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

27
static struct timekeeper timekeeper;
28

29 30 31
/* flag for if timekeeping is suspended */
int __read_mostly timekeeping_suspended;

32 33 34
/* Flag for if there is a persistent clock on this platform */
bool __read_mostly persistent_clock_exist = false;

35 36 37 38 39 40 41 42 43 44 45
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;
46
	tk->xtime_nsec = (u64)ts->tv_nsec << tk->shift;
47 48 49 50 51
}

static void tk_xtime_add(struct timekeeper *tk, const struct timespec *ts)
{
	tk->xtime_sec += ts->tv_sec;
52
	tk->xtime_nsec += (u64)ts->tv_nsec << tk->shift;
53
	tk_normalize_xtime(tk);
54
}
55

56 57 58 59 60 61 62 63 64 65 66 67 68 69
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);
70
	tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tk->tai_offset, 0));
71 72 73 74 75 76 77 78 79 80 81
}

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);
}

82 83 84 85 86 87 88 89 90 91
/**
 * 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!
 */
92
static void tk_setup_internals(struct timekeeper *tk, struct clocksource *clock)
93 94
{
	cycle_t interval;
95
	u64 tmp, ntpinterval;
96
	struct clocksource *old_clock;
97

98 99
	old_clock = tk->clock;
	tk->clock = clock;
100 101 102 103 104
	clock->cycle_last = clock->read(clock);

	/* Do the ns -> cycle conversion first, using original mult */
	tmp = NTP_INTERVAL_LENGTH;
	tmp <<= clock->shift;
105
	ntpinterval = tmp;
106 107
	tmp += clock->mult/2;
	do_div(tmp, clock->mult);
108 109 110 111
	if (tmp == 0)
		tmp = 1;

	interval = (cycle_t) tmp;
112
	tk->cycle_interval = interval;
113 114

	/* Go back from cycles -> shifted ns */
115 116 117
	tk->xtime_interval = (u64) interval * clock->mult;
	tk->xtime_remainder = ntpinterval - tk->xtime_interval;
	tk->raw_interval =
118
		((u64) interval * clock->mult) >> clock->shift;
119

120 121 122 123
	 /* if changing clocks, convert xtime_nsec shift units */
	if (old_clock) {
		int shift_change = clock->shift - old_clock->shift;
		if (shift_change < 0)
124
			tk->xtime_nsec >>= -shift_change;
125
		else
126
			tk->xtime_nsec <<= shift_change;
127
	}
128
	tk->shift = clock->shift;
129

130 131
	tk->ntp_error = 0;
	tk->ntp_error_shift = NTP_SCALE_SHIFT - clock->shift;
132 133 134 135 136 137

	/*
	 * The timekeeper keeps its own mult values for the currently
	 * active clocksource. These value will be adjusted via NTP
	 * to counteract clock drifting.
	 */
138
	tk->mult = clock->mult;
139
}
140

141
/* Timekeeper helper functions. */
142 143 144 145 146 147 148 149 150 151 152 153 154 155

#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

156
static inline s64 timekeeping_get_ns(struct timekeeper *tk)
157 158 159
{
	cycle_t cycle_now, cycle_delta;
	struct clocksource *clock;
160
	s64 nsec;
161 162

	/* read clocksource: */
163
	clock = tk->clock;
164 165 166 167 168
	cycle_now = clock->read(clock);

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

169 170
	nsec = cycle_delta * tk->mult + tk->xtime_nsec;
	nsec >>= tk->shift;
171

172 173
	/* If arch requires, add in get_arch_timeoffset() */
	return nsec + get_arch_timeoffset();
174 175
}

176
static inline s64 timekeeping_get_ns_raw(struct timekeeper *tk)
177 178 179
{
	cycle_t cycle_now, cycle_delta;
	struct clocksource *clock;
180
	s64 nsec;
181 182

	/* read clocksource: */
183
	clock = tk->clock;
184 185 186 187 188
	cycle_now = clock->read(clock);

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

189 190 191
	/* convert delta to nanoseconds. */
	nsec = clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);

192 193
	/* If arch requires, add in get_arch_timeoffset() */
	return nsec + get_arch_timeoffset();
194 195
}

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
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);

244
/* must hold write on timekeeper.lock */
245
static void timekeeping_update(struct timekeeper *tk, bool clearntp)
246 247
{
	if (clearntp) {
248
		tk->ntp_error = 0;
249 250
		ntp_clear();
	}
251
	update_vsyscall(tk);
252
	update_pvclock_gtod(tk);
253 254
}

