aboutsummaryrefslogtreecommitdiff
path: root/sys/kern/kern_rmlock.c
blob: 123b24845f8e058e82855cf98ca6c239e4c8f155 (plain) (blame)
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
/*-
 * SPDX-License-Identifier: BSD-3-Clause
 *
 * Copyright (c) 2007 Stephan Uphoff <ups@FreeBSD.org>
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. Neither the name of the author nor the names of any co-contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

/*
 * Machine independent bits of reader/writer lock implementation.
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

#include "opt_ddb.h"

#include <sys/param.h>
#include <sys/systm.h>

#include <sys/kernel.h>
#include <sys/kdb.h>
#include <sys/ktr.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/rmlock.h>
#include <sys/sched.h>
#include <sys/smp.h>
#include <sys/turnstile.h>
#include <sys/lock_profile.h>
#include <machine/cpu.h>

#ifdef DDB
#include <ddb/ddb.h>
#endif

/*
 * A cookie to mark destroyed rmlocks.  This is stored in the head of
 * rm_activeReaders.
 */
#define	RM_DESTROYED	((void *)0xdead)

#define	rm_destroyed(rm)						\
	(LIST_FIRST(&(rm)->rm_activeReaders) == RM_DESTROYED)

#define RMPF_ONQUEUE	1
#define RMPF_SIGNAL	2

#ifndef INVARIANTS
#define	_rm_assert(c, what, file, line)
#endif

static void	assert_rm(const struct lock_object *lock, int what);
#ifdef DDB
static void	db_show_rm(const struct lock_object *lock);
#endif
static void	lock_rm(struct lock_object *lock, uintptr_t how);
#ifdef KDTRACE_HOOKS
static int	owner_rm(const struct lock_object *lock, struct thread **owner);
#endif
static uintptr_t unlock_rm(struct lock_object *lock);

struct lock_class lock_class_rm = {
	.lc_name = "rm",
	.lc_flags = LC_SLEEPLOCK | LC_RECURSABLE,
	.lc_assert = assert_rm,
#ifdef DDB
	.lc_ddb_show = db_show_rm,
#endif
	.lc_lock = lock_rm,
	.lc_unlock = unlock_rm,
#ifdef KDTRACE_HOOKS
	.lc_owner = owner_rm,
#endif
};

struct lock_class lock_class_rm_sleepable = {
	.lc_name = "sleepable rm",
	.lc_flags = LC_SLEEPLOCK | LC_SLEEPABLE | LC_RECURSABLE,
	.lc_assert = assert_rm,
#ifdef DDB
	.lc_ddb_show = db_show_rm,
#endif
	.lc_lock = lock_rm,
	.lc_unlock = unlock_rm,
#ifdef KDTRACE_HOOKS
	.lc_owner = owner_rm,
#endif
};

static void
assert_rm(const struct lock_object *lock, int what)
{

	rm_assert((const struct rmlock *)lock, what);
}

static void
lock_rm(struct lock_object *lock, uintptr_t how)
{
	struct rmlock *rm;
	struct rm_priotracker *tracker;

	rm = (struct rmlock *)lock;
	if (how == 0)
		rm_wlock(rm);
	else {
		tracker = (struct rm_priotracker *)how;
		rm_rlock(rm, tracker);
	}
}

static uintptr_t
unlock_rm(struct lock_object *lock)
{
	struct thread *td;
	struct pcpu *pc;
	struct rmlock *rm;
	struct rm_queue *queue;
	struct rm_priotracker *tracker;
	uintptr_t how;

