[people/ms/linux.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:
 */
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:
 */
unsigned long __read_mostly sysctl_hung_task_check_interval_secs;

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:
 */
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(t->flags & (PF_FROZEN | PF_FREEZER_SKIP)))
	    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) {
		printk_prefer_direct_enter();

		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;

		printk_prefer_direct_exit();
	}

	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) {
		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;
		}
		/* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
		if (READ_ONCE(t->__state) == TASK_UNINTERRUPTIBLE)
			check_hung_task(t, timeout);
	}
 unlock:
	rcu_read_unlock();
	if (hung_task_show_lock) {
		printk_prefer_direct_enter();
		debug_show_all_locks();
		printk_prefer_direct_exit();
	}

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

	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 __user *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	*/
cd64647f	31	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	*/
	50	unsigned long __read_mostly sysctl_hung_task_check_interval_secs;
	51
270750db	52	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	*/
	75	unsigned int __read_mostly sysctl_hung_task_panic =
67fca000	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	*/
f9fab10b MSB	98	if (unlikely(t->flags & (PF_FROZEN \| PF_FREEZER_SKIP)))
	99	return;
	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) {
2bb2b7b5 JO	130	printk_prefer_direct_enter();
2bb2b7b5 JO	131
4ca5ede0 TH	132	if (sysctl_hung_task_warnings > 0)
4ca5ede0 TH	133	sysctl_hung_task_warnings--;
48a6d64e	134	pr_err("INFO: task %s:%d blocked for more than %ld seconds.\n",
b014beba	135	t->comm, t->pid, (jiffies - t->last_switch_time) / HZ);
48a6d64e JS	136	pr_err(" %s %s %.*s\n",
	137	print_tainted(), init_utsname()->release,
	138	(int)strcspn(init_utsname()->version, " "),
	139	init_utsname()->version);
	140	pr_err("\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
	141	" disables this message.\n");
	142	sched_show_task(t);
780cbcf2	143	hung_task_show_lock = true;
0ec9dc9b GP	144
	145	if (sysctl_hung_task_all_cpu_backtrace)
	146	hung_task_show_all_bt = true;
2bb2b7b5 JO	147
2bb2b7b5 JO	148	printk_prefer_direct_exit();
48a6d64e	149	}
e162b39a	150
e162b39a	151	touch_nmi_watchdog();
e162b39a MSB	152	}
e162b39a MSB	153
ce9dbe24 MSB	154	/*
	155	* To avoid extending the RCU grace period for an unbounded amount of time,
	156	* periodically exit the critical section and enter a new one.
	157	*
	158	* For preemptible RCU it is sufficient to call rcu_read_unlock in order
6a103b0d	159	* to exit the grace period. For classic RCU, a reschedule is required.
ce9dbe24	160	*/
6027ce49	161	static bool rcu_lock_break(struct task_struct g, struct task_struct t)
ce9dbe24	162	{
6027ce49 ON	163	bool can_cont;
6027ce49 ON	164
ce9dbe24 MSB	165	get_task_struct(g);
	166	get_task_struct(t);
	167	rcu_read_unlock();
	168	cond_resched();
	169	rcu_read_lock();
6027ce49	170	can_cont = pid_alive(g) && pid_alive(t);
ce9dbe24 MSB	171	put_task_struct(t);
ce9dbe24 MSB	172	put_task_struct(g);
6027ce49 ON	173
6027ce49 ON	174	return can_cont;
ce9dbe24 MSB	175	}
