posix-timers.c 36.2 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2
 * linux/kernel/posix-timers.c
L
Linus Torvalds 已提交
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
 *
 *
 * 2002-10-15  Posix Clocks & timers
 *                           by George Anzinger george@mvista.com
 *
 *			     Copyright (C) 2002 2003 by MontaVista Software.
 *
 * 2004-06-01  Fix CLOCK_REALTIME clock/timer TIMER_ABSTIME bug.
 *			     Copyright (C) 2004 Boris Hu
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or (at
 * your option) any later version.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * General Public License for more details.

 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *
 * MontaVista Software | 1237 East Arques Avenue | Sunnyvale | CA 94085 | USA
 */

/* These are all the functions necessary to implement
 * POSIX clocks & timers
 */
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/slab.h>
#include <linux/time.h>
A
Arjan van de Ven 已提交
37
#include <linux/mutex.h>
38
#include <linux/sched/task.h>
L
Linus Torvalds 已提交
39

40
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
41 42 43
#include <linux/list.h>
#include <linux/init.h>
#include <linux/compiler.h>
44
#include <linux/hash.h>
45
#include <linux/posix-clock.h>
L
Linus Torvalds 已提交
46 47 48 49
#include <linux/posix-timers.h>
#include <linux/syscalls.h>
#include <linux/wait.h>
#include <linux/workqueue.h>
50
#include <linux/export.h>
51
#include <linux/hashtable.h>
52
#include <linux/compat.h>
L
Linus Torvalds 已提交
53

54
#include "timekeeping.h"
55
#include "posix-timers.h"
56

L
Linus Torvalds 已提交
57
/*
58 59 60 61 62 63
 * Management arrays for POSIX timers. Timers are now kept in static hash table
 * with 512 entries.
 * Timer ids are allocated by local routine, which selects proper hash head by
 * key, constructed from current->signal address and per signal struct counter.
 * This keeps timer ids unique per process, but now they can intersect between
 * processes.
L
Linus Torvalds 已提交
64 65 66 67 68
 */

/*
 * Lets keep our timers in a slab cache :-)
 */
69
static struct kmem_cache *posix_timers_cache;
70 71 72

static DEFINE_HASHTABLE(posix_timers_hashtable, 9);
static DEFINE_SPINLOCK(hash_lock);
L
Linus Torvalds 已提交
73

74 75
static const struct k_clock * const posix_clocks[];
static const struct k_clock *clockid_to_kclock(const clockid_t id);
76
static const struct k_clock clock_realtime, clock_monotonic;
77

L
Linus Torvalds 已提交
78 79 80 81 82 83 84 85 86
/*
 * we assume that the new SIGEV_THREAD_ID shares no bits with the other
 * SIGEV values.  Here we put out an error if this assumption fails.
 */
#if SIGEV_THREAD_ID != (SIGEV_THREAD_ID & \
                       ~(SIGEV_SIGNAL | SIGEV_NONE | SIGEV_THREAD))
#error "SIGEV_THREAD_ID must not share bit with other SIGEV values!"
#endif

87 88 89 90 91 92 93 94
/*
 * parisc wants ENOTSUP instead of EOPNOTSUPP
 */
#ifndef ENOTSUP
# define ENANOSLEEP_NOTSUP EOPNOTSUPP
#else
# define ENANOSLEEP_NOTSUP ENOTSUP
#endif
L
Linus Torvalds 已提交
95 96 97 98 99 100 101 102 103 104 105 106 107

/*
 * The timer ID is turned into a timer address by idr_find().
 * Verifying a valid ID consists of:
 *
 * a) checking that idr_find() returns other than -1.
 * b) checking that the timer id matches the one in the timer itself.
 * c) that the timer owner is in the callers thread group.
 */

/*
 * CLOCKs: The POSIX standard calls for a couple of clocks and allows us
 *	    to implement others.  This structure defines the various
R
Richard Cochran 已提交
108
 *	    clocks.
L
Linus Torvalds 已提交
109 110 111 112 113 114 115 116 117
 *
 * RESOLUTION: Clock resolution is used to round up timer and interval
 *	    times, NOT to report clock times, which are reported with as
 *	    much resolution as the system can muster.  In some cases this
 *	    resolution may depend on the underlying clock hardware and
 *	    may not be quantifiable until run time, and only then is the
 *	    necessary code is written.	The standard says we should say
 *	    something about this issue in the documentation...
 *
R
Richard Cochran 已提交
118 119
 * FUNCTIONS: The CLOCKs structure defines possible functions to
 *	    handle various clock functions.
L
Linus Torvalds 已提交
120
 *
R
Richard Cochran 已提交
121 122 123 124
 *	    The standard POSIX timer management code assumes the
 *	    following: 1.) The k_itimer struct (sched.h) is used for
 *	    the timer.  2.) The list, it_lock, it_clock, it_id and
 *	    it_pid fields are not modified by timer code.
L
Linus Torvalds 已提交
125 126 127 128 129 130 131 132
 *
 * Permissions: It is assumed that the clock_settime() function defined
 *	    for each clock will take care of permission checks.	 Some
 *	    clocks may be set able by any user (i.e. local process
 *	    clocks) others not.	 Currently the only set able clock we
 *	    have is CLOCK_REALTIME and its high res counter part, both of
 *	    which we beg off on and pass to do_sys_settimeofday().
 */
N
Namhyung Kim 已提交
133 134 135 136 137 138 139
static struct k_itimer *__lock_timer(timer_t timer_id, unsigned long *flags);

#define lock_timer(tid, flags)						   \
({	struct k_itimer *__timr;					   \
	__cond_lock(&__timr->it_lock, __timr = __lock_timer(tid, flags));  \
	__timr;								   \
})
L
Linus Torvalds 已提交
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
static int hash(struct signal_struct *sig, unsigned int nr)
{
	return hash_32(hash32_ptr(sig) ^ nr, HASH_BITS(posix_timers_hashtable));
}

static struct k_itimer *__posix_timers_find(struct hlist_head *head,
					    struct signal_struct *sig,
					    timer_t id)
{
	struct k_itimer *timer;

	hlist_for_each_entry_rcu(timer, head, t_hash) {
		if ((timer->it_signal == sig) && (timer->it_id == id))
			return timer;
	}
	return NULL;
}

static struct k_itimer *posix_timer_by_id(timer_t id)
{
	struct signal_struct *sig = current->signal;
	struct hlist_head *head = &posix_timers_hashtable[hash(sig, id)];

	return __posix_timers_find(head, sig, id);
}

static int posix_timer_add(struct k_itimer *timer)
{
	struct signal_struct *sig = current->signal;
	int first_free_id = sig->posix_timer_id;
	struct hlist_head *head;
	int ret = -ENOENT;

	do {
		spin_lock(&hash_lock);
		head = &posix_timers_hashtable[hash(sig, sig->posix_timer_id)];
		if (!__posix_timers_find(head, sig, sig->posix_timer_id)) {
			hlist_add_head_rcu(&timer->t_hash, head);
			ret = sig->posix_timer_id;
		}
		if (++sig->posix_timer_id < 0)
			sig->posix_timer_id = 0;
		if ((sig->posix_timer_id == first_free_id) && (ret == -ENOENT))
			/* Loop over all possible ids completed */
			ret = -EAGAIN;
		spin_unlock(&hash_lock);
	} while (ret == -ENOENT);
	return ret;
}

