timer.c 45.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3
/*
 *  linux/kernel/timer.c
 *
4
 *  Kernel internal timers
L
Linus Torvalds 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *
 *  1997-01-28  Modified by Finn Arne Gangstad to make timers scale better.
 *
 *  1997-09-10  Updated NTP code according to technical memorandum Jan '96
 *              "A Kernel Model for Precision Timekeeping" by Dave Mills
 *  1998-12-24  Fixed a xtime SMP race (we need the xtime_lock rw spinlock to
 *              serialize accesses to xtime/lost_ticks).
 *                              Copyright (C) 1998  Andrea Arcangeli
 *  1999-03-10  Improved NTP compatibility by Ulrich Windl
 *  2002-05-31	Move sys_sysinfo here and make its locking sane, Robert Love
 *  2000-10-05  Implemented scalable SMP per-CPU timer handling.
 *                              Copyright (C) 2000, 2001, 2002  Ingo Molnar
 *              Designed by David S. Miller, Alexey Kuznetsov and Ingo Molnar
 */

#include <linux/kernel_stat.h>
23
#include <linux/export.h>
L
Linus Torvalds 已提交
24 25 26 27 28
#include <linux/interrupt.h>
#include <linux/percpu.h>
#include <linux/init.h>
#include <linux/mm.h>
#include <linux/swap.h>
29
#include <linux/pid_namespace.h>
L
Linus Torvalds 已提交
30 31 32 33 34 35 36
#include <linux/notifier.h>
#include <linux/thread_info.h>
#include <linux/time.h>
#include <linux/jiffies.h>
#include <linux/posix-timers.h>
#include <linux/cpu.h>
#include <linux/syscalls.h>
A
Adrian Bunk 已提交
37
#include <linux/delay.h>
38
#include <linux/tick.h>
39
#include <linux/kallsyms.h>
40
#include <linux/irq_work.h>
41
#include <linux/sched.h>
42
#include <linux/sched/sysctl.h>
43
#include <linux/slab.h>
44
#include <linux/compat.h>
L
Linus Torvalds 已提交
45 46 47 48 49 50 51

#include <asm/uaccess.h>
#include <asm/unistd.h>
#include <asm/div64.h>
#include <asm/timex.h>
#include <asm/io.h>

52 53
#include "tick-internal.h"

54 55 56
#define CREATE_TRACE_POINTS
#include <trace/events/timer.h>

A
Andi Kleen 已提交
57
__visible u64 jiffies_64 __cacheline_aligned_in_smp = INITIAL_JIFFIES;
T
Thomas Gleixner 已提交
58 59 60

EXPORT_SYMBOL(jiffies_64);

L
Linus Torvalds 已提交
61 62 63 64 65 66 67 68 69
/*
 * per-CPU timer vector definitions:
 */
#define TVN_BITS (CONFIG_BASE_SMALL ? 4 : 6)
#define TVR_BITS (CONFIG_BASE_SMALL ? 6 : 8)
#define TVN_SIZE (1 << TVN_BITS)
#define TVR_SIZE (1 << TVR_BITS)
#define TVN_MASK (TVN_SIZE - 1)
#define TVR_MASK (TVR_SIZE - 1)
70
#define MAX_TVAL ((unsigned long)((1ULL << (TVR_BITS + 4*TVN_BITS)) - 1))
L
Linus Torvalds 已提交
71

72
struct tvec {
73
	struct hlist_head vec[TVN_SIZE];
74
};
L
Linus Torvalds 已提交
75

76
struct tvec_root {
77
	struct hlist_head vec[TVR_SIZE];
78
};
L
Linus Torvalds 已提交
79

80
struct timer_base {
81 82
	spinlock_t lock;
	struct timer_list *running_timer;
83
	unsigned long clk;
84
	unsigned long next_timer;
85
	unsigned long active_timers;
86
	unsigned long all_timers;
87
	int cpu;
88
	bool migration_enabled;
89
	bool nohz_active;
90 91 92 93 94
	struct tvec_root tv1;
	struct tvec tv2;
	struct tvec tv3;
	struct tvec tv4;
	struct tvec tv5;
95
} ____cacheline_aligned;
L
Linus Torvalds 已提交
96

97

98
static DEFINE_PER_CPU(struct timer_base, timer_bases);
99

100 101 102
#if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
unsigned int sysctl_timer_migration = 1;

103
void timers_update_migration(bool update_nohz)
104 105 106 107 108
{
	bool on = sysctl_timer_migration && tick_nohz_active;
	unsigned int cpu;

	/* Avoid the loop, if nothing to update */
109
	if (this_cpu_read(timer_bases.migration_enabled) == on)
110 111 112
		return;

	for_each_possible_cpu(cpu) {
113
		per_cpu(timer_bases.migration_enabled, cpu) = on;
114
		per_cpu(hrtimer_bases.migration_enabled, cpu) = on;
115 116
		if (!update_nohz)
			continue;
117
		per_cpu(timer_bases.nohz_active, cpu) = true;
118
		per_cpu(hrtimer_bases.nohz_active, cpu) = true;
119 120 121 122 123 124 125 126 127 128 129 130 131
	}
}

int timer_migration_handler(struct ctl_table *table, int write,
			    void __user *buffer, size_t *lenp,
			    loff_t *ppos)
{
	static DEFINE_MUTEX(mutex);
	int ret;

	mutex_lock(&mutex);
	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	if (!ret && write)
132
		timers_update_migration(false);
133 134 135 136
	mutex_unlock(&mutex);
	return ret;
}

137
static inline struct timer_base *get_target_base(struct timer_base *base,
138 139 140
						int pinned)
{
	if (pinned || !base->migration_enabled)
141 142
		return this_cpu_ptr(&timer_bases);
	return per_cpu_ptr(&timer_bases, get_nohz_timer_target());
143 144
}
#else
145
static inline struct timer_base *get_target_base(struct timer_base *base,
146 147
						int pinned)
{
148
	return this_cpu_ptr(&timer_bases);
149 150 151
}
#endif

152 153
static unsigned long round_jiffies_common(unsigned long j, int cpu,
		bool force_up)
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
{
	int rem;
	unsigned long original = j;

	/*
	 * We don't want all cpus firing their timers at once hitting the
	 * same lock or cachelines, so we skew each extra cpu with an extra
	 * 3 jiffies. This 3 jiffies came originally from the mm/ code which
	 * already did this.
	 * The skew is done by adding 3*cpunr, then round, then subtract this
	 * extra offset again.
	 */
	j += cpu * 3;

	rem = j % HZ;

	/*
	 * If the target jiffie is just after a whole second (which can happen
	 * due to delays of the timer irq, long irq off times etc etc) then
	 * we should round down to the whole second, not up. Use 1/4th second
	 * as cutoff for this rounding as an extreme upper bound for this.
175
	 * But never round down if @force_up is set.
176
	 */
177
	if (rem < HZ/4 && !force_up) /* round down */
178 179 180 181 182 183 184
		j = j - rem;
	else /* round up */
		j = j - rem + HZ;

	/* now that we have rounded, subtract the extra skew again */
	j -= cpu * 3;

185 186 187 188 189
	/*
	 * Make sure j is still in the future. Otherwise return the
	 * unmodified value.
	 */
	return time_is_after_jiffies(j) ? j : original;
190
}
191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215

/**
 * __round_jiffies - function to round jiffies to a full second
 * @j: the time in (absolute) jiffies that should be rounded
 * @cpu: the processor number on which the timeout will happen
 *
 * __round_jiffies() rounds an absolute time in the future (in jiffies)
 * up or down to (approximately) full seconds. This is useful for timers
 * for which the exact time they fire does not matter too much, as long as
 * they fire approximately every X seconds.
 *
 * By rounding these timers to whole seconds, all such timers will fire
 * at the same time, rather than at various times spread out. The goal
 * of this is to have the CPU wake up less, which saves power.
 *
 * The exact rounding is skewed for each processor to avoid all
 * processors firing at the exact same time, which could lead
 * to lock contention or spurious cache line bouncing.
 *
 * The return value is the rounded version of the @j parameter.
 */
unsigned long __round_jiffies(unsigned long j, int cpu)
{
	return round_jiffies_common(j, cpu, false);
}
216 217 218 219 220 221 222
EXPORT_SYMBOL_GPL(__round_jiffies);

