clocksource.c 26.2 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
/*
 * linux/kernel/time/clocksource.c
 *
 * This file contains the functions which manage clocksource drivers.
 *
 * Copyright (C) 2004, 2005 IBM, John Stultz (johnstul@us.ibm.com)
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *
 * TODO WishList:
 *   o Allow clocksource drivers to be unregistered
 */

#include <linux/clocksource.h>
#include <linux/sysdev.h>
#include <linux/init.h>
#include <linux/module.h>
30
#include <linux/sched.h> /* for spin_unlock_irq() using preempt_count() m68k */
31
#include <linux/tick.h>
32
#include <linux/kthread.h>
33

34 35 36 37 38 39 40 41
void timecounter_init(struct timecounter *tc,
		      const struct cyclecounter *cc,
		      u64 start_tstamp)
{
	tc->cc = cc;
	tc->cycle_last = cc->read(cc);
	tc->nsec = start_tstamp;
}
42
EXPORT_SYMBOL_GPL(timecounter_init);
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

/**
 * timecounter_read_delta - get nanoseconds since last call of this function
 * @tc:         Pointer to time counter
 *
 * When the underlying cycle counter runs over, this will be handled
 * correctly as long as it does not run over more than once between
 * calls.
 *
 * The first call to this function for a new time counter initializes
 * the time tracking and returns an undefined result.
 */
static u64 timecounter_read_delta(struct timecounter *tc)
{
	cycle_t cycle_now, cycle_delta;
	u64 ns_offset;

	/* read cycle counter: */
	cycle_now = tc->cc->read(tc->cc);

	/* calculate the delta since the last timecounter_read_delta(): */
	cycle_delta = (cycle_now - tc->cycle_last) & tc->cc->mask;

	/* convert to nanoseconds: */
	ns_offset = cyclecounter_cyc2ns(tc->cc, cycle_delta);

	/* update time stamp of timecounter_read_delta() call: */
	tc->cycle_last = cycle_now;

	return ns_offset;
}

u64 timecounter_read(struct timecounter *tc)
{
	u64 nsec;

	/* increment time by nanoseconds since last call */
	nsec = timecounter_read_delta(tc);
	nsec += tc->nsec;
	tc->nsec = nsec;

	return nsec;
}
86
EXPORT_SYMBOL_GPL(timecounter_read);
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107

u64 timecounter_cyc2time(struct timecounter *tc,
			 cycle_t cycle_tstamp)
{
	u64 cycle_delta = (cycle_tstamp - tc->cycle_last) & tc->cc->mask;
	u64 nsec;

	/*
	 * Instead of always treating cycle_tstamp as more recent
	 * than tc->cycle_last, detect when it is too far in the
	 * future and treat it as old time stamp instead.
	 */
	if (cycle_delta > tc->cc->mask / 2) {
		cycle_delta = (tc->cycle_last - cycle_tstamp) & tc->cc->mask;
		nsec = tc->nsec - cyclecounter_cyc2ns(tc->cc, cycle_delta);
	} else {
		nsec = cyclecounter_cyc2ns(tc->cc, cycle_delta) + tc->nsec;
	}

	return nsec;
}
108
EXPORT_SYMBOL_GPL(timecounter_cyc2time);
109

110 111 112 113 114 115
/**
 * clocks_calc_mult_shift - calculate mult/shift factors for scaled math of clocks
 * @mult:	pointer to mult variable
 * @shift:	pointer to shift variable
 * @from:	frequency to convert from
 * @to:		frequency to convert to
116
 * @maxsec:	guaranteed runtime conversion range in seconds
117 118 119 120 121 122 123 124
 *
 * The function evaluates the shift/mult pair for the scaled math
 * operations of clocksources and clockevents.
 *
 * @to and @from are frequency values in HZ. For clock sources @to is
 * NSEC_PER_SEC == 1GHz and @from is the counter frequency. For clock
 * event @to is the counter frequency and @from is NSEC_PER_SEC.
 *
125
 * The @maxsec conversion range argument controls the time frame in
126 127 128 129 130 131 132 133
 * seconds which must be covered by the runtime conversion with the
 * calculated mult and shift factors. This guarantees that no 64bit
 * overflow happens when the input value of the conversion is
 * multiplied with the calculated mult factor. Larger ranges may
 * reduce the conversion accuracy by chosing smaller mult and shift
 * factors.
 */
void
134
clocks_calc_mult_shift(u32 *mult, u32 *shift, u32 from, u32 to, u32 maxsec)
135 136 137 138 139 140 141 142
{
	u64 tmp;
	u32 sft, sftacc= 32;

