Boot Linux faster!

Check our new training course

Boot Linux faster!

Check our new training course
and Creative Commons CC-BY-SA
lecture and lab materials

Bootlin logo

Elixir Cross Referencer

   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
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
/*
 * Copyright (c) 2018 Intel Corporation
 *
 * SPDX-License-Identifier: Apache-2.0
 */
#include <kernel.h>
#include <ksched.h>
#include <spinlock.h>
#include <sched_priq.h>
#include <wait_q.h>
#include <kswap.h>
#include <kernel_arch_func.h>
#include <syscall_handler.h>
#include <drivers/system_timer.h>
#include <stdbool.h>

#if defined(CONFIG_SCHED_DUMB)
#define _priq_run_add		z_priq_dumb_add
#define _priq_run_remove	z_priq_dumb_remove
# if defined(CONFIG_SCHED_CPU_MASK)
#  define _priq_run_best	_priq_dumb_mask_best
# else
#  define _priq_run_best	z_priq_dumb_best
# endif
#elif defined(CONFIG_SCHED_SCALABLE)
#define _priq_run_add		z_priq_rb_add
#define _priq_run_remove	z_priq_rb_remove
#define _priq_run_best		z_priq_rb_best
#elif defined(CONFIG_SCHED_MULTIQ)
#define _priq_run_add		z_priq_mq_add
#define _priq_run_remove	z_priq_mq_remove
#define _priq_run_best		z_priq_mq_best
#endif

#if defined(CONFIG_WAITQ_SCALABLE)
#define z_priq_wait_add		z_priq_rb_add
#define _priq_wait_remove	z_priq_rb_remove
#define _priq_wait_best		z_priq_rb_best
#elif defined(CONFIG_WAITQ_DUMB)
#define z_priq_wait_add		z_priq_dumb_add
#define _priq_wait_remove	z_priq_dumb_remove
#define _priq_wait_best		z_priq_dumb_best
#endif

/* the only struct z_kernel instance */
struct z_kernel _kernel;

static struct k_spinlock sched_spinlock;

#define LOCKED(lck) for (k_spinlock_key_t __i = {},			\
					  __key = k_spin_lock(lck);	\
			!__i.key;					\
			k_spin_unlock(lck, __key), __i.key = 1)

static inline int is_preempt(struct k_thread *thread)
{
#ifdef CONFIG_PREEMPT_ENABLED
	/* explanation in kernel_struct.h */
	return thread->base.preempt <= _PREEMPT_THRESHOLD;
#else
	return 0;
#endif
}

static inline int is_metairq(struct k_thread *thread)
{
#if CONFIG_NUM_METAIRQ_PRIORITIES > 0
	return (thread->base.prio - K_HIGHEST_THREAD_PRIO)
		< CONFIG_NUM_METAIRQ_PRIORITIES;
#else
	return 0;
#endif
}

#if CONFIG_ASSERT
static inline int is_thread_dummy(struct k_thread *thread)
{
	return !!(thread->base.thread_state & _THREAD_DUMMY);
}
#endif

static inline bool is_idle(struct k_thread *thread)
{
#ifdef CONFIG_SMP
	return thread->base.is_idle;
#else
	extern k_tid_t const _idle_thread;

	return thread == _idle_thread;
#endif
}

bool z_is_t1_higher_prio_than_t2(struct k_thread *t1, struct k_thread *t2)
{
	if (t1->base.prio < t2->base.prio) {
		return true;
	}

#ifdef CONFIG_SCHED_DEADLINE
	/* Note that we don't care about wraparound conditions.  The
	 * expectation is that the application will have arranged to
	 * block the threads, change their priorities or reset their
	 * deadlines when the job is complete.  Letting the deadlines
	 * go negative is fine and in fact prevents aliasing bugs.
	 */
	if (t1->base.prio == t2->base.prio) {
		int now = (int) k_cycle_get_32();
		int dt1 = t1->base.prio_deadline - now;
		int dt2 = t2->base.prio_deadline - now;

		return dt1 < dt2;
	}
#endif

	return false;
}

static ALWAYS_INLINE bool should_preempt(struct k_thread *th, int preempt_ok)
{
	/* Preemption is OK if it's being explicitly allowed by
	 * software state (e.g. the thread called k_yield())
	 */
	if (preempt_ok != 0) {
		return true;
	}

