watchdog.c 22.8 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
#include <uapi/linux/sched/types.h>
23
#include <linux/tick.h>
24
#include <linux/workqueue.h>
25
#include <linux/sched/clock.h>
26
#include <linux/sched/debug.h>
27 28

#include <asm/irq_regs.h>
29
#include <linux/kvm_para.h>
30
#include <linux/kthread.h>
31

32 33
static DEFINE_MUTEX(watchdog_proc_mutex);

34 35
#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR)
unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED|NMI_WATCHDOG_ENABLED;
36
#else
37
unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED;
38 39 40 41
#endif
int __read_mostly nmi_watchdog_enabled;
int __read_mostly soft_watchdog_enabled;
int __read_mostly watchdog_user_enabled;
42
int __read_mostly watchdog_thresh = 10;
43

44 45
#ifdef CONFIG_SMP
int __read_mostly sysctl_softlockup_all_cpu_backtrace;
46
int __read_mostly sysctl_hardlockup_all_cpu_backtrace;
47
#endif
48 49 50 51 52 53
static struct cpumask watchdog_cpumask __read_mostly;
unsigned long *watchdog_cpumask_bits = cpumask_bits(&watchdog_cpumask);

/* Helper for online, unparked cpus. */
#define for_each_watchdog_cpu(cpu) \
	for_each_cpu_and((cpu), cpu_online_mask, &watchdog_cpumask)
54

55 56
atomic_t watchdog_park_in_progress = ATOMIC_INIT(0);

57 58 59 60 61
/*
 * The 'watchdog_running' variable is set to 1 when the watchdog threads
 * are registered/started and is set to 0 when the watchdog threads are
 * unregistered/stopped, so it is an indicator whether the threads exist.
 */
62
static int __read_mostly watchdog_running;
63 64 65 66 67 68 69 70 71 72 73 74 75 76
/*
 * If a subsystem has a need to deactivate the watchdog temporarily, it
 * can use the suspend/resume interface to achieve this. The content of
 * the 'watchdog_suspended' variable reflects this state. Existing threads
 * are parked/unparked by the lockup_detector_{suspend|resume} functions
 * (see comment blocks pertaining to those functions for further details).
 *
 * 'watchdog_suspended' also prevents threads from being registered/started
 * or unregistered/stopped via parameters in /proc/sys/kernel, so the state
 * of 'watchdog_running' cannot change while the watchdog is deactivated
 * temporarily (see related code in 'proc' handlers).
 */
static int __read_mostly watchdog_suspended;

77
static u64 __read_mostly sample_period;
78 79 80 81 82 83

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);
84 85
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts);
static DEFINE_PER_CPU(unsigned long, soft_lockup_hrtimer_cnt);
86
static DEFINE_PER_CPU(struct task_struct *, softlockup_task_ptr_saved);
87
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved);
88
static unsigned long soft_lockup_nmi_warn;
89 90 91 92 93 94 95 96 97 98 99 100 101 102

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)
{
103
	watchdog_enabled = 0;
104 105 106 107 108 109
	return 1;
}
__setup("nowatchdog", nowatchdog_setup);

static int __init nosoftlockup_setup(char *str)
{
110
	watchdog_enabled &= ~SOFT_WATCHDOG_ENABLED;
111 112 113
	return 1;
}
__setup("nosoftlockup", nosoftlockup_setup);
114

115 116 117 118 119 120 121 122
#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);
123 124 125 126 127 128 129
static int __init hardlockup_all_cpu_backtrace_setup(char *str)
{
	sysctl_hardlockup_all_cpu_backtrace =
		!!simple_strtol(str, NULL, 0);
	return 1;
}
__setup("hardlockup_all_cpu_backtrace=", hardlockup_all_cpu_backtrace_setup);
130
#endif
131

132 133 134 135 136 137 138
/*
 * 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.
 */
139
static int get_softlockup_thresh(void)
140 141 142
{
	return watchdog_thresh * 2;
}
143 144 145 146 147 148

