summaryrefslogtreecommitdiff
path: root/sys/kern/kern_resource.c
blob: 8796b4833d352b7379b087c95c26fc41c96136c1 (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
/*	$OpenBSD: kern_resource.c,v 1.90 2024/10/03 10:20:05 claudio Exp $	*/
/*	$NetBSD: kern_resource.c,v 1.38 1996/10/23 07:19:38 matthias Exp $	*/

/*-
 * Copyright (c) 1982, 1986, 1991, 1993
 *	The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * 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 University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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.
 *
 *	@(#)kern_resource.c	8.5 (Berkeley) 1/21/94
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/file.h>
#include <sys/resourcevar.h>
#include <sys/pool.h>
#include <sys/proc.h>
#include <sys/ktrace.h>
#include <sys/sched.h>
#include <sys/signalvar.h>

#include <sys/mount.h>
#include <sys/syscallargs.h>

#include <uvm/uvm_extern.h>
#include <uvm/uvm.h>

/* Resource usage check interval in msec */
#define RUCHECK_INTERVAL	1000

/* SIGXCPU interval in seconds of process runtime */
#define SIGXCPU_INTERVAL	5

struct plimit	*lim_copy(struct plimit *);
struct plimit	*lim_write_begin(void);
void		 lim_write_commit(struct plimit *);

void	tuagg_sumup(struct tusage *, const struct tusage *);

/*
 * Patchable maximum data and stack limits.
 */
rlim_t maxdmap = MAXDSIZ;
rlim_t maxsmap = MAXSSIZ;

/*
 * Serializes resource limit updates.
 * This lock has to be held together with ps_mtx when updating
 * the process' ps_limit.
 */
struct rwlock rlimit_lock = RWLOCK_INITIALIZER("rlimitlk");

/*
 * Resource controls and accounting.
 */

int
sys_getpriority(struct proc *curp, void *v, register_t *retval)
{
	struct sys_getpriority_args /* {
		syscallarg(int) which;
		syscallarg(id_t) who;
	} */ *uap = v;
	struct process *pr;
	int low = NZERO + PRIO_MAX + 1;

	switch (SCARG(uap, which)) {

	case PRIO_PROCESS:
		if (SCARG(uap, who) == 0)
			pr = curp->p_p;
		else
			pr = prfind(SCARG(uap, who));
		if (pr == NULL)
			break;
		if (pr->ps_nice < low)
			low = pr->ps_nice;
		break;

	case PRIO_PGRP: {
		struct pgrp *pg;

		if (SCARG(uap, who) == 0)
			pg = curp->p_p->ps_pgrp;
		else if ((pg = pgfind(SCARG(uap, who))) == NULL)
			break;
		LIST_FOREACH(pr, &pg->pg_members, ps_pglist)
			if (pr->ps_nice < low)
				low = pr->ps_nice;
		break;
	}

	case PRIO_USER:
		if (SCARG(uap, who) == 0)
			SCARG(uap, who) = curp->p_ucred->cr_uid;
		LIST_FOREACH(pr, &allprocess, ps_list)
			if (pr->ps_ucred->cr_uid == SCARG(uap, who) &&
			    pr->ps_nice < low)
				low = pr->ps_nice;
		break;

	default:
		return (EINVAL);
	}
	if (low == NZERO + PRIO_MAX + 1)
		return (ESRCH);
	*retval = low - NZERO;
	return (0);
}

int
sys_setpriority(struct proc *curp, void *v, register_t *retval)
{
	struct sys_setpriority_args /* {
		syscallarg(int) which;
		syscallarg(id_t) who;
		syscallarg(int) prio;
	} */ *uap = v;
	struct process *pr;
	int found = 0, error = 0;

	switch (SCARG(uap, which)) {

	case PRIO_PROCESS:
		if (SCARG(uap, who) == 0)
			pr = curp->p_p;
		else
			pr = prfind(SCARG(uap, who));
		if (pr == NULL)
			break;
		error = donice(curp, pr, SCARG(uap, prio));
		found = 1;
		break;