	__ASSERT(_current != NULL, "");

	/* Or if we're pended/suspended/dummy (duh) */
	if (z_is_thread_prevented_from_running(_current)) {
		return true;
	}

	/* Edge case on ARM where a thread can be pended out of an
	 * interrupt handler before the "synchronous" swap starts
	 * context switching.  Platforms with atomic swap can never
	 * hit this.
	 */
	if (IS_ENABLED(CONFIG_SWAP_NONATOMIC)
	    && z_is_thread_timeout_active(th)) {
		return true;
	}

	/* Otherwise we have to be running a preemptible thread or
	 * switching to a metairq
	 */
	if (is_preempt(_current) || is_metairq(th)) {
		return true;
	}

	/* The idle threads can look "cooperative" if there are no
	 * preemptible priorities (this is sort of an API glitch).
	 * They must always be preemptible.
	 */
	if (!IS_ENABLED(CONFIG_PREEMPT_ENABLED) && is_idle(_current)) {
		return true;
	}

	return false;
}

#ifdef CONFIG_SCHED_CPU_MASK
static ALWAYS_INLINE struct k_thread *_priq_dumb_mask_best(sys_dlist_t *pq)
{
	/* With masks enabled we need to be prepared to walk the list
	 * looking for one we can run
	 */
	struct k_thread *t;

	SYS_DLIST_FOR_EACH_CONTAINER(pq, t, base.qnode_dlist) {
		if ((t->base.cpu_mask & BIT(_current_cpu->id)) != 0) {
			return t;
		}
	}
	return NULL;
}
#endif

static ALWAYS_INLINE struct k_thread *next_up(void)
{
#ifndef CONFIG_SMP
	/* In uniprocessor mode, we can leave the current thread in
	 * the queue (actually we have to, otherwise the assembly
	 * context switch code for all architectures would be
	 * responsible for putting it back in z_swap and ISR return!),
	 * which makes this choice simple.
	 */
	struct k_thread *th = _priq_run_best(&_kernel.ready_q.runq);

	return th ? th : _current_cpu->idle_thread;
#else

	/* Under SMP, the "cache" mechanism for selecting the next
	 * thread doesn't work, so we have more work to do to test
	 * _current against the best choice from the queue.
	 *
	 * Subtle note on "queued": in SMP mode, _current does not
	 * live in the queue, so this isn't exactly the same thing as
	 * "ready", it means "is _current already added back to the
	 * queue such that we don't want to re-add it".
	 */
	int queued = z_is_thread_queued(_current);
	int active = !z_is_thread_prevented_from_running(_current);

	/* Choose the best thread that is not current */
	struct k_thread *th = _priq_run_best(&_kernel.ready_q.runq);
	if (th == NULL) {
		th = _current_cpu->idle_thread;
	}

	if (active) {
		if (!queued &&
		    !z_is_t1_higher_prio_than_t2(th, _current)) {
			th = _current;
		}

		if (!should_preempt(th, _current_cpu->swap_ok)) {
			th = _current;
		}
	}

	/* Put _current back into the queue */
	if (th != _current && active && !is_idle(_current) && !queued) {
		_priq_run_add(&_kernel.ready_q.runq, _current);
		z_mark_thread_as_queued(_current);
	}

	/* Take the new _current out of the queue */
	if (z_is_thread_queued(th)) {
		_priq_run_remove(&_kernel.ready_q.runq, th);
	}
	z_mark_thread_as_not_queued(th);

	return th;
#endif
}

#ifdef CONFIG_TIMESLICING

static int slice_time;
static int slice_max_prio;

