xref: /netbsd-src/sys/kern/kern_time.c (revision deb6f0161a9109e7de9b519dc8dfb9478668dcdd)
1 /*	$NetBSD: kern_time.c,v 1.193 2018/11/29 17:40:12 maxv Exp $	*/
2 
3 /*-
4  * Copyright (c) 2000, 2004, 2005, 2007, 2008, 2009 The NetBSD Foundation, Inc.
5  * All rights reserved.
6  *
7  * This code is derived from software contributed to The NetBSD Foundation
8  * by Christopher G. Demetriou, and by Andrew Doran.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
20  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
21  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
23  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29  * POSSIBILITY OF SUCH DAMAGE.
30  */
31 
32 /*
33  * Copyright (c) 1982, 1986, 1989, 1993
34  *	The Regents of the University of California.  All rights reserved.
35  *
36  * Redistribution and use in source and binary forms, with or without
37  * modification, are permitted provided that the following conditions
38  * are met:
39  * 1. Redistributions of source code must retain the above copyright
40  *    notice, this list of conditions and the following disclaimer.
41  * 2. Redistributions in binary form must reproduce the above copyright
42  *    notice, this list of conditions and the following disclaimer in the
43  *    documentation and/or other materials provided with the distribution.
44  * 3. Neither the name of the University nor the names of its contributors
45  *    may be used to endorse or promote products derived from this software
46  *    without specific prior written permission.
47  *
48  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58  * SUCH DAMAGE.
59  *
60  *	@(#)kern_time.c	8.4 (Berkeley) 5/26/95
61  */
62 
63 #include <sys/cdefs.h>
64 __KERNEL_RCSID(0, "$NetBSD: kern_time.c,v 1.193 2018/11/29 17:40:12 maxv Exp $");
65 
66 #include <sys/param.h>
67 #include <sys/resourcevar.h>
68 #include <sys/kernel.h>
69 #include <sys/systm.h>
70 #include <sys/proc.h>
71 #include <sys/vnode.h>
72 #include <sys/signalvar.h>
73 #include <sys/syslog.h>
74 #include <sys/timetc.h>
75 #include <sys/timex.h>
76 #include <sys/kauth.h>
77 #include <sys/mount.h>
78 #include <sys/syscallargs.h>
79 #include <sys/cpu.h>
80 
81 static void	timer_intr(void *);
82 static void	itimerfire(struct ptimer *);
83 static void	itimerfree(struct ptimers *, int);
84 
85 kmutex_t	timer_lock;
86 
87 static void	*timer_sih;
88 static TAILQ_HEAD(, ptimer) timer_queue;
89 
90 struct pool ptimer_pool, ptimers_pool;
91 
92 #define	CLOCK_VIRTUAL_P(clockid)	\
93 	((clockid) == CLOCK_VIRTUAL || (clockid) == CLOCK_PROF)
94 
95 CTASSERT(ITIMER_REAL == CLOCK_REALTIME);
96 CTASSERT(ITIMER_VIRTUAL == CLOCK_VIRTUAL);
97 CTASSERT(ITIMER_PROF == CLOCK_PROF);
98 CTASSERT(ITIMER_MONOTONIC == CLOCK_MONOTONIC);
99 
100 #define	DELAYTIMER_MAX	32
101 
102 /*
103  * Initialize timekeeping.
104  */
105 void
106 time_init(void)
107 {
108 
109 	pool_init(&ptimer_pool, sizeof(struct ptimer), 0, 0, 0, "ptimerpl",
110 	    &pool_allocator_nointr, IPL_NONE);
111 	pool_init(&ptimers_pool, sizeof(struct ptimers), 0, 0, 0, "ptimerspl",
112 	    &pool_allocator_nointr, IPL_NONE);
113 }
114 
115 void
116 time_init2(void)
117 {
118 
119 	TAILQ_INIT(&timer_queue);
120 	mutex_init(&timer_lock, MUTEX_DEFAULT, IPL_SCHED);
121 	timer_sih = softint_establish(SOFTINT_CLOCK | SOFTINT_MPSAFE,
122 	    timer_intr, NULL);
123 }
124 
125 /* Time of day and interval timer support.
126  *
127  * These routines provide the kernel entry points to get and set
128  * the time-of-day and per-process interval timers.  Subroutines
129  * here provide support for adding and subtracting timeval structures
130  * and decrementing interval timers, optionally reloading the interval
131  * timers when they expire.
132  */
133 
134 /* This function is used by clock_settime and settimeofday */
135 static int
136 settime1(struct proc *p, const struct timespec *ts, bool check_kauth)
137 {
138 	struct timespec delta, now;
139 	int s;
140 
141 	/* WHAT DO WE DO ABOUT PENDING REAL-TIME TIMEOUTS??? */
142 	s = splclock();
143 	nanotime(&now);
144 	timespecsub(ts, &now, &delta);
145 
146 	if (check_kauth && kauth_authorize_system(kauth_cred_get(),
147 	    KAUTH_SYSTEM_TIME, KAUTH_REQ_SYSTEM_TIME_SYSTEM, __UNCONST(ts),
148 	    &delta, KAUTH_ARG(check_kauth ? false : true)) != 0) {
149 		splx(s);
150 		return (EPERM);
151 	}
152 
153 #ifdef notyet
154 	if ((delta.tv_sec < 86400) && securelevel > 0) { /* XXX elad - notyet */
155 		splx(s);
156 		return (EPERM);
157 	}
158 #endif
159 
160 	tc_setclock(ts);
161 
162 	timespecadd(&boottime, &delta, &boottime);
163 
164 	resettodr();
165 	splx(s);
166 
167 	return (0);
168 }
169 
170 int
171 settime(struct proc *p, struct timespec *ts)
172 {
173 	return (settime1(p, ts, true));
174 }
175 
176 /* ARGSUSED */
177 int
178 sys___clock_gettime50(struct lwp *l,
179     const struct sys___clock_gettime50_args *uap, register_t *retval)
180 {
181 	/* {
182 		syscallarg(clockid_t) clock_id;
183 		syscallarg(struct timespec *) tp;
184 	} */
185 	int error;
186 	struct timespec ats;
187 
188 	error = clock_gettime1(SCARG(uap, clock_id), &ats);
189 	if (error != 0)
190 		return error;
191 
192 	return copyout(&ats, SCARG(uap, tp), sizeof(ats));
193 }
194 
195 /* ARGSUSED */
196 int
197 sys___clock_settime50(struct lwp *l,
198     const struct sys___clock_settime50_args *uap, register_t *retval)
199 {
200 	/* {
201 		syscallarg(clockid_t) clock_id;
202 		syscallarg(const struct timespec *) tp;
203 	} */
204 	int error;
205 	struct timespec ats;
206 
207 	if ((error = copyin(SCARG(uap, tp), &ats, sizeof(ats))) != 0)
208 		return error;
209 
