watchdog.c 23 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
#define pr_fmt(fmt) "watchdog: " fmt
13

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
static DEFINE_MUTEX(watchdog_mutex);
33

34 35 36 37 38
int __read_mostly nmi_watchdog_enabled;

#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG)
unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED |
						NMI_WATCHDOG_ENABLED;
39
#else
40
unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED;
41
#endif
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57

#ifdef CONFIG_HARDLOCKUP_DETECTOR
/* boot commands */
/*
 * Should we panic when a soft-lockup or hard-lockup occurs:
 */
unsigned int __read_mostly hardlockup_panic =
			CONFIG_BOOTPARAM_HARDLOCKUP_PANIC_VALUE;
/*
 * 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().
 */
58
void __init hardlockup_detector_disable(void)
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79
{
	watchdog_enabled &= ~NMI_WATCHDOG_ENABLED;
}

static int __init hardlockup_panic_setup(char *str)
{
	if (!strncmp(str, "panic", 5))
		hardlockup_panic = 1;
	else if (!strncmp(str, "nopanic", 7))
		hardlockup_panic = 0;
	else if (!strncmp(str, "0", 1))
		watchdog_enabled &= ~NMI_WATCHDOG_ENABLED;
	else if (!strncmp(str, "1", 1))
		watchdog_enabled |= NMI_WATCHDOG_ENABLED;
	return 1;
}
__setup("nmi_watchdog=", hardlockup_panic_setup);

#endif

#ifdef CONFIG_SOFTLOCKUP_DETECTOR
80
int __read_mostly soft_watchdog_enabled;
81 82
#endif

83
int __read_mostly watchdog_user_enabled;
84
int __read_mostly watchdog_thresh = 10;
85

86 87
#ifdef CONFIG_SMP
int __read_mostly sysctl_softlockup_all_cpu_backtrace;
88
int __read_mostly sysctl_hardlockup_all_cpu_backtrace;
89
#endif
90
struct cpumask watchdog_cpumask __read_mostly;
91 92
unsigned long *watchdog_cpumask_bits = cpumask_bits(&watchdog_cpumask);

93 94 95 96 97
/*
 * 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.
 */
98
static int __read_mostly watchdog_running;
99 100 101 102

/*
 * These functions can be overridden if an architecture implements its
 * own hardlockup detector.
103 104 105 106
 *
 * watchdog_nmi_enable/disable can be implemented to start and stop when
 * softlockup watchdog threads start and stop. The arch must select the
 * SOFTLOCKUP_DETECTOR Kconfig.
107 108 109 110 111
 */
int __weak watchdog_nmi_enable(unsigned int cpu)
{
	return 0;
}
112

113 114
void __weak watchdog_nmi_disable(unsigned int cpu)
{
115
	hardlockup_detector_perf_disable();
116 117
}

118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
/*
 * watchdog_nmi_reconfigure can be implemented to be notified after any
 * watchdog configuration change. The arch hardlockup watchdog should
 * respond to the following variables:
 * - nmi_watchdog_enabled
 * - watchdog_thresh
 * - watchdog_cpumask
 * - sysctl_hardlockup_all_cpu_backtrace
 * - hardlockup_panic
 */
void __weak watchdog_nmi_reconfigure(void)
{
}


133 134 135 136 137 138
#ifdef CONFIG_SOFTLOCKUP_DETECTOR

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

139
static u64 __read_mostly sample_period;
140 141 142 143 144 145

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);
146 147
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts);
static DEFINE_PER_CPU(unsigned long, soft_lockup_hrtimer_cnt);
148
static DEFINE_PER_CPU(struct task_struct *, softlockup_task_ptr_saved);
149
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved);
150
static unsigned long soft_lockup_nmi_warn;
151 152 153 154 155 156 157 158 159 160 161 162 163 164

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)
{
165
	watchdog_enabled = 0;
166 167 168 169 170 171
	return 1;
}
__setup("nowatchdog", nowatchdog_setup);

static int __init nosoftlockup_setup(char *str)
{
172
	watchdog_enabled &= ~SOFT_WATCHDOG_ENABLED;
173 174 175
	return 1;
}
__setup("nosoftlockup", nosoftlockup_setup);
176

177 178 179 180 181 182 183 184
#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);
185
#ifdef CONFIG_HARDLOCKUP_DETECTOR
186 187 188 189 190 191 192
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);
193
#endif
194
#endif
195

196 197
static void __lockup_detector_cleanup(void);

198 199 200 201 202 203 204
/*
 * 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.
 */
205
static int get_softlockup_thresh(void)
206 207 208
{
	return watchdog_thresh * 2;
}
209 210 211 212 213 214

