watchdog.c 23.1 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
/*
 * 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
 */
128
void __weak watchdog_nmi_reconfigure(void) { }
129

130 131 132 133 134 135
#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)

136 137 138 139 140
/* Global variables, exported for sysctl */
unsigned int __read_mostly softlockup_panic =
			CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE;
int __read_mostly soft_watchdog_enabled;

141
static u64 __read_mostly sample_period;
142 143 144 145 146 147

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

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)
{
163
	watchdog_enabled = 0;
164 165 166 167 168 169
	return 1;
}
__setup("nowatchdog", nowatchdog_setup);

static int __init nosoftlockup_setup(char *str)
{
170
	watchdog_enabled &= ~SOFT_WATCHDOG_ENABLED;
171 172 173
	return 1;
}
__setup("nosoftlockup", nosoftlockup_setup);
174

175 176 177 178 179 180 181 182
#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);
183
#ifdef CONFIG_HARDLOCKUP_DETECTOR
184 185 186 187 188 189 190
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);
191
#endif
192
#endif
193

194 195
static void __lockup_detector_cleanup(void);

196 197 198 199 200 201 202
/*
 * 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.
 */
203
static int get_softlockup_thresh(void)
204 205 206
{
	return watchdog_thresh * 2;
}
207 208 209 210 211 212

/*
 * 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.
 */
213
static unsigned long get_timestamp(void)
214
{
215
	return running_clock() >> 30LL;  /* 2^30 ~= 10^9 */
216 217
}

218
static void set_sample_period(void)
219 220
{
	/*
221
	 * convert watchdog_thresh from seconds to ns
222 223 224 225
	 * 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
226
	 */
227
	sample_period = get_softlockup_thresh() * ((u64)NSEC_PER_SEC / 5);
228
	watchdog_update_hrtimer_threshold(sample_period);
229 230 231 232 233
}

/* Commands for resetting the watchdog */
static void __touch_watchdog(void)
{
234
	__this_cpu_write(watchdog_touch_ts, get_timestamp());
235 236
}

237 238 239 240 241 242 243 244 245
/**
 * 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)
246
{
247 248 249 250 251
	/*
	 * 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);
252
}
253 254 255 256

void touch_softlockup_watchdog(void)
{
	touch_softlockup_watchdog_sched();
257
	wq_watchdog_touch(raw_smp_processor_id());
258
}
259
EXPORT_SYMBOL(touch_softlockup_watchdog);
260

261
void touch_all_softlockup_watchdogs(void)
262 263 264 265 266 267 268 269
{
	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
	 */
270
	for_each_watchdog_cpu(cpu)
271
		per_cpu(watchdog_touch_ts, cpu) = 0;
272
	wq_watchdog_touch(-1);
273 274 275 276
}

void touch_softlockup_watchdog_sync(void)
{
277 278
	__this_cpu_write(softlockup_touch_sync, true);
	__this_cpu_write(watchdog_touch_ts, 0);
279 280
}

281
static int is_softlockup(unsigned long touch_ts)
282
{
283
	unsigned long now = get_timestamp();
284

285
	if ((watchdog_enabled & SOFT_WATCHDOG_ENABLED) && watchdog_thresh){
286 287 288 289
		/* Warn about unreasonable delays. */
		if (time_after(now, touch_ts + get_softlockup_thresh()))
			return now - touch_ts;
	}
290 291 292
	return 0;
}

293 294
/* watchdog detector functions */
bool is_hardlockup(void)
295
{
296
	unsigned long hrint = __this_cpu_read(hrtimer_interrupts);
297

298 299 300 301 302
	if (__this_cpu_read(hrtimer_interrupts_saved) == hrint)
		return true;

	__this_cpu_write(hrtimer_interrupts_saved, hrint);
	return false;
303
}
304 305

static void watchdog_interrupt_count(void)
306
{
307
	__this_cpu_inc(hrtimer_interrupts);
308
}
309

310 311 312
static int watchdog_enable_all_cpus(void);
static void watchdog_disable_all_cpus(void);

313 314 315
/* watchdog kicker functions */
static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer)
{
316
	unsigned long touch_ts = __this_cpu_read(watchdog_touch_ts);
317 318
	struct pt_regs *regs = get_irq_regs();
	int duration;
319
	int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace;
320

321
	if (!watchdog_enabled)
322 323
		return HRTIMER_NORESTART;

324 325 326 327
	/* kick the hardlockup detector */
	watchdog_interrupt_count();

	/* kick the softlockup detector */
328
	wake_up_process(__this_cpu_read(softlockup_watchdog));
329 330

