[linux-2.6-block.git] / kernel / hung_task.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 * Detect Hung Task
 *
 * kernel/hung_task.c - kernel thread for detecting tasks stuck in D state
 *
 */

#include <linux/mm.h>
#include <linux/cpu.h>
#include <linux/nmi.h>
#include <linux/init.h>
#include <linux/delay.h>
#include <linux/freezer.h>
#include <linux/kthread.h>
#include <linux/lockdep.h>
#include <linux/export.h>
#include <linux/panic_notifier.h>
#include <linux/sysctl.h>
#include <linux/suspend.h>
#include <linux/utsname.h>
#include <linux/sched/signal.h>
#include <linux/sched/debug.h>
#include <linux/sched/sysctl.h>
#include <linux/hung_task.h>

#include <trace/events/sched.h>

/*
 * The number of tasks checked:
 */
static int __read_mostly sysctl_hung_task_check_count = PID_MAX_LIMIT;

/*
 * Total number of tasks detected as hung since boot:
 */
static unsigned long __read_mostly sysctl_hung_task_detect_count;

/*
 * Limit number of tasks checked in a batch.
 *
 * This value controls the preemptibility of khungtaskd since preemption
 * is disabled during the critical section. It also controls the size of
 * the RCU grace period. So it needs to be upper-bound.
 */
#define HUNG_TASK_LOCK_BREAK (HZ / 10)

/*
 * Zero means infinite timeout - no checking done:
 */
unsigned long __read_mostly sysctl_hung_task_timeout_secs = CONFIG_DEFAULT_HUNG_TASK_TIMEOUT;
EXPORT_SYMBOL_GPL(sysctl_hung_task_timeout_secs);

/*
 * Zero (default value) means use sysctl_hung_task_timeout_secs:
 */
static unsigned long __read_mostly sysctl_hung_task_check_interval_secs;

static int __read_mostly sysctl_hung_task_warnings = 10;

static int __read_mostly did_panic;
static bool hung_task_show_lock;
static bool hung_task_call_panic;
static bool hung_task_show_all_bt;

static struct task_struct *watchdog_task;

#ifdef CONFIG_SMP
/*
 * Should we dump all CPUs backtraces in a hung task event?
 * Defaults to 0, can be changed via sysctl.
 */
static unsigned int __read_mostly sysctl_hung_task_all_cpu_backtrace;
#else
#define sysctl_hung_task_all_cpu_backtrace 0
#endif /* CONFIG_SMP */

/*
 * Should we panic (and reboot, if panic_timeout= is set) when a
 * hung task is detected:
 */
static unsigned int __read_mostly sysctl_hung_task_panic =
	IS_ENABLED(CONFIG_BOOTPARAM_HUNG_TASK_PANIC);

static int
hung_task_panic(struct notifier_block *this, unsigned long event, void *ptr)
{
	did_panic = 1;

	return NOTIFY_DONE;
}

static struct notifier_block panic_block = {
	.notifier_call = hung_task_panic,
};


#ifdef CONFIG_DETECT_HUNG_TASK_BLOCKER
static void debug_show_blocker(struct task_struct *task)
{
	struct task_struct *g, *t;
	unsigned long owner, blocker, blocker_type;

	RCU_LOCKDEP_WARN(!rcu_read_lock_held(), "No rcu lock held");

	blocker = READ_ONCE(task->blocker);
	if (!blocker)
		return;

	blocker_type = hung_task_get_blocker_type(blocker);

	switch (blocker_type) {
	case BLOCKER_TYPE_MUTEX:
		owner = mutex_get_owner(
			(struct mutex *)hung_task_blocker_to_lock(blocker));
		break;
	case BLOCKER_TYPE_SEM:
		owner = sem_last_holder(
			(struct semaphore *)hung_task_blocker_to_lock(blocker));
		break;
	default:
		WARN_ON_ONCE(1);
		return;
	}


	if (unlikely(!owner)) {
		switch (blocker_type) {
		case BLOCKER_TYPE_MUTEX:
			pr_err("INFO: task %s:%d is blocked on a mutex, but the owner is not found.\n",
			       task->comm, task->pid);
			break;
		case BLOCKER_TYPE_SEM:
			pr_err("INFO: task %s:%d is blocked on a semaphore, but the last holder is not found.\n",
			       task->comm, task->pid);
			break;
		}
		return;
	}

