timekeeping.c 44.2 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
#include "ntp_internal.h"
28

29
static struct timekeeper timekeeper;
30 31
static DEFINE_RAW_SPINLOCK(timekeeper_lock);
static seqcount_t timekeeper_seq;
32
static struct timekeeper shadow_timekeeper;
33

34 35 36
/* flag for if timekeeping is suspended */
int __read_mostly timekeeping_suspended;

37 38 39
/* Flag for if there is a persistent clock on this platform */
bool __read_mostly persistent_clock_exist = false;

40 41 42 43 44 45 46 47 48 49 50
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;
51
	tk->xtime_nsec = (u64)ts->tv_nsec << tk->shift;
52 53 54 55 56
}

static void tk_xtime_add(struct timekeeper *tk, const struct timespec *ts)
{
	tk->xtime_sec += ts->tv_sec;
57
	tk->xtime_nsec += (u64)ts->tv_nsec << tk->shift;
58
	tk_normalize_xtime(tk);
59
}
60

61 62 63 64 65 66 67 68 69 70 71 72 73 74
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);
75
	tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tk->tai_offset, 0));
76 77 78 79 80 81 82 83 84 85 86
}

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);
}

87 88 89 90 91 92 93 94 95 96
/**
 * 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!
 */
97
static void tk_setup_internals(struct timekeeper *tk, struct clocksource *clock)
98 99
{
	cycle_t interval;
100
	u64 tmp, ntpinterval;
101
	struct clocksource *old_clock;
102

103 104
	old_clock = tk->clock;
	tk->clock = clock;
105
	tk->cycle_last = clock->cycle_last = clock->read(clock);
106 107 108 109

	/* Do the ns -> cycle conversion first, using original mult */
	tmp = NTP_INTERVAL_LENGTH;
	tmp <<= clock->shift;
110
	ntpinterval = tmp;
111 112
	tmp += clock->mult/2;
	do_div(tmp, clock->mult);
113 114 115 116
	if (tmp == 0)
		tmp = 1;

	interval = (cycle_t) tmp;
117
	tk->cycle_interval = interval;
118 119

	/* Go back from cycles -> shifted ns */
120 121 122
	tk->xtime_interval = (u64) interval * clock->mult;
	tk->xtime_remainder = ntpinterval - tk->xtime_interval;
	tk->raw_interval =
123
		((u64) interval * clock->mult) >> clock->shift;
124

125 126 127 128
	 /* if changing clocks, convert xtime_nsec shift units */
	if (old_clock) {
		int shift_change = clock->shift - old_clock->shift;
		if (shift_change < 0)
129
			tk->xtime_nsec >>= -shift_change;
130
		else
131
			tk->xtime_nsec <<= shift_change;
132
	}
133
	tk->shift = clock->shift;
134

135 136
	tk->ntp_error = 0;
	tk->ntp_error_shift = NTP_SCALE_SHIFT - clock->shift;
137 138 139 140 141 142

	/*
	 * The timekeeper keeps its own mult values for the currently
	 * active clocksource. These value will be adjusted via NTP
	 * to counteract clock drifting.
	 */
143
	tk->mult = clock->mult;
144
}
145

146
/* Timekeeper helper functions. */
147 148 149 150 151 152 153 154 155 156 157 158 159 160

#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

161
static inline s64 timekeeping_get_ns(struct timekeeper *tk)
162 163 164
{
	cycle_t cycle_now, cycle_delta;
	struct clocksource *clock;
165
	s64 nsec;
166 167

	/* read clocksource: */
168
	clock = tk->clock;
169 170 171 172 173
	cycle_now = clock->read(clock);

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

174 175
	nsec = cycle_delta * tk->mult + tk->xtime_nsec;
	nsec >>= tk->shift;
176

177 178
	/* If arch requires, add in get_arch_timeoffset() */
	return nsec + get_arch_timeoffset();
179 180
}

181
static inline s64 timekeeping_get_ns_raw(struct timekeeper *tk)
182 183 184
{
	cycle_t cycle_now, cycle_delta;
	struct clocksource *clock;
185
	s64 nsec;
186 187

	/* read clocksource: */
188
	clock = tk->clock;
189 190 191 192 193
	cycle_now = clock->read(clock);

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

194 195 196
	/* convert delta to nanoseconds. */
	nsec = clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);

197 198
	/* If arch requires, add in get_arch_timeoffset() */
	return nsec + get_arch_timeoffset();
199 200
}

201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216
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
 */
int pvclock_gtod_register_notifier(struct notifier_block *nb)
{
	struct timekeeper *tk = &timekeeper;
	unsigned long flags;
	int ret;

217
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
218 219
	ret = raw_notifier_chain_register(&pvclock_gtod_chain, nb);
	update_pvclock_gtod(tk);
220
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
221 222 223 224 225 226 227 228 229 230 231 232 233 234

	return ret;
}
EXPORT_SYMBOL_GPL(pvclock_gtod_register_notifier);

/**
 * pvclock_gtod_unregister_notifier - unregister a pvclock
 * timedata update listener
 */
int pvclock_gtod_unregister_notifier(struct notifier_block *nb)
{
	unsigned long flags;
	int ret;

