aboutsummaryrefslogtreecommitdiff
path: root/SOURCES/cachy-bore.patch
blob: 9f0fcc46fd74c9f5654e7b32baad5ede7e8fada4 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
From 61efc6f62710e42ca623e209f901394d4f356cc9 Mon Sep 17 00:00:00 2001
From: Peter Jung <admin@ptr1337.dev>
Date: Thu, 30 May 2024 10:45:04 +0200
Subject: [PATCH] bore

Signed-off-by: Peter Jung <admin@ptr1337.dev>
---
 include/linux/sched.h   |  10 ++
 init/Kconfig            |  17 +++
 kernel/sched/core.c     | 143 +++++++++++++++++++++
 kernel/sched/debug.c    |  60 ++++++++-
 kernel/sched/fair.c     | 272 +++++++++++++++++++++++++++++++++++++---
 kernel/sched/features.h |  20 ++-
 kernel/sched/sched.h    |   7 ++
 7 files changed, 512 insertions(+), 17 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 3c2abbc587b49..e7bf3a034aa20 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -547,6 +547,16 @@ struct sched_entity {
 	u64				sum_exec_runtime;
 	u64				prev_sum_exec_runtime;
 	u64				vruntime;
+#ifdef CONFIG_SCHED_BORE
+	u64				burst_time;
+	u8				prev_burst_penalty;
+	u8				curr_burst_penalty;
+	u8				burst_penalty;
+	u8				burst_score;
+	u8				child_burst;
+	u32				child_burst_cnt;
+	u64				child_burst_last_cached;
+#endif // CONFIG_SCHED_BORE
 	s64				vlag;
 	u64				slice;
 
diff --git a/init/Kconfig b/init/Kconfig
index 664bedb9a71fb..6f9c7fc90707a 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -1279,6 +1279,23 @@ config CHECKPOINT_RESTORE
 
 	  If unsure, say N here.
 
+config SCHED_BORE
+	bool "Burst-Oriented Response Enhancer"
+	default y
+	help
+	  In Desktop and Mobile computing, one might prefer interactive
+	  tasks to keep responsive no matter what they run in the background.
+
+	  Enabling this kernel feature modifies the scheduler to discriminate
+	  tasks by their burst time (runtime since it last went sleeping or
+	  yielding state) and prioritize those that run less bursty.
+	  Such tasks usually include window compositor, widgets backend,
+	  terminal emulator, video playback, games and so on.
+	  With a little impact to scheduling fairness, it may improve
+	  responsiveness especially under heavy background workload.
+
+	  If unsure, say Y here.
+
 config SCHED_AUTOGROUP
 	bool "Automatic process group scheduling"
 	select CGROUPS
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index d211d40a2edc9..362df741dc85e 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4516,6 +4516,138 @@ int wake_up_state(struct task_struct *p, unsigned int state)
 	return try_to_wake_up(p, state, 0);
 }
 