ce9dbe24 MSB	176
e162b39a MSB	177	/*
	178	* Check whether a TASK_UNINTERRUPTIBLE does not get woken up for
	179	* a really long time (120 seconds). If that happens, print out
	180	* a warning.
	181	*/
603a148f	182	static void check_hung_uninterruptible_tasks(unsigned long timeout)
e162b39a MSB	183	{
e162b39a MSB	184	int max_count = sysctl_hung_task_check_count;
304ae427	185	unsigned long last_break = jiffies;
e162b39a MSB	186	struct task_struct g, t;
	187
	188	/*
	189	* If the system crashed already then all bets are off,
	190	* do not report extra hung tasks:
	191	*/
	192	if (test_taint(TAINT_DIE) \|\| did_panic)
	193	return;
	194
780cbcf2	195	hung_task_show_lock = false;
94be52dc	196	rcu_read_lock();
972fae69	197	for_each_process_thread(g, t) {
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	}
e162b39a	205	/* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
2f064a59	206	if (READ_ONCE(t->__state) == TASK_UNINTERRUPTIBLE)
17406b82	207	check_hung_task(t, timeout);
972fae69	208	}
e162b39a	209	unlock:
94be52dc	210	rcu_read_unlock();
2bb2b7b5 JO	211	if (hung_task_show_lock) {
2bb2b7b5 JO	212	printk_prefer_direct_enter();
780cbcf2	213	debug_show_all_locks();
2bb2b7b5 JO	214	printk_prefer_direct_exit();
2bb2b7b5 JO	215	}
0ec9dc9b GP	216
	217	if (hung_task_show_all_bt) {
	218	hung_task_show_all_bt = false;
2bb2b7b5	219	printk_prefer_direct_enter();
401c636a	220	trigger_all_cpu_backtrace();
2bb2b7b5	221	printk_prefer_direct_exit();
401c636a	222	}
0ec9dc9b GP	223
	224	if (hung_task_call_panic)
	225	panic("hung_task: blocked tasks");
e162b39a MSB	226	}
e162b39a MSB	227
b4aa14a6 TH	228	static long hung_timeout_jiffies(unsigned long last_checked,
b4aa14a6 TH	229	unsigned long timeout)
e162b39a MSB	230	{
e162b39a MSB	231	/* timeout of 0 will disable the watchdog */
b4aa14a6 TH	232	return timeout ? last_checked - jiffies + timeout * HZ :
b4aa14a6 TH	233	MAX_SCHEDULE_TIMEOUT;
e162b39a MSB	234	}
e162b39a MSB	235
bbe7a10e	236	#ifdef CONFIG_SYSCTL
e162b39a MSB	237	/*
	238	* Process updating of timeout sysctl
	239	*/
bbe7a10e XN	240	static int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
	241	void __user *buffer,
	242	size_t lenp, loff_t ppos)
e162b39a MSB	243	{
	244	int ret;
	245
8d65af78	246	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
e162b39a MSB	247
	248	if (ret \|\| !write)
	249	goto out;
	250
e162b39a MSB	251	wake_up_process(watchdog_task);
	252
	253	out:
	254	return ret;
	255	}
	256
bbe7a10e XN	257	/*
	258	* This is needed for proc_doulongvec_minmax of sysctl_hung_task_timeout_secs
	259	* and hung_task_check_interval_secs
	260	*/
	261	static const unsigned long hung_task_timeout_max = (LONG_MAX / HZ);
	262	static struct ctl_table hung_task_sysctls[] = {
	263	#ifdef CONFIG_SMP
	264	{
	265	.procname = "hung_task_all_cpu_backtrace",
	266	.data = &sysctl_hung_task_all_cpu_backtrace,
	267	.maxlen = sizeof(int),
	268	.mode = 0644,
	269	.proc_handler = proc_dointvec_minmax,
	270	.extra1 = SYSCTL_ZERO,
	271	.extra2 = SYSCTL_ONE,
	272	},
	273	#endif /* CONFIG_SMP */
	274	{
	275	.procname = "hung_task_panic",
	276	.data = &sysctl_hung_task_panic,
	277	.maxlen = sizeof(int),
	278	.mode = 0644,
	279	.proc_handler = proc_dointvec_minmax,
	280	.extra1 = SYSCTL_ZERO,
	281	.extra2 = SYSCTL_ONE,
	282	},
	283	{
	284	.procname = "hung_task_check_count",
	285	.data = &sysctl_hung_task_check_count,