255
/**
256
 * timekeeping_forward_now - update clock to the current time
257
 *
258 259 260
 * 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.
261
 */
262
static void timekeeping_forward_now(struct timekeeper *tk)
263 264
{
	cycle_t cycle_now, cycle_delta;
265
	struct clocksource *clock;
266
	s64 nsec;
267

268
	clock = tk->clock;
269
	cycle_now = clock->read(clock);
270
	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
271
	clock->cycle_last = cycle_now;
272

273
	tk->xtime_nsec += cycle_delta * tk->mult;
274

275 276
	/* If arch requires, add in get_arch_timeoffset() */
	tk->xtime_nsec += (u64)get_arch_timeoffset() << tk->shift;
277

278
	tk_normalize_xtime(tk);
279

280
	nsec = clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);
281
	timespec_add_ns(&tk->raw_time, nsec);
282 283 284
}

/**
285
 * __getnstimeofday - Returns the time of day in a timespec.
286 287
 * @ts:		pointer to the timespec to be set
 *
288 289
 * Updates the time of day in the timespec.
 * Returns 0 on success, or -ve when suspended (timespec will be undefined).
290
 */
291
int __getnstimeofday(struct timespec *ts)
292
{
293
	struct timekeeper *tk = &timekeeper;
294
	unsigned long seq;
295
	s64 nsecs = 0;
296 297

	do {
298
		seq = read_seqbegin(&tk->lock);
299

300
		ts->tv_sec = tk->xtime_sec;
301
		nsecs = timekeeping_get_ns(tk);
302

303
	} while (read_seqretry(&tk->lock, seq));
304

305
	ts->tv_nsec = 0;
306
	timespec_add_ns(ts, nsecs);
307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326

	/*
	 * 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));
327 328 329
}
EXPORT_SYMBOL(getnstimeofday);

330 331
ktime_t ktime_get(void)
{
332
	struct timekeeper *tk = &timekeeper;
333 334 335 336 337 338
	unsigned int seq;
	s64 secs, nsecs;

	WARN_ON(timekeeping_suspended);

	do {
339 340 341
		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;
342

343
	} while (read_seqretry(&tk->lock, seq));
344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361
	/*
	 * 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)
{
362
	struct timekeeper *tk = &timekeeper;
363
	struct timespec tomono;
364
	s64 nsec;
365 366 367 368 369
	unsigned int seq;

	WARN_ON(timekeeping_suspended);

	do {
370 371
		seq = read_seqbegin(&tk->lock);
		ts->tv_sec = tk->xtime_sec;
372
		nsec = timekeeping_get_ns(tk);
373
		tomono = tk->wall_to_monotonic;
374

375
	} while (read_seqretry(&tk->lock, seq));
376

377 378 379
	ts->tv_sec += tomono.tv_sec;
	ts->tv_nsec = 0;
	timespec_add_ns(ts, nsec + tomono.tv_nsec);
380 381 382
}
EXPORT_SYMBOL_GPL(ktime_get_ts);

J
John Stultz 已提交
383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412

/**
 * 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);


413 414 415 416 417 418 419 420 421 422 423 424 425 426
/**
 * 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);

427 428 429 430 431 432 433 434 435 436 437 438 439
#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)
{
440
	struct timekeeper *tk = &timekeeper;
441 442 443 444 445 446
	unsigned long seq;
	s64 nsecs_raw, nsecs_real;

	WARN_ON_ONCE(timekeeping_suspended);

	do {
447
		seq = read_seqbegin(&tk->lock);
448

449 450
		*ts_raw = tk->raw_time;
		ts_real->tv_sec = tk->xtime_sec;
451
		ts_real->tv_nsec = 0;
452

453 454
		nsecs_raw = timekeeping_get_ns_raw(tk);
		nsecs_real = timekeeping_get_ns(tk);
455

456
	} while (read_seqretry(&tk->lock, seq));
457 458 459 460 461 462 463 464

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

#endif /* CONFIG_NTP_PPS */

465 466 467 468
/**
 * do_gettimeofday - Returns the time of day in a timeval
 * @tv:		pointer to the timeval to be set
 *
469
 * NOTE: Users should be converted to using getnstimeofday()
470 471 472 473 474
 */
void do_gettimeofday(struct timeval *tv)
{
	struct timespec now;

475
	getnstimeofday(&now);
476 477 478 479
	tv->tv_sec = now.tv_sec;
	tv->tv_usec = now.tv_nsec/1000;
}
EXPORT_SYMBOL(do_gettimeofday);
480

