watchdog.c 20.6 KB
Newer Older
1 2 3 4 5
/*
 * Detect hard and soft lockups on a system
 *
 * started by Don Zickus, Copyright (C) 2010 Red Hat, Inc.
 *
6 7 8
 * Note: Most of this code is borrowed heavily from the original softlockup
 * detector, so thanks to Ingo for the initial implementation.
 * Some chunks also taken from the old x86-specific nmi watchdog code, thanks
9 10 11
 * to those contributors as well.
 */

12 13
#define pr_fmt(fmt) "NMI watchdog: " fmt

14 15 16 17 18 19
#include <linux/mm.h>
#include <linux/cpu.h>
#include <linux/nmi.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/sysctl.h>
20
#include <linux/smpboot.h>
21
#include <linux/sched/rt.h>
22 23

#include <asm/irq_regs.h>
24
#include <linux/kvm_para.h>
25 26
#include <linux/perf_event.h>

27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
/*
 * The run state of the lockup detectors is controlled by the content of the
 * 'watchdog_enabled' variable. Each lockup detector has its dedicated bit -
 * bit 0 for the hard lockup detector and bit 1 for the soft lockup detector.
 *
 * 'watchdog_user_enabled', 'nmi_watchdog_enabled' and 'soft_watchdog_enabled'
 * are variables that are only used as an 'interface' between the parameters
 * in /proc/sys/kernel and the internal state bits in 'watchdog_enabled'. The
 * 'watchdog_thresh' variable is handled differently because its value is not
 * boolean, and the lockup detectors are 'suspended' while 'watchdog_thresh'
 * is equal zero.
 */
#define NMI_WATCHDOG_ENABLED_BIT   0
#define SOFT_WATCHDOG_ENABLED_BIT  1
#define NMI_WATCHDOG_ENABLED      (1 << NMI_WATCHDOG_ENABLED_BIT)
#define SOFT_WATCHDOG_ENABLED     (1 << SOFT_WATCHDOG_ENABLED_BIT)

#ifdef CONFIG_HARDLOCKUP_DETECTOR
static unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED|NMI_WATCHDOG_ENABLED;
#else
static unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED;
#endif
int __read_mostly nmi_watchdog_enabled;
int __read_mostly soft_watchdog_enabled;
int __read_mostly watchdog_user_enabled;
52
int __read_mostly watchdog_thresh = 10;
53

54 55 56 57 58 59
#ifdef CONFIG_SMP
int __read_mostly sysctl_softlockup_all_cpu_backtrace;
#else
#define sysctl_softlockup_all_cpu_backtrace 0
#endif

60
static int __read_mostly watchdog_running;
61
static u64 __read_mostly sample_period;
62 63 64 65 66 67

static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts);
static DEFINE_PER_CPU(struct task_struct *, softlockup_watchdog);
static DEFINE_PER_CPU(struct hrtimer, watchdog_hrtimer);
static DEFINE_PER_CPU(bool, softlockup_touch_sync);
static DEFINE_PER_CPU(bool, soft_watchdog_warn);
68 69
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts);
static DEFINE_PER_CPU(unsigned long, soft_lockup_hrtimer_cnt);
70
static DEFINE_PER_CPU(struct task_struct *, softlockup_task_ptr_saved);
71
#ifdef CONFIG_HARDLOCKUP_DETECTOR
72 73
static DEFINE_PER_CPU(bool, hard_watchdog_warn);
static DEFINE_PER_CPU(bool, watchdog_nmi_touch);
74 75 76
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved);
static DEFINE_PER_CPU(struct perf_event *, watchdog_ev);
#endif
77
static unsigned long soft_lockup_nmi_warn;
78 79 80 81 82

/* boot commands */
/*
 * Should we panic when a soft-lockup or hard-lockup occurs:
 */
83
#ifdef CONFIG_HARDLOCKUP_DETECTOR
84 85
static int hardlockup_panic =
			CONFIG_BOOTPARAM_HARDLOCKUP_PANIC_VALUE;
86