/**
 * __round_jiffies_relative - function to round jiffies to a full second
 * @j: the time in (relative) jiffies that should be rounded
 * @cpu: the processor number on which the timeout will happen
 *
223
 * __round_jiffies_relative() rounds a time delta  in the future (in jiffies)
224 225 226 227 228 229 230 231 232 233 234 235
 * up or down to (approximately) full seconds. This is useful for timers
 * for which the exact time they fire does not matter too much, as long as
 * they fire approximately every X seconds.
 *
 * By rounding these timers to whole seconds, all such timers will fire
 * at the same time, rather than at various times spread out. The goal
 * of this is to have the CPU wake up less, which saves power.
 *
 * The exact rounding is skewed for each processor to avoid all
 * processors firing at the exact same time, which could lead
 * to lock contention or spurious cache line bouncing.
 *
236
 * The return value is the rounded version of the @j parameter.
237 238 239
 */
unsigned long __round_jiffies_relative(unsigned long j, int cpu)
{
240 241 242 243
	unsigned long j0 = jiffies;

	/* Use j0 because jiffies might change while we run */
	return round_jiffies_common(j + j0, cpu, false) - j0;
244 245 246 247 248 249 250
}
EXPORT_SYMBOL_GPL(__round_jiffies_relative);

/**
 * round_jiffies - function to round jiffies to a full second
 * @j: the time in (absolute) jiffies that should be rounded
 *
251
 * round_jiffies() rounds an absolute time in the future (in jiffies)
252 253 254 255 256 257 258 259
 * up or down to (approximately) full seconds. This is useful for timers
 * for which the exact time they fire does not matter too much, as long as
 * they fire approximately every X seconds.
 *
 * By rounding these timers to whole seconds, all such timers will fire
 * at the same time, rather than at various times spread out. The goal
 * of this is to have the CPU wake up less, which saves power.
 *
260
 * The return value is the rounded version of the @j parameter.
261 262 263
 */
unsigned long round_jiffies(unsigned long j)
{
264
	return round_jiffies_common(j, raw_smp_processor_id(), false);
265 266 267 268 269 270 271
}
EXPORT_SYMBOL_GPL(round_jiffies);

/**
 * round_jiffies_relative - function to round jiffies to a full second
 * @j: the time in (relative) jiffies that should be rounded
 *
272
 * round_jiffies_relative() rounds a time delta  in the future (in jiffies)
273 274 275 276 277 278 279 280
 * up or down to (approximately) full seconds. This is useful for timers
 * for which the exact time they fire does not matter too much, as long as
 * they fire approximately every X seconds.
 *
 * By rounding these timers to whole seconds, all such timers will fire
 * at the same time, rather than at various times spread out. The goal
 * of this is to have the CPU wake up less, which saves power.
 *
281
 * The return value is the rounded version of the @j parameter.
282 283 284 285 286 287 288
 */
unsigned long round_jiffies_relative(unsigned long j)
{
	return __round_jiffies_relative(j, raw_smp_processor_id());
}
EXPORT_SYMBOL_GPL(round_jiffies_relative);

289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353
/**
 * __round_jiffies_up - function to round jiffies up to a full second
 * @j: the time in (absolute) jiffies that should be rounded
 * @cpu: the processor number on which the timeout will happen
 *
 * This is the same as __round_jiffies() except that it will never
 * round down.  This is useful for timeouts for which the exact time
 * of firing does not matter too much, as long as they don't fire too
 * early.
 */
unsigned long __round_jiffies_up(unsigned long j, int cpu)
{
	return round_jiffies_common(j, cpu, true);
}
EXPORT_SYMBOL_GPL(__round_jiffies_up);

/**
 * __round_jiffies_up_relative - function to round jiffies up to a full second
 * @j: the time in (relative) jiffies that should be rounded
 * @cpu: the processor number on which the timeout will happen
 *
 * This is the same as __round_jiffies_relative() except that it will never
 * round down.  This is useful for timeouts for which the exact time
 * of firing does not matter too much, as long as they don't fire too
 * early.
 */
unsigned long __round_jiffies_up_relative(unsigned long j, int cpu)
{
	unsigned long j0 = jiffies;

	/* Use j0 because jiffies might change while we run */
	return round_jiffies_common(j + j0, cpu, true) - j0;
}
EXPORT_SYMBOL_GPL(__round_jiffies_up_relative);

/**
 * round_jiffies_up - function to round jiffies up to a full second
 * @j: the time in (absolute) jiffies that should be rounded
 *
 * This is the same as round_jiffies() except that it will never
 * round down.  This is useful for timeouts for which the exact time
 * of firing does not matter too much, as long as they don't fire too
 * early.
 */
unsigned long round_jiffies_up(unsigned long j)
{
	return round_jiffies_common(j, raw_smp_processor_id(), true);
}
EXPORT_SYMBOL_GPL(round_jiffies_up);

/**
 * round_jiffies_up_relative - function to round jiffies up to a full second
 * @j: the time in (relative) jiffies that should be rounded
 *
 * This is the same as round_jiffies_relative() except that it will never
 * round down.  This is useful for timeouts for which the exact time
 * of firing does not matter too much, as long as they don't fire too
 * early.
 */
unsigned long round_jiffies_up_relative(unsigned long j)
{
	return __round_jiffies_up_relative(j, raw_smp_processor_id());
}
EXPORT_SYMBOL_GPL(round_jiffies_up_relative);

354 355
/**
 * set_timer_slack - set the allowed slack for a timer
356
 * @timer: the timer to be modified
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372
 * @slack_hz: the amount of time (in jiffies) allowed for rounding
 *
 * Set the amount of time, in jiffies, that a certain timer has
 * in terms of slack. By setting this value, the timer subsystem
 * will schedule the actual timer somewhere between
 * the time mod_timer() asks for, and that time plus the slack.
 *
 * By setting the slack to -1, a percentage of the delay is used
 * instead.
 */
void set_timer_slack(struct timer_list *timer, int slack_hz)
{
	timer->slack = slack_hz;
}
EXPORT_SYMBOL_GPL(set_timer_slack);

373
static void
374
__internal_add_timer(struct timer_base *base, struct timer_list *timer)
L
Linus Torvalds 已提交
375 376
{
	unsigned long expires = timer->expires;
377
	unsigned long idx = expires - base->clk;
378
	struct hlist_head *vec;
L
Linus Torvalds 已提交
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396

	if (idx < TVR_SIZE) {
		int i = expires & TVR_MASK;
		vec = base->tv1.vec + i;
	} else if (idx < 1 << (TVR_BITS + TVN_BITS)) {
		int i = (expires >> TVR_BITS) & TVN_MASK;
		vec = base->tv2.vec + i;
	} else if (idx < 1 << (TVR_BITS + 2 * TVN_BITS)) {
		int i = (expires >> (TVR_BITS + TVN_BITS)) & TVN_MASK;
		vec = base->tv3.vec + i;
	} else if (idx < 1 << (TVR_BITS + 3 * TVN_BITS)) {
		int i = (expires >> (TVR_BITS + 2 * TVN_BITS)) & TVN_MASK;
		vec = base->tv4.vec + i;
	} else if ((signed long) idx < 0) {
		/*
		 * Can happen if you add a timer with expires == jiffies,
		 * or you set a timer to go off in the past
		 */
397
		vec = base->tv1.vec + (base->clk & TVR_MASK);
L
Linus Torvalds 已提交
398 399
	} else {
		int i;
400 401 402
		/* If the timeout is larger than MAX_TVAL (on 64-bit
		 * architectures or with CONFIG_BASE_SMALL=1) then we
		 * use the maximum timeout.
L
Linus Torvalds 已提交
403
		 */
404 405
		if (idx > MAX_TVAL) {
			idx = MAX_TVAL;
406
			expires = idx + base->clk;
L
Linus Torvalds 已提交
407 408 409 410
		}
		i = (expires >> (TVR_BITS + 3 * TVN_BITS)) & TVN_MASK;
		vec = base->tv5.vec + i;
	}
T
Thomas Gleixner 已提交
411