481 482 483 484 485 486
/**
 * 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
 */
487
int do_settimeofday(const struct timespec *tv)
488
{
489
	struct timekeeper *tk = &timekeeper;
490
	struct timespec ts_delta, xt;
491
	unsigned long flags;
492

493
	if (!timespec_valid_strict(tv))
494 495
		return -EINVAL;

496
	write_seqlock_irqsave(&tk->lock, flags);
497

498
	timekeeping_forward_now(tk);
499

500
	xt = tk_xtime(tk);
501 502 503
	ts_delta.tv_sec = tv->tv_sec - xt.tv_sec;
	ts_delta.tv_nsec = tv->tv_nsec - xt.tv_nsec;

504
	tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, ts_delta));
505

506
	tk_set_xtime(tk, tv);
507

508
	timekeeping_update(tk, true);
509

510
	write_sequnlock_irqrestore(&tk->lock, flags);
511 512 513 514 515 516 517 518

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

	return 0;
}
EXPORT_SYMBOL(do_settimeofday);

519 520 521 522 523 524 525 526
/**
 * 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)
{
527
	struct timekeeper *tk = &timekeeper;
528
	unsigned long flags;
529 530
	struct timespec tmp;
	int ret = 0;
531 532 533 534

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

535
	write_seqlock_irqsave(&tk->lock, flags);
536

537
	timekeeping_forward_now(tk);
538

539 540
	/* Make sure the proposed value is valid */
	tmp = timespec_add(tk_xtime(tk),  *ts);
541
	if (!timespec_valid_strict(&tmp)) {
542 543 544
		ret = -EINVAL;
		goto error;
	}
545

546 547
	tk_xtime_add(tk, ts);
	tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, *ts));
548

549
error: /* even if we error out, we forwarded the time, so call update */
550
	timekeeping_update(tk, true);
551

552
	write_sequnlock_irqrestore(&tk->lock, flags);
553 554 555 556

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

557
	return ret;
558 559 560
}
EXPORT_SYMBOL(timekeeping_inject_offset);

561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586

/**
 * 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;
587
	tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tai_offset, 0));
588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603
}

/**
 * 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);
}

604 605 606 607 608
/**
 * change_clocksource - Swaps clocksources if a new one is available
 *
 * Accumulates current time interval and initializes new clocksource
 */
609
static int change_clocksource(void *data)
610
{
611
	struct timekeeper *tk = &timekeeper;
612
	struct clocksource *new, *old;
613
	unsigned long flags;
614

615
	new = (struct clocksource *) data;
616

617
	write_seqlock_irqsave(&tk->lock, flags);
618

619
	timekeeping_forward_now(tk);
620
	if (!new->enable || new->enable(new) == 0) {
621 622
		old = tk->clock;
		tk_setup_internals(tk, new);
623 624 625
		if (old->disable)
			old->disable(old);
	}
626
	timekeeping_update(tk, true);
627

628
	write_sequnlock_irqrestore(&tk->lock, flags);
629

630 631
	return 0;
}
632

633 634 635 636 637 638 639 640 641
/**
 * 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)
{
642 643 644
	struct timekeeper *tk = &timekeeper;

	if (tk->clock == clock)
645
		return;
646
	stop_machine(change_clocksource, clock, NULL);
647 648
	tick_clock_notify();
}
649

650 651 652 653 654 655 656 657 658 659 660 661 662 663
/**
 * 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);
664

665 666 667 668 669 670 671 672
/**
 * 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)
{
673
	struct timekeeper *tk = &timekeeper;
674 675 676 677
	unsigned long seq;
	s64 nsecs;

	do {
678 679 680
		seq = read_seqbegin(&tk->lock);
		nsecs = timekeeping_get_ns_raw(tk);
		*ts = tk->raw_time;
681

682
	} while (read_seqretry(&tk->lock, seq));
683 684 685 686 687

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

688
/**
689
 * timekeeping_valid_for_hres - Check if timekeeping is suitable for hres
690
 */
691
int timekeeping_valid_for_hres(void)
692
{
693
	struct timekeeper *tk = &timekeeper;
694 695 696 697
	unsigned long seq;
	int ret;

	do {
698
		seq = read_seqbegin(&tk->lock);
699

700
		ret = tk->clock->flags & CLOCK_SOURCE_VALID_FOR_HRES;
701

702
	} while (read_seqretry(&tk->lock, seq));
703 704 705 706

	return ret;
}

707 708 709 710 711
/**
 * timekeeping_max_deferment - Returns max time the clocksource can be deferred
 */
u64 timekeeping_max_deferment(void)
{
712
	struct timekeeper *tk = &timekeeper;
J
John Stultz 已提交
713 714
	unsigned long seq;
	u64 ret;
715