L
Linus Torvalds 已提交
191 192 193 194 195
static inline void unlock_timer(struct k_itimer *timr, unsigned long flags)
{
	spin_unlock_irqrestore(&timr->it_lock, flags);
}

196
/* Get clock_realtime */
197
static int posix_clock_realtime_get(clockid_t which_clock, struct timespec64 *tp)
198
{
199
	ktime_get_real_ts64(tp);
200 201 202
	return 0;
}

203 204
/* Set clock_realtime */
static int posix_clock_realtime_set(const clockid_t which_clock,
205
				    const struct timespec64 *tp)
206
{
207
	return do_sys_settimeofday64(tp, NULL);
208 209
}

210 211 212 213 214 215
static int posix_clock_realtime_adj(const clockid_t which_clock,
				    struct timex *t)
{
	return do_adjtimex(t);
}

216 217 218
/*
 * Get monotonic time for posix timers
 */
219
static int posix_ktime_get_ts(clockid_t which_clock, struct timespec64 *tp)
220
{
221
	ktime_get_ts64(tp);
222 223
	return 0;
}
L
Linus Torvalds 已提交
224

225
/*
226
 * Get monotonic-raw time for posix timers
227
 */
228
static int posix_get_monotonic_raw(clockid_t which_clock, struct timespec64 *tp)
229
{
230
	getrawmonotonic64(tp);
231 232 233
	return 0;
}

234

235
static int posix_get_realtime_coarse(clockid_t which_clock, struct timespec64 *tp)
236
{
237
	*tp = current_kernel_time64();
238 239 240 241
	return 0;
}

static int posix_get_monotonic_coarse(clockid_t which_clock,
242
						struct timespec64 *tp)
243
{
244
	*tp = get_monotonic_coarse64();
245 246 247
	return 0;
}

248
static int posix_get_coarse_res(const clockid_t which_clock, struct timespec64 *tp)
249
{
250
	*tp = ktime_to_timespec64(KTIME_LOW_RES);
251 252
	return 0;
}
253

254
static int posix_get_boottime(const clockid_t which_clock, struct timespec64 *tp)
255
{
256
	get_monotonic_boottime64(tp);
257 258 259
	return 0;
}

260
static int posix_get_tai(clockid_t which_clock, struct timespec64 *tp)
J
John Stultz 已提交
261
{
262
	timekeeping_clocktai64(tp);
J
John Stultz 已提交
263 264
	return 0;
}
265

266
static int posix_get_hrtimer_res(clockid_t which_clock, struct timespec64 *tp)
267 268 269 270 271 272
{
	tp->tv_sec = 0;
	tp->tv_nsec = hrtimer_resolution;
	return 0;
}

L
Linus Torvalds 已提交
273 274 275 276 277 278
/*
 * Initialize everything, well, just everything in Posix clocks/timers ;)
 */
static __init int init_posix_timers(void)
{
	posix_timers_cache = kmem_cache_create("posix_timers_cache",
279 280
					sizeof (struct k_itimer), 0, SLAB_PANIC,
					NULL);
L
Linus Torvalds 已提交
281 282 283 284
	return 0;
}
__initcall(init_posix_timers);

285
static void common_hrtimer_rearm(struct k_itimer *timr)
L
Linus Torvalds 已提交
286
{
287 288
	struct hrtimer *timer = &timr->it.real.timer;

289
	if (!timr->it_interval)
L
Linus Torvalds 已提交
290 291
		return;

D
Davide Libenzi 已提交
292 293
	timr->it_overrun += (unsigned int) hrtimer_forward(timer,
						timer->base->get_time(),
294
						timr->it_interval);
295
	hrtimer_restart(timer);
L
Linus Torvalds 已提交
296 297 298 299 300 301 302 303 304 305
}

/*
 * This function is exported for use by the signal deliver code.  It is
 * called just prior to the info block being released and passes that
 * block to us.  It's function is to update the overrun entry AND to
 * restart the timer.  It should only be called if the timer is to be
 * restarted (i.e. we have flagged this in the sys_private entry of the
 * info block).
 *
L
Lucas De Marchi 已提交
306
 * To protect against the timer going away while the interrupt is queued,
L
Linus Torvalds 已提交
307 308
 * we require that the it_requeue_pending flag be set.
 */
309
void posixtimer_rearm(struct siginfo *info)
L
Linus Torvalds 已提交
310 311 312 313 314
{
	struct k_itimer *timr;
	unsigned long flags;

	timr = lock_timer(info->si_tid, &flags);
315 316
	if (!timr)
		return;
L
Linus Torvalds 已提交
317

318
	if (timr->it_requeue_pending == info->si_sys_private) {
319
		timr->kclock->timer_rearm(timr);
L
Linus Torvalds 已提交
320

321
		timr->it_active = 1;
322 323 324 325
		timr->it_overrun_last = timr->it_overrun;
		timr->it_overrun = -1;
		++timr->it_requeue_pending;

326
		info->si_overrun += timr->it_overrun_last;
327 328
	}

329
	unlock_timer(timr, flags);
L
Linus Torvalds 已提交
330 331
}

332
int posix_timer_event(struct k_itimer *timr, int si_private)
L
Linus Torvalds 已提交
333
{
334 335
	struct task_struct *task;
	int shared, ret = -1;
336 337
	/*
	 * FIXME: if ->sigq is queued we can race with
338
	 * dequeue_signal()->posixtimer_rearm().
339 340
	 *
	 * If dequeue_signal() sees the "right" value of
341
	 * si_sys_private it calls posixtimer_rearm().
342
	 * We re-queue ->sigq and drop ->it_lock().
343
	 * posixtimer_rearm() locks the timer
344 345 346
	 * and re-schedules it while ->sigq is pending.
	 * Not really bad, but not that we want.
	 */
L
Linus Torvalds 已提交
347 348
	timr->sigq->info.si_sys_private = si_private;

349 350 351 352 353 354 355
	rcu_read_lock();
	task = pid_task(timr->it_pid, PIDTYPE_PID);
	if (task) {
		shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID);
		ret = send_sigqueue(timr->sigq, task, shared);
	}
	rcu_read_unlock();
356 357
	/* If we failed to send the signal the timer stops. */
	return ret > 0;
L
Linus Torvalds 已提交
358 359 360 361 362 363 364 365 366
}

/*
 * This function gets called when a POSIX.1b interval timer expires.  It
 * is used as a callback from the kernel internal timer.  The
 * run_timer_list code ALWAYS calls with interrupts on.

