2010-05-07 21:11:44 +00:00
|
|
|
/*
|
|
|
|
* Detect hard and soft lockups on a system
|
|
|
|
*
|
|
|
|
* started by Don Zickus, Copyright (C) 2010 Red Hat, Inc.
|
|
|
|
*
|
2012-02-09 22:42:22 +00:00
|
|
|
* Note: Most of this code is borrowed heavily from the original softlockup
|
|
|
|
* detector, so thanks to Ingo for the initial implementation.
|
|
|
|
* Some chunks also taken from the old x86-specific nmi watchdog code, thanks
|
2010-05-07 21:11:44 +00:00
|
|
|
* to those contributors as well.
|
|
|
|
*/
|
|
|
|
|
2012-03-23 22:01:55 +00:00
|
|
|
#define pr_fmt(fmt) "NMI watchdog: " fmt
|
|
|
|
|
2010-05-07 21:11:44 +00:00
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/cpu.h>
|
|
|
|
#include <linux/nmi.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/sysctl.h>
|
2012-07-16 10:42:38 +00:00
|
|
|
#include <linux/smpboot.h>
|
2013-02-07 15:47:07 +00:00
|
|
|
#include <linux/sched/rt.h>
|
2017-02-01 17:07:51 +00:00
|
|
|
#include <uapi/linux/sched/types.h>
|
2015-06-24 23:55:45 +00:00
|
|
|
#include <linux/tick.h>
|
workqueue: implement lockup detector
Workqueue stalls can happen from a variety of usage bugs such as
missing WQ_MEM_RECLAIM flag or concurrency managed work item
indefinitely staying RUNNING. These stalls can be extremely difficult
to hunt down because the usual warning mechanisms can't detect
workqueue stalls and the internal state is pretty opaque.
To alleviate the situation, this patch implements workqueue lockup
detector. It periodically monitors all worker_pools periodically and,
if any pool failed to make forward progress longer than the threshold
duration, triggers warning and dumps workqueue state as follows.
BUG: workqueue lockup - pool cpus=0 node=0 flags=0x0 nice=0 stuck for 31s!
Showing busy workqueues and worker pools:
workqueue events: flags=0x0
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=17/256
pending: monkey_wrench_fn, e1000_watchdog, cache_reap, vmstat_shepherd, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, cgroup_release_agent
workqueue events_power_efficient: flags=0x80
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=2/256
pending: check_lifetime, neigh_periodic_work
workqueue cgroup_pidlist_destroy: flags=0x0
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1
pending: cgroup_pidlist_destroy_work_fn
...
The detection mechanism is controller through kernel parameter
workqueue.watchdog_thresh and can be updated at runtime through the
sysfs module parameter file.
v2: Decoupled from softlockup control knobs.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Don Zickus <dzickus@redhat.com>
Cc: Ulrich Obergfell <uobergfe@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Chris Mason <clm@fb.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
2015-12-08 16:28:04 +00:00
|
|
|
#include <linux/workqueue.h>
|
2017-02-01 15:36:40 +00:00
|
|
|
#include <linux/sched/clock.h>
|
2017-02-08 17:51:35 +00:00
|
|
|
#include <linux/sched/debug.h>
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
#include <asm/irq_regs.h>
|
2012-03-10 19:37:28 +00:00
|
|
|
#include <linux/kvm_para.h>
|
2015-09-04 22:45:15 +00:00
|
|
|
#include <linux/kthread.h>
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2015-05-18 09:31:50 +00:00
|
|
|
static DEFINE_MUTEX(watchdog_proc_mutex);
|
|
|
|
|
2016-12-14 23:06:21 +00:00
|
|
|
#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR)
|
|
|
|
unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED|NMI_WATCHDOG_ENABLED;
|
2015-04-14 22:43:55 +00:00
|
|
|
#else
|
2016-12-14 23:06:21 +00:00
|
|
|
unsigned long __read_mostly watchdog_enabled = SOFT_WATCHDOG_ENABLED;
|
2015-04-14 22:43:55 +00:00
|
|
|
#endif
|
|
|
|
int __read_mostly nmi_watchdog_enabled;
|
|
|
|
int __read_mostly soft_watchdog_enabled;
|
|
|
|
int __read_mostly watchdog_user_enabled;
|
2011-05-23 05:10:23 +00:00
|
|
|
int __read_mostly watchdog_thresh = 10;
|
2015-04-14 22:43:55 +00:00
|
|
|
|
2014-06-23 20:22:05 +00:00
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
int __read_mostly sysctl_softlockup_all_cpu_backtrace;
|
2015-11-06 02:44:41 +00:00
|
|
|
int __read_mostly sysctl_hardlockup_all_cpu_backtrace;
|
2014-06-23 20:22:05 +00:00
|
|
|
#endif
|
2015-06-24 23:55:45 +00:00
|
|
|
static struct cpumask watchdog_cpumask __read_mostly;
|
|
|
|
unsigned long *watchdog_cpumask_bits = cpumask_bits(&watchdog_cpumask);
|
|
|
|
|
|
|
|
/* Helper for online, unparked cpus. */
|
|
|
|
#define for_each_watchdog_cpu(cpu) \
|
|
|
|
for_each_cpu_and((cpu), cpu_online_mask, &watchdog_cpumask)
|
2014-06-23 20:22:05 +00:00
|
|
|
|
2017-01-24 23:17:53 +00:00
|
|
|
atomic_t watchdog_park_in_progress = ATOMIC_INIT(0);
|
|
|
|
|
2015-09-04 22:45:28 +00:00
|
|
|
/*
|
|
|
|
* The 'watchdog_running' variable is set to 1 when the watchdog threads
|
|
|
|
* are registered/started and is set to 0 when the watchdog threads are
|
|
|
|
* unregistered/stopped, so it is an indicator whether the threads exist.
|
|
|
|
*/
|
2013-05-19 18:45:15 +00:00
|
|
|
static int __read_mostly watchdog_running;
|
2015-09-04 22:45:28 +00:00
|
|
|
/*
|
|
|
|
* If a subsystem has a need to deactivate the watchdog temporarily, it
|
|
|
|
* can use the suspend/resume interface to achieve this. The content of
|
|
|
|
* the 'watchdog_suspended' variable reflects this state. Existing threads
|
|
|
|
* are parked/unparked by the lockup_detector_{suspend|resume} functions
|
|
|
|
* (see comment blocks pertaining to those functions for further details).