#ifdef CONFIG_SWAP_NONATOMIC
/* If z_swap() isn't atomic, then it's possible for a timer interrupt
 * to try to timeslice away _current after it has already pended
 * itself but before the corresponding context switch.  Treat that as
 * a noop condition in z_time_slice().
 */
static struct k_thread *pending_current;
#endif

static void reset_time_slice(void)
{
	/* Add the elapsed time since the last announced tick to the
	 * slice count, as we'll see those "expired" ticks arrive in a
	 * FUTURE z_time_slice() call.
	 */
	_current_cpu->slice_ticks = slice_time + z_clock_elapsed();

	z_set_timeout_expiry(slice_time, false);
}

void k_sched_time_slice_set(s32_t slice, int prio)
{
	LOCKED(&sched_spinlock) {
		_current_cpu->slice_ticks = 0;
		slice_time = z_ms_to_ticks(slice);
		slice_max_prio = prio;
		reset_time_slice();
	}
}

static inline int sliceable(struct k_thread *t)
{
	return is_preempt(t)
		&& !z_is_prio_higher(t->base.prio, slice_max_prio)
		&& !is_idle(t)
		&& !z_is_thread_timeout_active(t);
}

/* Called out of each timer interrupt */
void z_time_slice(int ticks)
{
#ifdef CONFIG_SWAP_NONATOMIC
	if (pending_current == _current) {
		reset_time_slice();
		return;
	}
	pending_current = NULL;
#endif

	if (slice_time && sliceable(_current)) {
		if (ticks >= _current_cpu->slice_ticks) {
			z_move_thread_to_end_of_prio_q(_current);
			reset_time_slice();
		} else {
			_current_cpu->slice_ticks -= ticks;
		}
	}
}
#else
static void reset_time_slice(void) { /* !CONFIG_TIMESLICING */ }
#endif

static void update_cache(int preempt_ok)
{
#ifndef CONFIG_SMP
	struct k_thread *th = next_up();

	if (should_preempt(th, preempt_ok)) {
		if (th != _current) {
			reset_time_slice();
		}
		_kernel.ready_q.cache = th;
	} else {
		_kernel.ready_q.cache = _current;
	}

#else
	/* The way this works is that the CPU record keeps its
	 * "cooperative swapping is OK" flag until the next reschedule
	 * call or context switch.  It doesn't need to be tracked per
	 * thread because if the thread gets preempted for whatever
	 * reason the scheduler will make the same decision anyway.
	 */
	_current_cpu->swap_ok = preempt_ok;
#endif
}

void z_add_thread_to_ready_q(struct k_thread *thread)
{
	LOCKED(&sched_spinlock) {
		_priq_run_add(&_kernel.ready_q.runq, thread);
		z_mark_thread_as_queued(thread);
		update_cache(0);
	}
}

void z_move_thread_to_end_of_prio_q(struct k_thread *thread)
{
	LOCKED(&sched_spinlock) {
		_priq_run_remove(&_kernel.ready_q.runq, thread);
		_priq_run_add(&_kernel.ready_q.runq, thread);
		z_mark_thread_as_queued(thread);
		update_cache(thread == _current);
	}
}

void z_remove_thread_from_ready_q(struct k_thread *thread)
{
	LOCKED(&sched_spinlock) {
		if (z_is_thread_queued(thread)) {
			_priq_run_remove(&_kernel.ready_q.runq, thread);
			z_mark_thread_as_not_queued(thread);
		}
		update_cache(thread == _current);
	}
}

static void pend(struct k_thread *thread, _wait_q_t *wait_q, s32_t timeout)
{
	z_remove_thread_from_ready_q(thread);
	z_mark_thread_as_pending(thread);

	if (wait_q != NULL) {
		thread->base.pended_on = wait_q;
		z_priq_wait_add(&wait_q->waitq, thread);
	}

	if (timeout != K_FOREVER) {
		s32_t ticks = _TICK_ALIGN + z_ms_to_ticks(timeout);

		z_add_thread_timeout(thread, ticks);
	}

	sys_trace_thread_pend(thread);
}

void z_pend_thread(struct k_thread *thread, _wait_q_t *wait_q, s32_t timeout)
{
	__ASSERT_NO_MSG(thread == _current || is_thread_dummy(thread));
	pend(thread, wait_q, timeout);
}

static _wait_q_t *pended_on(struct k_thread *thread)
{
	__ASSERT_NO_MSG(thread->base.pended_on);

	return thread->base.pended_on;
}

ALWAYS_INLINE struct k_thread *z_find_first_thread_to_unpend(_wait_q_t *wait_q,
						     struct k_thread *from)
{
	ARG_UNUSED(from);

	struct k_thread *ret = NULL;

