timekeeping.c 40.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10
/*
 *  linux/kernel/time/timekeeping.c
 *
 *  Kernel timekeeping code and accessor functions
 *
 *  This code was moved from linux/kernel/timer.c.
 *  Please see that file for copyright and history logs.
 *
 */

11
#include <linux/timekeeper_internal.h>
12 13 14 15 16
#include <linux/module.h>
#include <linux/interrupt.h>
#include <linux/percpu.h>
#include <linux/init.h>
#include <linux/mm.h>
17
#include <linux/sched.h>
18
#include <linux/syscore_ops.h>
19 20 21 22
#include <linux/clocksource.h>
#include <linux/jiffies.h>
#include <linux/time.h>
#include <linux/tick.h>
23
#include <linux/stop_machine.h>
24
#include <linux/pvclock_gtod.h>
25

26
#include "tick-internal.h"
27

28
static struct timekeeper timekeeper;
29

30 31 32
/* flag for if timekeeping is suspended */
int __read_mostly timekeeping_suspended;

33 34 35
/* Flag for if there is a persistent clock on this platform */
bool __read_mostly persistent_clock_exist = false;

36 37 38 39 40 41 42 43 44 45 46
static inline void tk_normalize_xtime(struct timekeeper *tk)
{
	while (tk->xtime_nsec >= ((u64)NSEC_PER_SEC << tk->shift)) {
		tk->xtime_nsec -= (u64)NSEC_PER_SEC << tk->shift;
		tk->xtime_sec++;
	}
}

static void tk_set_xtime(struct timekeeper *tk, const struct timespec *ts)
{
	tk->xtime_sec = ts->tv_sec;
47
	tk->xtime_nsec = (u64)ts->tv_nsec << tk->shift;
48 49 50 51 52
}

static void tk_xtime_add(struct timekeeper *tk, const struct timespec *ts)
{
	tk->xtime_sec += ts->tv_sec;
53
	tk->xtime_nsec += (u64)ts->tv_nsec << tk->shift;
54
	tk_normalize_xtime(tk);
55
}
56

57 58 59 60 61 62 63 64 65 66 67 68 69 70
static void tk_set_wall_to_mono(struct timekeeper *tk, struct timespec wtm)
{
	struct timespec tmp;

	/*
	 * Verify consistency of: offset_real = -wall_to_monotonic
	 * before modifying anything
	 */
	set_normalized_timespec(&tmp, -tk->wall_to_monotonic.tv_sec,
					-tk->wall_to_monotonic.tv_nsec);
	WARN_ON_ONCE(tk->offs_real.tv64 != timespec_to_ktime(tmp).tv64);
	tk->wall_to_monotonic = wtm;
	set_normalized_timespec(&tmp, -wtm.tv_sec, -wtm.tv_nsec);
	tk->offs_real = timespec_to_ktime(tmp);
71
	tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tk->tai_offset, 0));
72 73 74 75 76 77 78 79 80 81 82
}

static void tk_set_sleep_time(struct timekeeper *tk, struct timespec t)
{
	/* Verify consistency before modifying */
	WARN_ON_ONCE(tk->offs_boot.tv64 != timespec_to_ktime(tk->total_sleep_time).tv64);

	tk->total_sleep_time	= t;
	tk->offs_boot		= timespec_to_ktime(t);
}

83 84 85 86 87 88 89 90 91 92
/**
 * timekeeper_setup_internals - Set up internals to use clocksource clock.
 *
 * @clock:		Pointer to clocksource.
 *
 * Calculates a fixed cycle/nsec interval for a given clocksource/adjustment
 * pair and interval request.
 *
 * Unless you're the timekeeping code, you should not be using this!
 */
93
static void tk_setup_internals(struct timekeeper *tk, struct clocksource *clock)
94 95
{
	cycle_t interval;
96
	u64 tmp, ntpinterval;
97
	struct clocksource *old_clock;
98

99 100
	old_clock = tk->clock;
	tk->clock = clock;
101 102 103 104 105
	clock->cycle_last = clock->read(clock);

	/* Do the ns -> cycle conversion first, using original mult */
	tmp = NTP_INTERVAL_LENGTH;
	tmp <<= clock->shift;
106
	ntpinterval = tmp;
107 108
	tmp += clock->mult/2;
	do_div(tmp, clock->mult);
109 110 111 112
	if (tmp == 0)
		tmp = 1;

	interval = (cycle_t) tmp;
113
	tk->cycle_interval = interval;
114 115

	/* Go back from cycles -> shifted ns */
116 117 118
	tk->xtime_interval = (u64) interval * clock->mult;
	tk->xtime_remainder = ntpinterval - tk->xtime_interval;
	tk->raw_interval =
119
		((u64) interval * clock->mult) >> clock->shift;
120

121 122 123 124
	 /* if changing clocks, convert xtime_nsec shift units */
	if (old_clock) {
		int shift_change = clock->shift - old_clock->shift;
		if (shift_change < 0)
125
			tk->xtime_nsec >>= -shift_change;
126
		else
127
			tk->xtime_nsec <<= shift_change;
128
	}
129
	tk->shift = clock->shift;
130

131 132
	tk->ntp_error = 0;
	tk->ntp_error_shift = NTP_SCALE_SHIFT - clock->shift;
133 134 135 136 137 138

	/*
	 * The timekeeper keeps its own mult values for the currently
	 * active clocksource. These value will be adjusted via NTP
	 * to counteract clock drifting.
	 */
139
	tk->mult = clock->mult;
140
}
141

142
/* Timekeeper helper functions. */
143 144 145 146 147 148 149 150 151 152 153 154 155 156

#ifdef CONFIG_ARCH_USES_GETTIMEOFFSET
u32 (*arch_gettimeoffset)(void);

u32 get_arch_timeoffset(void)
{
	if (likely(arch_gettimeoffset))
		return arch_gettimeoffset();
	return 0;
}
#else
static inline u32 get_arch_timeoffset(void) { return 0; }
#endif

157
static inline s64 timekeeping_get_ns(struct timekeeper *tk)
158 159 160
{
	cycle_t cycle_now, cycle_delta;
	struct clocksource *clock;
161
	s64 nsec;
162 163

	/* read clocksource: */
164
	clock = tk->clock;
165 166 167 168 169
	cycle_now = clock->read(clock);

	/* calculate the delta since the last update_wall_time: */
	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;

170 171
	nsec = cycle_delta * tk->mult + tk->xtime_nsec;
	nsec >>= tk->shift;
172

173 174
	/* If arch requires, add in get_arch_timeoffset() */
	return nsec + get_arch_timeoffset();
175 176
}

177
static inline s64 timekeeping_get_ns_raw(struct timekeeper *tk)
178 179 180
{
	cycle_t cycle_now, cycle_delta;
	struct clocksource *clock;
181
	s64 nsec;
182 183

	/* read clocksource: */
184
	clock = tk->clock;
185 186 187 188 189
	cycle_now = clock->read(clock);

	/* calculate the delta since the last update_wall_time: */
	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;

190 191 192
	/* convert delta to nanoseconds. */
	nsec = clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);

193 194
	/* If arch requires, add in get_arch_timeoffset() */
	return nsec + get_arch_timeoffset();
195 196
}

197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244
static RAW_NOTIFIER_HEAD(pvclock_gtod_chain);

static void update_pvclock_gtod(struct timekeeper *tk)
{
	raw_notifier_call_chain(&pvclock_gtod_chain, 0, tk);
}