412
	hlist_add_head(&timer->entry, vec);
L
Linus Torvalds 已提交
413 414
}

415
static void internal_add_timer(struct timer_base *base, struct timer_list *timer)
416
{
417 418
	/* Advance base->jiffies, if the base is empty */
	if (!base->all_timers++)
419
		base->clk = jiffies;
420

421 422
	__internal_add_timer(base, timer);
	/*
423
	 * Update base->active_timers and base->next_timer
424
	 */
425
	if (!(timer->flags & TIMER_DEFERRABLE)) {
426 427
		if (!base->active_timers++ ||
		    time_before(timer->expires, base->next_timer))
428 429
			base->next_timer = timer->expires;
	}
430 431 432 433 434 435 436 437 438 439 440 441 442 443

	/*
	 * Check whether the other CPU is in dynticks mode and needs
	 * to be triggered to reevaluate the timer wheel.
	 * We are protected against the other CPU fiddling
	 * with the timer by holding the timer base lock. This also
	 * makes sure that a CPU on the way to stop its tick can not
	 * evaluate the timer wheel.
	 *
	 * Spare the IPI for deferrable timers on idle targets though.
	 * The next busy ticks will take care of it. Except full dynticks
	 * require special care against races with idle_cpu(), lets deal
	 * with that later.
	 */
444 445 446 447 448
	if (base->nohz_active) {
		if (!(timer->flags & TIMER_DEFERRABLE) ||
		    tick_nohz_full_cpu(base->cpu))
			wake_up_nohz_cpu(base->cpu);
	}
449 450
}

451 452 453 454 455 456 457 458 459 460
#ifdef CONFIG_TIMER_STATS
void __timer_stats_timer_set_start_info(struct timer_list *timer, void *addr)
{
	if (timer->start_site)
		return;

	timer->start_site = addr;
	memcpy(timer->start_comm, current->comm, TASK_COMM_LEN);
	timer->start_pid = current->pid;
}
461 462 463

static void timer_stats_account_timer(struct timer_list *timer)
{
464 465 466 467 468 469 470 471
	void *site;

	/*
	 * start_site can be concurrently reset by
	 * timer_stats_timer_clear_start_info()
	 */
	site = READ_ONCE(timer->start_site);
	if (likely(!site))
472
		return;
473

474
	timer_stats_update_stats(timer, timer->start_pid, site,
475 476
				 timer->function, timer->start_comm,
				 timer->flags);
477 478 479 480
}

#else
static void timer_stats_account_timer(struct timer_list *timer) {}
481 482
#endif

483 484 485 486
#ifdef CONFIG_DEBUG_OBJECTS_TIMERS

static struct debug_obj_descr timer_debug_descr;

487 488 489 490 491
static void *timer_debug_hint(void *addr)
{
	return ((struct timer_list *) addr)->function;
}

492 493 494 495 496 497 498 499
static bool timer_is_static_object(void *addr)
{
	struct timer_list *timer = addr;

	return (timer->entry.pprev == NULL &&
		timer->entry.next == TIMER_ENTRY_STATIC);
}

500 501 502
/*
 * fixup_init is called when:
 * - an active object is initialized
503
 */
504
static bool timer_fixup_init(void *addr, enum debug_obj_state state)
505 506 507 508 509 510 511
{
	struct timer_list *timer = addr;

	switch (state) {
	case ODEBUG_STATE_ACTIVE:
		del_timer_sync(timer);
		debug_object_init(timer, &timer_debug_descr);
512
		return true;
513
	default:
514
		return false;
515 516 517
	}
}

518 519 520 521 522 523
/* Stub timer callback for improperly used timers. */
static void stub_timer(unsigned long data)
{
	WARN_ON(1);
}

524 525 526
/*
 * fixup_activate is called when:
 * - an active object is activated
527
 * - an unknown non-static object is activated
528
 */
529
static bool timer_fixup_activate(void *addr, enum debug_obj_state state)
530 531 532 533 534
{
	struct timer_list *timer = addr;

	switch (state) {
	case ODEBUG_STATE_NOTAVAILABLE:
535 536
		setup_timer(timer, stub_timer, 0);
		return true;
537 538 539 540 541

	case ODEBUG_STATE_ACTIVE:
		WARN_ON(1);

	default:
542
		return false;
543 544 545 546 547 548 549
	}
}

/*
 * fixup_free is called when:
 * - an active object is freed
 */
550
static bool timer_fixup_free(void *addr, enum debug_obj_state state)
551 552 553 554 555 556 557
{
	struct timer_list *timer = addr;

	switch (state) {
	case ODEBUG_STATE_ACTIVE:
		del_timer_sync(timer);
		debug_object_free(timer, &timer_debug_descr);
558
		return true;
559
	default:
560
		return false;
561 562 563
	}
}

564 565 566 567
/*
 * fixup_assert_init is called when:
 * - an untracked/uninit-ed object is found
 */
568
static bool timer_fixup_assert_init(void *addr, enum debug_obj_state state)
569 570 571 572 573
{
	struct timer_list *timer = addr;

	switch (state) {
	case ODEBUG_STATE_NOTAVAILABLE:
574 575
		setup_timer(timer, stub_timer, 0);
		return true;
576
	default:
577
		return false;
578 579 580
	}
}

581
static struct debug_obj_descr timer_debug_descr = {
582 583
	.name			= "timer_list",
	.debug_hint		= timer_debug_hint,
584
	.is_static_object	= timer_is_static_object,
585 586 587 588
	.fixup_init		= timer_fixup_init,
	.fixup_activate		= timer_fixup_activate,
	.fixup_free		= timer_fixup_free,
	.fixup_assert_init	= timer_fixup_assert_init,
589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610
};

static inline void debug_timer_init(struct timer_list *timer)
{
	debug_object_init(timer, &timer_debug_descr);
}

static inline void debug_timer_activate(struct timer_list *timer)
{
	debug_object_activate(timer, &timer_debug_descr);
}

static inline void debug_timer_deactivate(struct timer_list *timer)
{
	debug_object_deactivate(timer, &timer_debug_descr);
}

static inline void debug_timer_free(struct timer_list *timer)
{
	debug_object_free(timer, &timer_debug_descr);
}

611 612 613 614 615
static inline void debug_timer_assert_init(struct timer_list *timer)
{
	debug_object_assert_init(timer, &timer_debug_descr);
}

T
Tejun Heo 已提交
616 617
static void do_init_timer(struct timer_list *timer, unsigned int flags,
			  const char *name, struct lock_class_key *key);
618

T
Tejun Heo 已提交
619 620
void init_timer_on_stack_key(struct timer_list *timer, unsigned int flags,
			     const char *name, struct lock_class_key *key)
621 622
{
	debug_object_init_on_stack(timer, &timer_debug_descr);
T
Tejun Heo 已提交
623
	do_init_timer(timer, flags, name, key);
624
}
625
EXPORT_SYMBOL_GPL(init_timer_on_stack_key);
626 627 628 629 630 631 632 633 634 635 636

void destroy_timer_on_stack(struct timer_list *timer)
{
	debug_object_free(timer, &timer_debug_descr);
}
EXPORT_SYMBOL_GPL(destroy_timer_on_stack);

#else
static inline void debug_timer_init(struct timer_list *timer) { }
static inline void debug_timer_activate(struct timer_list *timer) { }
static inline void debug_timer_deactivate(struct timer_list *timer) { }
637
static inline void debug_timer_assert_init(struct timer_list *timer) { }
638 639
#endif

640 641 642 643 644 645 646 647 648 649
static inline void debug_init(struct timer_list *timer)
{
	debug_timer_init(timer);
	trace_timer_init(timer);
}

static inline void
debug_activate(struct timer_list *timer, unsigned long expires)
{
	debug_timer_activate(timer);
650
	trace_timer_start(timer, expires, timer->flags);
651 652 653 654 655 656 657 658
}

static inline void debug_deactivate(struct timer_list *timer)
{
	debug_timer_deactivate(timer);
	trace_timer_cancel(timer);
}