	/* Ensure the owner information is correct. */
	for_each_process_thread(g, t) {
		if ((unsigned long)t != owner)
			continue;

		switch (blocker_type) {
		case BLOCKER_TYPE_MUTEX:
			pr_err("INFO: task %s:%d is blocked on a mutex likely owned by task %s:%d.\n",
			       task->comm, task->pid, t->comm, t->pid);
			break;
		case BLOCKER_TYPE_SEM:
			pr_err("INFO: task %s:%d blocked on a semaphore likely last held by task %s:%d\n",
			       task->comm, task->pid, t->comm, t->pid);
			break;
		}
		sched_show_task(t);
		return;
	}
}
#else
static inline void debug_show_blocker(struct task_struct *task)
{
}
#endif

static void check_hung_task(struct task_struct *t, unsigned long timeout)
{
	unsigned long switch_count = t->nvcsw + t->nivcsw;

	/*
	 * Ensure the task is not frozen.
	 * Also, skip vfork and any other user process that freezer should skip.
	 */
	if (unlikely(READ_ONCE(t->__state) & TASK_FROZEN))
		return;

	/*
	 * When a freshly created task is scheduled once, changes its state to
	 * TASK_UNINTERRUPTIBLE without having ever been switched out once, it
	 * musn't be checked.
	 */
	if (unlikely(!switch_count))
		return;

	if (switch_count != t->last_switch_count) {
		t->last_switch_count = switch_count;
		t->last_switch_time = jiffies;
		return;
	}
	if (time_is_after_jiffies(t->last_switch_time + timeout * HZ))
		return;

	/*
	 * This counter tracks the total number of tasks detected as hung
	 * since boot.
	 */
	sysctl_hung_task_detect_count++;

	trace_sched_process_hang(t);

	if (sysctl_hung_task_panic) {
		console_verbose();
		hung_task_show_lock = true;
		hung_task_call_panic = true;
	}

	/*
	 * Ok, the task did not get scheduled for more than 2 minutes,
	 * complain:
	 */
	if (sysctl_hung_task_warnings || hung_task_call_panic) {
		if (sysctl_hung_task_warnings > 0)
			sysctl_hung_task_warnings--;
		pr_err("INFO: task %s:%d blocked for more than %ld seconds.\n",
		       t->comm, t->pid, (jiffies - t->last_switch_time) / HZ);
		pr_err("      %s %s %.*s\n",
			print_tainted(), init_utsname()->release,
			(int)strcspn(init_utsname()->version, " "),
			init_utsname()->version);
		if (t->flags & PF_POSTCOREDUMP)
			pr_err("      Blocked by coredump.\n");
		pr_err("\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
			" disables this message.\n");
		sched_show_task(t);
		debug_show_blocker(t);
		hung_task_show_lock = true;

		if (sysctl_hung_task_all_cpu_backtrace)
			hung_task_show_all_bt = true;
		if (!sysctl_hung_task_warnings)
			pr_info("Future hung task reports are suppressed, see sysctl kernel.hung_task_warnings\n");
	}

	touch_nmi_watchdog();
}

/*
 * To avoid extending the RCU grace period for an unbounded amount of time,
 * periodically exit the critical section and enter a new one.
 *
 * For preemptible RCU it is sufficient to call rcu_read_unlock in order
 * to exit the grace period. For classic RCU, a reschedule is required.
 */
static bool rcu_lock_break(struct task_struct *g, struct task_struct *t)
{
	bool can_cont;

	get_task_struct(g);
	get_task_struct(t);
	rcu_read_unlock();
	cond_resched();
	rcu_read_lock();
	can_cont = pid_alive(g) && pid_alive(t);
	put_task_struct(t);
	put_task_struct(g);

	return can_cont;
}

/*
 * Check whether a TASK_UNINTERRUPTIBLE does not get woken up for
 * a really long time (120 seconds). If that happens, print out
 * a warning.
 */
static void check_hung_uninterruptible_tasks(unsigned long timeout)
{
	int max_count = sysctl_hung_task_check_count;
	unsigned long last_break = jiffies;
	struct task_struct *g, *t;