 * This code is for CLOCK_REALTIME* and CLOCK_MONOTONIC* timers.
 */
367
static enum hrtimer_restart posix_timer_fn(struct hrtimer *timer)
L
Linus Torvalds 已提交
368
{
369
	struct k_itimer *timr;
L
Linus Torvalds 已提交
370
	unsigned long flags;
371
	int si_private = 0;
372
	enum hrtimer_restart ret = HRTIMER_NORESTART;
L
Linus Torvalds 已提交
373

374
	timr = container_of(timer, struct k_itimer, it.real.timer);
L
Linus Torvalds 已提交
375 376
	spin_lock_irqsave(&timr->it_lock, flags);

377
	timr->it_active = 0;
378
	if (timr->it_interval != 0)
379
		si_private = ++timr->it_requeue_pending;
L
Linus Torvalds 已提交
380

381 382 383 384 385 386
	if (posix_timer_event(timr, si_private)) {
		/*
		 * signal was not sent because of sig_ignor
		 * we will not get a call back to restart it AND
		 * it should be restarted.
		 */
387
		if (timr->it_interval != 0) {
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
			ktime_t now = hrtimer_cb_get_time(timer);

			/*
			 * FIXME: What we really want, is to stop this
			 * timer completely and restart it in case the
			 * SIG_IGN is removed. This is a non trivial
			 * change which involves sighand locking
			 * (sigh !), which we don't want to do late in
			 * the release cycle.
			 *
			 * For now we just let timers with an interval
			 * less than a jiffie expire every jiffie to
			 * avoid softirq starvation in case of SIG_IGN
			 * and a very small interval, which would put
			 * the timer right back on the softirq pending
			 * list. By moving now ahead of time we trick
			 * hrtimer_forward() to expire the timer
			 * later, while we still maintain the overrun
			 * accuracy, but have some inconsistency in
			 * the timer_gettime() case. This is at least
			 * better than a starved softirq. A more
			 * complex fix which solves also another related
			 * inconsistency is already in the pipeline.
			 */
#ifdef CONFIG_HIGH_RES_TIMERS
			{
T
Thomas Gleixner 已提交
414
				ktime_t kj = NSEC_PER_SEC / HZ;
415

416
				if (timr->it_interval < kj)
417 418 419
					now = ktime_add(now, kj);
			}
#endif
D
Davide Libenzi 已提交
420
			timr->it_overrun += (unsigned int)
421
				hrtimer_forward(timer, now,
422
						timr->it_interval);
423
			ret = HRTIMER_RESTART;
424
			++timr->it_requeue_pending;
425
			timr->it_active = 1;
L
Linus Torvalds 已提交
426 427 428
		}
	}

429 430 431
	unlock_timer(timr, flags);
	return ret;
}
L
Linus Torvalds 已提交
432

433
static struct pid *good_sigevent(sigevent_t * event)
L
Linus Torvalds 已提交
434 435 436 437
{
	struct task_struct *rtn = current->group_leader;

	if ((event->sigev_notify & SIGEV_THREAD_ID ) &&
438
		(!(rtn = find_task_by_vpid(event->sigev_notify_thread_id)) ||
439
		 !same_thread_group(rtn, current) ||
L
Linus Torvalds 已提交
440 441 442 443 444 445 446
		 (event->sigev_notify & ~SIGEV_THREAD_ID) != SIGEV_SIGNAL))
		return NULL;

	if (((event->sigev_notify & ~SIGEV_THREAD_ID) != SIGEV_NONE) &&
	    ((event->sigev_signo <= 0) || (event->sigev_signo > SIGRTMAX)))
		return NULL;

447
	return task_pid(rtn);
L
Linus Torvalds 已提交
448 449 450 451 452
}

static struct k_itimer * alloc_posix_timer(void)
{
	struct k_itimer *tmr;
453
	tmr = kmem_cache_zalloc(posix_timers_cache, GFP_KERNEL);
L
Linus Torvalds 已提交
454 455 456 457
	if (!tmr)
		return tmr;
	if (unlikely(!(tmr->sigq = sigqueue_alloc()))) {
		kmem_cache_free(posix_timers_cache, tmr);
458
		return NULL;
L
Linus Torvalds 已提交
459
	}
460
	memset(&tmr->sigq->info, 0, sizeof(siginfo_t));
L
Linus Torvalds 已提交
461 462 463
	return tmr;
}

E
Eric Dumazet 已提交
464 465 466 467 468 469 470
static void k_itimer_rcu_free(struct rcu_head *head)
{
	struct k_itimer *tmr = container_of(head, struct k_itimer, it.rcu);

	kmem_cache_free(posix_timers_cache, tmr);
}

L
Linus Torvalds 已提交
471 472 473 474 475 476
#define IT_ID_SET	1
#define IT_ID_NOT_SET	0
static void release_posix_timer(struct k_itimer *tmr, int it_id_set)
{
	if (it_id_set) {
		unsigned long flags;
477 478 479
		spin_lock_irqsave(&hash_lock, flags);
		hlist_del_rcu(&tmr->t_hash);
		spin_unlock_irqrestore(&hash_lock, flags);
L
Linus Torvalds 已提交
480
	}
481
	put_pid(tmr->it_pid);
L
Linus Torvalds 已提交
482
	sigqueue_free(tmr->sigq);
E
Eric Dumazet 已提交
483
	call_rcu(&tmr->it.rcu, k_itimer_rcu_free);
L
Linus Torvalds 已提交
484 485
}

486 487 488 489 490 491
static int common_timer_create(struct k_itimer *new_timer)
{
	hrtimer_init(&new_timer->it.real.timer, new_timer->it_clock, 0);
	return 0;
}

L
Linus Torvalds 已提交
492
/* Create a POSIX.1b interval timer. */
493 494
static int do_timer_create(clockid_t which_clock, struct sigevent *event,
			   timer_t __user *created_timer_id)
L
Linus Torvalds 已提交
495
{
496
	const struct k_clock *kc = clockid_to_kclock(which_clock);
497
	struct k_itimer *new_timer;
498
	int error, new_timer_id;
L
Linus Torvalds 已提交
499 500
	int it_id_set = IT_ID_NOT_SET;

501
	if (!kc)
L
Linus Torvalds 已提交
502
		return -EINVAL;
503 504
	if (!kc->timer_create)
		return -EOPNOTSUPP;
L
Linus Torvalds 已提交
505 506 507 508 509 510

	new_timer = alloc_posix_timer();
	if (unlikely(!new_timer))
		return -EAGAIN;

	spin_lock_init(&new_timer->it_lock);
511 512 513
	new_timer_id = posix_timer_add(new_timer);
	if (new_timer_id < 0) {
		error = new_timer_id;
L
Linus Torvalds 已提交
514 515 516 517 518 519
		goto out;
	}

	it_id_set = IT_ID_SET;
	new_timer->it_id = (timer_t) new_timer_id;
	new_timer->it_clock = which_clock;
520
	new_timer->kclock = kc;
L
Linus Torvalds 已提交
521 522
	new_timer->it_overrun = -1;