87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
static bool hardlockup_detector_enabled = true;
/*
 * We may not want to enable hard lockup detection by default in all cases,
 * for example when running the kernel as a guest on a hypervisor. In these
 * cases this function can be called to disable hard lockup detection. This
 * function should only be executed once by the boot processor before the
 * kernel command line parameters are parsed, because otherwise it is not
 * possible to override this in hardlockup_panic_setup().
 */
void watchdog_enable_hardlockup_detector(bool val)
{
	hardlockup_detector_enabled = val;
}

bool watchdog_hardlockup_detector_is_enabled(void)
{
	return hardlockup_detector_enabled;
}

106 107 108 109
static int __init hardlockup_panic_setup(char *str)
{
	if (!strncmp(str, "panic", 5))
		hardlockup_panic = 1;
110 111
	else if (!strncmp(str, "nopanic", 7))
		hardlockup_panic = 0;
112
	else if (!strncmp(str, "0", 1))
113
		watchdog_user_enabled = 0;
114 115 116 117 118 119 120 121
	else if (!strncmp(str, "1", 1) || !strncmp(str, "2", 1)) {
		/*
		 * Setting 'nmi_watchdog=1' or 'nmi_watchdog=2' (legacy option)
		 * has the same effect.
		 */
		watchdog_user_enabled = 1;
		watchdog_enable_hardlockup_detector(true);
	}
122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139
	return 1;
}
__setup("nmi_watchdog=", hardlockup_panic_setup);
#endif

unsigned int __read_mostly softlockup_panic =
			CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE;

static int __init softlockup_panic_setup(char *str)
{
	softlockup_panic = simple_strtoul(str, NULL, 0);

	return 1;
}
__setup("softlockup_panic=", softlockup_panic_setup);

static int __init nowatchdog_setup(char *str)
{
140
	watchdog_user_enabled = 0;
141 142 143 144 145 146 147
	return 1;
}
__setup("nowatchdog", nowatchdog_setup);

/* deprecated */
static int __init nosoftlockup_setup(char *str)
{
148
	watchdog_user_enabled = 0;
149 150 151 152
	return 1;
}
__setup("nosoftlockup", nosoftlockup_setup);
/*  */
153 154 155 156 157 158 159 160 161
#ifdef CONFIG_SMP
static int __init softlockup_all_cpu_backtrace_setup(char *str)
{
	sysctl_softlockup_all_cpu_backtrace =
		!!simple_strtol(str, NULL, 0);
	return 1;
}
__setup("softlockup_all_cpu_backtrace=", softlockup_all_cpu_backtrace_setup);
#endif
162

163 164 165 166 167 168 169
/*
 * Hard-lockup warnings should be triggered after just a few seconds. Soft-
 * lockups can have false positives under extreme conditions. So we generally
 * want a higher threshold for soft lockups than for hard lockups. So we couple
 * the thresholds with a factor: we make the soft threshold twice the amount of
 * time the hard threshold is.
 */
170
static int get_softlockup_thresh(void)
171 172 173
{
	return watchdog_thresh * 2;
}
174 175 176 177 178 179

/*
 * Returns seconds, approximately.  We don't need nanosecond
 * resolution, and we don't need to waste time with a big divide when
 * 2^30ns == 1.074s.
 */
180
static unsigned long get_timestamp(void)
181
{
182
	return running_clock() >> 30LL;  /* 2^30 ~= 10^9 */
183 184
}

185
static void set_sample_period(void)
186 187
{
	/*
188
	 * convert watchdog_thresh from seconds to ns
189 190 191 192
	 * the divide by 5 is to give hrtimer several chances (two
	 * or three with the current relation between the soft
	 * and hard thresholds) to increment before the
	 * hardlockup detector generates a warning
193
	 */
194
	sample_period = get_softlockup_thresh() * ((u64)NSEC_PER_SEC / 5);
195 196 197 198 199
}

/* Commands for resetting the watchdog */
static void __touch_watchdog(void)
{
200
	__this_cpu_write(watchdog_touch_ts, get_timestamp());
201 202
}