+#ifdef CONFIG_SCHED_BORE
+extern u8   sched_burst_fork_atavistic;
+extern uint sched_burst_cache_lifetime;
+
+static void __init sched_init_bore(void) {
+	init_task.se.burst_time = 0;
+	init_task.se.prev_burst_penalty = 0;
+	init_task.se.curr_burst_penalty = 0;
+	init_task.se.burst_penalty = 0;
+	init_task.se.burst_score = 0;
+	init_task.se.child_burst_last_cached = 0;
+}
+
+void inline sched_fork_bore(struct task_struct *p) {
+	p->se.burst_time = 0;
+	p->se.curr_burst_penalty = 0;
+	p->se.burst_score = 0;
+	p->se.child_burst_last_cached = 0;
+}
+
+static u32 count_child_tasks(struct task_struct *p) {
+	struct task_struct *child;
+	u32 cnt = 0;
+	list_for_each_entry(child, &p->children, sibling) {cnt++;}
+	return cnt;
+}
+
+static inline bool task_is_inheritable(struct task_struct *p) {
+	return (p->sched_class == &fair_sched_class);
+}
+
+static inline bool child_burst_cache_expired(struct task_struct *p, u64 now) {
+	u64 expiration_time =
+		p->se.child_burst_last_cached + sched_burst_cache_lifetime;
+	return ((s64)(expiration_time - now) < 0);
+}
+
+static void __update_child_burst_cache(
+	struct task_struct *p, u32 cnt, u32 sum, u64 now) {
+	u8 avg = 0;
+	if (cnt) avg = sum / cnt;
+	p->se.child_burst = max(avg, p->se.burst_penalty);
+	p->se.child_burst_cnt = cnt;
+	p->se.child_burst_last_cached = now;
+}
+
+static inline void update_child_burst_direct(struct task_struct *p, u64 now) {
+	struct task_struct *child;
+	u32 cnt = 0;
+	u32 sum = 0;
+
+	list_for_each_entry(child, &p->children, sibling) {
+		if (!task_is_inheritable(child)) continue;
+		cnt++;
+		sum += child->se.burst_penalty;
+	}
+
+	__update_child_burst_cache(p, cnt, sum, now);
+}
+
+static inline u8 __inherit_burst_direct(struct task_struct *p, u64 now) {
+	struct task_struct *parent = p->real_parent;
+	if (child_burst_cache_expired(parent, now))
+		update_child_burst_direct(parent, now);
+
+	return parent->se.child_burst;
+}
+
+static void update_child_burst_topological(
+	struct task_struct *p, u64 now, u32 depth, u32 *acnt, u32 *asum) {
+	struct task_struct *child, *dec;
+	u32 cnt = 0, dcnt = 0;
+	u32 sum = 0;
+
+	list_for_each_entry(child, &p->children, sibling) {
+		dec = child;
+		while ((dcnt = count_child_tasks(dec)) == 1)
+			dec = list_first_entry(&dec->children, struct task_struct, sibling);
+		
+		if (!dcnt || !depth) {
+			if (!task_is_inheritable(dec)) continue;
+			cnt++;
+			sum += dec->se.burst_penalty;
+			continue;
+		}
+		if (!child_burst_cache_expired(dec, now)) {
+			cnt += dec->se.child_burst_cnt;
+			sum += (u32)dec->se.child_burst * dec->se.child_burst_cnt;
+			continue;
+		}
+		update_child_burst_topological(dec, now, depth - 1, &cnt, &sum);
+	}
+
+	__update_child_burst_cache(p, cnt, sum, now);
+	*acnt += cnt;
+	*asum += sum;
+}
+
+static inline u8 __inherit_burst_topological(struct task_struct *p, u64 now) {
+	struct task_struct *anc = p->real_parent;
+	u32 cnt = 0, sum = 0;
+
+	while (anc->real_parent != anc && count_child_tasks(anc) == 1)
+		anc = anc->real_parent;
+
+	if (child_burst_cache_expired(anc, now))
+		update_child_burst_topological(
+			anc, now, sched_burst_fork_atavistic - 1, &cnt, &sum);
+
+	return anc->se.child_burst;
+}
+
+static inline void inherit_burst(struct task_struct *p) {
+	u8 burst_cache;
+	u64 now = ktime_get_ns();
+
+	read_lock(&tasklist_lock);
+	burst_cache = likely(sched_burst_fork_atavistic)?
+		__inherit_burst_topological(p, now):
+		__inherit_burst_direct(p, now);
+	read_unlock(&tasklist_lock);
+
+	p->se.prev_burst_penalty = max(p->se.prev_burst_penalty, burst_cache);
+}
+
+static void sched_post_fork_bore(struct task_struct *p) {
+	if (p->sched_class == &fair_sched_class)
+		inherit_burst(p);
+	p->se.burst_penalty = p->se.prev_burst_penalty;
+}
+#endif // CONFIG_SCHED_BORE
+
 /*
  * Perform scheduler related setup for a newly forked process p.
  * p is forked by current.
@@ -4532,6 +4664,9 @@ static void __sched_fork(unsigned long clone_flags, struct task_struct *p)
 	p->se.prev_sum_exec_runtime	= 0;
 	p->se.nr_migrations		= 0;
 	p->se.vruntime			= 0;
+#ifdef CONFIG_SCHED_BORE
+	sched_fork_bore(p);
+#endif // CONFIG_SCHED_BORE
 	p->se.vlag			= 0;
 	p->se.slice			= sysctl_sched_base_slice;
 	INIT_LIST_HEAD(&p->se.group_node);
@@ -4848,6 +4983,9 @@ void sched_cgroup_fork(struct task_struct *p, struct kernel_clone_args *kargs)
 
 void sched_post_fork(struct task_struct *p)
 {
+#ifdef CONFIG_SCHED_BORE
+	sched_post_fork_bore(p);
+#endif // CONFIG_SCHED_BORE
 	uclamp_post_fork(p);
 }
 