/**
 * pvclock_gtod_register_notifier - register a pvclock timedata update listener
 *
 * Must hold write on timekeeper.lock
 */
int pvclock_gtod_register_notifier(struct notifier_block *nb)
{
	struct timekeeper *tk = &timekeeper;
	unsigned long flags;
	int ret;

	write_seqlock_irqsave(&tk->lock, flags);
	ret = raw_notifier_chain_register(&pvclock_gtod_chain, nb);
	/* update timekeeping data */
	update_pvclock_gtod(tk);
	write_sequnlock_irqrestore(&tk->lock, flags);

	return ret;
}
EXPORT_SYMBOL_GPL(pvclock_gtod_register_notifier);

/**
 * pvclock_gtod_unregister_notifier - unregister a pvclock
 * timedata update listener
 *
 * Must hold write on timekeeper.lock
 */
int pvclock_gtod_unregister_notifier(struct notifier_block *nb)
{
	struct timekeeper *tk = &timekeeper;
	unsigned long flags;
	int ret;

	write_seqlock_irqsave(&tk->lock, flags);
	ret = raw_notifier_chain_unregister(&pvclock_gtod_chain, nb);
	write_sequnlock_irqrestore(&tk->lock, flags);

	return ret;
}
EXPORT_SYMBOL_GPL(pvclock_gtod_unregister_notifier);

245
/* must hold write on timekeeper.lock */
246
static void timekeeping_update(struct timekeeper *tk, bool clearntp)
247 248
{
	if (clearntp) {
249
		tk->ntp_error = 0;
250 251
		ntp_clear();
	}
252
	update_vsyscall(tk);
253
	update_pvclock_gtod(tk);
254 255
}

256
/**
257
 * timekeeping_forward_now - update clock to the current time
258
 *
259 260 261
 * Forward the current clock to update its state since the last call to
 * update_wall_time(). This is useful before significant clock changes,
 * as it avoids having to deal with this time offset explicitly.
262
 */
263
static void timekeeping_forward_now(struct timekeeper *tk)
264 265
{
	cycle_t cycle_now, cycle_delta;
266
	struct clocksource *clock;
267
	s64 nsec;
268

269
	clock = tk->clock;
270
	cycle_now = clock->read(clock);
271
	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
272
	clock->cycle_last = cycle_now;
273

274
	tk->xtime_nsec += cycle_delta * tk->mult;
275

276 277
	/* If arch requires, add in get_arch_timeoffset() */
	tk->xtime_nsec += (u64)get_arch_timeoffset() << tk->shift;
278

279
	tk_normalize_xtime(tk);
280

281
	nsec = clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);
282
	timespec_add_ns(&tk->raw_time, nsec);
283 284 285
}

/**
286
 * __getnstimeofday - Returns the time of day in a timespec.
287 288
 * @ts:		pointer to the timespec to be set
 *
289 290
 * Updates the time of day in the timespec.
 * Returns 0 on success, or -ve when suspended (timespec will be undefined).
291
 */
292
int __getnstimeofday(struct timespec *ts)
293
{
294
	struct timekeeper *tk = &timekeeper;
295
	unsigned long seq;
296
	s64 nsecs = 0;
297 298

	do {
299
		seq = read_seqbegin(&tk->lock);
300

301
		ts->tv_sec = tk->xtime_sec;
302
		nsecs = timekeeping_get_ns(tk);
303

304
	} while (read_seqretry(&tk->lock, seq));
305

306
	ts->tv_nsec = 0;
307
	timespec_add_ns(ts, nsecs);
308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327

	/*
	 * Do not bail out early, in case there were callers still using
	 * the value, even in the face of the WARN_ON.
	 */
	if (unlikely(timekeeping_suspended))
		return -EAGAIN;
	return 0;
}
EXPORT_SYMBOL(__getnstimeofday);

/**
 * getnstimeofday - Returns the time of day in a timespec.
 * @ts:		pointer to the timespec to be set
 *
 * Returns the time of day in a timespec (WARN if suspended).
 */
void getnstimeofday(struct timespec *ts)
{
	WARN_ON(__getnstimeofday(ts));
328 329 330
}
EXPORT_SYMBOL(getnstimeofday);

331 332
ktime_t ktime_get(void)
{
333
	struct timekeeper *tk = &timekeeper;
334 335 336 337 338 339
	unsigned int seq;
	s64 secs, nsecs;

	WARN_ON(timekeeping_suspended);

	do {
340 341 342
		seq = read_seqbegin(&tk->lock);
		secs = tk->xtime_sec + tk->wall_to_monotonic.tv_sec;
		nsecs = timekeeping_get_ns(tk) + tk->wall_to_monotonic.tv_nsec;
343

344
	} while (read_seqretry(&tk->lock, seq));
345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
	/*
	 * Use ktime_set/ktime_add_ns to create a proper ktime on
	 * 32-bit architectures without CONFIG_KTIME_SCALAR.
	 */
	return ktime_add_ns(ktime_set(secs, 0), nsecs);
}
EXPORT_SYMBOL_GPL(ktime_get);

/**
 * ktime_get_ts - get the monotonic clock in timespec format
 * @ts:		pointer to timespec variable
 *
 * The function calculates the monotonic clock from the realtime
 * clock and the wall_to_monotonic offset and stores the result
 * in normalized timespec format in the variable pointed to by @ts.
 */
void ktime_get_ts(struct timespec *ts)
{
363
	struct timekeeper *tk = &timekeeper;
364
	struct timespec tomono;
365
	s64 nsec;
366 367 368 369 370
	unsigned int seq;

	WARN_ON(timekeeping_suspended);

	do {
371 372
		seq = read_seqbegin(&tk->lock);
		ts->tv_sec = tk->xtime_sec;
373
		nsec = timekeeping_get_ns(tk);
374
		tomono = tk->wall_to_monotonic;
375

376
	} while (read_seqretry(&tk->lock, seq));
377

378 379 380
	ts->tv_sec += tomono.tv_sec;
	ts->tv_nsec = 0;
	timespec_add_ns(ts, nsec + tomono.tv_nsec);
381 382 383
}
EXPORT_SYMBOL_GPL(ktime_get_ts);

J
John Stultz 已提交
384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413

/**
 * timekeeping_clocktai - Returns the TAI time of day in a timespec
 * @ts:		pointer to the timespec to be set
 *
 * Returns the time of day in a timespec.
 */
void timekeeping_clocktai(struct timespec *ts)
{
	struct timekeeper *tk = &timekeeper;
	unsigned long seq;
	u64 nsecs;

	WARN_ON(timekeeping_suspended);

	do {
		seq = read_seqbegin(&tk->lock);

		ts->tv_sec = tk->xtime_sec + tk->tai_offset;
		nsecs = timekeeping_get_ns(tk);

	} while (read_seqretry(&tk->lock, seq));

	ts->tv_nsec = 0;
	timespec_add_ns(ts, nsecs);

}
EXPORT_SYMBOL(timekeeping_clocktai);


414 415 416 417 418 419 420 421 422 423 424 425 426 427
/**
 * ktime_get_clocktai - Returns the TAI time of day in a ktime
 *
 * Returns the time of day in a ktime.
 */
ktime_t ktime_get_clocktai(void)
{
	struct timespec ts;

	timekeeping_clocktai(&ts);
	return timespec_to_ktime(ts);
}
EXPORT_SYMBOL(ktime_get_clocktai);

