| Age | Commit message (Collapse) | Author |
|---|
|
coff: implement enough of the incremental linker to pass behavior and incremental tests on Windows
|
|
This adds the following for passthrough to lld:
- `--print-gc-sections`
- `--print-icf-sections`
- `--print-map`
I am not adding these to the cache manifest, since it does not change
the produced artifacts.
Tested with an example from #11398: it successfully prints the resulting
map and the GC'd sections.
|
|
|
|
This is problematic because in practice it depends on whether the
compiler backend supports it too, as evidenced by the TODO comment about
LLVM not supporting some architectures that in fact do support tail
calls.
Instead this logic is organized strategically in src/target.zig, part of
the internal compiler source code, and the behavior tests in question
duplicate some logic for deciding whether to proceed with the test.
The proper place to expose this flag is in `@import("builtin")` - the
generated source file - so that third party compilers can advertise
whether they support tail calls.
|
|
|
|
From `copy_file_range(2)` errors:
ETXTBSY
Either fd_in or fd_out refers to an active swap file.
Same error will be used in the upcoming `ioctl_ficlonerange(2)`:
ETXTBSY
One of the files is a swap file. Swap files cannot share storage.
|
|
This commit enables `-u <symbol>` for ELF and `-include:<symbol>` for
COFF linkers for use internally. This means we do not expose these
flags to the users just yet, however, we make use of them internally
whenever required. One such use case is forcing inclusion of
`_tls_index` when linking for Windows with mingw and LTO and dead
code stripping enabled. This ensures we add `_tls_index` to the symbol
resolver as an undefined symbol and force the linker to include an atom
that provides it marking it a dead-code-stripping root - meaning it will
not be garbage collected by the linker no matter what.
|
|
stage1 is available behind the -fstage1 flag.
closes #89
|
|
This is one of the final remaining TODOs for the LLVM backend.
|
|
|
|
Firstly, opening a file handle is not really needed since we won't even
use it, and secondly, this can cause AccessDenied errors on Windows
when trying to move a directory from zig-cache/tmp/ to zig-cache/o/
since, without POSIX semantics, it is illegal to move directories
with open handles to any of its resources.
|
|
I encountered this error today when testing the self-hosted compiler on
Windows.
|
|
Now, each object file will store a mutable table of symbols that it
defines. Upon symbol resolution between object files, the symbol
will be updated with a globally allocated section ordinal and address
in virtual memory. If the object defines a globally available symbol,
its location only (comprising of the symbol index and object index)
will be stored in the globals map for easy access when relocating, etc.
This approach cleans up the symbol management significantly, and matches
the status quo used in zld/ELF.
Additionally, this makes scoping symbol stabs easier too as they are
now naturally contained within each object file.
|
|
* test/link: initial wasm support
This adds basic parsing and dumping of wasm section so they
can be tested using the new linker-test infrastructure.
* test/link: all wasm sections parsing and dumping
We now parse and dump all sections for the wasm binary format.
Currently, this only dumps the name of a custom section.
Later this should also dump symbol table, name, linking metadata and relocations.
All of those live within the custom sections.
* Add wasm linker test
This also fixes a parser mistake in reading the flags.
* test/link: implement linker tests wasm & fixes
Adds several test cases to test the wasm self-hosted linker.
This also introduces fixes that were caught during the implementation
of those tests.
* test-runner: obey omit_stage2 for standalone
When a standalone test requires stage2, but stage2 is omit
from the compiler, such test case will not be included as part
of the test suite that is being ran. This is to support CI's
where we omit stage2 to lower the memory usage.
|
|
* make the setting in the linker backend be non-optional; by this time
all defaults are supposed to be resolved.
* integrate with `zig cc`
* change the CLI parsing to match C compiler parsing, allowing
`--compress-debug-sections` alone to choose a default encoding of
zlib.
|
|
This argument is both a compiler and a linker flag. The linker flag was
not understood; now it is. Go likes to use it as a linker flag.