523
	if (event) {
524
		rcu_read_lock();
525
		new_timer->it_pid = get_pid(good_sigevent(event));
526
		rcu_read_unlock();
527
		if (!new_timer->it_pid) {
L
Linus Torvalds 已提交
528 529 530
			error = -EINVAL;
			goto out;
		}
531 532 533
		new_timer->it_sigev_notify     = event->sigev_notify;
		new_timer->sigq->info.si_signo = event->sigev_signo;
		new_timer->sigq->info.si_value = event->sigev_value;
L
Linus Torvalds 已提交
534
	} else {
535 536 537 538
		new_timer->it_sigev_notify     = SIGEV_SIGNAL;
		new_timer->sigq->info.si_signo = SIGALRM;
		memset(&new_timer->sigq->info.si_value, 0, sizeof(sigval_t));
		new_timer->sigq->info.si_value.sival_int = new_timer->it_id;
539
		new_timer->it_pid = get_pid(task_tgid(current));
L
Linus Torvalds 已提交
540 541
	}

542
	new_timer->sigq->info.si_tid   = new_timer->it_id;
543
	new_timer->sigq->info.si_code  = SI_TIMER;
544

545 546 547 548 549 550
	if (copy_to_user(created_timer_id,
			 &new_timer_id, sizeof (new_timer_id))) {
		error = -EFAULT;
		goto out;
	}

551
	error = kc->timer_create(new_timer);
552 553 554
	if (error)
		goto out;

555
	spin_lock_irq(&current->sighand->siglock);
556
	new_timer->it_signal = current->signal;
557 558
	list_add(&new_timer->list, &current->signal->posix_timers);
	spin_unlock_irq(&current->sighand->siglock);
559 560

	return 0;
561
	/*
L
Linus Torvalds 已提交
562 563 564 565 566 567
	 * In the case of the timer belonging to another task, after
	 * the task is unlocked, the timer is owned by the other task
	 * and may cease to exist at any time.  Don't use or modify
	 * new_timer after the unlock call.
	 */
out:
568
	release_posix_timer(new_timer, it_id_set);
L
Linus Torvalds 已提交
569 570 571
	return error;
}

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
SYSCALL_DEFINE3(timer_create, const clockid_t, which_clock,
		struct sigevent __user *, timer_event_spec,
		timer_t __user *, created_timer_id)
{
	if (timer_event_spec) {
		sigevent_t event;

		if (copy_from_user(&event, timer_event_spec, sizeof (event)))
			return -EFAULT;
		return do_timer_create(which_clock, &event, created_timer_id);
	}
	return do_timer_create(which_clock, NULL, created_timer_id);
}

#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE3(timer_create, clockid_t, which_clock,
		       struct compat_sigevent __user *, timer_event_spec,
		       timer_t __user *, created_timer_id)
{
	if (timer_event_spec) {
		sigevent_t event;

		if (get_compat_sigevent(&event, timer_event_spec))
			return -EFAULT;
		return do_timer_create(which_clock, &event, created_timer_id);
	}
	return do_timer_create(which_clock, NULL, created_timer_id);
}
#endif

L
Linus Torvalds 已提交
602 603 604 605 606 607 608
/*
 * Locking issues: We need to protect the result of the id look up until
 * we get the timer locked down so it is not deleted under us.  The
 * removal is done under the idr spinlock so we use that here to bridge
 * the find to the timer lock.  To avoid a dead lock, the timer id MUST
 * be release with out holding the timer lock.
 */
N
Namhyung Kim 已提交
609
static struct k_itimer *__lock_timer(timer_t timer_id, unsigned long *flags)
L
Linus Torvalds 已提交
610 611
{
	struct k_itimer *timr;
E
Eric Dumazet 已提交
612

613 614 615 616 617 618 619
	/*
	 * timer_t could be any type >= int and we want to make sure any
	 * @timer_id outside positive int range fails lookup.
	 */
	if ((unsigned long long)timer_id > INT_MAX)
		return NULL;

E
Eric Dumazet 已提交
620
	rcu_read_lock();
621
	timr = posix_timer_by_id(timer_id);
L
Linus Torvalds 已提交
622
	if (timr) {
E
Eric Dumazet 已提交
623
		spin_lock_irqsave(&timr->it_lock, *flags);
624
		if (timr->it_signal == current->signal) {
E
Eric Dumazet 已提交
625
			rcu_read_unlock();
626 627
			return timr;
		}
E
Eric Dumazet 已提交
628
		spin_unlock_irqrestore(&timr->it_lock, *flags);
629
	}
E
Eric Dumazet 已提交
630
	rcu_read_unlock();
L
Linus Torvalds 已提交
631

632
	return NULL;
L
Linus Torvalds 已提交
633 634
}

635 636 637 638 639 640 641 642 643 644 645 646 647 648
static ktime_t common_hrtimer_remaining(struct k_itimer *timr, ktime_t now)
{
	struct hrtimer *timer = &timr->it.real.timer;

	return __hrtimer_expires_remaining_adjusted(timer, now);
}

static int common_hrtimer_forward(struct k_itimer *timr, ktime_t now)
{
	struct hrtimer *timer = &timr->it.real.timer;

	return (int)hrtimer_forward(timer, now, timr->it_interval);
}

L
Linus Torvalds 已提交
649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
/*
 * Get the time remaining on a POSIX.1b interval timer.  This function
 * is ALWAYS called with spin_lock_irq on the timer, thus it must not
 * mess with irq.
 *
 * We have a couple of messes to clean up here.  First there is the case
 * of a timer that has a requeue pending.  These timers should appear to
 * be in the timer list with an expiry as if we were to requeue them
 * now.
 *
 * The second issue is the SIGEV_NONE timer which may be active but is
 * not really ever put in the timer list (to save system resources).
 * This timer may be expired, and if so, we will do it here.  Otherwise
 * it is the same as a requeue pending timer WRT to what we should
 * report.
 */
665
void common_timer_get(struct k_itimer *timr, struct itimerspec64 *cur_setting)
L
Linus Torvalds 已提交
666
{
667
	const struct k_clock *kc = timr->kclock;
668
	ktime_t now, remaining, iv;
669 670
	struct timespec64 ts64;
	bool sig_none;
L
Linus Torvalds 已提交
671

672
	sig_none = (timr->it_sigev_notify & ~SIGEV_THREAD_ID) == SIGEV_NONE;
673
	iv = timr->it_interval;
674

675
	/* interval timer ? */
676
	if (iv) {
677
		cur_setting->it_interval = ktime_to_timespec64(iv);
678 679 680 681 682 683 684 685
	} else if (!timr->it_active) {
		/*
		 * SIGEV_NONE oneshot timers are never queued. Check them
		 * below.
		 */
		if (!sig_none)
			return;
	}
686