/*
 * 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.
 */
149
static unsigned long get_timestamp(void)
150
{
151
	return running_clock() >> 30LL;  /* 2^30 ~= 10^9 */
152 153
}

154
static void set_sample_period(void)
155 156
{
	/*
157
	 * convert watchdog_thresh from seconds to ns
158 159 160 161
	 * 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
162
	 */
163
	sample_period = get_softlockup_thresh() * ((u64)NSEC_PER_SEC / 5);
164 165 166 167 168
}

/* Commands for resetting the watchdog */
static void __touch_watchdog(void)
{
169
	__this_cpu_write(watchdog_touch_ts, get_timestamp());
170 171
}

172 173 174 175 176 177 178 179 180
/**
 * touch_softlockup_watchdog_sched - touch watchdog on scheduler stalls
 *
 * Call when the scheduler may have stalled for legitimate reasons
 * preventing the watchdog task from executing - e.g. the scheduler
 * entering idle state.  This should only be used for scheduler events.
 * Use touch_softlockup_watchdog() for everything else.
 */
void touch_softlockup_watchdog_sched(void)
181
{
182 183 184 185 186
	/*
	 * 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);
187
}
188 189 190 191

void touch_softlockup_watchdog(void)
{
	touch_softlockup_watchdog_sched();
192
	wq_watchdog_touch(raw_smp_processor_id());
193
}
194
EXPORT_SYMBOL(touch_softlockup_watchdog);
195

196
void touch_all_softlockup_watchdogs(void)
197 198 199 200 201 202 203 204
{
	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
	 */
205
	for_each_watchdog_cpu(cpu)
206
		per_cpu(watchdog_touch_ts, cpu) = 0;
207
	wq_watchdog_touch(-1);
208 209 210 211
}

void touch_softlockup_watchdog_sync(void)
{
212 213
	__this_cpu_write(softlockup_touch_sync, true);
	__this_cpu_write(watchdog_touch_ts, 0);
214 215 216
}

/* watchdog detector functions */
217
bool is_hardlockup(void)
218
{
219
	unsigned long hrint = __this_cpu_read(hrtimer_interrupts);
220

221
	if (__this_cpu_read(hrtimer_interrupts_saved) == hrint)
222
		return true;
223

224
	__this_cpu_write(hrtimer_interrupts_saved, hrint);
225
	return false;
226 227
}

228
static int is_softlockup(unsigned long touch_ts)
229
{
230
	unsigned long now = get_timestamp();
231

232
	if ((watchdog_enabled & SOFT_WATCHDOG_ENABLED) && watchdog_thresh){
233 234 235 236
		/* Warn about unreasonable delays. */
		if (time_after(now, touch_ts + get_softlockup_thresh()))
			return now - touch_ts;
	}
237 238 239 240 241
	return 0;
}

static void watchdog_interrupt_count(void)
{
242
	__this_cpu_inc(hrtimer_interrupts);
243
}
244

245 246 247 248 249 250 251 252 253 254 255
/*
 * These two functions are mostly architecture specific
 * defining them as weak here.
 */
int __weak watchdog_nmi_enable(unsigned int cpu)
{
	return 0;
}
void __weak watchdog_nmi_disable(unsigned int cpu)
{
}
256

257 258 259
static int watchdog_enable_all_cpus(void);
static void watchdog_disable_all_cpus(void);

260 261 262
/* watchdog kicker functions */
static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer)
{
263
	unsigned long touch_ts = __this_cpu_read(watchdog_touch_ts);
264 265
	struct pt_regs *regs = get_irq_regs();
	int duration;
266
	int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace;
267

268 269 270
	if (atomic_read(&watchdog_park_in_progress) != 0)
		return HRTIMER_NORESTART;

271 272 273 274
	/* kick the hardlockup detector */
	watchdog_interrupt_count();

	/* kick the softlockup detector */
275
	wake_up_process(__this_cpu_read(softlockup_watchdog));
276 277