	rm = (struct rmlock *)lock;
	tracker = NULL;
	how = 0;
	rm_assert(rm, RA_LOCKED | RA_NOTRECURSED);
	if (rm_wowned(rm))
		rm_wunlock(rm);
	else {
		/*
		 * Find the right rm_priotracker structure for curthread.
		 * The guarantee about its uniqueness is given by the fact
		 * we already asserted the lock wasn't recursively acquired.
		 */
		critical_enter();
		td = curthread;
		pc = get_pcpu();
		for (queue = pc->pc_rm_queue.rmq_next;
		    queue != &pc->pc_rm_queue; queue = queue->rmq_next) {
			tracker = (struct rm_priotracker *)queue;
				if ((tracker->rmp_rmlock == rm) &&
				    (tracker->rmp_thread == td)) {
					how = (uintptr_t)tracker;
					break;
				}
		}
		KASSERT(tracker != NULL,
		    ("rm_priotracker is non-NULL when lock held in read mode"));
		critical_exit();
		rm_runlock(rm, tracker);
	}
	return (how);
}

#ifdef KDTRACE_HOOKS
static int
owner_rm(const struct lock_object *lock, struct thread **owner)
{
	const struct rmlock *rm;
	struct lock_class *lc;

	rm = (const struct rmlock *)lock;
	lc = LOCK_CLASS(&rm->rm_wlock_object);
	return (lc->lc_owner(&rm->rm_wlock_object, owner));
}
#endif

static struct mtx rm_spinlock;

MTX_SYSINIT(rm_spinlock, &rm_spinlock, "rm_spinlock", MTX_SPIN);

/*
 * Add or remove tracker from per-cpu list.
 *
 * The per-cpu list can be traversed at any time in forward direction from an
 * interrupt on the *local* cpu.
 */
static void inline
rm_tracker_add(struct pcpu *pc, struct rm_priotracker *tracker)
{
	struct rm_queue *next;

	/* Initialize all tracker pointers */
	tracker->rmp_cpuQueue.rmq_prev = &pc->pc_rm_queue;
	next = pc->pc_rm_queue.rmq_next;
	tracker->rmp_cpuQueue.rmq_next = next;

	/* rmq_prev is not used during froward traversal. */
	next->rmq_prev = &tracker->rmp_cpuQueue;

	/* Update pointer to first element. */
	pc->pc_rm_queue.rmq_next = &tracker->rmp_cpuQueue;
}

/*
 * Return a count of the number of trackers the thread 'td' already
 * has on this CPU for the lock 'rm'.
 */
static int
rm_trackers_present(const struct pcpu *pc, const struct rmlock *rm,
    const struct thread *td)
{
	struct rm_queue *queue;
	struct rm_priotracker *tracker;
	int count;

	count = 0;
	for (queue = pc->pc_rm_queue.rmq_next; queue != &pc->pc_rm_queue;
	    queue = queue->rmq_next) {
		tracker = (struct rm_priotracker *)queue;
		if ((tracker->rmp_rmlock == rm) && (tracker->rmp_thread == td))
			count++;
	}
	return (count);
}

static void inline
rm_tracker_remove(struct pcpu *pc, struct rm_priotracker *tracker)
{
	struct rm_queue *next, *prev;

	next = tracker->rmp_cpuQueue.rmq_next;
	prev = tracker->rmp_cpuQueue.rmq_prev;

	/* Not used during forward traversal. */
	next->rmq_prev = prev;

	/* Remove from list. */
	prev->rmq_next = next;
}

static void
rm_cleanIPI(void *arg)
{
	struct pcpu *pc;
	struct rmlock *rm = arg;
	struct rm_priotracker *tracker;
	struct rm_queue *queue;
	pc = get_pcpu();

	for (queue = pc->pc_rm_queue.rmq_next; queue != &pc->pc_rm_queue;
	    queue = queue->rmq_next) {
		tracker = (struct rm_priotracker *)queue;
		if (tracker->rmp_rmlock == rm && tracker->rmp_flags == 0) {
			tracker->rmp_flags = RMPF_ONQUEUE;
			mtx_lock_spin(&rm_spinlock);
			LIST_INSERT_HEAD(&rm->rm_activeReaders, tracker,
			    rmp_qentry);
			mtx_unlock_spin(&rm_spinlock);
		}
	}
}

