posix-cpu-timers.c 39.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * Implement CPU time clocks for the POSIX clock interface.
 */

#include <linux/sched.h>
#include <linux/posix-timers.h>
#include <linux/errno.h>
R
Roman Zippel 已提交
8 9
#include <linux/math64.h>
#include <asm/uaccess.h>
10
#include <linux/kernel_stat.h>
11
#include <trace/events/timer.h>
12
#include <linux/random.h>
13 14
#include <linux/tick.h>
#include <linux/workqueue.h>
L
Linus Torvalds 已提交
15

16
/*
17 18 19 20
 * Called after updating RLIMIT_CPU to run cpu timer and update
 * tsk->signal->cputime_expires expiration cache if necessary. Needs
 * siglock protection since other code may update expiration cache as
 * well.
21
 */
22
void update_rlimit_cpu(struct task_struct *task, unsigned long rlim_new)
23
{
24
	cputime_t cputime = secs_to_cputime(rlim_new);
25

26 27 28
	spin_lock_irq(&task->sighand->siglock);
	set_process_cpu_timer(task, CPUCLOCK_PROF, &cputime, NULL);
	spin_unlock_irq(&task->sighand->siglock);
29 30
}

31
static int check_clock(const clockid_t which_clock)
L
Linus Torvalds 已提交
32 33 34 35 36 37 38 39 40 41 42
{
	int error = 0;
	struct task_struct *p;
	const pid_t pid = CPUCLOCK_PID(which_clock);

	if (CPUCLOCK_WHICH(which_clock) >= CPUCLOCK_MAX)
		return -EINVAL;

	if (pid == 0)
		return 0;

43
	rcu_read_lock();
44
	p = find_task_by_vpid(pid);
45
	if (!p || !(CPUCLOCK_PERTHREAD(which_clock) ?
46
		   same_thread_group(p, current) : has_group_leader_pid(p))) {
L
Linus Torvalds 已提交
47 48
		error = -EINVAL;
	}
49
	rcu_read_unlock();
L
Linus Torvalds 已提交
50 51 52 53

	return error;
}

54
static inline unsigned long long
55
timespec_to_sample(const clockid_t which_clock, const struct timespec *tp)
L
Linus Torvalds 已提交
56
{
57 58 59
	unsigned long long ret;

	ret = 0;		/* high half always zero when .cpu used */
L
Linus Torvalds 已提交
60
	if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED) {
61
		ret = (unsigned long long)tp->tv_sec * NSEC_PER_SEC + tp->tv_nsec;
L
Linus Torvalds 已提交
62
	} else {
63
		ret = cputime_to_expires(timespec_to_cputime(tp));
L
Linus Torvalds 已提交
64 65 66 67
	}
	return ret;
}

68
static void sample_to_timespec(const clockid_t which_clock,
69
			       unsigned long long expires,
L
Linus Torvalds 已提交
70 71
			       struct timespec *tp)
{
R
Roman Zippel 已提交
72
	if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED)
73
		*tp = ns_to_timespec(expires);
R
Roman Zippel 已提交
74
	else
75
		cputime_to_timespec((__force cputime_t)expires, tp);
L
Linus Torvalds 已提交
76 77 78 79 80 81
}

/*
 * Update expiry time from increment, and increase overrun count,
 * given the current clock sample.
 */
82
static void bump_cpu_timer(struct k_itimer *timer,
83
			   unsigned long long now)
L
Linus Torvalds 已提交
84 85
{
	int i;
86
	unsigned long long delta, incr;
L
Linus Torvalds 已提交
87

88
	if (timer->it.cpu.incr == 0)
L
Linus Torvalds 已提交
89 90
		return;

91 92
	if (now < timer->it.cpu.expires)
		return;
L
Linus Torvalds 已提交
93

94 95
	incr = timer->it.cpu.incr;
	delta = now + incr - timer->it.cpu.expires;
L
Linus Torvalds 已提交
96

97 98 99 100 101 102 103 104 105 106 107
	/* Don't use (incr*2 < delta), incr*2 might overflow. */
	for (i = 0; incr < delta - incr; i++)
		incr = incr << 1;

	for (; i >= 0; incr >>= 1, i--) {
		if (delta < incr)
			continue;

		timer->it.cpu.expires += incr;
		timer->it_overrun += 1 << i;
		delta -= incr;
L
Linus Torvalds 已提交
108 109 110
	}
}

111 112 113 114 115 116 117 118 119 120 121 122 123 124 125
/**
 * task_cputime_zero - Check a task_cputime struct for all zero fields.
 *
 * @cputime:	The struct to compare.
 *
 * Checks @cputime to see if all fields are zero.  Returns true if all fields
 * are zero, false if any field is nonzero.
 */
static inline int task_cputime_zero(const struct task_cputime *cputime)
{
	if (!cputime->utime && !cputime->stime && !cputime->sum_exec_runtime)
		return 1;
	return 0;
}

126
static inline unsigned long long prof_ticks(struct task_struct *p)
L
Linus Torvalds 已提交
127
{
128 129 130 131
	cputime_t utime, stime;

	task_cputime(p, &utime, &stime);

132
	return cputime_to_expires(utime + stime);
L
Linus Torvalds 已提交
133
}
134
static inline unsigned long long virt_ticks(struct task_struct *p)
L
Linus Torvalds 已提交
135
{
136 137 138 139
	cputime_t utime;

	task_cputime(p, &utime, NULL);

140
	return cputime_to_expires(utime);
L
Linus Torvalds 已提交
141 142
}

143 144
static int
posix_cpu_clock_getres(const clockid_t which_clock, struct timespec *tp)
L
Linus Torvalds 已提交
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
{
	int error = check_clock(which_clock);
	if (!error) {
		tp->tv_sec = 0;
		tp->tv_nsec = ((NSEC_PER_SEC + HZ - 1) / HZ);
		if (CPUCLOCK_WHICH(which_clock) == CPUCLOCK_SCHED) {
			/*
			 * If sched_clock is using a cycle counter, we
			 * don't have any idea of its true resolution
			 * exported, but it is much more than 1s/HZ.
			 */
			tp->tv_nsec = 1;
		}
	}
	return error;
}

162 163
static int
posix_cpu_clock_set(const clockid_t which_clock, const struct timespec *tp)
L
Linus Torvalds 已提交
164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
{
	/*
	 * You can never reset a CPU clock, but we check for other errors
	 * in the call before failing with EPERM.
	 */
	int error = check_clock(which_clock);
	if (error == 0) {
		error = -EPERM;
	}
	return error;
}


/*
 * Sample a per-thread clock for the given task.
 */
180
static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p,
181
			    unsigned long long *sample)
L
Linus Torvalds 已提交
182 183 184 185 186
{
	switch (CPUCLOCK_WHICH(which_clock)) {
	default:
		return -EINVAL;
	case CPUCLOCK_PROF:
187
		*sample = prof_ticks(p);
L
Linus Torvalds 已提交
188 189
		break;
	case CPUCLOCK_VIRT:
190
		*sample = virt_ticks(p);
L
Linus Torvalds 已提交
191 192
		break;
	case CPUCLOCK_SCHED:
193
		*sample = task_sched_runtime(p);
L
Linus Torvalds 已提交
194 195 196 197 198
		break;
	}
	return 0;
}

199 200 201 202 203
/*
 * Set cputime to sum_cputime if sum_cputime > cputime. Use cmpxchg
 * to avoid race conditions with concurrent updates to cputime.
 */
static inline void __update_gt_cputime(atomic64_t *cputime, u64 sum_cputime)
204
{
205 206 207 208 209 210 211 212
	u64 curr_cputime;
retry:
	curr_cputime = atomic64_read(cputime);
	if (sum_cputime > curr_cputime) {
		if (atomic64_cmpxchg(cputime, curr_cputime, sum_cputime) != curr_cputime)
			goto retry;
	}
}
213