210 	return clock_settime1(l->l_proc, SCARG(uap, clock_id), &ats, true);
211 }
212 
213 
214 int
215 clock_settime1(struct proc *p, clockid_t clock_id, const struct timespec *tp,
216     bool check_kauth)
217 {
218 	int error;
219 
220 	switch (clock_id) {
221 	case CLOCK_REALTIME:
222 		if ((error = settime1(p, tp, check_kauth)) != 0)
223 			return (error);
224 		break;
225 	case CLOCK_MONOTONIC:
226 		return (EINVAL);	/* read-only clock */
227 	default:
228 		return (EINVAL);
229 	}
230 
231 	return 0;
232 }
233 
234 int
235 sys___clock_getres50(struct lwp *l, const struct sys___clock_getres50_args *uap,
236     register_t *retval)
237 {
238 	/* {
239 		syscallarg(clockid_t) clock_id;
240 		syscallarg(struct timespec *) tp;
241 	} */
242 	struct timespec ts;
243 	int error;
244 
245 	if ((error = clock_getres1(SCARG(uap, clock_id), &ts)) != 0)
246 		return error;
247 
248 	if (SCARG(uap, tp))
249 		error = copyout(&ts, SCARG(uap, tp), sizeof(ts));
250 
251 	return error;
252 }
253 
254 int
255 clock_getres1(clockid_t clock_id, struct timespec *ts)
256 {
257 
258 	switch (clock_id) {
259 	case CLOCK_REALTIME:
260 	case CLOCK_MONOTONIC:
261 		ts->tv_sec = 0;
262 		if (tc_getfrequency() > 1000000000)
263 			ts->tv_nsec = 1;
264 		else
265 			ts->tv_nsec = 1000000000 / tc_getfrequency();
266 		break;
267 	default:
268 		return EINVAL;
269 	}
270 
271 	return 0;
272 }
273 
274 /* ARGSUSED */
275 int
276 sys___nanosleep50(struct lwp *l, const struct sys___nanosleep50_args *uap,
277     register_t *retval)
278 {
279 	/* {
280 		syscallarg(struct timespec *) rqtp;
281 		syscallarg(struct timespec *) rmtp;
282 	} */
283 	struct timespec rmt, rqt;
284 	int error, error1;
285 
286 	error = copyin(SCARG(uap, rqtp), &rqt, sizeof(struct timespec));
287 	if (error)
288 		return (error);
289 
290 	error = nanosleep1(l, CLOCK_MONOTONIC, 0, &rqt,
291 	    SCARG(uap, rmtp) ? &rmt : NULL);
292 	if (SCARG(uap, rmtp) == NULL || (error != 0 && error != EINTR))
293 		return error;
294 
295 	error1 = copyout(&rmt, SCARG(uap, rmtp), sizeof(rmt));
296 	return error1 ? error1 : error;
297 }
298 
299 /* ARGSUSED */
300 int
301 sys_clock_nanosleep(struct lwp *l, const struct sys_clock_nanosleep_args *uap,
302     register_t *retval)
303 {
304 	/* {
305 		syscallarg(clockid_t) clock_id;
306 		syscallarg(int) flags;
307 		syscallarg(struct timespec *) rqtp;
308 		syscallarg(struct timespec *) rmtp;
309 	} */
310 	struct timespec rmt, rqt;
311 	int error, error1;
312 
313 	error = copyin(SCARG(uap, rqtp), &rqt, sizeof(struct timespec));
314 	if (error)
315 		goto out;
316 
317 	error = nanosleep1(l, SCARG(uap, clock_id), SCARG(uap, flags), &rqt,
318 	    SCARG(uap, rmtp) ? &rmt : NULL);
319 	if (SCARG(uap, rmtp) == NULL || (error != 0 && error != EINTR))
320 		goto out;
321 
322 	if ((SCARG(uap, flags) & TIMER_ABSTIME) == 0 &&
323 	    (error1 = copyout(&rmt, SCARG(uap, rmtp), sizeof(rmt))) != 0)
324 		error = error1;
325 out:
326 	*retval = error;
327 	return 0;
328 }
329 
330 int
331 nanosleep1(struct lwp *l, clockid_t clock_id, int flags, struct timespec *rqt,
332     struct timespec *rmt)
333 {
334 	struct timespec rmtstart;
335 	int error, timo;
336 
337 	if ((error = ts2timo(clock_id, flags, rqt, &timo, &rmtstart)) != 0) {
338 		if (error == ETIMEDOUT) {
339 			error = 0;
340 			if (rmt != NULL)
341 				rmt->tv_sec = rmt->tv_nsec = 0;
342 		}
343 		return error;
344 	}
345 
346 	/*
347 	 * Avoid inadvertently sleeping forever
348 	 */
349 	if (timo == 0)
350 		timo = 1;
351 again:
352 	error = kpause("nanoslp", true, timo, NULL);
353 	if (rmt != NULL || error == 0) {
354 		struct timespec rmtend;
355 		struct timespec t0;
356 		struct timespec *t;
357 
358 		(void)clock_gettime1(clock_id, &rmtend);
359 		t = (rmt != NULL) ? rmt : &t0;
360 		if (flags & TIMER_ABSTIME) {
361 			timespecsub(rqt, &rmtend, t);
362 		} else {
363 			timespecsub(&rmtend, &rmtstart, t);
364 			timespecsub(rqt, t, t);
365 		}
366 		if (t->tv_sec < 0)
367 			timespecclear(t);
368 		if (error == 0) {
369 			timo = tstohz(t);
370 			if (timo > 0)
371 				goto again;
372 		}
373 	}
374 
375 	if (error == ERESTART)
376 		error = EINTR;
377 	if (error == EWOULDBLOCK)
378 		error = 0;
379 
380 	return error;
381 }
382 
383 int
384 sys_clock_getcpuclockid2(struct lwp *l,
385     const struct sys_clock_getcpuclockid2_args *uap,
386     register_t *retval)
387 {
388 	/* {
389 		syscallarg(idtype_t idtype;
390 		syscallarg(id_t id);
391 		syscallarg(clockid_t *)clock_id;
392 	} */
393 	pid_t pid;
394 	lwpid_t lid;
395 	clockid_t clock_id;
396 	id_t id = SCARG(uap, id);
397 
398 	switch (SCARG(uap, idtype)) {
399 	case P_PID:
400 		pid = id == 0 ? l->l_proc->p_pid : id;
401 		clock_id = CLOCK_PROCESS_CPUTIME_ID | pid;
402 		break;
403 	case P_LWPID:
404 		lid = id == 0 ? l->l_lid : id;
405 		clock_id = CLOCK_THREAD_CPUTIME_ID | lid;
406 		break;
407 	default:
408 		return EINVAL;
409 	}
410 	return copyout(&clock_id, SCARG(uap, clock_id), sizeof(clock_id));
411 }
412 
413 /* ARGSUSED */
414 int
415 sys___gettimeofday50(struct lwp *l, const struct sys___gettimeofday50_args *uap,
416     register_t *retval)
417 {
418 	/* {
419 		syscallarg(struct timeval *) tp;
420 		syscallarg(void *) tzp;		really "struct timezone *";
421 	} */
422 	struct timeval atv;
423 	int error = 0;
424 	struct timezone tzfake;
425 
426 	if (SCARG(uap, tp)) {
427 		memset(&atv, 0, sizeof(atv));
428 		microtime(&atv);
429 		error = copyout(&atv, SCARG(uap, tp), sizeof(atv));
430 		if (error)
431 			return (error);
432 	}
433 	if (SCARG(uap, tzp)) {
434 		/*
435 		 * NetBSD has no kernel notion of time zone, so we just
436 		 * fake up a timezone struct and return it if demanded.