@@ -9931,6 +10069,11 @@ void __init sched_init(void)
 	BUG_ON(&dl_sched_class != &stop_sched_class + 1);
 #endif
 
+#ifdef CONFIG_SCHED_BORE
+	sched_init_bore();
+	printk(KERN_INFO "BORE (Burst-Oriented Response Enhancer) CPU Scheduler modification 5.2.0 by Masahito Suzuki");
+#endif // CONFIG_SCHED_BORE
+
 	wait_bit_init();
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 8d5d98a5834df..b1786126171e9 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -167,7 +167,52 @@ static const struct file_operations sched_feat_fops = {
 };
 
 #ifdef CONFIG_SMP
+#ifdef CONFIG_SCHED_BORE
+static ssize_t sched_min_base_slice_write(struct file *filp, const char __user *ubuf,
+				   size_t cnt, loff_t *ppos)
+{
+	char buf[16];
+	unsigned int value;
+
+	if (cnt > 15)
+		cnt = 15;
+
+	if (copy_from_user(&buf, ubuf, cnt))
+		return -EFAULT;
+	buf[cnt] = '\0';
+
+	if (kstrtouint(buf, 10, &value))
+		return -EINVAL;
 
+	if (!value)
+		return -EINVAL;
+
+	sysctl_sched_min_base_slice = value;
+	sched_update_min_base_slice();
+
+	*ppos += cnt;
+	return cnt;
+}
+
+static int sched_min_base_slice_show(struct seq_file *m, void *v)
+{
+	seq_printf(m, "%d\n", sysctl_sched_min_base_slice);
+	return 0;
+}
+
+static int sched_min_base_slice_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, sched_min_base_slice_show, NULL);
+}
+
+static const struct file_operations sched_min_base_slice_fops = {
+	.open		= sched_min_base_slice_open,
+	.write		= sched_min_base_slice_write,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+#else // !CONFIG_SCHED_BORE
 static ssize_t sched_scaling_write(struct file *filp, const char __user *ubuf,
 				   size_t cnt, loff_t *ppos)
 {
@@ -213,7 +258,7 @@ static const struct file_operations sched_scaling_fops = {
 	.llseek		= seq_lseek,
 	.release	= single_release,
 };
-
+#endif // CONFIG_SCHED_BORE
 #endif /* SMP */
 
 #ifdef CONFIG_PREEMPT_DYNAMIC
@@ -347,13 +392,20 @@ static __init int sched_init_debug(void)
 	debugfs_create_file("preempt", 0644, debugfs_sched, NULL, &sched_dynamic_fops);
 #endif
 
+#ifdef CONFIG_SCHED_BORE
+	debugfs_create_file("min_base_slice_ns", 0644, debugfs_sched, NULL, &sched_min_base_slice_fops);
+	debugfs_create_u32("base_slice_ns", 0400, debugfs_sched, &sysctl_sched_base_slice);
+#else // !CONFIG_SCHED_BORE
 	debugfs_create_u32("base_slice_ns", 0644, debugfs_sched, &sysctl_sched_base_slice);
+#endif // CONFIG_SCHED_BORE
 
 	debugfs_create_u32("latency_warn_ms", 0644, debugfs_sched, &sysctl_resched_latency_warn_ms);
 	debugfs_create_u32("latency_warn_once", 0644, debugfs_sched, &sysctl_resched_latency_warn_once);
 
 #ifdef CONFIG_SMP
+#if !defined(CONFIG_SCHED_BORE)
 	debugfs_create_file("tunable_scaling", 0644, debugfs_sched, NULL, &sched_scaling_fops);
+#endif // CONFIG_SCHED_BORE
 	debugfs_create_u32("migration_cost_ns", 0644, debugfs_sched, &sysctl_sched_migration_cost);
 	debugfs_create_u32("nr_migrate", 0644, debugfs_sched, &sysctl_sched_nr_migrate);
 