	/* .. and repeat */
278
	hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period));
279 280

	if (touch_ts == 0) {
281
		if (unlikely(__this_cpu_read(softlockup_touch_sync))) {
282 283 284 285
			/*
			 * If the time stamp was touched atomically
			 * make sure the scheduler tick is up to date.
			 */
286
			__this_cpu_write(softlockup_touch_sync, false);
287 288
			sched_clock_tick();
		}
289 290 291

		/* Clear the guest paused flag on watchdog reset */
		kvm_check_and_clear_guest_paused();
292 293 294 295 296 297 298 299 300 301
		__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
	 */
302
	duration = is_softlockup(touch_ts);
303
	if (unlikely(duration)) {
304 305 306 307 308 309 310 311
		/*
		 * 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;

312
		/* only warn once */
313 314 315 316 317 318 319 320 321 322 323 324 325 326
		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();
			}
327
			return HRTIMER_RESTART;
328
		}
329

330 331 332 333 334 335 336 337 338 339 340
		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;
			}
		}

341
		pr_emerg("BUG: soft lockup - CPU#%d stuck for %us! [%s:%d]\n",
342
			smp_processor_id(), duration,
343
			current->comm, task_pid_nr(current));
344
		__this_cpu_write(softlockup_task_ptr_saved, current);
345 346 347 348 349 350 351
		print_modules();
		print_irqtrace_events(current);
		if (regs)
			show_regs(regs);
		else
			dump_stack();

352 353 354 355 356 357 358 359 360 361 362
		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
363
		add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK);
364 365
		if (softlockup_panic)
			panic("softlockup: hung tasks");
366
		__this_cpu_write(soft_watchdog_warn, true);
367
	} else
368
		__this_cpu_write(soft_watchdog_warn, false);
369 370 371 372

	return HRTIMER_RESTART;
}

373 374 375
static void watchdog_set_prio(unsigned int policy, unsigned int prio)
{
	struct sched_param param = { .sched_priority = prio };
376

377 378 379 380
	sched_setscheduler(current, policy, &param);
}

static void watchdog_enable(unsigned int cpu)
381
{
382
	struct hrtimer *hrtimer = raw_cpu_ptr(&watchdog_hrtimer);
383

384 385 386 387
	/* kick off the timer for the hardlockup detector */
	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
	hrtimer->function = watchdog_timer_fn;

388 389
	/* Enable the perf event */
	watchdog_nmi_enable(cpu);
390 391

	/* done here because hrtimer_start can only pin to smp_processor_id() */
392
	hrtimer_start(hrtimer, ns_to_ktime(sample_period),
393 394
		      HRTIMER_MODE_REL_PINNED);

395 396 397 398
	/* initialize timestamp */
	watchdog_set_prio(SCHED_FIFO, MAX_RT_PRIO - 1);
	__touch_watchdog();
}
399

400 401
static void watchdog_disable(unsigned int cpu)
{
402
	struct hrtimer *hrtimer = raw_cpu_ptr(&watchdog_hrtimer);
403

404 405 406 407
	watchdog_set_prio(SCHED_NORMAL, 0);
	hrtimer_cancel(hrtimer);
	/* disable the perf event */
	watchdog_nmi_disable(cpu);
408 409
}

410 411 412 413 414
static void watchdog_cleanup(unsigned int cpu, bool online)
{
	watchdog_disable(cpu);
}

415 416 417 418 419 420 421 422 423
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.
 *
424
 * It only runs once every sample_period seconds (4 seconds by
425 426 427 428 429 430 431 432 433
 * 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();
434 435 436 437 438 439 440 441 442 443 444 445 446 447 448

	/*
	 * watchdog_nmi_enable() clears the NMI_WATCHDOG_ENABLED bit in the
	 * failure path. Check for failures that can occur asynchronously -
	 * for example, when CPUs are on-lined - and shut down the hardware
	 * perf event on each CPU accordingly.
	 *
	 * The only non-obvious place this bit can be cleared is through
	 * watchdog_nmi_enable(), so a pr_info() is placed there.  Placing a
	 * pr_info here would be too noisy as it would result in a message
	 * every few seconds if the hardlockup was disabled but the softlockup
	 * enabled.
	 */
	if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED))
		watchdog_nmi_disable(cpu);
449
}
450