|
|
|
|
*
|
|
|
|
* 'watchdog_suspended' also prevents threads from being registered/started
|
|
|
|
* or unregistered/stopped via parameters in /proc/sys/kernel, so the state
|
|
|
|
* of 'watchdog_running' cannot change while the watchdog is deactivated
|
|
|
|
* temporarily (see related code in 'proc' handlers).
|
|
|
|
*/
|
|
|
|
static int __read_mostly watchdog_suspended;
|
|
|
|
|
2012-12-17 23:59:50 +00:00
|
|
|
static u64 __read_mostly sample_period;
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts);
|
|
|
|
static DEFINE_PER_CPU(struct task_struct *, softlockup_watchdog);
|
|
|
|
static DEFINE_PER_CPU(struct hrtimer, watchdog_hrtimer);
|
|
|
|
static DEFINE_PER_CPU(bool, softlockup_touch_sync);
|
|
|
|
static DEFINE_PER_CPU(bool, soft_watchdog_warn);
|
2012-07-16 10:42:38 +00:00
|
|
|
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts);
|
|
|
|
static DEFINE_PER_CPU(unsigned long, soft_lockup_hrtimer_cnt);
|
2014-10-09 22:25:17 +00:00
|
|
|
static DEFINE_PER_CPU(struct task_struct *, softlockup_task_ptr_saved);
|
2010-05-07 21:11:44 +00:00
|
|
|
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved);
|
2014-06-23 20:22:05 +00:00
|
|
|
static unsigned long soft_lockup_nmi_warn;
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
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 __init nowatchdog_setup(char *str)
|
|
|
|
{
|
2015-04-14 22:44:13 +00:00
|
|
|
watchdog_enabled = 0;
|
2010-05-07 21:11:44 +00:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
__setup("nowatchdog", nowatchdog_setup);
|
|
|
|
|
|
|
|
static int __init nosoftlockup_setup(char *str)
|
|
|
|
{
|
2015-04-14 22:44:13 +00:00
|
|
|
watchdog_enabled &= ~SOFT_WATCHDOG_ENABLED;
|
2010-05-07 21:11:44 +00:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
__setup("nosoftlockup", nosoftlockup_setup);
|
2015-04-14 22:44:13 +00:00
|
|
|
|
2014-06-23 20:22:05 +00:00
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
static int __init softlockup_all_cpu_backtrace_setup(char *str)
|
|
|
|
{
|
|
|
|
sysctl_softlockup_all_cpu_backtrace =
|
|
|
|
!!simple_strtol(str, NULL, 0);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
__setup("softlockup_all_cpu_backtrace=", softlockup_all_cpu_backtrace_setup);
|
2015-11-06 02:44:41 +00:00
|
|
|
static int __init hardlockup_all_cpu_backtrace_setup(char *str)
|
|
|
|
{
|
|
|
|
sysctl_hardlockup_all_cpu_backtrace =
|
|
|
|
!!simple_strtol(str, NULL, 0);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
__setup("hardlockup_all_cpu_backtrace=", hardlockup_all_cpu_backtrace_setup);
|
2014-06-23 20:22:05 +00:00
|
|
|
#endif
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2011-05-23 05:10:23 +00:00
|
|
|
/*
|
|
|
|
* Hard-lockup warnings should be triggered after just a few seconds. Soft-
|
|
|
|
* lockups can have false positives under extreme conditions. So we generally
|
|
|
|
* want a higher threshold for soft lockups than for hard lockups. So we couple
|
|
|
|
* the thresholds with a factor: we make the soft threshold twice the amount of
|
|
|
|
* time the hard threshold is.
|
|
|
|
*/
|
2011-05-24 03:43:18 +00:00
|
|
|
static int get_softlockup_thresh(void)
|
2011-05-23 05:10:23 +00:00
|
|
|
{
|
|
|
|
return watchdog_thresh * 2;
|
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
2012-12-27 02:49:44 +00:00
|
|
|
static unsigned long get_timestamp(void)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2015-02-12 23:01:24 +00:00
|
|
|
return running_clock() >> 30LL; /* 2^30 ~= 10^9 */
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
2012-12-17 23:59:50 +00:00
|
|
|
static void set_sample_period(void)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
|
|
|
/*
|
2011-05-23 05:10:22 +00:00
|
|
|
* convert watchdog_thresh from seconds to ns
|
2012-02-09 22:42:22 +00:00
|
|
|
* the divide by 5 is to give hrtimer several chances (two
|
|
|
|
* or three with the current relation between the soft
|
|
|
|
* and hard thresholds) to increment before the
|
|
|
|
* hardlockup detector generates a warning
|
2010-05-07 21:11:44 +00:00
|
|
|
*/
|
2012-12-17 23:59:50 +00:00
|
|
|
sample_period = get_softlockup_thresh() * ((u64)NSEC_PER_SEC / 5);
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Commands for resetting the watchdog */
|
|
|
|
static void __touch_watchdog(void)
|
|
|
|
{
|
2012-12-27 02:49:44 +00:00
|
|
|
__this_cpu_write(watchdog_touch_ts, get_timestamp());
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
2015-12-08 16:28:04 +00:00
|
|
|
/**
|
|
|
|
* touch_softlockup_watchdog_sched - touch watchdog on scheduler stalls
|
|
|
|
*
|
|
|
|
* Call when the scheduler may have stalled for legitimate reasons
|
|
|
|
* preventing the watchdog task from executing - e.g. the scheduler
|
|
|
|
* entering idle state. This should only be used for scheduler events.
|
|
|
|
* Use touch_softlockup_watchdog() for everything else.
|
|
|
|
*/
|
|
|
|
void touch_softlockup_watchdog_sched(void)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2014-04-18 22:07:12 +00:00
|
|
|
/*
|
|
|
|
* Preemption can be enabled. It doesn't matter which CPU's timestamp
|
|
|
|
* gets zeroed here, so use the raw_ operation.
|
|
|
|
*/
|
|
|
|
raw_cpu_write(watchdog_touch_ts, 0);
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
2015-12-08 16:28:04 +00:00
|
|
|
|
|
|
|
void touch_softlockup_watchdog(void)
|
|
|
|
{
|
|
|
|
touch_softlockup_watchdog_sched();
|
workqueue: implement lockup detector
Workqueue stalls can happen from a variety of usage bugs such as
missing WQ_MEM_RECLAIM flag or concurrency managed work item
indefinitely staying RUNNING. These stalls can be extremely difficult
to hunt down because the usual warning mechanisms can't detect
workqueue stalls and the internal state is pretty opaque.