	/*
	 * If the system crashed already then all bets are off,
	 * do not report extra hung tasks:
	 */
	if (test_taint(TAINT_DIE) || did_panic)
		return;

	hung_task_show_lock = false;
	rcu_read_lock();
	for_each_process_thread(g, t) {
		unsigned int state;

		if (!max_count--)
			goto unlock;
		if (time_after(jiffies, last_break + HUNG_TASK_LOCK_BREAK)) {
			if (!rcu_lock_break(g, t))
				goto unlock;
			last_break = jiffies;
		}
		/*
		 * skip the TASK_KILLABLE tasks -- these can be killed
		 * skip the TASK_IDLE tasks -- those are genuinely idle
		 */
		state = READ_ONCE(t->__state);
		if ((state & TASK_UNINTERRUPTIBLE) &&
		    !(state & TASK_WAKEKILL) &&
		    !(state & TASK_NOLOAD))
			check_hung_task(t, timeout);
	}
 unlock:
	rcu_read_unlock();
	if (hung_task_show_lock)
		debug_show_all_locks();

	if (hung_task_show_all_bt) {
		hung_task_show_all_bt = false;
		trigger_all_cpu_backtrace();
	}

	if (hung_task_call_panic)
		panic("hung_task: blocked tasks");
}

static long hung_timeout_jiffies(unsigned long last_checked,
				 unsigned long timeout)
{
	/* timeout of 0 will disable the watchdog */
	return timeout ? last_checked - jiffies + timeout * HZ :
		MAX_SCHEDULE_TIMEOUT;
}

#ifdef CONFIG_SYSCTL
/*
 * Process updating of timeout sysctl
 */
static int proc_dohung_task_timeout_secs(const struct ctl_table *table, int write,
				  void *buffer,
				  size_t *lenp, loff_t *ppos)
{
	int ret;

	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);

	if (ret || !write)
		goto out;

	wake_up_process(watchdog_task);

 out:
	return ret;
}

/*
 * This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
 * and hung_task_check_interval_secs
 */
static const unsigned long hung_task_timeout_max = (LONG_MAX / HZ);
static const struct ctl_table hung_task_sysctls[] = {
#ifdef CONFIG_SMP
	{
		.procname	= "hung_task_all_cpu_backtrace",
		.data		= &sysctl_hung_task_all_cpu_backtrace,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
		.extra2		= SYSCTL_ONE,
	},
#endif /* CONFIG_SMP */
	{
		.procname	= "hung_task_panic",
		.data		= &sysctl_hung_task_panic,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
		.extra2		= SYSCTL_ONE,
	},
	{
		.procname	= "hung_task_check_count",
		.data		= &sysctl_hung_task_check_count,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
	},
	{
		.procname	= "hung_task_timeout_secs",
		.data		= &sysctl_hung_task_timeout_secs,
		.maxlen		= sizeof(unsigned long),
		.mode		= 0644,
		.proc_handler	= proc_dohung_task_timeout_secs,
		.extra2		= (void *)&hung_task_timeout_max,
	},
	{
		.procname	= "hung_task_check_interval_secs",
		.data		= &sysctl_hung_task_check_interval_secs,
		.maxlen		= sizeof(unsigned long),
		.mode		= 0644,
		.proc_handler	= proc_dohung_task_timeout_secs,
		.extra2		= (void *)&hung_task_timeout_max,
	},
	{
		.procname	= "hung_task_warnings",
		.data		= &sysctl_hung_task_warnings,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_NEG_ONE,
	},
	{
		.procname	= "hung_task_detect_count",
		.data		= &sysctl_hung_task_detect_count,
		.maxlen		= sizeof(unsigned long),
		.mode		= 0444,
		.proc_handler	= proc_doulongvec_minmax,
	},
};

static void __init hung_task_sysctl_init(void)
{
	register_sysctl_init("kernel", hung_task_sysctls);
}
#else
#define hung_task_sysctl_init() do { } while (0)
#endif /* CONFIG_SYSCTL */


static atomic_t reset_hung_task = ATOMIC_INIT(0);

void reset_hung_task_detector(void)
{
	atomic_set(&reset_hung_task, 1);
}
EXPORT_SYMBOL_GPL(reset_hung_task_detector);

static bool hung_detector_suspended;

static int hungtask_pm_notify(struct notifier_block *self,
			      unsigned long action, void *hcpu)
{
	switch (action) {
	case PM_SUSPEND_PREPARE:
	case PM_HIBERNATION_PREPARE:
	case PM_RESTORE_PREPARE:
		hung_detector_suspended = true;
		break;
	case PM_POST_SUSPEND:
	case PM_POST_HIBERNATION:
	case PM_POST_RESTORE:
		hung_detector_suspended = false;
		break;
	default:
		break;
	}
	return NOTIFY_OK;
}

