Commit 82607adc authored by Tejun Heo's avatar Tejun Heo

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: default avatarTejun Heo <tj@kernel.org>
Acked-by: default avatarDon 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>
parent 03e0d461
...@@ -4114,6 +4114,15 @@ bytes respectively. Such letter suffixes can also be entirely omitted. ...@@ -4114,6 +4114,15 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
or other driver-specific files in the or other driver-specific files in the
Documentation/watchdog/ directory. Documentation/watchdog/ directory.
workqueue.watchdog_thresh=
If CONFIG_WQ_WATCHDOG is configured, workqueue can
warn stall conditions and dump internal state to
help debugging. 0 disables workqueue stall
detection; otherwise, it's the stall threshold
duration in seconds. The default value is 30 and
it can be updated at runtime by writing to the
corresponding sysfs file.
workqueue.disable_numa workqueue.disable_numa
By default, all work items queued to unbound By default, all work items queued to unbound
workqueues are affine to the NUMA nodes they're workqueues are affine to the NUMA nodes they're
......
...@@ -618,4 +618,10 @@ static inline int workqueue_sysfs_register(struct workqueue_struct *wq) ...@@ -618,4 +618,10 @@ static inline int workqueue_sysfs_register(struct workqueue_struct *wq)
{ return 0; } { return 0; }
#endif /* CONFIG_SYSFS */ #endif /* CONFIG_SYSFS */
#ifdef CONFIG_WQ_WATCHDOG
void wq_watchdog_touch(int cpu);
#else /* CONFIG_WQ_WATCHDOG */
static inline void wq_watchdog_touch(int cpu) { }
#endif /* CONFIG_WQ_WATCHDOG */
#endif #endif
...@@ -20,6 +20,7 @@ ...@@ -20,6 +20,7 @@
#include <linux/smpboot.h> #include <linux/smpboot.h>
#include <linux/sched/rt.h> #include <linux/sched/rt.h>
#include <linux/tick.h> #include <linux/tick.h>
#include <linux/workqueue.h>
#include <asm/irq_regs.h> #include <asm/irq_regs.h>
#include <linux/kvm_para.h> #include <linux/kvm_para.h>
...@@ -245,6 +246,7 @@ void touch_softlockup_watchdog_sched(void) ...@@ -245,6 +246,7 @@ void touch_softlockup_watchdog_sched(void)
void touch_softlockup_watchdog(void) void touch_softlockup_watchdog(void)
{ {
touch_softlockup_watchdog_sched(); touch_softlockup_watchdog_sched();
wq_watchdog_touch(raw_smp_processor_id());
} }
EXPORT_SYMBOL(touch_softlockup_watchdog); EXPORT_SYMBOL(touch_softlockup_watchdog);
...@@ -259,6 +261,7 @@ void touch_all_softlockup_watchdogs(void) ...@@ -259,6 +261,7 @@ void touch_all_softlockup_watchdogs(void)
*/ */
for_each_watchdog_cpu(cpu) for_each_watchdog_cpu(cpu)
per_cpu(watchdog_touch_ts, cpu) = 0; per_cpu(watchdog_touch_ts, cpu) = 0;
wq_watchdog_touch(-1);
} }
#ifdef CONFIG_HARDLOCKUP_DETECTOR #ifdef CONFIG_HARDLOCKUP_DETECTOR
......