235
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
236
	ret = raw_notifier_chain_unregister(&pvclock_gtod_chain, nb);
237
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
238 239 240 241 242

	return ret;
}
EXPORT_SYMBOL_GPL(pvclock_gtod_unregister_notifier);

243
/* must hold timekeeper_lock */
244
static void timekeeping_update(struct timekeeper *tk, bool clearntp, bool mirror)
245 246
{
	if (clearntp) {
247
		tk->ntp_error = 0;
248 249
		ntp_clear();
	}
250
	update_vsyscall(tk);
251
	update_pvclock_gtod(tk);
252 253 254

	if (mirror)
		memcpy(&shadow_timekeeper, &timekeeper, sizeof(timekeeper));
255 256
}

257
/**
258
 * timekeeping_forward_now - update clock to the current time
259
 *
260 261 262
 * 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.
263
 */
264
static void timekeeping_forward_now(struct timekeeper *tk)
265 266
{
	cycle_t cycle_now, cycle_delta;
267
	struct clocksource *clock;
268
	s64 nsec;
269

270
	clock = tk->clock;
271
	cycle_now = clock->read(clock);
272
	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
273
	tk->cycle_last = clock->cycle_last = cycle_now;
274

275
	tk->xtime_nsec += cycle_delta * tk->mult;
276

277 278
	/* If arch requires, add in get_arch_timeoffset() */
	tk->xtime_nsec += (u64)get_arch_timeoffset() << tk->shift;
279

280
	tk_normalize_xtime(tk);
281

282
	nsec = clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);
283
	timespec_add_ns(&tk->raw_time, nsec);
284 285 286
}

/**
287
 * __getnstimeofday - Returns the time of day in a timespec.
288 289
 * @ts:		pointer to the timespec to be set
 *
290 291
 * Updates the time of day in the timespec.
 * Returns 0 on success, or -ve when suspended (timespec will be undefined).
292
 */
293
int __getnstimeofday(struct timespec *ts)
294
{
295
	struct timekeeper *tk = &timekeeper;
296
	unsigned long seq;
297
	s64 nsecs = 0;
298 299

	do {
300
		seq = read_seqcount_begin(&timekeeper_seq);
301

302
		ts->tv_sec = tk->xtime_sec;
303
		nsecs = timekeeping_get_ns(tk);
304

305
	} while (read_seqcount_retry(&timekeeper_seq, seq));
306

307
	ts->tv_nsec = 0;
308
	timespec_add_ns(ts, nsecs);
309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328

	/*
	 * 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));
329 330 331
}
EXPORT_SYMBOL(getnstimeofday);

332 333
ktime_t ktime_get(void)
{
334
	struct timekeeper *tk = &timekeeper;
335 336 337 338 339 340
	unsigned int seq;
	s64 secs, nsecs;

	WARN_ON(timekeeping_suspended);

	do {
341
		seq = read_seqcount_begin(&timekeeper_seq);
342 343
		secs = tk->xtime_sec + tk->wall_to_monotonic.tv_sec;
		nsecs = timekeeping_get_ns(tk) + tk->wall_to_monotonic.tv_nsec;
344

345
	} while (read_seqcount_retry(&timekeeper_seq, seq));
346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
	/*
	 * 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)
{
364
	struct timekeeper *tk = &timekeeper;
365
	struct timespec tomono;
366
	s64 nsec;
367 368 369 370 371
	unsigned int seq;

	WARN_ON(timekeeping_suspended);

	do {
372
		seq = read_seqcount_begin(&timekeeper_seq);
373
		ts->tv_sec = tk->xtime_sec;
374
		nsec = timekeeping_get_ns(tk);
375
		tomono = tk->wall_to_monotonic;
376

377
	} while (read_seqcount_retry(&timekeeper_seq, seq));
378

379 380 381
	ts->tv_sec += tomono.tv_sec;
	ts->tv_nsec = 0;
	timespec_add_ns(ts, nsec + tomono.tv_nsec);
382 383 384
}
EXPORT_SYMBOL_GPL(ktime_get_ts);

J
John Stultz 已提交
385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400

/**
 * 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 {
401
		seq = read_seqcount_begin(&timekeeper_seq);
J
John Stultz 已提交
402 403 404 405

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

406
	} while (read_seqcount_retry(&timekeeper_seq, seq));
J
John Stultz 已提交
407 408 409 410 411 412 413 414

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

}
EXPORT_SYMBOL(timekeeping_clocktai);


415 416 417 418 419 420 421 422 423 424 425 426 427 428
/**
 * 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);

429 430 431 432 433 434 435 436 437 438 439 440 441
#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)
{
442
	struct timekeeper *tk = &timekeeper;
443 444 445 446 447 448
	unsigned long seq;
	s64 nsecs_raw, nsecs_real;

	WARN_ON_ONCE(timekeeping_suspended);

	do {
449
		seq = read_seqcount_begin(&timekeeper_seq);
450

451 452
		*ts_raw = tk->raw_time;
		ts_real->tv_sec = tk->xtime_sec;
453
		ts_real->tv_nsec = 0;
454