437 		 */
438 		tzfake.tz_minuteswest = 0;
439 		tzfake.tz_dsttime = 0;
440 		error = copyout(&tzfake, SCARG(uap, tzp), sizeof(tzfake));
441 	}
442 	return (error);
443 }
444 
445 /* ARGSUSED */
446 int
447 sys___settimeofday50(struct lwp *l, const struct sys___settimeofday50_args *uap,
448     register_t *retval)
449 {
450 	/* {
451 		syscallarg(const struct timeval *) tv;
452 		syscallarg(const void *) tzp; really "const struct timezone *";
453 	} */
454 
455 	return settimeofday1(SCARG(uap, tv), true, SCARG(uap, tzp), l, true);
456 }
457 
458 int
459 settimeofday1(const struct timeval *utv, bool userspace,
460     const void *utzp, struct lwp *l, bool check_kauth)
461 {
462 	struct timeval atv;
463 	struct timespec ts;
464 	int error;
465 
466 	/* Verify all parameters before changing time. */
467 
468 	/*
469 	 * NetBSD has no kernel notion of time zone, and only an
470 	 * obsolete program would try to set it, so we log a warning.
471 	 */
472 	if (utzp)
473 		log(LOG_WARNING, "pid %d attempted to set the "
474 		    "(obsolete) kernel time zone\n", l->l_proc->p_pid);
475 
476 	if (utv == NULL)
477 		return 0;
478 
479 	if (userspace) {
480 		if ((error = copyin(utv, &atv, sizeof(atv))) != 0)
481 			return error;
482 		utv = &atv;
483 	}
484 
485 	TIMEVAL_TO_TIMESPEC(utv, &ts);
486 	return settime1(l->l_proc, &ts, check_kauth);
487 }
488 
489 int	time_adjusted;			/* set if an adjustment is made */
490 
491 /* ARGSUSED */
492 int
493 sys___adjtime50(struct lwp *l, const struct sys___adjtime50_args *uap,
494     register_t *retval)
495 {
496 	/* {
497 		syscallarg(const struct timeval *) delta;
498 		syscallarg(struct timeval *) olddelta;
499 	} */
500 	int error;
501 	struct timeval atv, oldatv;
502 
503 	if ((error = kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_TIME,
504 	    KAUTH_REQ_SYSTEM_TIME_ADJTIME, NULL, NULL, NULL)) != 0)
505 		return error;
506 
507 	if (SCARG(uap, delta)) {
508 		error = copyin(SCARG(uap, delta), &atv,
509 		    sizeof(*SCARG(uap, delta)));
510 		if (error)
511 			return (error);
512 	}
513 	adjtime1(SCARG(uap, delta) ? &atv : NULL,
514 	    SCARG(uap, olddelta) ? &oldatv : NULL, l->l_proc);
515 	if (SCARG(uap, olddelta))
516 		error = copyout(&oldatv, SCARG(uap, olddelta),
517 		    sizeof(*SCARG(uap, olddelta)));
518 	return error;
519 }
520 
521 void
522 adjtime1(const struct timeval *delta, struct timeval *olddelta, struct proc *p)
523 {
524 	extern int64_t time_adjtime;  /* in kern_ntptime.c */
525 
526 	if (olddelta) {
527 		mutex_spin_enter(&timecounter_lock);
528 		olddelta->tv_sec = time_adjtime / 1000000;
529 		olddelta->tv_usec = time_adjtime % 1000000;
530 		if (olddelta->tv_usec < 0) {
531 			olddelta->tv_usec += 1000000;
532 			olddelta->tv_sec--;
533 		}
534 		mutex_spin_exit(&timecounter_lock);
535 	}
536 
537 	if (delta) {
538 		mutex_spin_enter(&timecounter_lock);
539 		time_adjtime = delta->tv_sec * 1000000 + delta->tv_usec;
540 
541 		if (time_adjtime) {
542 			/* We need to save the system time during shutdown */
543 			time_adjusted |= 1;
544 		}
545 		mutex_spin_exit(&timecounter_lock);
546 	}
547 }
548 
549 /*
550  * Interval timer support. Both the BSD getitimer() family and the POSIX
551  * timer_*() family of routines are supported.
552  *
553  * All timers are kept in an array pointed to by p_timers, which is
554  * allocated on demand - many processes don't use timers at all. The
555  * first four elements in this array are reserved for the BSD timers:
556  * element 0 is ITIMER_REAL, element 1 is ITIMER_VIRTUAL, element
557  * 2 is ITIMER_PROF, and element 3 is ITIMER_MONOTONIC. The rest may be
558  * allocated by the timer_create() syscall.
559  *
560  * Realtime timers are kept in the ptimer structure as an absolute
561  * time; virtual time timers are kept as a linked list of deltas.
562  * Virtual time timers are processed in the hardclock() routine of
563  * kern_clock.c.  The real time timer is processed by a callout
564  * routine, called from the softclock() routine.  Since a callout may
565  * be delayed in real time due to interrupt processing in the system,
566  * it is possible for the real time timeout routine (realtimeexpire,
567  * given below), to be delayed in real time past when it is supposed
568  * to occur.  It does not suffice, therefore, to reload the real timer
569  * .it_value from the real time timers .it_interval.  Rather, we
570  * compute the next time in absolute time the timer should go off.  */
571 
572 /* Allocate a POSIX realtime timer. */
573 int
574 sys_timer_create(struct lwp *l, const struct sys_timer_create_args *uap,
575     register_t *retval)
576 {
577 	/* {
578 		syscallarg(clockid_t) clock_id;
579 		syscallarg(struct sigevent *) evp;
580 		syscallarg(timer_t *) timerid;
581 	} */
582 
583 	return timer_create1(SCARG(uap, timerid), SCARG(uap, clock_id),
584 	    SCARG(uap, evp), copyin, l);
585 }
586 
587 int
588 timer_create1(timer_t *tid, clockid_t id, struct sigevent *evp,
589     copyin_t fetch_event, struct lwp *l)
590 {
591 	int error;
592 	timer_t timerid;
593 	struct ptimers *pts;
594 	struct ptimer *pt;
595 	struct proc *p;
596 
597 	p = l->l_proc;
598 
599 	if ((u_int)id > CLOCK_MONOTONIC)
600 		return (EINVAL);
601 
602 	if ((pts = p->p_timers) == NULL)
603 		pts = timers_alloc(p);
604 
605 	pt = pool_get(&ptimer_pool, PR_WAITOK);
606 	memset(pt, 0, sizeof(*pt));
607 	if (evp != NULL) {
608 		if (((error =