To alleviate the situation, this patch implements workqueue lockup
detector. It periodically monitors all worker_pools periodically and,
if any pool failed to make forward progress longer than the threshold
duration, triggers warning and dumps workqueue state as follows.
BUG: workqueue lockup - pool cpus=0 node=0 flags=0x0 nice=0 stuck for 31s!
Showing busy workqueues and worker pools:
workqueue events: flags=0x0
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=17/256
pending: monkey_wrench_fn, e1000_watchdog, cache_reap, vmstat_shepherd, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, cgroup_release_agent
workqueue events_power_efficient: flags=0x80
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=2/256
pending: check_lifetime, neigh_periodic_work
workqueue cgroup_pidlist_destroy: flags=0x0
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1
pending: cgroup_pidlist_destroy_work_fn
...
The detection mechanism is controller through kernel parameter
workqueue.watchdog_thresh and can be updated at runtime through the
sysfs module parameter file.
v2: Decoupled from softlockup control knobs.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Don Zickus <dzickus@redhat.com>
Cc: Ulrich Obergfell <uobergfe@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Chris Mason <clm@fb.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
2015-12-08 16:28:04 +00:00
|
|
|
wq_watchdog_touch(raw_smp_processor_id());
|
2015-12-08 16:28:04 +00:00
|
|
|
}
|
2010-05-13 06:53:33 +00:00
|
|
|
EXPORT_SYMBOL(touch_softlockup_watchdog);
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2010-05-07 21:11:45 +00:00
|
|
|
void touch_all_softlockup_watchdogs(void)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
|
|
|
int cpu;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* this is done lockless
|
|
|
|
* do we care if a 0 races with a timestamp?
|
|
|
|
* all it means is the softlock check starts one cycle later
|
|
|
|
*/
|
2015-06-24 23:55:45 +00:00
|
|
|
for_each_watchdog_cpu(cpu)
|
2010-05-07 21:11:44 +00:00
|
|
|
per_cpu(watchdog_touch_ts, cpu) = 0;
|
workqueue: implement lockup detector
Workqueue stalls can happen from a variety of usage bugs such as
missing WQ_MEM_RECLAIM flag or concurrency managed work item
indefinitely staying RUNNING. These stalls can be extremely difficult
to hunt down because the usual warning mechanisms can't detect
workqueue stalls and the internal state is pretty opaque.
To alleviate the situation, this patch implements workqueue lockup
detector. It periodically monitors all worker_pools periodically and,
if any pool failed to make forward progress longer than the threshold
duration, triggers warning and dumps workqueue state as follows.
BUG: workqueue lockup - pool cpus=0 node=0 flags=0x0 nice=0 stuck for 31s!
Showing busy workqueues and worker pools:
workqueue events: flags=0x0
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=17/256
pending: monkey_wrench_fn, e1000_watchdog, cache_reap, vmstat_shepherd, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, release_one_tty, cgroup_release_agent
workqueue events_power_efficient: flags=0x80
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=2/256
pending: check_lifetime, neigh_periodic_work
workqueue cgroup_pidlist_destroy: flags=0x0
pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/1
pending: cgroup_pidlist_destroy_work_fn
...
The detection mechanism is controller through kernel parameter
workqueue.watchdog_thresh and can be updated at runtime through the
sysfs module parameter file.
v2: Decoupled from softlockup control knobs.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Don Zickus <dzickus@redhat.com>
Cc: Ulrich Obergfell <uobergfe@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Chris Mason <clm@fb.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
2015-12-08 16:28:04 +00:00
|
|
|
wq_watchdog_touch(-1);
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void touch_softlockup_watchdog_sync(void)
|
|
|
|
{
|
2014-08-17 17:30:34 +00:00
|
|
|
__this_cpu_write(softlockup_touch_sync, true);
|
|
|
|
__this_cpu_write(watchdog_touch_ts, 0);
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* watchdog detector functions */
|
2016-12-14 23:06:21 +00:00
|
|
|
bool is_hardlockup(void)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2010-12-08 15:22:55 +00:00
|
|
|
unsigned long hrint = __this_cpu_read(hrtimer_interrupts);
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2010-12-08 15:22:55 +00:00
|
|
|
if (__this_cpu_read(hrtimer_interrupts_saved) == hrint)
|
2015-11-06 02:44:24 +00:00
|
|
|
return true;
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2010-12-08 15:22:55 +00:00
|
|
|
__this_cpu_write(hrtimer_interrupts_saved, hrint);
|
2015-11-06 02:44:24 +00:00
|
|
|
return false;
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
2010-05-17 22:06:04 +00:00
|
|
|
static int is_softlockup(unsigned long touch_ts)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2012-12-27 02:49:44 +00:00
|
|
|
unsigned long now = get_timestamp();
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2015-11-06 02:44:56 +00:00
|
|
|
if ((watchdog_enabled & SOFT_WATCHDOG_ENABLED) && watchdog_thresh){
|
2015-04-14 22:44:13 +00:00
|
|
|
/* Warn about unreasonable delays. */
|
|
|
|
if (time_after(now, touch_ts + get_softlockup_thresh()))
|
|
|
|
return now - touch_ts;
|
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void watchdog_interrupt_count(void)
|
|
|
|
{
|
2010-12-08 15:22:55 +00:00
|
|
|
__this_cpu_inc(hrtimer_interrupts);
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
2012-07-16 10:42:38 +00:00
|
|
|
|
2016-12-14 23:06:24 +00:00
|
|
|
/*
|
|
|
|
* These two functions are mostly architecture specific
|
|
|
|
* defining them as weak here.
