pragma-pack"), flagpd1("fapplication-extension"), flagpd1("fapply-global-visibility-to-externs"), +flagpd1("fapprox-func"), flagpd1("fasm"), flagpd1("fasm-blocks"), flagpd1("fassociative-math"), @@ -2590,6 +2583,7 @@ flagpd1("fdiagnostics-show-note-include-stack"), flagpd1("fdiagnostics-show-option"), flagpd1("fdiagnostics-show-template-tree"), flagpd1("fdigraphs"), +flagpd1("fdirect-access-external-data"), flagpd1("fdisable-module-hash"), flagpd1("fdiscard-value-names"), flagpd1("fdollar-ok"), @@ -2610,14 +2604,13 @@ flagpd1("feliminate-unused-debug-types"), flagpd1("fembed-bitcode"), flagpd1("fembed-bitcode-marker"), flagpd1("femit-all-decls"), -flagpd1("femit-coverage-data"), -flagpd1("femit-coverage-notes"), flagpd1("femulated-tls"), flagpd1("fenable-matrix"), flagpd1("fencode-extended-block-signature"), sepd1("ferror-limit"), flagpd1("fescaping-block-tail-calls"), flagpd1("fexceptions"), +flagpd1("fexperimental-debug-variable-locations"), flagpd1("fexperimental-isel"), flagpd1("fexperimental-new-constant-interpreter"), flagpd1("fexperimental-new-pass-manager"), @@ -2631,6 +2624,7 @@ flagpd1("ffast-math"), flagpd1("ffat-lto-objects"), flagpd1("fcheck-new"), flagpd1("ffine-grained-bitfield-accesses"), +flagpd1("ffinite-loops"), flagpd1("ffinite-math-only"), flagpd1("finline-limit"), flagpd1("ffixed-form"), @@ -2691,7 +2685,10 @@ flagpd1("fgnu-inline-asm"), flagpd1("fgnu-keywords"), flagpd1("fgnu-runtime"), flagpd1("fgpu-allow-device-init"), +flagpd1("fgpu-defer-diag"), +flagpd1("fgpu-exclude-wrong-side-overloads"), flagpd1("fgpu-rdc"), +flagpd1("fhalf-no-semantic-interposition"), flagpd1("fheinous-gnu-extensions"), flagpd1("fhip-dump-offload-linker-script"), flagpd1("fhip-new-launch-api"), @@ -2731,6 +2728,7 @@ flagpd1("fjump-tables"), flagpd1("fkeep-static-consts"), flagpd1("flat_namespace"), flagpd1("flax-vector-conversions"), +flagpd1("flegacy-pass-manager"), flagpd1("flimit-debug-info"), .{ .name = "flto", @@ -2745,6 +2743,7 @@ flagpd1("flto-visibility-public-std"), sepd1("fmacro-backtrace-limit"), flagpd1("fmath-errno"), flagpd1("fmax-identifier-length"), +flagpd1("fmemory-profile"), flagpd1("fmerge-all-constants"), flagpd1("fmerge-constants"), flagpd1("fmerge-functions"), @@ -2797,7 +2796,8 @@ flagpd1("fnext-runtime"), .pd2 = false, .psl = false, }, -flagpd1("fno-no-access-control"), +flagpd1("fno-aapcs-bitfield-width"), +flagpd1("fno-access-control"), flagpd1("fno-addrsig"), flagpd1("fno-aggressive-function-elimination"), flagpd1("fno-align-commons"), @@ -2875,6 +2875,7 @@ flagpd1("fno-diagnostics-show-note-include-stack"), flagpd1("fno-diagnostics-show-option"), flagpd1("fno-diagnostics-use-presumed-location"), flagpd1("fno-digraphs"), +flagpd1("fno-direct-access-external-data"), flagpd1("fno-discard-value-names"), flagpd1("fno-dllexport-inlines"), flagpd1("fno-dollar-ok"), @@ -2901,6 +2902,7 @@ flagpd1("fno-fast-math"), flagpd1("fno-fat-lto-objects"), flagpd1("fno-check-new"), flagpd1("fno-fine-grained-bitfield-accesses"), +flagpd1("fno-finite-loops"), flagpd1("fno-finite-math-only"), flagpd1("fno-inline-limit"), flagpd1("fno-fixed-form"), @@ -2925,6 +2927,8 @@ flagpd1("fno-gnu89-inline"), flagpd1("fno-gnu-inline-asm"), flagpd1("fno-gnu-keywords"), flagpd1("fno-gpu-allow-device-init"), +flagpd1("fno-gpu-defer-diag"), +flagpd1("fno-gpu-exclude-wrong-side-overloads"), flagpd1("fno-gpu-rdc"), flagpd1("fno-hip-new-launch-api"), flagpd1("fno-honor-infinities"), @@ -2948,6 +2952,7 @@ flagpd1("fno-ivopts"), flagpd1("fno-jump-tables"), flagpd1("fno-keep-static-consts"), flagpd1("fno-lax-vector-conversions"), +flagpd1("fno-legacy-pass-manager"), flagpd1("fno-limit-debug-info"), .{ .name = "fno-lto", @@ -2962,6 +2967,7 @@ flagpd1("fno-math-builtin"), flagpd1("fno-math-errno"), flagpd1("fno-max-identifier-length"), flagpd1("fno-max-type-align"), +flagpd1("fno-memory-profile"), flagpd1("fno-merge-all-constants"), flagpd1("fno-merge-constants"), flagpd1("fno-module-file-deps"), @@ -3011,6 +3017,7 @@ flagpd1("fno-permissive"), flagpd1("fno-pic"), flagpd1("fno-pie"), flagpd1("fno-plt"), +flagpd1("fno-prebuilt-implicit-modules"), flagpd1("fno-prefetch-loop-arrays"), flagpd1("fno-preserve-as-comments"), flagpd1("fno-printf"), @@ -3027,6 +3034,7 @@ flagpd1("fno-profile-sample-use"), flagpd1("fno-profile-use"), flagpd1("fno-profile-values"), flagpd1("fno-protect-parens"), +flagpd1("fno-pseudo-probe-for-profiling"), flagpd1("fno-range-check"), flagpd1("fno-real-4-real-10"), flagpd1("fno-real-4-real-16"), @@ -3097,6 +3105,7 @@ flagpd1("fno-spec-constr-count"), flagpd1("fno-spell-checking"), flagpd1("fno-split-dwarf-inlining"), flagpd1("fno-split-lto-unit"), +flagpd1("fno-split-machine-functions"), flagpd1("fno-stack-arrays"), flagpd1("fno-stack-check"), flagpd1("fno-stack-clash-protection"), @@ -3147,6 +3156,8 @@ flagpd1("fno-vect-cost-model"), flagpd1("fno-vectorize"), flagpd1("fno-verbose-asm"), flagpd1("fno-virtual-function-elimination"), +flagpd1("fno-visibility-from-dllstorageclass"), +flagpd1("fno-visibility-inlines-hidden-static-local-var"), flagpd1("fno-wchar"), flagpd1("fno-web"), flagpd1("fno-whole-file"), @@ -3154,6 +3165,7 @@ flagpd1("fno-whole-program"), flagpd1("fno-whole-program-vtables"), flagpd1("fno-working-directory"), flagpd1("fno-wrapv"), +flagpd1("fno-xl-pragma-pack"), flagpd1("fno-xray-always-emit-customevents"), flagpd1("fno-xray-always-emit-typedevents"), flagpd1("fno-xray-function-index"), @@ -3205,6 +3217,7 @@ flagpd1("fpack-struct"), flagpd1("fpadding-on-unsigned-fixed-point"), flagpd1("fparse-all-comments"), flagpd1("fpascal-strings"), +flagpd1("fpass-by-value-is-noalias"), flagpd1("fpcc-struct-return"), flagpd1("fpch-codegen"), flagpd1("fpch-debuginfo"), @@ -3216,6 +3229,7 @@ flagpd1("fpermissive"), flagpd1("fpic"), flagpd1("fpie"), flagpd1("fplt"), +flagpd1("fprebuilt-implicit-modules"), flagpd1("fprefetch-loop-arrays"), flagpd1("fpreserve-as-comments"), flagpd1("fpreserve-vec3-type"), @@ -3234,6 +3248,7 @@ flagpd1("fprofile-sample-use"), flagpd1("fprofile-use"), flagpd1("fprofile-values"), flagpd1("fprotect-parens"), +flagpd1("fpseudo-probe-for-profiling"), .{ .name = "framework", .syntax = .separate, @@ -3333,6 +3348,7 @@ flagpd1("fspell-checking"), sepd1("fspell-checking-limit"), flagpd1("fsplit-dwarf-inlining"), flagpd1("fsplit-lto-unit"), +flagpd1("fsplit-machine-functions"), flagpd1("fsplit-stack"), flagpd1("fstack-arrays"), flagpd1("fstack-check"), @@ -3388,6 +3404,7 @@ flagpd1("funsigned-bitfields"), flagpd1("funsigned-char"), flagpd1("funswitch-loops"), flagpd1("funwind-tables"), +flagpd1("fuse-ctor-homing"), flagpd1("fuse-cxa-atexit"), flagpd1("fuse-init-array"), flagpd1("fuse-line-directives"), @@ -3400,8 +3417,10 @@ flagpd1("fvectorize"), flagpd1("fverbose-asm"), flagpd1("fvirtual-function-elimination"), sepd1("fvisibility"), +flagpd1("fvisibility-from-dllstorageclass"), flagpd1("fvisibility-global-new-delete-hidden"), flagpd1("fvisibility-inlines-hidden"), +flagpd1("fvisibility-inlines-hidden-static-local-var"), flagpd1("fvisibility-ms-compat"), flagpd1("fwasm-exceptions"), flagpd1("fweb"), @@ -3410,6 +3429,7 @@ flagpd1("fwhole-program"), flagpd1("fwhole-program-vtables"), flagpd1("fwrapv"), flagpd1("fwritable-strings"), +flagpd1("fxl-pragma-pack"), flagpd1("fxray-always-emit-customevents"), flagpd1("fxray-always-emit-typedevents"), flagpd1("fxray-function-index"), @@ -3442,6 +3462,8 @@ flagpd1("gcodeview"), flagpd1("gcodeview-ghash"), flagpd1("gcolumn-info"), flagpd1("gdwarf"), +flagpd1("gdwarf32"), +flagpd1("gdwarf64"), flagpd1("gdwarf-2"), flagpd1("gdwarf-3"), flagpd1("gdwarf-4"), @@ -3477,6 +3499,7 @@ flagpd1("gno-gnu-pubnames"), flagpd1("gno-inline-line-tables"), flagpd1("gno-pubnames"), flagpd1("gno-record-command-line"), +flagpd1("gno-split-dwarf"), flagpd1("gno-strict-dwarf"), .{ .name = "gpu-use-aux-triple-only", @@ -3531,6 +3554,8 @@ flagpd1("m64"), flagpd1("m80387"), flagpd1("mseses"), flagpd1("mabi=ieeelongdouble"), +flagpd1("mabi=vec-default"), +flagpd1("mabi=vec-extabi"), flagpd1("mabicalls"), flagpd1("madx"), flagpd1("maes"), @@ -3561,6 +3586,7 @@ flagpd1("mavx512vl"), flagpd1("mavx512vnni"), flagpd1("mavx512vp2intersect"), flagpd1("mavx512vpopcntdq"), +flagpd1("mavxvnni"), flagpd1("mbackchain"), flagpd1("mbig-endian"), flagpd1("mbmi"), @@ -3593,6 +3619,7 @@ flagpd1("mdouble-float"), flagpd1("mdsp"), flagpd1("mdspr2"), sepd1("meabi"), +flagpd1("mefpu2"), flagpd1("membedded-data"), flagpd1("menable-experimental-extensions"), flagpd1("menable-no-infs"), @@ -3626,10 +3653,12 @@ flagpd1("mglobal-merge"), flagpd1("mgpopt"), flagpd1("mhard-float"), flagpd1("mhvx"), +flagpd1("mhreset"), flagpd1("mhtm"), flagpd1("miamcu"), flagpd1("mieee-fp"), flagpd1("mieee-rnd-near"), +flagpd1("mignore-xcoff-visibility"), flagpd1("migrate"), flagpd1("no-finalize-removal"), flagpd1("no-ns-alloc-error"), @@ -3656,6 +3685,7 @@ flagpd1("mips64r5"), flagpd1("mips64r6"), flagpd1("misel"), flagpd1("mkernel"), +flagpd1("mkl"), flagpd1("mldc1-sdc1"), sepd1("mlimit-float-precision"), sepd1("mlink-bitcode-file"), @@ -3674,10 +3704,12 @@ flagpd1("mlvi-hardening"), flagpd1("mlwp"), flagpd1("mlzcnt"), flagpd1("mmadd4"), +flagpd1("mmark-bti-property"), flagpd1("mmemops"), flagpd1("mmfcrf"), flagpd1("mmfocrf"), flagpd1("mmicromips"), +flagpd1("mmma"), flagpd1("mmmx"), flagpd1("mmovbe"), flagpd1("mmovdir64b"), @@ -3717,6 +3749,7 @@ flagpd1("mno-avx512vl"), flagpd1("mno-avx512vnni"), flagpd1("mno-avx512vp2intersect"), flagpd1("mno-avx512vpopcntdq"), +flagpd1("mno-avxvnni"), flagpd1("mno-backchain"), flagpd1("mno-bmi"), flagpd1("mno-bmi2"), @@ -3755,6 +3788,7 @@ flagpd1("mno-ginv"), flagpd1("mno-global-merge"), flagpd1("mno-gpopt"), flagpd1("mno-hvx"), +flagpd1("mno-hreset"), flagpd1("mno-htm"), flagpd1("mno-iamcu"), flagpd1("mno-implicit-float"), @@ -3763,6 +3797,7 @@ flagpd1("mno-inline-all-stringops"), flagpd1("mno-invariant-function-descriptors"), flagpd1("mno-invpcid"), flagpd1("mno-isel"), +flagpd1("mno-kl"), flagpd1("mno-ldc1-sdc1"), flagpd1("mno-local-sdata"), flagpd1("mno-long-calls"), @@ -3777,6 +3812,7 @@ flagpd1("mno-mfcrf"), flagpd1("mno-mfocrf"), flagpd1("mno-micromips"), flagpd1("mno-mips16"), +flagpd1("mno-mma"), flagpd1("mno-mmx"), flagpd1("mno-movbe"), flagpd1("mno-movdir64b"), @@ -3796,6 +3832,7 @@ flagpd1("mno-nvs"), flagpd1("mno-odd-spreg"), flagpd1("mno-omit-leaf-frame-pointer"), flagpd1("mno-outline"), +flagpd1("mno-outline-atomics"), flagpd1("mno-packed-stack"), flagpd1("mno-packets"), flagpd1("mno-pascal-strings"), @@ -3813,7 +3850,6 @@ flagpd1("mno-prefetchwt1"), flagpd1("mno-prfchw"), flagpd1("mno-ptwrite"), flagpd1("mno-pure-code"), -flagpd1("mno-qpx"), flagpd1("mno-rdpid"), flagpd1("mno-rdrnd"), flagpd1("mno-rdseed"), @@ -3846,7 +3882,6 @@ flagpd1("mno-simd128"), flagpd1("mno-soft-float"), flagpd1("mno-spe"), flagpd1("mno-speculative-load-hardening"), -flagpd1("mno-sram-ecc"), flagpd1("mno-sse"), flagpd1("mno-sse2"), flagpd1("mno-sse3"), @@ -3862,8 +3897,10 @@ flagpd1("mno-tbm"), flagpd1("mno-thumb"), flagpd1("mno-tls-direct-seg-refs"), flagpd1("mno-tsxldtrk"), +flagpd1("mno-uintr"), flagpd1("mno-unaligned-access"), flagpd1("mno-unimplemented-simd128"), +flagpd1("mno-unsafe-fp-atomics"), flagpd1("mno-vaes"), flagpd1("mno-virt"), flagpd1("mno-vpclmulqdq"), @@ -3874,9 +3911,9 @@ flagpd1("mno-waitpkg"), flagpd1("mno-warn-nonportable-cfstrings"), flagpd1("mno-wavefrontsize64"), flagpd1("mno-wbnoinvd"), +flagpd1("mno-widekl"), flagpd1("mno-x87"), flagpd1("mno-xgot"), -flagpd1("mno-xnack"), flagpd1("mno-xop"), flagpd1("mno-xsave"), flagpd1("mno-xsavec"), @@ -3887,6 +3924,7 @@ flagpd1("mnocrc"), flagpd1("mno-direct-move"), flagpd1("mnontrapping-fptoint"), flagpd1("mnop-mcount"), +flagpd1("mno-paired-vector-memops"), flagpd1("mno-crypto"), flagpd1("mnvj"), flagpd1("mnvs"), @@ -3896,8 +3934,10 @@ flagpd1("module-file-deps"), flagpd1("module-file-info"), flagpd1("momit-leaf-frame-pointer"), flagpd1("moutline"), +flagpd1("moutline-atomics"), flagpd1("mpacked-stack"), flagpd1("mpackets"), +flagpd1("mpaired-vector-memops"), flagpd1("mpascal-strings"), flagpd1("mpclmul"), flagpd1("mpconfig"), @@ -3915,7 +3955,6 @@ flagpd1("mprfchw"), flagpd1("mptwrite"), flagpd1("mpure-code"), flagpd1("mqdsp6-compat"), -flagpd1("mqpx"), flagpd1("mrdpid"), flagpd1("mrdrnd"), flagpd1("mrdseed"), @@ -3958,13 +3997,13 @@ flagpd1("msgx"), flagpd1("msha"), flagpd1("mshstk"), flagpd1("msign-ext"), +flagpd1("msim"), flagpd1("msimd128"), flagpd1("msingle-float"), sepd1("msmall-data-limit"), flagpd1("msoft-float"), flagpd1("mspe"), flagpd1("mspeculative-load-hardening"), -flagpd1("msram-ecc"), flagpd1("msse"), flagpd1("msse2"), flagpd1("msse3"), @@ -3988,11 +4027,13 @@ sepd1("mtp"), flagpd1("mtsxldtrk"), flagpd1("mtune=?"), flagpd1("muclibc"), +flagpd1("muintr"), flagpd1("multi_module"), sepd1("multiply_defined"), sepd1("multiply_defined_unused"), flagpd1("munaligned-access"), flagpd1("munimplemented-simd128"), +flagpd1("munsafe-fp-atomics"), flagpd1("munwind-tables"), flagpd1("mv5"), flagpd1("mv55"), @@ -4012,10 +4053,10 @@ flagpd1("mwaitpkg"), flagpd1("mwarn-nonportable-cfstrings"), flagpd1("mwavefrontsize64"), flagpd1("mwbnoinvd"), +flagpd1("mwidekl"), flagpd1("mx32"), flagpd1("mx87"), flagpd1("mxgot"), -flagpd1("mxnack"), flagpd1("mxop"), flagpd1("mxsave"), flagpd1("mxsavec"), @@ -4455,7 +4496,7 @@ sepd1("target-feature"), }, sepd1("target-linker-version"), flagpd1("templight-dump"), -flagpd1("test-coverage"), +flagpd1("test-io"), flagpd1("time"), .{ .name = "traditional", @@ -4483,6 +4524,7 @@ flagpd1("time"), }, flagpd1("trim-egraph"), sepd1("triple"), +sepd1("tune-cpu"), flagpd1("twolevel_namespace"), flagpd1("twolevel_namespace_hints"), sepd1("umbrella"), @@ -4534,6 +4576,7 @@ flagpd1("whyload"), }, joinpd1("fsanitize-undefined-strip-path-components="), joinpd1("fopenmp-cuda-teams-reduction-recs-num="), +joinpd1("fvisibility-externs-nodllstorageclass="), joinpd1("analyzer-config-compatibility-mode="), joinpd1("ftrivial-auto-var-init-stop-after="), joinpd1("fpatchable-function-entry-offset="), @@ -4542,7 +4585,11 @@ joinpd1("fsanitize-address-field-padding="), joinpd1("fdiagnostics-hotness-threshold="), joinpd1("fsanitize-memory-track-origins="), joinpd1("mwatchos-simulator-version-min="), +joinpd1("fvisibility-externs-dllimport="), +joinpd1("fvisibility-nodllstorageclass="), +joinpd1("fxray-selected-function-group="), joinpd1("mappletvsimulator-version-min="), +joinpd1("mstack-protector-guard-offset="), joinpd1("fsanitize-coverage-whitelist="), joinpd1("fsanitize-coverage-blacklist="), joinpd1("fobjc-nonfragile-abi-version="), @@ -4575,7 +4622,16 @@ joinpd1("fopenmp-cuda-blocks-per-sm="), joinpd1("fsanitize-system-blacklist="), jspd1("fxray-instruction-threshold"), joinpd1("headerpad_max_install_names"), +.{ + .name = "libomptarget-nvptx-bc-path=", + .syntax = .joined, + .zig_equivalent = .other, + .pd1 = false, + .pd2 = true, + .psl = false, +}, joinpd1("mios-simulator-version-min="), +joinpd1("mstack-protector-guard-reg="), .{ .name = "include-with-prefix-after=", .syntax = .joined, @@ -4649,14 +4705,6 @@ joinpd1("ftime-trace-granularity="), jspd1("fxray-always-instrument="), jspd1("internal-externc-isystem"), .{ - .name = "libomptarget-nvptx-path=", - .syntax = .joined, - .zig_equivalent = .other, - .pd1 = false, - .pd2 = true, - .psl = false, -}, -.{ .name = "no-system-header-prefix=", .syntax = .joined, .zig_equivalent = .other, @@ -4686,6 +4734,7 @@ jspd1("fxray-never-instrument="), jspd1("interface-stub-version="), joinpd1("malign-branch-boundary="), joinpd1("mappletvos-version-min="), +joinpd1("mstack-protector-guard="), joinpd1("Wnonportable-cfstrings"), joinpd1("fbasic-block-sections="), joinpd1("fdefault-calling-conv="), @@ -4705,6 +4754,8 @@ joinpd1("foperator-arrow-depth="), joinpd1("fprebuilt-module-path="), joinpd1("fprofile-filter-files="), joinpd1("fspell-checking-limit="), +joinpd1("fvisibility-dllexport="), +joinpd1("fxray-function-groups="), joinpd1("miphoneos-version-min="), joinpd1("msmall-data-threshold="), joinpd1("Wlarge-by-value-copy="), @@ -4723,7 +4774,7 @@ joinpd1("fmodules-prune-after="), .psl = false, }, jspd1("iframeworkwithsysroot"), -joinpd1("mamdgpu-debugger-abi="), +joinpd1("mcode-object-version="), joinpd1("mpad-max-prefix-size="), joinpd1("mprefer-vector-width="), joinpd1("msign-return-address="), @@ -4767,6 +4818,7 @@ joinpd1("fmax-stack-var-size="), joinpd1("fmodules-cache-path="), joinpd1("fmodules-embed-file="), joinpd1("fprofile-instrument="), +joinpd1("fprofile-prefix-map="), joinpd1("fprofile-sample-use="), joinpd1("fsanitize-blacklist="), joinpd1("mmacosx-version-min="), @@ -4795,6 +4847,14 @@ joinpd1("fprofile-instr-use="), .psl = false, }, joinpd1("fthin-link-bitcode="), +.{ + .name = "gpu-instrument-lib=", + .syntax = .joined, + .zig_equivalent = .other, + .pd1 = false, + .pd2 = true, + .psl = false, +}, joinpd1("mbranch-protection="), joinpd1("mmacos-version-min="), joinpd1("pch-through-header="), @@ -4831,6 +4891,7 @@ joinpd1("target-sdk-version="), .pd2 = true, .psl = false, }, +joinpd1("fbinutils-version="), joinpd1("fclang-abi-compat="), joinpd1("fcompile-resource="), joinpd1("fdebug-prefix-map="), @@ -4848,6 +4909,7 @@ joinpd1("fmacro-prefix-map="), }, joinpd1("fobjc-abi-version="), joinpd1("foutput-class-dir="), +joinpd1("fproc-stat-report="), joinpd1("fprofile-generate="), joinpd1("frewrite-map-file="), .{ @@ -4881,10 +4943,12 @@ joinpd1("fconstexpr-steps="), joinpd1("ffile-prefix-map="), joinpd1("fmodule-map-file="), joinpd1("fobjc-arc-cxxlib="), +joinpd1("fproc-stat-report"), jspd1("iwithprefixbefore"), joinpd1("malign-functions="), joinpd1("mios-version-min="), joinpd1("mstack-alignment="), +joinpd1("msve-vector-bits="), .{ .name = "no-cuda-gpu-arch=", .syntax = .joined, @@ -4905,9 +4969,11 @@ joinpd1("analyzer-output="), }, joinpd1("debug-info-kind="), joinpd1("debugger-tuning="), +joinpd1("exception-model="), joinpd1("fcf-runtime-abi="), joinpd1("finit-character="), joinpd1("fmax-type-align="), +joinpd1("fmemory-profile="), joinpd1("fmessage-length="), .{ .name = "fopenmp-targets=", @@ -4918,6 +4984,7 @@ joinpd1("fmessage-length="), .psl = false, }, joinpd1("fopenmp-version="), +joinpd1("fprofile-update="), joinpd1("fshow-overloads="), joinpd1("ftemplate-depth-"), joinpd1("ftemplate-depth="), @@ -5067,6 +5134,7 @@ joinpd1("finit-integer="), joinpd1("finit-logical="), joinpd1("finline-limit="), joinpd1("fobjc-runtime="), +joinpd1("fprofile-list="), .{ .name = "gcc-toolchain=", .syntax = .joined, @@ -5104,6 +5172,7 @@ joinpd1("Wlarger-than="), .pd2 = true, .psl = false, }, +joinpd1("arcmt-action="), joinpd1("ast-dump-all="), .{ .name = "autocomplete=", @@ -5134,6 +5203,7 @@ joinpd1("fpass-plugin="), joinpd1("fprofile-dir="), joinpd1("fprofile-use="), joinpd1("frandom-seed="), +joinpd1("ftime-report="), joinpd1("gsplit-dwarf="), jspd1("isystem-after"), joinpd1("malign-jumps="), @@ -5294,6 +5364,14 @@ jspd1("sub_library"), .psl = true, }, .{ + .name = "vctoolsdir", + .syntax = .joined_or_separate, + .zig_equivalent = .other, + .pd1 = true, + .pd2 = false, + .psl = true, +}, +.{ .name = "classpath=", .syntax = .joined, .zig_equivalent = .other, @@ -5422,11 +5500,20 @@ jspd1("undefined"), .pd2 = true, .psl = false, }, +jspd1("dsym-dir"), joinpd1("fopenmp="), joinpd1("fplugin="), joinpd1("fuse-ld="), joinpd1("fveclib="), jspd1("isysroot"), +.{ + .name = "ld-path=", + .syntax = .joined, + .zig_equivalent = .other, + .pd1 = false, + .pd2 = true, + .psl = false, +}, joinpd1("mcmodel="), joinpd1("mconsole"), joinpd1("mdouble="), @@ -5652,6 +5739,14 @@ jspd1("Ttext"), .psl = true, }, .{ + .name = "tune:", + .syntax = .joined, + .zig_equivalent = .other, + .pd1 = true, + .pd2 = false, + .psl = true, +}, +.{ .name = "warn-", .syntax = .joined, .zig_equivalent = .other, @@ -5737,6 +5832,14 @@ joinpd1("mtp="), .psl = false, }, .{ + .name = "Fe:", + .syntax = .joined, + .zig_equivalent = .other, + .pd1 = true, + .pd2 = false, + .psl = true, +}, +.{ .name = "RTC", .syntax = .joined, .zig_equivalent = .other, diff --git a/src/codegen/llvm.zig b/src/codegen/llvm.zig index f087957f1d..8e38145be7 100644 --- a/src/codegen/llvm.zig +++ b/src/codegen/llvm.zig @@ -26,6 +26,7 @@ pub fn targetTriple(allocator: *Allocator, target: std.Target) ![:0]u8 { .avr => "avr", .bpfel => "bpfel", .bpfeb => "bpfeb", + .csky => "csky", .hexagon => "hexagon", .mips => "mips", .mipsel => "mipsel", @@ -33,6 +34,7 @@ pub fn targetTriple(allocator: *Allocator, target: std.Target) ![:0]u8 { .mips64el => "mips64el", .msp430 => "msp430", .powerpc => "powerpc", + .powerpcle => "powerpcle", .powerpc64 => "powerpc64", .powerpc64le => "powerpc64le", .r600 => "r600", @@ -90,11 +92,11 @@ pub fn targetTriple(allocator: *Allocator, target: std.Target) ![:0]u8 { .openbsd => "openbsd", .solaris => "solaris", .windows => "windows", + .zos => "zos", .haiku => "haiku", .minix => "minix", .rtems => "rtems", .nacl => "nacl", - .cnk => "cnk", .aix => "aix", .cuda => "cuda", .nvcl => "nvcl", @@ -125,6 +127,7 @@ pub fn targetTriple(allocator: *Allocator, target: std.Target) ![:0]u8 { .gnueabi => "gnueabi", .gnueabihf => "gnueabihf", .gnux32 => "gnux32", + .gnuilp32 => "gnuilp32", .code16 => "code16", .eabi => "eabi", .eabihf => "eabihf", diff --git a/src/codegen/llvm/bindings.zig b/src/codegen/llvm/bindings.zig index ccba3d9973..223d006d5c 100644 --- a/src/codegen/llvm/bindings.zig +++ b/src/codegen/llvm/bindings.zig @@ -495,6 +495,7 @@ pub const ObjectFormatType = extern enum(c_int) { Unknown, COFF, ELF, + GOFF, MachO, Wasm, XCOFF, @@ -515,97 +516,99 @@ extern fn ZigLLVMWriteArchive( ) bool; pub const OSType = extern enum(c_int) { - UnknownOS = 0, - Ananas = 1, - CloudABI = 2, - Darwin = 3, - DragonFly = 4, - FreeBSD = 5, - Fuchsia = 6, - IOS = 7, - KFreeBSD = 8, - Linux = 9, - Lv2 = 10, - MacOSX = 11, - NetBSD = 12, - OpenBSD = 13, - Solaris = 14, - Win32 = 15, - Haiku = 16, - Minix = 17, - RTEMS = 18, - NaCl = 19, - CNK = 20, - AIX = 21, - CUDA = 22, - NVCL = 23, - AMDHSA = 24, - PS4 = 25, - ELFIAMCU = 26, - TvOS = 27, - WatchOS = 28, - Mesa3D = 29, - Contiki = 30, - AMDPAL = 31, - HermitCore = 32, - Hurd = 33, - WASI = 34, - Emscripten = 35, + UnknownOS, + Ananas, + CloudABI, + Darwin, + DragonFly, + FreeBSD, + Fuchsia, + IOS, + KFreeBSD, + Linux, + Lv2, + MacOSX, + NetBSD, + OpenBSD, + Solaris, + Win32, + ZOS, + Haiku, + Minix, + RTEMS, + NaCl, + AIX, + CUDA, + NVCL, + AMDHSA, + PS4, + ELFIAMCU, + TvOS, + WatchOS, + Mesa3D, + Contiki, + AMDPAL, + HermitCore, + Hurd, + WASI, + Emscripten, }; pub const ArchType = extern enum(c_int) { - UnknownArch = 0, - arm = 1, - armeb = 2, - aarch64 = 3, - aarch64_be = 4, - aarch64_32 = 5, - arc = 6, - avr = 7, - bpfel = 