	case PRIO_PGRP: {
		struct pgrp *pg;
		 
		if (SCARG(uap, who) == 0)
			pg = curp->p_p->ps_pgrp;
		else if ((pg = pgfind(SCARG(uap, who))) == NULL)
			break;
		LIST_FOREACH(pr, &pg->pg_members, ps_pglist) {
			error = donice(curp, pr, SCARG(uap, prio));
			found = 1;
		}
		break;
	}

	case PRIO_USER:
		if (SCARG(uap, who) == 0)
			SCARG(uap, who) = curp->p_ucred->cr_uid;
		LIST_FOREACH(pr, &allprocess, ps_list)
			if (pr->ps_ucred->cr_uid == SCARG(uap, who)) {
				error = donice(curp, pr, SCARG(uap, prio));
				found = 1;
			}
		break;

	default:
		return (EINVAL);
	}
	if (!found)
		return (ESRCH);
	return (error);
}

int
donice(struct proc *curp, struct process *chgpr, int n)
{
	struct ucred *ucred = curp->p_ucred;
	struct proc *p;

	if (ucred->cr_uid != 0 && ucred->cr_ruid != 0 &&
	    ucred->cr_uid != chgpr->ps_ucred->cr_uid &&
	    ucred->cr_ruid != chgpr->ps_ucred->cr_uid)
		return (EPERM);
	if (n > PRIO_MAX)
		n = PRIO_MAX;
	if (n < PRIO_MIN)
		n = PRIO_MIN;
	n += NZERO;
	if (n < chgpr->ps_nice && suser(curp))
		return (EACCES);
	chgpr->ps_nice = n;
	mtx_enter(&chgpr->ps_mtx);
	SCHED_LOCK();
	TAILQ_FOREACH(p, &chgpr->ps_threads, p_thr_link) {
		setpriority(p, p->p_estcpu, n);
	}
	SCHED_UNLOCK();
	mtx_leave(&chgpr->ps_mtx);
	return (0);
}

int
sys_setrlimit(struct proc *p, void *v, register_t *retval)
{
	struct sys_setrlimit_args /* {
		syscallarg(int) which;
		syscallarg(const struct rlimit *) rlp;
	} */ *uap = v;
	struct rlimit alim;
	int error;

	error = copyin((caddr_t)SCARG(uap, rlp), (caddr_t)&alim,
		       sizeof (struct rlimit));
	if (error)
		return (error);
#ifdef KTRACE
	if (KTRPOINT(p, KTR_STRUCT))
		ktrrlimit(p, &alim);
#endif
	return (dosetrlimit(p, SCARG(uap, which), &alim));
}

int
dosetrlimit(struct proc *p, u_int which, struct rlimit *limp)
{
	struct rlimit *alimp;
	struct plimit *limit;
	rlim_t maxlim;
	int error;

	if (which >= RLIM_NLIMITS || limp->rlim_cur > limp->rlim_max)
		return (EINVAL);

	rw_enter_write(&rlimit_lock);

	alimp = &p->p_p->ps_limit->pl_rlimit[which];
	if (limp->rlim_max > alimp->rlim_max) {
		if ((error = suser(p)) != 0) {
			rw_exit_write(&rlimit_lock);
			return (error);
		}
	}

	/* Get exclusive write access to the limit structure. */
	limit = lim_write_begin();
	alimp = &limit->pl_rlimit[which];

	switch (which) {
	case RLIMIT_DATA:
		maxlim = maxdmap;
		break;
	case RLIMIT_STACK:
		maxlim = maxsmap;
		break;
	case RLIMIT_NOFILE:
		maxlim = atomic_load_int(&maxfiles);
		break;
	case RLIMIT_NPROC:
		maxlim = atomic_load_int(&maxprocess);
		break;
	default:
		maxlim = RLIM_INFINITY;
		break;
	}

	if (limp->rlim_max > maxlim)
		limp->rlim_max = maxlim;
	if (limp->rlim_cur > limp->rlim_max)
		limp->rlim_cur = limp->rlim_max;