455 456
		nsecs_raw = timekeeping_get_ns_raw(tk);
		nsecs_real = timekeeping_get_ns(tk);
457

458
	} while (read_seqcount_retry(&timekeeper_seq, seq));
459 460 461 462 463 464 465 466

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

#endif /* CONFIG_NTP_PPS */

467 468 469 470
/**
 * do_gettimeofday - Returns the time of day in a timeval
 * @tv:		pointer to the timeval to be set
 *
471
 * NOTE: Users should be converted to using getnstimeofday()
472 473 474 475 476
 */
void do_gettimeofday(struct timeval *tv)
{
	struct timespec now;

477
	getnstimeofday(&now);
478 479 480 481
	tv->tv_sec = now.tv_sec;
	tv->tv_usec = now.tv_nsec/1000;
}
EXPORT_SYMBOL(do_gettimeofday);
482

483 484 485 486 487 488
/**
 * 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
 */
489
int do_settimeofday(const struct timespec *tv)
490
{
491
	struct timekeeper *tk = &timekeeper;
492
	struct timespec ts_delta, xt;
493
	unsigned long flags;
494

495
	if (!timespec_valid_strict(tv))
496 497
		return -EINVAL;

498 499
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
500

501
	timekeeping_forward_now(tk);
502

503
	xt = tk_xtime(tk);
504 505 506
	ts_delta.tv_sec = tv->tv_sec - xt.tv_sec;
	ts_delta.tv_nsec = tv->tv_nsec - xt.tv_nsec;

507
	tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, ts_delta));
508

509
	tk_set_xtime(tk, tv);
510

511
	timekeeping_update(tk, true, true);
512

513 514
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
515 516 517 518 519 520 521 522

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

	return 0;
}
EXPORT_SYMBOL(do_settimeofday);

523 524 525 526 527 528 529 530
/**
 * 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)
{
531
	struct timekeeper *tk = &timekeeper;
532
	unsigned long flags;
533 534
	struct timespec tmp;
	int ret = 0;
535 536 537 538

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

539 540
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
541

542
	timekeeping_forward_now(tk);
543

544 545
	/* Make sure the proposed value is valid */
	tmp = timespec_add(tk_xtime(tk),  *ts);
546
	if (!timespec_valid_strict(&tmp)) {
547 548 549
		ret = -EINVAL;
		goto error;
	}
550

551 552
	tk_xtime_add(tk, ts);
	tk_set_wall_to_mono(tk, timespec_sub(tk->wall_to_monotonic, *ts));
553

554
error: /* even if we error out, we forwarded the time, so call update */
555
	timekeeping_update(tk, true, true);
556

557 558
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
559 560 561 562

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

563
	return ret;
564 565 566
}
EXPORT_SYMBOL(timekeeping_inject_offset);

567 568 569 570 571 572 573 574 575 576 577 578

/**
 * 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 {
579
		seq = read_seqcount_begin(&timekeeper_seq);
580
		ret = tk->tai_offset;
581
	} while (read_seqcount_retry(&timekeeper_seq, seq));
582 583 584 585 586 587 588 589

	return ret;
}

/**
 * __timekeeping_set_tai_offset - Lock free worker function
 *
 */
590
static void __timekeeping_set_tai_offset(struct timekeeper *tk, s32 tai_offset)
591 592
{
	tk->tai_offset = tai_offset;
593
	tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tai_offset, 0));
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;

605 606
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
607
	__timekeeping_set_tai_offset(tk, tai_offset);
608 609
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
610
	clock_was_set();
611 612
}

613 614 615 616 617
/**
 * change_clocksource - Swaps clocksources if a new one is available
 *
 * Accumulates current time interval and initializes new clocksource
 */
618
static int change_clocksource(void *data)
619
{
620
	struct timekeeper *tk = &timekeeper;
621
	struct clocksource *new, *old;
622
	unsigned long flags;
623

624
	new = (struct clocksource *) data;
625

626 627
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
628

629
	timekeeping_forward_now(tk);
630
	if (!new->enable || new->enable(new) == 0) {
631 632
		old = tk->clock;
		tk_setup_internals(tk, new);
633 634 635
		if (old->disable)
			old->disable(old);
	}
636
	timekeeping_update(tk, true, true);
637

638 639
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
640

641 642
	return 0;
}
643

644 645 646 647 648 649 650
/**
 * 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.
 */
651
int timekeeping_notify(struct clocksource *clock)
652
{
653 654 655
	struct timekeeper *tk = &timekeeper;

	if (tk->clock == clock)
656
		return 0;
657
	stop_machine(change_clocksource, clock, NULL);
658
	tick_clock_notify();
659
	return tk->clock == clock ? 0 : -1;
660
}
661

662 663 664 665 666 667 668 669 670 671 672 673 674 675
/**
 * 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);
676

677 678 679 680 681 682 683 684
/**
 * 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)
{
685
	struct timekeeper *tk = &timekeeper;
686 687 688 689
	unsigned long seq;
	s64 nsecs;

	do {
690
		seq = read_seqcount_begin(&timekeeper_seq);
691 692
		nsecs = timekeeping_get_ns_raw(tk);
		*ts = tk->raw_time;
693