	286	.maxlen = sizeof(int),
	287	.mode = 0644,
	288	.proc_handler = proc_dointvec_minmax,
	289	.extra1 = SYSCTL_ZERO,
	290	},
	291	{
	292	.procname = "hung_task_timeout_secs",
	293	.data = &sysctl_hung_task_timeout_secs,
	294	.maxlen = sizeof(unsigned long),
	295	.mode = 0644,
	296	.proc_handler = proc_dohung_task_timeout_secs,
	297	.extra2 = (void *)&hung_task_timeout_max,
	298	},
	299	{
	300	.procname = "hung_task_check_interval_secs",
	301	.data = &sysctl_hung_task_check_interval_secs,
	302	.maxlen = sizeof(unsigned long),
	303	.mode = 0644,
	304	.proc_handler = proc_dohung_task_timeout_secs,
	305	.extra2 = (void *)&hung_task_timeout_max,
	306	},
	307	{
	308	.procname = "hung_task_warnings",
	309	.data = &sysctl_hung_task_warnings,
	310	.maxlen = sizeof(int),
	311	.mode = 0644,
	312	.proc_handler = proc_dointvec_minmax,
	313	.extra1 = SYSCTL_NEG_ONE,
	314	},
	315	{}
	316	};
	317
	318	static void __init hung_task_sysctl_init(void)
	319	{
	320	register_sysctl_init("kernel", hung_task_sysctls);
321	}
322	#else
323	#define hung_task_sysctl_init() do { } while (0)
324	#endif /* CONFIG_SYSCTL */
325
326
8b414521 MT	327	static atomic_t reset_hung_task = ATOMIC_INIT(0);
	328
	329	void reset_hung_task_detector(void)
	330	{
	331	atomic_set(&reset_hung_task, 1);
	332	}
	333	EXPORT_SYMBOL_GPL(reset_hung_task_detector);
	334
a1c6ca3c VK	335	static bool hung_detector_suspended;
	336
	337	static int hungtask_pm_notify(struct notifier_block *self,
	338	unsigned long action, void *hcpu)
	339	{
	340	switch (action) {
	341	case PM_SUSPEND_PREPARE:
	342	case PM_HIBERNATION_PREPARE:
	343	case PM_RESTORE_PREPARE:
	344	hung_detector_suspended = true;
	345	break;
	346	case PM_POST_SUSPEND:
	347	case PM_POST_HIBERNATION:
	348	case PM_POST_RESTORE:
	349	hung_detector_suspended = false;
	350	break;
	351	default:
	352	break;
	353	}
	354	return NOTIFY_OK;
	355	}
	356
e162b39a MSB	357	/*
	358	* kthread which checks for tasks stuck in D state
	359	*/
	360	static int watchdog(void *dummy)
	361	{
b4aa14a6 TH	362	unsigned long hung_last_checked = jiffies;
b4aa14a6 TH	363
e162b39a	364	set_user_nice(current, 0);
e162b39a MSB	365
e162b39a MSB	366	for ( ; ; ) {
17406b82	367	unsigned long timeout = sysctl_hung_task_timeout_secs;
a2e51445 DV	368	unsigned long interval = sysctl_hung_task_check_interval_secs;
a2e51445 DV	369	long t;
603a148f	370
a2e51445 DV	371	if (interval == 0)
	372	interval = timeout;
	373	interval = min_t(unsigned long, interval, timeout);
	374	t = hung_timeout_jiffies(hung_last_checked, interval);
b4aa14a6	375	if (t <= 0) {
a1c6ca3c VK	376	if (!atomic_xchg(&reset_hung_task, 0) &&
a1c6ca3c VK	377	!hung_detector_suspended)
b4aa14a6 TH	378	check_hung_uninterruptible_tasks(timeout);
b4aa14a6 TH	379	hung_last_checked = jiffies;
8b414521	380	continue;
b4aa14a6 TH	381	}
b4aa14a6 TH	382	schedule_timeout_interruptible(t);
e162b39a MSB	383	}
	384
	385	return 0;
	386	}
	387
	388	static int __init hung_task_init(void)
	389	{
	390	atomic_notifier_chain_register(&panic_notifier_list, &panic_block);
a1c6ca3c VK	391
	392	/* Disable hung task detector on suspend */
	393	pm_notifier(hungtask_pm_notify, 0);
	394
e162b39a	395	watchdog_task = kthread_run(watchdog, NULL, "khungtaskd");
bbe7a10e	396	hung_task_sysctl_init();
e162b39a MSB	397
	398	return 0;
	399	}
c96d6660	400	subsys_initcall(hung_task_init);