428 429 430 431 432 433 434 435 436 437 438 439 440
#ifdef CONFIG_NTP_PPS

/**
 * getnstime_raw_and_real - get day and raw monotonic time in timespec format
 * @ts_raw:	pointer to the timespec to be set to raw monotonic time
 * @ts_real:	pointer to the timespec to be set to the time of day
 *
 * This function reads both the time of day and raw monotonic time at the
 * same time atomically and stores the resulting timestamps in timespec
 * format.
 */
void getnstime_raw_and_real(struct timespec *ts_raw, struct timespec *ts_real)
{
441
	struct timekeeper *tk = &timekeeper;
442 443 444 445 446 447
	unsigned long seq;
	s64 nsecs_raw, nsecs_real;

	WARN_ON_ONCE(timekeeping_suspended);

	do {
448
		seq = read_seqbegin(&tk->lock);
449

450 451
		*ts_raw = tk->raw_time;
		ts_real->tv_sec = tk->xtime_sec;
452
		ts_real->tv_nsec = 0;
453

454 455
		nsecs_raw = timekeeping_get_ns_raw(tk);
		nsecs_real = timekeeping_get_ns(tk);
456

457
	} while (read_seqretry(&tk->lock, seq));
458 459 460 461 462 463 464 465

	timespec_add_ns(ts_raw, nsecs_raw);
	timespec_add_ns(ts_real, nsecs_real);
}
EXPORT_SYMBOL(getnstime_raw_and_real);

#endif /* CONFIG_NTP_PPS */

466 467 468 469
/**
 * do_gettimeofday - Returns the time of day in a timeval
 * @tv:		pointer to the timeval to be set
 *
470
 * NOTE: Users should be converted to using getnstimeofday()
471 472 473 474 475
 */
void do_gettimeofday(struct timeval *tv)
{
	struct timespec now;

476
	getnstimeofday(&now);
477 478 479 480
	tv->tv_sec = now.tv_sec;
	tv->tv_usec = now.tv_nsec/1000;
}
EXPORT_SYMBOL(do_gettimeofday);
481

482 483 484 485 486 487
/**
 * do_settimeofday - Sets the time of day
 * @tv:		pointer to the timespec variable containing the new time
 *
 * Sets the time of day to the new time and update NTP and notify hrtimers
 */
488
int do_settimeofday(const struct timespec *tv)
489
{
490
	struct timekeeper *tk = &timekeeper;
491
	struct timespec ts_delta, xt;
492
	unsigned long flags;
493

494
	if (!timespec_valid_strict(tv))
495 496
		return -EINVAL;

497
	write_seqlock_irqsave(&tk->lock, flags);
498

499
	timekeeping_forward_now(tk);
500

501
	xt = tk_xtime(tk);
502 503 504
	ts_delta.tv_sec = tv->tv_sec - xt.tv_sec;
	ts_delta.tv_nsec = tv->tv_nsec - xt.tv_nsec;

505
	tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, ts_delta));
506

507
	tk_set_xtime(tk, tv);
508

509
	timekeeping_update(tk, true);
510

511
	write_sequnlock_irqrestore(&tk->lock, flags);
512 513 514 515 516 517 518 519

	/* signal hrtimers about time change */
	clock_was_set();

	return 0;
}
EXPORT_SYMBOL(do_settimeofday);

520 521 522 523 524 525 526 527
/**
 * timekeeping_inject_offset - Adds or subtracts from the current time.
 * @tv:		pointer to the timespec variable containing the offset
 *
 * Adds or subtracts an offset value from the current time.
 */
int timekeeping_inject_offset(struct timespec *ts)
{
528
	struct timekeeper *tk = &timekeeper;
529
	unsigned long flags;
530 531
	struct timespec tmp;
	int ret = 0;
532 533 534 535

	if ((unsigned long)ts->tv_nsec >= NSEC_PER_SEC)
		return -EINVAL;

536
	write_seqlock_irqsave(&tk->lock, flags);
537

538
	timekeeping_forward_now(tk);
539

540 541
	/* Make sure the proposed value is valid */
	tmp = timespec_add(tk_xtime(tk),  *ts);
542
	if (!timespec_valid_strict(&tmp)) {
543 544 545
		ret = -EINVAL;
		goto error;
	}
546

547 548
	tk_xtime_add(tk, ts);
	tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, *ts));
549

550
error: /* even if we error out, we forwarded the time, so call update */
551
	timekeeping_update(tk, true);
552

553
	write_sequnlock_irqrestore(&tk->lock, flags);
554 555 556 557

	/* signal hrtimers about time change */
	clock_was_set();

558
	return ret;
559 560 561
}
EXPORT_SYMBOL(timekeeping_inject_offset);

562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587

/**
 * timekeeping_get_tai_offset - Returns current TAI offset from UTC
 *
 */
s32 timekeeping_get_tai_offset(void)
{
	struct timekeeper *tk = &timekeeper;
	unsigned int seq;
	s32 ret;

	do {
		seq = read_seqbegin(&tk->lock);
		ret = tk->tai_offset;
	} while (read_seqretry(&tk->lock, seq));

	return ret;
}

/**
 * __timekeeping_set_tai_offset - Lock free worker function
 *
 */
void __timekeeping_set_tai_offset(struct timekeeper *tk, s32 tai_offset)
{
	tk->tai_offset = tai_offset;
588
	tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tai_offset, 0));
589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604
}

/**
 * timekeeping_set_tai_offset - Sets the current TAI offset from UTC
 *
 */
void timekeeping_set_tai_offset(s32 tai_offset)
{
	struct timekeeper *tk = &timekeeper;
	unsigned long flags;

	write_seqlock_irqsave(&tk->lock, flags);
	__timekeeping_set_tai_offset(tk, tai_offset);
	write_sequnlock_irqrestore(&tk->lock, flags);
}

605 606 607 608 609
/**
 * change_clocksource - Swaps clocksources if a new one is available
 *
 * Accumulates current time interval and initializes new clocksource
 */
610
static int change_clocksource(void *data)
611
{
612
	struct timekeeper *tk = &timekeeper;
613
	struct clocksource *new, *old;
614
	unsigned long flags;
615

616
	new = (struct clocksource *) data;
617

618
	write_seqlock_irqsave(&tk->lock, flags);
619

620
	timekeeping_forward_now(tk);
621
	if (!new->enable || new->enable(new) == 0) {
622 623
		old = tk->clock;
		tk_setup_internals(tk, new);
624 625 626
		if (old->disable)
			old->disable(old);
	}
627
	timekeeping_update(tk, true);
628

629
	write_sequnlock_irqrestore(&tk->lock, flags);
630

631 632
	return 0;
}
633

634 635 636 637 638 639 640 641 642
/**
 * timekeeping_notify - Install a new clock source
 * @clock:		pointer to the clock source
 *
 * This function is called from clocksource.c after a new, better clock
 * source has been registered. The caller holds the clocksource_mutex.
 */
void timekeeping_notify(struct clocksource *clock)
{
643 644 645
	struct timekeeper *tk = &timekeeper;

	if (tk->clock == clock)
646
		return;
647
	stop_machine(change_clocksource, clock, NULL);
648 649
	tick_clock_notify();
}
650