void
rm_init_flags(struct rmlock *rm, const char *name, int opts)
{
	struct lock_class *lc;
	int liflags, xflags;

	liflags = 0;
	if (!(opts & RM_NOWITNESS))
		liflags |= LO_WITNESS;
	if (opts & RM_RECURSE)
		liflags |= LO_RECURSABLE;
	if (opts & RM_NEW)
		liflags |= LO_NEW;
	rm->rm_writecpus = all_cpus;
	LIST_INIT(&rm->rm_activeReaders);
	if (opts & RM_SLEEPABLE) {
		liflags |= LO_SLEEPABLE;
		lc = &lock_class_rm_sleepable;
		xflags = (opts & RM_NEW ? SX_NEW : 0);
		sx_init_flags(&rm->rm_lock_sx, "rmlock_sx",
		    xflags | SX_NOWITNESS);
	} else {
		lc = &lock_class_rm;
		xflags = (opts & RM_NEW ? MTX_NEW : 0);
		mtx_init(&rm->rm_lock_mtx, name, "rmlock_mtx",
		    xflags | MTX_NOWITNESS);
	}
	lock_init(&rm->lock_object, lc, name, NULL, liflags);
}

void
rm_init(struct rmlock *rm, const char *name)
{

	rm_init_flags(rm, name, 0);
}

void
rm_destroy(struct rmlock *rm)
{

	rm_assert(rm, RA_UNLOCKED);
	LIST_FIRST(&rm->rm_activeReaders) = RM_DESTROYED;
	if (rm->lock_object.lo_flags & LO_SLEEPABLE)
		sx_destroy(&rm->rm_lock_sx);
	else
		mtx_destroy(&rm->rm_lock_mtx);
	lock_destroy(&rm->lock_object);
}

int
rm_wowned(const struct rmlock *rm)
{

	if (rm->lock_object.lo_flags & LO_SLEEPABLE)
		return (sx_xlocked(&rm->rm_lock_sx));
	else
		return (mtx_owned(&rm->rm_lock_mtx));
}

void
rm_sysinit(void *arg)
{
	struct rm_args *args = arg;

	rm_init(args->ra_rm, args->ra_desc);
}

void
rm_sysinit_flags(void *arg)
{
	struct rm_args_flags *args = arg;

	rm_init_flags(args->ra_rm, args->ra_desc, args->ra_opts);
}

static int
_rm_rlock_hard(struct rmlock *rm, struct rm_priotracker *tracker, int trylock)
{
	struct pcpu *pc;

	critical_enter();
	pc = get_pcpu();

	/* Check if we just need to do a proper critical_exit. */
	if (!CPU_ISSET(pc->pc_cpuid, &rm->rm_writecpus)) {
		critical_exit();
		return (1);
	}

	/* Remove our tracker from the per-cpu list. */
	rm_tracker_remove(pc, tracker);

	/* Check to see if the IPI granted us the lock after all. */
	if (tracker->rmp_flags) {
		/* Just add back tracker - we hold the lock. */
		rm_tracker_add(pc, tracker);
		critical_exit();
		return (1);
	}

	/*
	 * We allow readers to acquire a lock even if a writer is blocked if
	 * the lock is recursive and the reader already holds the lock.
	 */
	if ((rm->lock_object.lo_flags & LO_RECURSABLE) != 0) {
		/*
		 * Just grant the lock if this thread already has a tracker
		 * for this lock on the per-cpu queue.
		 */
		if (rm_trackers_present(pc, rm, curthread) != 0) {
			mtx_lock_spin(&rm_spinlock);
			LIST_INSERT_HEAD(&rm->rm_activeReaders, tracker,
			    rmp_qentry);
			tracker->rmp_flags = RMPF_ONQUEUE;
			mtx_unlock_spin(&rm_spinlock);
			rm_tracker_add(pc, tracker);
			critical_exit();
			return (1);
		}
	}