694
	} while (read_seqcount_retry(&timekeeper_seq, seq));
695 696 697 698 699

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

700
/**
701
 * timekeeping_valid_for_hres - Check if timekeeping is suitable for hres
702
 */
703
int timekeeping_valid_for_hres(void)
704
{
705
	struct timekeeper *tk = &timekeeper;
706 707 708 709
	unsigned long seq;
	int ret;

	do {
710
		seq = read_seqcount_begin(&timekeeper_seq);
711

712
		ret = tk->clock->flags & CLOCK_SOURCE_VALID_FOR_HRES;
713

714
	} while (read_seqcount_retry(&timekeeper_seq, seq));
715 716 717 718

	return ret;
}

719 720 721 722 723
/**
 * timekeeping_max_deferment - Returns max time the clocksource can be deferred
 */
u64 timekeeping_max_deferment(void)
{
724
	struct timekeeper *tk = &timekeeper;
J
John Stultz 已提交
725 726
	unsigned long seq;
	u64 ret;
727

J
John Stultz 已提交
728
	do {
729
		seq = read_seqcount_begin(&timekeeper_seq);
J
John Stultz 已提交
730

731
		ret = tk->clock->max_idle_ns;
J
John Stultz 已提交
732

733
	} while (read_seqcount_retry(&timekeeper_seq, seq));
J
John Stultz 已提交
734 735

	return ret;
736 737
}

738
/**
739
 * read_persistent_clock -  Return time from the persistent clock.
740 741
 *
 * Weak dummy function for arches that do not yet support it.
742 743
 * Reads the time from the battery backed persistent clock.
 * Returns a timespec with tv_sec=0 and tv_nsec=0 if unsupported.
744 745 746
 *
 *  XXX - Do be sure to remove it once all arches implement it.
 */
747
void __attribute__((weak)) read_persistent_clock(struct timespec *ts)
748
{
749 750
	ts->tv_sec = 0;
	ts->tv_nsec = 0;
751 752
}

753 754 755 756 757 758 759 760 761 762 763 764 765 766 767
/**
 * 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;
}

768 769 770 771 772
/*
 * timekeeping_init - Initializes the clocksource and common timekeeping values
 */
void __init timekeeping_init(void)
{
773
	struct timekeeper *tk = &timekeeper;
774
	struct clocksource *clock;
775
	unsigned long flags;
776
	struct timespec now, boot, tmp;
777 778

	read_persistent_clock(&now);
779

780
	if (!timespec_valid_strict(&now)) {
781 782 783 784
		pr_warn("WARNING: Persistent clock returned invalid value!\n"
			"         Check your CMOS/BIOS settings.\n");
		now.tv_sec = 0;
		now.tv_nsec = 0;
785 786
	} else if (now.tv_sec || now.tv_nsec)
		persistent_clock_exist = true;
787

788
	read_boot_clock(&boot);
789
	if (!timespec_valid_strict(&boot)) {
790 791 792 793 794
		pr_warn("WARNING: Boot clock returned invalid value!\n"
			"         Check your CMOS/BIOS settings.\n");
		boot.tv_sec = 0;
		boot.tv_nsec = 0;
	}
795

796 797
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
798 799
	ntp_init();

800
	clock = clocksource_default_clock();
801 802
	if (clock->enable)
		clock->enable(clock);
803
	tk_setup_internals(tk, clock);
804

805 806 807
	tk_set_xtime(tk, &now);
	tk->raw_time.tv_sec = 0;
	tk->raw_time.tv_nsec = 0;
808
	if (boot.tv_sec == 0 && boot.tv_nsec == 0)
809
		boot = tk_xtime(tk);
810

811
	set_normalized_timespec(&tmp, -boot.tv_sec, -boot.tv_nsec);
812
	tk_set_wall_to_mono(tk, tmp);
813 814 815

	tmp.tv_sec = 0;
	tmp.tv_nsec = 0;
816
	tk_set_sleep_time(tk, tmp);
817

818 819
	memcpy(&shadow_timekeeper, &timekeeper, sizeof(timekeeper));

820 821
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
822 823 824
}

/* time in seconds when suspend began */
825
static struct timespec timekeeping_suspend_time;
826

827 828 829 830 831 832 833
/**
 * __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.
 */
834 835
static void __timekeeping_inject_sleeptime(struct timekeeper *tk,
							struct timespec *delta)
836
{
837
	if (!timespec_valid_strict(delta)) {
838
		printk(KERN_WARNING "__timekeeping_inject_sleeptime: Invalid "
839 840 841
					"sleep delta value!\n");
		return;
	}
842
	tk_xtime_add(tk, delta);
843 844
	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));
845 846 847 848 849 850 851 852 853 854 855 856 857 858
}

/**
 * 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)
{
859
	struct timekeeper *tk = &timekeeper;
860
	unsigned long flags;
861

862 863 864 865 866
	/*
	 * Make sure we don't set the clock twice, as timekeeping_resume()
	 * already did it
	 */
	if (has_persistent_clock())
867 868
		return;

869 870
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
J
John Stultz 已提交
871

872
	timekeeping_forward_now(tk);
873

874
	__timekeeping_inject_sleeptime(tk, delta);
875

876
	timekeeping_update(tk, true, true);
877

878 879
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
880 881 882 883 884