609 		    (*fetch_event)(evp, &pt->pt_ev, sizeof(pt->pt_ev))) != 0) ||
610 		    ((pt->pt_ev.sigev_notify < SIGEV_NONE) ||
611 			(pt->pt_ev.sigev_notify > SIGEV_SA)) ||
612 			(pt->pt_ev.sigev_notify == SIGEV_SIGNAL &&
613 			 (pt->pt_ev.sigev_signo <= 0 ||
614 			  pt->pt_ev.sigev_signo >= NSIG))) {
615 			pool_put(&ptimer_pool, pt);
616 			return (error ? error : EINVAL);
617 		}
618 	}
619 
620 	/* Find a free timer slot, skipping those reserved for setitimer(). */
621 	mutex_spin_enter(&timer_lock);
622 	for (timerid = TIMER_MIN; timerid < TIMER_MAX; timerid++)
623 		if (pts->pts_timers[timerid] == NULL)
624 			break;
625 	if (timerid == TIMER_MAX) {
626 		mutex_spin_exit(&timer_lock);
627 		pool_put(&ptimer_pool, pt);
628 		return EAGAIN;
629 	}
630 	if (evp == NULL) {
631 		pt->pt_ev.sigev_notify = SIGEV_SIGNAL;
632 		switch (id) {
633 		case CLOCK_REALTIME:
634 		case CLOCK_MONOTONIC:
635 			pt->pt_ev.sigev_signo = SIGALRM;
636 			break;
637 		case CLOCK_VIRTUAL:
638 			pt->pt_ev.sigev_signo = SIGVTALRM;
639 			break;
640 		case CLOCK_PROF:
641 			pt->pt_ev.sigev_signo = SIGPROF;
642 			break;
643 		}
644 		pt->pt_ev.sigev_value.sival_int = timerid;
645 	}
646 	pt->pt_info.ksi_signo = pt->pt_ev.sigev_signo;
647 	pt->pt_info.ksi_errno = 0;
648 	pt->pt_info.ksi_code = 0;
649 	pt->pt_info.ksi_pid = p->p_pid;
650 	pt->pt_info.ksi_uid = kauth_cred_getuid(l->l_cred);
651 	pt->pt_info.ksi_value = pt->pt_ev.sigev_value;
652 	pt->pt_type = id;
653 	pt->pt_proc = p;
654 	pt->pt_overruns = 0;
655 	pt->pt_poverruns = 0;
656 	pt->pt_entry = timerid;
657 	pt->pt_queued = false;
658 	timespecclear(&pt->pt_time.it_value);
659 	if (!CLOCK_VIRTUAL_P(id))
660 		callout_init(&pt->pt_ch, CALLOUT_MPSAFE);
661 	else
662 		pt->pt_active = 0;
663 
664 	pts->pts_timers[timerid] = pt;
665 	mutex_spin_exit(&timer_lock);
666 
667 	return copyout(&timerid, tid, sizeof(timerid));
668 }
669 
670 /* Delete a POSIX realtime timer */
671 int
672 sys_timer_delete(struct lwp *l, const struct sys_timer_delete_args *uap,
673     register_t *retval)
674 {
675 	/* {
676 		syscallarg(timer_t) timerid;
677 	} */
678 	struct proc *p = l->l_proc;
679 	timer_t timerid;
680 	struct ptimers *pts;
681 	struct ptimer *pt, *ptn;
682 
683 	timerid = SCARG(uap, timerid);
684 	pts = p->p_timers;
685 
686 	if (pts == NULL || timerid < 2 || timerid >= TIMER_MAX)
687 		return (EINVAL);
688 
689 	mutex_spin_enter(&timer_lock);
690 	if ((pt = pts->pts_timers[timerid]) == NULL) {
691 		mutex_spin_exit(&timer_lock);
692 		return (EINVAL);
693 	}
694 	if (CLOCK_VIRTUAL_P(pt->pt_type)) {
695 		if (pt->pt_active) {
696 			ptn = LIST_NEXT(pt, pt_list);
697 			LIST_REMOVE(pt, pt_list);
698 			for ( ; ptn; ptn = LIST_NEXT(ptn, pt_list))
699 				timespecadd(&pt->pt_time.it_value,
700 				    &ptn->pt_time.it_value,
701 				    &ptn->pt_time.it_value);
702 			pt->pt_active = 0;
703 		}
704 	}
705 	itimerfree(pts, timerid);
706 
707 	return (0);
708 }
709 
710 /*
711  * Set up the given timer. The value in pt->pt_time.it_value is taken
712  * to be an absolute time for CLOCK_REALTIME/CLOCK_MONOTONIC timers and
713  * a relative time for CLOCK_VIRTUAL/CLOCK_PROF timers.
714  */
715 void
716 timer_settime(struct ptimer *pt)
717 {
718 	struct ptimer *ptn, *pptn;
719 	struct ptlist *ptl;
720 
721 	KASSERT(mutex_owned(&timer_lock));
722 
723 	if (!CLOCK_VIRTUAL_P(pt->pt_type)) {
724 		callout_halt(&pt->pt_ch, &timer_lock);
725 		if (timespecisset(&pt->pt_time.it_value)) {
726 			/*
727 			 * Don't need to check tshzto() return value, here.
728 			 * callout_reset() does it for us.
729 			 */
730 			callout_reset(&pt->pt_ch,
731 			    pt->pt_type == CLOCK_MONOTONIC ?
732 			    tshztoup(&pt->pt_time.it_value) :
733 			    tshzto(&pt->pt_time.it_value),
734 			    realtimerexpire, pt);
735 		}
736 	} else {
737 		if (pt->pt_active) {
738 			ptn = LIST_NEXT(pt, pt_list);
739 			LIST_REMOVE(pt, pt_list);
740 			for ( ; ptn; ptn = LIST_NEXT(ptn, pt_list))
741 				timespecadd(&pt->pt_time.it_value,
742 				    &ptn->pt_time.it_value,
743 				    &ptn->pt_time.it_value);
744 		}
745 		if (timespecisset(&pt->pt_time.it_value)) {
746 			if (pt->pt_type == CLOCK_VIRTUAL)
747 				ptl = &pt->pt_proc->p_timers->pts_virtual;
748 			else
749 				ptl = &pt->pt_proc->p_timers->pts_prof;
750 
751 			for (ptn = LIST_FIRST(ptl), pptn = NULL;
752 			     ptn && timespeccmp(&pt->pt_time.it_value,
753 				 &ptn->pt_time.it_value, >);
754 			     pptn = ptn, ptn = LIST_NEXT(ptn, pt_list))
755 				timespecsub(&pt->pt_time.it_value,
756 				    &ptn->pt_time.it_value,
757 				    &pt->pt_time.it_value);
758 
759 			if (pptn)
760 				LIST_INSERT_AFTER(pptn, pt, pt_list);
761 			else
762 				LIST_INSERT_HEAD(ptl, pt, pt_list);
763 
764 			for ( ; ptn ; ptn = LIST_NEXT(ptn, pt_list))
765 				timespecsub(&ptn->pt_time.it_value,
766 				    &pt->pt_time.it_value,
767 				    &ptn->pt_time.it_value);
768 
769 			pt->pt_active = 1;
770 		} else
771 			pt->pt_active = 0;
772 	}
773 }
774 
775 void
776 timer_gettime(struct ptimer *pt, struct itimerspec *aits)
777 {
778 	struct timespec now;
779 	struct ptimer *ptn;
780 
781 	KASSERT(mutex_owned(&timer_lock));
782 
783 	*aits = pt->pt_time;
784 	if (!CLOCK_VIRTUAL_P(pt->pt_type)) {
785 		/*
786 		 * Convert from absolute to relative time in .it_value
787 		 * part of real time timer.  If time for real time
788 		 * timer has passed return 0, else return difference
789 		 * between current time and time for the timer to go
790 		 * off.