687 688 689 690 691 692
	/*
	 * The timespec64 based conversion is suboptimal, but it's not
	 * worth to implement yet another callback.
	 */
	kc->clock_get(timr->it_clock, &ts64);
	now = timespec64_to_ktime(ts64);
693

694
	/*
695 696
	 * When a requeue is pending or this is a SIGEV_NONE timer move the
	 * expiry time forward by intervals, so expiry is > now.
697
	 */
698 699
	if (iv && (timr->it_requeue_pending & REQUEUE_PENDING || sig_none))
		timr->it_overrun += kc->timer_forward(timr, now);
700

701
	remaining = kc->timer_remaining(timr, now);
702
	/* Return 0 only, when the timer is expired and not pending */
T
Thomas Gleixner 已提交
703
	if (remaining <= 0) {
704 705 706 707
		/*
		 * A single shot SIGEV_NONE timer must return 0, when
		 * it is expired !
		 */
708
		if (!sig_none)
709
			cur_setting->it_value.tv_nsec = 1;
710
	} else {
711
		cur_setting->it_value = ktime_to_timespec64(remaining);
712
	}
L
Linus Torvalds 已提交
713 714 715
}

/* Get the time remaining on a POSIX.1b interval timer. */
716
static int do_timer_gettime(timer_t timer_id,  struct itimerspec64 *setting)
L
Linus Torvalds 已提交
717
{
718
	struct k_itimer *timr;
719
	const struct k_clock *kc;
L
Linus Torvalds 已提交
720
	unsigned long flags;
721
	int ret = 0;
L
Linus Torvalds 已提交
722 723 724 725 726

	timr = lock_timer(timer_id, &flags);
	if (!timr)
		return -EINVAL;

727
	memset(setting, 0, sizeof(*setting));
728
	kc = timr->kclock;
729 730 731
	if (WARN_ON_ONCE(!kc || !kc->timer_get))
		ret = -EINVAL;
	else
732
		kc->timer_get(timr, setting);
L
Linus Torvalds 已提交
733 734

	unlock_timer(timr, flags);
735 736
	return ret;
}
L
Linus Torvalds 已提交
737

738 739 740 741 742
/* Get the time remaining on a POSIX.1b interval timer. */
SYSCALL_DEFINE2(timer_gettime, timer_t, timer_id,
		struct itimerspec __user *, setting)
{
	struct itimerspec64 cur_setting64;
L
Linus Torvalds 已提交
743

744 745 746 747 748 749 750
	int ret = do_timer_gettime(timer_id, &cur_setting64);
	if (!ret) {
		struct itimerspec cur_setting;
		cur_setting = itimerspec64_to_itimerspec(&cur_setting64);
		if (copy_to_user(setting, &cur_setting, sizeof (cur_setting)))
			ret = -EFAULT;
	}
751
	return ret;
L
Linus Torvalds 已提交
752
}
753

754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770
#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE2(timer_gettime, timer_t, timer_id,
		       struct compat_itimerspec __user *, setting)
{
	struct itimerspec64 cur_setting64;

	int ret = do_timer_gettime(timer_id, &cur_setting64);
	if (!ret) {
		struct itimerspec cur_setting;
		cur_setting = itimerspec64_to_itimerspec(&cur_setting64);
		if (put_compat_itimerspec(setting, &cur_setting))
			ret = -EFAULT;
	}
	return ret;
}
#endif

L
Linus Torvalds 已提交
771 772 773 774 775 776
/*
 * Get the number of overruns of a POSIX.1b interval timer.  This is to
 * be the overrun of the timer last delivered.  At the same time we are
 * accumulating overruns on the next timer.  The overrun is frozen when
 * the signal is delivered, either at the notify time (if the info block
 * is not queued) or at the actual delivery time (as we are informed by
777
 * the call back to posixtimer_rearm().  So all we need to do is
L
Linus Torvalds 已提交
778 779
 * to pick up the frozen overrun.
 */
780
SYSCALL_DEFINE1(timer_getoverrun, timer_t, timer_id)
L
Linus Torvalds 已提交
781 782 783
{
	struct k_itimer *timr;
	int overrun;
784
	unsigned long flags;
L
Linus Torvalds 已提交
785 786 787 788 789 790 791 792 793 794 795

	timr = lock_timer(timer_id, &flags);
	if (!timr)
		return -EINVAL;

	overrun = timr->it_overrun_last;
	unlock_timer(timr, flags);

	return overrun;
}

796 797 798 799 800 801 802
static void common_hrtimer_arm(struct k_itimer *timr, ktime_t expires,
			       bool absolute, bool sigev_none)
{
	struct hrtimer *timer = &timr->it.real.timer;
	enum hrtimer_mode mode;

	mode = absolute ? HRTIMER_MODE_ABS : HRTIMER_MODE_REL;
803 804 805 806 807 808 809 810 811 812 813 814
	/*
	 * Posix magic: Relative CLOCK_REALTIME timers are not affected by
	 * clock modifications, so they become CLOCK_MONOTONIC based under the
	 * hood. See hrtimer_init(). Update timr->kclock, so the generic
	 * functions which use timr->kclock->clock_get() work.
	 *
	 * Note: it_clock stays unmodified, because the next timer_set() might
	 * use ABSTIME, so it needs to switch back.
	 */
	if (timr->it_clock == CLOCK_REALTIME)
		timr->kclock = absolute ? &clock_realtime : &clock_monotonic;

815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830
	hrtimer_init(&timr->it.real.timer, timr->it_clock, mode);
	timr->it.real.timer.function = posix_timer_fn;

	if (!absolute)
		expires = ktime_add_safe(expires, timer->base->get_time());
	hrtimer_set_expires(timer, expires);

	if (!sigev_none)
		hrtimer_start_expires(timer, HRTIMER_MODE_ABS);
}

static int common_hrtimer_try_to_cancel(struct k_itimer *timr)
{
	return hrtimer_try_to_cancel(&timr->it.real.timer);
}

L
Linus Torvalds 已提交
831
/* Set a POSIX.1b interval timer. */
832 833 834
int common_timer_set(struct k_itimer *timr, int flags,
		     struct itimerspec64 *new_setting,
		     struct itimerspec64 *old_setting)
L
Linus Torvalds 已提交
835
{
836 837 838
	const struct k_clock *kc = timr->kclock;
	bool sigev_none;
	ktime_t expires;
L
Linus Torvalds 已提交
839 840 841 842

	if (old_setting)
		common_timer_get(timr, old_setting);

843
	/* Prevent rearming by clearing the interval */
844
	timr->it_interval = 0;
L
Linus Torvalds 已提交
845
	/*
846 847
	 * Careful here. On SMP systems the timer expiry function could be
	 * active and spinning on timr->it_lock.
L
Linus Torvalds 已提交
848
	 */
849
	if (kc->timer_try_to_cancel(timr) < 0)
L
Linus Torvalds 已提交
850 851
		return TIMER_RETRY;