	sched_unpin();
	critical_exit();

	if (trylock) {
		if (rm->lock_object.lo_flags & LO_SLEEPABLE) {
			if (!sx_try_xlock(&rm->rm_lock_sx))
				return (0);
		} else {
			if (!mtx_trylock(&rm->rm_lock_mtx))
				return (0);
		}
	} else {
		if (rm->lock_object.lo_flags & LO_SLEEPABLE) {
			THREAD_SLEEPING_OK();
			sx_xlock(&rm->rm_lock_sx);
			THREAD_NO_SLEEPING();
		} else
			mtx_lock(&rm->rm_lock_mtx);
	}

	critical_enter();
	pc = get_pcpu();
	CPU_CLR(pc->pc_cpuid, &rm->rm_writecpus);
	rm_tracker_add(pc, tracker);
	sched_pin();
	critical_exit();

	if (rm->lock_object.lo_flags & LO_SLEEPABLE)
		sx_xunlock(&rm->rm_lock_sx);
	else
		mtx_unlock(&rm->rm_lock_mtx);

	return (1);
}

int
_rm_rlock(struct rmlock *rm, struct rm_priotracker *tracker, int trylock)
{
	struct thread *td = curthread;
	struct pcpu *pc;

	if (SCHEDULER_STOPPED())
		return (1);

	tracker->rmp_flags  = 0;
	tracker->rmp_thread = td;
	tracker->rmp_rmlock = rm;

	if (rm->lock_object.lo_flags & LO_SLEEPABLE)
		THREAD_NO_SLEEPING();

	td->td_critnest++;	/* critical_enter(); */

	__compiler_membar();

	pc = cpuid_to_pcpu[td->td_oncpu]; /* pcpu_find(td->td_oncpu); */

	rm_tracker_add(pc, tracker);

	sched_pin();

	__compiler_membar();

	td->td_critnest--;

	/*
	 * Fast path to combine two common conditions into a single
	 * conditional jump.
	 */
	if (0 == (td->td_owepreempt |
	    CPU_ISSET(pc->pc_cpuid, &rm->rm_writecpus)))
		return (1);

	/* We do not have a read token and need to acquire one. */
	return _rm_rlock_hard(rm, tracker, trylock);
}

static void
_rm_unlock_hard(struct thread *td,struct rm_priotracker *tracker)
{

	if (td->td_owepreempt) {
		td->td_critnest++;
		critical_exit();
	}

	if (!tracker->rmp_flags)
		return;

	mtx_lock_spin(&rm_spinlock);
	LIST_REMOVE(tracker, rmp_qentry);

	if (tracker->rmp_flags & RMPF_SIGNAL) {
		struct rmlock *rm;
		struct turnstile *ts;

		rm = tracker->rmp_rmlock;

		turnstile_chain_lock(&rm->lock_object);
		mtx_unlock_spin(&rm_spinlock);

		ts = turnstile_lookup(&rm->lock_object);

		turnstile_signal(ts, TS_EXCLUSIVE_QUEUE);
		turnstile_unpend(ts, TS_EXCLUSIVE_LOCK);
		turnstile_chain_unlock(&rm->lock_object);
	} else
		mtx_unlock_spin(&rm_spinlock);
}

void
_rm_runlock(struct rmlock *rm, struct rm_priotracker *tracker)
{
	struct pcpu *pc;
	struct thread *td = tracker->rmp_thread;

	if (SCHEDULER_STOPPED())
		return;

	td->td_critnest++;	/* critical_enter(); */
	pc = cpuid_to_pcpu[td->td_oncpu]; /* pcpu_find(td->td_oncpu); */
	rm_tracker_remove(pc, tracker);
	td->td_critnest--;
	sched_unpin();

	if (rm->lock_object.lo_flags & LO_SLEEPABLE)
		THREAD_SLEEPING_OK();

	if (0 == (td->td_owepreempt | tracker->rmp_flags))
		return;