8, - bpfeb = 9, - hexagon = 10, - mips = 11, - mipsel = 12, - mips64 = 13, - mips64el = 14, - msp430 = 15, - ppc = 16, - ppc64 = 17, - ppc64le = 18, - r600 = 19, - amdgcn = 20, - riscv32 = 21, - riscv64 = 22, - sparc = 23, - sparcv9 = 24, - sparcel = 25, - systemz = 26, - tce = 27, - tcele = 28, - thumb = 29, - thumbeb = 30, - x86 = 31, - x86_64 = 32, - xcore = 33, - nvptx = 34, - nvptx64 = 35, - le32 = 36, - le64 = 37, - amdil = 38, - amdil64 = 39, - hsail = 40, - hsail64 = 41, - spir = 42, - spir64 = 43, - kalimba = 44, - shave = 45, - lanai = 46, - wasm32 = 47, - wasm64 = 48, - renderscript32 = 49, - renderscript64 = 50, - ve = 51, + UnknownArch, + arm, + armeb, + aarch64, + aarch64_be, + aarch64_32, + arc, + avr, + bpfel, + bpfeb, + csky, + hexagon, + mips, + mipsel, + mips64, + mips64el, + msp430, + ppc, + ppcle, + ppc64, + ppc64le, + r600, + amdgcn, + riscv32, + riscv64, + sparc, + sparcv9, + sparcel, + systemz, + tce, + tcele, + thumb, + thumbeb, + x86, + x86_64, + xcore, + nvptx, + nvptx64, + le32, + le64, + amdil, + amdil64, + hsail, + hsail64, + spir, + spir64, + kalimba, + shave, + lanai, + wasm32, + wasm64, + renderscript32, + renderscript64, + ve, }; pub const ParseCommandLineOptions = ZigLLVMParseCommandLineOptions; diff --git a/src/glibc.zig b/src/glibc.zig index b154bd530d..96154c1690 100644 --- a/src/glibc.zig +++ b/src/glibc.zig @@ -807,26 +807,26 @@ pub fn buildSharedObjects(comp: *Compilation) !void { // .globl _Exit_2_2_5 // .type _Exit_2_2_5, %function; // .symver _Exit_2_2_5, _Exit@@GLIBC_2.2.5 - // .hidden _Exit_2_2_5 // _Exit_2_2_5: const ver_index = ver_list.versions[ver_i]; const ver = metadata.all_versions[ver_index]; const sym_name = libc_fn.name; // Default symbol version definition vs normal symbol version definition - const want_two_ats = chosen_def_ver_index != 255 and ver_index == chosen_def_ver_index; - const at_sign_str = "@@"[0 .. @boolToInt(want_two_ats) + @as(usize, 1)]; - + const want_default = chosen_def_ver_index != 255 and ver_index == chosen_def_ver_index; + const at_sign_str: []const u8 = if (want_default) "@@" else "@"; if (ver.patch == 0) { - const sym_plus_ver = try std.fmt.allocPrint( - arena, - "{s}_{d}_{d}", - .{ sym_name, ver.major, ver.minor }, - ); + const sym_plus_ver = if (want_default) + sym_name + else + try std.fmt.allocPrint( + arena, + "{s}_GLIBC_{d}_{d}", + .{ sym_name, ver.major, ver.minor }, + ); try zig_body.writer().print( \\.globl {s} \\.type {s}, %function; \\.symver {s}, {s}{s}GLIBC_{d}.{d} - \\.hidden {s} \\{s}: \\ , .{ @@ -838,19 +838,20 @@ pub fn buildSharedObjects(comp: *Compilation) !void { ver.major, ver.minor, sym_plus_ver, - sym_plus_ver, }); } else { - const sym_plus_ver = try std.fmt.allocPrint( - arena, - "{s}_{d}_{d}_{d}", - .{ sym_name, ver.major, ver.minor, ver.patch }, - ); + const sym_plus_ver = if (want_default) + sym_name + else + try std.fmt.allocPrint( + arena, + "{s}_GLIBC_{d}_{d}_{d}", + .{ sym_name, ver.major, ver.minor, ver.patch }, + ); try zig_body.writer().print( \\.globl {s} \\.type {s}, %function; \\.symver {s}, {s}{s}GLIBC_{d}.{d}.{d} - \\.hidden {s} \\{s}: \\ , .{ @@ -863,7 +864,6 @@ pub fn buildSharedObjects(comp: *Compilation) !void { ver.minor, ver.patch, sym_plus_ver, - sym_plus_ver, }); } } diff --git a/src/libcxx.zig b/src/libcxx.zig index 11989998e9..e235403d40 100644 --- a/src/libcxx.zig +++ b/src/libcxx.zig @@ -49,6 +49,7 @@ const libcxx_files = [_][]const u8{ "src/future.cpp", "src/hash.cpp", "src/ios.cpp", + "src/ios.instantiations.cpp", "src/iostream.cpp", "src/locale.cpp", "src/memory.cpp", @@ -127,6 +128,7 @@ pub fn buildLibCXX(comp: *Compilation) !void { try cflags.append("-DLIBCXX_BUILDING_LIBCXXABI"); try cflags.append("-D_LIBCXXABI_DISABLE_VISIBILITY_ANNOTATIONS"); try cflags.append("-D_LIBCPP_DISABLE_VISIBILITY_ANNOTATIONS"); + try cflags.append("-D_LIBCPP_HAS_NO_VENDOR_AVAILABILITY_ANNOTATIONS"); if (target.abi.isMusl()) { try cflags.append("-D_LIBCPP_HAS_MUSL_LIBC"); diff --git a/src/stage1/codegen.cpp b/src/stage1/codegen.cpp index 97daf249b1..ec953dfa13 100644 --- a/src/stage1/codegen.cpp +++ b/src/stage1/codegen.cpp @@ -490,14 +490,12 @@ static LLVMValueRef make_fn_llvm_value(CodeGen *g, ZigFn *fn) { if (fn->body_node != nullptr) { maybe_export_dll(g, llvm_fn, linkage); - bool want_fn_safety = g->build_mode != BuildModeFastRelease && + bool want_ssp_attrs = g->build_mode != BuildModeFastRelease && g->build_mode != BuildModeSmallRelease && - !fn->def_scope->safety_off; - if (want_fn_safety) { - if (g->link_libc) { - addLLVMFnAttr(llvm_fn, "sspstrong"); - addLLVMFnAttrStr(llvm_fn, "stack-protector-buffer-size", "4"); - } + g->link_libc; + if (want_ssp_attrs) { + addLLVMFnAttr(llvm_fn, "sspstrong"); + addLLVMFnAttrStr(llvm_fn, "stack-protector-buffer-size", "4"); } if (g->have_stack_probing && !fn->def_scope->safety_off) { addLLVMFnAttrStr(llvm_fn, "probe-stack", "__zig_probe_stack"); @@ -551,7 +549,7 @@ static LLVMValueRef make_fn_llvm_value(CodeGen *g, ZigFn *fn) { } else if (want_first_arg_sret(g, &fn_type->data.fn.fn_type_id)) { // Sret pointers must not be address 0 addLLVMArgAttr(llvm_fn, 0, "nonnull"); - addLLVMArgAttr(llvm_fn, 0, "sret"); + ZigLLVMAddSretAttr(llvm_fn, 0, get_llvm_type(g, return_type)); if (cc_want_sret_attr(cc)) { addLLVMArgAttr(llvm_fn, 0, "noalias"); } @@ -1597,35 +1595,16 @@ static const BuildBinOpFunc unsigned_op[3] = { LLVMBuildNUWAdd, LLVMBuildNUWSub, static LLVMValueRef gen_overflow_op(CodeGen *g, ZigType *operand_type, AddSubMul op, LLVMValueRef val1, LLVMValueRef val2) { - LLVMValueRef overflow_bit; - LLVMValueRef result; - + LLVMValueRef fn_val = get_int_overflow_fn(g, operand_type, op); + LLVMValueRef params[] = { + val1, + val2, + }; + LLVMValueRef result_struct = LLVMBuildCall(g->builder, fn_val, params, 2, ""); + LLVMValueRef result = LLVMBuildExtractValue(g->builder, result_struct, 0, ""); + LLVMValueRef overflow_bit = LLVMBuildExtractValue(g->builder, result_struct, 1, ""); if (operand_type->id == ZigTypeIdVector) { - ZigType *int_type = operand_type->data.vector.elem_type; - assert(int_type->id == ZigTypeIdInt); - LLVMTypeRef one_more_bit_int = LLVMIntType(int_type->data.integral.bit_count + 1); - LLVMTypeRef one_more_bit_int_vector = LLVMVectorType(one_more_bit_int, operand_type->data.vector.len); - const auto buildExtFn = int_type->data.integral.is_signed ? LLVMBuildSExt : LLVMBuildZExt; - LLVMValueRef extended1 = buildExtFn(g->builder, val1, one_more_bit_int_vector, ""); - LLVMValueRef extended2 = buildExtFn(g->builder, val2, one_more_bit_int_vector, ""); - LLVMValueRef extended_result = wrap_op[op](g->builder, extended1, extended2, ""); - result = LLVMBuildTrunc(g->builder, extended_result, get_llvm_type(g, operand_type), ""); - - LLVMValueRef re_extended_result = buildExtFn(g->builder, result, one_more_bit_int_vector, ""); - LLVMValueRef overflow_vector = LLVMBuildICmp(g->builder, LLVMIntNE, extended_result, re_extended_result, ""); - LLVMTypeRef bitcast_int_type = LLVMIntType(operand_type->data.vector.len); - LLVMValueRef bitcasted_overflow = LLVMBuildBitCast(g->builder, overflow_vector, bitcast_int_type, ""); - LLVMValueRef zero = LLVMConstNull(bitcast_int_type); - overflow_bit = LLVMBuildICmp(g->builder, LLVMIntNE, bitcasted_overflow, zero, ""); - } else { - LLVMValueRef fn_val = get_int_overflow_fn(g, operand_type, op); - LLVMValueRef params[] = { - val1, - val2, - }; - LLVMValueRef result_struct = LLVMBuildCall(g->builder, fn_val, params, 2, ""); - result = LLVMBuildExtractValue(g->builder, result_struct, 0, ""); - overflow_bit = LLVMBuildExtractValue(g->builder, result_struct, 1, ""); + overflow_bit = ZigLLVMBuildOrReduce(g->builder, overflow_bit); } LLVMBasicBlockRef fail_block = LLVMAppendBasicBlock(g->cur_fn_val, "OverflowFail"); @@ -1993,7 +1972,7 @@ static bool iter_function_params_c_abi(CodeGen *g, ZigType *fn_type, FnWalk *fn_ switch (fn_walk->id) { case FnWalkIdAttrs: if (abi_class != X64CABIClass_MEMORY_nobyval) { - ZigLLVMAddByValAttr(llvm_fn, fn_walk->data.attrs.gen_i + 1, get_llvm_type(g, ty)); + ZigLLVMAddByValAttr(llvm_fn, fn_walk->data.attrs.gen_i, get_llvm_type(g, ty)); addLLVMArgAttrInt(llvm_fn, fn_walk->data.attrs.gen_i, "align", get_abi_alignment(g, ty)); } else if (g->zig_target->arch == ZigLLVM_aarch64 || g->zig_target->arch == ZigLLVM_aarch64_be) @@ -4058,12 +4037,6 @@ static void gen_set_stack_pointer(CodeGen *g, LLVMValueRef aligned_end_addr) { LLVMBuildCall(g->builder, write_register_fn_val, params, 2, ""); } -static void set_call_instr_sret(CodeGen *g, LLVMValueRef call_instr) { - unsigned attr_kind_id = LLVMGetEnumAttributeKindForName("sret", 4); - LLVMAttributeRef sret_attr = LLVMCreateEnumAttribute(LLVMGetGlobalContext(), attr_kind_id, 0); - LLVMAddCallSiteAttribute(call_instr, 1, sret_attr); -} - static void render_async_spills(CodeGen *g) { ZigType *fn_type = g->cur_fn->type_entry; ZigType *import = get_scope_import(&g->cur_fn->fndef_scope->base); @@ -4587,7 +4560,7 @@ static LLVMValueRef ir_render_call(CodeGen *g, IrExecutableGen *executable, IrIn } else if (!ret_has_bits) { return nullptr; } else if (first_arg_ret) { - set_call_instr_sret(g, result); + ZigLLVMSetCallSret(result, get_llvm_type(g, src_return_type)); return result_loc; } else if (handle_is_ptr(g, src_return_type)) { LLVMValueRef store_instr = LLVMBuildStore(g->builder, result, result_loc); diff --git a/src/stage1/stage1.h b/src/stage1/stage1.h index 7fa1576d64..59632b9877 100644 --- a/src/stage1/stage1.h +++ b/src/stage1/stage1.h @@ -73,11 +73,11 @@ enum Os { OsOpenBSD, OsSolaris, OsWindows, + OsZOS, OsHaiku, OsMinix, OsRTEMS, OsNaCl, // Native Client - OsCNK, // BG/P Compute-Node Kernel OsAIX, OsCUDA, // NVIDIA CUDA OsNVCL, // NVIDIA OpenCL diff --git a/src/stage1/target.cpp b/src/stage1/target.cpp index f2fbc30da8..82b78e113e 100644 --- a/src/stage1/target.cpp +++ b/src/stage1/target.cpp @@ -23,6 +23,7 @@ static const ZigLLVM_ArchType arch_list[] = { ZigLLVM_avr, // AVR: Atmel AVR microcontroller ZigLLVM_bpfel, // eBPF or extended BPF or 64-bit BPF (little endian) ZigLLVM_bpfeb, // eBPF or extended BPF or 64-bit BPF (big endian) + ZigLLVM_csky, // CSKY: csky ZigLLVM_hexagon, // Hexagon: hexagon ZigLLVM_mips, // MIPS: mips, mipsallegrex, mipsr6 ZigLLVM_mipsel, // MIPSEL: mipsel, mipsallegrexe, mipsr6el @@ -30,6 +31,7 @@ static const ZigLLVM_ArchType arch_list[] = { ZigLLVM_mips64el, // MIPS64EL: mips64el, mips64r6el, mipsn32el, mipsn32r6el ZigLLVM_msp430, // MSP430: msp430 ZigLLVM_ppc, // PPC: powerpc + ZigLLVM_ppcle, // PPCLE: powerpc (little endian) ZigLLVM_ppc64, // PPC64: powerpc64, ppu ZigLLVM_ppc64le, // PPC64LE: powerpc64le ZigLLVM_r600, // R600: AMD GPUs HD2XXX - HD6XXX @@ -71,8 +73,6 @@ static const ZigLLVM_VendorType vendor_list[] = { ZigLLVM_Apple, ZigLLVM_PC, ZigLLVM_SCEI, - ZigLLVM_BGP, - ZigLLVM_BGQ, ZigLLVM_Freescale, ZigLLVM_IBM, ZigLLVM_ImaginationTechnologies, @@ -101,11 +101,11 @@ static const Os os_list[] = { OsOpenBSD, OsSolaris, OsWindows, + OsZOS, OsHaiku, OsMinix, OsRTEMS, OsNaCl, // Native Client - OsCNK, // BG/P Compute-Node Kernel OsAIX, OsCUDA, // NVIDIA CUDA OsNVCL, // NVIDIA OpenCL @@ -135,6 +135,7 @@ static const ZigLLVM_EnvironmentType abi_list[] = { ZigLLVM_GNUEABI, ZigLLVM_GNUEABIHF, ZigLLVM_GNUX32, + ZigLLVM_GNUILP32, ZigLLVM_CODE16, ZigLLVM_EABI, ZigLLVM_EABIHF, @@ -155,8 +156,10 @@ static const ZigLLVM_ObjectFormatType oformat_list[] = { ZigLLVM_UnknownObjectFormat, ZigLLVM_COFF, ZigLLVM_ELF, + ZigLLVM_GOFF, ZigLLVM_MachO, ZigLLVM_Wasm, + ZigLLVM_XCOFF, }; size_t target_oformat_count(void) { @@ -173,6 +176,7 @@ const char *target_oformat_name(ZigLLVM_ObjectFormatType oformat) { case ZigLLVM_UnknownObjectFormat: return "unknown"; case ZigLLVM_COFF: return "coff"; case ZigLLVM_ELF: return "elf"; + case ZigLLVM_GOFF: return "goff"; case ZigLLVM_MachO: return "macho"; case ZigLLVM_Wasm: return "wasm"; case ZigLLVM_XCOFF: return "xcoff"; @@ -240,6 +244,8 @@ ZigLLVM_OSType get_llvm_os_type(Os os_type) { case OsWindows: case OsUefi: return ZigLLVM_Win32; + case OsZOS: + return ZigLLVM_ZOS; case OsHaiku: return ZigLLVM_Haiku; case OsMinix: @@ -248,8 +254,6 @@ ZigLLVM_OSType get_llvm_os_type(Os os_type) { return ZigLLVM_RTEMS; case OsNaCl: return ZigLLVM_NaCl; - case OsCNK: - return ZigLLVM_CNK; case OsAIX: return ZigLLVM_AIX; case OsCUDA: @@ -306,11 +310,11 @@ const char *target_os_name(Os os_type) { case OsOpenBSD: case OsSolaris: case OsWindows: + case OsZOS: case OsHaiku: case OsMinix: case OsRTEMS: case OsNaCl: // Native Client - case OsCNK: // BG/P Compute-Node Kernel case OsAIX: case OsCUDA: // NVIDIA CUDA case OsNVCL: // NVIDIA OpenCL @@ -485,6 +489,7 @@ uint32_t target_arch_pointer_bit_width(ZigLLVM_ArchType arch) { case ZigLLVM_mipsel: case ZigLLVM_nvptx: case ZigLLVM_ppc: + case ZigLLVM_ppcle: case ZigLLVM_r600: case ZigLLVM_riscv32: case ZigLLVM_sparc: @@ -504,6 +509,7 @@ uint32_t target_arch_pointer_bit_width(ZigLLVM_ArchType arch) { case ZigLLVM_wasm32: case ZigLLVM_renderscript32: case ZigLLVM_aarch64_32: + case ZigLLVM_csky: return 32; case ZigLLVM_aarch64: @@ -550,6 +556,7 @@ uint32_t target_arch_largest_atomic_bits(ZigLLVM_ArchType arch) { case ZigLLVM_mipsel: case ZigLLVM_nvptx: case ZigLLVM_ppc: + case ZigLLVM_ppcle: case ZigLLVM_r600: case ZigLLVM_riscv32: case ZigLLVM_sparc: @@ -568,6 +575,7 @@ uint32_t target_arch_largest_atomic_bits(ZigLLVM_ArchType arch) { case ZigLLVM_shave: case ZigLLVM_wasm32: case ZigLLVM_renderscript32: + case ZigLLVM_csky: return 32; case ZigLLVM_aarch64: @@ -707,10 +715,10 @@ uint32_t target_c_type_size_in_bits(const ZigTarget *target, CIntType id) { case OsKFreeBSD: case OsLv2: case OsSolaris: + case OsZOS: case OsMinix: case OsRTEMS: case OsNaCl: - case OsCNK: case OsAIX: case OsCUDA: case OsNVCL: @@ -799,6 +807,7 @@ const char *arch_stack_pointer_register_name(ZigLLVM_ArchType arch) { case ZigLLVM_riscv64: case ZigLLVM_mipsel: case ZigLLVM_ppc: + case ZigLLVM_ppcle: case ZigLLVM_ppc64: case ZigLLVM_ppc64le: return "sp"; @@ -814,6 +823,7 @@ const char *arch_stack_pointer_register_name(ZigLLVM_ArchType arch) { case ZigLLVM_avr: case ZigLLVM_bpfeb: case ZigLLVM_bpfel: + case ZigLLVM_csky: case ZigLLVM_hexagon: case ZigLLVM_lanai: case ZigLLVM_hsail: @@ -868,6 +878,7 @@ bool target_is_arm(const ZigTarget *target) { case ZigLLVM_avr: case ZigLLVM_bpfeb: case ZigLLVM_bpfel: + case ZigLLVM_csky: case ZigLLVM_hexagon: case ZigLLVM_lanai: case ZigLLVM_hsail: @@ -901,6 +912,7 @@ bool target_is_arm(const ZigTarget *target) { case ZigLLVM_wasm64: case ZigLLVM_xcore: case ZigLLVM_ppc: + case ZigLLVM_ppcle: case ZigLLVM_ppc64: case ZigLLVM_ve: return false; @@ -951,11 +963,10 @@ ZigLLVM_EnvironmentType target_default_abi(ZigLLVM_ArchType arch, Os os) { case OsCloudABI: case OsLv2: case OsSolaris: - case OsHaiku: + case OsZOS: case OsMinix: case OsRTEMS: case OsNaCl: - case OsCNK: case OsAIX: case OsCUDA: case OsNVCL: @@ -979,6 +990,7 @@ ZigLLVM_EnvironmentType target_default_abi(ZigLLVM_ArchType arch, Os os) { case OsNetBSD: case OsDragonFly: case OsHurd: + case OsHaiku: return ZigLLVM_GNU; case OsUefi: case OsWindows: @@ -1094,6 +1106,7 @@ const char *target_libc_generic_name(const ZigTarget *target) { case ZigLLVM_GNUEABI: case ZigLLVM_GNUEABIHF: case ZigLLVM_GNUX32: + case ZigLLVM_GNUILP32: return "glibc"; case ZigLLVM_Musl: case ZigLLVM_MuslEABI: diff --git a/src/stage1/zig0.cpp b/src/stage1/zig0.cpp index bcc9dbc00a..e16f79455a 100644 --- a/src/stage1/zig0.cpp +++ b/src/stage1/zig0.cpp @@ -94,6 +94,8 @@ static Os get_zig_os_type(ZigLLVM_OSType os_type) { return OsSolaris; case ZigLLVM_Win32: return OsWindows; + case ZigLLVM_ZOS: + return OsZOS; case ZigLLVM_Haiku: return OsHaiku; case ZigLLVM_Minix: @@ -102,8 +104,6 @@ static Os get_zig_os_type(ZigLLVM_OSType os_type) { return OsRTEMS; case ZigLLVM_NaCl: return OsNaCl; - case ZigLLVM_CNK: - return OsCNK; case ZigLLVM_AIX: return OsAIX; case ZigLLVM_CUDA: diff --git a/src/target.zig b/src/target.zig index 5a884271cc..c5b87bb242 100644 --- a/src/target.zig +++ b/src/target.zig @@ -71,6 +71,7 @@ pub fn libCGenericName(target: std.Target) [:0]const u8 { .gnueabi, .gnueabihf, .gnux32, + .gnuilp32, => return "glibc", .musl, .musleabi, @@ -204,11 +205,11 @@ pub fn osToLLVM(os_tag: std.Target.Os.Tag) llvm.OSType { .netbsd => .NetBSD, .openbsd => .OpenBSD, .solaris => .Solaris, + .zos => .ZOS, .haiku => .Haiku, .minix => .Minix, .rtems => .RTEMS, .nacl => .NaCl, - .cnk => .CNK, .aix => .AIX, .cuda => .CUDA, .nvcl => .NVCL, @@ -238,6 +239,7 @@ pub fn archToLLVM(arch_tag: std.Target.Cpu.Arch) llvm.ArchType { .avr => .avr, .bpfel => .bpfel, .bpfeb => .bpfeb, + .csky => .csky, .hexagon => .hexagon, .mips => .mips, .mipsel => .mipsel, @@ -245,6 +247,7 @@ pub fn archToLLVM(arch_tag: std.Target.Cpu.Arch) llvm.ArchType { .mips64el => .mips64el, .msp430 => .msp430, .powerpc => .ppc, + .powerpcle => .ppcle, .powerpc64 => .ppc64, .powerpc64le => .ppc64le, .r600 => .r600, diff --git a/src/translate_c.zig b/src/translate_c.zig index ed80a13058..bb38e8ad55 100644 --- a/src/translate_c.zig +++ b/src/translate_c.zig @@ -2501,7 +2501,7 @@ fn transSwitchProngStmtInline( fn transConstantExpr(c: *Context, scope: *Scope, expr: *const clang.Expr, used: ResultUsed) TransError!Node { var result: clang.ExprEvalResult = undefined; - if (!expr.evaluateAsConstantExpr(&result, .EvaluateForCodeGen, c.clang_context)) + if (!expr.evaluateAsConstantExpr(&result, .Normal, c.clang_context)) return fail(c, error.UnsupportedTranslation, expr.getBeginLoc(), "invalid constant expression", .{}); switch (result.Val.getKind()) { diff --git a/src/type.zig b/src/type.zig index 38fe6dd3e6..30605ce67b 100644 --- a/src/type.zig +++ b/src/type.zig @@ -3646,11 +3646,11 @@ pub const CType = enum { .kfreebsd, .lv2, .solaris, + .zos, .haiku, .minix, .rtems, .nacl, - .cnk, .aix, .cuda, .nvcl, diff --git a/src/zig_clang.cpp b/src/zig_clang.cpp index 0ddb597371..6bc612bd8d 100644 --- a/src/zig_clang.cpp +++ b/src/zig_clang.cpp @@ -182,6 +182,8 @@ void ZigClang_detect_enum_CK(clang::CastKind x) { case clang::CK_IntegralToBoolean: case clang::CK_IntegralToFloating: case clang::CK_IntegralToPointer: + case clang::CK_FloatingToFixedPoint: + case clang::CK_FixedPointToFloating: case clang::CK_LValueBitCast: case clang::CK_LValueToRValueBitCast: case clang::CK_LValueToRValue: @@ -237,6 +239,8 @@ static_assert((clang::CastKind)ZigClangCK_VectorSplat == clang::CK_VectorSplat, static_assert((clang::CastKind)ZigClangCK_IntegralCast == clang::CK_IntegralCast, ""); static_assert((clang::CastKind)ZigClangCK_IntegralToBoolean == clang::CK_IntegralToBoolean, ""); static_assert((clang::CastKind)ZigClangCK_IntegralToFloating == clang::CK_IntegralToFloating, ""); +static_assert((clang::CastKind)ZigClangCK_FloatingToFixedPoint == clang::CK_FloatingToFixedPoint, ""); +static_assert((clang::CastKind)ZigClangCK_FixedPointToFloating == clang::CK_FixedPointToFloating, ""); static_assert((clang::CastKind)ZigClangCK_FixedPointCast == clang::CK_FixedPointCast, ""); static_assert((clang::CastKind)ZigClangCK_FixedPointToIntegral == clang::CK_FixedPointToIntegral, ""); static_assert((clang::CastKind)ZigClangCK_IntegralToFixedPoint == clang::CK_IntegralToFixedPoint, ""); @@ -928,6 +932,7 @@ void ZigClang_detect_enum_DeclKind(clang::Decl::Kind x) { case clang::Decl::MSGuid: case clang::Decl::OMPDeclareMapper: case clang::Decl::OMPDeclareReduction: + case clang::Decl::TemplateParamObject: case clang::Decl::UnresolvedUsingValue: case clang::Decl::OMPAllocate: case clang::Decl::OMPRequires: @@ -1013,6 +1018,7 @@ static_assert((clang::Decl::Kind)ZigClangDeclIndirectField == clang::Decl::Indir static_assert((clang::Decl::Kind)ZigClangDeclMSGuid == clang::Decl::MSGuid, ""); static_assert((clang::Decl::Kind)ZigClangDeclOMPDeclareMapper == clang::Decl::OMPDeclareMapper, ""); static_assert((clang::Decl::Kind)ZigClangDeclOMPDeclareReduction == clang::Decl::OMPDeclareReduction, ""); +static_assert((clang::Decl::Kind)ZigClangDeclTemplateParamObject == clang::Decl::TemplateParamObject, ""); static_assert((clang::Decl::Kind)ZigClangDeclUnresolvedUsingValue == clang::Decl::UnresolvedUsingValue, ""); static_assert((clang::Decl::Kind)ZigClangDeclOMPRequires == clang::Decl::OMPRequires, ""); static_assert((clang::Decl::Kind)ZigClangDeclOMPThreadPrivate == clang::Decl::OMPThreadPrivate, ""); @@ -1122,6 +1128,8 @@ void ZigClang_detect_enum_BuiltinTypeKind(clang::BuiltinType::Kind x) { case clang::BuiltinType::SveFloat64x4: case clang::BuiltinType::SveBFloat16x4: case clang::BuiltinType::SveBool: + case clang::BuiltinType::VectorQuad: + case clang::BuiltinType::VectorPair: case clang::BuiltinType::Void: case clang::BuiltinType::Bool: case clang::BuiltinType::Char_U: @@ -1295,6 +1303,8 @@ static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeSveFloat32x4 == clang static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeSveFloat64x4 == clang::BuiltinType::SveFloat64x4, ""); static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeSveBFloat16x4 == clang::BuiltinType::SveBFloat16x4, ""); static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeSveBool == clang::BuiltinType::SveBool, ""); +static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeVectorQuad == clang::BuiltinType::VectorQuad, ""); +static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeVectorPair == clang::BuiltinType::VectorPair, ""); static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeVoid == clang::BuiltinType::Void, ""); static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeBool == clang::BuiltinType::Bool, ""); static_assert((clang::BuiltinType::Kind)ZigClangBuiltinTypeChar_U == clang::BuiltinType::Char_U, ""); @@ -1517,15 +1527,19 @@ static_assert((clang::PreprocessedEntity::EntityKind)ZigClangPreprocessedEntity_ static_assert((clang::PreprocessedEntity::EntityKind)ZigClangPreprocessedEntity_InclusionDirectiveKind == clang::PreprocessedEntity::InclusionDirectiveKind, ""); -void ZigClang_detect_enum_ConstExprUsage(clang::Expr::ConstExprUsage x) { +void ZigClang_detect_enum_ConstantExprKind(clang::Expr::ConstantExprKind x) { switch (x) { - case clang::Expr::EvaluateForCodeGen: - case clang::Expr::EvaluateForMangling: + case clang::Expr::ConstantExprKind::Normal: + case clang::Expr::ConstantExprKind::NonClassTemplateArgument: + case clang::Expr::ConstantExprKind::ClassTemplateArgument: + case clang::Expr::ConstantExprKind::ImmediateInvocation: break; } } -static_assert((clang::Expr::ConstExprUsage)ZigClangExpr_EvaluateForCodeGen == clang::Expr::EvaluateForCodeGen, ""); -static_assert((clang::Expr::ConstExprUsage)ZigClangExpr_EvaluateForMangling == clang::Expr::EvaluateForMangling, ""); +static_assert((clang::Expr::ConstantExprKind)ZigClangExpr_ContantExprKind_Normal == clang::Expr::ConstantExprKind::Normal, ""); +static_assert((clang::Expr::ConstantExprKind)ZigClangExpr_ContantExprKind_NonClassTemplateArgument == clang::Expr::ConstantExprKind::NonClassTemplateArgument, ""); +static_assert((clang::Expr::ConstantExprKind)ZigClangExpr_ContantExprKind_ClassTemplateArgument == clang::Expr::ConstantExprKind::ClassTemplateArgument, ""); +static_assert((clang::Expr::ConstantExprKind)ZigClangExpr_ContantExprKind_ImmediateInvocation == clang::Expr::ConstantExprKind::ImmediateInvocation, ""); static_assert(sizeof(ZigClangAPValue) == sizeof(clang::APValue), ""); @@ -2143,12 +2157,12 @@ bool ZigClangExpr_EvaluateAsFloat(const ZigClangExpr *self, ZigClangAPFloat **re } bool ZigClangExpr_EvaluateAsConstantExpr(const ZigClangExpr *self, ZigClangExprEvalResult *result, - ZigClangExpr_ConstExprUsage usage, const struct ZigClangASTContext *ctx) + ZigClangExpr_ConstantExprKind kind, const struct ZigClangASTContext *ctx) { auto casted_self = reinterpret_cast<const clang::Expr *>(self); auto casted_ctx = reinterpret_cast<const clang::ASTContext *>(ctx); clang::Expr::EvalResult eval_result; - if (!casted_self->EvaluateAsConstantExpr(eval_result, (clang::Expr::ConstExprUsage)usage, *casted_ctx)) { + if (!casted_self->EvaluateAsConstantExpr(eval_result, *casted_ctx, (clang::Expr::ConstantExprKind)kind)) { return false; } *result = bitcast(eval_result); diff --git a/src/zig_clang.h b/src/zig_clang.h index b3717946ea..476feb2a1e 100644 --- a/src/zig_clang.h +++ b/src/zig_clang.h @@ -541,6 +541,8 @@ enum ZigClangCK { ZigClangCK_IntegralCast, ZigClangCK_IntegralToBoolean, ZigClangCK_IntegralToFloating, + ZigClangCK_FloatingToFixedPoint, + ZigClangCK_FixedPointToFloating, ZigClangCK_FixedPointCast, ZigClangCK_FixedPointToIntegral, ZigClangCK_IntegralToFixedPoint, @@ -647,6 +649,7 @@ enum ZigClangDeclKind { ZigClangDeclMSGuid, ZigClangDeclOMPDeclareMapper, ZigClangDeclOMPDeclareReduction, + ZigClangDeclTemplateParamObject, ZigClangDeclUnresolvedUsingValue, ZigClangDeclOMPAllocate, ZigClangDeclOMPRequires, @@ -757,6 +760,8 @@ enum ZigClangBuiltinTypeKind { ZigClangBuiltinTypeSveFloat64x4, ZigClangBuiltinTypeSveBFloat16x4, ZigClangBuiltinTypeSveBool, + ZigClangBuiltinTypeVectorQuad, + ZigClangBuiltinTypeVectorPair, ZigClangBuiltinTypeVoid, ZigClangBuiltinTypeBool, ZigClangBuiltinTypeChar_U, @@ -915,9 +920,11 @@ enum ZigClangPreprocessedEntity_EntityKind { ZigClangPreprocessedEntity_InclusionDirectiveKind, }; -enum ZigClangExpr_ConstExprUsage { - ZigClangExpr_EvaluateForCodeGen, - ZigClangExpr_EvaluateForMangling, +enum ZigClangExpr_ConstantExprKind { + ZigClangExpr_ContantExprKind_Normal, + ZigClangExpr_ContantExprKind_NonClassTemplateArgument, + ZigClangExpr_ContantExprKind_ClassTemplateArgument, + ZigClangExpr_ContantExprKind_ImmediateInvocation, }; enum ZigClangUnaryExprOrTypeTrait_Kind { @@ -1083,7 +1090,7 @@ ZIG_EXTERN_C bool ZigClangExpr_EvaluateAsBooleanCondition(const struct ZigClangE ZIG_EXTERN_C bool ZigClangExpr_EvaluateAsFloat(const struct ZigClangExpr *self, ZigClangAPFloat **result, const struct ZigClangASTContext *ctx); ZIG_EXTERN_C bool ZigClangExpr_EvaluateAsConstantExpr(const struct ZigClangExpr *, - struct ZigClangExprEvalResult *, ZigClangExpr_ConstExprUsage, const struct ZigClangASTContext *); + struct ZigClangExprEvalResult *, ZigClangExpr_ConstantExprKind, const struct ZigClangASTContext *); ZIG_EXTERN_C const ZigClangExpr *ZigClangInitListExpr_getInit(const ZigClangInitListExpr *, unsigned); ZIG_EXTERN_C const ZigClangExpr *ZigClangInitListExpr_getArrayFiller(const ZigClangInitListExpr *); diff --git a/src/zig_clang_cc1_main.cpp b/src/zig_clang_cc1_main.cpp index 0872015e0a..0918860015 100644 --- a/src/zig_clang_cc1_main.cpp +++ b/src/zig_clang_cc1_main.cpp @@ -251,8 +251,7 @@ int cc1_main(ArrayRef<const char *> Argv, const char *Argv0, void *MainAddr) { if (auto profilerOutput = Clang->createOutputFile(Path.str(), /*Binary=*/false, - /*RemoveFileOnSignal=*/false, "", - /*Extension=*/"json", + /*RemoveFileOnSignal=*/false, /*useTemporary=*/false)) { llvm::timeTraceProfilerWrite(*profilerOutput); diff --git a/src/zig_clang_cc1as_main.cpp b/src/zig_clang_cc1as_main.cpp index 77b99b2013..de71026fbf 100644 --- a/src/zig_clang_cc1as_main.cpp +++ b/src/zig_clang_cc1as_main.cpp @@ -221,19 +221,13 @@ bool AssemblerInvocation::CreateFromArgs(AssemblerInvocation &Opts, // Any DebugInfoKind implies GenDwarfForAssembly. Opts.GenDwarfForAssembly = Args.hasArg(OPT_debug_info_kind_EQ); - if (const Arg *A = Args.getLastArg(OPT_compress_debug_sections, - OPT_compress_debug_sections_EQ)) { - if (A->getOption().getID() == OPT_compress_debug_sections) { - // TODO: be more clever about the compression type auto-detection - Opts.CompressDebugSections = llvm::DebugCompressionType::GNU; - } else { - Opts.CompressDebugSections = - llvm::StringSwitch<llvm::DebugCompressionType>(A->getValue()) - .Case("none", llvm::DebugCompressionType::None) - .Case("zlib", llvm::DebugCompressionType::Z) - .Case("zlib-gnu", llvm::DebugCompressionType::GNU) - .Default(llvm::DebugCompressionType::None); - } + if (const Arg *A = Args.getLastArg(OPT_compress_debug_sections_EQ)) { + Opts.CompressDebugSections = + llvm::StringSwitch<llvm::DebugCompressionType>(A->getValue()) + .Case("none", llvm::DebugCompressionType::None) + .Case("zlib", llvm::DebugCompressionType::Z) + .Case("zlib-gnu", llvm::DebugCompressionType::GNU) + .Default(llvm::DebugCompressionType::None); } Opts.RelaxELFRelocations = Args.hasArg(OPT_mrelax_relocations); @@ -434,8 +428,11 @@ static bool ExecuteAssembler(AssemblerInvocation &Opts, std::unique_ptr<MCStreamer> Str; std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo()); + assert(MCII && "Unable to create instruction info!"); + std::unique_ptr<MCSubtargetInfo> STI( TheTarget->createMCSubtargetInfo(Opts.Triple, Opts.CPU, FS)); + assert(STI && "Unable to create subtarget info!"); raw_pwrite_stream *Out = FDOS.get(); std::unique_ptr<buffer_ostream> BOS; @@ -474,6 +471,8 @@ static bool ExecuteAssembler(AssemblerInvocation &Opts, TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx)); std::unique_ptr<MCAsmBackend> MAB( TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions)); + assert(MAB && "Unable to create asm backend!"); + std::unique_ptr<MCObjectWriter> OW = DwoOS ? MAB->createDwoObjectWriter(*Out, *DwoOS) : MAB->createObjectWriter(*Out); @@ -526,8 +525,8 @@ static bool ExecuteAssembler(AssemblerInvocation &Opts, Failed = Parser->Run(Opts.NoInitialTextSection); } - // Close Streamer first. - // It might have a reference to the output stream. + // Parser has a reference to the output stream (Str), so close Parser first. + Parser.reset(); Str.reset(); // Close the output stream early. BOS.reset(); diff --git a/src/zig_clang_driver.cpp b/src/zig_clang_driver.cpp index fbe407a06c..ac892f95e8 100644 --- a/src/zig_clang_driver.cpp +++ b/src/zig_clang_driver.cpp @@ -529,6 +529,13 @@ int ZigClang_main(int argc_, const char **argv_) { #ifdef _WIN32 IsCrash |= CommandRes == 3; #endif +#if LLVM_ON_UNIX + // When running in integrated-cc1 mode, the CrashRecoveryContext returns + // the same codes as if the program crashed. See section "Exit Status for + // Commands": + // https://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xcu_chap02.html + IsCrash |= CommandRes > 128; +#endif if (IsCrash) { TheDriver.generateCompilationDiagnostics(*C, *FailingCommand); break; diff --git a/src/zig_llvm.cpp b/src/zig_llvm.cpp index 280920ac74..5a7d60f83b 100644 --- a/src/zig_llvm.cpp +++ b/src/zig_llvm.cpp @@ -614,8 +614,9 @@ void ZigLLVMDisposeDIBuilder(ZigLLVMDIBuilder *dbuilder) { } void ZigLLVMSetCurrentDebugLocation(LLVMBuilderRef builder, int line, int column, ZigLLVMDIScope *scope) { - unwrap(builder)->SetCurrentDebugLocation(DebugLoc::get( - line, column, reinterpret_cast<DIScope*>(scope))); + DIScope* di_scope = reinterpret_cast<DIScope*>(scope); + DebugLoc debug_loc = DILocation::get(di_scope->getContext(), line, column, di_scope, nullptr, false); + unwrap(builder)->SetCurrentDebugLocation(debug_loc); } void ZigLLVMClearCurrentDebugLocation(LLVMBuilderRef builder) { @@ -776,7 +777,8 @@ LLVMValueRef ZigLLVMInsertDeclare(ZigLLVMDIBuilder *dibuilder, LLVMValueRef stor } ZigLLVMDILocation *ZigLLVMGetDebugLoc(unsigned line, unsigned col, ZigLLVMDIScope *scope) { - DebugLoc debug_loc = DebugLoc::get(line, col, reinterpret_cast<DIScope*>(scope), nullptr); + DIScope* di_scope = reinterpret_cast<DIScope*>(scope); + DebugLoc debug_loc = DILocation::get(di_scope->getContext(), line, col, di_scope, nullptr, false); return reinterpret_cast<ZigLLVMDILocation*>(debug_loc.get()); } @@ -796,7 +798,17 @@ void ZigLLVMAddByValAttr(LLVMValueRef fn_ref, unsigned ArgNo, LLVMTypeRef type_v AttrBuilder attr_builder; Type *llvm_type = unwrap<Type>(type_val); attr_builder.addByValAttr(llvm_type); - const AttributeList new_attr_set = attr_set.addAttributes(func->getContext(), ArgNo, attr_builder); + const AttributeList new_attr_set = attr_set.addAttributes(func->getContext(), ArgNo + 1, attr_builder); + func->setAttributes(new_attr_set); +} + +void ZigLLVMAddSretAttr(LLVMValueRef fn_ref, unsigned ArgNo, LLVMTypeRef type_val) { + Function *func = unwrap<Function>(fn_ref); + const AttributeList attr_set = func->getAttributes(); + AttrBuilder attr_builder; + Type *llvm_type = unwrap<Type>(type_val); + attr_builder.addStructRetAttr(llvm_type); + const AttributeList new_attr_set = attr_set.addAttributes(func->getContext(), ArgNo + 1, attr_builder); func->setAttributes(new_attr_set); } @@ -928,6 +940,15 @@ void ZigLLVMSetTailCall(LLVMValueRef Call) { unwrap<CallInst>(Call)->setTailCallKind(CallInst::TCK_MustTail); } +void ZigLLVMSetCallSret(LLVMValueRef Call, LLVMTypeRef return_type) { + const AttributeList attr_set = unwrap<CallInst>(Call)->getAttributes(); + AttrBuilder attr_builder; + Type *llvm_type = unwrap<Type>(return_type); + attr_builder.addStructRetAttr(llvm_type); + const AttributeList new_attr_set = attr_set.addAttributes(unwrap<CallInst>(Call)->getContext(), 1, attr_builder); + unwrap<CallInst>(Call)->setAttributes(new_attr_set); +} + void ZigLLVMFunctionSetPrefixData(LLVMValueRef function, LLVMValueRef data) { unwrap<Function>(function)->setPrefixData(unwrap<Constant>(data)); } @@ -1194,6 +1215,7 @@ static_assert((Triple::ArchType)ZigLLVM_arc == Triple::arc, ""); static_assert((Triple::ArchType)ZigLLVM_avr == Triple::avr, ""); static_assert((Triple::ArchType)ZigLLVM_bpfel == Triple::bpfel, ""); static_assert((Triple::ArchType)ZigLLVM_bpfeb == Triple::bpfeb, ""); +static_assert((Triple::ArchType)ZigLLVM_csky == Triple::csky, ""); static_assert((Triple::ArchType)ZigLLVM_hexagon == Triple::hexagon, ""); static_assert((Triple::ArchType)ZigLLVM_mips == Triple::mips, ""); static_assert((Triple::ArchType)ZigLLVM_mipsel == Triple::mipsel, ""); @@ -1201,6 +1223,7 @@ static_assert((Triple::ArchType)ZigLLVM_mips64 == Triple::mips64, ""); static_assert((Triple::ArchType)ZigLLVM_mips64el == Triple::mips64el, ""); static_assert((Triple::ArchType)ZigLLVM_msp430 == Triple::msp430, ""); static_assert((Triple::ArchType)ZigLLVM_ppc == Triple::ppc, ""); +static_assert((Triple::ArchType)ZigLLVM_ppcle == Triple::ppcle, ""); static_assert((Triple::ArchType)ZigLLVM_ppc64 == Triple::ppc64, ""); static_assert((Triple::ArchType)ZigLLVM_ppc64le == Triple::ppc64le, ""); static_assert((Triple::ArchType)ZigLLVM_r600 == Triple::r600, ""); @@ -1242,8 +1265,6 @@ static_assert((Triple::VendorType)ZigLLVM_UnknownVendor == Triple::UnknownVendor static_assert((Triple::VendorType)ZigLLVM_Apple == Triple::Apple, ""); static_assert((Triple::VendorType)ZigLLVM_PC == Triple::PC, ""); static_assert((Triple::VendorType)ZigLLVM_SCEI == Triple::SCEI, ""); -static_assert((Triple::VendorType)ZigLLVM_BGP == Triple::BGP, ""); -static_assert((Triple::VendorType)ZigLLVM_BGQ == Triple::BGQ, ""); static_assert((Triple::VendorType)ZigLLVM_Freescale == Triple::Freescale, ""); static_assert((Triple::VendorType)ZigLLVM_IBM == Triple::IBM, ""); static_assert((Triple::VendorType)ZigLLVM_ImaginationTechnologies == Triple::ImaginationTechnologies, ""); @@ -1275,11 +1296,11 @@ static_assert((Triple::OSType)ZigLLVM_NetBSD == Triple::NetBSD, ""); static_assert((Triple::OSType)ZigLLVM_OpenBSD == Triple::OpenBSD, ""); static_assert((Triple::OSType)ZigLLVM_Solaris == Triple::Solaris, ""); static_assert((Triple::OSType)ZigLLVM_Win32 == Triple::Win32, ""); +static_assert((Triple::OSType)ZigLLVM_ZOS == Triple::ZOS, ""); static_assert((Triple::OSType)ZigLLVM_Haiku == Triple::Haiku, ""); static_assert((Triple::OSType)ZigLLVM_Minix == Triple::Minix, ""); static_assert((Triple::OSType)ZigLLVM_RTEMS == Triple::RTEMS, ""); static_assert((Triple::OSType)ZigLLVM_NaCl == Triple::NaCl, ""); -static_assert((Triple::OSType)ZigLLVM_CNK == Triple::CNK, ""); static_assert((Triple::OSType)ZigLLVM_AIX == Triple::AIX, ""); static_assert((Triple::OSType)ZigLLVM_CUDA == Triple::CUDA, ""); static_assert((Triple::OSType)ZigLLVM_NVCL == Triple::NVCL, ""); @@ -1304,6 +1325,7 @@ static_assert((Triple::EnvironmentType)ZigLLVM_GNUABI64 == Triple::GNUABI64, "") static_assert((Triple::EnvironmentType)ZigLLVM_GNUEABI == Triple::GNUEABI, ""); static_assert((Triple::EnvironmentType)ZigLLVM_GNUEABIHF == Triple::GNUEABIHF, ""); static_assert((Triple::EnvironmentType)ZigLLVM_GNUX32 == Triple::GNUX32, ""); +static_assert((Triple::EnvironmentType)ZigLLVM_GNUILP32 == Triple::GNUILP32, ""); static_assert((Triple::EnvironmentType)ZigLLVM_CODE16 == Triple::CODE16, ""); static_assert((Triple::EnvironmentType)ZigLLVM_EABI == Triple::EABI, ""); static_assert((Triple::EnvironmentType)ZigLLVM_EABIHF == Triple::EABIHF, ""); @@ -1322,6 +1344,7 @@ static_assert((Triple::EnvironmentType)ZigLLVM_LastEnvironmentType == Triple::La static_assert((Triple::ObjectFormatType)ZigLLVM_UnknownObjectFormat == Triple::UnknownObjectFormat, ""); static_assert((Triple::ObjectFormatType)ZigLLVM_COFF == Triple::COFF, ""); static_assert((Triple::ObjectFormatType)ZigLLVM_ELF == Triple::ELF, ""); +static_assert((Triple::ObjectFormatType)ZigLLVM_GOFF == Triple::GOFF, ""); static_assert((Triple::ObjectFormatType)ZigLLVM_MachO == Triple::MachO, ""); static_assert((Triple::ObjectFormatType)ZigLLVM_Wasm == Triple::Wasm, ""); static_assert((Triple::ObjectFormatType)ZigLLVM_XCOFF == Triple::XCOFF, ""); diff --git a/src/zig_llvm.h b/src/zig_llvm.h index 504ea4ec01..0d08980835 100644 --- a/src/zig_llvm.h +++ b/src/zig_llvm.h @@ -264,11 +264,13 @@ ZIG_EXTERN_C struct ZigLLVMDILocation *ZigLLVMGetDebugLoc(unsigned line, unsigne ZIG_EXTERN_C void ZigLLVMSetFastMath(LLVMBuilderRef builder_wrapped, bool on_state); ZIG_EXTERN_C void ZigLLVMSetTailCall(LLVMValueRef Call); +ZIG_EXTERN_C void ZigLLVMSetCallSret(LLVMValueRef Call, LLVMTypeRef return_type); ZIG_EXTERN_C void ZigLLVMFunctionSetPrefixData(LLVMValueRef fn, LLVMValueRef data); ZIG_EXTERN_C void ZigLLVMFunctionSetCallingConv(LLVMValueRef function, enum ZigLLVM_CallingConv cc); ZIG_EXTERN_C void ZigLLVMAddFunctionAttr(LLVMValueRef fn, const char *attr_name, const char *attr_value); ZIG_EXTERN_C void ZigLLVMAddByValAttr(LLVMValueRef fn_ref, unsigned ArgNo, LLVMTypeRef type_val); +ZIG_EXTERN_C void ZigLLVMAddSretAttr(LLVMValueRef fn_ref, unsigned ArgNo, LLVMTypeRef type_val); ZIG_EXTERN_C void ZigLLVMAddFunctionAttrCold(LLVMValueRef fn); ZIG_EXTERN_C void ZigLLVMParseCommandLineOptions(size_t argc, const char *const *argv); @@ -288,6 +290,7 @@ enum ZigLLVM_ArchType { ZigLLVM_avr, // AVR: Atmel AVR microcontroller ZigLLVM_bpfel, // eBPF or extended BPF or 64-bit BPF (little endian) ZigLLVM_bpfeb, // eBPF or extended BPF or 64-bit BPF (big endian) + ZigLLVM_csky, // CSKY: csky ZigLLVM_hexagon, // Hexagon: hexagon ZigLLVM_mips, // MIPS: mips, mipsallegrex, mipsr6 ZigLLVM_mipsel, // MIPSEL: mipsel, mipsallegrexe, mipsr6el @@ -295,6 +298,7 @@ enum ZigLLVM_ArchType { ZigLLVM_mips64el, // MIPS64EL: mips64el, mips64r6el, mipsn32el, mipsn32r6el ZigLLVM_msp430, // MSP430: msp430 ZigLLVM_ppc, // PPC: powerpc + ZigLLVM_ppcle, // PPCLE: powerpc (little endian) ZigLLVM_ppc64, // PPC64: powerpc64, ppu ZigLLVM_ppc64le, // PPC64LE: powerpc64le ZigLLVM_r600, // R600: AMD GPUs HD2XXX - HD6XXX @@ -340,8 +344,6 @@ enum ZigLLVM_VendorType { ZigLLVM_Apple, ZigLLVM_PC, ZigLLVM_SCEI, - ZigLLVM_BGP, - ZigLLVM_BGQ, ZigLLVM_Freescale, ZigLLVM_IBM, ZigLLVM_ImaginationTechnologies, @@ -375,11 +377,11 @@ enum ZigLLVM_OSType { ZigLLVM_OpenBSD, ZigLLVM_Solaris, ZigLLVM_Win32, + ZigLLVM_ZOS, ZigLLVM_Haiku, ZigLLVM_Minix, ZigLLVM_RTEMS, ZigLLVM_NaCl, // Native Client - ZigLLVM_CNK, // BG/P Compute-Node Kernel ZigLLVM_AIX, ZigLLVM_CUDA, // NVIDIA CUDA ZigLLVM_NVCL, // NVIDIA OpenCL @@ -409,6 +411,7 @@ enum ZigLLVM_EnvironmentType { ZigLLVM_GNUEABI, ZigLLVM_GNUEABIHF, ZigLLVM_GNUX32, + ZigLLVM_GNUILP32, ZigLLVM_CODE16, ZigLLVM_EABI, ZigLLVM_EABIHF, @@ -432,6 +435,7 @@ enum ZigLLVM_ObjectFormatType { ZigLLVM_COFF, ZigLLVM_ELF, + ZigLLVM_GOFF, ZigLLVM_MachO, ZigLLVM_Wasm, ZigLLVM_XCOFF, diff --git a/test/stage1/behavior/vector.zig b/test/stage1/behavior/vector.zig index 5839376b5f..9dbdf02348 100644 --- a/test/stage1/behavior/vector.zig +++ b/test/stage1/behavior/vector.zig @@ -510,6 +510,28 @@ test "vector reduce operation" { const N = @typeInfo(@TypeOf(x)).Array.len; const TX = @typeInfo(@TypeOf(x)).Array.child; + // LLVM ERROR: Cannot select: intrinsic %llvm.aarch64.neon.fminnmv + // https://github.com/ziglang/zig/issues/8130 + // wasmtime: unknown import: `env::fminf` has not been defined + // https://github.com/ziglang/zig/issues/8131 + switch (std.builtin.arch) { + .aarch64 => switch (TX) { + f16 => switch (op) { + .Min, .Max, => return, + else => {}, + }, + else => {}, + }, + .wasm32 => switch (@typeInfo(TX)) { + .Float => switch (op) { + .Min, .Max, => return, + else => {}, + }, + else => {}, + }, + else => {}, + } + var r = @reduce(op, @as(Vector(N, TX), x)); switch (@typeInfo(TX)) { .Int, .Bool => expectEqual(expected, r), diff --git a/tools/update_cpu_features.zig b/tools/update_cpu_features.zig new file mode 100644 index 0000000000..7e3c636c31 --- /dev/null +++ b/tools/update_cpu_features.zig @@ -0,0 +1,1292 @@ +const std = @import("std"); +const fs = std.fs; +const mem = std.mem; +const json = std.json; +const assert = std.debug.assert; + +// All references to other features are based on "zig name" as the key. + +const FeatureOverride = struct { + llvm_name: []const u8, + /// If true, completely omit the feature; as if it does not exist. + omit: bool = false, + /// If true, omit the feature, but all the dependencies of the feature + /// are added in its place. + flatten: bool = false, + zig_name: ?[]const u8 = null, + desc: ?[]const u8 = null, + extra_deps: []const []const u8 = &.{}, +}; + +const Cpu = struct { + llvm_name: ?[]const u8, + zig_name: []const u8, + features: []const []const u8, +}; + +const Feature = struct { + llvm_name: ?[]const u8 = null, + zig_name: []const u8, + desc: []const u8, + deps: []const []const u8, + flatten: bool = false, +}; + +const LlvmTarget = struct { + zig_name: []const u8, + llvm_name: []const u8, + td_name: []const u8, + feature_overrides: []const FeatureOverride = &.{}, + extra_cpus: []const Cpu = &.{}, + extra_features: []const Feature = &.{}, + branch_quota: ?usize = null, +}; + +const llvm_targets = [_]LlvmTarget{ + .{ + .zig_name = "aarch64", + .llvm_name = "AArch64", + .td_name = "AArch64.td", + .branch_quota = 2000, + .feature_overrides = &.{ + .{ + .llvm_name = "CONTEXTIDREL2", + .zig_name = "contextidr_el2", + .desc = "Enable RW operand Context ID Register (EL2)", + }, + .{ + .llvm_name = "neoversee1", + .zig_name = "neoverse_e1", + }, + .{ + .llvm_name = "neoversen1", + .zig_name = "neoverse_n1", + }, + .{ + .llvm_name = "neoversen2", + .zig_name = "neoverse_n2", + }, + .{ + .llvm_name = "neoversev1", + .zig_name = "neoverse_v1", + }, + .{ + .llvm_name = "exynosm3", + .zig_name = "exynos_m3", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "exynosm4", + .zig_name = "exynos_m4", + }, + .{ + .llvm_name = "v8.1a", + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "a35", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "a53", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "a55", + .flatten = true, + }, + .{ + .llvm_name = "a57", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "a64fx", + .flatten = true, + }, + .{ + .llvm_name = "a72", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "a73", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "a75", + .flatten = true, + }, + .{ + .llvm_name = "a77", + .flatten = true, + }, + .{ + .llvm_name = "apple-a10", + .flatten = true, + }, + .{ + .llvm_name = "apple-a11", + .flatten = true, + }, + .{ + .llvm_name = "apple-a14", + .flatten = true, + }, + .{ + .llvm_name = "carmel", + .flatten = true, + }, + .{ + .llvm_name = "cortex-a78", + .flatten = true, + }, + .{ + .llvm_name = "cortex-x1", + .flatten = true, + }, + .{ + .llvm_name = "falkor", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "kryo", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "saphira", + .flatten = true, + }, + .{ + .llvm_name = "thunderx", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "thunderx2t99", + .flatten = true, + }, + .{ + .llvm_name = "thunderx3t110", + .flatten = true, + }, + .{ + .llvm_name = "thunderxt81", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "thunderxt83", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "thunderxt88", + .flatten = true, + .extra_deps = &.{"v8a"}, + }, + .{ + .llvm_name = "tsv110", + .flatten = true, + }, + }, + .extra_features = &.{ + .{ + .zig_name = "v8a", + .desc = "Support ARM v8a instructions", + .deps = &.{ "fp_armv8", "neon" }, + }, + }, + .extra_cpus = &.{ + .{ + .llvm_name = null, + .zig_name = "exynos_m1", + .features = &.{ + "crc", + "crypto", + "exynos_cheap_as_move", + "force_32bit_jump_tables", + "fuse_aes", + "perfmon", + "slow_misaligned_128store", + "slow_paired_128", + "use_postra_scheduler", + "use_reciprocal_square_root", + "v8a", + "zcz_fp", + }, + }, + .{ + .llvm_name = null, + .zig_name = "exynos_m2", + .features = &.{ + "crc", + "crypto", + "exynos_cheap_as_move", + "force_32bit_jump_tables", + "fuse_aes", + "perfmon", + "slow_misaligned_128store", + "slow_paired_128", + "use_postra_scheduler", + "v8a", + "zcz_fp", + }, + }, + }, + }, + .{ + .zig_name = "amdgpu", + .llvm_name = "AMDGPU", + .td_name = "AMDGPU.td", + .feature_overrides = &.{ + .{ + .llvm_name = "DumpCode", + .omit = true, + }, + .{ + .llvm_name = "dumpcode", + .omit = true, + }, + }, + }, + .