214
static void update_gt_cputime(struct task_cputime_atomic *cputime_atomic, struct task_cputime *sum)
215
{
216 217 218
	__update_gt_cputime(&cputime_atomic->utime, sum->utime);
	__update_gt_cputime(&cputime_atomic->stime, sum->stime);
	__update_gt_cputime(&cputime_atomic->sum_exec_runtime, sum->sum_exec_runtime);
219
}
220

221 222 223
/* Sample task_cputime_atomic values in "atomic_timers", store results in "times". */
static inline void sample_cputime_atomic(struct task_cputime *times,
					 struct task_cputime_atomic *atomic_times)
224
{
225 226 227
	times->utime = atomic64_read(&atomic_times->utime);
	times->stime = atomic64_read(&atomic_times->stime);
	times->sum_exec_runtime = atomic64_read(&atomic_times->sum_exec_runtime);
228 229 230 231 232 233 234
}

void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times)
{
	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
	struct task_cputime sum;

235 236
	/* Check if cputimer isn't running. This is accessed without locking. */
	if (!READ_ONCE(cputimer->running)) {
237 238 239
		/*
		 * The POSIX timer interface allows for absolute time expiry
		 * values through the TIMER_ABSTIME flag, therefore we have
240
		 * to synchronize the timer to the clock every time we start it.
241 242
		 */
		thread_group_cputime(tsk, &sum);
243
		update_gt_cputime(&cputimer->cputime_atomic, &sum);
244 245 246 247 248 249 250 251

		/*
		 * We're setting cputimer->running without a lock. Ensure
		 * this only gets written to in one operation. We set
		 * running after update_gt_cputime() as a small optimization,
		 * but barriers are not required because update_gt_cputime()
		 * can handle concurrent updates.
		 */
252
		WRITE_ONCE(cputimer->running, true);
253
	}
254
	sample_cputime_atomic(times, &cputimer->cputime_atomic);
255 256
}

L
Linus Torvalds 已提交
257 258
/*
 * Sample a process (thread group) clock for the given group_leader task.
259 260
 * Must be called with task sighand lock held for safe while_each_thread()
 * traversal.
L
Linus Torvalds 已提交
261
 */
262 263
static int cpu_clock_sample_group(const clockid_t which_clock,
				  struct task_struct *p,
264
				  unsigned long long *sample)
L
Linus Torvalds 已提交
265
{
266 267
	struct task_cputime cputime;

268
	switch (CPUCLOCK_WHICH(which_clock)) {
L
Linus Torvalds 已提交
269 270 271
	default:
		return -EINVAL;
	case CPUCLOCK_PROF:
272
		thread_group_cputime(p, &cputime);
273
		*sample = cputime_to_expires(cputime.utime + cputime.stime);
L
Linus Torvalds 已提交
274 275
		break;
	case CPUCLOCK_VIRT:
276
		thread_group_cputime(p, &cputime);
277
		*sample = cputime_to_expires(cputime.utime);
L
Linus Torvalds 已提交
278 279
		break;
	case CPUCLOCK_SCHED:
280
		thread_group_cputime(p, &cputime);
281
		*sample = cputime.sum_exec_runtime;
L
Linus Torvalds 已提交
282 283 284 285 286
		break;
	}
	return 0;
}

287 288 289 290 291 292 293 294 295 296 297
static int posix_cpu_clock_get_task(struct task_struct *tsk,
				    const clockid_t which_clock,
				    struct timespec *tp)
{
	int err = -EINVAL;
	unsigned long long rtn;

	if (CPUCLOCK_PERTHREAD(which_clock)) {
		if (same_thread_group(tsk, current))
			err = cpu_clock_sample(which_clock, tsk, &rtn);
	} else {
298
		if (tsk == current || thread_group_leader(tsk))
299 300 301 302 303 304 305 306 307
			err = cpu_clock_sample_group(which_clock, tsk, &rtn);
	}

	if (!err)
		sample_to_timespec(which_clock, rtn, tp);

	return err;
}

L
Linus Torvalds 已提交
308

309
static int posix_cpu_clock_get(const clockid_t which_clock, struct timespec *tp)
L
Linus Torvalds 已提交
310 311
{
	const pid_t pid = CPUCLOCK_PID(which_clock);
312
	int err = -EINVAL;
L
Linus Torvalds 已提交
313 314 315 316 317 318

	if (pid == 0) {
		/*
		 * Special case constant value for our own clocks.
		 * We don't have to do any lookup to find ourselves.
		 */
319
		err = posix_cpu_clock_get_task(current, which_clock, tp);
L
Linus Torvalds 已提交
320 321 322 323 324 325
	} else {
		/*
		 * Find the given PID, and validate that the caller
		 * should be able to see it.
		 */
		struct task_struct *p;
326
		rcu_read_lock();
327
		p = find_task_by_vpid(pid);
328 329
		if (p)
			err = posix_cpu_clock_get_task(p, which_clock, tp);
330
		rcu_read_unlock();
L
Linus Torvalds 已提交
331 332
	}

333
	return err;
L
Linus Torvalds 已提交
334 335 336 337
}

/*
 * Validate the clockid_t for a new CPU-clock timer, and initialize the timer.
338 339
 * This is called from sys_timer_create() and do_cpu_nanosleep() with the
 * new timer already all-zeros initialized.
L
Linus Torvalds 已提交
340
 */
341
static int posix_cpu_timer_create(struct k_itimer *new_timer)
L
Linus Torvalds 已提交
342 343 344 345 346 347 348 349 350 351
{
	int ret = 0;
	const pid_t pid = CPUCLOCK_PID(new_timer->it_clock);
	struct task_struct *p;

	if (CPUCLOCK_WHICH(new_timer->it_clock) >= CPUCLOCK_MAX)
		return -EINVAL;

	INIT_LIST_HEAD(&new_timer->it.cpu.entry);

352
	rcu_read_lock();
L
Linus Torvalds 已提交
353 354 355 356
	if (CPUCLOCK_PERTHREAD(new_timer->it_clock)) {
		if (pid == 0) {
			p = current;
		} else {
357
			p = find_task_by_vpid(pid);
358
			if (p && !same_thread_group(p, current))
L
Linus Torvalds 已提交
359 360 361 362 363 364
				p = NULL;
		}
	} else {
		if (pid == 0) {
			p = current->group_leader;
		} else {
365
			p = find_task_by_vpid(pid);
366
			if (p && !has_group_leader_pid(p))
L
Linus Torvalds 已提交
367 368 369 370 371 372 373 374 375
				p = NULL;
		}
	}
	new_timer->it.cpu.task = p;
	if (p) {
		get_task_struct(p);
	} else {
		ret = -EINVAL;
	}
376
	rcu_read_unlock();
L
Linus Torvalds 已提交
377 378 379 380 381 382 383 384 385 386

	return ret;
}

/*
 * Clean up a CPU-clock timer that is about to be destroyed.
 * This is called from timer deletion with the timer already locked.
 * If we return TIMER_RETRY, it's necessary to release the timer's lock
 * and try again.  (This happens when the timer is in the middle of firing.)
 */
387
static int posix_cpu_timer_del(struct k_itimer *timer)
L
Linus Torvalds 已提交
388
{
389
	int ret = 0;
390 391 392
	unsigned long flags;
	struct sighand_struct *sighand;
	struct task_struct *p = timer->it.cpu.task;
L
Linus Torvalds 已提交
393

