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

/*
 * Detect Soft Lockups
 *
 * started by Ingo Molnar, Copyright (C) 2005, 2006 Red Hat, Inc.
 *
 * this code detects soft lockups: incidents in where on a CPU
 * the kernel does not reschedule for 10 seconds or more.
 */
#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/notifier.h>
#include <linux/module.h>
#include <linux/sysctl.h>

#include <asm/irq_regs.h>

static DEFINE_SPINLOCK(print_lock);

static DEFINE_PER_CPU(unsigned long, touch_timestamp);
static DEFINE_PER_CPU(unsigned long, print_timestamp);
static DEFINE_PER_CPU(struct task_struct *, watchdog_task);

static int __read_mostly did_panic;
int __read_mostly softlockup_thresh = 60;

/*
 * Should we panic (and reboot, if panic_timeout= is set) when a
 * soft-lockup occurs:
 */
unsigned int __read_mostly softlockup_panic =
				CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE;

static int __init softlockup_panic_setup(char *str)
{
	softlockup_panic = simple_strtoul(str, NULL, 0);

	return 1;
}
__setup("softlockup_panic=", softlockup_panic_setup);

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

	return NOTIFY_DONE;
}

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

/*
 * 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.
 */
static unsigned long get_timestamp(int this_cpu)
{
	return cpu_clock(this_cpu) >> 30LL;  /* 2^30 ~= 10^9 */
}

static void __touch_softlockup_watchdog(void)
{
	int this_cpu = raw_smp_processor_id();

	__raw_get_cpu_var(touch_timestamp) = get_timestamp(this_cpu);
}

void touch_softlockup_watchdog(void)
{
	__raw_get_cpu_var(touch_timestamp) = 0;
}
EXPORT_SYMBOL(touch_softlockup_watchdog);

void touch_all_softlockup_watchdogs(void)
{
	int cpu;

	/* Cause each CPU to re-update its timestamp rather than complain */
	for_each_online_cpu(cpu)
		per_cpu(touch_timestamp, cpu) = 0;
}
EXPORT_SYMBOL(touch_all_softlockup_watchdogs);

int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
			     struct file *filp, void __user *buffer,
			     size_t *lenp, loff_t *ppos)
{
	touch_all_softlockup_watchdogs();
	return proc_dointvec_minmax(table, write, filp, buffer, lenp, ppos);
}

/*
 * This callback runs from the timer interrupt, and checks
 * whether the watchdog thread has hung or not:
 */
void softlockup_tick(void)
{
	int this_cpu = smp_processor_id();
	unsigned long touch_timestamp = per_cpu(touch_timestamp, this_cpu);
	unsigned long print_timestamp;
	struct pt_regs *regs = get_irq_regs();
	unsigned long now;

	/* Is detection switched off? */
	if (!per_cpu(watchdog_task, this_cpu) || softlockup_thresh <= 0) {
		/* Be sure we don't false trigger if switched back on */
		if (touch_timestamp)
			per_cpu(touch_timestamp, this_cpu) = 0;
		return;
	}

	if (touch_timestamp == 0) {
		__touch_softlockup_watchdog();
		return;
	}

	print_timestamp = per_cpu(print_timestamp, this_cpu);

	/* report at most once a second */
	if (print_timestamp == touch_timestamp || did_panic)
		return;

	/* do not print during early bootup: */
	if (unlikely(system_state != SYSTEM_RUNNING)) {
		__touch_softlockup_watchdog();
		return;
	}

	now = get_timestamp(this_cpu);

	/*
	 * Wake up the high-prio watchdog task twice per
	 * threshold timespan.
	 */
	if (now > touch_timestamp + softlockup_thresh/2)
		wake_up_process(per_cpu(watchdog_task, this_cpu));