852 853
	timr->it_active = 0;
	timr->it_requeue_pending = (timr->it_requeue_pending + 2) &
L
Linus Torvalds 已提交
854 855 856
		~REQUEUE_PENDING;
	timr->it_overrun_last = 0;

857
	/* Switch off the timer when it_value is zero */
858 859
	if (!new_setting->it_value.tv_sec && !new_setting->it_value.tv_nsec)
		return 0;
L
Linus Torvalds 已提交
860

861
	timr->it_interval = timespec64_to_ktime(new_setting->it_interval);
862 863
	expires = timespec64_to_ktime(new_setting->it_value);
	sigev_none = (timr->it_sigev_notify & ~SIGEV_THREAD_ID) == SIGEV_NONE;
864

865 866
	kc->timer_arm(timr, expires, flags & TIMER_ABSTIME, sigev_none);
	timr->it_active = !sigev_none;
L
Linus Torvalds 已提交
867 868 869
	return 0;
}

870 871 872
static int do_timer_settime(timer_t timer_id, int flags,
			    struct itimerspec64 *new_spec64,
			    struct itimerspec64 *old_spec64)
L
Linus Torvalds 已提交
873
{
874
	const struct k_clock *kc;
875
	struct k_itimer *timr;
876
	unsigned long flag;
877
	int error = 0;
L
Linus Torvalds 已提交
878

879 880
	if (!timespec64_valid(&new_spec64->it_interval) ||
	    !timespec64_valid(&new_spec64->it_value))
L
Linus Torvalds 已提交
881 882
		return -EINVAL;

883 884
	if (old_spec64)
		memset(old_spec64, 0, sizeof(*old_spec64));
L
Linus Torvalds 已提交
885 886 887 888 889
retry:
	timr = lock_timer(timer_id, &flag);
	if (!timr)
		return -EINVAL;

890
	kc = timr->kclock;
891 892 893
	if (WARN_ON_ONCE(!kc || !kc->timer_set))
		error = -EINVAL;
	else
894
		error = kc->timer_set(timr, flags, new_spec64, old_spec64);
L
Linus Torvalds 已提交
895 896 897

	unlock_timer(timr, flag);
	if (error == TIMER_RETRY) {
898
		old_spec64 = NULL;	// We already got the old time...
L
Linus Torvalds 已提交
899 900 901
		goto retry;
	}

902 903
	return error;
}
L
Linus Torvalds 已提交
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
/* Set a POSIX.1b interval timer */
SYSCALL_DEFINE4(timer_settime, timer_t, timer_id, int, flags,
		const struct itimerspec __user *, new_setting,
		struct itimerspec __user *, old_setting)
{
	struct itimerspec64 new_spec64, old_spec64;
	struct itimerspec64 *rtn = old_setting ? &old_spec64 : NULL;
	struct itimerspec new_spec;
	int error = 0;

	if (!new_setting)
		return -EINVAL;

	if (copy_from_user(&new_spec, new_setting, sizeof (new_spec)))
		return -EFAULT;
	new_spec64 = itimerspec_to_itimerspec64(&new_spec);

	error = do_timer_settime(timer_id, flags, &new_spec64, rtn);
	if (!error && old_setting) {
		struct itimerspec old_spec;
		old_spec = itimerspec64_to_itimerspec(&old_spec64);
		if (copy_to_user(old_setting, &old_spec, sizeof (old_spec)))
			error = -EFAULT;
	}
	return error;
}

#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE4(timer_settime, timer_t, timer_id, int, flags,
		       struct compat_itimerspec __user *, new,
		       struct compat_itimerspec __user *, old)
{
	struct itimerspec64 new_spec64, old_spec64;
	struct itimerspec64 *rtn = old ? &old_spec64 : NULL;
	struct itimerspec new_spec;
	int error = 0;

	if (!new)
		return -EINVAL;
	if (get_compat_itimerspec(&new_spec, new))
		return -EFAULT;

	new_spec64 = itimerspec_to_itimerspec64(&new_spec);
	error = do_timer_settime(timer_id, flags, &new_spec64, rtn);
	if (!error && old) {
		struct itimerspec old_spec;
		old_spec = itimerspec64_to_itimerspec(&old_spec64);
		if (put_compat_itimerspec(old, &old_spec))
			error = -EFAULT;
	}
L
Linus Torvalds 已提交
955 956
	return error;
}
957
#endif
L
Linus Torvalds 已提交
958

959
int common_timer_del(struct k_itimer *timer)
L
Linus Torvalds 已提交
960
{
961
	const struct k_clock *kc = timer->kclock;
962

963 964
	timer->it_interval = 0;
	if (kc->timer_try_to_cancel(timer) < 0)
L
Linus Torvalds 已提交
965
		return TIMER_RETRY;
966
	timer->it_active = 0;
L
Linus Torvalds 已提交
967 968 969 970 971
	return 0;
}

static inline int timer_delete_hook(struct k_itimer *timer)
{
972
	const struct k_clock *kc = timer->kclock;
973 974 975 976

	if (WARN_ON_ONCE(!kc || !kc->timer_del))
		return -EINVAL;
	return kc->timer_del(timer);
L
Linus Torvalds 已提交
977 978 979
}

/* Delete a POSIX.1b interval timer. */
980
SYSCALL_DEFINE1(timer_delete, timer_t, timer_id)
L
Linus Torvalds 已提交
981 982
{
	struct k_itimer *timer;
983
	unsigned long flags;
L
Linus Torvalds 已提交
984 985 986 987 988 989

retry_delete:
	timer = lock_timer(timer_id, &flags);
	if (!timer)
		return -EINVAL;

990
	if (timer_delete_hook(timer) == TIMER_RETRY) {
L
Linus Torvalds 已提交
991 992 993
		unlock_timer(timer, flags);
		goto retry_delete;
	}
994

L
Linus Torvalds 已提交
995 996 997 998 999 1000 1001
	spin_lock(&current->sighand->siglock);
	list_del(&timer->list);
	spin_unlock(&current->sighand->siglock);
	/*
	 * This keeps any tasks waiting on the spin lock from thinking
	 * they got something (see the lock code above).
	 */
1002
	timer->it_signal = NULL;
1003

L
Linus Torvalds 已提交
1004 1005 1006 1007
	unlock_timer(timer, flags);
	release_posix_timer(timer, IT_ID_SET);
	return 0;
}
1008

L
Linus Torvalds 已提交
1009 1010 1011
/*
 * return timer owned by the process, used by exit_itimers
 */
1012
static void itimer_delete(struct k_itimer *timer)
L
Linus Torvalds 已提交
1013 1014 1015 1016 1017 1018
{
	unsigned long flags;

retry_delete:
	spin_lock_irqsave(&timer->it_lock, flags);