	/*
	 * Calculate the shift factor which is limiting the conversion
	 * range:
	 */
143
	tmp = ((u64)maxsec * from) >> 32;
144 145 146 147 148 149 150 151 152 153 154
	while (tmp) {
		tmp >>=1;
		sftacc--;
	}

	/*
	 * Find the conversion shift/mult pair which has the best
	 * accuracy and fits the maxsec conversion range:
	 */
	for (sft = 32; sft > 0; sft--) {
		tmp = (u64) to << sft;
155
		tmp += from / 2;
156 157 158 159 160 161 162 163
		do_div(tmp, from);
		if ((tmp >> sftacc) == 0)
			break;
	}
	*mult = tmp;
	*shift = sft;
}

164 165
/*[Clocksource internal variables]---------
 * curr_clocksource:
166
 *	currently selected clocksource.
167 168
 * clocksource_list:
 *	linked list with the registered clocksources
169 170
 * clocksource_mutex:
 *	protects manipulations to curr_clocksource and the clocksource_list
171 172 173
 * override_name:
 *	Name of the user-specified clocksource.
 */
174
static struct clocksource *curr_clocksource;
175
static LIST_HEAD(clocksource_list);
176
static DEFINE_MUTEX(clocksource_mutex);
177
static char override_name[32];
178
static int finished_booting;
179

180
#ifdef CONFIG_CLOCKSOURCE_WATCHDOG
181 182
static void clocksource_watchdog_work(struct work_struct *work);

183 184 185
static LIST_HEAD(watchdog_list);
static struct clocksource *watchdog;
static struct timer_list watchdog_timer;
186
static DECLARE_WORK(watchdog_work, clocksource_watchdog_work);
187
static DEFINE_SPINLOCK(watchdog_lock);
188
static int watchdog_running;
189
static atomic_t watchdog_reset_pending;
T
Thomas Gleixner 已提交
190

191
static int clocksource_watchdog_kthread(void *data);
192
static void __clocksource_change_rating(struct clocksource *cs, int rating);
193

194
/*
195
 * Interval: 0.5sec Threshold: 0.0625s
196 197
 */
#define WATCHDOG_INTERVAL (HZ >> 1)
198
#define WATCHDOG_THRESHOLD (NSEC_PER_SEC >> 4)
199

200 201 202 203 204 205 206 207 208
static void clocksource_watchdog_work(struct work_struct *work)
{
	/*
	 * If kthread_run fails the next watchdog scan over the
	 * watchdog_list will find the unstable clock again.
	 */
	kthread_run(clocksource_watchdog_kthread, NULL, "kwatchdog");
}

209
static void __clocksource_unstable(struct clocksource *cs)
210 211
{
	cs->flags &= ~(CLOCK_SOURCE_VALID_FOR_HRES | CLOCK_SOURCE_WATCHDOG);
212
	cs->flags |= CLOCK_SOURCE_UNSTABLE;
213 214
	if (finished_booting)
		schedule_work(&watchdog_work);
215 216
}

217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245
static void clocksource_unstable(struct clocksource *cs, int64_t delta)
{
	printk(KERN_WARNING "Clocksource %s unstable (delta = %Ld ns)\n",
	       cs->name, delta);
	__clocksource_unstable(cs);
}

/**
 * clocksource_mark_unstable - mark clocksource unstable via watchdog
 * @cs:		clocksource to be marked unstable
 *
 * This function is called instead of clocksource_change_rating from
 * cpu hotplug code to avoid a deadlock between the clocksource mutex
 * and the cpu hotplug mutex. It defers the update of the clocksource
 * to the watchdog thread.
 */
void clocksource_mark_unstable(struct clocksource *cs)
{
	unsigned long flags;

	spin_lock_irqsave(&watchdog_lock, flags);
	if (!(cs->flags & CLOCK_SOURCE_UNSTABLE)) {
		if (list_empty(&cs->wd_list))
			list_add(&cs->wd_list, &watchdog_list);
		__clocksource_unstable(cs);
	}
	spin_unlock_irqrestore(&watchdog_lock, flags);
}