394
	WARN_ON_ONCE(p == NULL);
395

396 397 398 399 400 401
	/*
	 * Protect against sighand release/switch in exit/exec and process/
	 * thread timer list entry concurrent read/writes.
	 */
	sighand = lock_task_sighand(p, &flags);
	if (unlikely(sighand == NULL)) {
402 403 404 405
		/*
		 * We raced with the reaping of the task.
		 * The deletion should have cleared us off the list.
		 */
406
		WARN_ON_ONCE(!list_empty(&timer->it.cpu.entry));
407 408 409 410 411
	} else {
		if (timer->it.cpu.firing)
			ret = TIMER_RETRY;
		else
			list_del(&timer->it.cpu.entry);
412 413

		unlock_task_sighand(p, &flags);
L
Linus Torvalds 已提交
414
	}
415 416 417

	if (!ret)
		put_task_struct(p);
L
Linus Torvalds 已提交
418

419
	return ret;
L
Linus Torvalds 已提交
420 421
}

422
static void cleanup_timers_list(struct list_head *head)
423 424 425
{
	struct cpu_timer_list *timer, *next;

426
	list_for_each_entry_safe(timer, next, head, entry)
427 428 429
		list_del_init(&timer->entry);
}

L
Linus Torvalds 已提交
430 431 432 433 434 435
/*
 * Clean out CPU timers still ticking when a thread exited.  The task
 * pointer is cleared, and the expiry time is replaced with the residual
 * time for later timer_gettime calls to return.
 * This must be called with the siglock held.
 */
436
static void cleanup_timers(struct list_head *head)
L
Linus Torvalds 已提交
437
{
438 439 440
	cleanup_timers_list(head);
	cleanup_timers_list(++head);
	cleanup_timers_list(++head);
L
Linus Torvalds 已提交
441 442 443 444 445 446 447 448 449
}

/*
 * These are both called with the siglock held, when the current thread
 * is being reaped.  When the final (leader) thread in the group is reaped,
 * posix_cpu_timers_exit_group will be called after posix_cpu_timers_exit.
 */
void posix_cpu_timers_exit(struct task_struct *tsk)
{
450 451
	add_device_randomness((const void*) &tsk->se.sum_exec_runtime,
						sizeof(unsigned long long));
452
	cleanup_timers(tsk->cpu_timers);
L
Linus Torvalds 已提交
453 454 455 456

}
void posix_cpu_timers_exit_group(struct task_struct *tsk)
{
457
	cleanup_timers(tsk->signal->cpu_timers);
L
Linus Torvalds 已提交
458 459
}

460 461
static inline int expires_gt(cputime_t expires, cputime_t new_exp)
{
462
	return expires == 0 || expires > new_exp;
463 464
}

L
Linus Torvalds 已提交
465 466
/*
 * Insert the timer on the appropriate list before any timers that
467
 * expire later.  This must be called with the sighand lock held.
L
Linus Torvalds 已提交
468
 */
469
static void arm_timer(struct k_itimer *timer)
L
Linus Torvalds 已提交
470 471 472
{
	struct task_struct *p = timer->it.cpu.task;
	struct list_head *head, *listpos;
473
	struct task_cputime *cputime_expires;
L
Linus Torvalds 已提交
474 475 476
	struct cpu_timer_list *const nt = &timer->it.cpu;
	struct cpu_timer_list *next;

477 478 479 480 481 482 483
	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
		head = p->cpu_timers;
		cputime_expires = &p->cputime_expires;
	} else {
		head = p->signal->cpu_timers;
		cputime_expires = &p->signal->cputime_expires;
	}
L
Linus Torvalds 已提交
484 485 486
	head += CPUCLOCK_WHICH(timer->it_clock);

	listpos = head;
487
	list_for_each_entry(next, head, entry) {
488
		if (nt->expires < next->expires)
489 490
			break;
		listpos = &next->entry;
L
Linus Torvalds 已提交
491 492 493 494
	}
	list_add(&nt->entry, listpos);

	if (listpos == head) {
495
		unsigned long long exp = nt->expires;
496

L
Linus Torvalds 已提交
497
		/*
498 499 500 501
		 * We are the new earliest-expiring POSIX 1.b timer, hence
		 * need to update expiration cache. Take into account that
		 * for process timers we share expiration cache with itimers
		 * and RLIMIT_CPU and for thread timers with RLIMIT_RTTIME.
L
Linus Torvalds 已提交
502 503
		 */

504 505
		switch (CPUCLOCK_WHICH(timer->it_clock)) {
		case CPUCLOCK_PROF:
506 507
			if (expires_gt(cputime_expires->prof_exp, expires_to_cputime(exp)))
				cputime_expires->prof_exp = expires_to_cputime(exp);
508 509
			break;
		case CPUCLOCK_VIRT:
510 511
			if (expires_gt(cputime_expires->virt_exp, expires_to_cputime(exp)))
				cputime_expires->virt_exp = expires_to_cputime(exp);
512 513 514
			break;
		case CPUCLOCK_SCHED:
			if (cputime_expires->sched_exp == 0 ||
515 516
			    cputime_expires->sched_exp > exp)
				cputime_expires->sched_exp = exp;
517
			break;
L
Linus Torvalds 已提交
518
		}
519 520 521 522
		if (CPUCLOCK_PERTHREAD(timer->it_clock))
			tick_dep_set_task(p, TICK_DEP_BIT_POSIX_TIMER);
		else
			tick_dep_set_signal(p->signal, TICK_DEP_BIT_POSIX_TIMER);
L
Linus Torvalds 已提交
523 524 525 526 527 528 529 530
	}
}

/*
 * The timer is locked, fire it and arrange for its reload.
 */
static void cpu_timer_fire(struct k_itimer *timer)
{
531 532 533 534
	if ((timer->it_sigev_notify & ~SIGEV_THREAD_ID) == SIGEV_NONE) {
		/*
		 * User don't want any signal.
		 */
535
		timer->it.cpu.expires = 0;
536
	} else if (unlikely(timer->sigq == NULL)) {
L
Linus Torvalds 已提交
537 538 539 540 541
		/*
		 * This a special case for clock_nanosleep,
		 * not a normal timer from sys_timer_create.
		 */
		wake_up_process(timer->it_process);
542 543
		timer->it.cpu.expires = 0;
	} else if (timer->it.cpu.incr == 0) {
L
Linus Torvalds 已提交
544 545 546 547
		/*
		 * One-shot timer.  Clear it as soon as it's fired.
		 */
		posix_timer_event(timer, 0);
548
		timer->it.cpu.expires = 0;
L
Linus Torvalds 已提交
549 550 551 552 553 554 555 556 557 558 559
	} else if (posix_timer_event(timer, ++timer->it_requeue_pending)) {
		/*
		 * The signal did not get queued because the signal
		 * was ignored, so we won't get any callback to
		 * reload the timer.  But we need to keep it
		 * ticking in case the signal is deliverable next time.
		 */
		posix_cpu_timer_schedule(timer);
	}
}

560 561
/*
 * Sample a process (thread group) timer for the given group_leader task.
562 563
 * Must be called with task sighand lock held for safe while_each_thread()
 * traversal.
564 565 566
 */
static int cpu_timer_sample_group(const clockid_t which_clock,
				  struct task_struct *p,
567
				  unsigned long long *sample)
568 569 570 571 572 573 574 575
{
	struct task_cputime cputime;

	thread_group_cputimer(p, &cputime);
	switch (CPUCLOCK_WHICH(which_clock)) {
	default:
		return -EINVAL;
	case CPUCLOCK_PROF:
576
		*sample = cputime_to_expires(cputime.utime + cputime.stime);
577 578
		break;
	case CPUCLOCK_VIRT:
579
		*sample = cputime_to_expires(cputime.utime);
580 581
		break;
	case CPUCLOCK_SCHED:
582
		*sample = cputime.sum_exec_runtime;
583 584 585 586 587
		break;
	}
	return 0;
}