J
John Stultz 已提交
716
	do {
717
		seq = read_seqbegin(&tk->lock);
J
John Stultz 已提交
718

719
		ret = tk->clock->max_idle_ns;
J
John Stultz 已提交
720

721
	} while (read_seqretry(&tk->lock, seq));
J
John Stultz 已提交
722 723

	return ret;
724 725
}

726
/**
727
 * read_persistent_clock -  Return time from the persistent clock.
728 729
 *
 * Weak dummy function for arches that do not yet support it.
730 731
 * Reads the time from the battery backed persistent clock.
 * Returns a timespec with tv_sec=0 and tv_nsec=0 if unsupported.
732 733 734
 *
 *  XXX - Do be sure to remove it once all arches implement it.
 */
735
void __attribute__((weak)) read_persistent_clock(struct timespec *ts)
736
{
737 738
	ts->tv_sec = 0;
	ts->tv_nsec = 0;
739 740
}

741 742 743 744 745 746 747 748 749 750 751 752 753 754 755
/**
 * 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;
}

756 757 758 759 760
/*
 * timekeeping_init - Initializes the clocksource and common timekeeping values
 */
void __init timekeeping_init(void)
{
761
	struct timekeeper *tk = &timekeeper;
762
	struct clocksource *clock;
763
	unsigned long flags;
764
	struct timespec now, boot, tmp;
765 766

	read_persistent_clock(&now);
767

768
	if (!timespec_valid_strict(&now)) {
769 770 771 772
		pr_warn("WARNING: Persistent clock returned invalid value!\n"
			"         Check your CMOS/BIOS settings.\n");
		now.tv_sec = 0;
		now.tv_nsec = 0;
773 774
	} else if (now.tv_sec || now.tv_nsec)
		persistent_clock_exist = true;
775

776
	read_boot_clock(&boot);
777
	if (!timespec_valid_strict(&boot)) {
778 779 780 781 782
		pr_warn("WARNING: Boot clock returned invalid value!\n"
			"         Check your CMOS/BIOS settings.\n");
		boot.tv_sec = 0;
		boot.tv_nsec = 0;
	}
783

784
	seqlock_init(&tk->lock);
785

R
Roman Zippel 已提交
786
	ntp_init();
787

788
	write_seqlock_irqsave(&tk->lock, flags);
789
	clock = clocksource_default_clock();
790 791
	if (clock->enable)
		clock->enable(clock);
792
	tk_setup_internals(tk, clock);
793

794 795 796
	tk_set_xtime(tk, &now);
	tk->raw_time.tv_sec = 0;
	tk->raw_time.tv_nsec = 0;
797
	if (boot.tv_sec == 0 && boot.tv_nsec == 0)
798
		boot = tk_xtime(tk);
799

800
	set_normalized_timespec(&tmp, -boot.tv_sec, -boot.tv_nsec);
801
	tk_set_wall_to_mono(tk, tmp);
802 803 804

	tmp.tv_sec = 0;
	tmp.tv_nsec = 0;
805
	tk_set_sleep_time(tk, tmp);
806

807
	write_sequnlock_irqrestore(&tk->lock, flags);
808 809 810
}

/* time in seconds when suspend began */
811
static struct timespec timekeeping_suspend_time;
812

813 814 815 816 817 818 819
/**
 * __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.
 */
820 821
static void __timekeeping_inject_sleeptime(struct timekeeper *tk,
							struct timespec *delta)
822
{
823
	if (!timespec_valid_strict(delta)) {
824
		printk(KERN_WARNING "__timekeeping_inject_sleeptime: Invalid "
825 826 827
					"sleep delta value!\n");
		return;
	}
828
	tk_xtime_add(tk, delta);
829 830
	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));
831 832 833 834 835 836 837 838 839 840 841 842 843 844
}

/**
 * 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)
{
845
	struct timekeeper *tk = &timekeeper;
846
	unsigned long flags;
847

848 849 850 851 852
	/*
	 * Make sure we don't set the clock twice, as timekeeping_resume()
	 * already did it
	 */
	if (has_persistent_clock())
853 854
		return;