	/* Warn about unreasonable delays: */
	if (now <= (touch_timestamp + softlockup_thresh))
		return;

	per_cpu(print_timestamp, this_cpu) = touch_timestamp;

	spin_lock(&print_lock);
	printk(KERN_ERR "BUG: soft lockup - CPU#%d stuck for %lus! [%s:%d]\n",
			this_cpu, now - touch_timestamp,
			current->comm, task_pid_nr(current));
	print_modules();
	print_irqtrace_events(current);
	if (regs)
		show_regs(regs);
	else
		dump_stack();
	spin_unlock(&print_lock);

	if (softlockup_panic)
		panic("softlockup: hung tasks");
}

/*
 * Have a reasonable limit on the number of tasks checked:
 */
unsigned long __read_mostly sysctl_hung_task_check_count = 1024;

/*
 * Zero means infinite timeout - no checking done:
 */
unsigned long __read_mostly sysctl_hung_task_timeout_secs = 480;

unsigned long __read_mostly sysctl_hung_task_warnings = 10;

/*
 * Only do the hung-tasks check on one CPU:
 */
static int check_cpu __read_mostly = -1;

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

	if (t->flags & PF_FROZEN)
		return;

	if (switch_count != t->last_switch_count || !t->last_switch_timestamp) {
		t->last_switch_count = switch_count;
		t->last_switch_timestamp = now;
		return;
	}
	if ((long)(now - t->last_switch_timestamp) <
					sysctl_hung_task_timeout_secs)
		return;
	if (!sysctl_hung_task_warnings)
		return;
	sysctl_hung_task_warnings--;

	/*
	 * Ok, the task did not get scheduled for more than 2 minutes,
	 * complain:
	 */
	printk(KERN_ERR "INFO: task %s:%d blocked for more than "
			"%ld seconds.\n", t->comm, t->pid,
			sysctl_hung_task_timeout_secs);
	printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
			" disables this message.\n");
	sched_show_task(t);
	__debug_show_held_locks(t);

	t->last_switch_timestamp = now;
	touch_nmi_watchdog();

	if (softlockup_panic)
		panic("softlockup: blocked tasks");
}

/*
 * 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(int this_cpu)
{
	int max_count = sysctl_hung_task_check_count;
	unsigned long now = get_timestamp(this_cpu);
	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;

	read_lock(&tasklist_lock);
	do_each_thread(g, t) {
		if (!--max_count)
			goto unlock;
		/* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
		if (t->state == TASK_UNINTERRUPTIBLE)
			check_hung_task(t, now);
	} while_each_thread(g, t);
 unlock:
	read_unlock(&tasklist_lock);
}

/*
 * The watchdog thread - runs every second and touches the timestamp.
 */
static int watchdog(void *__bind_cpu)
{
	struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
	int this_cpu = (long)__bind_cpu;

	sched_setscheduler(current, SCHED_FIFO, &param);

	/* initialize timestamp */
	__touch_softlockup_watchdog();

	set_current_state(TASK_INTERRUPTIBLE);
	/*
	 * Run briefly once per second to reset the softlockup timestamp.
	 * If this gets delayed for more than 60 seconds then the
	 * debug-printout triggers in softlockup_tick().
	 */
	while (!kthread_should_stop()) {
		__touch_softlockup_watchdog();
		schedule();

		if (kthread_should_stop())
			break;

		if (this_cpu == check_cpu) {
			if (sysctl_hung_task_timeout_secs)
				check_hung_uninterruptible_tasks(this_cpu);
		}

		set_current_state(TASK_INTERRUPTIBLE);
	}
	__set_current_state(TASK_RUNNING);

	return 0;
}

/*
 * Create/destroy watchdog threads as CPUs come and go:
 */
static int __cpuinit
cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
{
	int hotcpu = (unsigned long)hcpu;
	struct task_struct *p;