@@ -595,6 +647,9 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p)
 		SPLIT_NS(schedstat_val_or_zero(p->stats.sum_sleep_runtime)),
 		SPLIT_NS(schedstat_val_or_zero(p->stats.sum_block_runtime)));
 
+#ifdef CONFIG_SCHED_BORE
+	SEQ_printf(m, " %2d", p->se.burst_score);
+#endif // CONFIG_SCHED_BORE
 #ifdef CONFIG_NUMA_BALANCING
 	SEQ_printf(m, " %d %d", task_node(p), task_numa_group_id(p));
 #endif
@@ -1068,6 +1123,9 @@ void proc_sched_show_task(struct task_struct *p, struct pid_namespace *ns,
 
 	P(se.load.weight);
 #ifdef CONFIG_SMP
+#ifdef CONFIG_SCHED_BORE
+	P(se.burst_score);
+#endif // CONFIG_SCHED_BORE
 	P(se.avg.load_sum);
 	P(se.avg.runnable_sum);
 	P(se.avg.util_sum);
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 213c94d027a4c..3c2d149d0235d 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -19,6 +19,9 @@
  *
  *  Adaptive scheduling granularity, math enhancements by Peter Zijlstra
  *  Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra
+ *
+ *  Burst-Oriented Response Enhancer (BORE) CPU Scheduler
+ *  Copyright (C) 2021-2024 Masahito Suzuki <firelzrd@gmail.com>
  */
 #include <linux/energy_model.h>
 #include <linux/mmap_lock.h>
@@ -64,20 +67,125 @@
  *   SCHED_TUNABLESCALING_LOG - scaled logarithmical, *1+ilog(ncpus)
  *   SCHED_TUNABLESCALING_LINEAR - scaled linear, *ncpus
  *
- * (default SCHED_TUNABLESCALING_LOG = *(1+ilog(ncpus))
+ * (BORE  default SCHED_TUNABLESCALING_NONE = *1 constant)
+ * (EEVDF default SCHED_TUNABLESCALING_LOG  = *(1+ilog(ncpus))
  */
+#ifdef CONFIG_SCHED_BORE
+unsigned int sysctl_sched_tunable_scaling = SCHED_TUNABLESCALING_NONE;
+#else // !CONFIG_SCHED_BORE
 unsigned int sysctl_sched_tunable_scaling = SCHED_TUNABLESCALING_LOG;
+#endif // CONFIG_SCHED_BORE
 
 /*
  * Minimal preemption granularity for CPU-bound tasks:
  *
- * (default: 0.75 msec * (1 + ilog(ncpus)), units: nanoseconds)
+ * (BORE  default: max(1 sec / HZ, min_base_slice) constant, units: nanoseconds)
+ * (EEVDF default: 0.75 msec * (1 + ilog(ncpus)), units: nanoseconds)
  */
+#ifdef CONFIG_SCHED_BORE
+unsigned int            sysctl_sched_base_slice = 1000000000ULL / HZ;
+static unsigned int configured_sched_base_slice = 1000000000ULL / HZ;
+unsigned int        sysctl_sched_min_base_slice =    2000000ULL;
+#else // !CONFIG_SCHED_BORE
 unsigned int sysctl_sched_base_slice			= 750000ULL;
 static unsigned int normalized_sysctl_sched_base_slice	= 750000ULL;
+#endif // CONFIG_SCHED_BORE
 
 const_debug unsigned int sysctl_sched_migration_cost	= 500000UL;
 