791 		 */
792 		if (timespecisset(&aits->it_value)) {
793 			if (pt->pt_type == CLOCK_REALTIME) {
794 				getnanotime(&now);
795 			} else { /* CLOCK_MONOTONIC */
796 				getnanouptime(&now);
797 			}
798 			if (timespeccmp(&aits->it_value, &now, <))
799 				timespecclear(&aits->it_value);
800 			else
801 				timespecsub(&aits->it_value, &now,
802 				    &aits->it_value);
803 		}
804 	} else if (pt->pt_active) {
805 		if (pt->pt_type == CLOCK_VIRTUAL)
806 			ptn = LIST_FIRST(&pt->pt_proc->p_timers->pts_virtual);
807 		else
808 			ptn = LIST_FIRST(&pt->pt_proc->p_timers->pts_prof);
809 		for ( ; ptn && ptn != pt; ptn = LIST_NEXT(ptn, pt_list))
810 			timespecadd(&aits->it_value,
811 			    &ptn->pt_time.it_value, &aits->it_value);
812 		KASSERT(ptn != NULL); /* pt should be findable on the list */
813 	} else
814 		timespecclear(&aits->it_value);
815 }
816 
817 
818 
819 /* Set and arm a POSIX realtime timer */
820 int
821 sys___timer_settime50(struct lwp *l,
822     const struct sys___timer_settime50_args *uap,
823     register_t *retval)
824 {
825 	/* {
826 		syscallarg(timer_t) timerid;
827 		syscallarg(int) flags;
828 		syscallarg(const struct itimerspec *) value;
829 		syscallarg(struct itimerspec *) ovalue;
830 	} */
831 	int error;
832 	struct itimerspec value, ovalue, *ovp = NULL;
833 
834 	if ((error = copyin(SCARG(uap, value), &value,
835 	    sizeof(struct itimerspec))) != 0)
836 		return (error);
837 
838 	if (SCARG(uap, ovalue))
839 		ovp = &ovalue;
840 
841 	if ((error = dotimer_settime(SCARG(uap, timerid), &value, ovp,
842 	    SCARG(uap, flags), l->l_proc)) != 0)
843 		return error;
844 
845 	if (ovp)
846 		return copyout(&ovalue, SCARG(uap, ovalue),
847 		    sizeof(struct itimerspec));
848 	return 0;
849 }
850 
851 int
852 dotimer_settime(int timerid, struct itimerspec *value,
853     struct itimerspec *ovalue, int flags, struct proc *p)
854 {
855 	struct timespec now;
856 	struct itimerspec val, oval;
857 	struct ptimers *pts;
858 	struct ptimer *pt;
859 	int error;
860 
861 	pts = p->p_timers;
862 
863 	if (pts == NULL || timerid < 2 || timerid >= TIMER_MAX)
864 		return EINVAL;
865 	val = *value;
866 	if ((error = itimespecfix(&val.it_value)) != 0 ||
867 	    (error = itimespecfix(&val.it_interval)) != 0)
868 		return error;
869 
870 	mutex_spin_enter(&timer_lock);
871 	if ((pt = pts->pts_timers[timerid]) == NULL) {
872 		mutex_spin_exit(&timer_lock);
873 		return EINVAL;
874 	}
875 
876 	oval = pt->pt_time;
877 	pt->pt_time = val;
878 
879 	/*
880 	 * If we've been passed a relative time for a realtime timer,
881 	 * convert it to absolute; if an absolute time for a virtual
882 	 * timer, convert it to relative and make sure we don't set it
883 	 * to zero, which would cancel the timer, or let it go
884 	 * negative, which would confuse the comparison tests.
885 	 */
886 	if (timespecisset(&pt->pt_time.it_value)) {
887 		if (!CLOCK_VIRTUAL_P(pt->pt_type)) {
888 			if ((flags & TIMER_ABSTIME) == 0) {
889 				if (pt->pt_type == CLOCK_REALTIME) {
890 					getnanotime(&now);
891 				} else { /* CLOCK_MONOTONIC */
892 					getnanouptime(&now);
893 				}
894 				timespecadd(&pt->pt_time.it_value, &now,
895 				    &pt->pt_time.it_value);
896 			}
897 		} else {
898 			if ((flags & TIMER_ABSTIME) != 0) {
899 				getnanotime(&now);
900 				timespecsub(&pt->pt_time.it_value, &now,
901 				    &pt->pt_time.it_value);
902 				if (!timespecisset(&pt->pt_time.it_value) ||
903 				    pt->pt_time.it_value.tv_sec < 0) {
904 					pt->pt_time.it_value.tv_sec = 0;
905 					pt->pt_time.it_value.tv_nsec = 1;
906 				}
907 			}
908 		}
909 	}
910 
911 	timer_settime(pt);
912 	mutex_spin_exit(&timer_lock);
913 
914 	if (ovalue)
915 		*ovalue = oval;
916 
917 	return (0);
918 }
919 
920 /* Return the time remaining until a POSIX timer fires. */
921 int
922 sys___timer_gettime50(struct lwp *l,
923     const struct sys___timer_gettime50_args *uap, register_t *retval)
924 {
925 	/* {
926 		syscallarg(timer_t) timerid;
927 		syscallarg(struct itimerspec *) value;
928 	} */
929 	struct itimerspec its;
930 	int error;
931 
932 	if ((error = dotimer_gettime(SCARG(uap, timerid), l->l_proc,
933 	    &its)) != 0)
934 		return error;
935 
936 	return copyout(&its, SCARG(uap, value), sizeof(its));
937 }
938 
939 int
940 dotimer_gettime(int timerid, struct proc *p, struct itimerspec *its)
941 {
942 	struct ptimer *pt;
943 	struct ptimers *pts;
944 
945 	pts = p->p_timers;
946 	if (pts == NULL || timerid < 2 || timerid >= TIMER_MAX)
947 		return (EINVAL);
948 	mutex_spin_enter(&timer_lock);
949 	if ((pt = pts->pts_timers[timerid]) == NULL) {
950 		mutex_spin_exit(&timer_lock);
951 		return (EINVAL);
952 	}
953 	timer_gettime(pt, its);
954 	mutex_spin_exit(&timer_lock);
955 
956 	return 0;
957 }
958 
959 /*
960  * Return the count of the number of times a periodic timer expired
961  * while a notification was already pending. The counter is reset when
962  * a timer expires and a notification can be posted.