651 652 653 654 655 656 657 658 659 660 661 662 663 664
/**
 * ktime_get_real - get the real (wall-) time in ktime_t format
 *
 * returns the time in ktime_t format
 */
ktime_t ktime_get_real(void)
{
	struct timespec now;

	getnstimeofday(&now);

	return timespec_to_ktime(now);
}
EXPORT_SYMBOL_GPL(ktime_get_real);
665

666 667 668 669 670 671 672 673
/**
 * getrawmonotonic - Returns the raw monotonic time in a timespec
 * @ts:		pointer to the timespec to be set
 *
 * Returns the raw monotonic time (completely un-modified by ntp)
 */
void getrawmonotonic(struct timespec *ts)
{
674
	struct timekeeper *tk = &timekeeper;
675 676 677 678
	unsigned long seq;
	s64 nsecs;

	do {
679 680 681
		seq = read_seqbegin(&tk->lock);
		nsecs = timekeeping_get_ns_raw(tk);
		*ts = tk->raw_time;
682

683
	} while (read_seqretry(&tk->lock, seq));
684 685 686 687 688

	timespec_add_ns(ts, nsecs);
}
EXPORT_SYMBOL(getrawmonotonic);

689
/**
690
 * timekeeping_valid_for_hres - Check if timekeeping is suitable for hres
691
 */
692
int timekeeping_valid_for_hres(void)
693
{
694
	struct timekeeper *tk = &timekeeper;
695 696 697 698
	unsigned long seq;
	int ret;

	do {
699
		seq = read_seqbegin(&tk->lock);
700

701
		ret = tk->clock->flags & CLOCK_SOURCE_VALID_FOR_HRES;
702

703
	} while (read_seqretry(&tk->lock, seq));
704 705 706 707

	return ret;
}

708 709 710 711 712
/**
 * timekeeping_max_deferment - Returns max time the clocksource can be deferred
 */
u64 timekeeping_max_deferment(void)
{
713
	struct timekeeper *tk = &timekeeper;
J
John Stultz 已提交
714 715
	unsigned long seq;
	u64 ret;
716

J
John Stultz 已提交
717
	do {
718
		seq = read_seqbegin(&tk->lock);
J
John Stultz 已提交
719

720
		ret = tk->clock->max_idle_ns;
J
John Stultz 已提交
721

722
	} while (read_seqretry(&tk->lock, seq));
J
John Stultz 已提交
723 724

	return ret;
725 726
}

727
/**
728
 * read_persistent_clock -  Return time from the persistent clock.
729 730
 *
 * Weak dummy function for arches that do not yet support it.
731 732
 * Reads the time from the battery backed persistent clock.
 * Returns a timespec with tv_sec=0 and tv_nsec=0 if unsupported.
733 734 735
 *
 *  XXX - Do be sure to remove it once all arches implement it.
 */
736
void __attribute__((weak)) read_persistent_clock(struct timespec *ts)
737
{
738 739
	ts->tv_sec = 0;
	ts->tv_nsec = 0;
740 741
}

742 743 744 745 746 747 748 749 750 751 752 753 754 755 756
/**
 * read_boot_clock -  Return time of the system start.
 *
 * Weak dummy function for arches that do not yet support it.
 * Function to read the exact time the system has been started.
 * Returns a timespec with tv_sec=0 and tv_nsec=0 if unsupported.
 *
 *  XXX - Do be sure to remove it once all arches implement it.
 */
void __attribute__((weak)) read_boot_clock(struct timespec *ts)
{
	ts->tv_sec = 0;
	ts->tv_nsec = 0;
}

757 758 759 760 761
/*
 * timekeeping_init - Initializes the clocksource and common timekeeping values
 */
void __init timekeeping_init(void)
{
762
	struct timekeeper *tk = &timekeeper;
763
	struct clocksource *clock;
764
	unsigned long flags;
765
	struct timespec now, boot, tmp;
766 767

	read_persistent_clock(&now);
768

769
	if (!timespec_valid_strict(&now)) {
770 771 772 773
		pr_warn("WARNING: Persistent clock returned invalid value!\n"
			"         Check your CMOS/BIOS settings.\n");
		now.tv_sec = 0;
		now.tv_nsec = 0;
774 775
	} else if (now.tv_sec || now.tv_nsec)
		persistent_clock_exist = true;
776

777
	read_boot_clock(&boot);
778
	if (!timespec_valid_strict(&boot)) {
779 780 781 782 783
		pr_warn("WARNING: Boot clock returned invalid value!\n"
			"         Check your CMOS/BIOS settings.\n");
		boot.tv_sec = 0;
		boot.tv_nsec = 0;
	}
784

785
	seqlock_init(&tk->lock);
786

R
Roman Zippel 已提交
787
	ntp_init();
788

789
	write_seqlock_irqsave(&tk->lock, flags);
790
	clock = clocksource_default_clock();
791 792
	if (clock->enable)
		clock->enable(clock);
793
	tk_setup_internals(tk, clock);
794

795 796 797
	tk_set_xtime(tk, &now);
	tk->raw_time.tv_sec = 0;
	tk->raw_time.tv_nsec = 0;
798
	if (boot.tv_sec == 0 && boot.tv_nsec == 0)
799
		boot = tk_xtime(tk);
800

801
	set_normalized_timespec(&tmp, -boot.tv_sec, -boot.tv_nsec);
802
	tk_set_wall_to_mono(tk, tmp);
803 804 805

	tmp.tv_sec = 0;
	tmp.tv_nsec = 0;
806
	tk_set_sleep_time(tk, tmp);
807

808
	write_sequnlock_irqrestore(&tk->lock, flags);
809 810 811
}

/* time in seconds when suspend began */
812
static struct timespec timekeeping_suspend_time;
813

814 815 816 817 818 819 820
/**
 * __timekeeping_inject_sleeptime - Internal function to add sleep interval
 * @delta: pointer to a timespec delta value
 *
 * Takes a timespec offset measuring a suspend interval and properly
 * adds the sleep offset to the timekeeping variables.
 */
821 822
static void __timekeeping_inject_sleeptime(struct timekeeper *tk,
							struct timespec *delta)
823
{
824
	if (!timespec_valid_strict(delta)) {
825
		printk(KERN_WARNING "__timekeeping_inject_sleeptime: Invalid "
826 827 828
					"sleep delta value!\n");
		return;
	}
829
	tk_xtime_add(tk, delta);
830 831
	tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, *delta));
	tk_set_sleep_time(tk, timespec_add(tk->total_sleep_time, *delta));
832 833 834 835 836 837 838 839 840 841 842 843 844 845
}

/**
 * timekeeping_inject_sleeptime - Adds suspend interval to timeekeeping values
 * @delta: pointer to a timespec delta value
 *
 * This hook is for architectures that cannot support read_persistent_clock
 * because their RTC/persistent clock is only accessible when irqs are enabled.
 *
 * This function should only be called by rtc_resume(), and allows
 * a suspend offset to be injected into the timekeeping values.
 */
void timekeeping_inject_sleeptime(struct timespec *delta)
{
846
	struct timekeeper *tk = &timekeeper;
847
	unsigned long flags;
848

849 850 851 852 853
	/*
	 * Make sure we don't set the clock twice, as timekeeping_resume()
	 * already did it
	 */
	if (has_persistent_clock())
854 855
		return;