...@@ -148,6 +148,8 @@ struct worker_pool { ...@@ -148,6 +148,8 @@ struct worker_pool {
int id; /* I: pool ID */ int id; /* I: pool ID */
unsigned int flags; /* X: flags */ unsigned int flags; /* X: flags */
unsigned long watchdog_ts; /* L: watchdog timestamp */
struct list_head worklist; /* L: list of pending works */ struct list_head worklist; /* L: list of pending works */
int nr_workers; /* L: total number of workers */ int nr_workers; /* L: total number of workers */
...@@ -1083,6 +1085,8 @@ static void pwq_activate_delayed_work(struct work_struct *work) ...@@ -1083,6 +1085,8 @@ static void pwq_activate_delayed_work(struct work_struct *work)
struct pool_workqueue *pwq = get_work_pwq(work); struct pool_workqueue *pwq = get_work_pwq(work);
trace_workqueue_activate_work(work); trace_workqueue_activate_work(work);
if (list_empty(&pwq->pool->worklist))
pwq->pool->watchdog_ts = jiffies;
move_linked_works(work, &pwq->pool->worklist, NULL); move_linked_works(work, &pwq->pool->worklist, NULL);
__clear_bit(WORK_STRUCT_DELAYED_BIT, work_data_bits(work)); __clear_bit(WORK_STRUCT_DELAYED_BIT, work_data_bits(work));
pwq->nr_active++; pwq->nr_active++;
...@@ -1385,6 +1389,8 @@ static void __queue_work(int cpu, struct workqueue_struct *wq, ...@@ -1385,6 +1389,8 @@ static void __queue_work(int cpu, struct workqueue_struct *wq,
trace_workqueue_activate_work(work); trace_workqueue_activate_work(work);
pwq->nr_active++; pwq->nr_active++;
worklist = &pwq->pool->worklist; worklist = &pwq->pool->worklist;
if (list_empty(worklist))
pwq->pool->watchdog_ts = jiffies;
} else { } else {
work_flags |= WORK_STRUCT_DELAYED; work_flags |= WORK_STRUCT_DELAYED;
worklist = &pwq->delayed_works; worklist = &pwq->delayed_works;
...@@ -2157,6 +2163,8 @@ static int worker_thread(void *__worker) ...@@ -2157,6 +2163,8 @@ static int worker_thread(void *__worker)
list_first_entry(&pool->worklist, list_first_entry(&pool->worklist,
struct work_struct, entry); struct work_struct, entry);
pool->watchdog_ts = jiffies;
if (likely(!(*work_data_bits(work) & WORK_STRUCT_LINKED))) { if (likely(!(*work_data_bits(work) & WORK_STRUCT_LINKED))) {
/* optimization path, not strictly necessary */ /* optimization path, not strictly necessary */
process_one_work(worker, work); process_one_work(worker, work);
...@@ -2240,6 +2248,7 @@ static int rescuer_thread(void *__rescuer) ...@@ -2240,6 +2248,7 @@ static int rescuer_thread(void *__rescuer)
struct pool_workqueue, mayday_node); struct pool_workqueue, mayday_node);
struct worker_pool *pool = pwq->pool; struct worker_pool *pool = pwq->pool;
struct work_struct *work, *n; struct work_struct *work, *n;
bool first = true;
__set_current_state(TASK_RUNNING); __set_current_state(TASK_RUNNING);
list_del_init(&pwq->mayday_node); list_del_init(&pwq->mayday_node);
...@@ -2256,9 +2265,14 @@ static int rescuer_thread(void *__rescuer) ...@@ -2256,9 +2265,14 @@ static int rescuer_thread(void *__rescuer)
* process'em. * process'em.
*/ */
WARN_ON_ONCE(!list_empty(scheduled)); WARN_ON_ONCE(!list_empty(scheduled));
list_for_each_entry_safe(work, n, &pool->worklist, entry) list_for_each_entry_safe(work, n, &pool->worklist, entry) {
if (get_work_pwq(work) == pwq) if (get_work_pwq(work) == pwq) {
if (first)
pool->watchdog_ts = jiffies;
move_linked_works(work, scheduled, &n); move_linked_works(work, scheduled, &n);
}
first = false;
}
if (!list_empty(scheduled)) { if (!list_empty(scheduled)) {
process_scheduled_works(rescuer); process_scheduled_works(rescuer);
...@@ -3104,6 +3118,7 @@ static int init_worker_pool(struct worker_pool *pool) ...@@ -3104,6 +3118,7 @@ static int init_worker_pool(struct worker_pool *pool)
pool->cpu = -1; pool->cpu = -1;
pool->node = NUMA_NO_NODE; pool->node = NUMA_NO_NODE;
pool->flags |= POOL_DISASSOCIATED; pool->flags |= POOL_DISASSOCIATED;
pool->watchdog_ts = jiffies;
INIT_LIST_HEAD(&pool->worklist); INIT_LIST_HEAD(&pool->worklist);
INIT_LIST_HEAD(&pool->idle_list); INIT_LIST_HEAD(&pool->idle_list);
hash_init(pool->busy_hash); hash_init(pool->busy_hash);
...@@ -4343,7 +4358,9 @@ void show_workqueue_state(void) ...@@ -4343,7 +4358,9 @@ void show_workqueue_state(void)
pr_info("pool %d:", pool->id); pr_info("pool %d:", pool->id);
pr_cont_pool_info(pool); pr_cont_pool_info(pool);
pr_cont(" workers=%d", pool->nr_workers); pr_cont(" hung=%us workers=%d",
jiffies_to_msecs(jiffies - pool->watchdog_ts) / 1000,
pool->nr_workers);
if (pool->manager) if (pool->manager)
pr_cont(" manager: %d", pr_cont(" manager: %d",
task_pid_nr(pool->manager->task)); task_pid_nr(pool->manager->task));
...@@ -5202,6 +5219,154 @@ static void workqueue_sysfs_unregister(struct workqueue_struct *wq) ...@@ -5202,6 +5219,154 @@ static void workqueue_sysfs_unregister(struct workqueue_struct *wq)
static void workqueue_sysfs_unregister(struct workqueue_struct *wq) { } static void workqueue_sysfs_unregister(struct workqueue_struct *wq) { }
#endif /* CONFIG_SYSFS */ #endif /* CONFIG_SYSFS */
/*
* Workqueue watchdog.