+#ifdef CONFIG_SCHED_BORE
+u8   __read_mostly sched_bore                   = 1;
+u8   __read_mostly sched_burst_smoothness_long  = 1;
+u8   __read_mostly sched_burst_smoothness_short = 0;
+u8   __read_mostly sched_burst_fork_atavistic   = 2;
+u8   __read_mostly sched_burst_penalty_offset   = 22;
+uint __read_mostly sched_burst_penalty_scale    = 1280;
+uint __read_mostly sched_burst_cache_lifetime   = 60000000;
+static int __maybe_unused sixty_four     = 64;
+static int __maybe_unused maxval_12_bits = 4095;
+
+#define MAX_BURST_PENALTY (39U <<2)
+
+static inline u32 log2plus1_u64_u32f8(u64 v) {
+	u32 msb = fls64(v);
+	s32 excess_bits = msb - 9;
+    u8 fractional = (0 <= excess_bits)? v >> excess_bits: v << -excess_bits;
+	return msb << 8 | fractional;
+}
+
+static inline u32 calc_burst_penalty(u64 burst_time) {
+	u32 greed, tolerance, penalty, scaled_penalty;
+	
+	greed = log2plus1_u64_u32f8(burst_time);
+	tolerance = sched_burst_penalty_offset << 8;
+	penalty = max(0, (s32)greed - (s32)tolerance);
+	scaled_penalty = penalty * sched_burst_penalty_scale >> 16;
+
+	return min(MAX_BURST_PENALTY, scaled_penalty);
+}
+
+static inline u64 scale_slice(u64 delta, struct sched_entity *se) {
+	return mul_u64_u32_shr(delta, sched_prio_to_wmult[se->burst_score], 22);
+}
+
+static inline u64 __unscale_slice(u64 delta, u8 score) {
+	return mul_u64_u32_shr(delta, sched_prio_to_weight[score], 10);
+}
+
+static inline u64 unscale_slice(u64 delta, struct sched_entity *se) {
+	return __unscale_slice(delta, se->burst_score);
+}
+
+void reweight_task(struct task_struct *p, int prio);
+
+static void update_burst_score(struct sched_entity *se) {
+	if (!entity_is_task(se)) return;
+	struct task_struct *p = task_of(se);
+	u8 prio = p->static_prio - MAX_RT_PRIO;
+	u8 prev_prio = min(39, prio + se->burst_score);
+
+	se->burst_score = se->burst_penalty >> 2;
+
+	u8 new_prio = min(39, prio + se->burst_score);
+	if (new_prio != prev_prio)
+		reweight_task(p, new_prio);
+}
+
+static void update_burst_penalty(struct sched_entity *se) {
+	se->curr_burst_penalty = calc_burst_penalty(se->burst_time);
+	se->burst_penalty = max(se->prev_burst_penalty, se->curr_burst_penalty);
+	update_burst_score(se);
+}
+
+static inline u32 binary_smooth(u32 new, u32 old) {
+  int increment = new - old;
+  return (0 <= increment)?
+    old + ( increment >> (int)sched_burst_smoothness_long):
+    old - (-increment >> (int)sched_burst_smoothness_short);
+}
+
+static void restart_burst(struct sched_entity *se) {
+	se->burst_penalty = se->prev_burst_penalty =
+		binary_smooth(se->curr_burst_penalty, se->prev_burst_penalty);
+	se->curr_burst_penalty = 0;
+	se->burst_time = 0;
+	update_burst_score(se);
+}
+
+static void restart_burst_rescale_deadline(struct sched_entity *se) {
+	s64 vscaled, wremain, vremain = se->deadline - se->vruntime;
+	u8 prev_score = se->burst_score;
+	restart_burst(se);
+	if (prev_score > se->burst_score) {
+		wremain = __unscale_slice(abs(vremain), prev_score);
+		vscaled = scale_slice(wremain, se);
+		if (unlikely(vremain < 0))
+			vscaled = -vscaled;
+		se->deadline = se->vruntime + vscaled;
+	}
+}
+#endif // CONFIG_SCHED_BORE
+
 int sched_thermal_decay_shift;
 static int __init setup_sched_thermal_decay_shift(char *str)
 {
@@ -137,6 +245,69 @@ static unsigned int sysctl_numa_balancing_promote_rate_limit = 65536;
 