L
Linus Torvalds 已提交
588 589 590 591 592 593
/*
 * Guts of sys_timer_settime for CPU timers.
 * This is called with the timer locked and interrupts disabled.
 * If we return TIMER_RETRY, it's necessary to release the timer's lock
 * and try again.  (This happens when the timer is in the middle of firing.)
 */
594
static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
595
			       struct itimerspec *new, struct itimerspec *old)
L
Linus Torvalds 已提交
596
{
597 598
	unsigned long flags;
	struct sighand_struct *sighand;
L
Linus Torvalds 已提交
599
	struct task_struct *p = timer->it.cpu.task;
600
	unsigned long long old_expires, new_expires, old_incr, val;
L
Linus Torvalds 已提交
601 602
	int ret;

603
	WARN_ON_ONCE(p == NULL);
L
Linus Torvalds 已提交
604 605 606 607

	new_expires = timespec_to_sample(timer->it_clock, &new->it_value);

	/*
608 609 610 611 612 613
	 * Protect against sighand release/switch in exit/exec and p->cpu_timers
	 * and p->signal->cpu_timers read/write in arm_timer()
	 */
	sighand = lock_task_sighand(p, &flags);
	/*
	 * If p has just been reaped, we can no
L
Linus Torvalds 已提交
614 615
	 * longer get any information about it at all.
	 */
616
	if (unlikely(sighand == NULL)) {
L
Linus Torvalds 已提交
617 618 619 620 621 622
		return -ESRCH;
	}

	/*
	 * Disarm any old timer after extracting its expiry time.
	 */
623
	WARN_ON_ONCE(!irqs_disabled());
624 625

	ret = 0;
626
	old_incr = timer->it.cpu.incr;
L
Linus Torvalds 已提交
627
	old_expires = timer->it.cpu.expires;
628 629 630 631 632
	if (unlikely(timer->it.cpu.firing)) {
		timer->it.cpu.firing = -1;
		ret = TIMER_RETRY;
	} else
		list_del_init(&timer->it.cpu.entry);
L
Linus Torvalds 已提交
633 634 635 636 637 638 639 640 641 642 643 644

	/*
	 * We need to sample the current value to convert the new
	 * value from to relative and absolute, and to convert the
	 * old value from absolute to relative.  To set a process
	 * timer, we need a sample to balance the thread expiry
	 * times (in arm_timer).  With an absolute time, we must
	 * check if it's already passed.  In short, we need a sample.
	 */
	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
		cpu_clock_sample(timer->it_clock, p, &val);
	} else {
645
		cpu_timer_sample_group(timer->it_clock, p, &val);
L
Linus Torvalds 已提交
646 647 648
	}

	if (old) {
649
		if (old_expires == 0) {
L
Linus Torvalds 已提交
650 651 652 653 654 655 656 657 658 659 660 661 662 663
			old->it_value.tv_sec = 0;
			old->it_value.tv_nsec = 0;
		} else {
			/*
			 * Update the timer in case it has
			 * overrun already.  If it has,
			 * we'll report it as having overrun
			 * and with the next reloaded timer
			 * already ticking, though we are
			 * swallowing that pending
			 * notification here to install the
			 * new setting.
			 */
			bump_cpu_timer(timer, val);
664 665
			if (val < timer->it.cpu.expires) {
				old_expires = timer->it.cpu.expires - val;
L
Linus Torvalds 已提交
666 667 668 669 670 671 672 673 674 675
				sample_to_timespec(timer->it_clock,
						   old_expires,
						   &old->it_value);
			} else {
				old->it_value.tv_nsec = 1;
				old->it_value.tv_sec = 0;
			}
		}
	}

676
	if (unlikely(ret)) {
L
Linus Torvalds 已提交
677 678 679 680 681 682
		/*
		 * We are colliding with the timer actually firing.
		 * Punt after filling in the timer's old value, and
		 * disable this firing since we are already reporting
		 * it as an overrun (thanks to bump_cpu_timer above).
		 */
683
		unlock_task_sighand(p, &flags);
L
Linus Torvalds 已提交
684 685 686
		goto out;
	}

687
	if (new_expires != 0 && !(timer_flags & TIMER_ABSTIME)) {
688
		new_expires += val;
L
Linus Torvalds 已提交
689 690 691 692 693 694 695 696
	}

	/*
	 * Install the new expiry time (or zero).
	 * For a timer with no notification action, we don't actually
	 * arm the timer (we'll just fake it for timer_gettime).
	 */
	timer->it.cpu.expires = new_expires;
697
	if (new_expires != 0 && val < new_expires) {
698
		arm_timer(timer);
L
Linus Torvalds 已提交
699 700
	}

701
	unlock_task_sighand(p, &flags);
L
Linus Torvalds 已提交
702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718
	/*
	 * Install the new reload setting, and
	 * set up the signal and overrun bookkeeping.
	 */
	timer->it.cpu.incr = timespec_to_sample(timer->it_clock,
						&new->it_interval);

	/*
	 * This acts as a modification timestamp for the timer,
	 * so any automatic reload attempt will punt on seeing
	 * that we have reset the timer manually.
	 */
	timer->it_requeue_pending = (timer->it_requeue_pending + 2) &
		~REQUEUE_PENDING;
	timer->it_overrun_last = 0;
	timer->it_overrun = -1;

719
	if (new_expires != 0 && !(val < new_expires)) {
L
Linus Torvalds 已提交
720 721 722 723 724 725 726 727 728 729 730 731
		/*
		 * The designated time already passed, so we notify
		 * immediately, even if the thread never runs to
		 * accumulate more time on this clock.
		 */
		cpu_timer_fire(timer);
	}

	ret = 0;
 out:
	if (old) {
		sample_to_timespec(timer->it_clock,
732
				   old_incr, &old->it_interval);
L
Linus Torvalds 已提交
733
	}
734

L
Linus Torvalds 已提交
735 736 737
	return ret;
}

738
static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp)
L
Linus Torvalds 已提交
739
{
740
	unsigned long long now;
L
Linus Torvalds 已提交
741 742
	struct task_struct *p = timer->it.cpu.task;

743 744
	WARN_ON_ONCE(p == NULL);

L
Linus Torvalds 已提交
745 746 747 748 749 750
	/*
	 * Easy part: convert the reload time.
	 */
	sample_to_timespec(timer->it_clock,
			   timer->it.cpu.incr, &itp->it_interval);

751
	if (timer->it.cpu.expires == 0) {	/* Timer not armed at all.  */
L
Linus Torvalds 已提交
752 753 754 755 756 757 758 759 760 761
		itp->it_value.tv_sec = itp->it_value.tv_nsec = 0;
		return;
	}

	/*
	 * Sample the clock to take the difference with the expiry time.
	 */
	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
		cpu_clock_sample(timer->it_clock, p, &now);
	} else {
762 763 764 765 766 767 768 769 770 771
		struct sighand_struct *sighand;
		unsigned long flags;