	_rm_unlock_hard(td, tracker);
}

void
_rm_wlock(struct rmlock *rm)
{
	struct rm_priotracker *prio;
	struct turnstile *ts;
	cpuset_t readcpus;

	if (SCHEDULER_STOPPED())
		return;

	if (rm->lock_object.lo_flags & LO_SLEEPABLE)
		sx_xlock(&rm->rm_lock_sx);
	else
		mtx_lock(&rm->rm_lock_mtx);

	if (CPU_CMP(&rm->rm_writecpus, &all_cpus)) {
		/* Get all read tokens back */
		readcpus = all_cpus;
		CPU_NAND(&readcpus, &rm->rm_writecpus);
		rm->rm_writecpus = all_cpus;

		/*
		 * Assumes rm->rm_writecpus update is visible on other CPUs
		 * before rm_cleanIPI is called.
		 */
#ifdef SMP
		smp_rendezvous_cpus(readcpus,
		    smp_no_rendezvous_barrier,
		    rm_cleanIPI,
		    smp_no_rendezvous_barrier,
		    rm);

#else
		rm_cleanIPI(rm);
#endif

		mtx_lock_spin(&rm_spinlock);
		while ((prio = LIST_FIRST(&rm->rm_activeReaders)) != NULL) {
			ts = turnstile_trywait(&rm->lock_object);
			prio->rmp_flags = RMPF_ONQUEUE | RMPF_SIGNAL;
			mtx_unlock_spin(&rm_spinlock);
			turnstile_wait(ts, prio->rmp_thread,
			    TS_EXCLUSIVE_QUEUE);
			mtx_lock_spin(&rm_spinlock);
		}
		mtx_unlock_spin(&rm_spinlock);
	}
}

void
_rm_wunlock(struct rmlock *rm)
{

	if (rm->lock_object.lo_flags & LO_SLEEPABLE)
		sx_xunlock(&rm->rm_lock_sx);
	else
		mtx_unlock(&rm->rm_lock_mtx);
}

#if LOCK_DEBUG > 0

void
_rm_wlock_debug(struct rmlock *rm, const char *file, int line)
{

	if (SCHEDULER_STOPPED())
		return;

	KASSERT(kdb_active != 0 || !TD_IS_IDLETHREAD(curthread),
	    ("rm_wlock() by idle thread %p on rmlock %s @ %s:%d",
	    curthread, rm->lock_object.lo_name, file, line));
	KASSERT(!rm_destroyed(rm),
	    ("rm_wlock() of destroyed rmlock @ %s:%d", file, line));
	_rm_assert(rm, RA_UNLOCKED, file, line);

	WITNESS_CHECKORDER(&rm->lock_object, LOP_NEWORDER | LOP_EXCLUSIVE,
	    file, line, NULL);

	_rm_wlock(rm);

	LOCK_LOG_LOCK("RMWLOCK", &rm->lock_object, 0, 0, file, line);
	WITNESS_LOCK(&rm->lock_object, LOP_EXCLUSIVE, file, line);
	TD_LOCKS_INC(curthread);
}

void
_rm_wunlock_debug(struct rmlock *rm, const char *file, int line)
{

	if (SCHEDULER_STOPPED())
		return;

	KASSERT(!rm_destroyed(rm),
	    ("rm_wunlock() of destroyed rmlock @ %s:%d", file, line));
	_rm_assert(rm, RA_WLOCKED, file, line);
	WITNESS_UNLOCK(&rm->lock_object, LOP_EXCLUSIVE, file, line);
	LOCK_LOG_LOCK("RMWUNLOCK", &rm->lock_object, 0, 0, file, line);
	_rm_wunlock(rm);
	TD_LOCKS_DEC(curthread);
}

int
_rm_rlock_debug(struct rmlock *rm, struct rm_priotracker *tracker,
    int trylock, const char *file, int line)
{

	if (SCHEDULER_STOPPED())
		return (1);