/*
 * kthread which checks for tasks stuck in D state
 */
static int watchdog(void *dummy)
{
	unsigned long hung_last_checked = jiffies;

	set_user_nice(current, 0);

	for ( ; ; ) {
		unsigned long timeout = sysctl_hung_task_timeout_secs;
		unsigned long interval = sysctl_hung_task_check_interval_secs;
		long t;

		if (interval == 0)
			interval = timeout;
		interval = min_t(unsigned long, interval, timeout);
		t = hung_timeout_jiffies(hung_last_checked, interval);
		if (t <= 0) {
			if (!atomic_xchg(&reset_hung_task, 0) &&
			    !hung_detector_suspended)
				check_hung_uninterruptible_tasks(timeout);
			hung_last_checked = jiffies;
			continue;
		}
		schedule_timeout_interruptible(t);
	}

	return 0;
}

static int __init hung_task_init(void)
{
	atomic_notifier_chain_register(&panic_notifier_list, &panic_block);

	/* Disable hung task detector on suspend */
	pm_notifier(hungtask_pm_notify, 0);

	watchdog_task = kthread_run(watchdog, NULL, "khungtaskd");
	hung_task_sysctl_init();

	return 0;
}
subsys_initcall(hung_task_init);
Commit	Line	Data
457c8996	1	// SPDX-License-Identifier: GPL-2.0-only
e162b39a MSB	2	/*
	3	* Detect Hung Task
	4	*
	5	* kernel/hung_task.c - kernel thread for detecting tasks stuck in D state
	6	*
	7	*/
	8
	9	#include <linux/mm.h>
	10	#include <linux/cpu.h>
	11	#include <linux/nmi.h>
	12	#include <linux/init.h>
	13	#include <linux/delay.h>
	14	#include <linux/freezer.h>
	15	#include <linux/kthread.h>
	16	#include <linux/lockdep.h>
9984de1a	17	#include <linux/export.h>
f39650de	18	#include <linux/panic_notifier.h>
e162b39a	19	#include <linux/sysctl.h>
a1c6ca3c	20	#include <linux/suspend.h>
41e85ce8	21	#include <linux/utsname.h>
3f07c014	22	#include <linux/sched/signal.h>
b17b0153	23	#include <linux/sched/debug.h>
a98eb6f1	24	#include <linux/sched/sysctl.h>
e711faaa	25	#include <linux/hung_task.h>
3f07c014	26
6a716c90	27	#include <trace/events/sched.h>
e162b39a MSB	28
e162b39a MSB	29	/*
ce9dbe24	30	* The number of tasks checked:
e162b39a	31	*/
882c5b26	32	static int __read_mostly sysctl_hung_task_check_count = PID_MAX_LIMIT;
ce9dbe24	33
03ecb24d LY	34	/*
	35	* Total number of tasks detected as hung since boot:
	36	*/
	37	static unsigned long __read_mostly sysctl_hung_task_detect_count;
	38
ce9dbe24 MSB	39	/*
	40	* Limit number of tasks checked in a batch.
	41	*
	42	* This value controls the preemptibility of khungtaskd since preemption
	43	* is disabled during the critical section. It also controls the size of
	44	* the RCU grace period. So it needs to be upper-bound.
	45	*/
304ae427	46	#define HUNG_TASK_LOCK_BREAK (HZ / 10)
e162b39a MSB	47
	48	/*
	49	* Zero means infinite timeout - no checking done:
	50	*/
e11feaa1	51	unsigned long __read_mostly sysctl_hung_task_timeout_secs = CONFIG_DEFAULT_HUNG_TASK_TIMEOUT;
5c3273ec	52	EXPORT_SYMBOL_GPL(sysctl_hung_task_timeout_secs);
e162b39a	53
a2e51445 DV	54	/*
	55	* Zero (default value) means use sysctl_hung_task_timeout_secs:
	56	*/
882c5b26	57	static unsigned long __read_mostly sysctl_hung_task_check_interval_secs;
a2e51445	58
882c5b26	59	static int __read_mostly sysctl_hung_task_warnings = 10;
e162b39a MSB	60
e162b39a MSB	61	static int __read_mostly did_panic;
780cbcf2	62	static bool hung_task_show_lock;
401c636a	63	static bool hung_task_call_panic;
0ec9dc9b	64	static bool hung_task_show_all_bt;
e162b39a MSB	65
	66	static struct task_struct *watchdog_task;
	67