/*
 * 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.
 */
215
static unsigned long get_timestamp(void)
216
{
217
	return running_clock() >> 30LL;  /* 2^30 ~= 10^9 */
218 219
}

220
static void set_sample_period(void)
221 222
{
	/*
223
	 * convert watchdog_thresh from seconds to ns
224 225 226 227
	 * 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
228
	 */
229
	sample_period = get_softlockup_thresh() * ((u64)NSEC_PER_SEC / 5);
230
	watchdog_update_hrtimer_threshold(sample_period);
231 232 233 234 235
}

/* Commands for resetting the watchdog */
static void __touch_watchdog(void)
{
236
	__this_cpu_write(watchdog_touch_ts, get_timestamp());
237 238
}

239 240 241 242 243 244 245 246 247
/**
 * 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)
248
{
249 250 251 252 253
	/*
	 * 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);
254
}
255 256 257 258

void touch_softlockup_watchdog(void)
{
	touch_softlockup_watchdog_sched();
259
	wq_watchdog_touch(raw_smp_processor_id());
260
}
261
EXPORT_SYMBOL(touch_softlockup_watchdog);
262

263
void touch_all_softlockup_watchdogs(void)
264 265 266 267 268 269 270 271
{
	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
	 */
272
	for_each_watchdog_cpu(cpu)
273
		per_cpu(watchdog_touch_ts, cpu) = 0;
274
	wq_watchdog_touch(-1);
275 276 277 278
}

void touch_softlockup_watchdog_sync(void)
{
279 280
	__this_cpu_write(softlockup_touch_sync, true);
	__this_cpu_write(watchdog_touch_ts, 0);
281 282
}

283
static int is_softlockup(unsigned long touch_ts)
284
{
285
	unsigned long now = get_timestamp();
286

287
	if ((watchdog_enabled & SOFT_WATCHDOG_ENABLED) && watchdog_thresh){
288 289 290 291
		/* Warn about unreasonable delays. */
		if (time_after(now, touch_ts + get_softlockup_thresh()))
			return now - touch_ts;
	}
292 293 294
	return 0;
}

295 296
/* watchdog detector functions */
bool is_hardlockup(void)
297
{
298
	unsigned long hrint = __this_cpu_read(hrtimer_interrupts);
299

300 301 302 303 304
	if (__this_cpu_read(hrtimer_interrupts_saved) == hrint)
		return true;

	__this_cpu_write(hrtimer_interrupts_saved, hrint);
	return false;
305
}
306 307

static void watchdog_interrupt_count(void)
308
{
309
	__this_cpu_inc(hrtimer_interrupts);
310
}
311

312 313 314
static int watchdog_enable_all_cpus(void);
static void watchdog_disable_all_cpus(void);

315 316 317
/* watchdog kicker functions */
static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer)
{
318
	unsigned long touch_ts = __this_cpu_read(watchdog_touch_ts);
319 320
	struct pt_regs *regs = get_irq_regs();
	int duration;
321
	int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace;
322

323
	if (!watchdog_enabled)
324 325
		return HRTIMER_NORESTART;

326 327 328 329
	/* kick the hardlockup detector */
	watchdog_interrupt_count();

	/* kick the softlockup detector */
330
	wake_up_process(__this_cpu_read(softlockup_watchdog));
331 332

	/* .. and repeat */
333
	hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period));
334 335

	if (touch_ts == 0) {
336
		if (unlikely(__this_cpu_read(softlockup_touch_sync))) {
337 338 339 340
			/*
			 * If the time stamp was touched atomically
			 * make sure the scheduler tick is up to date.
			 */
341
			__this_cpu_write(softlockup_touch_sync, false);
342 343
			sched_clock_tick();
		}
344 345 346

		/* Clear the guest paused flag on watchdog reset */
		kvm_check_and_clear_guest_paused();
347 348 349 350 351 352 353 354 355 356
		__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
	 */
357
	duration = is_softlockup(touch_ts);
358
	if (unlikely(duration)) {
359 360 361 362 363 364 365 366
		/*
		 * 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;

367
		/* only warn once */
368 369 370 371 372 373 374 375 376 377 378 379 380 381
		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();
			}
382
			return HRTIMER_RESTART;
383
		}
384

385 386 387 388 389 390 391 392 393 394 395
		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;
			}
		}

396
		pr_emerg("BUG: soft lockup - CPU#%d stuck for %us! [%s:%d]\n",
397
			smp_processor_id(), duration,
398
			current->comm, task_pid_nr(current));
399
		__this_cpu_write(softlockup_task_ptr_saved, current);
400 401 402 403 404 405 406
		print_modules();
		print_irqtrace_events(current);
		if (regs)
			show_regs(regs);
		else
			dump_stack();