		/*
		 * Protect against sighand release/switch in exit/exec and
		 * also make timer sampling safe if it ends up calling
		 * thread_group_cputime().
		 */
		sighand = lock_task_sighand(p, &flags);
		if (unlikely(sighand == NULL)) {
L
Linus Torvalds 已提交
772 773 774 775 776
			/*
			 * The process has been reaped.
			 * We can't even collect a sample any more.
			 * Call the timer disarmed, nothing else to do.
			 */
777
			timer->it.cpu.expires = 0;
778 779
			sample_to_timespec(timer->it_clock, timer->it.cpu.expires,
					   &itp->it_value);
780
			return;
L
Linus Torvalds 已提交
781
		} else {
782
			cpu_timer_sample_group(timer->it_clock, p, &now);
783
			unlock_task_sighand(p, &flags);
L
Linus Torvalds 已提交
784 785 786
		}
	}

787
	if (now < timer->it.cpu.expires) {
L
Linus Torvalds 已提交
788
		sample_to_timespec(timer->it_clock,
789
				   timer->it.cpu.expires - now,
L
Linus Torvalds 已提交
790 791 792 793 794 795 796 797 798 799 800
				   &itp->it_value);
	} else {
		/*
		 * The timer should have expired already, but the firing
		 * hasn't taken place yet.  Say it's just about to expire.
		 */
		itp->it_value.tv_nsec = 1;
		itp->it_value.tv_sec = 0;
	}
}

801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822
static unsigned long long
check_timers_list(struct list_head *timers,
		  struct list_head *firing,
		  unsigned long long curr)
{
	int maxfire = 20;

	while (!list_empty(timers)) {
		struct cpu_timer_list *t;

		t = list_first_entry(timers, struct cpu_timer_list, entry);

		if (!--maxfire || curr < t->expires)
			return t->expires;

		t->firing = 1;
		list_move_tail(&t->entry, firing);
	}

	return 0;
}

L
Linus Torvalds 已提交
823 824 825 826 827 828 829 830 831
/*
 * Check for any per-thread CPU timers that have fired and move them off
 * the tsk->cpu_timers[N] list onto the firing list.  Here we update the
 * tsk->it_*_expires values to reflect the remaining thread CPU timers.
 */
static void check_thread_timers(struct task_struct *tsk,
				struct list_head *firing)
{
	struct list_head *timers = tsk->cpu_timers;
832
	struct signal_struct *const sig = tsk->signal;
833 834
	struct task_cputime *tsk_expires = &tsk->cputime_expires;
	unsigned long long expires;
835
	unsigned long soft;
L
Linus Torvalds 已提交
836

837 838 839 840 841 842 843
	/*
	 * If cputime_expires is zero, then there are no active
	 * per thread CPU timers.
	 */
	if (task_cputime_zero(&tsk->cputime_expires))
		return;

844 845
	expires = check_timers_list(timers, firing, prof_ticks(tsk));
	tsk_expires->prof_exp = expires_to_cputime(expires);
L
Linus Torvalds 已提交
846

847 848
	expires = check_timers_list(++timers, firing, virt_ticks(tsk));
	tsk_expires->virt_exp = expires_to_cputime(expires);
L
Linus Torvalds 已提交
849

850 851
	tsk_expires->sched_exp = check_timers_list(++timers, firing,
						   tsk->se.sum_exec_runtime);
852 853 854 855

	/*
	 * Check for the special case thread timers.
	 */
856
	soft = READ_ONCE(sig->rlim[RLIMIT_RTTIME].rlim_cur);
857
	if (soft != RLIM_INFINITY) {
858
		unsigned long hard =
859
			READ_ONCE(sig->rlim[RLIMIT_RTTIME].rlim_max);
860

861 862
		if (hard != RLIM_INFINITY &&
		    tsk->rt.timeout > DIV_ROUND_UP(hard, USEC_PER_SEC/HZ)) {
863 864 865 866 867 868 869
			/*
			 * At the hard limit, we just die.
			 * No need to calculate anything else now.
			 */
			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
			return;
		}
870
		if (tsk->rt.timeout > DIV_ROUND_UP(soft, USEC_PER_SEC/HZ)) {
871 872 873
			/*
			 * At the soft limit, send a SIGXCPU every second.
			 */
874 875 876
			if (soft < hard) {
				soft += USEC_PER_SEC;
				sig->rlim[RLIMIT_RTTIME].rlim_cur = soft;
877
			}
878 879 880
			printk(KERN_INFO
				"RT Watchdog Timeout: %s[%d]\n",
				tsk->comm, task_pid_nr(tsk));
881 882 883
			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
		}
	}
884 885
	if (task_cputime_zero(tsk_expires))
		tick_dep_clear_task(tsk, TICK_DEP_BIT_POSIX_TIMER);
L
Linus Torvalds 已提交
886 887
}

888
static inline void stop_process_timers(struct signal_struct *sig)
889
{
890
	struct thread_group_cputimer *cputimer = &sig->cputimer;
891

892
	/* Turn off cputimer->running. This is done without locking. */
893
	WRITE_ONCE(cputimer->running, false);
894
	tick_dep_clear_signal(sig, TICK_DEP_BIT_POSIX_TIMER);
895 896
}

897 898
static u32 onecputick;

899
static void check_cpu_itimer(struct task_struct *tsk, struct cpu_itimer *it,
900 901
			     unsigned long long *expires,
			     unsigned long long cur_time, int signo)
902
{
903
	if (!it->expires)
904 905
		return;

906 907 908
	if (cur_time >= it->expires) {
		if (it->incr) {
			it->expires += it->incr;
909 910
			it->error += it->incr_error;
			if (it->error >= onecputick) {
911
				it->expires -= cputime_one_jiffy;
912 913
				it->error -= onecputick;
			}
914
		} else {
915
			it->expires = 0;
916
		}
917

918 919 920
		trace_itimer_expire(signo == SIGPROF ?
				    ITIMER_PROF : ITIMER_VIRTUAL,
				    tsk->signal->leader_pid, cur_time);
921 922 923
		__group_send_sig_info(signo, SEND_SIG_PRIV, tsk);
	}

924
	if (it->expires && (!*expires || it->expires < *expires)) {
925 926 927 928
		*expires = it->expires;
	}
}

L
Linus Torvalds 已提交
929 930 931 932 933 934 935 936 937
/*
 * Check for any per-thread CPU timers that have fired and move them
 * off the tsk->*_timers list onto the firing list.  Per-thread timers
 * have already been taken off.
 */
static void check_process_timers(struct task_struct *tsk,
				 struct list_head *firing)
{
	struct signal_struct *const sig = tsk->signal;
938
	unsigned long long utime, ptime, virt_expires, prof_expires;
939
	unsigned long long sum_sched_runtime, sched_expires;
L
Linus Torvalds 已提交
940
	struct list_head *timers = sig->cpu_timers;
941
	struct task_cputime cputime;
942
	unsigned long soft;
L
Linus Torvalds 已提交
943

944 945 946 947 948 949 950
	/*
	 * If cputimer is not running, then there are no active
	 * process wide timers (POSIX 1.b, itimers, RLIMIT_CPU).
	 */
	if (!READ_ONCE(tsk->signal->cputimer.running))
		return;

951 952 953 954 955 956
        /*
	 * Signify that a thread is checking for process timers.
	 * Write access to this field is protected by the sighand lock.
	 */
	sig->cputimer.checking_timer = true;

L
Linus Torvalds 已提交
957 958 959
	/*
	 * Collect the current process totals.
	 */
960
	thread_group_cputimer(tsk, &cputime);
961 962
	utime = cputime_to_expires(cputime.utime);
	ptime = utime + cputime_to_expires(cputime.stime);
963
	sum_sched_runtime = cputime.sum_exec_runtime;
L
Linus Torvalds 已提交
964