659 660 661 662 663
static inline void debug_assert_init(struct timer_list *timer)
{
	debug_timer_assert_init(timer);
}

T
Tejun Heo 已提交
664 665
static void do_init_timer(struct timer_list *timer, unsigned int flags,
			  const char *name, struct lock_class_key *key)
666
{
667
	timer->entry.pprev = NULL;
668
	timer->flags = flags | raw_smp_processor_id();
669
	timer->slack = -1;
670 671 672 673 674
#ifdef CONFIG_TIMER_STATS
	timer->start_site = NULL;
	timer->start_pid = -1;
	memset(timer->start_comm, 0, TASK_COMM_LEN);
#endif
675
	lockdep_init_map(&timer->lockdep_map, name, key, 0);
676
}
677 678

/**
R
Randy Dunlap 已提交
679
 * init_timer_key - initialize a timer
680
 * @timer: the timer to be initialized
T
Tejun Heo 已提交
681
 * @flags: timer flags
R
Randy Dunlap 已提交
682 683 684
 * @name: name of the timer
 * @key: lockdep class key of the fake lock used for tracking timer
 *       sync lock dependencies
685
 *
R
Randy Dunlap 已提交
686
 * init_timer_key() must be done to a timer prior calling *any* of the
687 688
 * other timer functions.
 */
T
Tejun Heo 已提交
689 690
void init_timer_key(struct timer_list *timer, unsigned int flags,
		    const char *name, struct lock_class_key *key)
691
{
692
	debug_init(timer);
T
Tejun Heo 已提交
693
	do_init_timer(timer, flags, name, key);
694
}
695
EXPORT_SYMBOL(init_timer_key);
696

697
static inline void detach_timer(struct timer_list *timer, bool clear_pending)
698
{
699
	struct hlist_node *entry = &timer->entry;
700

701
	debug_deactivate(timer);
702

703
	__hlist_del(entry);
704
	if (clear_pending)
705 706
		entry->pprev = NULL;
	entry->next = LIST_POISON2;
707 708
}

709
static inline void
710
detach_expired_timer(struct timer_list *timer, struct timer_base *base)
711 712
{
	detach_timer(timer, true);
713
	if (!(timer->flags & TIMER_DEFERRABLE))
714
		base->active_timers--;
715
	base->all_timers--;
716 717
}

718
static int detach_if_pending(struct timer_list *timer, struct timer_base *base,
719 720 721 722 723 724
			     bool clear_pending)
{
	if (!timer_pending(timer))
		return 0;

	detach_timer(timer, clear_pending);
725
	if (!(timer->flags & TIMER_DEFERRABLE)) {
726
		base->active_timers--;
727
		if (timer->expires == base->next_timer)
728
			base->next_timer = base->clk;
729
	}
730 731
	/* If this was the last timer, advance base->jiffies */
	if (!--base->all_timers)
732
		base->clk = jiffies;
733 734 735
	return 1;
}

736
/*
737
 * We are using hashed locking: holding per_cpu(timer_bases).lock
738 739 740 741 742 743
 * means that all timers which are tied to this base via timer->base are
 * locked, and the base itself is locked too.
 *
 * So __run_timers/migrate_timers can safely modify all timers which could
 * be found on ->tvX lists.
 *
744 745
 * When the timer's base is locked and removed from the list, the
 * TIMER_MIGRATING flag is set, FIXME
746
 */
747
static struct timer_base *lock_timer_base(struct timer_list *timer,
748
					unsigned long *flags)
749
	__acquires(timer->base->lock)
750 751
{
	for (;;) {
752
		u32 tf = timer->flags;
753
		struct timer_base *base;
754 755

		if (!(tf & TIMER_MIGRATING)) {
756
			base = per_cpu_ptr(&timer_bases, tf & TIMER_CPUMASK);
757
			spin_lock_irqsave(&base->lock, *flags);
758
			if (timer->flags == tf)
759 760 761 762 763 764 765
				return base;
			spin_unlock_irqrestore(&base->lock, *flags);
		}
		cpu_relax();
	}
}

I
Ingo Molnar 已提交
766
static inline int
767
__mod_timer(struct timer_list *timer, unsigned long expires, bool pending_only)
L
Linus Torvalds 已提交
768
{
769
	struct timer_base *base, *new_base;
L
Linus Torvalds 已提交
770
	unsigned long flags;
771
	int ret = 0;
L
Linus Torvalds 已提交
772

773
	timer_stats_timer_set_start_info(timer);
L
Linus Torvalds 已提交
774 775
	BUG_ON(!timer->function);

776 777
	base = lock_timer_base(timer, &flags);

778 779 780
	ret = detach_if_pending(timer, base, false);
	if (!ret && pending_only)
		goto out_unlock;
781

782
	debug_activate(timer, expires);
783

784
	new_base = get_target_base(base, timer->flags & TIMER_PINNED);
785

786
	if (base != new_base) {
L
Linus Torvalds 已提交
787
		/*
788 789 790 791 792
		 * We are trying to schedule the timer on the local CPU.
		 * However we can't change timer's base while it is running,
		 * otherwise del_timer_sync() can't detect that the timer's
		 * handler yet has not finished. This also guarantees that
		 * the timer is serialized wrt itself.
L
Linus Torvalds 已提交
793
		 */
794
		if (likely(base->running_timer != timer)) {
795
			/* See the comment in lock_timer_base() */
796 797
			timer->flags |= TIMER_MIGRATING;

798
			spin_unlock(&base->lock);
799 800
			base = new_base;
			spin_lock(&base->lock);
801 802
			WRITE_ONCE(timer->flags,
				   (timer->flags & ~TIMER_BASEMASK) | base->cpu);
L
Linus Torvalds 已提交
803 804 805 806
		}
	}

	timer->expires = expires;
807
	internal_add_timer(base, timer);
I
Ingo Molnar 已提交
808 809

out_unlock:
810
	spin_unlock_irqrestore(&base->lock, flags);
L
Linus Torvalds 已提交
811 812 813 814

	return ret;
}

815
/**
I
Ingo Molnar 已提交
816 817 818
 * mod_timer_pending - modify a pending timer's timeout
 * @timer: the pending timer to be modified
 * @expires: new timeout in jiffies
L
Linus Torvalds 已提交
819
 *
I
Ingo Molnar 已提交
820 821 822 823
 * mod_timer_pending() is the same for pending timers as mod_timer(),
 * but will not re-activate and modify already deleted timers.
 *
 * It is useful for unserialized use of timers.
L
Linus Torvalds 已提交
824
 */
I
Ingo Molnar 已提交
825
int mod_timer_pending(struct timer_list *timer, unsigned long expires)
L
Linus Torvalds 已提交
826
{
827
	return __mod_timer(timer, expires, true);
L
Linus Torvalds 已提交
828
}
I
Ingo Molnar 已提交
829
EXPORT_SYMBOL(mod_timer_pending);
L
Linus Torvalds 已提交
830

831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846
/*
 * Decide where to put the timer while taking the slack into account
 *
 * Algorithm:
 *   1) calculate the maximum (absolute) time
 *   2) calculate the highest bit where the expires and new max are different
 *   3) use this bit to make a mask
 *   4) use the bitmask to round down the maximum time, so that all last
 *      bits are zeros
 */
static inline
unsigned long apply_slack(struct timer_list *timer, unsigned long expires)
{
	unsigned long expires_limit, mask;
	int bit;

847
	if (timer->slack >= 0) {
848
		expires_limit = expires + timer->slack;
849
	} else {
850 851 852 853
		long delta = expires - jiffies;

		if (delta < 256)
			return expires;
854

855
		expires_limit = expires + delta / 256;
856
	}
857 858 859 860
	mask = expires ^ expires_limit;
	if (mask == 0)
		return expires;

