hung_task.c 5.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/*
 * 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>
16
#include <linux/export.h>
17
#include <linux/sysctl.h>
18
#include <linux/utsname.h>
19 20

/*
21
 * The number of tasks checked:
22
 */
23
int __read_mostly sysctl_hung_task_check_count = PID_MAX_LIMIT;
24 25 26 27 28 29 30 31 32

/*
 * 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_BATCHING 1024
33 34 35 36

/*
 * Zero means infinite timeout - no checking done:
 */
37
unsigned long __read_mostly sysctl_hung_task_timeout_secs = CONFIG_DEFAULT_HUNG_TASK_TIMEOUT;
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71

unsigned long __read_mostly sysctl_hung_task_warnings = 10;

static int __read_mostly did_panic;

static struct task_struct *watchdog_task;

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

static int __init hung_task_panic_setup(char *str)
{
	sysctl_hung_task_panic = simple_strtoul(str, NULL, 0);

	return 1;
}
__setup("hung_task_panic=", hung_task_panic_setup);

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,
};

72
static void check_hung_task(struct task_struct *t, unsigned long timeout)
73 74 75
{
	unsigned long switch_count = t->nvcsw + t->nivcsw;

76 77
	/*
	 * Ensure the task is not frozen.
78
	 * Also, skip vfork and any other user process that freezer should skip.
79
	 */
80 81 82 83 84 85 86 87 88
	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))
89 90
		return;

91
	if (switch_count != t->last_switch_count) {
92 93 94 95 96 97 98 99 100 101 102
		t->last_switch_count = switch_count;
		return;
	}
	if (!sysctl_hung_task_warnings)
		return;
	sysctl_hung_task_warnings--;

	/*
	 * Ok, the task did not get scheduled for more than 2 minutes,
	 * complain:
	 */
103 104 105 106 107 108 109 110
	pr_err("INFO: task %s:%d blocked for more than %ld seconds.\n",
		t->comm, t->pid, timeout);
	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");
111
	sched_show_task(t);
112
	debug_show_held_locks(t);
113 114 115

	touch_nmi_watchdog();

116 117
	if (sysctl_hung_task_panic) {
		trigger_all_cpu_backtrace();
118
		panic("hung_task: blocked tasks");
119
	}
120 121
}

122 123 124 125 126
/*
 * 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
127
 * to exit the grace period. For classic RCU, a reschedule is required.
128
 */
129
static bool rcu_lock_break(struct task_struct *g, struct task_struct *t)
130
{
131 132
	bool can_cont;

133 134 135 136 137
	get_task_struct(g);
	get_task_struct(t);
	rcu_read_unlock();
	cond_resched();
	rcu_read_lock();
138
	can_cont = pid_alive(g) && pid_alive(t);
139 140
	put_task_struct(t);
	put_task_struct(g);
141 142

	return can_cont;
143 144
}

145 146 147 148 149
/*
 * Check whether a TASK_UNINTERRUPTIBLE does not get woken up for
 * a really long time (120 seconds). If that happens, print out
 * a warning.
 */
150
static void check_hung_uninterruptible_tasks(unsigned long timeout)
151 152
{
	int max_count = sysctl_hung_task_check_count;
153
	int batch_count = HUNG_TASK_BATCHING;
154 155 156 157 158 159 160 161 162
	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;

163
	rcu_read_lock();
164
	do_each_thread(g, t) {
165
		if (!max_count--)
166
			goto unlock;
167 168
		if (!--batch_count) {
			batch_count = HUNG_TASK_BATCHING;
169
			if (!rcu_lock_break(g, t))
170 171
				goto unlock;
		}
172 173
		/* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
		if (t->state == TASK_UNINTERRUPTIBLE)
174
			check_hung_task(t, timeout);
175 176
	} while_each_thread(g, t);
 unlock:
177
	rcu_read_unlock();
178 179
}

180
static unsigned long timeout_jiffies(unsigned long timeout)
181 182
{
	/* timeout of 0 will disable the watchdog */
183
	return timeout ? timeout * HZ : MAX_SCHEDULE_TIMEOUT;
184 185 186 187 188 189
}

/*
 * Process updating of timeout sysctl
 */
int proc_dohung_task_timeout_secs(struct ctl_table *table, int write,
190
				  void __user *buffer,
191 192 193 194
				  size_t *lenp, loff_t *ppos)
{
	int ret;

195
	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213

	if (ret || !write)
		goto out;

	wake_up_process(watchdog_task);

 out:
	return ret;
}

/*
 * kthread which checks for tasks stuck in D state
 */
static int watchdog(void *dummy)
{
	set_user_nice(current, 0);

	for ( ; ; ) {
214
		unsigned long timeout = sysctl_hung_task_timeout_secs;
215

216 217
		while (schedule_timeout_interruptible(timeout_jiffies(timeout)))
			timeout = sysctl_hung_task_timeout_secs;
218

219
		check_hung_uninterruptible_tasks(timeout);
220 221 222 223 224 225 226 227 228 229 230 231 232 233
	}

	return 0;
}

static int __init hung_task_init(void)
{
	atomic_notifier_chain_register(&panic_notifier_list, &panic_block);
	watchdog_task = kthread_run(watchdog, NULL, "khungtaskd");

	return 0;
}

module_init(hung_task_init);