	if (which == RLIMIT_CPU && limp->rlim_cur != RLIM_INFINITY &&
	    alimp->rlim_cur == RLIM_INFINITY)
		timeout_add_msec(&p->p_p->ps_rucheck_to, RUCHECK_INTERVAL);

	if (which == RLIMIT_STACK) {
		/*
		 * Stack is allocated to the max at exec time with only
		 * "rlim_cur" bytes accessible.  If stack limit is going
		 * up make more accessible, if going down make inaccessible.
		 */
		if (limp->rlim_cur != alimp->rlim_cur) {
			vaddr_t addr;
			vsize_t size;
			vm_prot_t prot;
			struct vmspace *vm = p->p_vmspace;

			if (limp->rlim_cur > alimp->rlim_cur) {
				prot = PROT_READ | PROT_WRITE;
				size = limp->rlim_cur - alimp->rlim_cur;
#ifdef MACHINE_STACK_GROWS_UP
				addr = (vaddr_t)vm->vm_maxsaddr +
				    alimp->rlim_cur;
#else
				addr = (vaddr_t)vm->vm_minsaddr -
				    limp->rlim_cur;
#endif
			} else {
				prot = PROT_NONE;
				size = alimp->rlim_cur - limp->rlim_cur;
#ifdef MACHINE_STACK_GROWS_UP
				addr = (vaddr_t)vm->vm_maxsaddr +
				    limp->rlim_cur;
#else
				addr = (vaddr_t)vm->vm_minsaddr -
				    alimp->rlim_cur;
#endif
			}
			addr = trunc_page(addr);
			size = round_page(size);
			KERNEL_LOCK();
			(void) uvm_map_protect(&vm->vm_map, addr,
			    addr+size, prot, UVM_ET_STACK, FALSE, FALSE);
			KERNEL_UNLOCK();
		}
	}

	*alimp = *limp;

	lim_write_commit(limit);
	rw_exit_write(&rlimit_lock);

	return (0);
}

int
sys_getrlimit(struct proc *p, void *v, register_t *retval)
{
	struct sys_getrlimit_args /* {
		syscallarg(int) which;
		syscallarg(struct rlimit *) rlp;
	} */ *uap = v;
	struct plimit *limit;
	struct rlimit alimp;
	int error;

	if (SCARG(uap, which) < 0 || SCARG(uap, which) >= RLIM_NLIMITS)
		return (EINVAL);
	limit = lim_read_enter();
	alimp = limit->pl_rlimit[SCARG(uap, which)];
	lim_read_leave(limit);
	error = copyout(&alimp, SCARG(uap, rlp), sizeof(struct rlimit));
#ifdef KTRACE
	if (error == 0 && KTRPOINT(p, KTR_STRUCT))
		ktrrlimit(p, &alimp);
#endif
	return (error);
}

/* Add the counts from *from to *tu, ensuring a consistent read of *from. */ 
void
tuagg_sumup(struct tusage *tu, const struct tusage *from)
{
	struct tusage	tmp;
	uint64_t	enter, leave;

	enter = from->tu_gen;
	for (;;) {
		/* the generation number is odd during an update */
		while (enter & 1) {
			CPU_BUSY_CYCLE();
			enter = from->tu_gen;
		}

		membar_consumer();
		tmp = *from;
		membar_consumer();
		leave = from->tu_gen;

		if (enter == leave)
			break;
		enter = leave;
	}

	tu->tu_uticks += tmp.tu_uticks;
	tu->tu_sticks += tmp.tu_sticks;
	tu->tu_iticks += tmp.tu_iticks;
	timespecadd(&tu->tu_runtime, &tmp.tu_runtime, &tu->tu_runtime);
}

void
tuagg_get_proc(struct tusage *tu, struct proc *p)
{
	memset(tu, 0, sizeof(*tu));
	tuagg_sumup(tu, &p->p_tu);
}

void
tuagg_get_process(struct tusage *tu, struct process *pr)
{
	struct proc *q;

	memset(tu, 0, sizeof(*tu));

	mtx_enter(&pr->ps_mtx);
	tuagg_sumup(tu, &pr->ps_tu);
	/* add on all living threads */
	TAILQ_FOREACH(q, &pr->ps_threads, p_thr_link)
		tuagg_sumup(tu, &q->p_tu);
	mtx_leave(&pr->ps_mtx);
}