	LOCKED(&sched_spinlock) {
		ret = _priq_wait_best(&wait_q->waitq);
	}

	return ret;
}

ALWAYS_INLINE void z_unpend_thread_no_timeout(struct k_thread *thread)
{
	LOCKED(&sched_spinlock) {
		_priq_wait_remove(&pended_on(thread)->waitq, thread);
		z_mark_thread_as_not_pending(thread);
	}

	thread->base.pended_on = NULL;
}

#ifdef CONFIG_SYS_CLOCK_EXISTS
/* Timeout handler for *_thread_timeout() APIs */
void z_thread_timeout(struct _timeout *to)
{
	struct k_thread *th = CONTAINER_OF(to, struct k_thread, base.timeout);

	if (th->base.pended_on != NULL) {
		z_unpend_thread_no_timeout(th);
	}
	z_mark_thread_as_started(th);
	z_ready_thread(th);
}
#endif

int z_pend_curr_irqlock(u32_t key, _wait_q_t *wait_q, s32_t timeout)
{
	pend(_current, wait_q, timeout);

#if defined(CONFIG_TIMESLICING) && defined(CONFIG_SWAP_NONATOMIC)
	pending_current = _current;

	int ret = z_swap_irqlock(key);
	LOCKED(&sched_spinlock) {
		if (pending_current == _current) {
			pending_current = NULL;
		}
	}
	return ret;
#else
	return z_swap_irqlock(key);
#endif
}

int z_pend_curr(struct k_spinlock *lock, k_spinlock_key_t key,
	       _wait_q_t *wait_q, s32_t timeout)
{
#if defined(CONFIG_TIMESLICING) && defined(CONFIG_SWAP_NONATOMIC)
	pending_current = _current;
#endif
	pend(_current, wait_q, timeout);
	return z_swap(lock, key);
}

struct k_thread *z_unpend_first_thread(_wait_q_t *wait_q)
{
	struct k_thread *t = z_unpend1_no_timeout(wait_q);

	if (t != NULL) {
		(void)z_abort_thread_timeout(t);
	}

	return t;
}

void z_unpend_thread(struct k_thread *thread)
{
	z_unpend_thread_no_timeout(thread);
	(void)z_abort_thread_timeout(thread);
}

void z_thread_priority_set(struct k_thread *thread, int prio)
{
	bool need_sched = 0;

	LOCKED(&sched_spinlock) {
		need_sched = z_is_thread_ready(thread);

		if (need_sched) {
			_priq_run_remove(&_kernel.ready_q.runq, thread);
			thread->base.prio = prio;
			_priq_run_add(&_kernel.ready_q.runq, thread);
			update_cache(1);
		} else {
			thread->base.prio = prio;
		}
	}
	sys_trace_thread_priority_set(thread);

	if (IS_ENABLED(CONFIG_SMP) &&
	    !IS_ENABLED(CONFIG_SCHED_IPI_SUPPORTED)) {
		z_sched_ipi();
	}

	if (need_sched && _current->base.sched_locked == 0) {
		z_reschedule_unlocked();
	}
}

static inline int resched(void)
{
#ifdef CONFIG_SMP
	if (!_current_cpu->swap_ok) {
		return 0;
	}
	_current_cpu->swap_ok = 0;
#endif

	return !z_is_in_isr();
}

void z_reschedule(struct k_spinlock *lock, k_spinlock_key_t key)
{
	if (resched()) {
		z_swap(lock, key);
	} else {
		k_spin_unlock(lock, key);
	}
}

void z_reschedule_irqlock(u32_t key)
{
	if (resched()) {
		z_swap_irqlock(key);
	} else {
		irq_unlock(key);
	}
}

void k_sched_lock(void)
{
	LOCKED(&sched_spinlock) {
		z_sched_lock();
	}
}

void k_sched_unlock(void)
{
#ifdef CONFIG_PREEMPT_ENABLED
	__ASSERT(_current->base.sched_locked != 0, "");
	__ASSERT(!z_is_in_isr(), "");

	LOCKED(&sched_spinlock) {
		++_current->base.sched_locked;
		update_cache(1);
	}

	K_DEBUG("scheduler unlocked (%p:%d)\n",
		_current, _current->base.sched_locked);

	z_reschedule_unlocked();
#endif
}

#ifdef CONFIG_SMP
struct k_thread *z_get_next_ready_thread(void)
{
	struct k_thread *ret = 0;