246 247
static void clocksource_watchdog(unsigned long data)
{
248
	struct clocksource *cs;
249 250
	cycle_t csnow, wdnow;
	int64_t wd_nsec, cs_nsec;
251
	int next_cpu, reset_pending;
252 253

	spin_lock(&watchdog_lock);
254 255
	if (!watchdog_running)
		goto out;
256

257 258
	reset_pending = atomic_read(&watchdog_reset_pending);

259 260 261
	list_for_each_entry(cs, &watchdog_list, wd_list) {

		/* Clocksource already marked unstable? */
262
		if (cs->flags & CLOCK_SOURCE_UNSTABLE) {
263 264
			if (finished_booting)
				schedule_work(&watchdog_work);
265
			continue;
266
		}
267

268
		local_irq_disable();
269
		csnow = cs->read(cs);
270 271
		wdnow = watchdog->read(watchdog);
		local_irq_enable();
T
Thomas Gleixner 已提交
272

273
		/* Clocksource initialized ? */
274 275
		if (!(cs->flags & CLOCK_SOURCE_WATCHDOG) ||
		    atomic_read(&watchdog_reset_pending)) {
276
			cs->flags |= CLOCK_SOURCE_WATCHDOG;
277 278
			cs->wd_last = wdnow;
			cs->cs_last = csnow;
T
Thomas Gleixner 已提交
279 280 281
			continue;
		}

282 283 284 285
		wd_nsec = clocksource_cyc2ns((wdnow - cs->wd_last) & watchdog->mask,
					     watchdog->mult, watchdog->shift);

		cs_nsec = clocksource_cyc2ns((csnow - cs->cs_last) &
286
					     cs->mask, cs->mult, cs->shift);
287 288 289
		cs->cs_last = csnow;
		cs->wd_last = wdnow;

290 291 292
		if (atomic_read(&watchdog_reset_pending))
			continue;

293
		/* Check the deviation from the watchdog clocksource. */
294
		if ((abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD)) {
295 296 297 298 299 300 301 302 303 304 305 306 307 308
			clocksource_unstable(cs, cs_nsec - wd_nsec);
			continue;
		}

		if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) &&
		    (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS) &&
		    (watchdog->flags & CLOCK_SOURCE_IS_CONTINUOUS)) {
			cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
			/*
			 * We just marked the clocksource as highres-capable,
			 * notify the rest of the system as well so that we
			 * transition into high-res mode:
			 */
			tick_clock_notify();
309 310 311
		}
	}

312 313 314 315 316 317 318
	/*
	 * We only clear the watchdog_reset_pending, when we did a
	 * full cycle through all clocksources.
	 */
	if (reset_pending)
		atomic_dec(&watchdog_reset_pending);

319 320 321 322 323 324 325 326 327
	/*
	 * Cycle through CPUs to check if the CPUs stay synchronized
	 * to each other.
	 */
	next_cpu = cpumask_next(raw_smp_processor_id(), cpu_online_mask);
	if (next_cpu >= nr_cpu_ids)
		next_cpu = cpumask_first(cpu_online_mask);
	watchdog_timer.expires += WATCHDOG_INTERVAL;
	add_timer_on(&watchdog_timer, next_cpu);
328
out:
329 330
	spin_unlock(&watchdog_lock);
}
331

332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350
static inline void clocksource_start_watchdog(void)
{
	if (watchdog_running || !watchdog || list_empty(&watchdog_list))
		return;
	init_timer(&watchdog_timer);
	watchdog_timer.function = clocksource_watchdog;
	watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL;
	add_timer_on(&watchdog_timer, cpumask_first(cpu_online_mask));
	watchdog_running = 1;
}

static inline void clocksource_stop_watchdog(void)
{
	if (!watchdog_running || (watchdog && !list_empty(&watchdog_list)))
		return;
	del_timer(&watchdog_timer);
	watchdog_running = 0;
}

351 352 353 354 355 356 357 358
static inline void clocksource_reset_watchdog(void)
{
	struct clocksource *cs;

	list_for_each_entry(cs, &watchdog_list, wd_list)
		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
}

T
Thomas Gleixner 已提交
359 360
static void clocksource_resume_watchdog(void)
{
361
	atomic_inc(&watchdog_reset_pending);
T
Thomas Gleixner 已提交
362 363
}