407 408 409 410 411 412 413 414 415 416 417
		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
418
		add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK);
419 420
		if (softlockup_panic)
			panic("softlockup: hung tasks");
421
		__this_cpu_write(soft_watchdog_warn, true);
422
	} else
423
		__this_cpu_write(soft_watchdog_warn, false);
424 425 426 427

	return HRTIMER_RESTART;
}

428 429 430
static void watchdog_set_prio(unsigned int policy, unsigned int prio)
{
	struct sched_param param = { .sched_priority = prio };
431

432 433 434 435
	sched_setscheduler(current, policy, &param);
}

static void watchdog_enable(unsigned int cpu)
436
{
437
	struct hrtimer *hrtimer = this_cpu_ptr(&watchdog_hrtimer);
438

439 440 441 442
	/*
	 * Start the timer first to prevent the NMI watchdog triggering
	 * before the timer has a chance to fire.
	 */
443 444
	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
	hrtimer->function = watchdog_timer_fn;
445 446
	hrtimer_start(hrtimer, ns_to_ktime(sample_period),
		      HRTIMER_MODE_REL_PINNED);
447

448 449
	/* Initialize timestamp */
	__touch_watchdog();
450 451
	/* Enable the perf event */
	watchdog_nmi_enable(cpu);
452

453 454
	watchdog_set_prio(SCHED_FIFO, MAX_RT_PRIO - 1);
}
455

456 457
static void watchdog_disable(unsigned int cpu)
{
458
	struct hrtimer *hrtimer = this_cpu_ptr(&watchdog_hrtimer);
459

460
	watchdog_set_prio(SCHED_NORMAL, 0);
461 462 463 464 465
	/*
	 * Disable the perf event first. That prevents that a large delay
	 * between disabling the timer and disabling the perf event causes
	 * the perf NMI to detect a false positive.
	 */
466
	watchdog_nmi_disable(cpu);
467
	hrtimer_cancel(hrtimer);
468 469
}

470 471 472 473 474
static void watchdog_cleanup(unsigned int cpu, bool online)
{
	watchdog_disable(cpu);
}

475 476 477 478 479 480 481 482 483
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.
 *
484
 * It only runs once every sample_period seconds (4 seconds by
485 486 487 488 489 490 491 492 493 494
 * 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();
}
495

496 497 498 499 500 501 502 503 504 505 506
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,
};

507 508
/*
 * park all watchdog threads that are specified in 'watchdog_cpumask'
509 510 511 512 513 514
 *
 * 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.
515 516 517
 *
 * This function may only be called in a context that is protected against
 * races with CPU hotplug - for example, via get_online_cpus().
518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
 */
static int watchdog_park_threads(void)
{
	int cpu, ret = 0;

	for_each_watchdog_cpu(cpu) {
		ret = kthread_park(per_cpu(softlockup_watchdog, cpu));
		if (ret)
			break;
	}
	return ret;
}

/*
 * unpark all watchdog threads that are specified in 'watchdog_cpumask'
533 534 535
 *
 * This function may only be called in a context that is protected against
 * races with CPU hotplug - for example, via get_online_cpus().
536 537 538 539 540 541 542 543 544
 */
static void watchdog_unpark_threads(void)
{
	int cpu;

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

545
static int update_watchdog_all_cpus(void)
546
{
547 548 549 550 551 552
	int ret;

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

553
	watchdog_unpark_threads();
554 555

	return 0;
556 557
}

558
static int watchdog_enable_all_cpus(void)
559
{
560
	int err = 0;
561

562
	if (!watchdog_running) {
563 564
		err = smpboot_register_percpu_thread_cpumask(&watchdog_threads,
							     &watchdog_cpumask);
565 566
		if (err)
			pr_err("Failed to create watchdog threads, disabled\n");
567
		else
568
			watchdog_running = 1;
569 570 571 572 573
	} else {
		/*
		 * Enable/disable the lockup detectors or
		 * change the sample period 'on the fly'.
		 */
574 575 576 577 578 579
		err = update_watchdog_all_cpus();

		if (err) {
			watchdog_disable_all_cpus();
			pr_err("Failed to update lockup detectors, disabled\n");
		}
580
	}
581

582 583 584
	if (err)
		watchdog_enabled = 0;

585
	return err;
586 587 588 589
}

static void watchdog_disable_all_cpus(void)
{
590 591
	if (watchdog_running) {
		watchdog_running = 0;
592
		smpboot_unregister_percpu_thread(&watchdog_threads);
593
	}
594 595
}

596 597 598 599 600 601 602 603
#ifdef CONFIG_SYSCTL
static int watchdog_update_cpus(void)
{
	return smpboot_update_cpumask_percpu_thread(
		    &watchdog_threads, &watchdog_cpumask);
}
#endif