451 452 453 454 455 456 457 458 459 460 461
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,
};

462 463
/*
 * park all watchdog threads that are specified in 'watchdog_cpumask'
464 465 466 467 468 469
 *
 * This function returns an error if kthread_park() of a watchdog thread
 * fails. In this situation, the watchdog threads of some CPUs can already
 * be parked and the watchdog threads of other CPUs can still be runnable.
 * Callers are expected to handle this special condition as appropriate in
 * their context.
470 471 472
 *
 * This function may only be called in a context that is protected against
 * races with CPU hotplug - for example, via get_online_cpus().
473 474 475 476 477
 */
static int watchdog_park_threads(void)
{
	int cpu, ret = 0;

478 479
	atomic_set(&watchdog_park_in_progress, 1);

480 481 482 483 484 485
	for_each_watchdog_cpu(cpu) {
		ret = kthread_park(per_cpu(softlockup_watchdog, cpu));
		if (ret)
			break;
	}

486 487
	atomic_set(&watchdog_park_in_progress, 0);

488 489 490 491 492
	return ret;
}

/*
 * unpark all watchdog threads that are specified in 'watchdog_cpumask'
493 494 495
 *
 * This function may only be called in a context that is protected against
 * races with CPU hotplug - for example, via get_online_cpus().
496 497 498 499 500 501 502 503 504
 */
static void watchdog_unpark_threads(void)
{
	int cpu;

	for_each_watchdog_cpu(cpu)
		kthread_unpark(per_cpu(softlockup_watchdog, cpu));
}

505 506 507
/*
 * Suspend the hard and soft lockup detector by parking the watchdog threads.
 */
508
int lockup_detector_suspend(void)
509 510 511
{
	int ret = 0;

512
	get_online_cpus();
513 514 515 516 517 518
	mutex_lock(&watchdog_proc_mutex);
	/*
	 * Multiple suspend requests can be active in parallel (counted by
	 * the 'watchdog_suspended' variable). If the watchdog threads are
	 * running, the first caller takes care that they will be parked.
	 * The state of 'watchdog_running' cannot change while a suspend
519
	 * request is active (see related code in 'proc' handlers).
520 521 522 523 524 525
	 */
	if (watchdog_running && !watchdog_suspended)
		ret = watchdog_park_threads();

	if (ret == 0)
		watchdog_suspended++;
526 527 528 529 530
	else {
		watchdog_disable_all_cpus();
		pr_err("Failed to suspend lockup detectors, disabled\n");
		watchdog_enabled = 0;
	}
531 532 533 534 535 536 537 538 539

	mutex_unlock(&watchdog_proc_mutex);

	return ret;
}

/*
 * Resume the hard and soft lockup detector by unparking the watchdog threads.
 */
540
void lockup_detector_resume(void)
541 542 543 544 545 546 547 548 549 550 551 552
{
	mutex_lock(&watchdog_proc_mutex);

	watchdog_suspended--;
	/*
	 * The watchdog threads are unparked if they were previously running
	 * and if there is no more active suspend request.
	 */
	if (watchdog_running && !watchdog_suspended)
		watchdog_unpark_threads();

	mutex_unlock(&watchdog_proc_mutex);
553
	put_online_cpus();
554 555
}

556
static int update_watchdog_all_cpus(void)
557
{
558 559 560 561 562 563
	int ret;

	ret = watchdog_park_threads();
	if (ret)
		return ret;

564
	watchdog_unpark_threads();
565 566

	return 0;
567 568
}

569
static int watchdog_enable_all_cpus(void)
570
{
571
	int err = 0;
572

573
	if (!watchdog_running) {
574 575
		err = smpboot_register_percpu_thread_cpumask(&watchdog_threads,
							     &watchdog_cpumask);
576 577
		if (err)
			pr_err("Failed to create watchdog threads, disabled\n");
578
		else
579
			watchdog_running = 1;
580 581 582 583 584
	} else {
		/*
		 * Enable/disable the lockup detectors or
		 * change the sample period 'on the fly'.
		 */
585 586 587 588 589 590
		err = update_watchdog_all_cpus();

		if (err) {
			watchdog_disable_all_cpus();
			pr_err("Failed to update lockup detectors, disabled\n");
		}
591
	}
592