	switch (action) {
	case CPU_UP_PREPARE:
	case CPU_UP_PREPARE_FROZEN:
		BUG_ON(per_cpu(watchdog_task, hotcpu));
		p = kthread_create(watchdog, hcpu, "watchdog/%d", hotcpu);
		if (IS_ERR(p)) {
			printk(KERN_ERR "watchdog for %i failed\n", hotcpu);
			return NOTIFY_BAD;
		}
		per_cpu(touch_timestamp, hotcpu) = 0;
		per_cpu(watchdog_task, hotcpu) = p;
		kthread_bind(p, hotcpu);
		break;
	case CPU_ONLINE:
	case CPU_ONLINE_FROZEN:
		check_cpu = cpumask_any(cpu_online_mask);
		wake_up_process(per_cpu(watchdog_task, hotcpu));
		break;
#ifdef CONFIG_HOTPLUG_CPU
	case CPU_DOWN_PREPARE:
	case CPU_DOWN_PREPARE_FROZEN:
		if (hotcpu == check_cpu) {
			/* Pick any other online cpu. */
			check_cpu = cpumask_any_but(cpu_online_mask, hotcpu);
		}
		break;

	case CPU_UP_CANCELED:
	case CPU_UP_CANCELED_FROZEN:
		if (!per_cpu(watchdog_task, hotcpu))
			break;
		/* Unbind so it can run.  Fall thru. */
		kthread_bind(per_cpu(watchdog_task, hotcpu),
			     cpumask_any(cpu_online_mask));
	case CPU_DEAD:
	case CPU_DEAD_FROZEN:
		p = per_cpu(watchdog_task, hotcpu);
		per_cpu(watchdog_task, hotcpu) = NULL;
		kthread_stop(p);
		break;
#endif /* CONFIG_HOTPLUG_CPU */
	}
	return NOTIFY_OK;
}

static struct notifier_block __cpuinitdata cpu_nfb = {
	.notifier_call = cpu_callback
};

static int __initdata nosoftlockup;

static int __init nosoftlockup_setup(char *str)
{
	nosoftlockup = 1;
	return 1;
}
__setup("nosoftlockup", nosoftlockup_setup);

static int __init spawn_softlockup_task(void)
{
	void *cpu = (void *)(long)smp_processor_id();
	int err;

	if (nosoftlockup)
		return 0;

	err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu);
	if (err == NOTIFY_BAD) {
		BUG();
		return 1;
	}
	cpu_callback(&cpu_nfb, CPU_ONLINE, cpu);
	register_cpu_notifier(&cpu_nfb);

	atomic_notifier_chain_register(&panic_notifier_list, &panic_block);