856
	write_seqlock_irqsave(&tk->lock, flags);
J
John Stultz 已提交
857

858
	timekeeping_forward_now(tk);
859

860
	__timekeeping_inject_sleeptime(tk, delta);
861

862
	timekeeping_update(tk, true);
863

864
	write_sequnlock_irqrestore(&tk->lock, flags);
865 866 867 868 869

	/* signal hrtimers about time change */
	clock_was_set();
}

870 871 872 873 874 875 876
/**
 * timekeeping_resume - Resumes the generic timekeeping subsystem.
 *
 * This is for the generic clocksource timekeeping.
 * xtime/wall_to_monotonic/jiffies/etc are
 * still managed by arch specific suspend/resume code.
 */
877
static void timekeeping_resume(void)
878
{
879
	struct timekeeper *tk = &timekeeper;
880
	struct clocksource *clock = tk->clock;
881
	unsigned long flags;
882 883 884
	struct timespec ts_new, ts_delta;
	cycle_t cycle_now, cycle_delta;
	bool suspendtime_found = false;
885

886
	read_persistent_clock(&ts_new);
887

888
	clockevents_resume();
889 890
	clocksource_resume();

891
	write_seqlock_irqsave(&tk->lock, flags);
892

893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932
	/*
	 * After system resumes, we need to calculate the suspended time and
	 * compensate it for the OS time. There are 3 sources that could be
	 * used: Nonstop clocksource during suspend, persistent clock and rtc
	 * device.
	 *
	 * One specific platform may have 1 or 2 or all of them, and the
	 * preference will be:
	 *	suspend-nonstop clocksource -> persistent clock -> rtc
	 * The less preferred source will only be tried if there is no better
	 * usable source. The rtc part is handled separately in rtc core code.
	 */
	cycle_now = clock->read(clock);
	if ((clock->flags & CLOCK_SOURCE_SUSPEND_NONSTOP) &&
		cycle_now > clock->cycle_last) {
		u64 num, max = ULLONG_MAX;
		u32 mult = clock->mult;
		u32 shift = clock->shift;
		s64 nsec = 0;

		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;

		/*
		 * "cycle_delta * mutl" may cause 64 bits overflow, if the
		 * suspended time is too long. In that case we need do the
		 * 64 bits math carefully
		 */
		do_div(max, mult);
		if (cycle_delta > max) {
			num = div64_u64(cycle_delta, max);
			nsec = (((u64) max * mult) >> shift) * num;
			cycle_delta -= num * max;
		}
		nsec += ((u64) cycle_delta * mult) >> shift;

		ts_delta = ns_to_timespec(nsec);
		suspendtime_found = true;
	} else if (timespec_compare(&ts_new, &timekeeping_suspend_time) > 0) {
		ts_delta = timespec_sub(ts_new, timekeeping_suspend_time);
		suspendtime_found = true;
933
	}
934 935 936 937 938 939

	if (suspendtime_found)
		__timekeeping_inject_sleeptime(tk, &ts_delta);

	/* Re-base the last cycle value */
	clock->cycle_last = cycle_now;
940
	tk->ntp_error = 0;
941
	timekeeping_suspended = 0;
942 943
	timekeeping_update(tk, false);
	write_sequnlock_irqrestore(&tk->lock, flags);
944 945 946 947 948 949

	touch_softlockup_watchdog();

	clockevents_notify(CLOCK_EVT_NOTIFY_RESUME, NULL);

	/* Resume hrtimers */
950
	hrtimers_resume();
951 952
}

953
static int timekeeping_suspend(void)
954
{
955
	struct timekeeper *tk = &timekeeper;
956
	unsigned long flags;
957 958
	struct timespec		delta, delta_delta;
	static struct timespec	old_delta;
959

960
	read_persistent_clock(&timekeeping_suspend_time);
961

962 963
	write_seqlock_irqsave(&tk->lock, flags);
	timekeeping_forward_now(tk);
964
	timekeeping_suspended = 1;
965 966 967 968 969 970 971

	/*
	 * To avoid drift caused by repeated suspend/resumes,
	 * which each can add ~1 second drift error,
	 * try to compensate so the difference in system time
	 * and persistent_clock time stays close to constant.
	 */
972
	delta = timespec_sub(tk_xtime(tk), timekeeping_suspend_time);
973 974 975 976 977 978 979 980 981 982 983 984
	delta_delta = timespec_sub(delta, old_delta);
	if (abs(delta_delta.tv_sec)  >= 2) {
		/*
		 * if delta_delta is too large, assume time correction
		 * has occured and set old_delta to the current delta.
		 */
		old_delta = delta;
	} else {
		/* Otherwise try to adjust old_system to compensate */
		timekeeping_suspend_time =
			timespec_add(timekeeping_suspend_time, delta_delta);
	}
985
	write_sequnlock_irqrestore(&tk->lock, flags);
986 987

	clockevents_notify(CLOCK_EVT_NOTIFY_SUSPEND, NULL);
M
Magnus Damm 已提交
988
	clocksource_suspend();
989
	clockevents_suspend();
990 991 992 993 994

	return 0;
}

/* sysfs resume/suspend bits for timekeeping */
995
static struct syscore_ops timekeeping_syscore_ops = {
996 997 998 999
	.resume		= timekeeping_resume,
	.suspend	= timekeeping_suspend,
};

1000
static int __init timekeeping_init_ops(void)
1001
{
1002 1003
	register_syscore_ops(&timekeeping_syscore_ops);
	return 0;
1004 1005
}

1006
device_initcall(timekeeping_init_ops);
1007 1008 1009 1010 1011

/*
 * If the error is already larger, we look ahead even further
 * to compensate for late or lost adjustments.
 */
1012 1013
static __always_inline int timekeeping_bigadjust(struct timekeeper *tk,
						 s64 error, s64 *interval,
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025
						 s64 *offset)
{
	s64 tick_error, i;
	u32 look_ahead, adj;
	s32 error2, mult;

	/*
	 * Use the current error value to determine how much to look ahead.
	 * The larger the error the slower we adjust for it to avoid problems
	 * with losing too many ticks, otherwise we would overadjust and
	 * produce an even larger error.  The smaller the adjustment the
	 * faster we try to adjust for it, as lost ticks can do less harm
L
Li Zefan 已提交
1026
	 * here.  This is tuned so that an error of about 1 msec is adjusted
1027 1028
	 * within about 1 sec (or 2^20 nsec in 2^SHIFT_HZ ticks).
	 */
1029
	error2 = tk->ntp_error >> (NTP_SCALE_SHIFT + 22 - 2 * SHIFT_HZ);
1030 1031 1032 1033 1034 1035 1036 1037
	error2 = abs(error2);
	for (look_ahead = 0; error2 > 0; look_ahead++)
		error2 >>= 2;

	/*
	 * Now calculate the error in (1 << look_ahead) ticks, but first
	 * remove the single look ahead already included in the error.
	 */
1038 1039
	tick_error = ntp_tick_length() >> (tk->ntp_error_shift + 1);
	tick_error -= tk->xtime_interval >> 1;
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063
	error = ((error - tick_error) >> look_ahead) + tick_error;

	/* Finally calculate the adjustment shift value.  */
	i = *interval;
	mult = 1;
	if (error < 0) {
		error = -error;
		*interval = -*interval;
		*offset = -*offset;
		mult = -1;
	}
	for (adj = 0; error > i; adj++)
		error >>= 1;

	*interval <<= adj;
	*offset <<= adj;
	return mult << adj;
}

/*
 * Adjust the multiplier to reduce the error value,
 * this is optimized for the most common adjustments of -1,0,1,
 * for other values we can do a bit more work.
 */
1064
static void timekeeping_adjust(struct timekeeper *tk, s64 offset)
1065
{
1066
	s64 error, interval = tk->cycle_interval;
1067 1068
	int adj;