{ + .zig_name = "arc", + .llvm_name = "ARC", + .td_name = "ARC.td", + }, + .{ + .zig_name = "arm", + .llvm_name = "ARM", + .td_name = "ARM.td", + .branch_quota = 10000, + .extra_cpus = &.{ + .{ + .llvm_name = "generic", + .zig_name = "baseline", + .features = &.{"v7a"}, + }, + .{ + .llvm_name = null, + .zig_name = "exynos_m1", + .features = &.{ "v8a", "exynos" }, + }, + .{ + .llvm_name = null, + .zig_name = "exynos_m2", + .features = &.{ "v8a", "exynos" }, + }, + }, + .feature_overrides = &.{ + .{ + .llvm_name = "cortex-a78", + .flatten = true, + }, + .{ + .llvm_name = "r5", + .flatten = true, + }, + .{ + .llvm_name = "r52", + .flatten = true, + }, + .{ + .llvm_name = "r7", + .flatten = true, + }, + .{ + .llvm_name = "m7", + .flatten = true, + }, + .{ + .llvm_name = "krait", + .flatten = true, + }, + .{ + .llvm_name = "kryo", + .flatten = true, + }, + .{ + .llvm_name = "cortex-x1", + .flatten = true, + }, + .{ + .llvm_name = "neoverse-v1", + .flatten = true, + }, + .{ + .llvm_name = "a5", + .flatten = true, + }, + .{ + .llvm_name = "a7", + .flatten = true, + }, + .{ + .llvm_name = "a8", + .flatten = true, + }, + .{ + .llvm_name = "a9", + .flatten = true, + }, + .{ + .llvm_name = "a12", + .flatten = true, + }, + .{ + .llvm_name = "a15", + .flatten = true, + }, + .{ + .llvm_name = "a17", + .flatten = true, + }, + .{ + .llvm_name = "a32", + .flatten = true, + }, + .{ + .llvm_name = "a35", + .flatten = true, + }, + .{ + .llvm_name = "a53", + .flatten = true, + }, + .{ + .llvm_name = "a55", + .flatten = true, + }, + .{ + .llvm_name = "a57", + .flatten = true, + }, + .{ + .llvm_name = "a72", + .flatten = true, + }, + .{ + .llvm_name = "a73", + .flatten = true, + }, + .{ + .llvm_name = "a75", + .flatten = true, + }, + .{ + .llvm_name = "a77", + .flatten = true, + }, + .{ + .llvm_name = "a78c", + .flatten = true, + }, + .{ + .llvm_name = "armv2", + .zig_name = "v2", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv2a", + .zig_name = "v2a", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv3", + .zig_name = "v3", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv3m", + .zig_name = "v3m", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv4", + .zig_name = "v4", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv4t", + .zig_name = "v4t", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv5t", + .zig_name = "v5t", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv5te", + .zig_name = "v5te", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv5tej", + .zig_name = "v5tej", + .extra_deps = &.{"strict_align"}, + }, + .{ + .llvm_name = "armv6", + .zig_name = "v6", + }, + .{ + .llvm_name = "armv6-m", + .zig_name = "v6m", + }, + .{ + .llvm_name = "armv6j", + .zig_name = "v6j", + }, + .{ + .llvm_name = "armv6k", + .zig_name = "v6k", + }, + .{ + .llvm_name = "armv6kz", + .zig_name = "v6kz", + }, + .{ + .llvm_name = "armv6s-m", + .zig_name = "v6sm", + }, + .{ + .llvm_name = "armv6t2", + .zig_name = "v6t2", + }, + .{ + .llvm_name = "armv7-a", + .zig_name = "v7a", + }, + .{ + .llvm_name = "armv7-m", + .zig_name = "v7m", + }, + .{ + .llvm_name = "armv7-r", + .zig_name = "v7r", + }, + .{ + .llvm_name = "armv7e-m", + .zig_name = "v7em", + }, + .{ + .llvm_name = "armv7k", + .zig_name = "v7k", + }, + .{ + .llvm_name = "armv7s", + .zig_name = "v7s", + }, + .{ + .llvm_name = "armv7ve", + .zig_name = "v7ve", + }, + .{ + .llvm_name = "armv8.1-a", + .zig_name = "v8_1a", + }, + .{ + .llvm_name = "armv8.1-m.main", + .zig_name = "v8_1m_main", + }, + .{ + .llvm_name = "armv8.2-a", + .zig_name = "v8_2a", + }, + .{ + .llvm_name = "armv8.3-a", + .zig_name = "v8_3a", + }, + .{ + .llvm_name = "armv8.4-a", + .zig_name = "v8_4a", + }, + .{ + .llvm_name = "armv8.5-a", + .zig_name = "v8_5a", + }, + .{ + .llvm_name = "armv8.6-a", + .zig_name = "v8_6a", + }, + .{ + .llvm_name = "armv8.7-a", + .zig_name = "v8_7a", + }, + .{ + .llvm_name = "armv8-a", + .zig_name = "v8a", + }, + .{ + .llvm_name = "armv8-m.base", + .zig_name = "v8m", + }, + .{ + .llvm_name = "armv8-m.main", + .zig_name = "v8m_main", + }, + .{ + .llvm_name = "armv8-r", + .zig_name = "v8r", + }, + .{ + .llvm_name = "v4t", + .zig_name = "has_v4t", + }, + .{ + .llvm_name = "v5t", + .zig_name = "has_v5t", + }, + .{ + .llvm_name = "v5te", + .zig_name = "has_v5te", + }, + .{ + .llvm_name = "v6", + .zig_name = "has_v6", + }, + .{ + .llvm_name = "v6k", + .zig_name = "has_v6k", + }, + .{ + .llvm_name = "v6m", + .zig_name = "has_v6m", + }, + .{ + .llvm_name = "v6t2", + .zig_name = "has_v6t2", + }, + .{ + .llvm_name = "v7", + .zig_name = "has_v7", + }, + .{ + .llvm_name = "v7clrex", + .zig_name = "has_v7clrex", + }, + .{ + .llvm_name = "v8", + .zig_name = "has_v8", + }, + .{ + .llvm_name = "v8m", + .zig_name = "has_v8m", + }, + .{ + .llvm_name = "v8m.main", + .zig_name = "has_v8m_main", + }, + .{ + .llvm_name = "v8.1a", + .zig_name = "has_v8_1a", + }, + .{ + .llvm_name = "v8.1m.main", + .zig_name = "has_v8_1m_main", + }, + .{ + .llvm_name = "v8.2a", + .zig_name = "has_v8_2a", + }, + .{ + .llvm_name = "v8.3a", + .zig_name = "has_v8_3a", + }, + .{ + .llvm_name = "v8.4a", + .zig_name = "has_v8_4a", + }, + .{ + .llvm_name = "v8.5a", + .zig_name = "has_v8_5a", + }, + .{ + .llvm_name = "v8.6a", + .zig_name = "has_v8_6a", + }, + .{ + .llvm_name = "v8.7a", + .zig_name = "has_v8_7a", + }, + }, + }, + .{ + .zig_name = "avr", + .llvm_name = "AVR", + .td_name = "AVR.td", + }, + .{ + .zig_name = "bpf", + .llvm_name = "BPF", + .td_name = "BPF.td", + }, + .{ + .zig_name = "csky", + .llvm_name = "CSKY", + .td_name = "CSKY.td", + }, + .{ + .zig_name = "hexagon", + .llvm_name = "Hexagon", + .td_name = "Hexagon.td", + }, + .{ + .zig_name = "lanai", + .llvm_name = "Lanai", + .td_name = "Lanai.td", + }, + .{ + .zig_name = "msp430", + .llvm_name = "MSP430", + .td_name = "MSP430.td", + }, + .{ + .zig_name = "mips", + .llvm_name = "Mips", + .td_name = "Mips.td", + }, + .{ + .zig_name = "nvptx", + .llvm_name = "NVPTX", + .td_name = "NVPTX.td", + }, + .{ + .zig_name = "powerpc", + .llvm_name = "PowerPC", + .td_name = "PPC.td", + }, + .{ + .zig_name = "riscv", + .llvm_name = "RISCV", + .td_name = "RISCV.td", + .extra_cpus = &.{ + .{ + .llvm_name = null, + .zig_name = "baseline_rv32", + .features = &.{ "a", "c", "d", "f", "m" }, + }, + .{ + .llvm_name = null, + .zig_name = "baseline_rv64", + .features = &.{ "64bit", "a", "c", "d", "f", "m" }, + }, + }, + }, + .{ + .zig_name = "sparc", + .llvm_name = "Sparc", + .td_name = "Sparc.td", + }, + .{ + .zig_name = "systemz", + .llvm_name = "SystemZ", + .td_name = "SystemZ.td", + }, + .{ + .zig_name = "ve", + .llvm_name = "VE", + .td_name = "VE.td", + }, + .{ + .zig_name = "wasm", + .llvm_name = "WebAssembly", + .td_name = "WebAssembly.td", + }, + .{ + .zig_name = "x86", + .llvm_name = "X86", + .td_name = "X86.td", + .feature_overrides = &.{ + .{ + .llvm_name = "64bit-mode", + .omit = true, + }, + .{ + .llvm_name = "i386", + .zig_name = "_i386", + }, + .{ + .llvm_name = "i486", + .zig_name = "_i486", + }, + .{ + .llvm_name = "i586", + .zig_name = "_i586", + }, + .{ + .llvm_name = "i686", + .zig_name = "_i686", + }, + }, + }, + .{ + .zig_name = "xcore", + .llvm_name = "XCore", + .td_name = "XCore.td", + }, +}; + +pub fn main() anyerror!void { + var arena_state = std.heap.ArenaAllocator.init(std.heap.page_allocator); + defer arena_state.deinit(); + const arena = &arena_state.allocator; + + const args = try std.process.argsAlloc(arena); + if (args.len <= 1) { + usageAndExit(std.io.getStdErr(), args[0], 1); + } + if (std.mem.eql(u8, args[1], "--help")) { + usageAndExit(std.io.getStdOut(), args[0], 0); + } + if (args.len < 4) { + usageAndExit(std.io.getStdErr(), args[0], 1); + } + + const llvm_tblgen_exe = args[1]; + if (std.mem.startsWith(u8, llvm_tblgen_exe, "-")) { + usageAndExit(std.io.getStdErr(), args[0], 1); + } + + const llvm_src_root = args[2]; + if (std.mem.startsWith(u8, llvm_src_root, "-")) { + usageAndExit(std.io.getStdErr(), args[0], 1); + } + + const zig_src_root = args[3]; + if (std.mem.startsWith(u8, zig_src_root, "-")) { + usageAndExit(std.io.getStdErr(), args[0], 1); + } + + var zig_src_dir = try fs.cwd().openDir(zig_src_root, .{}); + defer zig_src_dir.close(); + + var progress = std.Progress{}; + const root_progress = try progress.start("", llvm_targets.len); + defer root_progress.end(); + + if (std.builtin.single_threaded) { + for (llvm_targets) |llvm_target| { + try processOneTarget(Job{ + .llvm_tblgen_exe = llvm_tblgen_exe, + .llvm_src_root = llvm_src_root, + .zig_src_dir = zig_src_dir, + .root_progress = root_progress, + .llvm_target = llvm_target, + }); + } + } else { + var threads = try arena.alloc(*std.Thread, llvm_targets.len); + for (llvm_targets) |llvm_target, i| { + threads[i] = try std.Thread.spawn(processOneTarget, .{ + .llvm_tblgen_exe = llvm_tblgen_exe, + .llvm_src_root = llvm_src_root, + .zig_src_dir = zig_src_dir, + .root_progress = root_progress, + .llvm_target = llvm_target, + }); + } + for (threads) |thread| { + thread.wait(); + } + } +} + +const Job = struct { + llvm_tblgen_exe: []const u8, + llvm_src_root: []const u8, + zig_src_dir: std.fs.Dir, + root_progress: *std.Progress.Node, + llvm_target: LlvmTarget, +}; + +fn processOneTarget(job: Job) anyerror!void { + const llvm_target = job.llvm_target; + + var arena_state = std.heap.ArenaAllocator.init(std.heap.page_allocator); + defer arena_state.deinit(); + const arena = &arena_state.allocator; + + var progress_node = job.root_progress.start(llvm_target.zig_name, 3); + progress_node.activate(); + defer progress_node.end(); + + var tblgen_progress = progress_node.start("invoke llvm-tblgen", 0); + tblgen_progress.activate(); + + const child_args = [_][]const u8{ + job.llvm_tblgen_exe, + "--dump-json", + try std.fmt.allocPrint(arena, "{s}/llvm/lib/Target/{s}/{s}", .{ + job.llvm_src_root, + llvm_target.llvm_name, + llvm_target.td_name, + }), + try std.fmt.allocPrint(arena, "-I={s}/llvm/include", .{job.llvm_src_root}), + try std.fmt.allocPrint(arena, "-I={s}/llvm/lib/Target/{s}", .{ + job.llvm_src_root, llvm_target.llvm_name, + }), + }; + + const child_result = try std.ChildProcess.exec(.{ + .allocator = arena, + .argv = &child_args, + .max_output_bytes = 200 * 1024 * 1024, + }); + tblgen_progress.end(); + if (child_result.stderr.len != 0) { + std.debug.warn("{s}\n", .{child_result.stderr}); + } + + const json_text = switch (child_result.term) { + .Exited => |code| if (code == 0) child_result.stdout else { + std.debug.warn("llvm-tblgen exited with code {d}\n", .{code}); + std.process.exit(1); + }, + else => { + std.debug.warn("llvm-tblgen crashed\n", .