0ec9dc9b GP	68	#ifdef CONFIG_SMP
	69	/*
	70	* Should we dump all CPUs backtraces in a hung task event?
	71	* Defaults to 0, can be changed via sysctl.
	72	*/
bbe7a10e XN	73	static unsigned int __read_mostly sysctl_hung_task_all_cpu_backtrace;
	74	#else
	75	#define sysctl_hung_task_all_cpu_backtrace 0
0ec9dc9b GP	76	#endif /* CONFIG_SMP */
0ec9dc9b GP	77
e162b39a MSB	78	/*
	79	* Should we panic (and reboot, if panic_timeout= is set) when a
	80	* hung task is detected:
	81	*/
882c5b26 TR	82	static unsigned int __read_mostly sysctl_hung_task_panic =
882c5b26 TR	83	IS_ENABLED(CONFIG_BOOTPARAM_HUNG_TASK_PANIC);
e162b39a	84
e162b39a MSB	85	static int
	86	hung_task_panic(struct notifier_block this, unsigned long event, void ptr)
	87	{
	88	did_panic = 1;
	89
	90	return NOTIFY_DONE;
	91	}
	92
	93	static struct notifier_block panic_block = {
	94	.notifier_call = hung_task_panic,
	95	};
	96
3cf67d61 MHG	97
	98	#ifdef CONFIG_DETECT_HUNG_TASK_BLOCKER
	99	static void debug_show_blocker(struct task_struct *task)
	100	{
	101	struct task_struct g, t;
194a9b9e	102	unsigned long owner, blocker, blocker_type;
3cf67d61 MHG	103
	104	RCU_LOCKDEP_WARN(!rcu_read_lock_held(), "No rcu lock held");
	105
e711faaa	106	blocker = READ_ONCE(task->blocker);
194a9b9e	107	if (!blocker)
3cf67d61 MHG	108	return;
3cf67d61 MHG	109
194a9b9e LY	110	blocker_type = hung_task_get_blocker_type(blocker);
	111
	112	switch (blocker_type) {
	113	case BLOCKER_TYPE_MUTEX:
	114	owner = mutex_get_owner(
	115	(struct mutex *)hung_task_blocker_to_lock(blocker));
	116	break;
	117	case BLOCKER_TYPE_SEM:
	118	owner = sem_last_holder(
	119	(struct semaphore *)hung_task_blocker_to_lock(blocker));
	120	break;
	121	default:
	122	WARN_ON_ONCE(1);
	123	return;
	124	}
	125
e711faaa	126
3cf67d61	127	if (unlikely(!owner)) {
194a9b9e LY	128	switch (blocker_type) {
	129	case BLOCKER_TYPE_MUTEX:
	130	pr_err("INFO: task %s:%d is blocked on a mutex, but the owner is not found.\n",
	131	task->comm, task->pid);
	132	break;
	133	case BLOCKER_TYPE_SEM:
	134	pr_err("INFO: task %s:%d is blocked on a semaphore, but the last holder is not found.\n",
	135	task->comm, task->pid);
	136	break;
	137	}
3cf67d61 MHG	138	return;
	139	}
	140
	141	/* Ensure the owner information is correct. */
	142	for_each_process_thread(g, t) {
194a9b9e LY	143	if ((unsigned long)t != owner)
	144	continue;
	145
	146	switch (blocker_type) {
	147	case BLOCKER_TYPE_MUTEX:
3cf67d61	148	pr_err("INFO: task %s:%d is blocked on a mutex likely owned by task %s:%d.\n",
194a9b9e LY	149	task->comm, task->pid, t->comm, t->pid);
	150	break;
	151	case BLOCKER_TYPE_SEM:
	152	pr_err("INFO: task %s:%d blocked on a semaphore likely last held by task %s:%d\n",
	153	task->comm, task->pid, t->comm, t->pid);
	154	break;
3cf67d61	155	}
194a9b9e LY	156	sched_show_task(t);
194a9b9e LY	157	return;
3cf67d61 MHG	158	}
	159	}
	160	#else
	161	static inline void debug_show_blocker(struct task_struct *task)
	162	{
	163	}
	164	#endif
	165
17406b82	166	static void check_hung_task(struct task_struct *t, unsigned long timeout)
e162b39a MSB	167	{
	168	unsigned long switch_count = t->nvcsw + t->nivcsw;
	169
cf2592f5 FW	170	/*
cf2592f5 FW	171	* Ensure the task is not frozen.
f9fab10b	172	* Also, skip vfork and any other user process that freezer should skip.
cf2592f5	173	*/
fdf756f7	174	if (unlikely(READ_ONCE(t->__state) & TASK_FROZEN))
f5d39b02	175	return;
f9fab10b MSB	176
	177	/*