 #ifdef CONFIG_SYSCTL
 static struct ctl_table sched_fair_sysctls[] = {
+#ifdef CONFIG_SCHED_BORE
+	{
+		.procname	= "sched_bore",
+		.data		= &sched_bore,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler = proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ONE,
+		.extra2		= SYSCTL_ONE,
+	},
+	{
+		.procname	= "sched_burst_smoothness_long",
+		.data		= &sched_burst_smoothness_long,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler = proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= SYSCTL_ONE,
+	},
+	{
+		.procname	= "sched_burst_smoothness_short",
+		.data		= &sched_burst_smoothness_short,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler = proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= SYSCTL_ONE,
+	},
+	{
+		.procname	= "sched_burst_fork_atavistic",
+		.data		= &sched_burst_fork_atavistic,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler = proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= SYSCTL_THREE,
+	},
+	{
+		.procname	= "sched_burst_penalty_offset",
+		.data		= &sched_burst_penalty_offset,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler = proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= &sixty_four,
+	},
+	{
+		.procname	= "sched_burst_penalty_scale",
+		.data		= &sched_burst_penalty_scale,
+		.maxlen		= sizeof(uint),
+		.mode		= 0644,
+		.proc_handler = proc_douintvec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= &maxval_12_bits,
+	},
+	{
+		.procname	= "sched_burst_cache_lifetime",
+		.data		= &sched_burst_cache_lifetime,
+		.maxlen		= sizeof(uint),
+		.mode		= 0644,
+		.proc_handler = proc_douintvec,
+	},
+#endif // CONFIG_SCHED_BORE
 #ifdef CONFIG_CFS_BANDWIDTH
 	{
 		.procname       = "sched_cfs_bandwidth_slice_us",
@@ -195,6 +366,13 @@ static inline void update_load_set(struct load_weight *lw, unsigned long w)
  *
  * This idea comes from the SD scheduler of Con Kolivas:
  */
+#ifdef CONFIG_SCHED_BORE
+static void update_sysctl(void) {
+	sysctl_sched_base_slice =
+		max(sysctl_sched_min_base_slice, configured_sched_base_slice);
+}
+void sched_update_min_base_slice(void) { update_sysctl(); }
+#else // !CONFIG_SCHED_BORE
 static unsigned int get_update_sysctl_factor(void)
 {
 	unsigned int cpus = min_t(unsigned int, num_online_cpus(), 8);
@@ -225,6 +403,7 @@ static void update_sysctl(void)
 	SET_SYSCTL(sched_base_slice);
 #undef SET_SYSCTL
 }
+#endif // CONFIG_SCHED_BORE
 
 void __init sched_init_granularity(void)
 {
@@ -702,6 +881,9 @@ static s64 entity_lag(u64 avruntime, struct sched_entity *se)
 
 	vlag = avruntime - se->vruntime;
 	limit = calc_delta_fair(max_t(u64, 2*se->slice, TICK_NSEC), se);
+#ifdef CONFIG_SCHED_BORE
+	limit >>= 1;
+#endif // CONFIG_SCHED_BORE
 
 	return clamp(vlag, -limit, limit);
 }
@@ -862,6 +1044,39 @@ struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq)
 	return __node_2_se(left);
 }
 
+static inline bool pick_curr(struct cfs_rq *cfs_rq,
+			     struct sched_entity *curr, struct sched_entity *wakee)
+{
+	/*
+	 * Nothing to preserve...
+	 */
+	if (!curr || !sched_feat(RESPECT_SLICE))
+		return false;
+
+	/*
+	 * Allow preemption at the 0-lag point -- even if not all of the slice
+	 * is consumed. Note: placement of positive lag can push V left and render
+	 * @curr instantly ineligible irrespective the time on-cpu.
+	 */
+	if (sched_feat(RUN_TO_PARITY) && !entity_eligible(cfs_rq, curr))
+		return false;
+
+	/*
+	 * Don't preserve @curr when the @wakee has a shorter slice and earlier
+	 * deadline. IOW, explicitly allow preemption.
+	 */
+	if (sched_feat(PREEMPT_SHORT) && wakee &&
+	    wakee->slice < curr->slice &&
+	    (s64)(wakee->deadline - curr->deadline) < 0)
+		return false;
+
+	/*
+	 * Preserve @curr to allow it to finish its first slice.
+	 * See the HACK in set_next_entity().
+	 */
+	return curr->vlag == curr->deadline;
+}
+
 /*
  * Earliest Eligible Virtual Deadline First
  *
@@ -881,28 +1096,27 @@ struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq)
  *
  * Which allows tree pruning through eligibility.
  */
-static struct sched_entity *pick_eevdf(struct cfs_rq *cfs_rq)
+static struct sched_entity *pick_eevdf(struct cfs_rq *cfs_rq, struct sched_entity *wakee)
 {
 	struct rb_node *node = cfs_rq->tasks_timeline.rb_root.rb_node;
 	struct sched_entity *se = __pick_first_entity(cfs_rq);
 	struct sched_entity *curr = cfs_rq->curr;
 	struct sched_entity *best = NULL;
 