/*
 * Update the process ps_tu usage with the values from proc p while
 * doing so the times for proc p are reset.
 * This requires that p is either curproc or SDEAD and that the
 * IPL is higher than IPL_STATCLOCK. ps_mtx uses IPL_HIGH so
 * this should always be the case.
 */
void
tuagg_add_process(struct process *pr, struct proc *p)
{
	MUTEX_ASSERT_LOCKED(&pr->ps_mtx);
	KASSERT(curproc == p || p->p_stat == SDEAD);

	tu_enter(&pr->ps_tu);
	tuagg_sumup(&pr->ps_tu, &p->p_tu);
	tu_leave(&pr->ps_tu);

	/* Now reset CPU time usage for the thread. */
	timespecclear(&p->p_tu.tu_runtime);
	p->p_tu.tu_uticks = p->p_tu.tu_sticks = p->p_tu.tu_iticks = 0;
}

/*
 * Transform the running time and tick information in a struct tusage
 * into user, system, and interrupt time usage.
 */
void
calctsru(struct tusage *tup, struct timespec *up, struct timespec *sp,
    struct timespec *ip)
{
	u_quad_t st, ut, it;

	st = tup->tu_sticks;
	ut = tup->tu_uticks;
	it = tup->tu_iticks;

	if (st + ut + it == 0) {
		timespecclear(up);
		timespecclear(sp);
		if (ip != NULL)
			timespecclear(ip);
		return;
	}

	st = st * 1000000000 / stathz;
	sp->tv_sec = st / 1000000000;
	sp->tv_nsec = st % 1000000000;
	ut = ut * 1000000000 / stathz;
	up->tv_sec = ut / 1000000000;
	up->tv_nsec = ut % 1000000000;
	if (ip != NULL) {
		it = it * 1000000000 / stathz;
		ip->tv_sec = it / 1000000000;
		ip->tv_nsec = it % 1000000000;
	}
}

void
calcru(struct tusage *tup, struct timeval *up, struct timeval *sp,
    struct timeval *ip)
{
	struct timespec u, s, i;

	calctsru(tup, &u, &s, ip != NULL ? &i : NULL);
	TIMESPEC_TO_TIMEVAL(up, &u);
	TIMESPEC_TO_TIMEVAL(sp, &s);
	if (ip != NULL)
		TIMESPEC_TO_TIMEVAL(ip, &i);
}

int
sys_getrusage(struct proc *p, void *v, register_t *retval)
{
	struct sys_getrusage_args /* {
		syscallarg(int) who;
		syscallarg(struct rusage *) rusage;
	} */ *uap = v;
	struct rusage ru;
	int error;

	error = dogetrusage(p, SCARG(uap, who), &ru);
	if (error == 0) {
		error = copyout(&ru, SCARG(uap, rusage), sizeof(ru));
#ifdef KTRACE
		if (error == 0 && KTRPOINT(p, KTR_STRUCT))
			ktrrusage(p, &ru);
#endif
	}
	return (error);
}

int
dogetrusage(struct proc *p, int who, struct rusage *rup)
{
	struct process *pr = p->p_p;
	struct proc *q;
	struct tusage tu = { 0 };

	KERNEL_ASSERT_LOCKED();

	switch (who) {
	case RUSAGE_SELF:
		/* start with the sum of dead threads, if any */
		if (pr->ps_ru != NULL)
			*rup = *pr->ps_ru;
		else
			memset(rup, 0, sizeof(*rup));
		tuagg_sumup(&tu, &pr->ps_tu);