855
	write_seqlock_irqsave(&tk->lock, flags);
J
John Stultz 已提交
856

857
	timekeeping_forward_now(tk);
858

859
	__timekeeping_inject_sleeptime(tk, delta);
860

861
	timekeeping_update(tk, true);
862

863
	write_sequnlock_irqrestore(&tk->lock, flags);
864 865 866 867 868

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

869 870 871 872 873 874 875
/**
 * 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.
 */
876
static void timekeeping_resume(void)
877
{
878
	struct timekeeper *tk = &timekeeper;
879
	struct clocksource *clock = tk->clock;
880
	unsigned long flags;
881 882 883
	struct timespec ts_new, ts_delta;
	cycle_t cycle_now, cycle_delta;
	bool suspendtime_found = false;
884

885
	read_persistent_clock(&ts_new);
886

887
	clockevents_resume();
888 889
	clocksource_resume();

890
	write_seqlock_irqsave(&tk->lock, flags);
891

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
	/*
	 * 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;
932
	}
933 934 935 936 937 938

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

	/* Re-base the last cycle value */
	clock->cycle_last = cycle_now;
939
	tk->ntp_error = 0;
940
	timekeeping_suspended = 0;
941 942
	timekeeping_update(tk, false);
	write_sequnlock_irqrestore(&tk->lock, flags);
943 944 945 946 947 948

	touch_softlockup_watchdog();

	clockevents_notify(CLOCK_EVT_NOTIFY_RESUME, NULL);

	/* Resume hrtimers */
949
	hrtimers_resume();
950 951
}

952
static int timekeeping_suspend(void)
953
{
954
	struct timekeeper *tk = &timekeeper;
955
	unsigned long flags;
956 957
	struct timespec		delta, delta_delta;
	static struct timespec	old_delta;
958

959
	read_persistent_clock(&timekeeping_suspend_time);
960

961 962
	write_seqlock_irqsave(&tk->lock, flags);
	timekeeping_forward_now(tk);
963
	timekeeping_suspended = 1;
964 965 966 967 968 969 970

	/*
	 * 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.
	 */
971
	delta = timespec_sub(tk_xtime(tk), timekeeping_suspend_time);
972 973 974 975 976 977 978 979 980 981 982 983
	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);
	}
984
	write_sequnlock_irqrestore(&tk->lock, flags);
985 986

	clockevents_notify(CLOCK_EVT_NOTIFY_SUSPEND, NULL);
M
Magnus Damm 已提交
987
	clocksource_suspend();
988
	clockevents_suspend();
989 990 991 992 993

	return 0;
}

/* sysfs resume/suspend bits for timekeeping */
994
static struct syscore_ops timekeeping_syscore_ops = {
995 996 997 998
	.resume		= timekeeping_resume,
	.suspend	= timekeeping_suspend,
};

999
static int __init timekeeping_init_ops(void)
1000
{
1001 1002
	register_syscore_ops(&timekeeping_syscore_ops);
	return 0;
1003 1004
}

1005
device_initcall(timekeeping_init_ops);
1006 1007 1008 1009 1010

/*
 * If the error is already larger, we look ahead even further
 * to compensate for late or lost adjustments.
 */
1011 1012
static __always_inline int timekeeping_bigadjust(struct timekeeper *tk,
						 s64 error, s64 *interval,
1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024
						 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
1025
	 * here.  This is tuned so that an error of about 1 msec is adjusted
1026 1027
	 * within about 1 sec (or 2^20 nsec in 2^SHIFT_HZ ticks).
	 */
1028
	error2 = tk->ntp_error >> (NTP_SCALE_SHIFT + 22 - 2 * SHIFT_HZ);
1029 1030 1031 1032 1033 1034 1035 1036
	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.
	 */
1037 1038
	tick_error = ntp_tick_length() >> (tk->ntp_error_shift + 1);
	tick_error -= tk->xtime_interval >> 1;
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
	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.
 */
1063
static void timekeeping_adjust(struct timekeeper *tk, s64 offset)
1064
{
1065
	s64 error, interval = tk->cycle_interval;
1066 1067
	int adj;