1069
	/*
1070
	 * The point of this is to check if the error is greater than half
1071 1072 1073 1074 1075
	 * an interval.
	 *
	 * First we shift it down from NTP_SHIFT to clocksource->shifted nsecs.
	 *
	 * Note we subtract one in the shift, so that error is really error*2.
1076 1077
	 * This "saves" dividing(shifting) interval twice, but keeps the
	 * (error > interval) comparison as still measuring if error is
1078
	 * larger than half an interval.
1079
	 *
1080
	 * Note: It does not "save" on aggravation when reading the code.
1081
	 */
1082
	error = tk->ntp_error >> (tk->ntp_error_shift - 1);
1083
	if (error > interval) {
1084 1085
		/*
		 * We now divide error by 4(via shift), which checks if
1086
		 * the error is greater than twice the interval.
1087 1088 1089
		 * If it is greater, we need a bigadjust, if its smaller,
		 * we can adjust by 1.
		 */
1090
		error >>= 2;
1091 1092 1093 1094 1095
		/*
		 * XXX - In update_wall_time, we round up to the next
		 * nanosecond, and store the amount rounded up into
		 * the error. This causes the likely below to be unlikely.
		 *
1096
		 * The proper fix is to avoid rounding up by using
1097
		 * the high precision tk->xtime_nsec instead of
1098 1099 1100
		 * xtime.tv_nsec everywhere. Fixing this will take some
		 * time.
		 */
1101 1102 1103
		if (likely(error <= interval))
			adj = 1;
		else
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119
			adj = timekeeping_bigadjust(tk, error, &interval, &offset);
	} else {
		if (error < -interval) {
			/* See comment above, this is just switched for the negative */
			error >>= 2;
			if (likely(error >= -interval)) {
				adj = -1;
				interval = -interval;
				offset = -offset;
			} else {
				adj = timekeeping_bigadjust(tk, error, &interval, &offset);
			}
		} else {
			goto out_adjust;
		}
	}
1120

1121 1122
	if (unlikely(tk->clock->maxadj &&
		(tk->mult + adj > tk->clock->mult + tk->clock->maxadj))) {
1123 1124
		printk_once(KERN_WARNING
			"Adjusting %s more than 11%% (%ld vs %ld)\n",
1125 1126
			tk->clock->name, (long)tk->mult + adj,
			(long)tk->clock->mult + tk->clock->maxadj);
1127
	}
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176
	/*
	 * So the following can be confusing.
	 *
	 * To keep things simple, lets assume adj == 1 for now.
	 *
	 * When adj != 1, remember that the interval and offset values
	 * have been appropriately scaled so the math is the same.
	 *
	 * The basic idea here is that we're increasing the multiplier
	 * by one, this causes the xtime_interval to be incremented by
	 * one cycle_interval. This is because:
	 *	xtime_interval = cycle_interval * mult
	 * So if mult is being incremented by one:
	 *	xtime_interval = cycle_interval * (mult + 1)
	 * Its the same as:
	 *	xtime_interval = (cycle_interval * mult) + cycle_interval
	 * Which can be shortened to:
	 *	xtime_interval += cycle_interval
	 *
	 * So offset stores the non-accumulated cycles. Thus the current
	 * time (in shifted nanoseconds) is:
	 *	now = (offset * adj) + xtime_nsec
	 * Now, even though we're adjusting the clock frequency, we have
	 * to keep time consistent. In other words, we can't jump back
	 * in time, and we also want to avoid jumping forward in time.
	 *
	 * So given the same offset value, we need the time to be the same
	 * both before and after the freq adjustment.
	 *	now = (offset * adj_1) + xtime_nsec_1
	 *	now = (offset * adj_2) + xtime_nsec_2
	 * So:
	 *	(offset * adj_1) + xtime_nsec_1 =
	 *		(offset * adj_2) + xtime_nsec_2
	 * And we know:
	 *	adj_2 = adj_1 + 1
	 * So:
	 *	(offset * adj_1) + xtime_nsec_1 =
	 *		(offset * (adj_1+1)) + xtime_nsec_2
	 *	(offset * adj_1) + xtime_nsec_1 =
	 *		(offset * adj_1) + offset + xtime_nsec_2
	 * Canceling the sides:
	 *	xtime_nsec_1 = offset + xtime_nsec_2
	 * Which gives us:
	 *	xtime_nsec_2 = xtime_nsec_1 - offset
	 * Which simplfies to:
	 *	xtime_nsec -= offset
	 *
	 * XXX - TODO: Doc ntp_error calculation.
	 */
1177 1178 1179 1180
	tk->mult += adj;
	tk->xtime_interval += interval;
	tk->xtime_nsec -= offset;
	tk->ntp_error -= (interval - offset) << tk->ntp_error_shift;
1181

1182
out_adjust:
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
	/*
	 * It may be possible that when we entered this function, xtime_nsec
	 * was very small.  Further, if we're slightly speeding the clocksource
	 * in the code above, its possible the required corrective factor to
	 * xtime_nsec could cause it to underflow.
	 *
	 * Now, since we already accumulated the second, cannot simply roll
	 * the accumulated second back, since the NTP subsystem has been
	 * notified via second_overflow. So instead we push xtime_nsec forward
	 * by the amount we underflowed, and add that amount into the error.
	 *
	 * We'll correct this error next time through this function, when
	 * xtime_nsec is not as small.
	 */
1197 1198 1199 1200
	if (unlikely((s64)tk->xtime_nsec < 0)) {
		s64 neg = -(s64)tk->xtime_nsec;
		tk->xtime_nsec = 0;
		tk->ntp_error += neg << tk->ntp_error_shift;
1201 1202
	}

1203 1204
}

1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
/**
 * accumulate_nsecs_to_secs - Accumulates nsecs into secs
 *
 * Helper function that accumulates a the nsecs greater then a second
 * from the xtime_nsec field to the xtime_secs field.
 * It also calls into the NTP code to handle leapsecond processing.
 *
 */
static inline void accumulate_nsecs_to_secs(struct timekeeper *tk)
{
	u64 nsecps = (u64)NSEC_PER_SEC << tk->shift;

	while (tk->xtime_nsec >= nsecps) {
		int leap;

		tk->xtime_nsec -= nsecps;
		tk->xtime_sec++;

		/* Figure out if its a leap sec and apply if needed */
		leap = second_overflow(tk->xtime_sec);
1225 1226 1227 1228
		if (unlikely(leap)) {
			struct timespec ts;

			tk->xtime_sec += leap;
1229

1230 1231 1232 1233 1234
			ts.tv_sec = leap;
			ts.tv_nsec = 0;
			tk_set_wall_to_mono(tk,
				timespec_sub(tk->wall_to_monotonic, ts));

1235 1236
			__timekeeping_set_tai_offset(tk, tk->tai_offset - leap);

1237 1238
			clock_was_set_delayed();
		}
1239 1240 1241
	}
}

1242 1243 1244 1245 1246 1247 1248 1249 1250
/**
 * logarithmic_accumulation - shifted accumulation of cycles
 *
 * This functions accumulates a shifted interval of cycles into
 * into a shifted interval nanoseconds. Allows for O(log) accumulation
 * loop.
 *
 * Returns the unconsumed cycles.
 */
1251 1252
static cycle_t logarithmic_accumulation(struct timekeeper *tk, cycle_t offset,
						u32 shift)
1253
{
T
Thomas Gleixner 已提交
1254
	cycle_t interval = tk->cycle_interval << shift;
1255
	u64 raw_nsecs;
1256