861
	bit = __fls(mask);
862

863
	mask = (1UL << bit) - 1;
864 865 866 867 868 869

	expires_limit = expires_limit & ~(mask);

	return expires_limit;
}

870
/**
L
Linus Torvalds 已提交
871 872
 * mod_timer - modify a timer's timeout
 * @timer: the timer to be modified
873
 * @expires: new timeout in jiffies
L
Linus Torvalds 已提交
874
 *
875
 * mod_timer() is a more efficient way to update the expire field of an
L
Linus Torvalds 已提交
876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891
 * active timer (if the timer is inactive it will be activated)
 *
 * mod_timer(timer, expires) is equivalent to:
 *
 *     del_timer(timer); timer->expires = expires; add_timer(timer);
 *
 * Note that if there are multiple unserialized concurrent users of the
 * same timer, then mod_timer() is the only safe way to modify the timeout,
 * since add_timer() cannot modify an already running timer.
 *
 * The function returns whether it has modified a pending timer or not.
 * (ie. mod_timer() of an inactive timer returns 0, mod_timer() of an
 * active timer returns 1.)
 */
int mod_timer(struct timer_list *timer, unsigned long expires)
{
892 893
	expires = apply_slack(timer, expires);

L
Linus Torvalds 已提交
894 895 896 897 898
	/*
	 * This is a common optimization triggered by the
	 * networking code - if the timer is re-modified
	 * to be the same thing then just return:
	 */
899
	if (timer_pending(timer) && timer->expires == expires)
L
Linus Torvalds 已提交
900 901
		return 1;

902
	return __mod_timer(timer, expires, false);
L
Linus Torvalds 已提交
903 904 905
}
EXPORT_SYMBOL(mod_timer);

I
Ingo Molnar 已提交
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
/**
 * add_timer - start a timer
 * @timer: the timer to be added
 *
 * The kernel will do a ->function(->data) callback from the
 * timer interrupt at the ->expires point in the future. The
 * current time is 'jiffies'.
 *
 * The timer's ->expires, ->function (and if the handler uses it, ->data)
 * fields must be set prior calling this function.
 *
 * Timers with an ->expires field in the past will be executed in the next
 * timer tick.
 */
void add_timer(struct timer_list *timer)
{
	BUG_ON(timer_pending(timer));
	mod_timer(timer, timer->expires);
}
EXPORT_SYMBOL(add_timer);

/**
 * add_timer_on - start a timer on a particular CPU
 * @timer: the timer to be added
 * @cpu: the CPU to start it on
 *
 * This is not very scalable on SMP. Double adds are not possible.
 */
void add_timer_on(struct timer_list *timer, int cpu)
{
936 937
	struct timer_base *new_base = per_cpu_ptr(&timer_bases, cpu);
	struct timer_base *base;
I
Ingo Molnar 已提交
938 939 940 941
	unsigned long flags;

	timer_stats_timer_set_start_info(timer);
	BUG_ON(timer_pending(timer) || !timer->function);
942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958

	/*
	 * If @timer was on a different CPU, it should be migrated with the
	 * old base locked to prevent other operations proceeding with the
	 * wrong base locked.  See lock_timer_base().
	 */
	base = lock_timer_base(timer, &flags);
	if (base != new_base) {
		timer->flags |= TIMER_MIGRATING;

		spin_unlock(&base->lock);
		base = new_base;
		spin_lock(&base->lock);
		WRITE_ONCE(timer->flags,
			   (timer->flags & ~TIMER_BASEMASK) | cpu);
	}

959
	debug_activate(timer, timer->expires);
I
Ingo Molnar 已提交
960 961 962
	internal_add_timer(base, timer);
	spin_unlock_irqrestore(&base->lock, flags);
}
A
Andi Kleen 已提交
963
EXPORT_SYMBOL_GPL(add_timer_on);
I
Ingo Molnar 已提交
964

965
/**
L
Linus Torvalds 已提交
966 967 968 969 970 971 972 973 974 975 976 977
 * del_timer - deactive a timer.
 * @timer: the timer to be deactivated
 *
 * del_timer() deactivates a timer - this works on both active and inactive
 * timers.
 *
 * The function returns whether it has deactivated a pending timer or not.
 * (ie. del_timer() of an inactive timer returns 0, del_timer() of an
 * active timer returns 1.)
 */
int del_timer(struct timer_list *timer)
{
978
	struct timer_base *base;
L
Linus Torvalds 已提交
979
	unsigned long flags;
980
	int ret = 0;
L
Linus Torvalds 已提交
981

982 983
	debug_assert_init(timer);

984
	timer_stats_timer_clear_start_info(timer);
985 986
	if (timer_pending(timer)) {
		base = lock_timer_base(timer, &flags);
987
		ret = detach_if_pending(timer, base, true);
L
Linus Torvalds 已提交
988 989 990
		spin_unlock_irqrestore(&base->lock, flags);
	}

991
	return ret;
L
Linus Torvalds 已提交
992 993 994
}
EXPORT_SYMBOL(del_timer);

995 996 997 998
/**
 * try_to_del_timer_sync - Try to deactivate a timer
 * @timer: timer do del
 *
999 1000 1001 1002 1003
 * This function tries to deactivate a timer. Upon successful (ret >= 0)
 * exit the timer is not queued and the handler is not running on any CPU.
 */
int try_to_del_timer_sync(struct timer_list *timer)
{
1004
	struct timer_base *base;
1005 1006 1007
	unsigned long flags;
	int ret = -1;

1008 1009
	debug_assert_init(timer);

1010 1011
	base = lock_timer_base(timer, &flags);

1012 1013 1014
	if (base->running_timer != timer) {
		timer_stats_timer_clear_start_info(timer);
		ret = detach_if_pending(timer, base, true);
1015 1016 1017 1018 1019
	}
	spin_unlock_irqrestore(&base->lock, flags);

	return ret;
}
1020 1021
EXPORT_SYMBOL(try_to_del_timer_sync);

1022
#ifdef CONFIG_SMP
1023
/**
L
Linus Torvalds 已提交
1024 1025 1026 1027 1028 1029 1030
 * del_timer_sync - deactivate a timer and wait for the handler to finish.
 * @timer: the timer to be deactivated
 *
 * This function only differs from del_timer() on SMP: besides deactivating
 * the timer it also makes sure the handler has finished executing on other
 * CPUs.
 *
1031
 * Synchronization rules: Callers must prevent restarting of the timer,
L
Linus Torvalds 已提交
1032
 * otherwise this function is meaningless. It must not be called from
T
Tejun Heo 已提交
1033 1034 1035 1036
 * interrupt contexts unless the timer is an irqsafe one. The caller must
 * not hold locks which would prevent completion of the timer's
 * handler. The timer's handler must not call add_timer_on(). Upon exit the
 * timer is not queued and the handler is not running on any CPU.
L
Linus Torvalds 已提交
1037
 *
T
Tejun Heo 已提交
1038 1039 1040
 * Note: For !irqsafe timers, you must not hold locks that are held in
 *   interrupt context while calling this function. Even if the lock has
 *   nothing to do with the timer in question.  Here's why:
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056
 *
 *    CPU0                             CPU1
 *    ----                             ----
 *                                   <SOFTIRQ>
 *                                   call_timer_fn();
 *                                     base->running_timer = mytimer;
 *  spin_lock_irq(somelock);
 *                                     <IRQ>
 *                                        spin_lock(somelock);
 *  del_timer_sync(mytimer);
 *   while (base->running_timer == mytimer);
 *
 * Now del_timer_sync() will never return and never release somelock.
 * The interrupt on the other CPU is waiting to grab somelock but
 * it has interrupted the softirq that CPU0 is waiting to finish.
 *
L
Linus Torvalds 已提交
1057 1058 1059 1060
 * The function returns whether it has deactivated a pending timer or not.
 */
int del_timer_sync(struct timer_list *timer)
{
1061
#ifdef CONFIG_LOCKDEP
1062 1063
	unsigned long flags;

1064 1065 1066 1067
	/*
	 * If lockdep gives a backtrace here, please reference
	 * the synchronization rules above.
	 */
1068
	local_irq_save(flags);
1069 1070
	lock_map_acquire(&timer->lockdep_map);
	lock_map_release(&timer->lockdep_map);
1071
	local_irq_restore(flags);
1072
#endif
1073 1074 1075 1076
	/*
	 * don't use it in hardirq context, because it
	 * could lead to deadlock.
	 */
1077
	WARN_ON(in_irq() && !(timer->flags & TIMER_IRQSAFE));
1078 1079 1080 1081
	for (;;) {
		int ret = try_to_del_timer_sync(timer);
		if (ret >= 0)
			return ret;
1082
		cpu_relax();
1083
	}
L
Linus Torvalds 已提交
1084
}
1085
EXPORT_SYMBOL(del_timer_sync);
L
Linus Torvalds 已提交
1086 1087
#endif