Tested with sqlite3. The size difference is significant, and I confirmed
gdb understands both binaries.
zlib: 3.66MB
------------
CC="zig cc"
CFLAGS="-Wl,--compress-debug-sections=zlib -O2"
./configure --disable-tcl
make
FILE SIZE VM SIZE
-------------- --------------
39.1% 1.43Mi 88.4% 1.43Mi .text
19.6% 734Ki 0.0% 0 .debug_info
16.4% 613Ki 0.0% 0 .debug_loc
13.1% 492Ki 0.0% 0 .debug_line
4.2% 157Ki 9.5% 157Ki .rodata
2.3% 87.6Ki 0.0% 0 .debug_ranges
1.5% 56.2Ki 0.0% 0 .symtab
1.1% 40.2Ki 0.0% 0 .strtab
1.0% 38.2Ki 0.0% 0 .debug_str
0.7% 26.2Ki 0.0% 0 .debug_frame
0.4% 15.3Ki 0.9% 15.3Ki .data
0.1% 4.71Ki 0.3% 4.71Ki .dynsym
0.1% 3.65Ki 0.2% 3.26Ki [16 Others]
0.1% 2.55Ki 0.2% 2.55Ki .rela.plt
0.1% 2.12Ki 0.0% 0 [ELF Section Headers]
0.0% 0 0.1% 2.02Ki .bss
0.0% 1.84Ki 0.1% 1.84Ki .dynstr
0.0% 1.72Ki 0.1% 1.72Ki .plt
0.0% 1.58Ki 0.1% 1.58Ki .hash
0.0% 1.17Ki 0.0% 0 .debug_abbrev
0.0% 1.01Ki 0.1% 1.01Ki .rela.dyn
100.0% 3.66Mi 100.0% 1.62Mi TOTAL
none: 8.56MB
------------
CC="zig cc" CFLAGS="-O2" ./configure --disable-tcl
make
FILE SIZE VM SIZE
-------------- --------------
41.1% 3.52Mi 0.0% 0 .debug_loc
18.5% 1.59Mi 0.0% 0 .debug_info
16.7% 1.43Mi 88.4% 1.43Mi .text
11.8% 1.01Mi 0.0% 0 .debug_line
5.9% 515Ki 0.0% 0 .debug_ranges
1.8% 157Ki 9.5% 157Ki .rodata
1.3% 118Ki 0.0% 0 .debug_frame
1.3% 110Ki 0.0% 0 .debug_str
0.6% 56.2Ki 0.0% 0 .symtab
0.5% 40.2Ki 0.0% 0 .strtab
0.2% 15.3Ki 0.9% 15.3Ki .data
0.1% 4.71Ki 0.3% 4.71Ki .dynsym
0.0% 3.64Ki 0.2% 3.26Ki [16 Others]
0.0% 2.98Ki 0.0% 0 .debug_abbrev
0.0% 2.55Ki 0.2% 2.55Ki .rela.plt
0.0% 2.12Ki 0.0% 0 [ELF Section Headers]
0.0% 0 0.1% 2.02Ki .bss
0.0% 1.84Ki 0.1% 1.84Ki .dynstr
0.0% 1.72Ki 0.1% 1.72Ki .plt
0.0% 1.58Ki 0.1% 1.58Ki .hash
0.0% 1.01Ki 0.1% 1.01Ki .rela.dyn
100.0% 8.56Mi 100.0% 1.62Mi TOTAL
|
|
|
|
Future improvement: make plain error notes actually render as notes
rather than errors, but keep them as errors for the case of
sub-compilation errors, e.g. when compiler-rt has compilation errors.
|
|
|
|
Parse `-weak-lx` and `-weak_framework x` in the CLI.
|
|
MachO linker now handles `-needed-l<name>`, `-needed_library=<name>`
and `-needed_framework=<name>`. While on macOS `-l` is equivalent
to `-needed-l`, and `-framework` to `-needed_framework`, it can be
used to the same effect as on Linux if combined with `-dead_strip_dylibs`.
This commit also adds handling for `-needed_library` which is macOS
specific flag only (in addition to `-needed-l`).
Finally, in order to leverage new linker testing harness, this commit
added ability to specify lowering to those flags via `build.zig`:
`linkSystemLibraryNeeded` (and related), and `linkFrameworkNeeded`.
|
|
|
|
|
|
Includes both traditiona and incremental codepaths with one caveat that
in incremental case, the requested size cannot be smaller than the
default padding size due to prealloc required due to incremental nature
of linking.
Also parse `-headerpad_max_install_names`, however, not actionable just yet -
missing implementation.
|
|
|
|
This adds clarification to the getGlobalSymbol doc comments,
as well as renames the `addExternFn` function for MachO to `getGlobalSymbol`.
This function will now be called from 'src/link.zig' as well.
Finally, this also enables compiling zig's libc using LLVM even though
the `fno-LLVM` flag is given.
|
|
|
|
Pass `-pagezero_size` to the MachO linker. This is the final
"unsupported linker arg" that I could chase that CGo uses. After this
and #11874 we may be able to fail on an "unsupported linker arg" instead
of emiting a warning.