	return 0;
}
early_initcall(spawn_softlockup_task);
Commit	Line	Data
8446f1d3 IM	1	/*
	2	* Detect Soft Lockups
	3	*
6687a97d	4	* started by Ingo Molnar, Copyright (C) 2005, 2006 Red Hat, Inc.
8446f1d3 IM	5	*
	6	* this code detects soft lockups: incidents in where on a CPU
	7	* the kernel does not reschedule for 10 seconds or more.
	8	*/
8446f1d3 IM	9	#include <linux/mm.h>
8446f1d3 IM	10	#include <linux/cpu.h>
82a1fcb9	11	#include <linux/nmi.h>
8446f1d3 IM	12	#include <linux/init.h>
8446f1d3 IM	13	#include <linux/delay.h>
83144186	14	#include <linux/freezer.h>
8446f1d3	15	#include <linux/kthread.h>
8d5be7f4	16	#include <linux/lockdep.h>
8446f1d3 IM	17	#include <linux/notifier.h>
8446f1d3 IM	18	#include <linux/module.h>
baf48f65	19	#include <linux/sysctl.h>
8446f1d3	20
43581a10 IM	21	#include <asm/irq_regs.h>
43581a10 IM	22
8446f1d3 IM	23	static DEFINE_SPINLOCK(print_lock);
8446f1d3 IM	24
6687a97d IM	25	static DEFINE_PER_CPU(unsigned long, touch_timestamp);
6687a97d IM	26	static DEFINE_PER_CPU(unsigned long, print_timestamp);
8446f1d3 IM	27	static DEFINE_PER_CPU(struct task_struct *, watchdog_task);
8446f1d3 IM	28
90739081	29	static int __read_mostly did_panic;
9383d967	30	int __read_mostly softlockup_thresh = 60;
6687a97d	31
9c44bc03 IM	32	/*
	33	* Should we panic (and reboot, if panic_timeout= is set) when a
	34	* soft-lockup occurs:
	35	*/
	36	unsigned int __read_mostly softlockup_panic =
	37	CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE;
	38
	39	static int __init softlockup_panic_setup(char *str)
	40	{
	41	softlockup_panic = simple_strtoul(str, NULL, 0);
	42
	43	return 1;
	44	}
	45	__setup("softlockup_panic=", softlockup_panic_setup);
	46
6687a97d IM	47	static int
6687a97d IM	48	softlock_panic(struct notifier_block this, unsigned long event, void ptr)
8446f1d3 IM	49	{
	50	did_panic = 1;
	51
	52	return NOTIFY_DONE;
	53	}
	54
	55	static struct notifier_block panic_block = {
	56	.notifier_call = softlock_panic,
	57	};
	58
966812dc JF	59	/*
	60	* Returns seconds, approximately. We don't need nanosecond
	61	* resolution, and we don't need to waste time with a big divide when
	62	* 2^30ns == 1.074s.
	63	*/
a3b13c23	64	static unsigned long get_timestamp(int this_cpu)
966812dc	65	{
82a1fcb9	66	return cpu_clock(this_cpu) >> 30LL; /* 2^30 ~= 10^9 */
966812dc JF	67	}
966812dc JF	68
8c2238ea	69	static void __touch_softlockup_watchdog(void)
8446f1d3	70	{
a3b13c23 IM	71	int this_cpu = raw_smp_processor_id();
	72
	73	__raw_get_cpu_var(touch_timestamp) = get_timestamp(this_cpu);
8446f1d3	74	}
8c2238ea JW	75
	76	void touch_softlockup_watchdog(void)
	77	{
	78	__raw_get_cpu_var(touch_timestamp) = 0;
	79	}
8446f1d3 IM	80	EXPORT_SYMBOL(touch_softlockup_watchdog);
8446f1d3 IM	81
04c9167f JF	82	void touch_all_softlockup_watchdogs(void)
	83	{
	84	int cpu;
	85
	86	/* Cause each CPU to re-update its timestamp rather than complain */
	87	for_each_online_cpu(cpu)
	88	per_cpu(touch_timestamp, cpu) = 0;
	89	}
	90	EXPORT_SYMBOL(touch_all_softlockup_watchdogs);
	91
baf48f65 MSB	92	int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
	93	struct file filp, void __user buffer,