364
static void clocksource_enqueue_watchdog(struct clocksource *cs)
365 366 367 368 369
{
	unsigned long flags;

	spin_lock_irqsave(&watchdog_lock, flags);
	if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) {
370
		/* cs is a clocksource to be watched. */
371
		list_add(&cs->wd_list, &watchdog_list);
372
		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
373
	} else {
374
		/* cs is a watchdog. */
375
		if (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS)
376
			cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
377
		/* Pick the best watchdog. */
378 379 380
		if (!watchdog || cs->rating > watchdog->rating) {
			watchdog = cs;
			/* Reset watchdog cycles */
381
			clocksource_reset_watchdog();
382 383
		}
	}
384 385
	/* Check if the watchdog timer needs to be started. */
	clocksource_start_watchdog();
386 387
	spin_unlock_irqrestore(&watchdog_lock, flags);
}
388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415

static void clocksource_dequeue_watchdog(struct clocksource *cs)
{
	struct clocksource *tmp;
	unsigned long flags;

	spin_lock_irqsave(&watchdog_lock, flags);
	if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) {
		/* cs is a watched clocksource. */
		list_del_init(&cs->wd_list);
	} else if (cs == watchdog) {
		/* Reset watchdog cycles */
		clocksource_reset_watchdog();
		/* Current watchdog is removed. Find an alternative. */
		watchdog = NULL;
		list_for_each_entry(tmp, &clocksource_list, list) {
			if (tmp == cs || tmp->flags & CLOCK_SOURCE_MUST_VERIFY)
				continue;
			if (!watchdog || tmp->rating > watchdog->rating)
				watchdog = tmp;
		}
	}
	cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
	/* Check if the watchdog timer needs to be stopped. */
	clocksource_stop_watchdog();
	spin_unlock_irqrestore(&watchdog_lock, flags);
}

416
static int clocksource_watchdog_kthread(void *data)
417 418 419
{
	struct clocksource *cs, *tmp;
	unsigned long flags;
420
	LIST_HEAD(unstable);
421

422
	mutex_lock(&clocksource_mutex);
423 424 425 426
	spin_lock_irqsave(&watchdog_lock, flags);
	list_for_each_entry_safe(cs, tmp, &watchdog_list, wd_list)
		if (cs->flags & CLOCK_SOURCE_UNSTABLE) {
			list_del_init(&cs->wd_list);
427
			list_add(&cs->wd_list, &unstable);
428 429 430
		}
	/* Check if the watchdog timer needs to be stopped. */
	clocksource_stop_watchdog();
431 432 433 434 435
	spin_unlock_irqrestore(&watchdog_lock, flags);

	/* Needs to be done outside of watchdog lock */
	list_for_each_entry_safe(cs, tmp, &unstable, wd_list) {
		list_del_init(&cs->wd_list);
436
		__clocksource_change_rating(cs, 0);
437
	}
438
	mutex_unlock(&clocksource_mutex);
439
	return 0;
440 441
}

442 443 444
#else /* CONFIG_CLOCKSOURCE_WATCHDOG */

static void clocksource_enqueue_watchdog(struct clocksource *cs)
445 446 447 448
{
	if (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS)
		cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
}
T
Thomas Gleixner 已提交
449

450
static inline void clocksource_dequeue_watchdog(struct clocksource *cs) { }
T
Thomas Gleixner 已提交
451
static inline void clocksource_resume_watchdog(void) { }
452
static inline int clocksource_watchdog_kthread(void *data) { return 0; }
453 454

#endif /* CONFIG_CLOCKSOURCE_WATCHDOG */
455

M
Magnus Damm 已提交
456 457 458 459 460 461 462 463 464 465 466 467
/**
 * clocksource_suspend - suspend the clocksource(s)
 */
void clocksource_suspend(void)
{
	struct clocksource *cs;

	list_for_each_entry_reverse(cs, &clocksource_list, list)
		if (cs->suspend)
			cs->suspend(cs);
}

T
Thomas Gleixner 已提交
468 469 470 471 472
/**
 * clocksource_resume - resume the clocksource(s)
 */
void clocksource_resume(void)
{
473
	struct clocksource *cs;
T
Thomas Gleixner 已提交
474

475
	list_for_each_entry(cs, &clocksource_list, list)
T
Thomas Gleixner 已提交
476
		if (cs->resume)
477
			cs->resume(cs);
T
Thomas Gleixner 已提交
478 479 480 481

	clocksource_resume_watchdog();
}

J
Jason Wessel 已提交
482 483 484 485
/**
 * clocksource_touch_watchdog - Update watchdog
 *
 * Update the watchdog after exception contexts such as kgdb so as not
486 487
 * to incorrectly trip the watchdog. This might fail when the kernel
 * was stopped in code which holds watchdog_lock.
J
Jason Wessel 已提交
488 489 490 491 492 493
 */
void clocksource_touch_watchdog(void)
{
	clocksource_resume_watchdog();
}