	/* .. and repeat */
331
	hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period));
332 333

	if (touch_ts == 0) {
334
		if (unlikely(__this_cpu_read(softlockup_touch_sync))) {
335 336 337 338
			/*
			 * If the time stamp was touched atomically
			 * make sure the scheduler tick is up to date.
			 */
339
			__this_cpu_write(softlockup_touch_sync, false);
340 341
			sched_clock_tick();
		}
342 343 344

		/* Clear the guest paused flag on watchdog reset */
		kvm_check_and_clear_guest_paused();
345 346 347 348 349 350 351 352 353 354
		__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
	 */
355
	duration = is_softlockup(touch_ts);
356
	if (unlikely(duration)) {
357 358 359 360 361 362 363 364
		/*
		 * 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;

365
		/* only warn once */
366 367 368 369 370 371 372 373 374 375 376 377 378 379
		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();
			}
380
			return HRTIMER_RESTART;
381
		}
382

383 384 385 386 387 388 389 390 391 392 393
		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;
			}
		}

394
		pr_emerg("BUG: soft lockup - CPU#%d stuck for %us! [%s:%d]\n",
395
			smp_processor_id(), duration,
396
			current->comm, task_pid_nr(current));
397
		__this_cpu_write(softlockup_task_ptr_saved, current);
398 399 400 401 402 403 404
		print_modules();
		print_irqtrace_events(current);
		if (regs)
			show_regs(regs);
		else
			dump_stack();

405 406 407 408 409 410 411 412 413 414 415
		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
416
		add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK);
417 418
		if (softlockup_panic)
			panic("softlockup: hung tasks");
419
		__this_cpu_write(soft_watchdog_warn, true);
420
	} else
421
		__this_cpu_write(soft_watchdog_warn, false);
422 423 424 425

	return HRTIMER_RESTART;
}

426 427 428
static void watchdog_set_prio(unsigned int policy, unsigned int prio)
{
	struct sched_param param = { .sched_priority = prio };
429

430 431 432 433
	sched_setscheduler(current, policy, &param);
}

static void watchdog_enable(unsigned int cpu)
434
{
435
	struct hrtimer *hrtimer = this_cpu_ptr(&watchdog_hrtimer);
436

437 438 439 440
	/*
	 * Start the timer first to prevent the NMI watchdog triggering
	 * before the timer has a chance to fire.
	 */
441 442
	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
	hrtimer->function = watchdog_timer_fn;
443 444
	hrtimer_start(hrtimer, ns_to_ktime(sample_period),
		      HRTIMER_MODE_REL_PINNED);
445

446 447
	/* Initialize timestamp */
	__touch_watchdog();
448 449
	/* Enable the perf event */
	watchdog_nmi_enable(cpu);
450

451 452
	watchdog_set_prio(SCHED_FIFO, MAX_RT_PRIO - 1);
}
453

454 455
static void watchdog_disable(unsigned int cpu)
{
456
	struct hrtimer *hrtimer = this_cpu_ptr(&watchdog_hrtimer);
457

458
	watchdog_set_prio(SCHED_NORMAL, 0);
459 460 461 462 463
	/*
	 * 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.
	 */
464
	watchdog_nmi_disable(cpu);
465
	hrtimer_cancel(hrtimer);
466 467
}

468 469 470 471 472
static void watchdog_cleanup(unsigned int cpu, bool online)
{
	watchdog_disable(cpu);
}

473 474 475 476 477 478 479 480 481
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.
 *
482
 * It only runs once every sample_period seconds (4 seconds by
483 484 485 486 487 488 489 490 491 492
 * 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();
}
493

494 495 496 497 498 499 500 501 502 503 504
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,
};

505 506
/*
 * park all watchdog threads that are specified in 'watchdog_cpumask'
507 508 509 510 511 512
 *
 * 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.
513 514 515
 *
 * This function may only be called in a context that is protected against
 * races with CPU hotplug - for example, via get_online_cpus().
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530
 */
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'
531 532 533
 *
 * This function may only be called in a context that is protected against
 * races with CPU hotplug - for example, via get_online_cpus().
534 535 536 537 538 539 540 541 542
 */
static void watchdog_unpark_threads(void)
{
	int cpu;

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

543
static int update_watchdog_all_cpus(void)
544
{
545 546 547 548 549 550
	int ret;

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

551
	watchdog_unpark_threads();
552 553

	return 0;
554 555
}

556
static int watchdog_enable_all_cpus(void)
557
{
558
	int err = 0;
559

560
	if (!watchdog_running) {
561 562
		err = smpboot_register_percpu_thread_cpumask(&watchdog_threads,
							     &watchdog_cpumask);
563 564
		if (err)
			pr_err("Failed to create watchdog threads, disabled\n");
565
		else
566
			watchdog_running = 1;
567 568 569 570 571
	} else {
		/*
		 * Enable/disable the lockup detectors or
		 * change the sample period 'on the fly'.
		 */
572 573 574 575 576 577
		err = update_watchdog_all_cpus();

		if (err) {
			watchdog_disable_all_cpus();
			pr_err("Failed to update lockup detectors, disabled\n");
		}
578
	}
579