|
|
|
|
*/
|
|
|
|
int __weak watchdog_nmi_enable(unsigned int cpu)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
void __weak watchdog_nmi_disable(unsigned int cpu)
|
|
|
|
{
|
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2015-11-06 02:44:30 +00:00
|
|
|
static int watchdog_enable_all_cpus(void);
|
|
|
|
static void watchdog_disable_all_cpus(void);
|
|
|
|
|
2010-05-07 21:11:44 +00:00
|
|
|
/* watchdog kicker functions */
|
|
|
|
static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer)
|
|
|
|
{
|
2010-12-08 15:22:55 +00:00
|
|
|
unsigned long touch_ts = __this_cpu_read(watchdog_touch_ts);
|
2010-05-07 21:11:44 +00:00
|
|
|
struct pt_regs *regs = get_irq_regs();
|
|
|
|
int duration;
|
2014-06-23 20:22:05 +00:00
|
|
|
int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace;
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2017-01-24 23:17:53 +00:00
|
|
|
if (atomic_read(&watchdog_park_in_progress) != 0)
|
|
|
|
return HRTIMER_NORESTART;
|
|
|
|
|
2010-05-07 21:11:44 +00:00
|
|
|
/* kick the hardlockup detector */
|
|
|
|
watchdog_interrupt_count();
|
|
|
|
|
|
|
|
/* kick the softlockup detector */
|
2010-12-08 15:22:55 +00:00
|
|
|
wake_up_process(__this_cpu_read(softlockup_watchdog));
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
/* .. and repeat */
|
2012-12-17 23:59:50 +00:00
|
|
|
hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period));
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
if (touch_ts == 0) {
|
2010-12-08 15:22:55 +00:00
|
|
|
if (unlikely(__this_cpu_read(softlockup_touch_sync))) {
|
2010-05-07 21:11:44 +00:00
|
|
|
/*
|
|
|
|
* If the time stamp was touched atomically
|
|
|
|
* make sure the scheduler tick is up to date.
|
|
|
|
*/
|
2010-12-08 15:22:55 +00:00
|
|
|
__this_cpu_write(softlockup_touch_sync, false);
|
2010-05-07 21:11:44 +00:00
|
|
|
sched_clock_tick();
|
|
|
|
}
|
2012-03-10 19:37:28 +00:00
|
|
|
|
|
|
|
/* Clear the guest paused flag on watchdog reset */
|
|
|
|
kvm_check_and_clear_guest_paused();
|
2010-05-07 21:11:44 +00:00
|
|
|
__touch_watchdog();
|
|
|
|
return HRTIMER_RESTART;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* check for a softlockup
|
|
|
|
* This is done by making sure a high priority task is
|
|
|
|
* being scheduled. The task touches the watchdog to
|
|
|
|
* indicate it is getting cpu time. If it hasn't then
|
|
|
|
* this is a good indication some task is hogging the cpu
|
|
|
|
*/
|
2010-05-17 22:06:04 +00:00
|
|
|
duration = is_softlockup(touch_ts);
|
2010-05-07 21:11:44 +00:00
|
|
|
if (unlikely(duration)) {
|
2012-03-10 19:37:28 +00:00
|
|
|
/*
|
|
|
|
* If a virtual machine is stopped by the host it can look to
|
|
|
|
* the watchdog like a soft lockup, check to see if the host
|
|
|
|
* stopped the vm before we issue the warning
|
|
|
|
*/
|
|
|
|
if (kvm_check_and_clear_guest_paused())
|
|
|
|
return HRTIMER_RESTART;
|
|
|
|
|
2010-05-07 21:11:44 +00:00
|
|
|
/* only warn once */
|
2014-10-09 22:25:17 +00:00
|
|
|
if (__this_cpu_read(soft_watchdog_warn) == true) {
|
|
|
|
/*
|
|
|
|
* When multiple processes are causing softlockups the
|
|
|
|
* softlockup detector only warns on the first one
|
|
|
|
* because the code relies on a full quiet cycle to
|
|
|
|
* re-arm. The second process prevents the quiet cycle
|
|
|
|
* and never gets reported. Use task pointers to detect
|
|
|
|
* this.
|
|
|
|
*/
|
|
|
|
if (__this_cpu_read(softlockup_task_ptr_saved) !=
|
|
|
|
current) {
|
|
|
|
__this_cpu_write(soft_watchdog_warn, false);
|
|
|
|
__touch_watchdog();
|
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
return HRTIMER_RESTART;
|
2014-10-09 22:25:17 +00:00
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2014-06-23 20:22:05 +00:00
|
|
|
if (softlockup_all_cpu_backtrace) {
|
|
|
|
/* Prevent multiple soft-lockup reports if one cpu is already
|
|
|
|
* engaged in dumping cpu back traces
|
|
|
|
*/
|
|
|
|
if (test_and_set_bit(0, &soft_lockup_nmi_warn)) {
|
|
|
|
/* Someone else will report us. Let's give up */
|
|
|
|
__this_cpu_write(soft_watchdog_warn, true);
|
|
|
|
return HRTIMER_RESTART;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-08-06 23:04:03 +00:00
|
|
|
pr_emerg("BUG: soft lockup - CPU#%d stuck for %us! [%s:%d]\n",
|
2010-05-17 22:06:04 +00:00
|
|
|
smp_processor_id(), duration,
|
2010-05-07 21:11:44 +00:00
|
|
|
current->comm, task_pid_nr(current));
|
2014-10-09 22:25:17 +00:00
|
|
|
__this_cpu_write(softlockup_task_ptr_saved, current);
|
2010-05-07 21:11:44 +00:00
|
|
|
print_modules();
|
|
|
|
print_irqtrace_events(current);
|
|
|
|
if (regs)
|
|
|
|
show_regs(regs);
|
|
|
|
else
|
|
|
|
dump_stack();
|
|
|
|
|
2014-06-23 20:22:05 +00:00
|
|
|
if (softlockup_all_cpu_backtrace) {
|
|
|
|
/* Avoid generating two back traces for current
|
|
|
|
* given that one is already made above
|
|
|
|
*/
|
|
|
|
trigger_allbutself_cpu_backtrace();
|
|
|
|
|
|
|
|
clear_bit(0, &soft_lockup_nmi_warn);
|
|
|
|
/* Barrier to sync with other cpus */
|
|
|
|
smp_mb__after_atomic();
|
|
|
|
}
|
|
|
|
|
2014-08-08 21:22:31 +00:00
|
|
|
add_taint(TAINT_SOFTLOCKUP, LOCKDEP_STILL_OK);
|
2010-05-07 21:11:44 +00:00
|
|
|
if (softlockup_panic)
|
|
|
|
panic("softlockup: hung tasks");
|
2010-12-08 15:22:55 +00:00
|
|
|
__this_cpu_write(soft_watchdog_warn, true);
|
2010-05-07 21:11:44 +00:00
|
|
|
} else
|
2010-12-08 15:22:55 +00:00
|
|
|
__this_cpu_write(soft_watchdog_warn, false);
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
return HRTIMER_RESTART;
|
|
|
|
}
|
|
|
|
|
2012-07-16 10:42:38 +00:00
|
|
|