	94	size_t lenp, loff_t ppos)
	95	{
	96	touch_all_softlockup_watchdogs();
	97	return proc_dointvec_minmax(table, write, filp, buffer, lenp, ppos);
	98	}
	99
8446f1d3 IM	100	/*
	101	* This callback runs from the timer interrupt, and checks
	102	* whether the watchdog thread has hung or not:
	103	*/
6687a97d	104	void softlockup_tick(void)
8446f1d3 IM	105	{
8446f1d3 IM	106	int this_cpu = smp_processor_id();
6687a97d	107	unsigned long touch_timestamp = per_cpu(touch_timestamp, this_cpu);
966812dc	108	unsigned long print_timestamp;
43581a10	109	struct pt_regs *regs = get_irq_regs();
966812dc	110	unsigned long now;
8446f1d3	111
9383d967 DS	112	/* Is detection switched off? */
	113	if (!per_cpu(watchdog_task, this_cpu) \|\| softlockup_thresh <= 0) {
	114	/* Be sure we don't false trigger if switched back on */
	115	if (touch_timestamp)
	116	per_cpu(touch_timestamp, this_cpu) = 0;
	117	return;
	118	}
	119
04c9167f	120	if (touch_timestamp == 0) {
8c2238ea	121	__touch_softlockup_watchdog();
966812dc	122	return;
04c9167f	123	}
966812dc JF	124
	125	print_timestamp = per_cpu(print_timestamp, this_cpu);
	126
	127	/* report at most once a second */
3e2f69fd	128	if (print_timestamp == touch_timestamp \|\| did_panic)
8446f1d3 IM	129	return;
8446f1d3 IM	130
6687a97d IM	131	/* do not print during early bootup: */
6687a97d IM	132	if (unlikely(system_state != SYSTEM_RUNNING)) {
8c2238ea	133	__touch_softlockup_watchdog();
8446f1d3	134	return;
6687a97d	135	}
8446f1d3	136
a3b13c23	137	now = get_timestamp(this_cpu);
966812dc	138
dd7a1e56 JW	139	/*
	140	* Wake up the high-prio watchdog task twice per
	141	* threshold timespan.
	142	*/
	143	if (now > touch_timestamp + softlockup_thresh/2)
ed50d6cb PZ	144	wake_up_process(per_cpu(watchdog_task, this_cpu));
ed50d6cb PZ	145
82a1fcb9	146	/* Warn about unreasonable delays: */
c4f3b63f	147	if (now <= (touch_timestamp + softlockup_thresh))
43581a10	148	return;
8446f1d3	149
43581a10 IM	150	per_cpu(print_timestamp, this_cpu) = touch_timestamp;
	151
	152	spin_lock(&print_lock);
c4f3b63f RT	153	printk(KERN_ERR "BUG: soft lockup - CPU#%d stuck for %lus! [%s:%d]\n",
c4f3b63f RT	154	this_cpu, now - touch_timestamp,
ba25f9dc	155	current->comm, task_pid_nr(current));
688c9175	156	print_modules();
8d5be7f4	157	print_irqtrace_events(current);
43581a10 IM	158	if (regs)
	159	show_regs(regs);
	160	else
6687a97d	161	dump_stack();
43581a10	162	spin_unlock(&print_lock);
9c44bc03 IM	163
	164	if (softlockup_panic)
	165	panic("softlockup: hung tasks");
8446f1d3 IM	166	}
8446f1d3 IM	167
82a1fcb9 IM	168	/*
	169	* Have a reasonable limit on the number of tasks checked:
	170	*/
90739081	171	unsigned long __read_mostly sysctl_hung_task_check_count = 1024;
82a1fcb9 IM	172
	173	/*
	174	* Zero means infinite timeout - no checking done:
	175	*/
fb822db4	176	unsigned long __read_mostly sysctl_hung_task_timeout_secs = 480;
82a1fcb9	177
90739081	178	unsigned long __read_mostly sysctl_hung_task_warnings = 10;
82a1fcb9 IM	179
	180	/*
	181	* Only do the hung-tasks check on one CPU:
	182	*/
	183	static int check_cpu __read_mostly = -1;
	184