965 966 967
	prof_expires = check_timers_list(timers, firing, ptime);
	virt_expires = check_timers_list(++timers, firing, utime);
	sched_expires = check_timers_list(++timers, firing, sum_sched_runtime);
L
Linus Torvalds 已提交
968 969 970 971

	/*
	 * Check for the special case process timers.
	 */
972 973 974 975
	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_PROF], &prof_expires, ptime,
			 SIGPROF);
	check_cpu_itimer(tsk, &sig->it[CPUCLOCK_VIRT], &virt_expires, utime,
			 SIGVTALRM);
976
	soft = READ_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur);
977
	if (soft != RLIM_INFINITY) {
L
Linus Torvalds 已提交
978
		unsigned long psecs = cputime_to_secs(ptime);
979
		unsigned long hard =
980
			READ_ONCE(sig->rlim[RLIMIT_CPU].rlim_max);
L
Linus Torvalds 已提交
981
		cputime_t x;
982
		if (psecs >= hard) {
L
Linus Torvalds 已提交
983 984 985 986 987 988 989
			/*
			 * At the hard limit, we just die.
			 * No need to calculate anything else now.
			 */
			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
			return;
		}
990
		if (psecs >= soft) {
L
Linus Torvalds 已提交
991 992 993 994
			/*
			 * At the soft limit, send a SIGXCPU every second.
			 */
			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
995 996 997
			if (soft < hard) {
				soft++;
				sig->rlim[RLIMIT_CPU].rlim_cur = soft;
L
Linus Torvalds 已提交
998 999
			}
		}
1000
		x = secs_to_cputime(soft);
1001
		if (!prof_expires || x < prof_expires) {
L
Linus Torvalds 已提交
1002 1003 1004 1005
			prof_expires = x;
		}
	}

1006 1007
	sig->cputime_expires.prof_exp = expires_to_cputime(prof_expires);
	sig->cputime_expires.virt_exp = expires_to_cputime(virt_expires);
1008 1009 1010
	sig->cputime_expires.sched_exp = sched_expires;
	if (task_cputime_zero(&sig->cputime_expires))
		stop_process_timers(sig);
1011 1012

	sig->cputimer.checking_timer = false;
L
Linus Torvalds 已提交
1013 1014 1015 1016 1017 1018 1019 1020
}

/*
 * This is called from the signal code (via do_schedule_next_timer)
 * when the last timer signal was delivered and we have to reload the timer.
 */
void posix_cpu_timer_schedule(struct k_itimer *timer)
{
1021 1022
	struct sighand_struct *sighand;
	unsigned long flags;
L
Linus Torvalds 已提交
1023
	struct task_struct *p = timer->it.cpu.task;
1024
	unsigned long long now;
L
Linus Torvalds 已提交
1025

1026
	WARN_ON_ONCE(p == NULL);
L
Linus Torvalds 已提交
1027 1028 1029 1030 1031 1032 1033

	/*
	 * Fetch the current sample and update the timer's expiry time.
	 */
	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
		cpu_clock_sample(timer->it_clock, p, &now);
		bump_cpu_timer(timer, now);
1034
		if (unlikely(p->exit_state))
1035
			goto out;
1036

1037 1038 1039 1040
		/* Protect timer list r/w in arm_timer() */
		sighand = lock_task_sighand(p, &flags);
		if (!sighand)
			goto out;
L
Linus Torvalds 已提交
1041
	} else {
1042 1043 1044 1045 1046 1047
		/*
		 * Protect arm_timer() and timer sampling in case of call to
		 * thread_group_cputime().
		 */
		sighand = lock_task_sighand(p, &flags);
		if (unlikely(sighand == NULL)) {
L
Linus Torvalds 已提交
1048 1049 1050 1051
			/*
			 * The process has been reaped.
			 * We can't even collect a sample any more.
			 */
1052
			timer->it.cpu.expires = 0;
1053
			goto out;
L
Linus Torvalds 已提交
1054
		} else if (unlikely(p->exit_state) && thread_group_empty(p)) {
1055
			unlock_task_sighand(p, &flags);
1056
			/* Optimizations: if the process is dying, no need to rearm */
1057
			goto out;
L
Linus Torvalds 已提交
1058
		}
1059
		cpu_timer_sample_group(timer->it_clock, p, &now);
L
Linus Torvalds 已提交
1060
		bump_cpu_timer(timer, now);
1061
		/* Leave the sighand locked for the call below.  */
L
Linus Torvalds 已提交
1062 1063 1064 1065 1066
	}

	/*
	 * Now re-arm for the new expiry time.
	 */
1067
	WARN_ON_ONCE(!irqs_disabled());
1068
	arm_timer(timer);
1069
	unlock_task_sighand(p, &flags);
1070 1071 1072 1073 1074

out:
	timer->it_overrun_last = timer->it_overrun;
	timer->it_overrun = -1;
	++timer->it_requeue_pending;
L
Linus Torvalds 已提交
1075 1076
}

1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
/**
 * task_cputime_expired - Compare two task_cputime entities.
 *
 * @sample:	The task_cputime structure to be checked for expiration.
 * @expires:	Expiration times, against which @sample will be checked.
 *
 * Checks @sample against @expires to see if any field of @sample has expired.
 * Returns true if any field of the former is greater than the corresponding
 * field of the latter if the latter field is set.  Otherwise returns false.
 */
static inline int task_cputime_expired(const struct task_cputime *sample,
					const struct task_cputime *expires)
{
1090
	if (expires->utime && sample->utime >= expires->utime)
1091
		return 1;
1092
	if (expires->stime && sample->utime + sample->stime >= expires->stime)
1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
		return 1;
	if (expires->sum_exec_runtime != 0 &&
	    sample->sum_exec_runtime >= expires->sum_exec_runtime)
		return 1;
	return 0;
}

/**
 * fastpath_timer_check - POSIX CPU timers fast path.
 *
 * @tsk:	The task (thread) being checked.
 *
1105 1106 1107 1108
 * Check the task and thread group timers.  If both are zero (there are no
 * timers set) return false.  Otherwise snapshot the task and thread group
 * timers and compare them with the corresponding expiration times.  Return
 * true if a timer has expired, else return false.
1109
 */
1110
static inline int fastpath_timer_check(struct task_struct *tsk)
1111
{
1112
	struct signal_struct *sig;
1113 1114

	if (!task_cputime_zero(&tsk->cputime_expires)) {
1115
		struct task_cputime task_sample;
1116

1117 1118
		task_cputime(tsk, &task_sample.utime, &task_sample.stime);
		task_sample.sum_exec_runtime = tsk->se.sum_exec_runtime;
1119 1120 1121
		if (task_cputime_expired(&task_sample, &tsk->cputime_expires))
			return 1;
	}
1122 1123

	sig = tsk->signal;
1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
	/*
	 * Check if thread group timers expired when the cputimer is
	 * running and no other thread in the group is already checking
	 * for thread group cputimers. These fields are read without the
	 * sighand lock. However, this is fine because this is meant to
	 * be a fastpath heuristic to determine whether we should try to
	 * acquire the sighand lock to check/handle timers.
	 *
	 * In the worst case scenario, if 'running' or 'checking_timer' gets
	 * set but the current thread doesn't see the change yet, we'll wait
	 * until the next thread in the group gets a scheduler interrupt to
	 * handle the timer. This isn't an issue in practice because these
	 * types of delays with signals actually getting sent are expected.
	 */
	if (READ_ONCE(sig->cputimer.running) &&
	    !READ_ONCE(sig->cputimer.checking_timer)) {
1140 1141
		struct task_cputime group_sample;