		/* add on all living threads */
		TAILQ_FOREACH(q, &pr->ps_threads, p_thr_link) {
			ruadd(rup, &q->p_ru);
			tuagg_sumup(&tu, &q->p_tu);
		}

		calcru(&tu, &rup->ru_utime, &rup->ru_stime, NULL);
		break;

	case RUSAGE_THREAD:
		*rup = p->p_ru;
		calcru(&p->p_tu, &rup->ru_utime, &rup->ru_stime, NULL);
		break;

	case RUSAGE_CHILDREN:
		*rup = pr->ps_cru;
		break;

	default:
		return (EINVAL);
	}
	return (0);
}

void
ruadd(struct rusage *ru, const struct rusage *ru2)
{
	long *ip;
	const long *ip2;
	int i;

	timeradd(&ru->ru_utime, &ru2->ru_utime, &ru->ru_utime);
	timeradd(&ru->ru_stime, &ru2->ru_stime, &ru->ru_stime);
	if (ru->ru_maxrss < ru2->ru_maxrss)
		ru->ru_maxrss = ru2->ru_maxrss;
	ip = &ru->ru_first; ip2 = &ru2->ru_first;
	for (i = &ru->ru_last - &ru->ru_first; i >= 0; i--)
		*ip++ += *ip2++;
}

/*
 * Check if the process exceeds its cpu resource allocation.
 * If over max, kill it.
 */
void
rucheck(void *arg)
{
	struct rlimit rlim;
	struct tusage tu = { 0 };
	struct process *pr = arg;
	struct proc *q;
	time_t runtime;

	KERNEL_ASSERT_LOCKED();

	mtx_enter(&pr->ps_mtx);
	rlim = pr->ps_limit->pl_rlimit[RLIMIT_CPU];
	tuagg_sumup(&tu, &pr->ps_tu);
	TAILQ_FOREACH(q, &pr->ps_threads, p_thr_link)
		tuagg_sumup(&tu, &q->p_tu);
	mtx_leave(&pr->ps_mtx);

	runtime = tu.tu_runtime.tv_sec;

	if ((rlim_t)runtime >= rlim.rlim_cur) {
		if ((rlim_t)runtime >= rlim.rlim_max) {
			prsignal(pr, SIGKILL);
		} else if (runtime >= pr->ps_nextxcpu) {
			prsignal(pr, SIGXCPU);
			pr->ps_nextxcpu = runtime + SIGXCPU_INTERVAL;
		}
	}

	timeout_add_msec(&pr->ps_rucheck_to, RUCHECK_INTERVAL);
}

struct pool plimit_pool;

void
lim_startup(struct plimit *limit0)
{
	rlim_t lim;
	int i;

	pool_init(&plimit_pool, sizeof(struct plimit), 0, IPL_MPFLOOR,
	    PR_WAITOK, "plimitpl", NULL);

	for (i = 0; i < nitems(limit0->pl_rlimit); i++)
		limit0->pl_rlimit[i].rlim_cur =
		    limit0->pl_rlimit[i].rlim_max = RLIM_INFINITY;
	limit0->pl_rlimit[RLIMIT_NOFILE].rlim_cur = NOFILE;
	limit0->pl_rlimit[RLIMIT_NOFILE].rlim_max = MIN(NOFILE_MAX,
	    (maxfiles - NOFILE > NOFILE) ? maxfiles - NOFILE : NOFILE);
	limit0->pl_rlimit[RLIMIT_NPROC].rlim_cur = MAXUPRC;
	lim = ptoa(uvmexp.free);
	limit0->pl_rlimit[RLIMIT_RSS].rlim_max = lim;
	lim = ptoa(64*1024);		/* Default to very low */
	limit0->pl_rlimit[RLIMIT_MEMLOCK].rlim_max = lim;
	limit0->pl_rlimit[RLIMIT_MEMLOCK].rlim_cur = lim / 3;
	refcnt_init(&limit0->pl_refcnt);
}