	LOCKED(&sched_spinlock) {
		ret = next_up();
	}

	return ret;
}
#endif

/* Just a wrapper around _current = xxx with tracing */
static inline void set_current(struct k_thread *new_thread)
{
#ifdef CONFIG_TRACING
	sys_trace_thread_switched_out();
#endif
	_current = new_thread;
#ifdef CONFIG_TRACING
	sys_trace_thread_switched_in();
#endif
}

#ifdef CONFIG_USE_SWITCH
void *z_get_next_switch_handle(void *interrupted)
{
	_current->switch_handle = interrupted;

#ifdef CONFIG_SMP
	LOCKED(&sched_spinlock) {
		struct k_thread *th = next_up();

		if (_current != th) {
			reset_time_slice();
			_current_cpu->swap_ok = 0;
			set_current(th);
#ifdef SPIN_VALIDATE
			/* Changed _current!  Update the spinlock
			 * bookeeping so the validation doesn't get
			 * confused when the "wrong" thread tries to
			 * release the lock.
			 */
			z_spin_lock_set_owner(&sched_spinlock);
#endif
		}
	}
#else
	set_current(z_get_next_ready_thread());
#endif

	/* Some architectures don't have a working IPI, so the best we
	 * can do there is check the abort status of the current
	 * thread here on ISR exit
	 */
	if (IS_ENABLED(CONFIG_SMP) &&
	    !IS_ENABLED(CONFIG_SCHED_IPI_SUPPORTED)) {
		z_sched_ipi();
	}

	z_check_stack_sentinel();

	return _current->switch_handle;
}
#endif

ALWAYS_INLINE void z_priq_dumb_add(sys_dlist_t *pq, struct k_thread *thread)
{
	struct k_thread *t;

	__ASSERT_NO_MSG(!is_idle(thread));

	SYS_DLIST_FOR_EACH_CONTAINER(pq, t, base.qnode_dlist) {
		if (z_is_t1_higher_prio_than_t2(thread, t)) {
			sys_dlist_insert(&t->base.qnode_dlist,
					 &thread->base.qnode_dlist);
			return;
		}
	}

	sys_dlist_append(pq, &thread->base.qnode_dlist);
}

void z_priq_dumb_remove(sys_dlist_t *pq, struct k_thread *thread)
{
#if defined(CONFIG_SWAP_NONATOMIC) && defined(CONFIG_SCHED_DUMB)
	if (pq == &_kernel.ready_q.runq && thread == _current &&
	    z_is_thread_prevented_from_running(thread)) {
		return;
	}
#endif

	__ASSERT_NO_MSG(!is_idle(thread));

	sys_dlist_remove(&thread->base.qnode_dlist);
}

struct k_thread *z_priq_dumb_best(sys_dlist_t *pq)
{
	struct k_thread *t = NULL;
	sys_dnode_t *n = sys_dlist_peek_head(pq);

	if (n != NULL) {
		t = CONTAINER_OF(n, struct k_thread, base.qnode_dlist);
	}
	return t;
}

bool z_priq_rb_lessthan(struct rbnode *a, struct rbnode *b)
{
	struct k_thread *ta, *tb;

	ta = CONTAINER_OF(a, struct k_thread, base.qnode_rb);
	tb = CONTAINER_OF(b, struct k_thread, base.qnode_rb);

	if (z_is_t1_higher_prio_than_t2(ta, tb)) {
		return true;
	} else if (z_is_t1_higher_prio_than_t2(tb, ta)) {
		return false;
	} else {
		return ta->base.order_key < tb->base.order_key ? 1 : 0;
	}
}

void z_priq_rb_add(struct _priq_rb *pq, struct k_thread *thread)
{
	struct k_thread *t;

	__ASSERT_NO_MSG(!is_idle(thread));

	thread->base.order_key = pq->next_order_key++;