static void watchdog_set_prio(unsigned int policy, unsigned int prio)
|
|
|
|
{
|
|
|
|
struct sched_param param = { .sched_priority = prio };
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2012-07-16 10:42:38 +00:00
|
|
|
sched_setscheduler(current, policy, ¶m);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void watchdog_enable(unsigned int cpu)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2014-08-17 17:30:34 +00:00
|
|
|
struct hrtimer *hrtimer = raw_cpu_ptr(&watchdog_hrtimer);
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2012-12-19 19:51:31 +00:00
|
|
|
/* kick off the timer for the hardlockup detector */
|
|
|
|
hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
|
|
|
|
hrtimer->function = watchdog_timer_fn;
|
|
|
|
|
2012-07-16 10:42:38 +00:00
|
|
|
/* Enable the perf event */
|
|
|
|
watchdog_nmi_enable(cpu);
|
2010-05-07 21:11:44 +00:00
|
|
|
|
|
|
|
/* done here because hrtimer_start can only pin to smp_processor_id() */
|
2012-12-17 23:59:50 +00:00
|
|
|
hrtimer_start(hrtimer, ns_to_ktime(sample_period),
|
2010-05-07 21:11:44 +00:00
|
|
|
HRTIMER_MODE_REL_PINNED);
|
|
|
|
|
2012-07-16 10:42:38 +00:00
|
|
|
/* initialize timestamp */
|
|
|
|
watchdog_set_prio(SCHED_FIFO, MAX_RT_PRIO - 1);
|
|
|
|
__touch_watchdog();
|
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2012-07-16 10:42:38 +00:00
|
|
|
static void watchdog_disable(unsigned int cpu)
|
|
|
|
{
|
2014-08-17 17:30:34 +00:00
|
|
|
struct hrtimer *hrtimer = raw_cpu_ptr(&watchdog_hrtimer);
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2012-07-16 10:42:38 +00:00
|
|
|
watchdog_set_prio(SCHED_NORMAL, 0);
|
|
|
|
hrtimer_cancel(hrtimer);
|
|
|
|
/* disable the perf event */
|
|
|
|
watchdog_nmi_disable(cpu);
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
2013-06-06 13:42:53 +00:00
|
|
|
static void watchdog_cleanup(unsigned int cpu, bool online)
|
|
|
|
{
|
|
|
|
watchdog_disable(cpu);
|
|
|
|
}
|
|
|
|
|
2012-07-16 10:42:38 +00:00
|
|
|
static int watchdog_should_run(unsigned int cpu)
|
|
|
|
{
|
|
|
|
return __this_cpu_read(hrtimer_interrupts) !=
|
|
|
|
__this_cpu_read(soft_lockup_hrtimer_cnt);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The watchdog thread function - touches the timestamp.
|
|
|
|
*
|
2012-12-17 23:59:50 +00:00
|
|
|
* It only runs once every sample_period seconds (4 seconds by
|
2012-07-16 10:42:38 +00:00
|
|
|
* default) to reset the softlockup timestamp. If this gets delayed
|
|
|
|
* for more than 2*watchdog_thresh seconds then the debug-printout
|
|
|
|
* triggers in watchdog_timer_fn().
|
|
|
|
*/
|
|
|
|
static void watchdog(unsigned int cpu)
|
|
|
|
{
|
|
|
|
__this_cpu_write(soft_lockup_hrtimer_cnt,
|
|
|
|
__this_cpu_read(hrtimer_interrupts));
|
|
|
|
__touch_watchdog();
|
2015-04-14 22:44:10 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* watchdog_nmi_enable() clears the NMI_WATCHDOG_ENABLED bit in the
|
|
|
|
* failure path. Check for failures that can occur asynchronously -
|
|
|
|
* for example, when CPUs are on-lined - and shut down the hardware
|
|
|
|
* perf event on each CPU accordingly.
|
|
|
|
*
|
|
|
|
* The only non-obvious place this bit can be cleared is through
|
|
|
|
* watchdog_nmi_enable(), so a pr_info() is placed there. Placing a
|
|
|
|
* pr_info here would be too noisy as it would result in a message
|
|
|
|
* every few seconds if the hardlockup was disabled but the softlockup
|
|
|
|
* enabled.
|
|
|
|
*/
|
|
|
|
if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED))
|
|
|
|
watchdog_nmi_disable(cpu);
|
2012-07-16 10:42:38 +00:00
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2013-06-06 13:42:53 +00:00
|
|
|
static struct smp_hotplug_thread watchdog_threads = {
|
|
|
|
.store = &softlockup_watchdog,
|
|
|
|
.thread_should_run = watchdog_should_run,
|
|
|
|
.thread_fn = watchdog,
|
|
|
|
.thread_comm = "watchdog/%u",
|
|
|
|
.setup = watchdog_enable,
|
|
|
|
.cleanup = watchdog_cleanup,
|
|
|
|
.park = watchdog_disable,
|
|
|
|
.unpark = watchdog_enable,
|
|
|
|
};
|
|
|
|
|
2015-09-04 22:45:15 +00:00
|
|
|
/*
|
|
|
|
* park all watchdog threads that are specified in 'watchdog_cpumask'
|
2015-11-06 02:44:39 +00:00
|
|
|
*
|
|
|
|
* This function returns an error if kthread_park() of a watchdog thread
|
|
|
|
* fails. In this situation, the watchdog threads of some CPUs can already
|
|
|
|
* be parked and the watchdog threads of other CPUs can still be runnable.
|
|
|
|
* Callers are expected to handle this special condition as appropriate in
|
|
|
|
* their context.
|
2015-11-06 02:44:53 +00:00
|
|
|
*
|
|
|
|
* This function may only be called in a context that is protected against
|
|
|
|
* races with CPU hotplug - for example, via get_online_cpus().
|
2015-09-04 22:45:15 +00:00
|
|
|
*/
|
|
|
|
static int watchdog_park_threads(void)
|
|
|
|
{
|
|
|
|
int cpu, ret = 0;
|
|
|
|
|
2017-01-24 23:17:53 +00:00
|
|
|
atomic_set(&watchdog_park_in_progress, 1);
|
|
|
|
|
2015-09-04 22:45:15 +00:00
|
|
|
for_each_watchdog_cpu(cpu) {
|
|
|
|
ret = kthread_park(per_cpu(softlockup_watchdog, cpu));
|
|
|
|
if (ret)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2017-01-24 23:17:53 +00:00
|
|
|
atomic_set(&watchdog_park_in_progress, 0);
|
|
|
|
|
2015-09-04 22:45:15 +00:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* unpark all watchdog threads that are specified in 'watchdog_cpumask'
|
2015-11-06 02:44:53 +00:00
|
|
|
*
|
|
|
|
* This function may only be called in a context that is protected against
|
|
|
|
* races with CPU hotplug - for example, via get_online_cpus().