593 594 595
	if (err)
		watchdog_enabled = 0;

596
	return err;
597 598 599 600
}

static void watchdog_disable_all_cpus(void)
{
601 602
	if (watchdog_running) {
		watchdog_running = 0;
603
		smpboot_unregister_percpu_thread(&watchdog_threads);
604
	}
605 606
}

607 608
#ifdef CONFIG_SYSCTL

609
/*
610 611 612 613 614 615 616 617 618 619 620 621 622 623
 * 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)
624
		err = watchdog_enable_all_cpus();
625 626 627 628 629 630 631
	else
		watchdog_disable_all_cpus();

	return err;

}

632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649
/*
 * common function for watchdog, nmi_watchdog and soft_watchdog parameter
 *
 * caller             | table->data points to | 'which' contains the flag(s)
 * -------------------|-----------------------|-----------------------------
 * proc_watchdog      | watchdog_user_enabled | NMI_WATCHDOG_ENABLED or'ed
 *                    |                       | with SOFT_WATCHDOG_ENABLED
 * -------------------|-----------------------|-----------------------------
 * proc_nmi_watchdog  | nmi_watchdog_enabled  | NMI_WATCHDOG_ENABLED
 * -------------------|-----------------------|-----------------------------
 * proc_soft_watchdog | soft_watchdog_enabled | SOFT_WATCHDOG_ENABLED
 */
static int proc_watchdog_common(int which, struct ctl_table *table, int write,
				void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int err, old, new;
	int *watchdog_param = (int *)table->data;

650
	get_online_cpus();
651 652
	mutex_lock(&watchdog_proc_mutex);

653 654 655 656 657 658
	if (watchdog_suspended) {
		/* no parameter changes allowed while watchdog is suspended */
		err = -EAGAIN;
		goto out;
	}

659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
	/*
	 * If the parameter is being read return the state of the corresponding
	 * bit(s) in 'watchdog_enabled', else update 'watchdog_enabled' and the
	 * run state of the lockup detectors.
	 */
	if (!write) {
		*watchdog_param = (watchdog_enabled & which) != 0;
		err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
	} else {
		err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
		if (err)
			goto out;

		/*
		 * There is a race window between fetching the current value
		 * from 'watchdog_enabled' and storing the new value. During
		 * this race window, watchdog_nmi_enable() can sneak in and
		 * clear the NMI_WATCHDOG_ENABLED bit in 'watchdog_enabled'.
		 * The 'cmpxchg' detects this race and the loop retries.
		 */
		do {
			old = watchdog_enabled;
			/*
			 * If the parameter value is not zero set the
			 * corresponding bit(s), else clear it(them).
			 */
			if (*watchdog_param)
				new = old | which;
			else
				new = old & ~which;
		} while (cmpxchg(&watchdog_enabled, old, new) != old);

		/*
692 693 694 695 696
		 * Update the run state of the lockup detectors. There is _no_
		 * need to check the value returned by proc_watchdog_update()
		 * and to restore the previous value of 'watchdog_enabled' as
		 * both lockup detectors are disabled if proc_watchdog_update()
		 * returns an error.
697
		 */
698 699 700
		if (old == new)
			goto out;

701 702 703 704
		err = proc_watchdog_update();
	}
out:
	mutex_unlock(&watchdog_proc_mutex);
705
	put_online_cpus();
706 707 708
	return err;
}

709 710 711 712 713 714 715 716 717 718 719 720
/*
 * /proc/sys/kernel/watchdog
 */
int proc_watchdog(struct ctl_table *table, int write,
		  void __user *buffer, size_t *lenp, loff_t *ppos)
{
	return proc_watchdog_common(NMI_WATCHDOG_ENABLED|SOFT_WATCHDOG_ENABLED,
				    table, write, buffer, lenp, ppos);
}