963  */
964 int
965 sys_timer_getoverrun(struct lwp *l, const struct sys_timer_getoverrun_args *uap,
966     register_t *retval)
967 {
968 	/* {
969 		syscallarg(timer_t) timerid;
970 	} */
971 	struct proc *p = l->l_proc;
972 	struct ptimers *pts;
973 	int timerid;
974 	struct ptimer *pt;
975 
976 	timerid = SCARG(uap, timerid);
977 
978 	pts = p->p_timers;
979 	if (pts == NULL || timerid < 2 || timerid >= TIMER_MAX)
980 		return (EINVAL);
981 	mutex_spin_enter(&timer_lock);
982 	if ((pt = pts->pts_timers[timerid]) == NULL) {
983 		mutex_spin_exit(&timer_lock);
984 		return (EINVAL);
985 	}
986 	*retval = pt->pt_poverruns;
987 	if (*retval >= DELAYTIMER_MAX)
988 		*retval = DELAYTIMER_MAX;
989 	mutex_spin_exit(&timer_lock);
990 
991 	return (0);
992 }
993 
994 /*
995  * Real interval timer expired:
996  * send process whose timer expired an alarm signal.
997  * If time is not set up to reload, then just return.
998  * Else compute next time timer should go off which is > current time.
999  * This is where delay in processing this timeout causes multiple
1000  * SIGALRM calls to be compressed into one.
1001  */
1002 void
1003 realtimerexpire(void *arg)
1004 {
1005 	uint64_t last_val, next_val, interval, now_ns;
1006 	struct timespec now, next;
1007 	struct ptimer *pt;
1008 	int backwards;
1009 
1010 	pt = arg;
1011 
1012 	mutex_spin_enter(&timer_lock);
1013 	itimerfire(pt);
1014 
1015 	if (!timespecisset(&pt->pt_time.it_interval)) {
1016 		timespecclear(&pt->pt_time.it_value);
1017 		mutex_spin_exit(&timer_lock);
1018 		return;
1019 	}
1020 
1021 	if (pt->pt_type == CLOCK_MONOTONIC) {
1022 		getnanouptime(&now);
1023 	} else {
1024 		getnanotime(&now);
1025 	}
1026 	backwards = (timespeccmp(&pt->pt_time.it_value, &now, >));
1027 	timespecadd(&pt->pt_time.it_value, &pt->pt_time.it_interval, &next);
1028 	/* Handle the easy case of non-overflown timers first. */
1029 	if (!backwards && timespeccmp(&next, &now, >)) {
1030 		pt->pt_time.it_value = next;
1031 	} else {
1032 		now_ns = timespec2ns(&now);
1033 		last_val = timespec2ns(&pt->pt_time.it_value);
1034 		interval = timespec2ns(&pt->pt_time.it_interval);
1035 
1036 		next_val = now_ns +
1037 		    (now_ns - last_val + interval - 1) % interval;
1038 
1039 		if (backwards)
1040 			next_val += interval;
1041 		else
1042 			pt->pt_overruns += (now_ns - last_val) / interval;
1043 
1044 		pt->pt_time.it_value.tv_sec = next_val / 1000000000;
1045 		pt->pt_time.it_value.tv_nsec = next_val % 1000000000;
1046 	}
1047 
1048 	/*
1049 	 * Don't need to check tshzto() return value, here.
1050 	 * callout_reset() does it for us.
1051 	 */
1052 	callout_reset(&pt->pt_ch, pt->pt_type == CLOCK_MONOTONIC ?
1053 	    tshztoup(&pt->pt_time.it_value) : tshzto(&pt->pt_time.it_value),
1054 	    realtimerexpire, pt);
1055 	mutex_spin_exit(&timer_lock);
1056 }
1057 
1058 /* BSD routine to get the value of an interval timer. */
1059 /* ARGSUSED */
1060 int
1061 sys___getitimer50(struct lwp *l, const struct sys___getitimer50_args *uap,
1062     register_t *retval)
1063 {
1064 	/* {
1065 		syscallarg(int) which;
1066 		syscallarg(struct itimerval *) itv;
1067 	} */
1068 	struct proc *p = l->l_proc;
1069 	struct itimerval aitv;
1070 	int error;
1071 
1072 	memset(&aitv, 0, sizeof(aitv));
1073 	error = dogetitimer(p, SCARG(uap, which), &aitv);
1074 	if (error)
1075 		return error;
1076 	return (copyout(&aitv, SCARG(uap, itv), sizeof(struct itimerval)));
1077 }
1078 
1079 int
1080 dogetitimer(struct proc *p, int which, struct itimerval *itvp)
1081 {
1082 	struct ptimers *pts;
1083 	struct ptimer *pt;
1084 	struct itimerspec its;
1085 
1086 	if ((u_int)which > ITIMER_MONOTONIC)
1087 		return (EINVAL);
1088 
1089 	mutex_spin_enter(&timer_lock);
1090 	pts = p->p_timers;
1091 	if (pts == NULL || (pt = pts->pts_timers[which]) == NULL) {
1092 		timerclear(&itvp->it_value);
1093 		timerclear(&itvp->it_interval);
1094 	} else {
1095 		timer_gettime(pt, &its);
1096 		TIMESPEC_TO_TIMEVAL(&itvp->it_value, &its.it_value);
1097 		TIMESPEC_TO_TIMEVAL(&itvp->it_interval, &its.it_interval);
1098 	}
1099 	mutex_spin_exit(&timer_lock);
1100 
1101 	return 0;
1102 }
1103 
1104 /* BSD routine to set/arm an interval timer. */
1105 /* ARGSUSED */
1106 int
1107 sys___setitimer50(struct lwp *l, const struct sys___setitimer50_args *uap,
1108     register_t *retval)
1109 {
1110 	/* {
1111 		syscallarg(int) which;
1112 		syscallarg(const struct itimerval *) itv;
1113 		syscallarg(struct itimerval *) oitv;
1114 	} */
1115 	struct proc *p = l->l_proc;
1116 	int which = SCARG(uap, which);
1117 	struct sys___getitimer50_args getargs;
1118 	const struct itimerval *itvp;
1119 	struct itimerval aitv;
1120 	int error;
1121 
1122 	if ((u_int)which > ITIMER_MONOTONIC)
1123 		return (EINVAL);
1124 	itvp = SCARG(uap, itv);
1125 	if (itvp &&
1126 	    (error = copyin(itvp, &aitv, sizeof(struct itimerval))) != 0)
1127 		return (error);
1128 	if (SCARG(uap, oitv) != NULL) {
1129 		SCARG(&getargs, which) = which;
1130 		SCARG(&getargs, itv) = SCARG(uap, oitv);
1131 		if ((error = sys___getitimer50(l, &getargs, retval)) != 0)
1132 			return (error);
1133 	}
1134 	if (itvp == 0)
1135 		return (0);
1136 
1137 	return dosetitimer(p, which, &aitv);
1138 }
1139 
1140 int
1141 dosetitimer(struct proc *p, int which, struct itimerval *itvp)
1142 {
1143 	struct timespec now;
1144 	struct ptimers *pts;
1145 	struct ptimer *pt, *spare;
1146 
1147 	KASSERT((u_int)which <= CLOCK_MONOTONIC);
1148 	if (itimerfix(&itvp->it_value) || itimerfix(&itvp->it_interval))
1149 		return (EINVAL);
1150 
1151 	/*
1152 	 * Don't bother allocating data structures if the process just
1153 	 * wants to clear the timer.