	/* Renumber at wraparound.  This is tiny code, and in practice
	 * will almost never be hit on real systems.  BUT on very
	 * long-running systems where a priq never completely empties
	 * AND that contains very large numbers of threads, it can be
	 * a latency glitch to loop over all the threads like this.
	 */
	if (!pq->next_order_key) {
		RB_FOR_EACH_CONTAINER(&pq->tree, t, base.qnode_rb) {
			t->base.order_key = pq->next_order_key++;
		}
	}

	rb_insert(&pq->tree, &thread->base.qnode_rb);
}

void z_priq_rb_remove(struct _priq_rb *pq, struct k_thread *thread)
{
#if defined(CONFIG_SWAP_NONATOMIC) && defined(CONFIG_SCHED_SCALABLE)
	if (pq == &_kernel.ready_q.runq && thread == _current &&
	    z_is_thread_prevented_from_running(thread)) {
		return;
	}
#endif
	__ASSERT_NO_MSG(!is_idle(thread));

	rb_remove(&pq->tree, &thread->base.qnode_rb);

	if (!pq->tree.root) {
		pq->next_order_key = 0;
	}
}

struct k_thread *z_priq_rb_best(struct _priq_rb *pq)
{
	struct k_thread *t = NULL;
	struct rbnode *n = rb_get_min(&pq->tree);

	if (n != NULL) {
		t = CONTAINER_OF(n, struct k_thread, base.qnode_rb);
	}
	return t;
}

#ifdef CONFIG_SCHED_MULTIQ
# if (K_LOWEST_THREAD_PRIO - K_HIGHEST_THREAD_PRIO) > 31
# error Too many priorities for multiqueue scheduler (max 32)
# endif
#endif

ALWAYS_INLINE void z_priq_mq_add(struct _priq_mq *pq, struct k_thread *thread)
{
	int priority_bit = thread->base.prio - K_HIGHEST_THREAD_PRIO;

	sys_dlist_append(&pq->queues[priority_bit], &thread->base.qnode_dlist);
	pq->bitmask |= (1 << priority_bit);
}

ALWAYS_INLINE void z_priq_mq_remove(struct _priq_mq *pq, struct k_thread *thread)
{
#if defined(CONFIG_SWAP_NONATOMIC) && defined(CONFIG_SCHED_MULTIQ)
	if (pq == &_kernel.ready_q.runq && thread == _current &&
	    z_is_thread_prevented_from_running(thread)) {
		return;
	}
#endif
	int priority_bit = thread->base.prio - K_HIGHEST_THREAD_PRIO;

	sys_dlist_remove(&thread->base.qnode_dlist);
	if (sys_dlist_is_empty(&pq->queues[priority_bit])) {
		pq->bitmask &= ~(1 << priority_bit);
	}
}

struct k_thread *z_priq_mq_best(struct _priq_mq *pq)
{
	if (!pq->bitmask) {
		return NULL;
	}

	struct k_thread *t = NULL;
	sys_dlist_t *l = &pq->queues[__builtin_ctz(pq->bitmask)];
	sys_dnode_t *n = sys_dlist_peek_head(l);

	if (n != NULL) {
		t = CONTAINER_OF(n, struct k_thread, base.qnode_dlist);
	}
	return t;
}

int z_unpend_all(_wait_q_t *wait_q)
{
	int need_sched = 0;
	struct k_thread *th;

	while ((th = z_waitq_head(wait_q)) != NULL) {
		z_unpend_thread(th);
		z_ready_thread(th);
		need_sched = 1;
	}

	return need_sched;
}

void z_sched_init(void)
{
#ifdef CONFIG_SCHED_DUMB
	sys_dlist_init(&_kernel.ready_q.runq);
#endif

#ifdef CONFIG_SCHED_SCALABLE
	_kernel.ready_q.runq = (struct _priq_rb) {
		.tree = {
			.lessthan_fn = z_priq_rb_lessthan,
		}
	};
#endif

#ifdef CONFIG_SCHED_MULTIQ
	for (int i = 0; i < ARRAY_SIZE(_kernel.ready_q.runq.queues); i++) {
		sys_dlist_init(&_kernel.ready_q.runq.queues[i]);
	}
#endif