	185	static void check_hung_task(struct task_struct *t, unsigned long now)
	186	{
	187	unsigned long switch_count = t->nvcsw + t->nivcsw;
	188
	189	if (t->flags & PF_FROZEN)
	190	return;
	191
	192	if (switch_count != t->last_switch_count \|\| !t->last_switch_timestamp) {
	193	t->last_switch_count = switch_count;
	194	t->last_switch_timestamp = now;
	195	return;
	196	}
	197	if ((long)(now - t->last_switch_timestamp) <
	198	sysctl_hung_task_timeout_secs)
	199	return;
20195546	200	if (!sysctl_hung_task_warnings)
82a1fcb9 IM	201	return;
	202	sysctl_hung_task_warnings--;
	203
	204	/*
	205	* Ok, the task did not get scheduled for more than 2 minutes,
	206	* complain:
	207	*/
	208	printk(KERN_ERR "INFO: task %s:%d blocked for more than "
	209	"%ld seconds.\n", t->comm, t->pid,
	210	sysctl_hung_task_timeout_secs);
	211	printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
	212	" disables this message.\n");
	213	sched_show_task(t);
	214	__debug_show_held_locks(t);
	215
	216	t->last_switch_timestamp = now;
	217	touch_nmi_watchdog();
9c44bc03 IM	218
	219	if (softlockup_panic)
	220	panic("softlockup: blocked tasks");
82a1fcb9 IM	221	}
	222
	223	/*
	224	* Check whether a TASK_UNINTERRUPTIBLE does not get woken up for
	225	* a really long time (120 seconds). If that happens, print out
	226	* a warning.
	227	*/
	228	static void check_hung_uninterruptible_tasks(int this_cpu)
	229	{
	230	int max_count = sysctl_hung_task_check_count;
	231	unsigned long now = get_timestamp(this_cpu);
	232	struct task_struct g, t;
	233
	234	/*
	235	* If the system crashed already then all bets are off,
	236	* do not report extra hung tasks:
	237	*/
25ddbb18	238	if (test_taint(TAINT_DIE) \|\| did_panic)
82a1fcb9 IM	239	return;
	240
	241	read_lock(&tasklist_lock);
	242	do_each_thread(g, t) {
	243	if (!--max_count)
ed50d6cb	244	goto unlock;
cbaed698 ON	245	/* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
cbaed698 ON	246	if (t->state == TASK_UNINTERRUPTIBLE)
82a1fcb9 IM	247	check_hung_task(t, now);
82a1fcb9 IM	248	} while_each_thread(g, t);
ed50d6cb	249	unlock:
82a1fcb9 IM	250	read_unlock(&tasklist_lock);
	251	}
	252
8446f1d3 IM	253	/*
	254	* The watchdog thread - runs every second and touches the timestamp.
	255	*/
a5f2ce3c	256	static int watchdog(void *__bind_cpu)
8446f1d3	257	{
02fb6149	258	struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
82a1fcb9	259	int this_cpu = (long)__bind_cpu;
8446f1d3 IM	260
8446f1d3 IM	261	sched_setscheduler(current, SCHED_FIFO, &param);
8446f1d3	262
966812dc	263	/* initialize timestamp */
8c2238ea	264	__touch_softlockup_watchdog();
966812dc	265
7be2a03e	266	set_current_state(TASK_INTERRUPTIBLE);
8446f1d3	267	/*
6687a97d	268	* Run briefly once per second to reset the softlockup timestamp.
82a1fcb9	269	* If this gets delayed for more than 60 seconds then the
6687a97d	270	* debug-printout triggers in softlockup_tick().
8446f1d3 IM	271	*/
8446f1d3 IM	272	while (!kthread_should_stop()) {
8c2238ea	273	__touch_softlockup_watchdog();
ed50d6cb PZ	274	schedule();
	275
	276	if (kthread_should_stop())
	277	break;
82a1fcb9	278
7be2a03e DA	279	if (this_cpu == check_cpu) {