1019
	if (timer_delete_hook(timer) == TIMER_RETRY) {
L
Linus Torvalds 已提交
1020 1021 1022 1023 1024 1025 1026 1027
		unlock_timer(timer, flags);
		goto retry_delete;
	}
	list_del(&timer->list);
	/*
	 * This keeps any tasks waiting on the spin lock from thinking
	 * they got something (see the lock code above).
	 */
1028
	timer->it_signal = NULL;
1029

L
Linus Torvalds 已提交
1030 1031 1032 1033 1034
	unlock_timer(timer, flags);
	release_posix_timer(timer, IT_ID_SET);
}

/*
1035
 * This is called by do_exit or de_thread, only when there are no more
L
Linus Torvalds 已提交
1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047
 * references to the shared signal_struct.
 */
void exit_itimers(struct signal_struct *sig)
{
	struct k_itimer *tmr;

	while (!list_empty(&sig->posix_timers)) {
		tmr = list_entry(sig->posix_timers.next, struct k_itimer, list);
		itimer_delete(tmr);
	}
}

1048 1049
SYSCALL_DEFINE2(clock_settime, const clockid_t, which_clock,
		const struct timespec __user *, tp)
L
Linus Torvalds 已提交
1050
{
1051
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1052
	struct timespec64 new_tp;
L
Linus Torvalds 已提交
1053

1054
	if (!kc || !kc->clock_set)
L
Linus Torvalds 已提交
1055
		return -EINVAL;
1056

1057
	if (get_timespec64(&new_tp, tp))
L
Linus Torvalds 已提交
1058 1059
		return -EFAULT;

1060
	return kc->clock_set(which_clock, &new_tp);
L
Linus Torvalds 已提交
1061 1062
}

1063 1064
SYSCALL_DEFINE2(clock_gettime, const clockid_t, which_clock,
		struct timespec __user *,tp)
L
Linus Torvalds 已提交
1065
{
1066
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1067
	struct timespec64 kernel_tp;
L
Linus Torvalds 已提交
1068 1069
	int error;

1070
	if (!kc)
L
Linus Torvalds 已提交
1071
		return -EINVAL;
1072

1073
	error = kc->clock_get(which_clock, &kernel_tp);
1074

1075
	if (!error && put_timespec64(&kernel_tp, tp))
L
Linus Torvalds 已提交
1076 1077 1078 1079 1080
		error = -EFAULT;

	return error;
}

1081 1082 1083
SYSCALL_DEFINE2(clock_adjtime, const clockid_t, which_clock,
		struct timex __user *, utx)
{
1084
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097
	struct timex ktx;
	int err;

	if (!kc)
		return -EINVAL;
	if (!kc->clock_adj)
		return -EOPNOTSUPP;

	if (copy_from_user(&ktx, utx, sizeof(ktx)))
		return -EFAULT;

	err = kc->clock_adj(which_clock, &ktx);

1098
	if (err >= 0 && copy_to_user(utx, &ktx, sizeof(ktx)))
1099 1100 1101 1102 1103
		return -EFAULT;

	return err;
}

1104 1105 1106 1107
SYSCALL_DEFINE2(clock_getres, const clockid_t, which_clock,
		struct timespec __user *, tp)
{
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1108
	struct timespec64 rtn_tp;
1109 1110 1111 1112 1113
	int error;

	if (!kc)
		return -EINVAL;

1114
	error = kc->clock_getres(which_clock, &rtn_tp);
1115

1116
	if (!error && tp && put_timespec64(&rtn_tp, tp))
1117 1118 1119 1120 1121
		error = -EFAULT;

	return error;
}

1122 1123
#ifdef CONFIG_COMPAT

1124 1125 1126 1127
COMPAT_SYSCALL_DEFINE2(clock_settime, clockid_t, which_clock,
		       struct compat_timespec __user *, tp)
{
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1128
	struct timespec64 ts;
1129 1130 1131 1132

	if (!kc || !kc->clock_set)
		return -EINVAL;

1133
	if (compat_get_timespec64(&ts, tp))
1134 1135
		return -EFAULT;

1136
	return kc->clock_set(which_clock, &ts);
1137 1138 1139 1140 1141 1142
}

COMPAT_SYSCALL_DEFINE2(clock_gettime, clockid_t, which_clock,
		       struct compat_timespec __user *, tp)
{
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1143 1144
	struct timespec64 ts;
	int err;
1145 1146 1147 1148

	if (!kc)
		return -EINVAL;

1149
	err = kc->clock_get(which_clock, &ts);
1150

1151 1152
	if (!err && compat_put_timespec64(&ts, tp))
		err = -EFAULT;
1153

1154
	return err;
1155 1156
}

1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
COMPAT_SYSCALL_DEFINE2(clock_adjtime, clockid_t, which_clock,
		       struct compat_timex __user *, utp)
{
	const struct k_clock *kc = clockid_to_kclock(which_clock);
	struct timex ktx;
	int err;

	if (!kc)
		return -EINVAL;
	if (!kc->clock_adj)
		return -EOPNOTSUPP;

	err = compat_get_timex(&ktx, utp);
	if (err)
		return err;

	err = kc->clock_adj(which_clock, &ktx);

	if (err >= 0)
		err = compat_put_timex(utp, &ktx);

	return err;
}

1181 1182
COMPAT_SYSCALL_DEFINE2(clock_getres, clockid_t, which_clock,
		       struct compat_timespec __user *, tp)
L
Linus Torvalds 已提交
1183
{
1184
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1185 1186
	struct timespec64 ts;
	int err;
L
Linus Torvalds 已提交
1187

1188
	if (!kc)
L
Linus Torvalds 已提交
1189 1190
		return -EINVAL;

1191 1192 1193
	err = kc->clock_getres(which_clock, &ts);
	if (!err && tp && compat_put_timespec64(&ts, tp))
		return -EFAULT;
L
Linus Torvalds 已提交
1194

1195
	return err;
L
Linus Torvalds 已提交
1196
}
1197

1198
#endif
L
Linus Torvalds 已提交
1199

1200 1201 1202 1203
/*
 * nanosleep for monotonic and realtime clocks
 */
static int common_nsleep(const clockid_t which_clock, int flags,
1204
			 const struct timespec64 *rqtp)
1205
{
1206
	return hrtimer_nanosleep(rqtp, flags & TIMER_ABSTIME ?
1207 1208
				 HRTIMER_MODE_ABS : HRTIMER_MODE_REL,
				 which_clock);
1209
}
L
Linus Torvalds 已提交
1210

1211 1212 1213
SYSCALL_DEFINE4(clock_nanosleep, const clockid_t, which_clock, int, flags,
		const struct timespec __user *, rqtp,
		struct timespec __user *, rmtp)
L
Linus Torvalds 已提交
1214
{
1215
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1216
	struct timespec64 t64;
L
Linus Torvalds 已提交
1217 1218
	struct timespec t;