203
void touch_softlockup_watchdog(void)
204
{
205 206 207 208 209
	/*
	 * Preemption can be enabled.  It doesn't matter which CPU's timestamp
	 * gets zeroed here, so use the raw_ operation.
	 */
	raw_cpu_write(watchdog_touch_ts, 0);
210
}
211
EXPORT_SYMBOL(touch_softlockup_watchdog);
212

213
void touch_all_softlockup_watchdogs(void)
214 215 216 217 218 219 220 221 222 223 224 225
{
	int cpu;

	/*
	 * this is done lockless
	 * do we care if a 0 races with a timestamp?
	 * all it means is the softlock check starts one cycle later
	 */
	for_each_online_cpu(cpu)
		per_cpu(watchdog_touch_ts, cpu) = 0;
}

226
#ifdef CONFIG_HARDLOCKUP_DETECTOR
227 228
void touch_nmi_watchdog(void)
{
229 230 231 232 233 234 235
	/*
	 * Using __raw here because some code paths have
	 * preemption enabled.  If preemption is enabled
	 * then interrupts should be enabled too, in which
	 * case we shouldn't have to worry about the watchdog
	 * going off.
	 */
236
	raw_cpu_write(watchdog_nmi_touch, true);
237
	touch_softlockup_watchdog();
238 239 240
}
EXPORT_SYMBOL(touch_nmi_watchdog);

241 242
#endif

243 244
void touch_softlockup_watchdog_sync(void)
{
245 246
	__this_cpu_write(softlockup_touch_sync, true);
	__this_cpu_write(watchdog_touch_ts, 0);
247 248
}

249
#ifdef CONFIG_HARDLOCKUP_DETECTOR
250
/* watchdog detector functions */
251
static int is_hardlockup(void)
252
{
253
	unsigned long hrint = __this_cpu_read(hrtimer_interrupts);
254

255
	if (__this_cpu_read(hrtimer_interrupts_saved) == hrint)
256 257
		return 1;

258
	__this_cpu_write(hrtimer_interrupts_saved, hrint);
259 260 261 262
	return 0;
}
#endif

263
static int is_softlockup(unsigned long touch_ts)
264
{
265
	unsigned long now = get_timestamp();
266 267

	/* Warn about unreasonable delays: */
268
	if (time_after(now, touch_ts + get_softlockup_thresh()))
269 270 271 272 273
		return now - touch_ts;

	return 0;
}

274
#ifdef CONFIG_HARDLOCKUP_DETECTOR
275

276 277 278 279 280 281 282 283 284
static struct perf_event_attr wd_hw_attr = {
	.type		= PERF_TYPE_HARDWARE,
	.config		= PERF_COUNT_HW_CPU_CYCLES,
	.size		= sizeof(struct perf_event_attr),
	.pinned		= 1,
	.disabled	= 1,
};

/* Callback function for perf event subsystem */
285
static void watchdog_overflow_callback(struct perf_event *event,
286 287 288
		 struct perf_sample_data *data,
		 struct pt_regs *regs)
{
289 290 291
	/* Ensure the watchdog never gets throttled */
	event->hw.interrupts = 0;

292 293
	if (__this_cpu_read(watchdog_nmi_touch) == true) {
		__this_cpu_write(watchdog_nmi_touch, false);
294 295 296 297 298 299 300 301 302
		return;
	}

	/* check for a hardlockup
	 * This is done by making sure our timer interrupt
	 * is incrementing.  The timer interrupt should have
	 * fired multiple times before we overflow'd.  If it hasn't
	 * then this is a good indication the cpu is stuck
	 */
303 304 305
	if (is_hardlockup()) {
		int this_cpu = smp_processor_id();

306
		/* only print hardlockups once */
307
		if (__this_cpu_read(hard_watchdog_warn) == true)
308 309 310
			return;

		if (hardlockup_panic)
311 312
			panic("Watchdog detected hard LOCKUP on cpu %d",
			      this_cpu);
313
		else
314 315
			WARN(1, "Watchdog detected hard LOCKUP on cpu %d",
			     this_cpu);
316