1088
static int cascade(struct timer_base *base, struct tvec *tv, int index)
L
Linus Torvalds 已提交
1089 1090
{
	/* cascade all the timers from tv up one level */
1091 1092 1093
	struct timer_list *timer;
	struct hlist_node *tmp;
	struct hlist_head tv_list;
1094

1095
	hlist_move_list(tv->vec + index, &tv_list);
L
Linus Torvalds 已提交
1096 1097

	/*
1098 1099
	 * We are removing _all_ timers from the list, so we
	 * don't have to detach them individually.
L
Linus Torvalds 已提交
1100
	 */
1101
	hlist_for_each_entry_safe(timer, tmp, &tv_list, entry) {
1102 1103
		/* No accounting, while moving them */
		__internal_add_timer(base, timer);
L
Linus Torvalds 已提交
1104 1105 1106 1107 1108
	}

	return index;
}

1109 1110 1111
static void call_timer_fn(struct timer_list *timer, void (*fn)(unsigned long),
			  unsigned long data)
{
1112
	int count = preempt_count();
1113 1114 1115 1116 1117 1118 1119 1120 1121

#ifdef CONFIG_LOCKDEP
	/*
	 * It is permissible to free the timer from inside the
	 * function that is called from it, this we need to take into
	 * account for lockdep too. To avoid bogus "held lock freed"
	 * warnings as well as problems when looking into
	 * timer->lockdep_map, make a copy and use that here.
	 */
1122 1123 1124
	struct lockdep_map lockdep_map;

	lockdep_copy_map(&lockdep_map, &timer->lockdep_map);
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
#endif
	/*
	 * Couple the lock chain with the lock chain at
	 * del_timer_sync() by acquiring the lock_map around the fn()
	 * call here and in del_timer_sync().
	 */
	lock_map_acquire(&lockdep_map);

	trace_timer_expire_entry(timer);
	fn(data);
	trace_timer_expire_exit(timer);

	lock_map_release(&lockdep_map);

1139
	if (count != preempt_count()) {
1140
		WARN_ONCE(1, "timer: %pF preempt leak: %08x -> %08x\n",
1141
			  fn, count, preempt_count());
1142 1143 1144 1145 1146 1147
		/*
		 * Restore the preempt count. That gives us a decent
		 * chance to survive and extract information. If the
		 * callback kept a lock held, bad luck, but not worse
		 * than the BUG() we had.
		 */
1148
		preempt_count_set(count);
1149 1150 1151
	}
}

1152
#define INDEX(N) ((base->clk >> (TVR_BITS + (N) * TVN_BITS)) & TVN_MASK)
1153 1154

/**
L
Linus Torvalds 已提交
1155 1156 1157 1158 1159 1160
 * __run_timers - run all expired timers (if any) on this CPU.
 * @base: the timer vector to be processed.
 *
 * This function cascades all vectors and executes all expired timer
 * vectors.
 */
1161
static inline void __run_timers(struct timer_base *base)
L
Linus Torvalds 已提交
1162 1163 1164
{
	struct timer_list *timer;

1165
	spin_lock_irq(&base->lock);
1166

1167
	while (time_after_eq(jiffies, base->clk)) {
1168 1169
		struct hlist_head work_list;
		struct hlist_head *head = &work_list;
1170 1171 1172
		int index;

		if (!base->all_timers) {
1173
			base->clk = jiffies;
1174 1175 1176
			break;
		}

1177
		index = base->clk & TVR_MASK;
1178

L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184 1185 1186
		/*
		 * Cascade timers:
		 */
		if (!index &&
			(!cascade(base, &base->tv2, INDEX(0))) &&
				(!cascade(base, &base->tv3, INDEX(1))) &&
					!cascade(base, &base->tv4, INDEX(2)))
			cascade(base, &base->tv5, INDEX(3));
1187
		++base->clk;
1188 1189
		hlist_move_list(base->tv1.vec + index, head);
		while (!hlist_empty(head)) {
L
Linus Torvalds 已提交
1190 1191
			void (*fn)(unsigned long);
			unsigned long data;
T
Tejun Heo 已提交
1192
			bool irqsafe;
L
Linus Torvalds 已提交
1193

1194
			timer = hlist_entry(head->first, struct timer_list, entry);
1195 1196
			fn = timer->function;
			data = timer->data;
1197
			irqsafe = timer->flags & TIMER_IRQSAFE;
L
Linus Torvalds 已提交
1198

1199 1200
			timer_stats_account_timer(timer);

1201
			base->running_timer = timer;
1202
			detach_expired_timer(timer, base);
1203

T
Tejun Heo 已提交
1204 1205 1206 1207 1208 1209 1210 1211 1212
			if (irqsafe) {
				spin_unlock(&base->lock);
				call_timer_fn(timer, fn, data);
				spin_lock(&base->lock);
			} else {
				spin_unlock_irq(&base->lock);
				call_timer_fn(timer, fn, data);
				spin_lock_irq(&base->lock);
			}
L
Linus Torvalds 已提交
1213 1214
		}
	}
1215
	base->running_timer = NULL;
1216
	spin_unlock_irq(&base->lock);
L
Linus Torvalds 已提交
1217 1218
}

1219
#ifdef CONFIG_NO_HZ_COMMON
L
Linus Torvalds 已提交
1220 1221
/*
 * Find out when the next timer event is due to happen. This
R
Randy Dunlap 已提交
1222 1223
 * is used on S/390 to stop all activity when a CPU is idle.
 * This function needs to be called with interrupts disabled.
L
Linus Torvalds 已提交
1224
 */
1225
static unsigned long __next_timer_interrupt(struct timer_base *base)
L
Linus Torvalds 已提交
1226
{
1227 1228
	unsigned long clk = base->clk;
	unsigned long expires = clk + NEXT_TIMER_MAX_DELTA;
1229
	int index, slot, array, found = 0;
L
Linus Torvalds 已提交
1230
	struct timer_list *nte;
1231
	struct tvec *varray[4];
L
Linus Torvalds 已提交
1232 1233

	/* Look for timer events in tv1. */
1234
	index = slot = clk & TVR_MASK;
L
Linus Torvalds 已提交
1235
	do {
1236
		hlist_for_each_entry(nte, base->tv1.vec + slot, entry) {
1237
			if (nte->flags & TIMER_DEFERRABLE)
1238
				continue;
1239

1240
			found = 1;
L
Linus Torvalds 已提交
1241
			expires = nte->expires;
1242 1243 1244 1245
			/* Look at the cascade bucket(s)? */
			if (!index || slot < index)
				goto cascade;
			return expires;
L
Linus Torvalds 已提交
1246
		}
1247 1248 1249 1250 1251 1252
		slot = (slot + 1) & TVR_MASK;
	} while (slot != index);

cascade:
	/* Calculate the next cascade event */
	if (index)
1253 1254
		clk += TVR_SIZE - index;
	clk >>= TVR_BITS;
L
Linus Torvalds 已提交
1255 1256 1257 1258 1259 1260

	/* Check tv2-tv5. */
	varray[0] = &base->tv2;
	varray[1] = &base->tv3;
	varray[2] = &base->tv4;
	varray[3] = &base->tv5;
1261 1262

	for (array = 0; array < 4; array++) {
1263
		struct tvec *varp = varray[array];
1264

1265
		index = slot = clk & TVN_MASK;
L
Linus Torvalds 已提交
1266
		do {
1267
			hlist_for_each_entry(nte, varp->vec + slot, entry) {
1268
				if (nte->flags & TIMER_DEFERRABLE)
1269 1270
					continue;