|
2015-09-04 22:45:15 +00:00
|
|
|
*/
|
|
|
|
static void watchdog_unpark_threads(void)
|
|
|
|
{
|
|
|
|
int cpu;
|
|
|
|
|
|
|
|
for_each_watchdog_cpu(cpu)
|
|
|
|
kthread_unpark(per_cpu(softlockup_watchdog, cpu));
|
|
|
|
}
|
|
|
|
|
2015-09-04 22:45:18 +00:00
|
|
|
/*
|
|
|
|
* Suspend the hard and soft lockup detector by parking the watchdog threads.
|
|
|
|
*/
|
2015-09-04 22:45:28 +00:00
|
|
|
int lockup_detector_suspend(void)
|
2015-09-04 22:45:18 +00:00
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
|
2015-11-06 02:44:47 +00:00
|
|
|
get_online_cpus();
|
2015-09-04 22:45:18 +00:00
|
|
|
mutex_lock(&watchdog_proc_mutex);
|
|
|
|
/*
|
|
|
|
* Multiple suspend requests can be active in parallel (counted by
|
|
|
|
* the 'watchdog_suspended' variable). If the watchdog threads are
|
|
|
|
* running, the first caller takes care that they will be parked.
|
|
|
|
* The state of 'watchdog_running' cannot change while a suspend
|
2015-09-04 22:45:28 +00:00
|
|
|
* request is active (see related code in 'proc' handlers).
|
2015-09-04 22:45:18 +00:00
|
|
|
*/
|
|
|
|
if (watchdog_running && !watchdog_suspended)
|
|
|
|
ret = watchdog_park_threads();
|
|
|
|
|
|
|
|
if (ret == 0)
|
|
|
|
watchdog_suspended++;
|
2015-11-06 02:44:36 +00:00
|
|
|
else {
|
|
|
|
watchdog_disable_all_cpus();
|
|
|
|
pr_err("Failed to suspend lockup detectors, disabled\n");
|
|
|
|
watchdog_enabled = 0;
|
|
|
|
}
|
2015-09-04 22:45:18 +00:00
|
|
|
|
|
|
|
mutex_unlock(&watchdog_proc_mutex);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Resume the hard and soft lockup detector by unparking the watchdog threads.
|
|
|
|
*/
|
2015-09-04 22:45:28 +00:00
|
|
|
void lockup_detector_resume(void)
|
2015-09-04 22:45:18 +00:00
|
|
|
{
|
|
|
|
mutex_lock(&watchdog_proc_mutex);
|
|
|
|
|
|
|
|
watchdog_suspended--;
|
|
|
|
/*
|
|
|
|
* The watchdog threads are unparked if they were previously running
|
|
|
|
* and if there is no more active suspend request.
|
|
|
|
*/
|
|
|
|
if (watchdog_running && !watchdog_suspended)
|
|
|
|
watchdog_unpark_threads();
|
|
|
|
|
|
|
|
mutex_unlock(&watchdog_proc_mutex);
|
2015-11-06 02:44:47 +00:00
|
|
|
put_online_cpus();
|
2015-09-04 22:45:18 +00:00
|
|
|
}
|
|
|
|
|
2015-11-06 02:44:33 +00:00
|
|
|
static int update_watchdog_all_cpus(void)
|
2013-09-24 22:27:30 +00:00
|
|
|
{
|
2015-11-06 02:44:33 +00:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = watchdog_park_threads();
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2015-09-04 22:45:21 +00:00
|
|
|
watchdog_unpark_threads();
|
2015-11-06 02:44:33 +00:00
|
|
|
|
|
|
|
return 0;
|
2013-09-24 22:27:30 +00:00
|
|
|
}
|
|
|
|
|
2015-04-14 22:44:16 +00:00
|
|
|
static int watchdog_enable_all_cpus(void)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2013-06-06 13:42:53 +00:00
|
|
|
int err = 0;
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2013-05-19 18:45:15 +00:00
|
|
|
if (!watchdog_running) {
|
2015-09-04 22:45:06 +00:00
|
|
|
err = smpboot_register_percpu_thread_cpumask(&watchdog_threads,
|
|
|
|
&watchdog_cpumask);
|
2013-06-06 13:42:53 +00:00
|
|
|
if (err)
|
|
|
|
pr_err("Failed to create watchdog threads, disabled\n");
|
2015-09-04 22:45:06 +00:00
|
|
|
else
|
2013-05-19 18:45:15 +00:00
|
|
|
watchdog_running = 1;
|
2015-04-14 22:44:16 +00:00
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* Enable/disable the lockup detectors or
|
|
|
|
* change the sample period 'on the fly'.
|
|
|
|
*/
|
2015-11-06 02:44:33 +00:00
|
|
|
err = update_watchdog_all_cpus();
|
|
|
|
|
|
|
|
if (err) {
|
|
|
|
watchdog_disable_all_cpus();
|
|
|
|
pr_err("Failed to update lockup detectors, disabled\n");
|
|
|
|
}
|
2012-07-16 10:42:38 +00:00
|
|
|
}
|
2013-06-06 13:42:53 +00:00
|
|
|
|
2015-11-06 02:44:33 +00:00
|
|
|
if (err)
|
|
|
|
watchdog_enabled = 0;
|
|
|
|
|
2013-06-06 13:42:53 +00:00
|
|
|
return err;
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void watchdog_disable_all_cpus(void)
|
|
|
|
{
|
2013-05-19 18:45:15 +00:00
|
|
|
if (watchdog_running) {
|
|
|
|
watchdog_running = 0;
|
2013-06-06 13:42:53 +00:00
|
|
|
smpboot_unregister_percpu_thread(&watchdog_threads);
|
2012-07-16 10:42:38 +00:00
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
|
|
|
|
2015-11-06 02:44:30 +00:00
|
|
|
#ifdef CONFIG_SYSCTL
|
|
|
|
|
2010-05-07 21:11:44 +00:00
|
|
|
/*
|
2015-04-14 22:43:58 +00:00
|
|
|
* Update the run state of the lockup detectors.