	280	if (sysctl_hung_task_timeout_secs)
	281	check_hung_uninterruptible_tasks(this_cpu);
	282	}
ed50d6cb	283
7be2a03e	284	set_current_state(TASK_INTERRUPTIBLE);
8446f1d3	285	}
7be2a03e	286	__set_current_state(TASK_RUNNING);
8446f1d3 IM	287
	288	return 0;
	289	}
	290
	291	/*
	292	* Create/destroy watchdog threads as CPUs come and go:
	293	*/
8c78f307	294	static int __cpuinit
8446f1d3 IM	295	cpu_callback(struct notifier_block nfb, unsigned long action, void hcpu)
	296	{
	297	int hotcpu = (unsigned long)hcpu;
	298	struct task_struct *p;
	299
	300	switch (action) {
	301	case CPU_UP_PREPARE:
8bb78442	302	case CPU_UP_PREPARE_FROZEN:
8446f1d3 IM	303	BUG_ON(per_cpu(watchdog_task, hotcpu));
	304	p = kthread_create(watchdog, hcpu, "watchdog/%d", hotcpu);
	305	if (IS_ERR(p)) {
a5f2ce3c	306	printk(KERN_ERR "watchdog for %i failed\n", hotcpu);
8446f1d3 IM	307	return NOTIFY_BAD;
8446f1d3 IM	308	}
a5f2ce3c IM	309	per_cpu(touch_timestamp, hotcpu) = 0;
a5f2ce3c IM	310	per_cpu(watchdog_task, hotcpu) = p;
8446f1d3	311	kthread_bind(p, hotcpu);
a5f2ce3c	312	break;
8446f1d3	313	case CPU_ONLINE:
8bb78442	314	case CPU_ONLINE_FROZEN:
f1fc057c	315	check_cpu = cpumask_any(cpu_online_mask);
8446f1d3 IM	316	wake_up_process(per_cpu(watchdog_task, hotcpu));
	317	break;
	318	#ifdef CONFIG_HOTPLUG_CPU
82a1fcb9 IM	319	case CPU_DOWN_PREPARE:
	320	case CPU_DOWN_PREPARE_FROZEN:
	321	if (hotcpu == check_cpu) {
41c7bb95 RR	322	/* Pick any other online cpu. */
41c7bb95 RR	323	check_cpu = cpumask_any_but(cpu_online_mask, hotcpu);
82a1fcb9 IM	324	}
82a1fcb9 IM	325	break;
ed50d6cb PZ	326
	327	case CPU_UP_CANCELED:
	328	case CPU_UP_CANCELED_FROZEN:
	329	if (!per_cpu(watchdog_task, hotcpu))
	330	break;
	331	/* Unbind so it can run. Fall thru. */
	332	kthread_bind(per_cpu(watchdog_task, hotcpu),
f1fc057c	333	cpumask_any(cpu_online_mask));
8446f1d3	334	case CPU_DEAD:
8bb78442	335	case CPU_DEAD_FROZEN:
8446f1d3 IM	336	p = per_cpu(watchdog_task, hotcpu);
	337	per_cpu(watchdog_task, hotcpu) = NULL;
	338	kthread_stop(p);
	339	break;
	340	#endif /* CONFIG_HOTPLUG_CPU */
a5f2ce3c	341	}
8446f1d3 IM	342	return NOTIFY_OK;
	343	}
	344
8c78f307	345	static struct notifier_block __cpuinitdata cpu_nfb = {
8446f1d3 IM	346	.notifier_call = cpu_callback
	347	};
	348
7babe8db EGM	349	static int __initdata nosoftlockup;
	350
	351	static int __init nosoftlockup_setup(char *str)
	352	{
	353	nosoftlockup = 1;
	354	return 1;
	355	}
	356	__setup("nosoftlockup", nosoftlockup_setup);
	357
	358	static int __init spawn_softlockup_task(void)
8446f1d3 IM	359	{
8446f1d3 IM	360	void cpu = (void )(long)smp_processor_id();
7babe8db	361	int err;
8446f1d3	362
7babe8db EGM	363	if (nosoftlockup)
	364	return 0;
	365
	366	err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu);
	367	if (err == NOTIFY_BAD) {
	368	BUG();
	369	return 1;
	370	}
8446f1d3 IM	371	cpu_callback(&cpu_nfb, CPU_ONLINE, cpu);
	372	register_cpu_notifier(&cpu_nfb);
	373
e041c683	374	atomic_notifier_chain_register(&panic_notifier_list, &panic_block);
7babe8db EGM	375
7babe8db EGM	376	return 0;
8446f1d3	377	}
7babe8db	378	early_initcall(spawn_softlockup_task);