+	if (curr && !curr->on_rq)
+		curr = NULL;
+
 	/*
 	 * We can safely skip eligibility check if there is only one entity
 	 * in this cfs_rq, saving some cycles.
 	 */
 	if (cfs_rq->nr_running == 1)
-		return curr && curr->on_rq ? curr : se;
-
-	if (curr && (!curr->on_rq || !entity_eligible(cfs_rq, curr)))
-		curr = NULL;
+		return curr ?: se;
 
 	/*
-	 * Once selected, run a task until it either becomes non-eligible or
-	 * until it gets a new slice. See the HACK in set_next_entity().
+	 * Preserve @curr to let it finish its slice.
 	 */
-	if (sched_feat(RUN_TO_PARITY) && curr && curr->vlag == curr->deadline)
+	if (pick_curr(cfs_rq, curr, wakee))
 		return curr;
 
 	/* Pick the leftmost entity if it's eligible */
@@ -961,6 +1175,7 @@ struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq)
  * Scheduling class statistics methods:
  */
 #ifdef CONFIG_SMP
+#if !defined(CONFIG_SCHED_BORE)
 int sched_update_scaling(void)
 {
 	unsigned int factor = get_update_sysctl_factor();
@@ -972,6 +1187,7 @@ int sched_update_scaling(void)
 
 	return 0;
 }
+#endif // CONFIG_SCHED_BORE
 #endif
 #endif
 
@@ -1171,7 +1387,13 @@ static void update_curr(struct cfs_rq *cfs_rq)
 	if (unlikely(delta_exec <= 0))
 		return;
 
+#ifdef CONFIG_SCHED_BORE
+	curr->burst_time += delta_exec;
+	update_burst_penalty(curr);
+	curr->vruntime += max(1ULL, calc_delta_fair(delta_exec, curr));
+#else // !CONFIG_SCHED_BORE
 	curr->vruntime += calc_delta_fair(delta_exec, curr);
+#endif // CONFIG_SCHED_BORE
 	update_deadline(cfs_rq, curr);
 	update_min_vruntime(cfs_rq);
 
@@ -5183,6 +5405,9 @@ place_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
 	 *
 	 * EEVDF: placement strategy #1 / #2
 	 */
+#ifdef CONFIG_SCHED_BORE
+	if (se->vlag)
+#endif // CONFIG_SCHED_BORE
 	if (sched_feat(PLACE_LAG) && cfs_rq->nr_running) {
 		struct sched_entity *curr = cfs_rq->curr;
 		unsigned long load;
@@ -5472,7 +5697,7 @@ pick_next_entity(struct cfs_rq *cfs_rq)
 	    cfs_rq->next && entity_eligible(cfs_rq, cfs_rq->next))
 		return cfs_rq->next;
 
-	return pick_eevdf(cfs_rq);
+	return pick_eevdf(cfs_rq, NULL);
 }
 
 static bool check_cfs_rq_runtime(struct cfs_rq *cfs_rq);
@@ -6835,6 +7060,14 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags)
 	bool was_sched_idle = sched_idle_rq(rq);
 
 	util_est_dequeue(&rq->cfs, p);
+#ifdef CONFIG_SCHED_BORE
+	if (task_sleep) {
+		cfs_rq = cfs_rq_of(se);
+		if (cfs_rq->curr == se)
+			update_curr(cfs_rq);
+		restart_burst(se);
+	}
+#endif // CONFIG_SCHED_BORE
 
 	for_each_sched_entity(se) {
 		cfs_rq = cfs_rq_of(se);
@@ -8369,10 +8602,7 @@ static void check_preempt_wakeup_fair(struct rq *rq, struct task_struct *p, int
 	cfs_rq = cfs_rq_of(se);
 	update_curr(cfs_rq);
 