{}); + std.process.exit(1); + }, + }; + + var json_parse_progress = progress_node.start("parse JSON", 0); + json_parse_progress.activate(); + + var parser = json.Parser.init(arena, false); + const tree = try parser.parse(json_text); + json_parse_progress.end(); + + var render_progress = progress_node.start("render zig code", 0); + render_progress.activate(); + + const root_map = &tree.root.Object; + var features_table = std.StringHashMap(Feature).init(arena); + var all_features = std.ArrayList(Feature).init(arena); + var all_cpus = std.ArrayList(Cpu).init(arena); + { + var it = root_map.iterator(); + root_it: while (it.next()) |kv| { + if (kv.key.len == 0) continue; + if (kv.key[0] == '!') continue; + if (kv.value != .Object) continue; + if (hasSuperclass(&kv.value.Object, "SubtargetFeature")) { + const llvm_name = kv.value.Object.get("Name").?.String; + if (llvm_name.len == 0) continue; + + var zig_name = try llvmNameToZigName(arena, llvm_name); + var desc = kv.value.Object.get("Desc").?.String; + var deps = std.ArrayList([]const u8).init(arena); + var omit = false; + var flatten = false; + const implies = kv.value.Object.get("Implies").?.Array; + for (implies.items) |imply| { + const other_key = imply.Object.get("def").?.String; + const other_obj = &root_map.getEntry(other_key).?.value.Object; + const other_llvm_name = other_obj.get("Name").?.String; + const other_zig_name = (try llvmNameToZigNameOmit( + arena, + llvm_target, + other_llvm_name, + )) orelse continue; + try deps.append(other_zig_name); + } + for (llvm_target.feature_overrides) |feature_override| { + if (mem.eql(u8, llvm_name, feature_override.llvm_name)) { + if (feature_override.omit) { + // Still put the feature into the table so that we can + // expand dependencies for the feature overrides marked `flatten`. + omit = true; + } + if (feature_override.flatten) { + flatten = true; + } + if (feature_override.zig_name) |override_name| { + zig_name = override_name; + } + if (feature_override.desc) |override_desc| { + desc = override_desc; + } + for (feature_override.extra_deps) |extra_dep| { + try deps.append(extra_dep); + } + break; + } + } + const feature: Feature = .{ + .llvm_name = llvm_name, + .zig_name = zig_name, + .desc = desc, + .deps = deps.items, + .flatten = flatten, + }; + try features_table.put(zig_name, feature); + if (!omit and !flatten) { + try all_features.append(feature); + } + } + if (hasSuperclass(&kv.value.Object, "Processor")) { + const llvm_name = kv.value.Object.get("Name").?.String; + if (llvm_name.len == 0) continue; + + var zig_name = try llvmNameToZigName(arena, llvm_name); + var deps = std.ArrayList([]const u8).init(arena); + const features = kv.value.Object.get("Features").?.Array; + for (features.items) |feature| { + const feature_key = feature.Object.get("def").?.String; + const feature_obj = &root_map.getEntry(feature_key).?.value.Object; + const feature_llvm_name = feature_obj.get("Name").?.String; + if (feature_llvm_name.len == 0) continue; + const feature_zig_name = (try llvmNameToZigNameOmit( + arena, + llvm_target, + feature_llvm_name, + )) orelse continue; + try deps.append(feature_zig_name); + } + const tune_features = kv.value.Object.get("TuneFeatures").?.Array; + for (tune_features.items) |feature| { + const feature_key = feature.Object.get("def").?.String; + const feature_obj = &root_map.getEntry(feature_key).?.value.Object; + const feature_llvm_name = feature_obj.get("Name").?.String; + if (feature_llvm_name.len == 0) continue; + const feature_zig_name = (try llvmNameToZigNameOmit( + arena, + llvm_target, + feature_llvm_name, + )) orelse continue; + try deps.append(feature_zig_name); + } + for (llvm_target.feature_overrides) |feature_override| { + if (mem.eql(u8, llvm_name, feature_override.llvm_name)) { + if (feature_override.omit) { + continue :root_it; + } + if (feature_override.zig_name) |override_name| { + zig_name = override_name; + } + for (feature_override.extra_deps) |extra_dep| { + try deps.append(extra_dep); + } + break; + } + } + try all_cpus.append(.{ + .llvm_name = llvm_name, + .zig_name = zig_name, + .features = deps.items, + }); + } + } + } + for (llvm_target.extra_features) |extra_feature| { + try features_table.put(extra_feature.zig_name, extra_feature); + try all_features.append(extra_feature); + } + for (llvm_target.extra_cpus) |extra_cpu| { + try all_cpus.append(extra_cpu); + } + std.sort.sort(Feature, all_features.items, {}, featureLessThan); + std.sort.sort(Cpu, all_cpus.items, {}, cpuLessThan); + + const target_sub_path = try fs.path.join(arena, &.{ "lib", "std", "target" }); + var target_dir = try job.zig_src_dir.makeOpenPath(target_sub_path, .{}); + defer target_dir.close(); + + const zig_code_basename = try std.fmt.allocPrint(arena, "{s}.zig", .{llvm_target.zig_name}); + + if (all_features.items.len == 0) { + // We represent this with an empty file. + try target_dir.deleteTree(zig_code_basename); + return; + } + + var zig_code_file = try target_dir.createFile(zig_code_basename, .{}); + defer zig_code_file.close(); + + var bw = std.io.bufferedWriter(zig_code_file.writer()); + const w = bw.writer(); + + try w.writeAll( + \\//! This file is auto-generated by tools/update_cpu_features.zig. + \\ + \\const std = @import("../std.zig"); + \\const CpuFeature = std.Target.Cpu.Feature; + \\const CpuModel = std.Target.Cpu.Model; + \\ + \\pub const Feature = enum { + \\ + ); + + for (all_features.items) |feature| { + try w.print(" {},\n", .{std.zig.fmtId(feature.zig_name)}); + } + + try w.writeAll( + \\}; + \\ + \\pub usingnamespace CpuFeature.feature_set_fns(Feature); + \\ + \\pub const all_features = blk: { + \\ + ); + if (llvm_target.branch_quota) |branch_quota| { + try w.print(" @setEvalBranchQuota({d});\n", .{branch_quota}); + } + try w.writeAll( + \\ const len = @typeInfo(Feature).Enum.fields.len; + \\ std.debug.assert(len <= CpuFeature.Set.needed_bit_count); + \\ var result: [len]CpuFeature = undefined; + \\ + ); + + for (all_features.items) |feature| { + if (feature.llvm_name) |llvm_name| { + try w.print( + \\ result[@enumToInt(Feature.{})] = .{{ + \\ .llvm_name = "{}", + \\ .description = "{}", + \\ .dependencies = featureSet(&[_]Feature{{ + , + .{ + std.zig.fmtId(feature.zig_name), + std.zig.fmtEscapes(llvm_name), + std.zig.fmtEscapes(feature.desc), + }, + ); + } else { + try w.print( + \\ result[@enumToInt(Feature.{})] = .{{ + \\ .llvm_name = null, + \\ .description = "{}", + \\ .dependencies = featureSet(&[_]Feature{{ + , + .{ + std.zig.fmtId(feature.zig_name), + std.zig.fmtEscapes(feature.desc), + }, + ); + } + var deps_set = std.StringHashMap(void).init(arena); + for (feature.deps) |dep| { + try putDep(&deps_set, features_table, dep); + } + try pruneFeatures(arena, features_table, &deps_set); + var dependencies = std.ArrayList([]const u8).init(arena); + { + var it = deps_set.iterator(); + while (it.next()) |entry| { + try dependencies.append(entry.key); + } + } + std.sort.sort([]const u8, dependencies.items, {}, asciiLessThan); + + if (dependencies.items.len == 0) { + try w.writeAll( + \\}), + \\ }; + \\ + ); + } else { + try w.writeAll("\n"); + for (dependencies.items) |dep| { + try w.print(" .{},\n", .{std.zig.fmtId(dep)}); + } + try w.writeAll( + \\ }), + \\ }; + \\ + ); + } + } + try w.writeAll( + \\ const ti = @typeInfo(Feature); + \\ for (result) |*elem, i| { + \\ elem.index = i; + \\ elem.name = ti.Enum.fields[i].name; + \\ } + \\ break :blk result; + \\}; + \\ + \\pub const cpu = struct { + \\ + ); + for (all_cpus.items) |cpu| { + var deps_set = std.StringHashMap(void).init(arena); + for (cpu.features) |feature_zig_name| { + try putDep(&deps_set, features_table, feature_zig_name); + } + try pruneFeatures(arena, features_table, &deps_set); + var cpu_features = std.ArrayList([]const u8).init(arena); + { + var it = deps_set.iterator(); + while (it.next()) |entry| { + try cpu_features.append(entry.key); + } + } + std.sort.sort([]const u8, cpu_features.items, {}, asciiLessThan); + if (cpu.llvm_name) |llvm_name| { + try w.print( + \\ pub const {} = CpuModel{{ + \\ .name = "{}", + \\ .llvm_name = "{}", + \\ .features = featureSet(&[_]Feature{{ + , .{ + std.zig.fmtId(cpu.zig_name), + std.zig.fmtEscapes(cpu.zig_name), + std.zig.fmtEscapes(llvm_name), + }); + } else { + try w.print( + \\ pub const {} = CpuModel{{ + \\ .name = "{}", + \\ .llvm_name = null, + \\ .features = featureSet(&[_]Feature{{ + , .{ + std.zig.fmtId(cpu.zig_name), + std.zig.fmtEscapes(cpu.zig_name), + }); + } + if (cpu_features.items.len == 0) { + try w.writeAll( + \\}), + \\ }; + \\ + ); + } else { + try w.writeAll("\n"); + for (cpu_features.items) |feature_zig_name| { + try w.print(" .{},\n", .{std.zig.fmtId(feature_zig_name)}); + } + try w.writeAll( + \\ }), + \\ }; + \\ + ); + } + } + + try w.writeAll( + \\}; + \\ + ); + try bw.flush(); + + render_progress.end(); +} + +fn usageAndExit(file: fs.File, arg0: []const u8, code: u8) noreturn { + file.writer().print( + \\Usage: {s} /path/to/llvm-tblgen /path/git/llvm-project /path/git/zig + \\ + \\Updates lib/std/target/<target>.zig from llvm/lib/Target/<Target>/<Target>.td . + \\ + \\On a less beefy system, or when debugging, compile with --single-threaded. + \\ + , .{arg0}) catch std.process.exit(1); + std.process.exit(code); +} + +fn featureLessThan(context: void, a: Feature, b: Feature) bool { + return std.ascii.lessThanIgnoreCase(a.zig_name, b.zig_name); +} + +fn cpuLessThan(context: void, a: Cpu, b: Cpu) bool { + return std.ascii.lessThanIgnoreCase(a.zig_name, b.zig_name); +} + +fn asciiLessThan(context: void, a: []const u8, b: []const u8) bool { + return std.ascii.lessThanIgnoreCase(a, b); +} + +fn llvmNameToZigName(arena: *mem.Allocator, llvm_name: []const u8) ![]const u8 { + const duped = try arena.dupe(u8, llvm_name); + for (duped) |*byte| switch (byte.*) { + '-', '.' => byte.* = '_', + else => continue, + }; + return duped; +} + +fn llvmNameToZigNameOmit( + arena: *mem.Allocator, + llvm_target: LlvmTarget, + llvm_name: []const u8, +) !?[]const u8 { + for (llvm_target.feature_overrides) |feature_override| { + if (mem.eql(u8, feature_override.llvm_name, llvm_name)) { + if (feature_override.omit) return null; + return feature_override.zig_name orelse break; + } + } + return try llvmNameToZigName(arena, llvm_name); +} + +fn hasSuperclass(obj: *json.ObjectMap, class_name: []const u8) bool { + const superclasses_json = obj.get("!superclasses") orelse return false; + for (superclasses_json.Array.items) |superclass_json| { + const superclass = superclass_json.String; + if (std.mem.eql(u8, superclass, class_name)) { + return true; + } + } + return false; +} + +fn pruneFeatures( + arena: *mem.Allocator, + features_table: std.StringHashMap(Feature), + deps_set: *std.StringHashMap(void), +) !void { + // For each element, recursively iterate over the dependencies and add + // everything we find to a "deletion set". + // Then, iterate over the deletion set and delete all that stuff from `deps_set`. + var deletion_set = std.StringHashMap(void).init(arena); + { + var it = deps_set.iterator(); + while (it.next()) |entry| { + const feature = features_table.get(entry.key).?; + try walkFeatures(features_table, &deletion_set, feature); + } + } + { + var it = deletion_set.iterator(); + while (it.next()) |entry| { + _ = deps_set.remove(entry.key); + } + } +} + +fn walkFeatures( + features_table: std.StringHashMap(Feature), + deletion_set: *std.StringHashMap(void), + feature: Feature, +) error{OutOfMemory}!void { + for (feature.deps) |dep| { + try deletion_set.put(dep, {}); + const other_feature = features_table.get(dep).?; + try walkFeatures(features_table, deletion_set, other_feature); + } +} + +fn putDep( + deps_set: *std.StringHashMap(void), + features_table: std.StringHashMap(Feature), + zig_feature_name: []const u8, +) error{OutOfMemory}!void { + const feature = features_table.get(zig_feature_name).?; + if (feature.flatten) { + for (feature.deps) |dep| { + try putDep(deps_set, features_table, dep); + } + } else { + try deps_set.put(zig_feature_name, {}); + } +} |