1257
	/* If the offset is smaller then a shifted interval, do nothing */
T
Thomas Gleixner 已提交
1258
	if (offset < interval)
1259 1260 1261
		return offset;

	/* Accumulate one shifted interval */
T
Thomas Gleixner 已提交
1262 1263
	offset -= interval;
	tk->clock->cycle_last += interval;
1264

1265 1266
	tk->xtime_nsec += tk->xtime_interval << shift;
	accumulate_nsecs_to_secs(tk);
1267

1268
	/* Accumulate raw time */
1269
	raw_nsecs = (u64)tk->raw_interval << shift;
1270
	raw_nsecs += tk->raw_time.tv_nsec;
1271 1272 1273
	if (raw_nsecs >= NSEC_PER_SEC) {
		u64 raw_secs = raw_nsecs;
		raw_nsecs = do_div(raw_secs, NSEC_PER_SEC);
1274
		tk->raw_time.tv_sec += raw_secs;
1275
	}
1276
	tk->raw_time.tv_nsec = raw_nsecs;
1277 1278

	/* Accumulate error between NTP and clock interval */
1279 1280 1281
	tk->ntp_error += ntp_tick_length() << shift;
	tk->ntp_error -= (tk->xtime_interval + tk->xtime_remainder) <<
						(tk->ntp_error_shift + shift);
1282 1283 1284 1285

	return offset;
}

1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312
#ifdef CONFIG_GENERIC_TIME_VSYSCALL_OLD
static inline void old_vsyscall_fixup(struct timekeeper *tk)
{
	s64 remainder;

	/*
	* Store only full nanoseconds into xtime_nsec after rounding
	* it up and add the remainder to the error difference.
	* XXX - This is necessary to avoid small 1ns inconsistnecies caused
	* by truncating the remainder in vsyscalls. However, it causes
	* additional work to be done in timekeeping_adjust(). Once
	* the vsyscall implementations are converted to use xtime_nsec
	* (shifted nanoseconds), and CONFIG_GENERIC_TIME_VSYSCALL_OLD
	* users are removed, this can be killed.
	*/
	remainder = tk->xtime_nsec & ((1ULL << tk->shift) - 1);
	tk->xtime_nsec -= remainder;
	tk->xtime_nsec += 1ULL << tk->shift;
	tk->ntp_error += remainder << tk->ntp_error_shift;

}
#else
#define old_vsyscall_fixup(tk)
#endif



1313 1314 1315 1316
/**
 * update_wall_time - Uses the current clocksource to increment the wall time
 *
 */
1317
static void update_wall_time(void)
1318
{
1319
	struct clocksource *clock;
1320
	struct timekeeper *tk = &timekeeper;
1321
	cycle_t offset;
1322
	int shift = 0, maxshift;
J
John Stultz 已提交
1323 1324
	unsigned long flags;

1325
	write_seqlock_irqsave(&tk->lock, flags);
1326 1327 1328

	/* Make sure we're fully resumed: */
	if (unlikely(timekeeping_suspended))
J
John Stultz 已提交
1329
		goto out;
1330

1331
	clock = tk->clock;
J
John Stultz 已提交
1332 1333

#ifdef CONFIG_ARCH_USES_GETTIMEOFFSET
1334
	offset = tk->cycle_interval;
J
John Stultz 已提交
1335 1336
#else
	offset = (clock->read(clock) - clock->cycle_last) & clock->mask;
1337 1338
#endif

1339 1340 1341 1342
	/* Check if there's really nothing to do */
	if (offset < tk->cycle_interval)
		goto out;

1343 1344 1345 1346
	/*
	 * With NO_HZ we may have to accumulate many cycle_intervals
	 * (think "ticks") worth of time at once. To do this efficiently,
	 * we calculate the largest doubling multiple of cycle_intervals
1347
	 * that is smaller than the offset.  We then accumulate that
1348 1349
	 * chunk in one go, and then try to consume the next smaller
	 * doubled multiple.
1350
	 */
1351
	shift = ilog2(offset) - ilog2(tk->cycle_interval);
1352
	shift = max(0, shift);
1353
	/* Bound shift to one less than what overflows tick_length */
1354
	maxshift = (64 - (ilog2(ntp_tick_length())+1)) - 1;
1355
	shift = min(shift, maxshift);
1356 1357 1358
	while (offset >= tk->cycle_interval) {
		offset = logarithmic_accumulation(tk, offset, shift);
		if (offset < tk->cycle_interval<<shift)
1359
			shift--;
1360 1361 1362
	}

	/* correct the clock when NTP error is too big */
1363
	timekeeping_adjust(tk, offset);
1364

J
John Stultz 已提交
1365
	/*
1366 1367 1368 1369
	 * XXX This can be killed once everyone converts
	 * to the new update_vsyscall.
	 */
	old_vsyscall_fixup(tk);
1370

J
John Stultz 已提交
1371 1372
	/*
	 * Finally, make sure that after the rounding
1373
	 * xtime_nsec isn't larger than NSEC_PER_SEC
J
John Stultz 已提交
1374
	 */
1375
	accumulate_nsecs_to_secs(tk);
L
Linus Torvalds 已提交
1376

1377
	timekeeping_update(tk, false);
J
John Stultz 已提交
1378 1379

out:
1380
	write_sequnlock_irqrestore(&tk->lock, flags);
J
John Stultz 已提交
1381

1382
}
T
Tomas Janousek 已提交
1383 1384 1385 1386 1387

/**
 * getboottime - Return the real time of system boot.
 * @ts:		pointer to the timespec to be set
 *
1388
 * Returns the wall-time of boot in a timespec.
T
Tomas Janousek 已提交
1389 1390 1391 1392 1393 1394 1395 1396
 *
 * This is based on the wall_to_monotonic offset and the total suspend
 * time. Calls to settimeofday will affect the value returned (which
 * basically means that however wrong your real time clock is at boot time,
 * you get the right time here).
 */
void getboottime(struct timespec *ts)
{
1397
	struct timekeeper *tk = &timekeeper;
1398
	struct timespec boottime = {
1399 1400 1401 1402
		.tv_sec = tk->wall_to_monotonic.tv_sec +
				tk->total_sleep_time.tv_sec,
		.tv_nsec = tk->wall_to_monotonic.tv_nsec +
				tk->total_sleep_time.tv_nsec
1403
	};
1404 1405

	set_normalized_timespec(ts, -boottime.tv_sec, -boottime.tv_nsec);
T
Tomas Janousek 已提交
1406
}
1407
EXPORT_SYMBOL_GPL(getboottime);
T
Tomas Janousek 已提交
1408

1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419
/**
 * get_monotonic_boottime - Returns monotonic time since boot
 * @ts:		pointer to the timespec to be set
 *
 * Returns the monotonic time since boot in a timespec.
 *
 * This is similar to CLOCK_MONTONIC/ktime_get_ts, but also
 * includes the time spent in suspend.
 */
void get_monotonic_boottime(struct timespec *ts)
{
1420
	struct timekeeper *tk = &timekeeper;
1421
	struct timespec tomono, sleep;
1422
	s64 nsec;
1423 1424 1425 1426 1427
	unsigned int seq;