317
		__this_cpu_write(hard_watchdog_warn, true);
318 319 320
		return;
	}

321
	__this_cpu_write(hard_watchdog_warn, false);
322 323
	return;
}
324 325
#endif /* CONFIG_HARDLOCKUP_DETECTOR */

326 327
static void watchdog_interrupt_count(void)
{
328
	__this_cpu_inc(hrtimer_interrupts);
329
}
330 331 332

static int watchdog_nmi_enable(unsigned int cpu);
static void watchdog_nmi_disable(unsigned int cpu);
333 334 335 336

/* watchdog kicker functions */
static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer)
{
337
	unsigned long touch_ts = __this_cpu_read(watchdog_touch_ts);
338 339
	struct pt_regs *regs = get_irq_regs();
	int duration;
340
	int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace;
341 342 343 344 345

	/* kick the hardlockup detector */
	watchdog_interrupt_count();

	/* kick the softlockup detector */
346
	wake_up_process(__this_cpu_read(softlockup_watchdog));
347 348

	/* .. and repeat */
349
	hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period));
350 351

	if (touch_ts == 0) {
352
		if (unlikely(__this_cpu_read(softlockup_touch_sync))) {
353 354 355 356
			/*
			 * If the time stamp was touched atomically
			 * make sure the scheduler tick is up to date.
			 */
357
			__this_cpu_write(softlockup_touch_sync, false);
358 359
			sched_clock_tick();
		}
360 361 362

		/* Clear the guest paused flag on watchdog reset */
		kvm_check_and_clear_guest_paused();
363 364 365 366 367 368 369 370 371 372
		__touch_watchdog();
		return HRTIMER_RESTART;
	}

	/* check for a softlockup
	 * This is done by making sure a high priority task is
	 * being scheduled.  The task touches the watchdog to
	 * indicate it is getting cpu time.  If it hasn't then
	 * this is a good indication some task is hogging the cpu
	 */
373
	duration = is_softlockup(touch_ts);
374
	if (unlikely(duration)) {
375 376 377 378 379 380 381 382
		/*
		 * If a virtual machine is stopped by the host it can look to
		 * the watchdog like a soft lockup, check to see if the host
		 * stopped the vm before we issue the warning
		 */
		if (kvm_check_and_clear_guest_paused())
			return HRTIMER_RESTART;

383
		/* only warn once */
384 385 386 387 388 389 390 391 392 393 394 395 396 397
		if (__this_cpu_read(soft_watchdog_warn) == true) {
			/*
			 * When multiple processes are causing softlockups the
			 * softlockup detector only warns on the first one
			 * because the code relies on a full quiet cycle to
			 * re-arm.  The second process prevents the quiet cycle
			 * and never gets reported.  Use task pointers to detect
			 * this.
			 */
			if (__this_cpu_read(softlockup_task_ptr_saved) !=
			    current) {
				__this_cpu_write(soft_watchdog_warn, false);
				__touch_watchdog();
			}
398
			return HRTIMER_RESTART;
399
		}
400

401 402 403 404 405 406 407 408 409 410 411
		if (softlockup_all_cpu_backtrace) {
			/* Prevent multiple soft-lockup reports if one cpu is already
			 * engaged in dumping cpu back traces
			 */
			if (test_and_set_bit(0, &soft_lockup_nmi_warn)) {
				/* Someone else will report us. Let's give up */
				__this_cpu_write(soft_watchdog_warn, true);
				return HRTIMER_RESTART;
			}
		}

412
		pr_emerg("BUG: soft lockup - CPU#%d stuck for %us! [%s:%d]\n",
413
			smp_processor_id(), duration,
414
			current->comm, task_pid_nr(current));
415
		__this_cpu_write(softlockup_task_ptr_saved, current);
416 417 418 419 420 421 422
		print_modules();
		print_irqtrace_events(current);
		if (regs)
			show_regs(regs);
		else
			dump_stack();