494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509
/**
 * clocksource_max_adjustment- Returns max adjustment amount
 * @cs:         Pointer to clocksource
 *
 */
static u32 clocksource_max_adjustment(struct clocksource *cs)
{
	u64 ret;
	/*
	 * We won't try to correct for more then 11% adjustments (110,000 ppm),
	 */
	ret = (u64)cs->mult * 11;
	do_div(ret,100);
	return (u32)ret;
}

510 511 512 513 514 515 516 517 518 519 520 521
/**
 * clocksource_max_deferment - Returns max time the clocksource can be deferred
 * @cs:         Pointer to clocksource
 *
 */
static u64 clocksource_max_deferment(struct clocksource *cs)
{
	u64 max_nsecs, max_cycles;

	/*
	 * Calculate the maximum number of cycles that we can pass to the
	 * cyc2ns function without overflowing a 64-bit signed result. The
522 523 524 525 526 527 528
	 * maximum number of cycles is equal to ULLONG_MAX/(cs->mult+cs->maxadj)
	 * which is equivalent to the below.
	 * max_cycles < (2^63)/(cs->mult + cs->maxadj)
	 * max_cycles < 2^(log2((2^63)/(cs->mult + cs->maxadj)))
	 * max_cycles < 2^(log2(2^63) - log2(cs->mult + cs->maxadj))
	 * max_cycles < 2^(63 - log2(cs->mult + cs->maxadj))
	 * max_cycles < 1 << (63 - log2(cs->mult + cs->maxadj))
529 530 531 532
	 * Please note that we add 1 to the result of the log2 to account for
	 * any rounding errors, ensure the above inequality is satisfied and
	 * no overflow will occur.
	 */
533
	max_cycles = 1ULL << (63 - (ilog2(cs->mult + cs->maxadj) + 1));
534 535 536 537

	/*
	 * The actual maximum number of cycles we can defer the clocksource is
	 * determined by the minimum of max_cycles and cs->mask.
538 539
	 * Note: Here we subtract the maxadj to make sure we don't sleep for
	 * too long if there's a large negative adjustment.
540 541
	 */
	max_cycles = min_t(u64, max_cycles, (u64) cs->mask);
542 543
	max_nsecs = clocksource_cyc2ns(max_cycles, cs->mult - cs->maxadj,
					cs->shift);
544 545 546 547 548 549 550 551 552 553

	/*
	 * To ensure that the clocksource does not wrap whilst we are idle,
	 * limit the time the clocksource can be deferred by 12.5%. Please
	 * note a margin of 12.5% is used because this can be computed with
	 * a shift, versus say 10% which would require division.
	 */
	return max_nsecs - (max_nsecs >> 5);
}

J
John Stultz 已提交
554
#ifndef CONFIG_ARCH_USES_GETTIMEOFFSET
555 556

/**
557
 * clocksource_select - Select the best clocksource available
558
 *
559
 * Private function. Must hold clocksource_mutex when called.
560
 *
561 562
 * Select the clocksource with the best rating, or the clocksource,
 * which is selected by userspace override.
563
 */
564
static void clocksource_select(void)
565
{
566
	struct clocksource *best, *cs;
567

568
	if (!finished_booting || list_empty(&clocksource_list))
569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592
		return;
	/* First clocksource on the list has the best rating. */
	best = list_first_entry(&clocksource_list, struct clocksource, list);
	/* Check for the override clocksource. */
	list_for_each_entry(cs, &clocksource_list, list) {
		if (strcmp(cs->name, override_name) != 0)
			continue;
		/*
		 * Check to make sure we don't switch to a non-highres
		 * capable clocksource if the tick code is in oneshot
		 * mode (highres or nohz)
		 */
		if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) &&
		    tick_oneshot_mode_active()) {
			/* Override clocksource cannot be used. */
			printk(KERN_WARNING "Override clocksource %s is not "
			       "HRT compatible. Cannot switch while in "
			       "HRT/NOHZ mode\n", cs->name);
			override_name[0] = 0;
		} else
			/* Override clocksource can be used. */
			best = cs;
		break;
	}
593 594 595 596 597
	if (curr_clocksource != best) {
		printk(KERN_INFO "Switching to clocksource %s\n", best->name);
		curr_clocksource = best;
		timekeeping_notify(curr_clocksource);
	}
598
}
599