Test case:
zig=/code/zig/build/zig
CGO_ENABLED=1 GOOS=darwin GOARCH=amd64 CC="$zig cc -target x86_64-macos" CXX="$zig c++ -target x86_64-macos" go build -a -ldflags "-s -w" cgo.go
I compiled a trivial CGo program and executed it on an amd64 Darwin
host.
To be honest, I am not entirely sure what this is doing. This feels
right after reading what this argument does in LLVM sources, but I am by
no means qualified to make MachO pull requests. Will take feedback.
|
|
|
|
|
|
Note that the current documentation for the `-z noexecstack` is
incorrect. This indicates that an object *does not* require an
executable stack.
This is actually the default of LLD, and there has never been a way to
override this default by passing `-z execstack` to LLD.
This commit removes the redundant `-z noexecstack` option from
zig build-exe/build-lib/build-obj and ignores the option if passed
to zig cc for compatibility.
As far as I can tell, there is no reason for code to require an
executable stack. This option only exists because the stack was
originally executable by default and some programs came to depend
on that behavior. Instead, mprotect(2) may be used to make memory
pages executable.
|
|
closes #3047
|
|
Warnings about non-implemented `-z nocopyreloc` are common when
compiling go code (including Go's tests themselves). Let's just
make it stop complaining.
|
|
|
|
Rather than allocating Decl objects with an Allocator, we instead allocate
them with a SegmentedList. This provides four advantages:
* Stable memory so that one thread can access a Decl object while another
thread allocates additional Decl objects from this list.
* It allows us to use u32 indexes to reference Decl objects rather than
pointers, saving memory in Type, Value, and dependency sets.
* Using integers to reference Decl objects rather than pointers makes
serialization trivial.
* It provides a unique integer to be used for anonymous symbol names,
avoiding multi-threaded contention on an atomic counter.
|
|
This gives us insight as to what is happening when we are waiting for
things such as LLVM emit object and LLD linking.
|
|
This way we avoid the infamous SIGKILL on arm64 macos.
|
|
* The `@bitCast` workaround is removed in favor of `@ptrCast` properly
doing element casting for slice element types. This required an
enhancement both to stage1 and stage2.
* stage1 incorrectly accepts `.{}` instead of `{}`. stage2 code that
abused this is fixed.
* Make some parameters comptime to support functions in switch
expressions (as opposed to making them function pointers).
* Avoid relying on local temporaries being mutable.
* Workarounds for when stage1 and stage2 disagree on function pointer
types.
* Workaround recursive formatting bug with a `@panic("TODO")`.
* Remove unreachable `else` prongs for some inferred error sets.
All in effort towards #89.
|
|
Previously it would fail as `renameW` do not ever fail with
`PathAlreadyExists`.
As a workaround we check for dest dir existence before rename
on Windows.
|
|
|
|
Add a `target` parameter to every function that deals with Type and
Value.
|
|
With this change, we can now bake in entitlements into the binary.
Additionally, I see this as the first step towards full code signature
support which includes baking in Apple issued certificates for
redistribution, etc.
|
|
Also, fix premature exit in `link.File.makeWritable` in case we
are running M1 but executing binaries using Rosetta2.
|
|
This commit also adds a bunch of bindings for debug info.
|
|
Hook up Elf and MachO linkers to the new solution.
|
|
To unify the wasm backend with the other backends, we will now call `generateSymbol` to
lower a Decl into bytes. This means we also have to change some function signatures
to comply with the linker interface.
Since the general purpose generateSymbol is less featureful than wasm's, some tests are
temporarily disabled.
|
|
This requires generating an addend for the target relocation as
the field pointer might point at a field inner to the container.
|
|
We now correctly implement exporting decls. This means it is possible to export
a decl with a different name than the decl that is doing the export.
This also sets the symbols with the correct flags, so when we emit a relocatable
object file, a linker can correctly resolve symbols and/or export the symbol to the host environment.
This commit also includes fixes to ensure relocations have the correct offset to how other
linkers will expect the offset, rather than what we use internally.
Other linkers accept the offset, relative to the section.
Internally we use an offset relative to the atom.
|
|
|
|
In `getDeclVAddr`, it may happen that the target `Decl` has not
been allocated space in virtual memory. In this case, we store a
relocation in the linker-global table which we will iterate over
when flushing the module, and fill in any missing address in the
final binary. Note that for optimisation, if the address was resolved
at the time of a call to `getDeclVAddr`, we skip relocating this
atom.
This commit also adds the glue code for lowering const slices in
the ARM backend.
|