1271
				found = 1;
L
Linus Torvalds 已提交
1272 1273
				if (time_before(nte->expires, expires))
					expires = nte->expires;
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288
			}
			/*
			 * Do we still search for the first timer or are
			 * we looking up the cascade buckets ?
			 */
			if (found) {
				/* Look at the cascade bucket(s)? */
				if (!index || slot < index)
					break;
				return expires;
			}
			slot = (slot + 1) & TVN_MASK;
		} while (slot != index);

		if (index)
1289 1290
			clk += TVN_SIZE - index;
		clk >>= TVN_BITS;
L
Linus Torvalds 已提交
1291
	}
1292 1293
	return expires;
}
1294

1295 1296 1297 1298
/*
 * Check, if the next hrtimer event is before the next timer wheel
 * event:
 */
1299
static u64 cmp_next_hrtimer_event(u64 basem, u64 expires)
1300
{
1301
	u64 nextevt = hrtimer_get_next_event();
1302

1303
	/*
1304 1305
	 * If high resolution timers are enabled
	 * hrtimer_get_next_event() returns KTIME_MAX.
1306
	 */
1307 1308
	if (expires <= nextevt)
		return expires;
1309 1310

	/*
1311 1312
	 * If the next timer is already expired, return the tick base
	 * time so the tick is fired immediately.
1313
	 */
1314 1315
	if (nextevt <= basem)
		return basem;
1316

1317
	/*
1318 1319 1320 1321 1322 1323
	 * Round up to the next jiffie. High resolution timers are
	 * off, so the hrtimers are expired in the tick and we need to
	 * make sure that this tick really expires the timer to avoid
	 * a ping pong of the nohz stop code.
	 *
	 * Use DIV_ROUND_UP_ULL to prevent gcc calling __divdi3
1324
	 */
1325
	return DIV_ROUND_UP_ULL(nextevt, TICK_NSEC) * TICK_NSEC;
L
Linus Torvalds 已提交
1326
}
1327 1328

/**
1329 1330 1331 1332 1333 1334
 * get_next_timer_interrupt - return the time (clock mono) of the next timer
 * @basej:	base time jiffies
 * @basem:	base time clock monotonic
 *
 * Returns the tick aligned clock monotonic time of the next pending
 * timer or KTIME_MAX if no timer is pending.
1335
 */
1336
u64 get_next_timer_interrupt(unsigned long basej, u64 basem)
1337
{
1338
	struct timer_base *base = this_cpu_ptr(&timer_bases);
1339 1340
	u64 expires = KTIME_MAX;
	unsigned long nextevt;
1341

1342 1343 1344 1345 1346
	/*
	 * Pretend that there is no timer pending if the cpu is offline.
	 * Possible pending timers will be migrated later to an active cpu.
	 */
	if (cpu_is_offline(smp_processor_id()))
1347 1348
		return expires;

1349
	spin_lock(&base->lock);
1350
	if (base->active_timers) {
1351
		if (time_before_eq(base->next_timer, base->clk))
1352
			base->next_timer = __next_timer_interrupt(base);
1353 1354 1355 1356 1357
		nextevt = base->next_timer;
		if (time_before_eq(nextevt, basej))
			expires = basem;
		else
			expires = basem + (nextevt - basej) * TICK_NSEC;
1358
	}
1359 1360
	spin_unlock(&base->lock);

1361
	return cmp_next_hrtimer_event(basem, expires);
1362
}
L
Linus Torvalds 已提交
1363 1364 1365
#endif

/*
D
Daniel Walker 已提交
1366
 * Called from the timer interrupt handler to charge one tick to the current
L
Linus Torvalds 已提交
1367 1368 1369 1370 1371 1372 1373
 * process.  user_tick is 1 if the tick is user time, 0 for system.
 */
void update_process_times(int user_tick)
{
	struct task_struct *p = current;

	/* Note: this timer irq context must be accounted for as well. */
1374
	account_process_tick(p, user_tick);
L
Linus Torvalds 已提交
1375
	run_local_timers();
1376
	rcu_check_callbacks(user_tick);
1377 1378
#ifdef CONFIG_IRQ_WORK
	if (in_irq())
1379
		irq_work_tick();
1380
#endif
L
Linus Torvalds 已提交
1381
	scheduler_tick();
1382
	run_posix_cpu_timers(p);
L
Linus Torvalds 已提交
1383 1384 1385 1386 1387 1388 1389
}

/*
 * This function runs timers and the timer-tq in bottom half context.
 */
static void run_timer_softirq(struct softirq_action *h)
{
1390
	struct timer_base *base = this_cpu_ptr(&timer_bases);
L
Linus Torvalds 已提交
1391

1392
	if (time_after_eq(jiffies, base->clk))
L
Linus Torvalds 已提交
1393 1394 1395 1396 1397 1398 1399 1400
		__run_timers(base);
}

/*
 * Called by the local, per-CPU timer interrupt on SMP.
 */
void run_local_timers(void)
{
1401
	hrtimer_run_queues();
L
Linus Torvalds 已提交
1402 1403 1404 1405 1406 1407 1408 1409 1410
	raise_softirq(TIMER_SOFTIRQ);
}

#ifdef __ARCH_WANT_SYS_ALARM

/*
 * For backwards compatibility?  This can be done in libc so Alpha
 * and all newer ports shouldn't need it.
 */
1411
SYSCALL_DEFINE1(alarm, unsigned int, seconds)
L
Linus Torvalds 已提交
1412
{
1413
	return alarm_setitimer(seconds);
L
Linus Torvalds 已提交
1414 1415 1416 1417 1418 1419
}

#endif

static void process_timeout(unsigned long __data)
{
1420
	wake_up_process((struct task_struct *)__data);
L
Linus Torvalds 已提交
1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448
}

/**
 * schedule_timeout - sleep until timeout
 * @timeout: timeout value in jiffies
 *
 * Make the current task sleep until @timeout jiffies have
 * elapsed. The routine will return immediately unless
 * the current task state has been set (see set_current_state()).
 *
 * You can set the task state as follows -
 *
 * %TASK_UNINTERRUPTIBLE - at least @timeout jiffies are guaranteed to
 * pass before the routine returns. The routine will return 0
 *
 * %TASK_INTERRUPTIBLE - the routine may return early if a signal is
 * delivered to the current task. In this case the remaining time
 * in jiffies will be returned, or 0 if the timer expired in time
 *
 * The current task state is guaranteed to be TASK_RUNNING when this
 * routine returns.
 *
 * Specifying a @timeout value of %MAX_SCHEDULE_TIMEOUT will schedule
 * the CPU away without a bound on the timeout. In this case the return
 * value will be %MAX_SCHEDULE_TIMEOUT.
 *
 * In all cases the return value is guaranteed to be non-negative.
 */
1449
signed long __sched schedule_timeout(signed long timeout)
L
Linus Torvalds 已提交
1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473
{
	struct timer_list timer;
	unsigned long expire;

	switch (timeout)
	{
	case MAX_SCHEDULE_TIMEOUT:
		/*
		 * These two special cases are useful to be comfortable
		 * in the caller. Nothing more. We could take
		 * MAX_SCHEDULE_TIMEOUT from one of the negative value
		 * but I' d like to return a valid offset (>=0) to allow
		 * the caller to do everything it want with the retval.
		 */
		schedule();
		goto out;
	default:
		/*
		 * Another bit of PARANOID. Note that the retval will be
		 * 0 since no piece of kernel is supposed to do a check
		 * for a negative retval of schedule_timeout() (since it
		 * should never happens anyway). You just have the printk()
		 * that will tell you if something is gone wrong and where.
		 */
1474
		if (timeout < 0) {
L
Linus Torvalds 已提交
1475
			printk(KERN_ERR "schedule_timeout: wrong timeout "
1476 1477
				"value %lx\n", timeout);
			dump_stack();
L
Linus Torvalds 已提交
1478 1479 1480 1481 1482 1483 1484
			current->state = TASK_RUNNING;
			goto out;
		}
	}

	expire = timeout + jiffies;

1485
	setup_timer_on_stack(&timer, process_timeout, (unsigned long)current);
1486
	__mod_timer(&timer, expire, false);
L
Linus Torvalds 已提交
1487 1488 1489
	schedule();
	del_singleshot_timer_sync(&timer);