1154 	 */
1155 	spare = NULL;
1156 	pts = p->p_timers;
1157  retry:
1158 	if (!timerisset(&itvp->it_value) && (pts == NULL ||
1159 	    pts->pts_timers[which] == NULL))
1160 		return (0);
1161 	if (pts == NULL)
1162 		pts = timers_alloc(p);
1163 	mutex_spin_enter(&timer_lock);
1164 	pt = pts->pts_timers[which];
1165 	if (pt == NULL) {
1166 		if (spare == NULL) {
1167 			mutex_spin_exit(&timer_lock);
1168 			spare = pool_get(&ptimer_pool, PR_WAITOK);
1169 			memset(spare, 0, sizeof(*spare));
1170 			goto retry;
1171 		}
1172 		pt = spare;
1173 		spare = NULL;
1174 		pt->pt_ev.sigev_notify = SIGEV_SIGNAL;
1175 		pt->pt_ev.sigev_value.sival_int = which;
1176 		pt->pt_overruns = 0;
1177 		pt->pt_proc = p;
1178 		pt->pt_type = which;
1179 		pt->pt_entry = which;
1180 		pt->pt_queued = false;
1181 		if (pt->pt_type == CLOCK_REALTIME)
1182 			callout_init(&pt->pt_ch, CALLOUT_MPSAFE);
1183 		else
1184 			pt->pt_active = 0;
1185 
1186 		switch (which) {
1187 		case ITIMER_REAL:
1188 		case ITIMER_MONOTONIC:
1189 			pt->pt_ev.sigev_signo = SIGALRM;
1190 			break;
1191 		case ITIMER_VIRTUAL:
1192 			pt->pt_ev.sigev_signo = SIGVTALRM;
1193 			break;
1194 		case ITIMER_PROF:
1195 			pt->pt_ev.sigev_signo = SIGPROF;
1196 			break;
1197 		}
1198 		pts->pts_timers[which] = pt;
1199 	}
1200 
1201 	TIMEVAL_TO_TIMESPEC(&itvp->it_value, &pt->pt_time.it_value);
1202 	TIMEVAL_TO_TIMESPEC(&itvp->it_interval, &pt->pt_time.it_interval);
1203 
1204 	if (timespecisset(&pt->pt_time.it_value)) {
1205 		/* Convert to absolute time */
1206 		/* XXX need to wrap in splclock for timecounters case? */
1207 		switch (which) {
1208 		case ITIMER_REAL:
1209 			getnanotime(&now);
1210 			timespecadd(&pt->pt_time.it_value, &now,
1211 			    &pt->pt_time.it_value);
1212 			break;
1213 		case ITIMER_MONOTONIC:
1214 			getnanouptime(&now);
1215 			timespecadd(&pt->pt_time.it_value, &now,
1216 			    &pt->pt_time.it_value);
1217 			break;
1218 		default:
1219 			break;
1220 		}
1221 	}
1222 	timer_settime(pt);
1223 	mutex_spin_exit(&timer_lock);
1224 	if (spare != NULL)
1225 		pool_put(&ptimer_pool, spare);
1226 
1227 	return (0);
1228 }
1229 
1230 /* Utility routines to manage the array of pointers to timers. */
1231 struct ptimers *
1232 timers_alloc(struct proc *p)
1233 {
1234 	struct ptimers *pts;
1235 	int i;
1236 
1237 	pts = pool_get(&ptimers_pool, PR_WAITOK);
1238 	LIST_INIT(&pts->pts_virtual);
1239 	LIST_INIT(&pts->pts_prof);
1240 	for (i = 0; i < TIMER_MAX; i++)
1241 		pts->pts_timers[i] = NULL;
1242 	mutex_spin_enter(&timer_lock);
1243 	if (p->p_timers == NULL) {
1244 		p->p_timers = pts;
1245 		mutex_spin_exit(&timer_lock);
1246 		return pts;
1247 	}
1248 	mutex_spin_exit(&timer_lock);
1249 	pool_put(&ptimers_pool, pts);
1250 	return p->p_timers;
1251 }
1252 
1253 /*
1254  * Clean up the per-process timers. If "which" is set to TIMERS_ALL,
1255  * then clean up all timers and free all the data structures. If
1256  * "which" is set to TIMERS_POSIX, only clean up the timers allocated
1257  * by timer_create(), not the BSD setitimer() timers, and only free the
1258  * structure if none of those remain.
1259  */
1260 void
1261 timers_free(struct proc *p, int which)
1262 {
1263 	struct ptimers *pts;
1264 	struct ptimer *ptn;
1265 	struct timespec ts;
1266 	int i;
1267 
1268 	if (p->p_timers == NULL)
1269 		return;
1270 
1271 	pts = p->p_timers;
1272 	mutex_spin_enter(&timer_lock);
1273 	if (which == TIMERS_ALL) {
1274 		p->p_timers = NULL;
1275 		i = 0;
1276 	} else {
1277 		timespecclear(&ts);
1278 		for (ptn = LIST_FIRST(&pts->pts_virtual);
1279 		     ptn && ptn != pts->pts_timers[ITIMER_VIRTUAL];
1280 		     ptn = LIST_NEXT(ptn, pt_list)) {
1281 			KASSERT(ptn->pt_type == CLOCK_VIRTUAL);
1282 			timespecadd(&ts, &ptn->pt_time.it_value, &ts);
1283 		}
1284 		LIST_FIRST(&pts->pts_virtual) = NULL;
1285 		if (ptn) {
1286 			KASSERT(ptn->pt_type == CLOCK_VIRTUAL);
1287 			timespecadd(&ts, &ptn->pt_time.it_value,
1288 			    &ptn->pt_time.it_value);
1289 			LIST_INSERT_HEAD(&pts->pts_virtual, ptn, pt_list);
1290 		}
1291 		timespecclear(&ts);
1292 		for (ptn = LIST_FIRST(&pts->pts_prof);
1293 		     ptn && ptn != pts->pts_timers[ITIMER_PROF];
1294 		     ptn = LIST_NEXT(ptn, pt_list)) {
1295 			KASSERT(ptn->pt_type == CLOCK_PROF);
1296 			timespecadd(&ts, &ptn->pt_time.it_value, &ts);
1297 		}
1298 		LIST_FIRST(&pts->pts_prof) = NULL;
1299 		if (ptn) {
1300 			KASSERT(ptn->pt_type == CLOCK_PROF);
1301 			timespecadd(&ts, &ptn->pt_time.it_value,
1302 			    &ptn->pt_time.it_value);
1303 			LIST_INSERT_HEAD(&pts->pts_prof, ptn, pt_list);
1304 		}
1305 		i = TIMER_MIN;
1306 	}
1307 	for ( ; i < TIMER_MAX; i++) {
1308 		if (pts->pts_timers[i] != NULL) {
1309 			itimerfree(pts, i);
1310 			mutex_spin_enter(&timer_lock);
1311 		}
1312 	}
1313 	if (pts->pts_timers[0] == NULL && pts->pts_timers[1] == NULL &&
1314 	    pts->pts_timers[2] == NULL && pts->pts_timers[3] == NULL) {
1315 		p->p_timers = NULL;
1316 		mutex_spin_exit(&timer_lock);
1317 		pool_put(&ptimers_pool, pts);
1318 	} else
1319 		mutex_spin_exit(&timer_lock);
1320 }
1321 
1322 static void
1323 itimerfree(struct ptimers *pts, int index)
1324 {
1325 	struct ptimer *pt;
1326 
1327 	KASSERT(mutex_owned(&timer_lock));
1328 
1329 	pt = pts->pts_timers[index];
1330 	pts->pts_timers[index] = NULL;
1331 	if (!CLOCK_VIRTUAL_P(pt->pt_type))
1332 		callout_halt(&pt->pt_ch, &timer_lock);
1333 	if (pt->pt_queued)
1334 		TAILQ_REMOVE(&timer_queue, pt, pt_chain);
1335 	mutex_spin_exit(&timer_lock);
1336 	if (!CLOCK_VIRTUAL_P(pt->pt_type))
1337 		callout_destroy(&pt->pt_ch);
1338 	pool_put(&ptimer_pool, pt);
1339 }
1340 
1341 /*
1342  * Decrement an interval timer by a specified number
1343  * of nanoseconds, which must be less than a second,
1344  * i.e. < 1000000000.  If the timer expires, then reload
1345  * it.  In this case, carry over (nsec - old value) to
1346  * reduce the value reloaded into the timer so that
1347  * the timer does not drift.  This routine assumes
1348  * that it is called in a context where the timers
1349  * on which it is operating cannot change in value.