580 581 582
	if (err)
		watchdog_enabled = 0;

583
	return err;
584 585 586 587
}

static void watchdog_disable_all_cpus(void)
{
588 589
	if (watchdog_running) {
		watchdog_running = 0;
590
		smpboot_unregister_percpu_thread(&watchdog_threads);
591
	}
592 593
}

594 595 596 597 598 599 600
#else /* CONFIG_SOFTLOCKUP_DETECTOR */
static inline int watchdog_park_threads(void) { return 0; }
static inline void watchdog_unpark_threads(void) { }
static inline int watchdog_enable_all_cpus(void) { return 0; }
static inline void watchdog_disable_all_cpus(void) { }
static inline void set_sample_period(void) { }
#endif /* !CONFIG_SOFTLOCKUP_DETECTOR */
601

602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619
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);
}

620 621 622 623 624 625 626 627 628 629 630
/**
 * 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;
}

631 632
#ifdef CONFIG_SYSCTL

633
/*
634 635 636 637 638 639 640 641 642 643 644 645 646 647
 * 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)
648
		err = watchdog_enable_all_cpus();
649 650 651
	else
		watchdog_disable_all_cpus();

652 653
	watchdog_nmi_reconfigure();

654 655
	__lockup_detector_cleanup();

656 657 658 659
	return err;

}

660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677
/*
 * 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;

678
	cpu_hotplug_disable();
679
	mutex_lock(&watchdog_mutex);
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713

	/*
	 * 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);

		/*
714 715 716 717 718
		 * 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.
719
		 */
720 721 722
		if (old == new)
			goto out;

723 724 725
		err = proc_watchdog_update();
	}
out:
726
	mutex_unlock(&watchdog_mutex);
727
	cpu_hotplug_enable();
728 729 730
	return err;
}

731 732 733 734 735 736 737 738 739 740 741 742
/*
 * /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
743
 */
744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759
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);
}
760

761 762 763 764 765
/*
 * /proc/sys/kernel/watchdog_thresh
 */
int proc_watchdog_thresh(struct ctl_table *table, int write,
			 void __user *buffer, size_t *lenp, loff_t *ppos)
766
{
767
	int err, old, new;
768

769
	cpu_hotplug_disable();
770
	mutex_lock(&watchdog_mutex);
771

772
	old = ACCESS_ONCE(watchdog_thresh);
773
	err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
774

775
	if (err || !write)
776
		goto out;
777

778
	/*
779
	 * Update the sample period. Restore on failure.
780
	 */
781 782 783 784
	new = ACCESS_ONCE(watchdog_thresh);
	if (old == new)
		goto out;

785 786
	set_sample_period();
	err = proc_watchdog_update();
787
	if (err) {
788
		watchdog_thresh = old;
789 790
		set_sample_period();
	}
791
out:
792
	mutex_unlock(&watchdog_mutex);
793
	cpu_hotplug_enable();
794
	return err;
795
}
796

797 798 799 800 801 802 803 804 805
static int watchdog_update_cpus(void)
{
	if (IS_ENABLED(CONFIG_SOFTLOCKUP_DETECTOR)) {
		return smpboot_update_cpumask_percpu_thread(&watchdog_threads,
							    &watchdog_cpumask);
	}
	return 0;
}

806 807 808 809 810 811 812 813 814 815 816
/*
 * 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;

817
	cpu_hotplug_disable();
818
	mutex_lock(&watchdog_mutex);
819

820 821 822 823 824 825 826 827 828 829 830 831
	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.
			 */
832
			if (watchdog_update_cpus() != 0)
833 834
				pr_err("cpumask update failed\n");
		}
835 836

		watchdog_nmi_reconfigure();
837
		__lockup_detector_cleanup();
838
	}
839

840
	mutex_unlock(&watchdog_mutex);
841
	cpu_hotplug_enable();
842 843 844
	return err;
}

845 846
#endif /* CONFIG_SYSCTL */

847
void __init lockup_detector_init(void)
848
{
849
	set_sample_period();
850

851 852
#ifdef CONFIG_NO_HZ_FULL
	if (tick_nohz_full_enabled()) {
853 854
		pr_info("Disabling watchdog on nohz_full cores by default\n");
		cpumask_copy(&watchdog_cpumask, housekeeping_mask);
855 856 857 858 859 860
	} else
		cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
#else
	cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
#endif

861
	if (watchdog_enabled)
862
		watchdog_enable_all_cpus();
863
}