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

885 886 887 888 889 890 891
/**
 * 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.
 */
892
static void timekeeping_resume(void)
893
{
894
	struct timekeeper *tk = &timekeeper;
895
	struct clocksource *clock = tk->clock;
896
	unsigned long flags;
897 898 899
	struct timespec ts_new, ts_delta;
	cycle_t cycle_now, cycle_delta;
	bool suspendtime_found = false;
900

901
	read_persistent_clock(&ts_new);
902

903
	clockevents_resume();
904 905
	clocksource_resume();

906 907
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
908

909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948
	/*
	 * 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;
949
	}
950 951 952 953 954

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

	/* Re-base the last cycle value */
955
	tk->cycle_last = clock->cycle_last = cycle_now;
956
	tk->ntp_error = 0;
957
	timekeeping_suspended = 0;
958
	timekeeping_update(tk, false, true);
959 960
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
961 962 963 964 965 966

	touch_softlockup_watchdog();

	clockevents_notify(CLOCK_EVT_NOTIFY_RESUME, NULL);

	/* Resume hrtimers */
967
	hrtimers_resume();
968 969
}

970
static int timekeeping_suspend(void)
971
{
972
	struct timekeeper *tk = &timekeeper;
973
	unsigned long flags;
974 975
	struct timespec		delta, delta_delta;
	static struct timespec	old_delta;
976

977
	read_persistent_clock(&timekeeping_suspend_time);
978

979 980
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);
981
	timekeeping_forward_now(tk);
982
	timekeeping_suspended = 1;
983 984 985 986 987 988 989

	/*
	 * 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.
	 */
990
	delta = timespec_sub(tk_xtime(tk), timekeeping_suspend_time);
991 992 993 994 995 996 997 998 999 1000 1001 1002
	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);
	}
1003 1004
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
1005 1006

	clockevents_notify(CLOCK_EVT_NOTIFY_SUSPEND, NULL);
M
Magnus Damm 已提交
1007
	clocksource_suspend();
1008
	clockevents_suspend();
1009 1010 1011 1012 1013

	return 0;
}

/* sysfs resume/suspend bits for timekeeping */
1014
static struct syscore_ops timekeeping_syscore_ops = {
1015 1016 1017 1018
	.resume		= timekeeping_resume,
	.suspend	= timekeeping_suspend,
};

1019
static int __init timekeeping_init_ops(void)
1020
{
1021 1022
	register_syscore_ops(&timekeeping_syscore_ops);
	return 0;
1023 1024
}

1025
device_initcall(timekeeping_init_ops);
1026 1027 1028 1029 1030

/*
 * If the error is already larger, we look ahead even further
 * to compensate for late or lost adjustments.
 */
1031 1032
static __always_inline int timekeeping_bigadjust(struct timekeeper *tk,
						 s64 error, s64 *interval,
1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
						 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
1045
	 * here.  This is tuned so that an error of about 1 msec is adjusted
1046 1047
	 * within about 1 sec (or 2^20 nsec in 2^SHIFT_HZ ticks).
	 */
1048
	error2 = tk->ntp_error >> (NTP_SCALE_SHIFT + 22 - 2 * SHIFT_HZ);
1049 1050 1051 1052 1053 1054 1055 1056
	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.
	 */
1057 1058
	tick_error = ntp_tick_length() >> (tk->ntp_error_shift + 1);
	tick_error -= tk->xtime_interval >> 1;
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
	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.
 */
1083
static void timekeeping_adjust(struct timekeeper *tk, s64 offset)
1084
{
1085
	s64 error, interval = tk->cycle_interval;
1086 1087
	int adj;

1088
	/*
1089
	 * The point of this is to check if the error is greater than half
1090 1091 1092 1093 1094
	 * 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.
1095 1096
	 * This "saves" dividing(shifting) interval twice, but keeps the
	 * (error > interval) comparison as still measuring if error is
1097
	 * larger than half an interval.
1098
	 *
1099
	 * Note: It does not "save" on aggravation when reading the code.
1100
	 */
1101
	error = tk->ntp_error >> (tk->ntp_error_shift - 1);
1102
	if (error > interval) {
1103 1104
		/*
		 * We now divide error by 4(via shift), which checks if
1105
		 * the error is greater than twice the interval.
1106 1107 1108
		 * If it is greater, we need a bigadjust, if its smaller,
		 * we can adjust by 1.
		 */
1109
		error >>= 2;
1110 1111 1112 1113 1114
		/*
		 * 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.
		 *
1115
		 * The proper fix is to avoid rounding up by using
1116
		 * the high precision tk->xtime_nsec instead of
1117 1118 1119
		 * xtime.tv_nsec everywhere. Fixing this will take some
		 * time.
		 */
1120 1121 1122
		if (likely(error <= interval))
			adj = 1;
		else
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
			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;
		}
	}
1139