-	/*
-	 * XXX pick_eevdf(cfs_rq) != se ?
-	 */
-	if (pick_eevdf(cfs_rq) == pse)
+	if (pick_eevdf(cfs_rq, pse) == pse)
 		goto preempt;
 
 	return;
@@ -8590,16 +8820,25 @@ static void yield_task_fair(struct rq *rq)
 	/*
 	 * Are we the only task in the tree?
 	 */
+#if !defined(CONFIG_SCHED_BORE)
 	if (unlikely(rq->nr_running == 1))
 		return;
 
 	clear_buddies(cfs_rq, se);
+#endif // CONFIG_SCHED_BORE
 
 	update_rq_clock(rq);
 	/*
 	 * Update run-time statistics of the 'current'.
 	 */
 	update_curr(cfs_rq);
+#ifdef CONFIG_SCHED_BORE
+	restart_burst_rescale_deadline(se);
+	if (unlikely(rq->nr_running == 1))
+		return;
+
+	clear_buddies(cfs_rq, se);
+#endif // CONFIG_SCHED_BORE
 	/*
 	 * Tell update_rq_clock() that we've just updated,
 	 * so we don't do microscopic update in schedule()
@@ -12660,6 +12899,9 @@ static void task_fork_fair(struct task_struct *p)
 	curr = cfs_rq->curr;
 	if (curr)
 		update_curr(cfs_rq);
+#ifdef CONFIG_SCHED_BORE
+	update_burst_score(se);
+#endif // CONFIG_SCHED_BORE
 	place_entity(cfs_rq, se, ENQUEUE_INITIAL);
 	rq_unlock(rq, &rf);
 }
diff --git a/kernel/sched/features.h b/kernel/sched/features.h
index 143f55df890b1..bfeb9f65383d9 100644
--- a/kernel/sched/features.h
+++ b/kernel/sched/features.h
@@ -5,8 +5,26 @@
  * sleep+wake cycles. EEVDF placement strategy #1, #2 if disabled.
  */
 SCHED_FEAT(PLACE_LAG, true)
+/*
+ * Give new tasks half a slice to ease into the competition.
+ */
 SCHED_FEAT(PLACE_DEADLINE_INITIAL, true)
-SCHED_FEAT(RUN_TO_PARITY, true)
+/*
+ * Inhibit (wakeup) preemption until the current task has exhausted its slice.
+ */
+#ifdef CONFIG_SCHED_BORE
+SCHED_FEAT(RESPECT_SLICE, false)
+#else // !CONFIG_SCHED_BORE
+SCHED_FEAT(RESPECT_SLICE, true)
+#endif // CONFIG_SCHED_BORE
+/*
+ * Relax RESPECT_SLICE to allow preemption once current has reached 0-lag.
+ */
+SCHED_FEAT(RUN_TO_PARITY, false)
+/*
+ * Allow tasks with a shorter slice to disregard RESPECT_SLICE
+ */
+SCHED_FEAT(PREEMPT_SHORT, true)
 
 /*
  * Prefer to schedule the task we woke last (assuming it failed
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index ae50f212775e5..8c976d27f6e9c 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1967,7 +1967,11 @@ static inline void dirty_sched_domain_sysctl(int cpu)
 }
 #endif
 
+#ifdef CONFIG_SCHED_BORE
+extern void sched_update_min_base_slice(void);
+#else // !CONFIG_SCHED_BORE
 extern int sched_update_scaling(void);
+#endif // CONFIG_SCHED_BORE
 
 static inline const struct cpumask *task_user_cpus(struct task_struct *p)
 {
@@ -2554,6 +2558,9 @@ extern const_debug unsigned int sysctl_sched_nr_migrate;
 extern const_debug unsigned int sysctl_sched_migration_cost;
 
 extern unsigned int sysctl_sched_base_slice;
+#ifdef CONFIG_SCHED_BORE
+extern unsigned int sysctl_sched_min_base_slice;
+#endif // CONFIG_SCHED_BORE
 
 #ifdef CONFIG_SCHED_DEBUG
 extern int sysctl_resched_latency_warn_ms;
-- 
2.45.1