1142
		sample_cputime_atomic(&group_sample, &sig->cputimer.cputime_atomic);
1143

1144 1145 1146
		if (task_cputime_expired(&group_sample, &sig->cputime_expires))
			return 1;
	}
1147

1148
	return 0;
1149 1150
}

L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159
/*
 * This is called from the timer interrupt handler.  The irq handler has
 * already updated our counts.  We need to check if any timers fire now.
 * Interrupts are disabled.
 */
void run_posix_cpu_timers(struct task_struct *tsk)
{
	LIST_HEAD(firing);
	struct k_itimer *timer, *next;
1160
	unsigned long flags;
L
Linus Torvalds 已提交
1161

1162
	WARN_ON_ONCE(!irqs_disabled());
L
Linus Torvalds 已提交
1163 1164

	/*
1165
	 * The fast path checks that there are no expired thread or thread
1166
	 * group timers.  If that's so, just return.
L
Linus Torvalds 已提交
1167
	 */
1168
	if (!fastpath_timer_check(tsk))
1169
		return;
1170

1171 1172
	if (!lock_task_sighand(tsk, &flags))
		return;
1173 1174 1175 1176 1177 1178
	/*
	 * Here we take off tsk->signal->cpu_timers[N] and
	 * tsk->cpu_timers[N] all the timers that are firing, and
	 * put them on the firing list.
	 */
	check_thread_timers(tsk, &firing);
1179 1180

	check_process_timers(tsk, &firing);
L
Linus Torvalds 已提交
1181

1182 1183 1184 1185 1186 1187 1188 1189
	/*
	 * We must release these locks before taking any timer's lock.
	 * There is a potential race with timer deletion here, as the
	 * siglock now protects our private firing list.  We have set
	 * the firing flag in each timer, so that a deletion attempt
	 * that gets the timer lock before we do will give it up and
	 * spin until we've taken care of that timer below.
	 */
1190
	unlock_task_sighand(tsk, &flags);
L
Linus Torvalds 已提交
1191 1192 1193

	/*
	 * Now that all the timers on our list have the firing flag,
L
Lucas De Marchi 已提交
1194
	 * no one will touch their list entries but us.  We'll take
L
Linus Torvalds 已提交
1195 1196 1197 1198
	 * each timer's lock before clearing its firing flag, so no
	 * timer call will interfere.
	 */
	list_for_each_entry_safe(timer, next, &firing, it.cpu.entry) {
1199 1200
		int cpu_firing;

L
Linus Torvalds 已提交
1201 1202
		spin_lock(&timer->it_lock);
		list_del_init(&timer->it.cpu.entry);
1203
		cpu_firing = timer->it.cpu.firing;
L
Linus Torvalds 已提交
1204 1205 1206 1207 1208 1209
		timer->it.cpu.firing = 0;
		/*
		 * The firing flag is -1 if we collided with a reset
		 * of the timer, which already reported this
		 * almost-firing as an overrun.  So don't generate an event.
		 */
1210
		if (likely(cpu_firing >= 0))
L
Linus Torvalds 已提交
1211 1212 1213 1214 1215 1216
			cpu_timer_fire(timer);
		spin_unlock(&timer->it_lock);
	}
}

/*
1217
 * Set one of the process-wide special case CPU timers or RLIMIT_CPU.
1218
 * The tsk->sighand->siglock must be held by the caller.
L
Linus Torvalds 已提交
1219 1220 1221 1222
 */
void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
			   cputime_t *newval, cputime_t *oldval)
{
1223
	unsigned long long now;
L
Linus Torvalds 已提交
1224

1225
	WARN_ON_ONCE(clock_idx == CPUCLOCK_SCHED);
1226
	cpu_timer_sample_group(clock_idx, tsk, &now);
L
Linus Torvalds 已提交
1227 1228

	if (oldval) {
1229 1230 1231 1232 1233
		/*
		 * We are setting itimer. The *oldval is absolute and we update
		 * it to be relative, *newval argument is relative and we update
		 * it to be absolute.
		 */
1234
		if (*oldval) {
1235
			if (*oldval <= now) {
L
Linus Torvalds 已提交
1236
				/* Just about to fire. */
1237
				*oldval = cputime_one_jiffy;
L
Linus Torvalds 已提交
1238
			} else {
1239
				*oldval -= now;
L
Linus Torvalds 已提交
1240 1241 1242
			}
		}

1243
		if (!*newval)
1244
			return;
1245
		*newval += now;
L
Linus Torvalds 已提交
1246 1247 1248
	}

	/*
1249 1250
	 * Update expiration cache if we are the earliest timer, or eventually
	 * RLIMIT_CPU limit is earlier than prof_exp cpu timer expire.
L
Linus Torvalds 已提交
1251
	 */
1252 1253 1254
	switch (clock_idx) {
	case CPUCLOCK_PROF:
		if (expires_gt(tsk->signal->cputime_expires.prof_exp, *newval))
1255
			tsk->signal->cputime_expires.prof_exp = *newval;
1256 1257 1258
		break;
	case CPUCLOCK_VIRT:
		if (expires_gt(tsk->signal->cputime_expires.virt_exp, *newval))
1259
			tsk->signal->cputime_expires.virt_exp = *newval;
1260
		break;
L
Linus Torvalds 已提交
1261
	}
1262 1263

	tick_dep_set_signal(tsk->signal, TICK_DEP_BIT_POSIX_TIMER);
L
Linus Torvalds 已提交
1264 1265
}

1266 1267
static int do_cpu_nanosleep(const clockid_t which_clock, int flags,
			    struct timespec *rqtp, struct itimerspec *it)
L
Linus Torvalds 已提交
1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
{
	struct k_itimer timer;
	int error;

	/*
	 * Set up a temporary timer and then wait for it to go off.
	 */
	memset(&timer, 0, sizeof timer);
	spin_lock_init(&timer.it_lock);
	timer.it_clock = which_clock;
	timer.it_overrun = -1;
	error = posix_cpu_timer_create(&timer);
	timer.it_process = current;
	if (!error) {
		static struct itimerspec zero_it;
1283 1284 1285

		memset(it, 0, sizeof *it);
		it->it_value = *rqtp;
L
Linus Torvalds 已提交
1286 1287

		spin_lock_irq(&timer.it_lock);
1288
		error = posix_cpu_timer_set(&timer, flags, it, NULL);
L
Linus Torvalds 已提交
1289 1290 1291 1292 1293 1294
		if (error) {
			spin_unlock_irq(&timer.it_lock);
			return error;
		}

		while (!signal_pending(current)) {
1295
			if (timer.it.cpu.expires == 0) {
L
Linus Torvalds 已提交
1296
				/*
1297 1298
				 * Our timer fired and was reset, below
				 * deletion can not fail.
L
Linus Torvalds 已提交
1299
				 */
1300
				posix_cpu_timer_del(&timer);
L
Linus Torvalds 已提交
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
				spin_unlock_irq(&timer.it_lock);
				return 0;
			}

			/*
			 * Block until cpu_timer_fire (or a signal) wakes us.
			 */
			__set_current_state(TASK_INTERRUPTIBLE);
			spin_unlock_irq(&timer.it_lock);
			schedule();
			spin_lock_irq(&timer.it_lock);
		}

		/*
		 * We were interrupted by a signal.
		 */
		sample_to_timespec(which_clock, timer.it.cpu.expires, rqtp);
1318 1319 1320 1321 1322 1323 1324
		error = posix_cpu_timer_set(&timer, 0, &zero_it, it);
		if (!error) {
			/*
			 * Timer is now unarmed, deletion can not fail.
			 */
			posix_cpu_timer_del(&timer);
		}
L
Linus Torvalds 已提交
1325 1326
		spin_unlock_irq(&timer.it_lock);