	178	* When a freshly created task is scheduled once, changes its state to
	179	* TASK_UNINTERRUPTIBLE without having ever been switched out once, it
	180	* musn't be checked.
	181	*/
	182	if (unlikely(!switch_count))
e162b39a MSB	183	return;
e162b39a MSB	184
17406b82	185	if (switch_count != t->last_switch_count) {
e162b39a	186	t->last_switch_count = switch_count;
a2e51445	187	t->last_switch_time = jiffies;
e162b39a MSB	188	return;
e162b39a MSB	189	}
a2e51445 DV	190	if (time_is_after_jiffies(t->last_switch_time + timeout * HZ))
a2e51445 DV	191	return;
6a716c90	192
03ecb24d LY	193	/*
	194	* This counter tracks the total number of tasks detected as hung
	195	* since boot.
	196	*/
	197	sysctl_hung_task_detect_count++;
	198
6a716c90 ON	199	trace_sched_process_hang(t);
6a716c90 ON	200
168e06f7 LC	201	if (sysctl_hung_task_panic) {
	202	console_verbose();
	203	hung_task_show_lock = true;
	204	hung_task_call_panic = true;
	205	}
270750db	206
e162b39a MSB	207	/*
	208	* Ok, the task did not get scheduled for more than 2 minutes,
	209	* complain:
	210	*/
abd8ac05	211	if (sysctl_hung_task_warnings \|\| hung_task_call_panic) {
4ca5ede0 TH	212	if (sysctl_hung_task_warnings > 0)
4ca5ede0 TH	213	sysctl_hung_task_warnings--;
48a6d64e	214	pr_err("INFO: task %s:%d blocked for more than %ld seconds.\n",
b014beba	215	t->comm, t->pid, (jiffies - t->last_switch_time) / HZ);
48a6d64e JS	216	pr_err(" %s %s %.*s\n",
	217	print_tainted(), init_utsname()->release,
	218	(int)strcspn(init_utsname()->version, " "),
	219	init_utsname()->version);
65ef17aa OK	220	if (t->flags & PF_POSTCOREDUMP)
65ef17aa OK	221	pr_err(" Blocked by coredump.\n");
48a6d64e JS	222	pr_err("\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
	223	" disables this message.\n");
	224	sched_show_task(t);
3cf67d61	225	debug_show_blocker(t);
780cbcf2	226	hung_task_show_lock = true;
0ec9dc9b GP	227
	228	if (sysctl_hung_task_all_cpu_backtrace)
	229	hung_task_show_all_bt = true;
b1f712b3	230	if (!sysctl_hung_task_warnings)
b1f712b3	231	pr_info("Future hung task reports are suppressed, see sysctl kernel.hung_task_warnings\n");
48a6d64e	232	}
e162b39a	233
e162b39a	234	touch_nmi_watchdog();
e162b39a MSB	235	}
e162b39a MSB	236
ce9dbe24 MSB	237	/*
	238	* To avoid extending the RCU grace period for an unbounded amount of time,
	239	* periodically exit the critical section and enter a new one.
	240	*
	241	* For preemptible RCU it is sufficient to call rcu_read_unlock in order
6a103b0d	242	* to exit the grace period. For classic RCU, a reschedule is required.
ce9dbe24	243	*/
6027ce49	244	static bool rcu_lock_break(struct task_struct g, struct task_struct t)
ce9dbe24	245	{
6027ce49 ON	246	bool can_cont;
6027ce49 ON	247
ce9dbe24 MSB	248	get_task_struct(g);
	249	get_task_struct(t);
	250	rcu_read_unlock();
	251	cond_resched();
	252	rcu_read_lock();
6027ce49	253	can_cont = pid_alive(g) && pid_alive(t);
ce9dbe24 MSB	254	put_task_struct(t);
ce9dbe24 MSB	255	put_task_struct(g);
6027ce49 ON	256
6027ce49 ON	257	return can_cont;
ce9dbe24 MSB	258	}
ce9dbe24 MSB	259
e162b39a MSB	260	/*
	261	* Check whether a TASK_UNINTERRUPTIBLE does not get woken up for
	262	* a really long time (120 seconds). If that happens, print out
	263	* a warning.
	264	*/
603a148f	265	static void check_hung_uninterruptible_tasks(unsigned long timeout)
e162b39a MSB	266	{
e162b39a MSB	267	int max_count = sysctl_hung_task_check_count;