|
|
|
|
*/
|
|
|
|
static int proc_watchdog_update(void)
|
|
|
|
{
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Watchdog threads won't be started if they are already active.
|
|
|
|
* The 'watchdog_running' variable in watchdog_*_all_cpus() takes
|
|
|
|
* care of this. If those threads are already active, the sample
|
|
|
|
* period will be updated and the lockup detectors will be enabled
|
|
|
|
* or disabled 'on the fly'.
|
|
|
|
*/
|
|
|
|
if (watchdog_enabled && watchdog_thresh)
|
2015-04-14 22:44:16 +00:00
|
|
|
err = watchdog_enable_all_cpus();
|
2015-04-14 22:43:58 +00:00
|
|
|
else
|
|
|
|
watchdog_disable_all_cpus();
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2015-04-14 22:44:05 +00:00
|
|
|
/*
|
|
|
|
* common function for watchdog, nmi_watchdog and soft_watchdog parameter
|
|
|
|
*
|
|
|
|
* caller | table->data points to | 'which' contains the flag(s)
|
|
|
|
* -------------------|-----------------------|-----------------------------
|
|
|
|
* proc_watchdog | watchdog_user_enabled | NMI_WATCHDOG_ENABLED or'ed
|
|
|
|
* | | with SOFT_WATCHDOG_ENABLED
|
|
|
|
* -------------------|-----------------------|-----------------------------
|
|
|
|
* proc_nmi_watchdog | nmi_watchdog_enabled | NMI_WATCHDOG_ENABLED
|
|
|
|
* -------------------|-----------------------|-----------------------------
|
|
|
|
* proc_soft_watchdog | soft_watchdog_enabled | SOFT_WATCHDOG_ENABLED
|
|
|
|
*/
|
|
|
|
static int proc_watchdog_common(int which, struct ctl_table *table, int write,
|
|
|
|
void __user *buffer, size_t *lenp, loff_t *ppos)
|
|
|
|
{
|
|
|
|
int err, old, new;
|
|
|
|
int *watchdog_param = (int *)table->data;
|
|
|
|
|
2015-11-06 02:44:50 +00:00
|
|
|
get_online_cpus();
|
2015-04-14 22:44:05 +00:00
|
|
|
mutex_lock(&watchdog_proc_mutex);
|
|
|
|
|
2015-09-04 22:45:18 +00:00
|
|
|
if (watchdog_suspended) {
|
|
|
|
/* no parameter changes allowed while watchdog is suspended */
|
|
|
|
err = -EAGAIN;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2015-04-14 22:44:05 +00:00
|
|
|
/*
|
|
|
|
* If the parameter is being read return the state of the corresponding
|
|
|
|
* bit(s) in 'watchdog_enabled', else update 'watchdog_enabled' and the
|
|
|
|
* run state of the lockup detectors.
|
|
|
|
*/
|
|
|
|
if (!write) {
|
|
|
|
*watchdog_param = (watchdog_enabled & which) != 0;
|
|
|
|
err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
|
|
|
|
} else {
|
|
|
|
err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
|
|
|
|
if (err)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* There is a race window between fetching the current value
|
|
|
|
* from 'watchdog_enabled' and storing the new value. During
|
|
|
|
* this race window, watchdog_nmi_enable() can sneak in and
|
|
|
|
* clear the NMI_WATCHDOG_ENABLED bit in 'watchdog_enabled'.
|
|
|
|
* The 'cmpxchg' detects this race and the loop retries.
|
|
|
|
*/
|
|
|
|
do {
|
|
|
|
old = watchdog_enabled;
|
|
|
|
/*
|
|
|
|
* If the parameter value is not zero set the
|
|
|
|
* corresponding bit(s), else clear it(them).
|
|
|
|
*/
|
|
|
|
if (*watchdog_param)
|
|
|
|
new = old | which;
|
|
|
|
else
|
|
|
|
new = old & ~which;
|
|
|
|
} while (cmpxchg(&watchdog_enabled, old, new) != old);
|
|
|
|
|
|
|
|
/*
|
2015-11-06 02:44:33 +00:00
|
|
|
* Update the run state of the lockup detectors. There is _no_
|
|
|
|
* need to check the value returned by proc_watchdog_update()
|
|
|
|
* and to restore the previous value of 'watchdog_enabled' as
|
|
|
|
* both lockup detectors are disabled if proc_watchdog_update()
|
|
|
|
* returns an error.
|
2015-04-14 22:44:05 +00:00
|
|
|
*/
|
2016-03-17 21:17:23 +00:00
|
|
|
if (old == new)
|
|
|
|
goto out;
|
|
|
|
|
2015-04-14 22:44:05 +00:00
|
|
|
err = proc_watchdog_update();
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
mutex_unlock(&watchdog_proc_mutex);
|
2015-11-06 02:44:50 +00:00
|
|
|
put_online_cpus();
|
2015-04-14 22:44:05 +00:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2015-04-14 22:44:08 +00:00
|
|
|
/*
|
|
|
|
* /proc/sys/kernel/watchdog
|
|
|
|
*/
|
|
|
|
int proc_watchdog(struct ctl_table *table, int write,
|
|
|
|
void __user *buffer, size_t *lenp, loff_t *ppos)
|
|
|
|
{
|
|
|
|
return proc_watchdog_common(NMI_WATCHDOG_ENABLED|SOFT_WATCHDOG_ENABLED,
|
|
|
|
table, write, buffer, lenp, ppos);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* /proc/sys/kernel/nmi_watchdog
|
2010-05-07 21:11:44 +00:00
|
|
|
*/
|
2015-04-14 22:44:08 +00:00
|
|
|
int proc_nmi_watchdog(struct ctl_table *table, int write,
|
|
|
|
void __user *buffer, size_t *lenp, loff_t *ppos)
|
|
|
|
{
|
|
|
|
return proc_watchdog_common(NMI_WATCHDOG_ENABLED,
|
|
|
|
table, write, buffer, lenp, ppos);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* /proc/sys/kernel/soft_watchdog
|
|
|
|
*/
|
|
|
|
int proc_soft_watchdog(struct ctl_table *table, int write,
|
|
|
|
void __user *buffer, size_t *lenp, loff_t *ppos)
|
|
|
|
{
|
|
|
|
return proc_watchdog_common(SOFT_WATCHDOG_ENABLED,
|
|
|
|
table, write, buffer, lenp, ppos);
|
|
|
|
}
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2015-04-14 22:44:08 +00:00
|
|
|
/*
|
|
|
|
* /proc/sys/kernel/watchdog_thresh
|
|
|
|
*/
|
|
|
|
int proc_watchdog_thresh(struct ctl_table *table, int write,
|
|
|
|
void __user *buffer, size_t *lenp, loff_t *ppos)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2016-03-17 21:17:23 +00:00
|
|
|
int err, old, new;
|
2010-05-07 21:11:44 +00:00
|
|
|
|
2015-11-06 02:44:50 +00:00
|
|
|
get_online_cpus();
|
2013-09-24 22:27:29 +00:00
|
|
|
mutex_lock(&watchdog_proc_mutex);
|
2012-07-16 10:42:38 +00:00
|
|
|
|
2015-09-04 22:45:18 +00:00
|
|
|
if (watchdog_suspended) {
|
|
|
|
/* no parameter changes allowed while watchdog is suspended */
|
|
|
|
err = -EAGAIN;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2015-04-14 22:44:08 +00:00
|
|
|
old = ACCESS_ONCE(watchdog_thresh);
|
2013-06-06 13:42:53 +00:00
|
|
|
err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
|
2015-04-14 22:44:08 +00:00
|
|
|
|
2013-06-06 13:42:53 +00:00
|
|
|
if (err || !write)
|
2013-09-24 22:27:29 +00:00
|
|
|
goto out;
|
2011-05-23 05:10:21 +00:00
|
|
|
|
2013-03-12 18:44:08 +00:00
|
|
|
/*
|
2015-11-06 02:44:27 +00:00
|
|
|
* Update the sample period. Restore on failure.