423 424 425 426 427 428 429 430 431 432 433
		if (softlockup_all_cpu_backtrace) {
			/* Avoid generating two back traces for current
			 * given that one is already made above
			 */
			trigger_allbutself_cpu_backtrace();

			clear_bit(0, &soft_lockup_nmi_warn);
			/* Barrier to sync with other cpus */
			smp_mb__after_atomic();
		}

Josh Hunt's avatar
Josh Hunt committed
434
		add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK);
435 436
		if (softlockup_panic)
			panic("softlockup: hung tasks");
437
		__this_cpu_write(soft_watchdog_warn, true);
438
	} else
439
		__this_cpu_write(soft_watchdog_warn, false);
440 441 442 443

	return HRTIMER_RESTART;
}

444 445 446
static void watchdog_set_prio(unsigned int policy, unsigned int prio)
{
	struct sched_param param = { .sched_priority = prio };
447

448 449 450 451
	sched_setscheduler(current, policy, &param);
}

static void watchdog_enable(unsigned int cpu)
452
{
453
	struct hrtimer *hrtimer = raw_cpu_ptr(&watchdog_hrtimer);
454

455 456 457 458
	/* kick off the timer for the hardlockup detector */
	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
	hrtimer->function = watchdog_timer_fn;

459 460
	/* Enable the perf event */
	watchdog_nmi_enable(cpu);
461 462

	/* done here because hrtimer_start can only pin to smp_processor_id() */
463
	hrtimer_start(hrtimer, ns_to_ktime(sample_period),
464 465
		      HRTIMER_MODE_REL_PINNED);

466 467 468 469
	/* initialize timestamp */
	watchdog_set_prio(SCHED_FIFO, MAX_RT_PRIO - 1);
	__touch_watchdog();
}
470

471 472
static void watchdog_disable(unsigned int cpu)
{
473
	struct hrtimer *hrtimer = raw_cpu_ptr(&watchdog_hrtimer);
474

475 476 477 478
	watchdog_set_prio(SCHED_NORMAL, 0);
	hrtimer_cancel(hrtimer);
	/* disable the perf event */
	watchdog_nmi_disable(cpu);
479 480
}

481 482 483 484 485
static void watchdog_cleanup(unsigned int cpu, bool online)
{
	watchdog_disable(cpu);
}

486 487 488 489 490 491 492 493 494
static int watchdog_should_run(unsigned int cpu)
{
	return __this_cpu_read(hrtimer_interrupts) !=
		__this_cpu_read(soft_lockup_hrtimer_cnt);
}

/*
 * The watchdog thread function - touches the timestamp.
 *
495
 * It only runs once every sample_period seconds (4 seconds by
496 497 498 499 500 501 502 503 504 505
 * default) to reset the softlockup timestamp. If this gets delayed
 * for more than 2*watchdog_thresh seconds then the debug-printout
 * triggers in watchdog_timer_fn().
 */
static void watchdog(unsigned int cpu)
{
	__this_cpu_write(soft_lockup_hrtimer_cnt,
			 __this_cpu_read(hrtimer_interrupts));
	__touch_watchdog();
}
506

507
#ifdef CONFIG_HARDLOCKUP_DETECTOR
508 509 510 511 512 513 514
/*
 * People like the simple clean cpu node info on boot.
 * Reduce the watchdog noise by only printing messages
 * that are different from what cpu0 displayed.
 */
static unsigned long cpu0_err;

515
static int watchdog_nmi_enable(unsigned int cpu)
516 517 518 519
{
	struct perf_event_attr *wd_attr;
	struct perf_event *event = per_cpu(watchdog_ev, cpu);

520 521 522 523 524 525 526 527 528
	/*
	 * Some kernels need to default hard lockup detection to
	 * 'disabled', for example a guest on a hypervisor.
	 */
	if (!watchdog_hardlockup_detector_is_enabled()) {
		event = ERR_PTR(-ENOENT);
		goto handle_err;
	}

529 530 531 532 533 534 535 536 537
	/* is it already setup and enabled? */
	if (event && event->state > PERF_EVENT_STATE_OFF)
		goto out;