1140 1141
	if (unlikely(tk->clock->maxadj &&
		(tk->mult + adj > tk->clock->mult + tk->clock->maxadj))) {
1142 1143
		printk_once(KERN_WARNING
			"Adjusting %s more than 11%% (%ld vs %ld)\n",
1144 1145
			tk->clock->name, (long)tk->mult + adj,
			(long)tk->clock->mult + tk->clock->maxadj);
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 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195
	/*
	 * 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.
	 */
1196 1197 1198 1199
	tk->mult += adj;
	tk->xtime_interval += interval;
	tk->xtime_nsec -= offset;
	tk->ntp_error -= (interval - offset) << tk->ntp_error_shift;
1200

1201
out_adjust:
1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
	/*
	 * 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.
	 */
1216 1217 1218 1219
	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;
1220 1221
	}

1222 1223
}

1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243
/**
 * 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);
1244 1245 1246 1247
		if (unlikely(leap)) {
			struct timespec ts;

			tk->xtime_sec += leap;
1248

1249 1250 1251 1252 1253
			ts.tv_sec = leap;
			ts.tv_nsec = 0;
			tk_set_wall_to_mono(tk,
				timespec_sub(tk->wall_to_monotonic, ts));

1254 1255
			__timekeeping_set_tai_offset(tk, tk->tai_offset - leap);

1256 1257
			clock_was_set_delayed();
		}
1258 1259 1260
	}
}

1261 1262 1263 1264 1265 1266 1267 1268 1269
/**
 * 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.
 */
1270 1271
static cycle_t logarithmic_accumulation(struct timekeeper *tk, cycle_t offset,
						u32 shift)
1272
{
T
Thomas Gleixner 已提交
1273
	cycle_t interval = tk->cycle_interval << shift;
1274
	u64 raw_nsecs;
1275

1276
	/* If the offset is smaller then a shifted interval, do nothing */
T
Thomas Gleixner 已提交
1277
	if (offset < interval)
1278 1279 1280
		return offset;

	/* Accumulate one shifted interval */
T
Thomas Gleixner 已提交
1281
	offset -= interval;
1282
	tk->cycle_last += interval;
1283

1284 1285
	tk->xtime_nsec += tk->xtime_interval << shift;
	accumulate_nsecs_to_secs(tk);
1286

1287
	/* Accumulate raw time */
1288
	raw_nsecs = (u64)tk->raw_interval << shift;
1289
	raw_nsecs += tk->raw_time.tv_nsec;
1290 1291 1292
	if (raw_nsecs >= NSEC_PER_SEC) {
		u64 raw_secs = raw_nsecs;
		raw_nsecs = do_div(raw_secs, NSEC_PER_SEC);
1293
		tk->raw_time.tv_sec += raw_secs;
1294
	}
1295
	tk->raw_time.tv_nsec = raw_nsecs;
1296 1297

	/* Accumulate error between NTP and clock interval */
1298 1299 1300
	tk->ntp_error += ntp_tick_length() << shift;
	tk->ntp_error -= (tk->xtime_interval + tk->xtime_remainder) <<
						(tk->ntp_error_shift + shift);
1301 1302 1303 1304

	return offset;
}

1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
#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



1332 1333 1334 1335
/**
 * update_wall_time - Uses the current clocksource to increment the wall time
 *
 */
1336
static void update_wall_time(void)
1337
{
1338
	struct clocksource *clock;
1339 1340
	struct timekeeper *real_tk = &timekeeper;
	struct timekeeper *tk = &shadow_timekeeper;
1341
	cycle_t offset;
1342
	int shift = 0, maxshift;
J
John Stultz 已提交
1343 1344
	unsigned long flags;

1345
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
1346 1347 1348

	/* Make sure we're fully resumed: */
	if (unlikely(timekeeping_suspended))
J
John Stultz 已提交
1349
		goto out;
1350

1351
	clock = real_tk->clock;
J
John Stultz 已提交
1352 1353

#ifdef CONFIG_ARCH_USES_GETTIMEOFFSET
1354
	offset = real_tk->cycle_interval;
J
John Stultz 已提交
1355 1356
#else
	offset = (clock->read(clock) - clock->cycle_last) & clock->mask;
1357 1358
#endif

1359
	/* Check if there's really nothing to do */
1360
	if (offset < real_tk->cycle_interval)
1361 1362
		goto out;

1363 1364 1365 1366
	/*
	 * 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
1367
	 * that is smaller than the offset.  We then accumulate that
1368 1369
	 * chunk in one go, and then try to consume the next smaller
	 * doubled multiple.
1370
	 */
1371
	shift = ilog2(offset) - ilog2(tk->cycle_interval);
1372
	shift = max(0, shift);
1373
	/* Bound shift to one less than what overflows tick_length */
1374
	maxshift = (64 - (ilog2(ntp_tick_length())+1)) - 1;
1375
	shift = min(shift, maxshift);
1376 1377 1378
	while (offset >= tk->cycle_interval) {
		offset = logarithmic_accumulation(tk, offset, shift);
		if (offset < tk->cycle_interval<<shift)
1379
			shift--;
1380 1381 1382
	}

	/* correct the clock when NTP error is too big */
1383
	timekeeping_adjust(tk, offset);
1384

J
John Stultz 已提交
1385
	/*
1386 1387 1388 1389
	 * XXX This can be killed once everyone converts
	 * to the new update_vsyscall.
	 */
	old_vsyscall_fixup(tk);
1390