/*
 * /proc/sys/kernel/nmi_watchdog
721
 */
722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737
int proc_nmi_watchdog(struct ctl_table *table, int write,
		      void __user *buffer, size_t *lenp, loff_t *ppos)
{
	return proc_watchdog_common(NMI_WATCHDOG_ENABLED,
				    table, write, buffer, lenp, ppos);
}

/*
 * /proc/sys/kernel/soft_watchdog
 */
int proc_soft_watchdog(struct ctl_table *table, int write,
			void __user *buffer, size_t *lenp, loff_t *ppos)
{
	return proc_watchdog_common(SOFT_WATCHDOG_ENABLED,
				    table, write, buffer, lenp, ppos);
}
738

739 740 741 742 743
/*
 * /proc/sys/kernel/watchdog_thresh
 */
int proc_watchdog_thresh(struct ctl_table *table, int write,
			 void __user *buffer, size_t *lenp, loff_t *ppos)
744
{
745
	int err, old, new;
746

747
	get_online_cpus();
748
	mutex_lock(&watchdog_proc_mutex);
749

750 751 752 753 754 755
	if (watchdog_suspended) {
		/* no parameter changes allowed while watchdog is suspended */
		err = -EAGAIN;
		goto out;
	}

756
	old = ACCESS_ONCE(watchdog_thresh);
757
	err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
758

759
	if (err || !write)
760
		goto out;
761

762
	/*
763
	 * Update the sample period. Restore on failure.
764
	 */
765 766 767 768
	new = ACCESS_ONCE(watchdog_thresh);
	if (old == new)
		goto out;

769 770
	set_sample_period();
	err = proc_watchdog_update();
771
	if (err) {
772
		watchdog_thresh = old;
773 774
		set_sample_period();
	}
775 776
out:
	mutex_unlock(&watchdog_proc_mutex);
777
	put_online_cpus();
778
	return err;
779
}
780 781 782 783 784 785 786 787 788 789 790 791

/*
 * The cpumask is the mask of possible cpus that the watchdog can run
 * on, not the mask of cpus it is actually running on.  This allows the
 * user to specify a mask that will include cpus that have not yet
 * been brought online, if desired.
 */
int proc_watchdog_cpumask(struct ctl_table *table, int write,
			  void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int err;

792
	get_online_cpus();
793
	mutex_lock(&watchdog_proc_mutex);
794 795 796 797 798 799 800

	if (watchdog_suspended) {
		/* no parameter changes allowed while watchdog is suspended */
		err = -EAGAIN;
		goto out;
	}

801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817
	err = proc_do_large_bitmap(table, write, buffer, lenp, ppos);
	if (!err && write) {
		/* Remove impossible cpus to keep sysctl output cleaner. */
		cpumask_and(&watchdog_cpumask, &watchdog_cpumask,
			    cpu_possible_mask);

		if (watchdog_running) {
			/*
			 * Failure would be due to being unable to allocate
			 * a temporary cpumask, so we are likely not in a
			 * position to do much else to make things better.
			 */
			if (smpboot_update_cpumask_percpu_thread(
				    &watchdog_threads, &watchdog_cpumask) != 0)
				pr_err("cpumask update failed\n");
		}
	}
818
out:
819
	mutex_unlock(&watchdog_proc_mutex);
820
	put_online_cpus();
821 822 823
	return err;
}

824 825
#endif /* CONFIG_SYSCTL */

826
void __init lockup_detector_init(void)
827
{
828
	set_sample_period();
829

830 831
#ifdef CONFIG_NO_HZ_FULL
	if (tick_nohz_full_enabled()) {
832 833
		pr_info("Disabling watchdog on nohz_full cores by default\n");
		cpumask_copy(&watchdog_cpumask, housekeeping_mask);
834 835 836 837 838 839
	} else
		cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
#else
	cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
#endif

840
	if (watchdog_enabled)
841
		watchdog_enable_all_cpus();
842
}