	WARN_ON(timekeeping_suspended);

	do {
1428 1429
		seq = read_seqbegin(&tk->lock);
		ts->tv_sec = tk->xtime_sec;
1430
		nsec = timekeeping_get_ns(tk);
1431 1432
		tomono = tk->wall_to_monotonic;
		sleep = tk->total_sleep_time;
1433

1434
	} while (read_seqretry(&tk->lock, seq));
1435

1436 1437 1438
	ts->tv_sec += tomono.tv_sec + sleep.tv_sec;
	ts->tv_nsec = 0;
	timespec_add_ns(ts, nsec + tomono.tv_nsec + sleep.tv_nsec);
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458
}
EXPORT_SYMBOL_GPL(get_monotonic_boottime);

/**
 * ktime_get_boottime - Returns monotonic time since boot in a ktime
 *
 * Returns the monotonic time since boot in a ktime
 *
 * This is similar to CLOCK_MONTONIC/ktime_get, but also
 * includes the time spent in suspend.
 */
ktime_t ktime_get_boottime(void)
{
	struct timespec ts;

	get_monotonic_boottime(&ts);
	return timespec_to_ktime(ts);
}
EXPORT_SYMBOL_GPL(ktime_get_boottime);

T
Tomas Janousek 已提交
1459 1460 1461 1462 1463 1464
/**
 * monotonic_to_bootbased - Convert the monotonic time to boot based.
 * @ts:		pointer to the timespec to be converted
 */
void monotonic_to_bootbased(struct timespec *ts)
{
1465 1466 1467
	struct timekeeper *tk = &timekeeper;

	*ts = timespec_add(*ts, tk->total_sleep_time);
T
Tomas Janousek 已提交
1468
}
1469
EXPORT_SYMBOL_GPL(monotonic_to_bootbased);
1470

1471 1472
unsigned long get_seconds(void)
{
1473 1474 1475
	struct timekeeper *tk = &timekeeper;

	return tk->xtime_sec;
1476 1477 1478
}
EXPORT_SYMBOL(get_seconds);

1479 1480
struct timespec __current_kernel_time(void)
{
1481 1482 1483
	struct timekeeper *tk = &timekeeper;

	return tk_xtime(tk);
1484
}
1485

1486 1487
struct timespec current_kernel_time(void)
{
1488
	struct timekeeper *tk = &timekeeper;
1489 1490 1491 1492
	struct timespec now;
	unsigned long seq;

	do {
1493
		seq = read_seqbegin(&tk->lock);
L
Linus Torvalds 已提交
1494

1495 1496
		now = tk_xtime(tk);
	} while (read_seqretry(&tk->lock, seq));
1497 1498 1499 1500

	return now;
}
EXPORT_SYMBOL(current_kernel_time);
1501 1502 1503

struct timespec get_monotonic_coarse(void)
{
1504
	struct timekeeper *tk = &timekeeper;
1505 1506 1507 1508
	struct timespec now, mono;
	unsigned long seq;

	do {
1509
		seq = read_seqbegin(&tk->lock);
L
Linus Torvalds 已提交
1510

1511 1512 1513
		now = tk_xtime(tk);
		mono = tk->wall_to_monotonic;
	} while (read_seqretry(&tk->lock, seq));
1514 1515 1516 1517 1518

	set_normalized_timespec(&now, now.tv_sec + mono.tv_sec,
				now.tv_nsec + mono.tv_nsec);
	return now;
}
1519 1520

/*
1521
 * Must hold jiffies_lock
1522 1523 1524 1525 1526 1527 1528
 */
void do_timer(unsigned long ticks)
{
	jiffies_64 += ticks;
	update_wall_time();
	calc_global_load(ticks);
}
1529 1530

/**
1531 1532
 * get_xtime_and_monotonic_and_sleep_offset() - get xtime, wall_to_monotonic,
 *    and sleep offsets.
1533 1534
 * @xtim:	pointer to timespec to be set with xtime
 * @wtom:	pointer to timespec to be set with wall_to_monotonic
1535
 * @sleep:	pointer to timespec to be set with time in suspend
1536
 */
1537 1538
void get_xtime_and_monotonic_and_sleep_offset(struct timespec *xtim,
				struct timespec *wtom, struct timespec *sleep)
1539
{
1540
	struct timekeeper *tk = &timekeeper;
1541 1542 1543
	unsigned long seq;

	do {
1544 1545 1546 1547 1548
		seq = read_seqbegin(&tk->lock);
		*xtim = tk_xtime(tk);
		*wtom = tk->wall_to_monotonic;
		*sleep = tk->total_sleep_time;
	} while (read_seqretry(&tk->lock, seq));
1549
}
T
Torben Hohn 已提交
1550

1551 1552 1553 1554 1555 1556 1557 1558 1559
#ifdef CONFIG_HIGH_RES_TIMERS
/**
 * ktime_get_update_offsets - hrtimer helper
 * @offs_real:	pointer to storage for monotonic -> realtime offset
 * @offs_boot:	pointer to storage for monotonic -> boottime offset
 *
 * Returns current monotonic time and updates the offsets
 * Called from hrtimer_interupt() or retrigger_next_event()
 */
1560 1561
ktime_t ktime_get_update_offsets(ktime_t *offs_real, ktime_t *offs_boot,
							ktime_t *offs_tai)
1562
{
1563
	struct timekeeper *tk = &timekeeper;
1564 1565 1566 1567 1568
	ktime_t now;
	unsigned int seq;
	u64 secs, nsecs;

	do {
1569
		seq = read_seqbegin(&tk->lock);
1570

1571 1572
		secs = tk->xtime_sec;
		nsecs = timekeeping_get_ns(tk);
1573

1574 1575
		*offs_real = tk->offs_real;
		*offs_boot = tk->offs_boot;
1576
		*offs_tai = tk->offs_tai;
1577
	} while (read_seqretry(&tk->lock, seq));
1578 1579 1580 1581 1582 1583 1584

	now = ktime_add_ns(ktime_set(secs, 0), nsecs);
	now = ktime_sub(now, *offs_real);
	return now;
}
#endif

1585 1586 1587 1588 1589
/**
 * ktime_get_monotonic_offset() - get wall_to_monotonic in ktime_t format
 */
ktime_t ktime_get_monotonic_offset(void)
{
1590
	struct timekeeper *tk = &timekeeper;
1591 1592 1593 1594
	unsigned long seq;
	struct timespec wtom;

	do {
1595 1596 1597
		seq = read_seqbegin(&tk->lock);
		wtom = tk->wall_to_monotonic;
	} while (read_seqretry(&tk->lock, seq));
J
John Stultz 已提交
1598

1599 1600
	return timespec_to_ktime(wtom);
}
1601 1602
EXPORT_SYMBOL_GPL(ktime_get_monotonic_offset);

T
Torben Hohn 已提交
1603 1604 1605 1606 1607 1608 1609 1610
/**
 * xtime_update() - advances the timekeeping infrastructure
 * @ticks:	number of ticks, that have elapsed since the last call.
 *
 * Must be called with interrupts disabled.
 */
void xtime_update(unsigned long ticks)
{
1611
	write_seqlock(&jiffies_lock);
T
Torben Hohn 已提交
1612
	do_timer(ticks);
1613
	write_sequnlock(&jiffies_lock);
T
Torben Hohn 已提交
1614
}