1068
	/*
1069
	 * The point of this is to check if the error is greater than half
1070 1071 1072 1073 1074
	 * 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.
1075 1076
	 * This "saves" dividing(shifting) interval twice, but keeps the
	 * (error > interval) comparison as still measuring if error is
1077
	 * larger than half an interval.
1078
	 *
1079
	 * Note: It does not "save" on aggravation when reading the code.
1080
	 */
1081
	error = tk->ntp_error >> (tk->ntp_error_shift - 1);
1082
	if (error > interval) {
1083 1084
		/*
		 * We now divide error by 4(via shift), which checks if
1085
		 * the error is greater than twice the interval.
1086 1087 1088
		 * If it is greater, we need a bigadjust, if its smaller,
		 * we can adjust by 1.
		 */
1089
		error >>= 2;
1090 1091 1092 1093 1094
		/*
		 * 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.
		 *
1095
		 * The proper fix is to avoid rounding up by using
1096
		 * the high precision tk->xtime_nsec instead of
1097 1098 1099
		 * xtime.tv_nsec everywhere. Fixing this will take some
		 * time.
		 */
1100 1101 1102
		if (likely(error <= interval))
			adj = 1;
		else
1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118
			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;
		}
	}
1119

1120 1121
	if (unlikely(tk->clock->maxadj &&
		(tk->mult + adj > tk->clock->mult + tk->clock->maxadj))) {
1122 1123
		printk_once(KERN_WARNING
			"Adjusting %s more than 11%% (%ld vs %ld)\n",
1124 1125
			tk->clock->name, (long)tk->mult + adj,
			(long)tk->clock->mult + tk->clock->maxadj);
1126
	}
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
	/*
	 * 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.
	 */
1176 1177 1178 1179
	tk->mult += adj;
	tk->xtime_interval += interval;
	tk->xtime_nsec -= offset;
	tk->ntp_error -= (interval - offset) << tk->ntp_error_shift;
1180

1181
out_adjust:
1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195
	/*
	 * 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.
	 */
1196 1197 1198 1199
	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;
1200 1201
	}

1202 1203
}

1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
/**
 * 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);
1224 1225 1226 1227
		if (unlikely(leap)) {
			struct timespec ts;

			tk->xtime_sec += leap;
1228

1229 1230 1231 1232 1233
			ts.tv_sec = leap;
			ts.tv_nsec = 0;
			tk_set_wall_to_mono(tk,
				timespec_sub(tk->wall_to_monotonic, ts));

1234 1235
			__timekeeping_set_tai_offset(tk, tk->tai_offset - leap);

1236 1237
			clock_was_set_delayed();
		}
1238 1239 1240
	}
}

1241 1242 1243 1244 1245 1246 1247 1248 1249
/**
 * 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.
 */
1250 1251
static cycle_t logarithmic_accumulation(struct timekeeper *tk, cycle_t offset,
						u32 shift)
1252
{
1253
	u64 raw_nsecs;
1254

1255 1256
	/* If the offset is smaller then a shifted interval, do nothing */
	if (offset < tk->cycle_interval<<shift)
1257 1258 1259
		return offset;

	/* Accumulate one shifted interval */
1260 1261
	offset -= tk->cycle_interval << shift;
	tk->clock->cycle_last += tk->cycle_interval << shift;
1262

1263 1264
	tk->xtime_nsec += tk->xtime_interval << shift;
	accumulate_nsecs_to_secs(tk);
1265

1266
	/* Accumulate raw time */
1267
	raw_nsecs = (u64)tk->raw_interval << shift;
1268
	raw_nsecs += tk->raw_time.tv_nsec;
1269 1270 1271
	if (raw_nsecs >= NSEC_PER_SEC) {
		u64 raw_secs = raw_nsecs;
		raw_nsecs = do_div(raw_secs, NSEC_PER_SEC);
1272
		tk->raw_time.tv_sec += raw_secs;
1273
	}
1274
	tk->raw_time.tv_nsec = raw_nsecs;
1275 1276

	/* Accumulate error between NTP and clock interval */
1277 1278 1279
	tk->ntp_error += ntp_tick_length() << shift;
	tk->ntp_error -= (tk->xtime_interval + tk->xtime_remainder) <<
						(tk->ntp_error_shift + shift);
1280 1281 1282 1283

	return offset;
}

1284 1285 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
#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



1311 1312 1313 1314
/**
 * update_wall_time - Uses the current clocksource to increment the wall time
 *
 */
1315
static void update_wall_time(void)
1316
{
1317
	struct clocksource *clock;
1318
	struct timekeeper *tk = &timekeeper;
1319
	cycle_t offset;
1320
	int shift = 0, maxshift;
J
John Stultz 已提交
1321 1322
	unsigned long flags;

1323
	write_seqlock_irqsave(&tk->lock, flags);
1324 1325 1326

	/* Make sure we're fully resumed: */
	if (unlikely(timekeeping_suspended))
J
John Stultz 已提交
1327
		goto out;
1328