604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622
#else /* SOFTLOCKUP */
static int watchdog_park_threads(void)
{
	return 0;
}

static void watchdog_unpark_threads(void)
{
}

static int watchdog_enable_all_cpus(void)
{
	return 0;
}

static void watchdog_disable_all_cpus(void)
{
}

623 624 625 626 627 628 629
#ifdef CONFIG_SYSCTL
static int watchdog_update_cpus(void)
{
	return 0;
}
#endif

630 631 632 633 634
static void set_sample_period(void)
{
}
#endif /* SOFTLOCKUP */

635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652
static void __lockup_detector_cleanup(void)
{
	lockdep_assert_held(&watchdog_mutex);
	hardlockup_detector_perf_cleanup();
}

/**
 * lockup_detector_cleanup - Cleanup after cpu hotplug or sysctl changes
 *
 * Caller must not hold the cpu hotplug rwsem.
 */
void lockup_detector_cleanup(void)
{
	mutex_lock(&watchdog_mutex);
	__lockup_detector_cleanup();
	mutex_unlock(&watchdog_mutex);
}

653 654 655 656 657 658 659 660 661 662 663
/**
 * lockup_detector_soft_poweroff - Interface to stop lockup detector(s)
 *
 * Special interface for parisc. It prevents lockup detector warnings from
 * the default pm_poweroff() function which busy loops forever.
 */
void lockup_detector_soft_poweroff(void)
{
	watchdog_enabled = 0;
}

664 665
#ifdef CONFIG_SYSCTL

666
/*
667 668 669 670 671 672 673 674 675 676 677 678 679 680
 * 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)
681
		err = watchdog_enable_all_cpus();
682 683 684
	else
		watchdog_disable_all_cpus();

685 686
	watchdog_nmi_reconfigure();

687 688
	__lockup_detector_cleanup();

689 690 691 692
	return err;

}

693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710
/*
 * 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;

711
	cpu_hotplug_disable();
712
	mutex_lock(&watchdog_mutex);
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746

	/*
	 * 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);

		/*
747 748 749 750 751
		 * 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.
752
		 */
753 754 755
		if (old == new)
			goto out;

756 757 758
		err = proc_watchdog_update();
	}
out:
759
	mutex_unlock(&watchdog_mutex);
760
	cpu_hotplug_enable();
761 762 763
	return err;
}

764 765 766 767 768 769 770 771 772 773 774 775
/*
 * /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
776
 */
777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
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);
}
793

794 795 796 797 798
/*
 * /proc/sys/kernel/watchdog_thresh
 */
int proc_watchdog_thresh(struct ctl_table *table, int write,
			 void __user *buffer, size_t *lenp, loff_t *ppos)
799
{
800
	int err, old, new;
801

802
	cpu_hotplug_disable();
803
	mutex_lock(&watchdog_mutex);
804

805
	old = ACCESS_ONCE(watchdog_thresh);
806
	err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
807

808
	if (err || !write)
809
		goto out;
810

811
	/*
812
	 * Update the sample period. Restore on failure.
813
	 */
814 815 816 817
	new = ACCESS_ONCE(watchdog_thresh);
	if (old == new)
		goto out;

818 819
	set_sample_period();
	err = proc_watchdog_update();
820
	if (err) {
821
		watchdog_thresh = old;
822 823
		set_sample_period();
	}
824
out:
825
	mutex_unlock(&watchdog_mutex);
826
	cpu_hotplug_enable();
827
	return err;
828
}
829 830 831 832 833 834 835 836 837 838 839 840

/*
 * 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;

841
	cpu_hotplug_disable();
842
	mutex_lock(&watchdog_mutex);
843

844 845 846 847 848 849 850 851 852 853 854 855
	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.
			 */
856
			if (watchdog_update_cpus() != 0)
857 858
				pr_err("cpumask update failed\n");
		}
859 860

		watchdog_nmi_reconfigure();
861
		__lockup_detector_cleanup();
862
	}
863

864
	mutex_unlock(&watchdog_mutex);
865
	cpu_hotplug_enable();
866 867 868
	return err;
}

869 870
#endif /* CONFIG_SYSCTL */

871
void __init lockup_detector_init(void)
872
{
873
	set_sample_period();
874

875 876
#ifdef CONFIG_NO_HZ_FULL
	if (tick_nohz_full_enabled()) {
877 878
		pr_info("Disabling watchdog on nohz_full cores by default\n");
		cpumask_copy(&watchdog_cpumask, housekeeping_mask);
879 880 881 882 883 884
	} else
		cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
#else
	cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
#endif

885
	if (watchdog_enabled)
886
		watchdog_enable_all_cpus();
887
}