|
2013-03-12 18:44:08 +00:00
|
|
|
*/
|
2016-03-17 21:17:23 +00:00
|
|
|
new = ACCESS_ONCE(watchdog_thresh);
|
|
|
|
if (old == new)
|
|
|
|
goto out;
|
|
|
|
|
2015-04-14 22:44:08 +00:00
|
|
|
set_sample_period();
|
|
|
|
err = proc_watchdog_update();
|
2015-11-06 02:44:27 +00:00
|
|
|
if (err) {
|
2015-04-14 22:44:08 +00:00
|
|
|
watchdog_thresh = old;
|
2015-11-06 02:44:27 +00:00
|
|
|
set_sample_period();
|
|
|
|
}
|
2013-09-24 22:27:29 +00:00
|
|
|
out:
|
|
|
|
mutex_unlock(&watchdog_proc_mutex);
|
2015-11-06 02:44:50 +00:00
|
|
|
put_online_cpus();
|
2013-06-06 13:42:53 +00:00
|
|
|
return err;
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|
2015-06-24 23:55:45 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The cpumask is the mask of possible cpus that the watchdog can run
|
|
|
|
* on, not the mask of cpus it is actually running on. This allows the
|
|
|
|
* user to specify a mask that will include cpus that have not yet
|
|
|
|
* been brought online, if desired.
|
|
|
|
*/
|
|
|
|
int proc_watchdog_cpumask(struct ctl_table *table, int write,
|
|
|
|
void __user *buffer, size_t *lenp, loff_t *ppos)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
|
2015-11-06 02:44:50 +00:00
|
|
|
get_online_cpus();
|
2015-06-24 23:55:45 +00:00
|
|
|
mutex_lock(&watchdog_proc_mutex);
|
2015-09-04 22:45:18 +00:00
|
|
|
|
|
|
|
if (watchdog_suspended) {
|
|
|
|
/* no parameter changes allowed while watchdog is suspended */
|
|
|
|
err = -EAGAIN;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2015-06-24 23:55:45 +00:00
|
|
|
err = proc_do_large_bitmap(table, write, buffer, lenp, ppos);
|
|
|
|
if (!err && write) {
|
|
|
|
/* Remove impossible cpus to keep sysctl output cleaner. */
|
|
|
|
cpumask_and(&watchdog_cpumask, &watchdog_cpumask,
|
|
|
|
cpu_possible_mask);
|
|
|
|
|
|
|
|
if (watchdog_running) {
|
|
|
|
/*
|
|
|
|
* Failure would be due to being unable to allocate
|
|
|
|
* a temporary cpumask, so we are likely not in a
|
|
|
|
* position to do much else to make things better.
|
|
|
|
*/
|
|
|
|
if (smpboot_update_cpumask_percpu_thread(
|
|
|
|
&watchdog_threads, &watchdog_cpumask) != 0)
|
|
|
|
pr_err("cpumask update failed\n");
|
|
|
|
}
|
|
|
|
}
|
2015-09-04 22:45:18 +00:00
|
|
|
out:
|
2015-06-24 23:55:45 +00:00
|
|
|
mutex_unlock(&watchdog_proc_mutex);
|
2015-11-06 02:44:50 +00:00
|
|
|
put_online_cpus();
|
2015-06-24 23:55:45 +00:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2010-05-07 21:11:44 +00:00
|
|
|
#endif /* CONFIG_SYSCTL */
|
|
|
|
|
2010-11-25 17:38:29 +00:00
|
|
|
void __init lockup_detector_init(void)
|
2010-05-07 21:11:44 +00:00
|
|
|
{
|
2012-12-17 23:59:50 +00:00
|
|
|
set_sample_period();
|
2013-06-06 13:42:53 +00:00
|
|
|
|
2015-06-24 23:55:45 +00:00
|
|
|
#ifdef CONFIG_NO_HZ_FULL
|
|
|
|
if (tick_nohz_full_enabled()) {
|
2015-09-04 22:45:09 +00:00
|
|
|
pr_info("Disabling watchdog on nohz_full cores by default\n");
|
|
|
|
cpumask_copy(&watchdog_cpumask, housekeeping_mask);
|
2015-06-24 23:55:45 +00:00
|
|
|
} else
|
|
|
|
cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
|
|
|
|
#else
|
|
|
|
cpumask_copy(&watchdog_cpumask, cpu_possible_mask);
|
|
|
|
#endif
|
|
|
|
|
2015-04-14 22:44:13 +00:00
|
|
|
if (watchdog_enabled)
|
2015-04-14 22:44:16 +00:00
|
|
|
watchdog_enable_all_cpus();
|
2010-05-07 21:11:44 +00:00
|
|
|
}
|