J
John Stultz 已提交
600
#else /* !CONFIG_ARCH_USES_GETTIMEOFFSET */
601 602 603 604 605

static inline void clocksource_select(void) { }

#endif

606 607 608 609 610 611 612 613 614
/*
 * clocksource_done_booting - Called near the end of core bootup
 *
 * Hack to avoid lots of clocksource churn at boot time.
 * We use fs_initcall because we want this to start before
 * device_initcall but after subsys_initcall.
 */
static int __init clocksource_done_booting(void)
{
615 616 617 618
	mutex_lock(&clocksource_mutex);
	curr_clocksource = clocksource_default_clock();
	mutex_unlock(&clocksource_mutex);

619
	finished_booting = 1;
620 621 622 623 624 625

	/*
	 * Run the watchdog first to eliminate unstable clock sources
	 */
	clocksource_watchdog_kthread(NULL);

626
	mutex_lock(&clocksource_mutex);
627
	clocksource_select();
628
	mutex_unlock(&clocksource_mutex);
629 630 631 632
	return 0;
}
fs_initcall(clocksource_done_booting);

633 634
/*
 * Enqueue the clocksource sorted by rating
635
 */
636
static void clocksource_enqueue(struct clocksource *cs)
637
{
638 639
	struct list_head *entry = &clocksource_list;
	struct clocksource *tmp;
640

641
	list_for_each_entry(tmp, &clocksource_list, list)
642
		/* Keep track of the place, where to insert */
643 644 645
		if (tmp->rating >= cs->rating)
			entry = &tmp->list;
	list_add(&cs->list, entry);
646 647
}

648
/**
649
 * __clocksource_updatefreq_scale - Used update clocksource with new freq
650 651 652 653
 * @t:		clocksource to be registered
 * @scale:	Scale factor multiplied against freq to get clocksource hz
 * @freq:	clocksource frequency (cycles per second) divided by scale
 *
654
 * This should only be called from the clocksource->enable() method.
655 656
 *
 * This *SHOULD NOT* be called directly! Please use the
657
 * clocksource_updatefreq_hz() or clocksource_updatefreq_khz helper functions.
658
 */
659
void __clocksource_updatefreq_scale(struct clocksource *cs, u32 scale, u32 freq)
660
{
661
	u64 sec;
662
	/*
663 664 665 666 667 668 669 670
	 * Calc the maximum number of seconds which we can run before
	 * wrapping around. For clocksources which have a mask > 32bit
	 * we need to limit the max sleep time to have a good
	 * conversion precision. 10 minutes is still a reasonable
	 * amount. That results in a shift value of 24 for a
	 * clocksource with mask >= 40bit and f >= 4GHz. That maps to
	 * ~ 0.06ppm granularity for NTP. We apply the same 12.5%
	 * margin as we do in clocksource_max_deferment()
671
	 */
672 673 674 675 676 677 678 679
	sec = (cs->mask - (cs->mask >> 5));
	do_div(sec, freq);
	do_div(sec, scale);
	if (!sec)
		sec = 1;
	else if (sec > 600 && cs->mask > UINT_MAX)
		sec = 600;

680
	clocks_calc_mult_shift(&cs->mult, &cs->shift, freq,
681
			       NSEC_PER_SEC / scale, sec * scale);
682 683 684 685 686 687 688 689 690 691 692 693 694 695

	/*
	 * for clocksources that have large mults, to avoid overflow.
	 * Since mult may be adjusted by ntp, add an safety extra margin
	 *
	 */
	cs->maxadj = clocksource_max_adjustment(cs);
	while ((cs->mult + cs->maxadj < cs->mult)
		|| (cs->mult - cs->maxadj > cs->mult)) {
		cs->mult >>= 1;
		cs->shift--;
		cs->maxadj = clocksource_max_adjustment(cs);
	}

696
	cs->max_idle_ns = clocksource_max_deferment(cs);
697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713
}
EXPORT_SYMBOL_GPL(__clocksource_updatefreq_scale);

