[linux-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 <trace/events/sched.h>

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

/*
 * 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;

/*
 * 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,
};

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;

	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) {
		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);
		pr_err("\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
			" disables this message.\n");
		sched_show_task(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(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 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,
	},
	{}
};

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