#ifdef CONFIG_TIMESLICING
	k_sched_time_slice_set(CONFIG_TIMESLICE_SIZE,
		CONFIG_TIMESLICE_PRIORITY);
#endif
}

int z_impl_k_thread_priority_get(k_tid_t thread)
{
	return thread->base.prio;
}

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER1_SIMPLE(k_thread_priority_get, K_OBJ_THREAD,
			  struct k_thread *);
#endif

void z_impl_k_thread_priority_set(k_tid_t tid, int prio)
{
	/*
	 * Use NULL, since we cannot know what the entry point is (we do not
	 * keep track of it) and idle cannot change its priority.
	 */
	Z_ASSERT_VALID_PRIO(prio, NULL);
	__ASSERT(!z_is_in_isr(), "");

	struct k_thread *thread = (struct k_thread *)tid;

	z_thread_priority_set(thread, prio);
}

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER(k_thread_priority_set, thread_p, prio)
{
	struct k_thread *thread = (struct k_thread *)thread_p;

	Z_OOPS(Z_SYSCALL_OBJ(thread, K_OBJ_THREAD));
	Z_OOPS(Z_SYSCALL_VERIFY_MSG(_is_valid_prio(prio, NULL),
				    "invalid thread priority %d", (int)prio));
	Z_OOPS(Z_SYSCALL_VERIFY_MSG((s8_t)prio >= thread->base.prio,
				    "thread priority may only be downgraded (%d < %d)",
				    prio, thread->base.prio));

	z_impl_k_thread_priority_set((k_tid_t)thread, prio);
	return 0;
}
#endif

#ifdef CONFIG_SCHED_DEADLINE
void z_impl_k_thread_deadline_set(k_tid_t tid, int deadline)
{
	struct k_thread *th = tid;

	LOCKED(&sched_spinlock) {
		th->base.prio_deadline = k_cycle_get_32() + deadline;
		if (z_is_thread_queued(th)) {
			_priq_run_remove(&_kernel.ready_q.runq, th);
			_priq_run_add(&_kernel.ready_q.runq, th);
		}
	}
}

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER(k_thread_deadline_set, thread_p, deadline)
{
	struct k_thread *thread = (struct k_thread *)thread_p;

	Z_OOPS(Z_SYSCALL_OBJ(thread, K_OBJ_THREAD));
	Z_OOPS(Z_SYSCALL_VERIFY_MSG(deadline > 0,
				    "invalid thread deadline %d",
				    (int)deadline));

	z_impl_k_thread_deadline_set((k_tid_t)thread, deadline);
	return 0;
}
#endif
#endif

void z_impl_k_yield(void)
{
	__ASSERT(!z_is_in_isr(), "");

	if (!is_idle(_current)) {
		LOCKED(&sched_spinlock) {
			if (!IS_ENABLED(CONFIG_SMP) ||
			    z_is_thread_queued(_current)) {
				_priq_run_remove(&_kernel.ready_q.runq,
						 _current);
				_priq_run_add(&_kernel.ready_q.runq,
					      _current);
			}
			update_cache(1);
		}
	}
	z_swap_unlocked();
}

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER0_SIMPLE_VOID(k_yield);
#endif

s32_t z_impl_k_sleep(s32_t duration)
{
#ifdef CONFIG_MULTITHREADING
	u32_t expected_wakeup_time;
	s32_t ticks;

	__ASSERT(!z_is_in_isr(), "");
	__ASSERT(duration != K_FOREVER, "");

	K_DEBUG("thread %p for %d ns\n", _current, duration);

	/* wait of 0 ms is treated as a 'yield' */
	if (duration == 0) {
		k_yield();
		return 0;
	}

	ticks = _TICK_ALIGN + z_ms_to_ticks(duration);
	expected_wakeup_time = ticks + z_tick_get_32();

	/* Spinlock purely for local interrupt locking to prevent us
	 * from being interrupted while _current is in an intermediate
	 * state.  Should unify this implementation with pend().
	 */
	struct k_spinlock local_lock = {};
	k_spinlock_key_t key = k_spin_lock(&local_lock);

#if defined(CONFIG_TIMESLICING) && defined(CONFIG_SWAP_NONATOMIC)
	pending_current = _current;
#endif
	z_remove_thread_from_ready_q(_current);
	z_add_thread_timeout(_current, ticks);