1329
	clock = tk->clock;
J
John Stultz 已提交
1330 1331

#ifdef CONFIG_ARCH_USES_GETTIMEOFFSET
1332
	offset = tk->cycle_interval;
J
John Stultz 已提交
1333 1334
#else
	offset = (clock->read(clock) - clock->cycle_last) & clock->mask;
1335 1336
#endif

1337 1338 1339 1340
	/* Check if there's really nothing to do */
	if (offset < tk->cycle_interval)
		goto out;

1341 1342 1343 1344
	/*
	 * 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
1345
	 * that is smaller than the offset.  We then accumulate that
1346 1347
	 * chunk in one go, and then try to consume the next smaller
	 * doubled multiple.
1348
	 */
1349
	shift = ilog2(offset) - ilog2(tk->cycle_interval);
1350
	shift = max(0, shift);
1351
	/* Bound shift to one less than what overflows tick_length */
1352
	maxshift = (64 - (ilog2(ntp_tick_length())+1)) - 1;
1353
	shift = min(shift, maxshift);
1354 1355 1356
	while (offset >= tk->cycle_interval) {
		offset = logarithmic_accumulation(tk, offset, shift);
		if (offset < tk->cycle_interval<<shift)
1357
			shift--;
1358 1359 1360
	}

	/* correct the clock when NTP error is too big */
1361
	timekeeping_adjust(tk, offset);
1362

J
John Stultz 已提交
1363
	/*
1364 1365 1366 1367
	 * XXX This can be killed once everyone converts
	 * to the new update_vsyscall.
	 */
	old_vsyscall_fixup(tk);
1368

J
John Stultz 已提交
1369 1370
	/*
	 * Finally, make sure that after the rounding
1371
	 * xtime_nsec isn't larger than NSEC_PER_SEC
J
John Stultz 已提交
1372
	 */
1373
	accumulate_nsecs_to_secs(tk);
L
Linus Torvalds 已提交
1374

1375
	timekeeping_update(tk, false);
J
John Stultz 已提交
1376 1377

out:
1378
	write_sequnlock_irqrestore(&tk->lock, flags);
J
John Stultz 已提交
1379

1380
}
T
Tomas Janousek 已提交
1381 1382 1383 1384 1385

/**
 * getboottime - Return the real time of system boot.
 * @ts:		pointer to the timespec to be set
 *
1386
 * Returns the wall-time of boot in a timespec.
T
Tomas Janousek 已提交
1387 1388 1389 1390 1391 1392 1393 1394
 *
 * 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)
{
1395
	struct timekeeper *tk = &timekeeper;
1396
	struct timespec boottime = {
1397 1398 1399 1400
		.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
1401
	};
1402 1403

	set_normalized_timespec(ts, -boottime.tv_sec, -boottime.tv_nsec);
T
Tomas Janousek 已提交
1404
}
1405
EXPORT_SYMBOL_GPL(getboottime);
T
Tomas Janousek 已提交
1406

1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
/**
 * 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)
{
1418
	struct timekeeper *tk = &timekeeper;
1419
	struct timespec tomono, sleep;
1420
	s64 nsec;
1421 1422 1423 1424 1425
	unsigned int seq;

	WARN_ON(timekeeping_suspended);

	do {
1426 1427
		seq = read_seqbegin(&tk->lock);
		ts->tv_sec = tk->xtime_sec;
1428
		nsec = timekeeping_get_ns(tk);
1429 1430
		tomono = tk->wall_to_monotonic;
		sleep = tk->total_sleep_time;
1431

1432
	} while (read_seqretry(&tk->lock, seq));
1433

1434 1435 1436
	ts->tv_sec += tomono.tv_sec + sleep.tv_sec;
	ts->tv_nsec = 0;
	timespec_add_ns(ts, nsec + tomono.tv_nsec + sleep.tv_nsec);
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
}
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
1457 1458 1459 1460 1461 1462
/**
 * 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)
{
1463 1464 1465
	struct timekeeper *tk = &timekeeper;

	*ts = timespec_add(*ts, tk->total_sleep_time);
T
Tomas Janousek 已提交
1466
}
1467
EXPORT_SYMBOL_GPL(monotonic_to_bootbased);
1468

1469 1470
unsigned long get_seconds(void)
{
1471 1472 1473
	struct timekeeper *tk = &timekeeper;

	return tk->xtime_sec;
1474 1475 1476
}
EXPORT_SYMBOL(get_seconds);