/**
 * __clocksource_register_scale - Used to install new clocksources
 * @t:		clocksource to be registered
 * @scale:	Scale factor multiplied against freq to get clocksource hz
 * @freq:	clocksource frequency (cycles per second) divided by scale
 *
 * Returns -EBUSY if registration fails, zero otherwise.
 *
 * This *SHOULD NOT* be called directly! Please use the
 * clocksource_register_hz() or clocksource_register_khz helper functions.
 */
int __clocksource_register_scale(struct clocksource *cs, u32 scale, u32 freq)
{

714
	/* Initialize mult/shift and max_idle_ns */
715
	__clocksource_updatefreq_scale(cs, scale, freq);
716

717
	/* Add clocksource to the clcoksource list */
718 719 720
	mutex_lock(&clocksource_mutex);
	clocksource_enqueue(cs);
	clocksource_enqueue_watchdog(cs);
721
	clocksource_select();
722 723 724 725 726 727
	mutex_unlock(&clocksource_mutex);
	return 0;
}
EXPORT_SYMBOL_GPL(__clocksource_register_scale);


728
/**
729
 * clocksource_register - Used to install new clocksources
730 731 732 733
 * @t:		clocksource to be registered
 *
 * Returns -EBUSY if registration fails, zero otherwise.
 */
734
int clocksource_register(struct clocksource *cs)
735
{
736 737 738 739 740 741
	/* calculate max adjustment for given mult/shift */
	cs->maxadj = clocksource_max_adjustment(cs);
	WARN_ONCE(cs->mult + cs->maxadj < cs->mult,
		"Clocksource %s might overflow on 11%% adjustment\n",
		cs->name);

742 743 744
	/* calculate max idle time permitted for this clocksource */
	cs->max_idle_ns = clocksource_max_deferment(cs);

745
	mutex_lock(&clocksource_mutex);
746
	clocksource_enqueue(cs);
747
	clocksource_enqueue_watchdog(cs);
748
	clocksource_select();
749
	mutex_unlock(&clocksource_mutex);
750
	return 0;
751
}
752
EXPORT_SYMBOL(clocksource_register);
753

754 755 756 757 758 759 760 761
static void __clocksource_change_rating(struct clocksource *cs, int rating)
{
	list_del(&cs->list);
	cs->rating = rating;
	clocksource_enqueue(cs);
	clocksource_select();
}

762
/**
763
 * clocksource_change_rating - Change the rating of a registered clocksource
764
 */
765
void clocksource_change_rating(struct clocksource *cs, int rating)
766
{
767
	mutex_lock(&clocksource_mutex);
768
	__clocksource_change_rating(cs, rating);
769
	mutex_unlock(&clocksource_mutex);
770
}
771
EXPORT_SYMBOL(clocksource_change_rating);
772

773 774 775 776 777
/**
 * clocksource_unregister - remove a registered clocksource
 */
void clocksource_unregister(struct clocksource *cs)
{
778
	mutex_lock(&clocksource_mutex);
779
	clocksource_dequeue_watchdog(cs);
780
	list_del(&cs->list);
781
	clocksource_select();
782
	mutex_unlock(&clocksource_mutex);
783
}
784
EXPORT_SYMBOL(clocksource_unregister);
785

786
#ifdef CONFIG_SYSFS
787 788 789 790 791 792 793 794
/**
 * sysfs_show_current_clocksources - sysfs interface for current clocksource
 * @dev:	unused
 * @buf:	char buffer to be filled with clocksource list
 *
 * Provides sysfs interface for listing current clocksource.
 */
static ssize_t
795 796
sysfs_show_current_clocksources(struct sys_device *dev,
				struct sysdev_attribute *attr, char *buf)
797
{
798
	ssize_t count = 0;
799

800
	mutex_lock(&clocksource_mutex);
801
	count = snprintf(buf, PAGE_SIZE, "%s\n", curr_clocksource->name);
802
	mutex_unlock(&clocksource_mutex);
803

804
	return count;
805 806 807 808 809 810 811 812 813
}

/**
 * sysfs_override_clocksource - interface for manually overriding clocksource
 * @dev:	unused
 * @buf:	name of override clocksource
 * @count:	length of buffer
 *
 * Takes input from sysfs interface for manually overriding the default
814
 * clocksource selection.
815 816
 */
static ssize_t sysfs_override_clocksource(struct sys_device *dev,
817
					  struct sysdev_attribute *attr,
818 819 820
					  const char *buf, size_t count)
{
	size_t ret = count;
821