J
John Stultz 已提交
1391 1392
	/*
	 * Finally, make sure that after the rounding
1393
	 * xtime_nsec isn't larger than NSEC_PER_SEC
J
John Stultz 已提交
1394
	 */
1395
	accumulate_nsecs_to_secs(tk);
L
Linus Torvalds 已提交
1396

1397
	write_seqcount_begin(&timekeeper_seq);
1398 1399
	/* Update clock->cycle_last with the new value */
	clock->cycle_last = tk->cycle_last;
1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411
	/*
	 * Update the real timekeeper.
	 *
	 * We could avoid this memcpy by switching pointers, but that
	 * requires changes to all other timekeeper usage sites as
	 * well, i.e. move the timekeeper pointer getter into the
	 * spinlocked/seqcount protected sections. And we trade this
	 * memcpy under the timekeeper_seq against one before we start
	 * updating.
	 */
	memcpy(real_tk, tk, sizeof(*tk));
	timekeeping_update(real_tk, false, false);
1412
	write_seqcount_end(&timekeeper_seq);
1413
out:
1414
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
1415
}
T
Tomas Janousek 已提交
1416 1417 1418 1419 1420

/**
 * getboottime - Return the real time of system boot.
 * @ts:		pointer to the timespec to be set
 *
1421
 * Returns the wall-time of boot in a timespec.
T
Tomas Janousek 已提交
1422 1423 1424 1425 1426 1427 1428 1429
 *
 * 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)
{
1430
	struct timekeeper *tk = &timekeeper;
1431
	struct timespec boottime = {
1432 1433 1434 1435
		.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
1436
	};
1437 1438

	set_normalized_timespec(ts, -boottime.tv_sec, -boottime.tv_nsec);
T
Tomas Janousek 已提交
1439
}
1440
EXPORT_SYMBOL_GPL(getboottime);
T
Tomas Janousek 已提交
1441

1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
/**
 * 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)
{
1453
	struct timekeeper *tk = &timekeeper;
1454
	struct timespec tomono, sleep;
1455
	s64 nsec;
1456 1457 1458 1459 1460
	unsigned int seq;

	WARN_ON(timekeeping_suspended);

	do {
1461
		seq = read_seqcount_begin(&timekeeper_seq);
1462
		ts->tv_sec = tk->xtime_sec;
1463
		nsec = timekeeping_get_ns(tk);
1464 1465
		tomono = tk->wall_to_monotonic;
		sleep = tk->total_sleep_time;
1466

1467
	} while (read_seqcount_retry(&timekeeper_seq, seq));
1468

1469 1470 1471
	ts->tv_sec += tomono.tv_sec + sleep.tv_sec;
	ts->tv_nsec = 0;
	timespec_add_ns(ts, nsec + tomono.tv_nsec + sleep.tv_nsec);
1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491
}
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
1492 1493 1494 1495 1496 1497
/**
 * 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)
{
1498 1499 1500
	struct timekeeper *tk = &timekeeper;

	*ts = timespec_add(*ts, tk->total_sleep_time);
T
Tomas Janousek 已提交
1501
}
1502
EXPORT_SYMBOL_GPL(monotonic_to_bootbased);
1503

1504 1505
unsigned long get_seconds(void)
{
1506 1507 1508
	struct timekeeper *tk = &timekeeper;

	return tk->xtime_sec;
1509 1510 1511
}
EXPORT_SYMBOL(get_seconds);

1512 1513
struct timespec __current_kernel_time(void)
{
1514 1515 1516
	struct timekeeper *tk = &timekeeper;

	return tk_xtime(tk);
1517
}
1518

1519 1520
struct timespec current_kernel_time(void)
{
1521
	struct timekeeper *tk = &timekeeper;
1522 1523 1524 1525
	struct timespec now;
	unsigned long seq;

	do {
1526
		seq = read_seqcount_begin(&timekeeper_seq);
L
Linus Torvalds 已提交
1527

1528
		now = tk_xtime(tk);
1529
	} while (read_seqcount_retry(&timekeeper_seq, seq));
1530 1531 1532 1533

	return now;
}
EXPORT_SYMBOL(current_kernel_time);
1534 1535 1536

struct timespec get_monotonic_coarse(void)
{
1537
	struct timekeeper *tk = &timekeeper;
1538 1539 1540 1541
	struct timespec now, mono;
	unsigned long seq;

	do {
1542
		seq = read_seqcount_begin(&timekeeper_seq);
L
Linus Torvalds 已提交
1543

1544 1545
		now = tk_xtime(tk);
		mono = tk->wall_to_monotonic;
1546
	} while (read_seqcount_retry(&timekeeper_seq, seq));
1547 1548 1549 1550 1551

	set_normalized_timespec(&now, now.tv_sec + mono.tv_sec,
				now.tv_nsec + mono.tv_nsec);
	return now;
}
1552 1553

/*
1554
 * Must hold jiffies_lock
1555 1556 1557 1558 1559 1560 1561
 */
void do_timer(unsigned long ticks)
{
	jiffies_64 += ticks;
	update_wall_time();
	calc_global_load(ticks);
}
1562 1563