1477 1478
struct timespec __current_kernel_time(void)
{
1479 1480 1481
	struct timekeeper *tk = &timekeeper;

	return tk_xtime(tk);
1482
}
1483

1484 1485
struct timespec current_kernel_time(void)
{
1486
	struct timekeeper *tk = &timekeeper;
1487 1488 1489 1490
	struct timespec now;
	unsigned long seq;

	do {
1491
		seq = read_seqbegin(&tk->lock);
L
Linus Torvalds 已提交
1492

1493 1494
		now = tk_xtime(tk);
	} while (read_seqretry(&tk->lock, seq));
1495 1496 1497 1498

	return now;
}
EXPORT_SYMBOL(current_kernel_time);
1499 1500 1501

struct timespec get_monotonic_coarse(void)
{
1502
	struct timekeeper *tk = &timekeeper;
1503 1504 1505 1506
	struct timespec now, mono;
	unsigned long seq;

	do {
1507
		seq = read_seqbegin(&tk->lock);
L
Linus Torvalds 已提交
1508

1509 1510 1511
		now = tk_xtime(tk);
		mono = tk->wall_to_monotonic;
	} while (read_seqretry(&tk->lock, seq));
1512 1513 1514 1515 1516

	set_normalized_timespec(&now, now.tv_sec + mono.tv_sec,
				now.tv_nsec + mono.tv_nsec);
	return now;
}
1517 1518

/*
1519
 * Must hold jiffies_lock
1520 1521 1522 1523 1524 1525 1526
 */
void do_timer(unsigned long ticks)
{
	jiffies_64 += ticks;
	update_wall_time();
	calc_global_load(ticks);
}
1527 1528

/**
1529 1530
 * get_xtime_and_monotonic_and_sleep_offset() - get xtime, wall_to_monotonic,
 *    and sleep offsets.
1531 1532
 * @xtim:	pointer to timespec to be set with xtime
 * @wtom:	pointer to timespec to be set with wall_to_monotonic
1533
 * @sleep:	pointer to timespec to be set with time in suspend
1534
 */
1535 1536
void get_xtime_and_monotonic_and_sleep_offset(struct timespec *xtim,
				struct timespec *wtom, struct timespec *sleep)
1537
{
1538
	struct timekeeper *tk = &timekeeper;
1539 1540 1541
	unsigned long seq;

	do {
1542 1543 1544 1545 1546
		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));
1547
}
T
Torben Hohn 已提交
1548

1549 1550 1551 1552 1553 1554 1555 1556 1557
#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()
 */
1558 1559
ktime_t ktime_get_update_offsets(ktime_t *offs_real, ktime_t *offs_boot,
							ktime_t *offs_tai)
1560
{
1561
	struct timekeeper *tk = &timekeeper;
1562 1563 1564 1565 1566
	ktime_t now;
	unsigned int seq;
	u64 secs, nsecs;

	do {
1567
		seq = read_seqbegin(&tk->lock);
1568

1569 1570
		secs = tk->xtime_sec;
		nsecs = timekeeping_get_ns(tk);
1571

1572 1573
		*offs_real = tk->offs_real;
		*offs_boot = tk->offs_boot;
1574
		*offs_tai = tk->offs_tai;
1575
	} while (read_seqretry(&tk->lock, seq));
1576 1577 1578 1579 1580 1581 1582

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

1583 1584 1585 1586 1587
/**
 * ktime_get_monotonic_offset() - get wall_to_monotonic in ktime_t format
 */
ktime_t ktime_get_monotonic_offset(void)
{
1588
	struct timekeeper *tk = &timekeeper;
1589 1590 1591 1592
	unsigned long seq;
	struct timespec wtom;

	do {
1593 1594 1595
		seq = read_seqbegin(&tk->lock);
		wtom = tk->wall_to_monotonic;
	} while (read_seqretry(&tk->lock, seq));
J
John Stultz 已提交
1596

1597 1598
	return timespec_to_ktime(wtom);
}
1599 1600
EXPORT_SYMBOL_GPL(ktime_get_monotonic_offset);

T
Torben Hohn 已提交
1601 1602 1603 1604 1605 1606 1607 1608
/**
 * 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)
{
1609
	write_seqlock(&jiffies_lock);
T
Torben Hohn 已提交
1610
	do_timer(ticks);
1611
	write_sequnlock(&jiffies_lock);
T
Torben Hohn 已提交
1612
}