1490 1491 1492
	/* Remove the timer from the object tracker */
	destroy_timer_on_stack(&timer);

L
Linus Torvalds 已提交
1493 1494 1495 1496 1497 1498 1499
	timeout = expire - jiffies;

 out:
	return timeout < 0 ? 0 : timeout;
}
EXPORT_SYMBOL(schedule_timeout);

1500 1501 1502 1503
/*
 * We can use __set_current_state() here because schedule_timeout() calls
 * schedule() unconditionally.
 */
1504 1505
signed long __sched schedule_timeout_interruptible(signed long timeout)
{
A
Andrew Morton 已提交
1506 1507
	__set_current_state(TASK_INTERRUPTIBLE);
	return schedule_timeout(timeout);
1508 1509 1510
}
EXPORT_SYMBOL(schedule_timeout_interruptible);

M
Matthew Wilcox 已提交
1511 1512 1513 1514 1515 1516 1517
signed long __sched schedule_timeout_killable(signed long timeout)
{
	__set_current_state(TASK_KILLABLE);
	return schedule_timeout(timeout);
}
EXPORT_SYMBOL(schedule_timeout_killable);

1518 1519
signed long __sched schedule_timeout_uninterruptible(signed long timeout)
{
A
Andrew Morton 已提交
1520 1521
	__set_current_state(TASK_UNINTERRUPTIBLE);
	return schedule_timeout(timeout);
1522 1523 1524
}
EXPORT_SYMBOL(schedule_timeout_uninterruptible);

1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
/*
 * Like schedule_timeout_uninterruptible(), except this task will not contribute
 * to load average.
 */
signed long __sched schedule_timeout_idle(signed long timeout)
{
	__set_current_state(TASK_IDLE);
	return schedule_timeout(timeout);
}
EXPORT_SYMBOL(schedule_timeout_idle);

L
Linus Torvalds 已提交
1536
#ifdef CONFIG_HOTPLUG_CPU
1537
static void migrate_timer_list(struct timer_base *new_base, struct hlist_head *head)
L
Linus Torvalds 已提交
1538 1539
{
	struct timer_list *timer;
1540
	int cpu = new_base->cpu;
L
Linus Torvalds 已提交
1541

1542 1543
	while (!hlist_empty(head)) {
		timer = hlist_entry(head->first, struct timer_list, entry);
1544
		/* We ignore the accounting on the dying cpu */
1545
		detach_timer(timer, false);
1546
		timer->flags = (timer->flags & ~TIMER_BASEMASK) | cpu;
L
Linus Torvalds 已提交
1547 1548 1549 1550
		internal_add_timer(new_base, timer);
	}
}

1551
static void migrate_timers(int cpu)
L
Linus Torvalds 已提交
1552
{
1553 1554
	struct timer_base *old_base;
	struct timer_base *new_base;
L
Linus Torvalds 已提交
1555 1556 1557
	int i;

	BUG_ON(cpu_online(cpu));
1558 1559
	old_base = per_cpu_ptr(&timer_bases, cpu);
	new_base = get_cpu_ptr(&timer_bases);
1560 1561 1562 1563 1564
	/*
	 * The caller is globally serialized and nobody else
	 * takes two locks at once, deadlock is not possible.
	 */
	spin_lock_irq(&new_base->lock);
1565
	spin_lock_nested(&old_base->lock, SINGLE_DEPTH_NESTING);
1566 1567

	BUG_ON(old_base->running_timer);
L
Linus Torvalds 已提交
1568 1569

	for (i = 0; i < TVR_SIZE; i++)
1570 1571 1572 1573 1574 1575 1576 1577
		migrate_timer_list(new_base, old_base->tv1.vec + i);
	for (i = 0; i < TVN_SIZE; i++) {
		migrate_timer_list(new_base, old_base->tv2.vec + i);
		migrate_timer_list(new_base, old_base->tv3.vec + i);
		migrate_timer_list(new_base, old_base->tv4.vec + i);
		migrate_timer_list(new_base, old_base->tv5.vec + i);
	}

1578 1579 1580
	old_base->active_timers = 0;
	old_base->all_timers = 0;

1581
	spin_unlock(&old_base->lock);
1582
	spin_unlock_irq(&new_base->lock);
1583
	put_cpu_ptr(&timer_bases);
L
Linus Torvalds 已提交
1584 1585
}

1586
static int timer_cpu_notify(struct notifier_block *self,
L
Linus Torvalds 已提交
1587 1588
				unsigned long action, void *hcpu)
{
1589
	switch (action) {
L
Linus Torvalds 已提交
1590
	case CPU_DEAD:
1591
	case CPU_DEAD_FROZEN:
1592
		migrate_timers((long)hcpu);
L
Linus Torvalds 已提交
1593 1594 1595 1596
		break;
	default:
		break;
	}
1597

L
Linus Torvalds 已提交
1598 1599 1600
	return NOTIFY_OK;
}

1601 1602 1603 1604 1605 1606 1607
static inline void timer_register_cpu_notifier(void)
{
	cpu_notifier(timer_cpu_notify, 0);
}
#else
static inline void timer_register_cpu_notifier(void) { }
#endif /* CONFIG_HOTPLUG_CPU */
L
Linus Torvalds 已提交
1608

1609
static void __init init_timer_cpu(int cpu)
1610
{
1611
	struct timer_base *base = per_cpu_ptr(&timer_bases, cpu);
1612

1613 1614 1615
	base->cpu = cpu;
	spin_lock_init(&base->lock);

1616 1617
	base->clk = jiffies;
	base->next_timer = base->clk;
1618 1619 1620
}

static void __init init_timer_cpus(void)
L
Linus Torvalds 已提交
1621
{
1622 1623
	int cpu;

1624 1625
	for_each_possible_cpu(cpu)
		init_timer_cpu(cpu);
1626
}
1627

1628 1629 1630
void __init init_timers(void)
{
	init_timer_cpus();
1631
	init_timer_stats();
1632
	timer_register_cpu_notifier();
1633
	open_softirq(TIMER_SOFTIRQ, run_timer_softirq);
L
Linus Torvalds 已提交
1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
}

/**
 * msleep - sleep safely even with waitqueue interruptions
 * @msecs: Time in milliseconds to sleep for
 */
void msleep(unsigned int msecs)
{
	unsigned long timeout = msecs_to_jiffies(msecs) + 1;

1644 1645
	while (timeout)
		timeout = schedule_timeout_uninterruptible(timeout);
L
Linus Torvalds 已提交
1646 1647 1648 1649 1650
}

EXPORT_SYMBOL(msleep);

/**
1651
 * msleep_interruptible - sleep waiting for signals
L
Linus Torvalds 已提交
1652 1653 1654 1655 1656 1657
 * @msecs: Time in milliseconds to sleep for
 */
unsigned long msleep_interruptible(unsigned int msecs)
{
	unsigned long timeout = msecs_to_jiffies(msecs) + 1;

1658 1659
	while (timeout && !signal_pending(current))
		timeout = schedule_timeout_interruptible(timeout);
L
Linus Torvalds 已提交
1660 1661 1662 1663
	return jiffies_to_msecs(timeout);
}

EXPORT_SYMBOL(msleep_interruptible);
1664

1665
static void __sched do_usleep_range(unsigned long min, unsigned long max)
1666 1667
{
	ktime_t kmin;
1668
	u64 delta;
1669 1670

	kmin = ktime_set(0, min * NSEC_PER_USEC);
1671
	delta = (u64)(max - min) * NSEC_PER_USEC;
1672
	schedule_hrtimeout_range(&kmin, delta, HRTIMER_MODE_REL);
1673 1674 1675 1676 1677 1678 1679
}

/**
 * usleep_range - Drop in replacement for udelay where wakeup is flexible
 * @min: Minimum time in usecs to sleep
 * @max: Maximum time in usecs to sleep
 */
1680
void __sched usleep_range(unsigned long min, unsigned long max)
1681 1682 1683 1684 1685
{
	__set_current_state(TASK_UNINTERRUPTIBLE);
	do_usleep_range(min, max);
}
EXPORT_SYMBOL(usleep_range);