/*
 * Make a copy of the plimit structure.
 * We share these structures copy-on-write after fork,
 * and copy when a limit is changed.
 */
struct plimit *
lim_copy(struct plimit *lim)
{
	struct plimit *newlim;

	newlim = pool_get(&plimit_pool, PR_WAITOK);
	memcpy(newlim->pl_rlimit, lim->pl_rlimit,
	    sizeof(struct rlimit) * RLIM_NLIMITS);
	refcnt_init(&newlim->pl_refcnt);
	return (newlim);
}

void
lim_free(struct plimit *lim)
{
	if (refcnt_rele(&lim->pl_refcnt) == 0)
		return;
	pool_put(&plimit_pool, lim);
}

void
lim_fork(struct process *parent, struct process *child)
{
	struct plimit *limit;

	mtx_enter(&parent->ps_mtx);
	limit = parent->ps_limit;
	refcnt_take(&limit->pl_refcnt);
	mtx_leave(&parent->ps_mtx);

	child->ps_limit = limit;

	if (limit->pl_rlimit[RLIMIT_CPU].rlim_cur != RLIM_INFINITY)
		timeout_add_msec(&child->ps_rucheck_to, RUCHECK_INTERVAL);
}

/*
 * Return an exclusive write reference to the process' resource limit structure.
 * The caller has to release the structure by calling lim_write_commit().
 *
 * This invalidates any plimit read reference held by the calling thread.
 */
struct plimit *
lim_write_begin(void)
{
	struct plimit *limit;
	struct proc *p = curproc;

	rw_assert_wrlock(&rlimit_lock);

	if (p->p_limit != NULL)
		lim_free(p->p_limit);
	p->p_limit = NULL;

	/*
	 * It is safe to access ps_limit here without holding ps_mtx
	 * because rlimit_lock excludes other writers.
	 */

	limit = p->p_p->ps_limit;
	if (P_HASSIBLING(p) || refcnt_shared(&limit->pl_refcnt))
		limit = lim_copy(limit);

	return (limit);
}

/*
 * Finish exclusive write access to the plimit structure.
 * This makes the structure visible to other threads in the process.
 */
void
lim_write_commit(struct plimit *limit)
{
	struct plimit *olimit;
	struct proc *p = curproc;

	rw_assert_wrlock(&rlimit_lock);

	if (limit != p->p_p->ps_limit) {
		mtx_enter(&p->p_p->ps_mtx);
		olimit = p->p_p->ps_limit;
		p->p_p->ps_limit = limit;
		mtx_leave(&p->p_p->ps_mtx);

		lim_free(olimit);
	}
}

/*
 * Begin read access to the process' resource limit structure.
 * The access has to be finished by calling lim_read_leave().
 *
 * Sections denoted by lim_read_enter() and lim_read_leave() cannot nest.
 */
struct plimit *
lim_read_enter(void)
{
	struct plimit *limit;
	struct proc *p = curproc;
	struct process *pr = p->p_p;

	/*
	 * This thread might not observe the latest value of ps_limit
	 * if another thread updated the limits very recently on another CPU.
	 * However, the anomaly should disappear quickly, especially if
	 * there is any synchronization activity between the threads (or
	 * the CPUs).
	 */

	limit = p->p_limit;
	if (limit != pr->ps_limit) {
		mtx_enter(&pr->ps_mtx);
		limit = pr->ps_limit;
		refcnt_take(&limit->pl_refcnt);
		mtx_leave(&pr->ps_mtx);
		if (p->p_limit != NULL)
			lim_free(p->p_limit);
		p->p_limit = limit;
	}
	KASSERT(limit != NULL);
	return (limit);
}

/*
 * Get the value of the resource limit in given process.
 */
rlim_t
lim_cur_proc(struct proc *p, int which)
{
	struct process *pr = p->p_p;
	rlim_t val;

	KASSERT(which >= 0 && which < RLIM_NLIMITS);

	mtx_enter(&pr->ps_mtx);
	val = pr->ps_limit->pl_rlimit[which].rlim_cur;
	mtx_leave(&pr->ps_mtx);
	return (val);
}