1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
		while (error == TIMER_RETRY) {
			/*
			 * We need to handle case when timer was or is in the
			 * middle of firing. In other cases we already freed
			 * resources.
			 */
			spin_lock_irq(&timer.it_lock);
			error = posix_cpu_timer_del(&timer);
			spin_unlock_irq(&timer.it_lock);
		}

1338
		if ((it->it_value.tv_sec | it->it_value.tv_nsec) == 0) {
L
Linus Torvalds 已提交
1339 1340 1341 1342 1343 1344
			/*
			 * It actually did fire already.
			 */
			return 0;
		}

1345 1346 1347 1348 1349 1350
		error = -ERESTART_RESTARTBLOCK;
	}

	return error;
}

1351 1352 1353 1354
static long posix_cpu_nsleep_restart(struct restart_block *restart_block);

static int posix_cpu_nsleep(const clockid_t which_clock, int flags,
			    struct timespec *rqtp, struct timespec __user *rmtp)
1355
{
1356
	struct restart_block *restart_block = &current->restart_block;
1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
	struct itimerspec it;
	int error;

	/*
	 * Diagnose required errors first.
	 */
	if (CPUCLOCK_PERTHREAD(which_clock) &&
	    (CPUCLOCK_PID(which_clock) == 0 ||
	     CPUCLOCK_PID(which_clock) == current->pid))
		return -EINVAL;

	error = do_cpu_nanosleep(which_clock, flags, rqtp, &it);

	if (error == -ERESTART_RESTARTBLOCK) {

1372
		if (flags & TIMER_ABSTIME)
1373
			return -ERESTARTNOHAND;
L
Linus Torvalds 已提交
1374
		/*
1375 1376 1377
		 * Report back to the user the time still remaining.
		 */
		if (rmtp && copy_to_user(rmtp, &it.it_value, sizeof *rmtp))
L
Linus Torvalds 已提交
1378 1379
			return -EFAULT;

1380
		restart_block->fn = posix_cpu_nsleep_restart;
1381
		restart_block->nanosleep.clockid = which_clock;
1382 1383
		restart_block->nanosleep.rmtp = rmtp;
		restart_block->nanosleep.expires = timespec_to_ns(rqtp);
L
Linus Torvalds 已提交
1384 1385 1386 1387
	}
	return error;
}

1388
static long posix_cpu_nsleep_restart(struct restart_block *restart_block)
L
Linus Torvalds 已提交
1389
{
1390
	clockid_t which_clock = restart_block->nanosleep.clockid;
1391
	struct timespec t;
1392 1393
	struct itimerspec it;
	int error;
1394

1395
	t = ns_to_timespec(restart_block->nanosleep.expires);
1396

1397 1398 1399
	error = do_cpu_nanosleep(which_clock, TIMER_ABSTIME, &t, &it);

	if (error == -ERESTART_RESTARTBLOCK) {
1400
		struct timespec __user *rmtp = restart_block->nanosleep.rmtp;
1401
		/*
1402 1403 1404
		 * Report back to the user the time still remaining.
		 */
		if (rmtp && copy_to_user(rmtp, &it.it_value, sizeof *rmtp))
1405 1406
			return -EFAULT;

1407
		restart_block->nanosleep.expires = timespec_to_ns(&t);
1408 1409 1410
	}
	return error;

L
Linus Torvalds 已提交
1411 1412 1413 1414 1415
}

#define PROCESS_CLOCK	MAKE_PROCESS_CPUCLOCK(0, CPUCLOCK_SCHED)
#define THREAD_CLOCK	MAKE_THREAD_CPUCLOCK(0, CPUCLOCK_SCHED)

1416 1417
static int process_cpu_clock_getres(const clockid_t which_clock,
				    struct timespec *tp)
L
Linus Torvalds 已提交
1418 1419 1420
{
	return posix_cpu_clock_getres(PROCESS_CLOCK, tp);
}
1421 1422
static int process_cpu_clock_get(const clockid_t which_clock,
				 struct timespec *tp)
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429 1430
{
	return posix_cpu_clock_get(PROCESS_CLOCK, tp);
}
static int process_cpu_timer_create(struct k_itimer *timer)
{
	timer->it_clock = PROCESS_CLOCK;
	return posix_cpu_timer_create(timer);
}
1431
static int process_cpu_nsleep(const clockid_t which_clock, int flags,
1432 1433
			      struct timespec *rqtp,
			      struct timespec __user *rmtp)
L
Linus Torvalds 已提交
1434
{
1435
	return posix_cpu_nsleep(PROCESS_CLOCK, flags, rqtp, rmtp);
L
Linus Torvalds 已提交
1436
}
1437 1438 1439 1440
static long process_cpu_nsleep_restart(struct restart_block *restart_block)
{
	return -EINVAL;
}
1441 1442
static int thread_cpu_clock_getres(const clockid_t which_clock,
				   struct timespec *tp)
L
Linus Torvalds 已提交
1443 1444 1445
{
	return posix_cpu_clock_getres(THREAD_CLOCK, tp);
}
1446 1447
static int thread_cpu_clock_get(const clockid_t which_clock,
				struct timespec *tp)
L
Linus Torvalds 已提交
1448 1449 1450 1451 1452 1453 1454 1455 1456
{
	return posix_cpu_clock_get(THREAD_CLOCK, tp);
}
static int thread_cpu_timer_create(struct k_itimer *timer)
{
	timer->it_clock = THREAD_CLOCK;
	return posix_cpu_timer_create(timer);
}

1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468
struct k_clock clock_posix_cpu = {
	.clock_getres	= posix_cpu_clock_getres,
	.clock_set	= posix_cpu_clock_set,
	.clock_get	= posix_cpu_clock_get,
	.timer_create	= posix_cpu_timer_create,
	.nsleep		= posix_cpu_nsleep,
	.nsleep_restart	= posix_cpu_nsleep_restart,
	.timer_set	= posix_cpu_timer_set,
	.timer_del	= posix_cpu_timer_del,
	.timer_get	= posix_cpu_timer_get,
};

L
Linus Torvalds 已提交
1469 1470 1471
static __init int init_posix_cpu_timers(void)
{
	struct k_clock process = {
1472 1473 1474 1475 1476
		.clock_getres	= process_cpu_clock_getres,
		.clock_get	= process_cpu_clock_get,
		.timer_create	= process_cpu_timer_create,
		.nsleep		= process_cpu_nsleep,
		.nsleep_restart	= process_cpu_nsleep_restart,
L
Linus Torvalds 已提交
1477 1478
	};
	struct k_clock thread = {
1479 1480 1481
		.clock_getres	= thread_cpu_clock_getres,
		.clock_get	= thread_cpu_clock_get,
		.timer_create	= thread_cpu_timer_create,
L
Linus Torvalds 已提交
1482
	};
1483
	struct timespec ts;
L
Linus Torvalds 已提交
1484

1485 1486
	posix_timers_register_clock(CLOCK_PROCESS_CPUTIME_ID, &process);
	posix_timers_register_clock(CLOCK_THREAD_CPUTIME_ID, &thread);
L
Linus Torvalds 已提交
1487

1488
	cputime_to_timespec(cputime_one_jiffy, &ts);
1489 1490 1491
	onecputick = ts.tv_nsec;
	WARN_ON(ts.tv_sec != 0);

L
Linus Torvalds 已提交
1492 1493 1494
	return 0;
}
__initcall(init_posix_cpu_timers);