	(void)z_swap(&local_lock, key);

	ticks = expected_wakeup_time - z_tick_get_32();
	if (ticks > 0) {
		return __ticks_to_ms(ticks);
	}
#endif

	return 0;
}

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER(k_sleep, duration)
{
	/* FIXME there were some discussions recently on whether we should
	 * relax this, thread would be unscheduled until k_wakeup issued
	 */
	Z_OOPS(Z_SYSCALL_VERIFY_MSG(duration != K_FOREVER,
				    "sleeping forever not allowed"));

	return z_impl_k_sleep(duration);
}
#endif

void z_impl_k_wakeup(k_tid_t thread)
{
	if (z_is_thread_pending(thread)) {
		return;
	}

	if (z_abort_thread_timeout(thread) < 0) {
		return;
	}

	z_ready_thread(thread);

	if (!z_is_in_isr()) {
		z_reschedule_unlocked();
	}

	if (IS_ENABLED(CONFIG_SMP) &&
	    !IS_ENABLED(CONFIG_SCHED_IPI_SUPPORTED)) {
		z_sched_ipi();
	}
}

#ifdef CONFIG_SMP
/* Called out of the scheduler interprocessor interrupt.  All it does
 * is flag the current thread as dead if it needs to abort, so the ISR
 * return into something else and the other thread which called
 * k_thread_abort() can finish its work knowing the thing won't be
 * rescheduled.
 */
void z_sched_ipi(void)
{
	LOCKED(&sched_spinlock) {
		if (_current->base.thread_state & _THREAD_ABORTING) {
			_current->base.thread_state |= _THREAD_DEAD;
			_current_cpu->swap_ok = true;
		}
	}
}

void z_sched_abort(struct k_thread *thread)
{
	if (thread == _current) {
		z_remove_thread_from_ready_q(thread);
		return;
	}

	/* First broadcast an IPI to the other CPUs so they can stop
	 * it locally.  Not all architectures support that, alas.  If
	 * we don't have it, we need to wait for some other interrupt.
	 */
	thread->base.thread_state |= _THREAD_ABORTING;
#ifdef CONFIG_SCHED_IPI_SUPPORTED
	z_arch_sched_ipi();
#endif

	/* Wait for it to be flagged dead either by the CPU it was
	 * running on or because we caught it idle in the queue
	 */
	while ((thread->base.thread_state & _THREAD_DEAD) == 0) {
		LOCKED(&sched_spinlock) {
			if (z_is_thread_queued(thread)) {
				_current->base.thread_state |= _THREAD_DEAD;
				_priq_run_remove(&_kernel.ready_q.runq, thread);
				z_mark_thread_as_not_queued(thread);
			}
		}
	}
}
#endif

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER1_SIMPLE_VOID(k_wakeup, K_OBJ_THREAD, k_tid_t);
#endif

k_tid_t z_impl_k_current_get(void)
{
	return _current;
}

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER0_SIMPLE(k_current_get);
#endif

int z_impl_k_is_preempt_thread(void)
{
	return !z_is_in_isr() && is_preempt(_current);
}

#ifdef CONFIG_USERSPACE
Z_SYSCALL_HANDLER0_SIMPLE(k_is_preempt_thread);
#endif

#ifdef CONFIG_SCHED_CPU_MASK
# ifdef CONFIG_SMP
/* Right now we use a single byte for this mask */
BUILD_ASSERT_MSG(CONFIG_MP_NUM_CPUS <= 8, "Too many CPUs for mask word");
# endif


static int cpu_mask_mod(k_tid_t t, u32_t enable_mask, u32_t disable_mask)
{
	int ret = 0;

	LOCKED(&sched_spinlock) {
		if (z_is_thread_prevented_from_running(t)) {
			t->base.cpu_mask |= enable_mask;
			t->base.cpu_mask  &= ~disable_mask;
		} else {
			ret = -EINVAL;
		}
	}
	return ret;
}

int k_thread_cpu_mask_clear(k_tid_t thread)
{
	return cpu_mask_mod(thread, 0, 0xffffffff);
}

int k_thread_cpu_mask_enable_all(k_tid_t thread)
{
	return cpu_mask_mod(thread, 0xffffffff, 0);
}

int k_thread_cpu_mask_enable(k_tid_t thread, int cpu)
{
	return cpu_mask_mod(thread, BIT(cpu), 0);
}

int k_thread_cpu_mask_disable(k_tid_t thread, int cpu)
{
	return cpu_mask_mod(thread, 0, BIT(cpu));
}

#endif /* CONFIG_SCHED_CPU_MASK */