822 823 824 825 826 827 828 829
	/* strings from sysfs write are not 0 terminated! */
	if (count >= sizeof(override_name))
		return -EINVAL;

	/* strip of \n: */
	if (buf[count-1] == '\n')
		count--;

830
	mutex_lock(&clocksource_mutex);
831

832 833
	if (count > 0)
		memcpy(override_name, buf, count);
834
	override_name[count] = 0;
835
	clocksource_select();
836

837
	mutex_unlock(&clocksource_mutex);
838 839 840 841 842 843 844 845 846 847 848 849

	return ret;
}

/**
 * sysfs_show_available_clocksources - sysfs interface for listing clocksource
 * @dev:	unused
 * @buf:	char buffer to be filled with clocksource list
 *
 * Provides sysfs interface for listing registered clocksources
 */
static ssize_t
850 851 852
sysfs_show_available_clocksources(struct sys_device *dev,
				  struct sysdev_attribute *attr,
				  char *buf)
853
{
854
	struct clocksource *src;
855
	ssize_t count = 0;
856

857
	mutex_lock(&clocksource_mutex);
858
	list_for_each_entry(src, &clocksource_list, list) {
859 860 861 862 863 864
		/*
		 * Don't show non-HRES clocksource if the tick code is
		 * in one shot mode (highres=on or nohz=on)
		 */
		if (!tick_oneshot_mode_active() ||
		    (src->flags & CLOCK_SOURCE_VALID_FOR_HRES))
865
			count += snprintf(buf + count,
866 867
				  max((ssize_t)PAGE_SIZE - count, (ssize_t)0),
				  "%s ", src->name);
868
	}
869
	mutex_unlock(&clocksource_mutex);
870

871 872
	count += snprintf(buf + count,
			  max((ssize_t)PAGE_SIZE - count, (ssize_t)0), "\n");
873

874
	return count;
875 876 877 878 879
}

/*
 * Sysfs setup bits:
 */
880
static SYSDEV_ATTR(current_clocksource, 0644, sysfs_show_current_clocksources,
D
Daniel Walker 已提交
881
		   sysfs_override_clocksource);
882

883
static SYSDEV_ATTR(available_clocksource, 0444,
D
Daniel Walker 已提交
884
		   sysfs_show_available_clocksources, NULL);
885 886

static struct sysdev_class clocksource_sysclass = {
887
	.name = "clocksource",
888 889 890 891 892 893 894
};

static struct sys_device device_clocksource = {
	.id	= 0,
	.cls	= &clocksource_sysclass,
};

895
static int __init init_clocksource_sysfs(void)
896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912
{
	int error = sysdev_class_register(&clocksource_sysclass);

	if (!error)
		error = sysdev_register(&device_clocksource);
	if (!error)
		error = sysdev_create_file(
				&device_clocksource,
				&attr_current_clocksource);
	if (!error)
		error = sysdev_create_file(
				&device_clocksource,
				&attr_available_clocksource);
	return error;
}

device_initcall(init_clocksource_sysfs);
913
#endif /* CONFIG_SYSFS */
914 915 916 917 918 919 920 921 922 923

/**
 * boot_override_clocksource - boot clock override
 * @str:	override name
 *
 * Takes a clocksource= boot argument and uses it
 * as the clocksource override name.
 */
static int __init boot_override_clocksource(char* str)
{
924
	mutex_lock(&clocksource_mutex);
925 926
	if (str)
		strlcpy(override_name, str, sizeof(override_name));
927
	mutex_unlock(&clocksource_mutex);
928 929 930 931 932 933 934 935 936 937 938 939 940 941
	return 1;
}

__setup("clocksource=", boot_override_clocksource);

/**
 * boot_override_clock - Compatibility layer for deprecated boot option
 * @str:	override name
 *
 * DEPRECATED! Takes a clock= boot argument and uses it
 * as the clocksource override name
 */
static int __init boot_override_clock(char* str)
{
942 943 944 945 946 947 948
	if (!strcmp(str, "pmtmr")) {
		printk("Warning: clock=pmtmr is deprecated. "
			"Use clocksource=acpi_pm.\n");
		return boot_override_clocksource("acpi_pm");
	}
	printk("Warning! clock= boot option is deprecated. "
		"Use clocksource=xyz\n");
949 950 951 952
	return boot_override_clocksource(str);
}

__setup("clock=", boot_override_clock);