304ae427	268	unsigned long last_break = jiffies;
e162b39a MSB	269	struct task_struct g, t;
	270
	271	/*
	272	* If the system crashed already then all bets are off,
	273	* do not report extra hung tasks:
	274	*/
	275	if (test_taint(TAINT_DIE) \|\| did_panic)
	276	return;
	277
780cbcf2	278	hung_task_show_lock = false;
94be52dc	279	rcu_read_lock();
972fae69	280	for_each_process_thread(g, t) {
5aec788a PZ	281	unsigned int state;
5aec788a PZ	282
e5af0226	283	if (!max_count--)
e162b39a	284	goto unlock;
304ae427	285	if (time_after(jiffies, last_break + HUNG_TASK_LOCK_BREAK)) {
6027ce49	286	if (!rcu_lock_break(g, t))
ce9dbe24	287	goto unlock;
304ae427	288	last_break = jiffies;
ce9dbe24	289	}
fdf756f7 PZ	290	/*
	291	* skip the TASK_KILLABLE tasks -- these can be killed
	292	* skip the TASK_IDLE tasks -- those are genuinely idle
	293	*/
5aec788a PZ	294	state = READ_ONCE(t->__state);
5aec788a PZ	295	if ((state & TASK_UNINTERRUPTIBLE) &&
fdf756f7 PZ	296	!(state & TASK_WAKEKILL) &&
fdf756f7 PZ	297	!(state & TASK_NOLOAD))
17406b82	298	check_hung_task(t, timeout);
972fae69	299	}
e162b39a	300	unlock:
94be52dc	301	rcu_read_unlock();
07a22b61	302	if (hung_task_show_lock)
780cbcf2	303	debug_show_all_locks();
0ec9dc9b GP	304
	305	if (hung_task_show_all_bt) {
	306	hung_task_show_all_bt = false;
401c636a	307	trigger_all_cpu_backtrace();
401c636a	308	}
0ec9dc9b GP	309
	310	if (hung_task_call_panic)
	311	panic("hung_task: blocked tasks");
e162b39a MSB	312	}
e162b39a MSB	313
b4aa14a6 TH	314	static long hung_timeout_jiffies(unsigned long last_checked,
b4aa14a6 TH	315	unsigned long timeout)
e162b39a MSB	316	{
e162b39a MSB	317	/* timeout of 0 will disable the watchdog */
b4aa14a6 TH	318	return timeout ? last_checked - jiffies + timeout * HZ :
b4aa14a6 TH	319	MAX_SCHEDULE_TIMEOUT;
e162b39a MSB	320	}
e162b39a MSB	321
bbe7a10e	322	#ifdef CONFIG_SYSCTL
e162b39a MSB	323	/*
	324	* Process updating of timeout sysctl
	325	*/
78eb4ea2	326	static int proc_dohung_task_timeout_secs(const struct ctl_table *table, int write,
591c32bd	327	void *buffer,
bbe7a10e	328	size_t lenp, loff_t ppos)
e162b39a MSB	329	{
	330	int ret;
	331
8d65af78	332	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
e162b39a MSB	333
	334	if (ret \|\| !write)
	335	goto out;
	336
e162b39a MSB	337	wake_up_process(watchdog_task);
	338
	339	out:
	340	return ret;
	341	}
	342
bbe7a10e XN	343	/*
	344	* This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
	345	* and hung_task_check_interval_secs
	346	*/
	347	static const unsigned long hung_task_timeout_max = (LONG_MAX / HZ);
1751f872	348	static const struct ctl_table hung_task_sysctls[] = {
bbe7a10e XN	349	#ifdef CONFIG_SMP
	350	{
	351	.procname = "hung_task_all_cpu_backtrace",
	352	.data = &sysctl_hung_task_all_cpu_backtrace,
	353	.maxlen = sizeof(int),
	354	.mode = 0644,
	355	.proc_handler = proc_dointvec_minmax,
	356	.extra1 = SYSCTL_ZERO,
	357	.extra2 = SYSCTL_ONE,
	358	},
	359	#endif /* CONFIG_SMP */
	360	{
	361	.procname = "hung_task_panic",
	362	.data = &sysctl_hung_task_panic,
	363	.maxlen = sizeof(int),
	364	.mode = 0644,
	365	.proc_handler = proc_dointvec_minmax,
	366	.extra1 = SYSCTL_ZERO,
	367	.extra2 = SYSCTL_ONE,
	368	},
	369	{
	370	.procname = "hung_task_check_count",
	371	.data = &sysctl_hung_task_check_count,
	372	.maxlen = sizeof(int),
	373	.mode = 0644,
	374	.proc_handler = proc_dointvec_minmax,
	375	.extra1 = SYSCTL_ZERO,
	376	},
	377	{
	378	.procname = "hung_task_timeout_secs",