#ifdef INVARIANTS
	if (!(rm->lock_object.lo_flags & LO_RECURSABLE) && !trylock) {
		critical_enter();
		KASSERT(rm_trackers_present(get_pcpu(), rm,
		    curthread) == 0,
		    ("rm_rlock: recursed on non-recursive rmlock %s @ %s:%d\n",
		    rm->lock_object.lo_name, file, line));
		critical_exit();
	}
#endif
	KASSERT(kdb_active != 0 || !TD_IS_IDLETHREAD(curthread),
	    ("rm_rlock() by idle thread %p on rmlock %s @ %s:%d",
	    curthread, rm->lock_object.lo_name, file, line));
	KASSERT(!rm_destroyed(rm),
	    ("rm_rlock() of destroyed rmlock @ %s:%d", file, line));
	if (!trylock) {
		KASSERT(!rm_wowned(rm),
		    ("rm_rlock: wlock already held for %s @ %s:%d",
		    rm->lock_object.lo_name, file, line));
		WITNESS_CHECKORDER(&rm->lock_object, LOP_NEWORDER, file, line,
		    NULL);
	}

	if (_rm_rlock(rm, tracker, trylock)) {
		if (trylock)
			LOCK_LOG_TRY("RMRLOCK", &rm->lock_object, 0, 1, file,
			    line);
		else
			LOCK_LOG_LOCK("RMRLOCK", &rm->lock_object, 0, 0, file,
			    line);
		WITNESS_LOCK(&rm->lock_object, 0, file, line);
		TD_LOCKS_INC(curthread);
		return (1);
	} else if (trylock)
		LOCK_LOG_TRY("RMRLOCK", &rm->lock_object, 0, 0, file, line);

	return (0);
}

void
_rm_runlock_debug(struct rmlock *rm, struct rm_priotracker *tracker,
    const char *file, int line)
{

	if (SCHEDULER_STOPPED())
		return;

	KASSERT(!rm_destroyed(rm),
	    ("rm_runlock() of destroyed rmlock @ %s:%d", file, line));
	_rm_assert(rm, RA_RLOCKED, file, line);
	WITNESS_UNLOCK(&rm->lock_object, 0, file, line);
	LOCK_LOG_LOCK("RMRUNLOCK", &rm->lock_object, 0, 0, file, line);
	_rm_runlock(rm, tracker);
	TD_LOCKS_DEC(curthread);
}

#else

/*
 * Just strip out file and line arguments if no lock debugging is enabled in
 * the kernel - we are called from a kernel module.
 */
void
_rm_wlock_debug(struct rmlock *rm, const char *file, int line)
{

	_rm_wlock(rm);
}

void
_rm_wunlock_debug(struct rmlock *rm, const char *file, int line)
{

	_rm_wunlock(rm);
}

int
_rm_rlock_debug(struct rmlock *rm, struct rm_priotracker *tracker,
    int trylock, const char *file, int line)
{

	return _rm_rlock(rm, tracker, trylock);
}

void
_rm_runlock_debug(struct rmlock *rm, struct rm_priotracker *tracker,
    const char *file, int line)
{

	_rm_runlock(rm, tracker);
}

#endif

#ifdef INVARIANT_SUPPORT
#ifndef INVARIANTS
#undef _rm_assert
#endif

/*
 * Note that this does not need to use witness_assert() for read lock
 * assertions since an exact count of read locks held by this thread
 * is computable.
 */
void
_rm_assert(const struct rmlock *rm, int what, const char *file, int line)
{
	int count;