	/* it is setup but not enabled */
	if (event != NULL)
		goto out_enable;

	wd_attr = &wd_hw_attr;
538
	wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh);
539 540

	/* Try to register using hardware perf events */
541
	event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback, NULL);
542

543
handle_err:
544 545 546 547
	/* save cpu0 error for future comparision */
	if (cpu == 0 && IS_ERR(event))
		cpu0_err = PTR_ERR(event);

548
	if (!IS_ERR(event)) {
549 550 551
		/* only print for cpu0 or different than cpu0 */
		if (cpu == 0 || cpu0_err)
			pr_info("enabled on all CPUs, permanently consumes one hw-PMU counter.\n");
552 553 554
		goto out_save;
	}

555 556 557
	/* skip displaying the same error again */
	if (cpu > 0 && (PTR_ERR(event) == cpu0_err))
		return PTR_ERR(event);
558 559 560

	/* vary the KERN level based on the returned errno */
	if (PTR_ERR(event) == -EOPNOTSUPP)
561
		pr_info("disabled (cpu%i): not supported (no LAPIC?)\n", cpu);
562
	else if (PTR_ERR(event) == -ENOENT)
563
		pr_warn("disabled (cpu%i): hardware events not enabled\n",
564
			 cpu);
565
	else
566 567
		pr_err("disabled (cpu%i): unable to create perf event: %ld\n",
			cpu, PTR_ERR(event));
568
	return PTR_ERR(event);
569 570 571 572 573 574 575 576 577 578

	/* success path */
out_save:
	per_cpu(watchdog_ev, cpu) = event;
out_enable:
	perf_event_enable(per_cpu(watchdog_ev, cpu));
out:
	return 0;
}

579
static void watchdog_nmi_disable(unsigned int cpu)
580 581 582 583 584 585 586 587 588 589
{
	struct perf_event *event = per_cpu(watchdog_ev, cpu);

	if (event) {
		perf_event_disable(event);
		per_cpu(watchdog_ev, cpu) = NULL;

		/* should be in cleanup, but blocks oprofile */
		perf_event_release_kernel(event);
	}
590 591 592 593
	if (cpu == 0) {
		/* watchdog_nmi_enable() expects this to be zero initially. */
		cpu0_err = 0;
	}
594 595
}
#else
596 597
static int watchdog_nmi_enable(unsigned int cpu) { return 0; }
static void watchdog_nmi_disable(unsigned int cpu) { return; }
598
#endif /* CONFIG_HARDLOCKUP_DETECTOR */
599

600 601 602 603 604 605 606 607 608 609 610
static struct smp_hotplug_thread watchdog_threads = {
	.store			= &softlockup_watchdog,
	.thread_should_run	= watchdog_should_run,
	.thread_fn		= watchdog,
	.thread_comm		= "watchdog/%u",
	.setup			= watchdog_enable,
	.cleanup		= watchdog_cleanup,
	.park			= watchdog_disable,
	.unpark			= watchdog_enable,
};

611 612
static void restart_watchdog_hrtimer(void *info)
{
613
	struct hrtimer *hrtimer = raw_cpu_ptr(&watchdog_hrtimer);
614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638
	int ret;

	/*
	 * No need to cancel and restart hrtimer if it is currently executing
	 * because it will reprogram itself with the new period now.
	 * We should never see it unqueued here because we are running per-cpu
	 * with interrupts disabled.
	 */
	ret = hrtimer_try_to_cancel(hrtimer);
	if (ret == 1)
		hrtimer_start(hrtimer, ns_to_ktime(sample_period),
				HRTIMER_MODE_REL_PINNED);
}

static void update_timers(int cpu)
{
	/*
	 * Make sure that perf event counter will adopt to a new
	 * sampling period. Updating the sampling period directly would
	 * be much nicer but we do not have an API for that now so
	 * let's use a big hammer.
	 * Hrtimer will adopt the new period on the next tick but this
	 * might be late already so we have to restart the timer as well.
	 */
	watchdog_nmi_disable(cpu);
639
	smp_call_function_single(cpu, restart_watchdog_hrtimer, NULL, 1);
640 641 642 643 644 645 646 647 648 649 650 651 652 653
	watchdog_nmi_enable(cpu);
}

static void update_timers_all_cpus(void)
{
	int cpu;

	get_online_cpus();
	for_each_online_cpu(cpu)
		update_timers(cpu);
	put_online_cpus();
}

static int watchdog_enable_all_cpus(bool sample_period_changed)
654
{
655
	int err = 0;
656