/**
1564 1565
 * get_xtime_and_monotonic_and_sleep_offset() - get xtime, wall_to_monotonic,
 *    and sleep offsets.
1566 1567
 * @xtim:	pointer to timespec to be set with xtime
 * @wtom:	pointer to timespec to be set with wall_to_monotonic
1568
 * @sleep:	pointer to timespec to be set with time in suspend
1569
 */
1570 1571
void get_xtime_and_monotonic_and_sleep_offset(struct timespec *xtim,
				struct timespec *wtom, struct timespec *sleep)
1572
{
1573
	struct timekeeper *tk = &timekeeper;
1574 1575 1576
	unsigned long seq;

	do {
1577
		seq = read_seqcount_begin(&timekeeper_seq);
1578 1579 1580
		*xtim = tk_xtime(tk);
		*wtom = tk->wall_to_monotonic;
		*sleep = tk->total_sleep_time;
1581
	} while (read_seqcount_retry(&timekeeper_seq, seq));
1582
}
T
Torben Hohn 已提交
1583

1584 1585 1586 1587 1588 1589 1590 1591 1592
#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()
 */
1593 1594
ktime_t ktime_get_update_offsets(ktime_t *offs_real, ktime_t *offs_boot,
							ktime_t *offs_tai)
1595
{
1596
	struct timekeeper *tk = &timekeeper;
1597 1598 1599 1600 1601
	ktime_t now;
	unsigned int seq;
	u64 secs, nsecs;

	do {
1602
		seq = read_seqcount_begin(&timekeeper_seq);
1603

1604 1605
		secs = tk->xtime_sec;
		nsecs = timekeeping_get_ns(tk);
1606

1607 1608
		*offs_real = tk->offs_real;
		*offs_boot = tk->offs_boot;
1609
		*offs_tai = tk->offs_tai;
1610
	} while (read_seqcount_retry(&timekeeper_seq, seq));
1611 1612 1613 1614 1615 1616 1617

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

1618 1619 1620 1621 1622
/**
 * ktime_get_monotonic_offset() - get wall_to_monotonic in ktime_t format
 */
ktime_t ktime_get_monotonic_offset(void)
{
1623
	struct timekeeper *tk = &timekeeper;
1624 1625 1626 1627
	unsigned long seq;
	struct timespec wtom;

	do {
1628
		seq = read_seqcount_begin(&timekeeper_seq);
1629
		wtom = tk->wall_to_monotonic;
1630
	} while (read_seqcount_retry(&timekeeper_seq, seq));
J
John Stultz 已提交
1631

1632 1633
	return timespec_to_ktime(wtom);
}
1634 1635
EXPORT_SYMBOL_GPL(ktime_get_monotonic_offset);

1636 1637 1638 1639 1640
/**
 * do_adjtimex() - Accessor function to NTP __do_adjtimex function
 */
int do_adjtimex(struct timex *txc)
{
1641
	struct timekeeper *tk = &timekeeper;
1642
	unsigned long flags;
1643
	struct timespec ts;
1644
	s32 orig_tai, tai;
1645 1646 1647 1648 1649 1650 1651
	int ret;

	/* Validate the data before disabling interrupts */
	ret = ntp_validate_timex(txc);
	if (ret)
		return ret;

1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662
	if (txc->modes & ADJ_SETOFFSET) {
		struct timespec delta;
		delta.tv_sec  = txc->time.tv_sec;
		delta.tv_nsec = txc->time.tv_usec;
		if (!(txc->modes & ADJ_NANO))
			delta.tv_nsec *= 1000;
		ret = timekeeping_inject_offset(&delta);
		if (ret)
			return ret;
	}

1663 1664
	getnstimeofday(&ts);

1665 1666 1667
	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);

1668
	orig_tai = tai = tk->tai_offset;
1669
	ret = __do_adjtimex(txc, &ts, &tai);
1670

1671 1672 1673 1674
	if (tai != orig_tai) {
		__timekeeping_set_tai_offset(tk, tai);
		clock_was_set_delayed();
	}
1675 1676 1677
	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);

1678 1679
	return ret;
}
1680 1681 1682 1683 1684 1685 1686

#ifdef CONFIG_NTP_PPS
/**
 * hardpps() - Accessor function to NTP __hardpps function
 */
void hardpps(const struct timespec *phase_ts, const struct timespec *raw_ts)
{
1687 1688 1689 1690 1691
	unsigned long flags;

	raw_spin_lock_irqsave(&timekeeper_lock, flags);
	write_seqcount_begin(&timekeeper_seq);

1692
	__hardpps(phase_ts, raw_ts);
1693 1694 1695

	write_seqcount_end(&timekeeper_seq);
	raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
1696 1697 1698 1699
}
EXPORT_SYMBOL(hardpps);
#endif

T
Torben Hohn 已提交
1700 1701 1702 1703 1704 1705 1706 1707
/**
 * 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)
{
1708
	write_seqlock(&jiffies_lock);
T
Torben Hohn 已提交
1709
	do_timer(ticks);
1710
	write_sequnlock(&jiffies_lock);
T
Torben Hohn 已提交
1711
}