1350  */
1351 static int
1352 itimerdecr(struct ptimer *pt, int nsec)
1353 {
1354 	struct itimerspec *itp;
1355 
1356 	KASSERT(mutex_owned(&timer_lock));
1357 	KASSERT(CLOCK_VIRTUAL_P(pt->pt_type));
1358 
1359 	itp = &pt->pt_time;
1360 	if (itp->it_value.tv_nsec < nsec) {
1361 		if (itp->it_value.tv_sec == 0) {
1362 			/* expired, and already in next interval */
1363 			nsec -= itp->it_value.tv_nsec;
1364 			goto expire;
1365 		}
1366 		itp->it_value.tv_nsec += 1000000000;
1367 		itp->it_value.tv_sec--;
1368 	}
1369 	itp->it_value.tv_nsec -= nsec;
1370 	nsec = 0;
1371 	if (timespecisset(&itp->it_value))
1372 		return (1);
1373 	/* expired, exactly at end of interval */
1374 expire:
1375 	if (timespecisset(&itp->it_interval)) {
1376 		itp->it_value = itp->it_interval;
1377 		itp->it_value.tv_nsec -= nsec;
1378 		if (itp->it_value.tv_nsec < 0) {
1379 			itp->it_value.tv_nsec += 1000000000;
1380 			itp->it_value.tv_sec--;
1381 		}
1382 		timer_settime(pt);
1383 	} else
1384 		itp->it_value.tv_nsec = 0;		/* sec is already 0 */
1385 	return (0);
1386 }
1387 
1388 static void
1389 itimerfire(struct ptimer *pt)
1390 {
1391 
1392 	KASSERT(mutex_owned(&timer_lock));
1393 
1394 	/*
1395 	 * XXX Can overrun, but we don't do signal queueing yet, anyway.
1396 	 * XXX Relying on the clock interrupt is stupid.
1397 	 */
1398 	if (pt->pt_ev.sigev_notify != SIGEV_SIGNAL || pt->pt_queued) {
1399 		return;
1400 	}
1401 	TAILQ_INSERT_TAIL(&timer_queue, pt, pt_chain);
1402 	pt->pt_queued = true;
1403 	softint_schedule(timer_sih);
1404 }
1405 
1406 void
1407 timer_tick(lwp_t *l, bool user)
1408 {
1409 	struct ptimers *pts;
1410 	struct ptimer *pt;
1411 	proc_t *p;
1412 
1413 	p = l->l_proc;
1414 	if (p->p_timers == NULL)
1415 		return;
1416 
1417 	mutex_spin_enter(&timer_lock);
1418 	if ((pts = l->l_proc->p_timers) != NULL) {
1419 		/*
1420 		 * Run current process's virtual and profile time, as needed.
1421 		 */
1422 		if (user && (pt = LIST_FIRST(&pts->pts_virtual)) != NULL)
1423 			if (itimerdecr(pt, tick * 1000) == 0)
1424 				itimerfire(pt);
1425 		if ((pt = LIST_FIRST(&pts->pts_prof)) != NULL)
1426 			if (itimerdecr(pt, tick * 1000) == 0)
1427 				itimerfire(pt);
1428 	}
1429 	mutex_spin_exit(&timer_lock);
1430 }
1431 
1432 static void
1433 timer_intr(void *cookie)
1434 {
1435 	ksiginfo_t ksi;
1436 	struct ptimer *pt;
1437 	proc_t *p;
1438 
1439 	mutex_enter(proc_lock);
1440 	mutex_spin_enter(&timer_lock);
1441 	while ((pt = TAILQ_FIRST(&timer_queue)) != NULL) {
1442 		TAILQ_REMOVE(&timer_queue, pt, pt_chain);
1443 		KASSERT(pt->pt_queued);
1444 		pt->pt_queued = false;
1445 
1446 		if (pt->pt_proc->p_timers == NULL) {
1447 			/* Process is dying. */
1448 			continue;
1449 		}
1450 		p = pt->pt_proc;
1451 		if (pt->pt_ev.sigev_notify != SIGEV_SIGNAL) {
1452 			continue;
1453 		}
1454 		if (sigismember(&p->p_sigpend.sp_set, pt->pt_ev.sigev_signo)) {
1455 			pt->pt_overruns++;
1456 			continue;
1457 		}
1458 
1459 		KSI_INIT(&ksi);
1460 		ksi.ksi_signo = pt->pt_ev.sigev_signo;
1461 		ksi.ksi_code = SI_TIMER;
1462 		ksi.ksi_value = pt->pt_ev.sigev_value;
1463 		pt->pt_poverruns = pt->pt_overruns;
1464 		pt->pt_overruns = 0;
1465 		mutex_spin_exit(&timer_lock);
1466 		kpsignal(p, &ksi, NULL);
1467 		mutex_spin_enter(&timer_lock);
1468 	}
1469 	mutex_spin_exit(&timer_lock);
1470 	mutex_exit(proc_lock);
1471 }
1472 
1473 /*
1474  * Check if the time will wrap if set to ts.
1475  *
1476  * ts - timespec describing the new time
1477  * delta - the delta between the current time and ts
1478  */
1479 bool
1480 time_wraps(struct timespec *ts, struct timespec *delta)
1481 {
1482 
1483 	/*
1484 	 * Don't allow the time to be set forward so far it
1485 	 * will wrap and become negative, thus allowing an
1486 	 * attacker to bypass the next check below.  The
1487 	 * cutoff is 1 year before rollover occurs, so even
1488 	 * if the attacker uses adjtime(2) to move the time
1489 	 * past the cutoff, it will take a very long time
1490 	 * to get to the wrap point.
1491 	 */
1492 	if ((ts->tv_sec > LLONG_MAX - 365*24*60*60) ||
1493 	    (delta->tv_sec < 0 || delta->tv_nsec < 0))
1494 		return true;
1495 
1496 	return false;
1497 }
1498