657
	if (!watchdog_running) {
658 659 660 661
		err = smpboot_register_percpu_thread(&watchdog_threads);
		if (err)
			pr_err("Failed to create watchdog threads, disabled\n");
		else
662
			watchdog_running = 1;
663 664
	} else if (sample_period_changed) {
		update_timers_all_cpus();
665
	}
666 667

	return err;
668 669
}

670 671 672
/* prepare/enable/disable routines */
/* sysctl functions */
#ifdef CONFIG_SYSCTL
673 674
static void watchdog_disable_all_cpus(void)
{
675 676
	if (watchdog_running) {
		watchdog_running = 0;
677
		smpboot_unregister_percpu_thread(&watchdog_threads);
678
	}
679 680
}

681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
/*
 * Update the run state of the lockup detectors.
 */
static int proc_watchdog_update(void)
{
	int err = 0;

	/*
	 * Watchdog threads won't be started if they are already active.
	 * The 'watchdog_running' variable in watchdog_*_all_cpus() takes
	 * care of this. If those threads are already active, the sample
	 * period will be updated and the lockup detectors will be enabled
	 * or disabled 'on the fly'.
	 */
	if (watchdog_enabled && watchdog_thresh)
		err = watchdog_enable_all_cpus(true);
	else
		watchdog_disable_all_cpus();

	return err;

}

704 705
static DEFINE_MUTEX(watchdog_proc_mutex);

706
/*
707
 * proc handler for /proc/sys/kernel/nmi_watchdog,watchdog_thresh
708 709
 */

710 711
int proc_dowatchdog(struct ctl_table *table, int write,
		    void __user *buffer, size_t *lenp, loff_t *ppos)
712
{
713
	int err, old_thresh, old_enabled;
714
	bool old_hardlockup;
715

716
	mutex_lock(&watchdog_proc_mutex);
717
	old_thresh = ACCESS_ONCE(watchdog_thresh);
718
	old_enabled = ACCESS_ONCE(watchdog_user_enabled);
719
	old_hardlockup = watchdog_hardlockup_detector_is_enabled();
720

721 722
	err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
	if (err || !write)
723
		goto out;
724

725
	set_sample_period();
726 727
	/*
	 * Watchdog threads shouldn't be enabled if they are
728
	 * disabled. The 'watchdog_running' variable check in
729 730
	 * watchdog_*_all_cpus() function takes care of this.
	 */
731 732 733 734 735 736 737
	if (watchdog_user_enabled && watchdog_thresh) {
		/*
		 * Prevent a change in watchdog_thresh accidentally overriding
		 * the enablement of the hardlockup detector.
		 */
		if (watchdog_user_enabled != old_enabled)
			watchdog_enable_hardlockup_detector(true);
738
		err = watchdog_enable_all_cpus(old_thresh != watchdog_thresh);
739
	} else
740 741
		watchdog_disable_all_cpus();

742 743 744
	/* Restore old values on failure */
	if (err) {
		watchdog_thresh = old_thresh;
745
		watchdog_user_enabled = old_enabled;
746
		watchdog_enable_hardlockup_detector(old_hardlockup);
747
	}
748 749
out:
	mutex_unlock(&watchdog_proc_mutex);
750
	return err;
751 752 753
}
#endif /* CONFIG_SYSCTL */

754
void __init lockup_detector_init(void)
755
{
756
	set_sample_period();
757

758
	if (watchdog_user_enabled)
759
		watchdog_enable_all_cpus(false);
760
}