	if (panicstr != NULL)
		return;
	switch (what) {
	case RA_LOCKED:
	case RA_LOCKED | RA_RECURSED:
	case RA_LOCKED | RA_NOTRECURSED:
	case RA_RLOCKED:
	case RA_RLOCKED | RA_RECURSED:
	case RA_RLOCKED | RA_NOTRECURSED:
		/*
		 * Handle the write-locked case.  Unlike other
		 * primitives, writers can never recurse.
		 */
		if (rm_wowned(rm)) {
			if (what & RA_RLOCKED)
				panic("Lock %s exclusively locked @ %s:%d\n",
				    rm->lock_object.lo_name, file, line);
			if (what & RA_RECURSED)
				panic("Lock %s not recursed @ %s:%d\n",
				    rm->lock_object.lo_name, file, line);
			break;
		}

		critical_enter();
		count = rm_trackers_present(get_pcpu(), rm, curthread);
		critical_exit();

		if (count == 0)
			panic("Lock %s not %slocked @ %s:%d\n",
			    rm->lock_object.lo_name, (what & RA_RLOCKED) ?
			    "read " : "", file, line);
		if (count > 1) {
			if (what & RA_NOTRECURSED)
				panic("Lock %s recursed @ %s:%d\n",
				    rm->lock_object.lo_name, file, line);
		} else if (what & RA_RECURSED)
			panic("Lock %s not recursed @ %s:%d\n",
			    rm->lock_object.lo_name, file, line);
		break;
	case RA_WLOCKED:
		if (!rm_wowned(rm))
			panic("Lock %s not exclusively locked @ %s:%d\n",
			    rm->lock_object.lo_name, file, line);
		break;
	case RA_UNLOCKED:
		if (rm_wowned(rm))
			panic("Lock %s exclusively locked @ %s:%d\n",
			    rm->lock_object.lo_name, file, line);

		critical_enter();
		count = rm_trackers_present(get_pcpu(), rm, curthread);
		critical_exit();

		if (count != 0)
			panic("Lock %s read locked @ %s:%d\n",
			    rm->lock_object.lo_name, file, line);
		break;
	default:
		panic("Unknown rm lock assertion: %d @ %s:%d", what, file,
		    line);
	}
}
#endif /* INVARIANT_SUPPORT */

#ifdef DDB
static void
print_tracker(struct rm_priotracker *tr)
{
	struct thread *td;

	td = tr->rmp_thread;
	db_printf("   thread %p (tid %d, pid %d, \"%s\") {", td, td->td_tid,
	    td->td_proc->p_pid, td->td_name);
	if (tr->rmp_flags & RMPF_ONQUEUE) {
		db_printf("ONQUEUE");
		if (tr->rmp_flags & RMPF_SIGNAL)
			db_printf(",SIGNAL");
	} else
		db_printf("0");
	db_printf("}\n");
}

static void
db_show_rm(const struct lock_object *lock)
{
	struct rm_priotracker *tr;
	struct rm_queue *queue;
	const struct rmlock *rm;
	struct lock_class *lc;
	struct pcpu *pc;

	rm = (const struct rmlock *)lock;
	db_printf(" writecpus: ");
	ddb_display_cpuset(__DEQUALIFY(const cpuset_t *, &rm->rm_writecpus));
	db_printf("\n");
	db_printf(" per-CPU readers:\n");
	STAILQ_FOREACH(pc, &cpuhead, pc_allcpu)
		for (queue = pc->pc_rm_queue.rmq_next;
		    queue != &pc->pc_rm_queue; queue = queue->rmq_next) {
			tr = (struct rm_priotracker *)queue;
			if (tr->rmp_rmlock == rm)
				print_tracker(tr);
		}
	db_printf(" active readers:\n");
	LIST_FOREACH(tr, &rm->rm_activeReaders, rmp_qentry)
		print_tracker(tr);
	lc = LOCK_CLASS(&rm->rm_wlock_object);
	db_printf("Backing write-lock (%s):\n", lc->lc_name);
	lc->lc_ddb_show(&rm->rm_wlock_object);
}
#endif