1219
	if (!kc)
L
Linus Torvalds 已提交
1220
		return -EINVAL;
1221 1222
	if (!kc->nsleep)
		return -ENANOSLEEP_NOTSUP;
L
Linus Torvalds 已提交
1223 1224 1225 1226

	if (copy_from_user(&t, rqtp, sizeof (struct timespec)))
		return -EFAULT;

1227 1228
	t64 = timespec_to_timespec64(t);
	if (!timespec64_valid(&t64))
L
Linus Torvalds 已提交
1229
		return -EINVAL;
1230 1231
	if (flags & TIMER_ABSTIME)
		rmtp = NULL;
1232
	current->restart_block.nanosleep.type = rmtp ? TT_NATIVE : TT_NONE;
1233
	current->restart_block.nanosleep.rmtp = rmtp;
L
Linus Torvalds 已提交
1234

1235
	return kc->nsleep(which_clock, flags, &t64);
L
Linus Torvalds 已提交
1236
}
1237

1238 1239 1240 1241
#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE4(clock_nanosleep, clockid_t, which_clock, int, flags,
		       struct compat_timespec __user *, rqtp,
		       struct compat_timespec __user *, rmtp)
1242
{
1243
	const struct k_clock *kc = clockid_to_kclock(which_clock);
1244 1245
	struct timespec64 t64;
	struct timespec t;
1246

1247
	if (!kc)
1248
		return -EINVAL;
1249 1250 1251 1252 1253
	if (!kc->nsleep)
		return -ENANOSLEEP_NOTSUP;

	if (compat_get_timespec(&t, rqtp))
		return -EFAULT;
1254

1255 1256 1257 1258 1259 1260 1261 1262 1263
	t64 = timespec_to_timespec64(t);
	if (!timespec64_valid(&t64))
		return -EINVAL;
	if (flags & TIMER_ABSTIME)
		rmtp = NULL;
	current->restart_block.nanosleep.type = rmtp ? TT_COMPAT : TT_NONE;
	current->restart_block.nanosleep.compat_rmtp = rmtp;

	return kc->nsleep(which_clock, flags, &t64);
1264
}
1265
#endif
1266 1267

static const struct k_clock clock_realtime = {
1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
	.clock_getres		= posix_get_hrtimer_res,
	.clock_get		= posix_clock_realtime_get,
	.clock_set		= posix_clock_realtime_set,
	.clock_adj		= posix_clock_realtime_adj,
	.nsleep			= common_nsleep,
	.timer_create		= common_timer_create,
	.timer_set		= common_timer_set,
	.timer_get		= common_timer_get,
	.timer_del		= common_timer_del,
	.timer_rearm		= common_hrtimer_rearm,
	.timer_forward		= common_hrtimer_forward,
	.timer_remaining	= common_hrtimer_remaining,
	.timer_try_to_cancel	= common_hrtimer_try_to_cancel,
	.timer_arm		= common_hrtimer_arm,
1282 1283 1284
};

static const struct k_clock clock_monotonic = {
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
	.clock_getres		= posix_get_hrtimer_res,
	.clock_get		= posix_ktime_get_ts,
	.nsleep			= common_nsleep,
	.timer_create		= common_timer_create,
	.timer_set		= common_timer_set,
	.timer_get		= common_timer_get,
	.timer_del		= common_timer_del,
	.timer_rearm		= common_hrtimer_rearm,
	.timer_forward		= common_hrtimer_forward,
	.timer_remaining	= common_hrtimer_remaining,
	.timer_try_to_cancel	= common_hrtimer_try_to_cancel,
	.timer_arm		= common_hrtimer_arm,
1297 1298 1299
};

static const struct k_clock clock_monotonic_raw = {
1300 1301
	.clock_getres		= posix_get_hrtimer_res,
	.clock_get		= posix_get_monotonic_raw,
1302 1303 1304
};

static const struct k_clock clock_realtime_coarse = {
1305 1306
	.clock_getres		= posix_get_coarse_res,
	.clock_get		= posix_get_realtime_coarse,
1307 1308 1309
};

static const struct k_clock clock_monotonic_coarse = {
1310 1311
	.clock_getres		= posix_get_coarse_res,
	.clock_get		= posix_get_monotonic_coarse,
1312 1313 1314
};

static const struct k_clock clock_tai = {
1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
	.clock_getres		= posix_get_hrtimer_res,
	.clock_get		= posix_get_tai,
	.nsleep			= common_nsleep,
	.timer_create		= common_timer_create,
	.timer_set		= common_timer_set,
	.timer_get		= common_timer_get,
	.timer_del		= common_timer_del,
	.timer_rearm		= common_hrtimer_rearm,
	.timer_forward		= common_hrtimer_forward,
	.timer_remaining	= common_hrtimer_remaining,
	.timer_try_to_cancel	= common_hrtimer_try_to_cancel,
	.timer_arm		= common_hrtimer_arm,
1327 1328 1329
};

static const struct k_clock clock_boottime = {
1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341
	.clock_getres		= posix_get_hrtimer_res,
	.clock_get		= posix_get_boottime,
	.nsleep			= common_nsleep,
	.timer_create		= common_timer_create,
	.timer_set		= common_timer_set,
	.timer_get		= common_timer_get,
	.timer_del		= common_timer_del,
	.timer_rearm		= common_hrtimer_rearm,
	.timer_forward		= common_hrtimer_forward,
	.timer_remaining	= common_hrtimer_remaining,
	.timer_try_to_cancel	= common_hrtimer_try_to_cancel,
	.timer_arm		= common_hrtimer_arm,
1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
};

static const struct k_clock * const posix_clocks[] = {
	[CLOCK_REALTIME]		= &clock_realtime,
	[CLOCK_MONOTONIC]		= &clock_monotonic,
	[CLOCK_PROCESS_CPUTIME_ID]	= &clock_process,
	[CLOCK_THREAD_CPUTIME_ID]	= &clock_thread,
	[CLOCK_MONOTONIC_RAW]		= &clock_monotonic_raw,
	[CLOCK_REALTIME_COARSE]		= &clock_realtime_coarse,
	[CLOCK_MONOTONIC_COARSE]	= &clock_monotonic_coarse,
	[CLOCK_BOOTTIME]		= &clock_boottime,
	[CLOCK_REALTIME_ALARM]		= &alarm_clock,
	[CLOCK_BOOTTIME_ALARM]		= &alarm_clock,
	[CLOCK_TAI]			= &clock_tai,
};

static const struct k_clock *clockid_to_kclock(const clockid_t id)
{
	if (id < 0)
		return (id & CLOCKFD_MASK) == CLOCKFD ?
			&clock_posix_dynamic : &clock_posix_cpu;

	if (id >= ARRAY_SIZE(posix_clocks) || !posix_clocks[id])
		return NULL;
	return posix_clocks[id];
}