	379	.data = &sysctl_hung_task_timeout_secs,
	380	.maxlen = sizeof(unsigned long),
	381	.mode = 0644,
	382	.proc_handler = proc_dohung_task_timeout_secs,
	383	.extra2 = (void *)&hung_task_timeout_max,
	384	},
	385	{
	386	.procname = "hung_task_check_interval_secs",
	387	.data = &sysctl_hung_task_check_interval_secs,
	388	.maxlen = sizeof(unsigned long),
	389	.mode = 0644,
	390	.proc_handler = proc_dohung_task_timeout_secs,
	391	.extra2 = (void *)&hung_task_timeout_max,
	392	},
	393	{
	394	.procname = "hung_task_warnings",
	395	.data = &sysctl_hung_task_warnings,
	396	.maxlen = sizeof(int),
	397	.mode = 0644,
	398	.proc_handler = proc_dointvec_minmax,
	399	.extra1 = SYSCTL_NEG_ONE,
	400	},
03ecb24d LY	401	{
	402	.procname = "hung_task_detect_count",
	403	.data = &sysctl_hung_task_detect_count,
	404	.maxlen = sizeof(unsigned long),
	405	.mode = 0444,
	406	.proc_handler = proc_doulongvec_minmax,
	407	},
bbe7a10e XN	408	};
	409
	410	static void __init hung_task_sysctl_init(void)
	411	{
	412	register_sysctl_init("kernel", hung_task_sysctls);
	413	}
	414	#else
	415	#define hung_task_sysctl_init() do { } while (0)
	416	#endif /* CONFIG_SYSCTL */
	417
	418
8b414521 MT	419	static atomic_t reset_hung_task = ATOMIC_INIT(0);
	420
	421	void reset_hung_task_detector(void)
	422	{
	423	atomic_set(&reset_hung_task, 1);
	424	}
	425	EXPORT_SYMBOL_GPL(reset_hung_task_detector);
	426
a1c6ca3c VK	427	static bool hung_detector_suspended;
	428
	429	static int hungtask_pm_notify(struct notifier_block *self,
	430	unsigned long action, void *hcpu)
	431	{
	432	switch (action) {
	433	case PM_SUSPEND_PREPARE:
	434	case PM_HIBERNATION_PREPARE:
	435	case PM_RESTORE_PREPARE:
	436	hung_detector_suspended = true;
	437	break;
	438	case PM_POST_SUSPEND:
	439	case PM_POST_HIBERNATION:
	440	case PM_POST_RESTORE:
	441	hung_detector_suspended = false;
	442	break;
	443	default:
	444	break;
	445	}
	446	return NOTIFY_OK;
	447	}
	448
e162b39a MSB	449	/*
	450	* kthread which checks for tasks stuck in D state
	451	*/
	452	static int watchdog(void *dummy)
	453	{
b4aa14a6 TH	454	unsigned long hung_last_checked = jiffies;
b4aa14a6 TH	455
e162b39a	456	set_user_nice(current, 0);
e162b39a MSB	457
e162b39a MSB	458	for ( ; ; ) {
17406b82	459	unsigned long timeout = sysctl_hung_task_timeout_secs;
a2e51445 DV	460	unsigned long interval = sysctl_hung_task_check_interval_secs;
a2e51445 DV	461	long t;
603a148f	462
a2e51445 DV	463	if (interval == 0)
	464	interval = timeout;
	465	interval = min_t(unsigned long, interval, timeout);
	466	t = hung_timeout_jiffies(hung_last_checked, interval);
b4aa14a6	467	if (t <= 0) {
a1c6ca3c VK	468	if (!atomic_xchg(&reset_hung_task, 0) &&
a1c6ca3c VK	469	!hung_detector_suspended)
b4aa14a6 TH	470	check_hung_uninterruptible_tasks(timeout);
b4aa14a6 TH	471	hung_last_checked = jiffies;
8b414521	472	continue;
b4aa14a6 TH	473	}
b4aa14a6 TH	474	schedule_timeout_interruptible(t);
e162b39a MSB	475	}
	476
	477	return 0;
	478	}
	479
	480	static int __init hung_task_init(void)
	481	{
	482	atomic_notifier_chain_register(&panic_notifier_list, &panic_block);
a1c6ca3c VK	483
	484	/* Disable hung task detector on suspend */
	485	pm_notifier(hungtask_pm_notify, 0);
	486
e162b39a	487	watchdog_task = kthread_run(watchdog, NULL, "khungtaskd");
bbe7a10e	488	hung_task_sysctl_init();
e162b39a MSB	489
	490	return 0;
	491	}
c96d6660	492	subsys_initcall(hung_task_init);