*
* Stall may be caused by various bugs - missing WQ_MEM_RECLAIM, illegal
* flush dependency, a concurrency managed work item which stays RUNNING
* indefinitely. Workqueue stalls can be very difficult to debug as the
* usual warning mechanisms don't trigger and internal workqueue state is
* largely opaque.
*
* Workqueue watchdog monitors all worker pools periodically and dumps
* state if some pools failed to make forward progress for a while where
* forward progress is defined as the first item on ->worklist changing.
*
* This mechanism is controlled through the kernel parameter
* "workqueue.watchdog_thresh" which can be updated at runtime through the
* corresponding sysfs parameter file.
*/
#ifdef CONFIG_WQ_WATCHDOG
static void wq_watchdog_timer_fn(unsigned long data);
static unsigned long wq_watchdog_thresh = 30;
static struct timer_list wq_watchdog_timer =
TIMER_DEFERRED_INITIALIZER(wq_watchdog_timer_fn, 0, 0);
static unsigned long wq_watchdog_touched = INITIAL_JIFFIES;
static DEFINE_PER_CPU(unsigned long, wq_watchdog_touched_cpu) = INITIAL_JIFFIES;
static void wq_watchdog_reset_touched(void)
{
int cpu;
wq_watchdog_touched = jiffies;
for_each_possible_cpu(cpu)
per_cpu(wq_watchdog_touched_cpu, cpu) = jiffies;
}
static void wq_watchdog_timer_fn(unsigned long data)
{
unsigned long thresh = READ_ONCE(wq_watchdog_thresh) * HZ;
bool lockup_detected = false;
struct worker_pool *pool;
int pi;
if (!thresh)
return;
rcu_read_lock();
for_each_pool(pool, pi) {
unsigned long pool_ts, touched, ts;
if (list_empty(&pool->worklist))
continue;
/* get the latest of pool and touched timestamps */
pool_ts = READ_ONCE(pool->watchdog_ts);
touched = READ_ONCE(wq_watchdog_touched);
if (time_after(pool_ts, touched))
ts = pool_ts;
else
ts = touched;
if (pool->cpu >= 0) {
unsigned long cpu_touched =
READ_ONCE(per_cpu(wq_watchdog_touched_cpu,
pool->cpu));
if (time_after(cpu_touched, ts))
ts = cpu_touched;
}
/* did we stall? */
if (time_after(jiffies, ts + thresh)) {
lockup_detected = true;
pr_emerg("BUG: workqueue lockup - pool");
pr_cont_pool_info(pool);
pr_cont(" stuck for %us!\n",
jiffies_to_msecs(jiffies - pool_ts) / 1000);
}
}
rcu_read_unlock();
if (lockup_detected)
show_workqueue_state();
wq_watchdog_reset_touched();
mod_timer(&wq_watchdog_timer, jiffies + thresh);
}
void wq_watchdog_touch(int cpu)
{
if (cpu >= 0)
per_cpu(wq_watchdog_touched_cpu, cpu) = jiffies;
else
wq_watchdog_touched = jiffies;
}
static void wq_watchdog_set_thresh(unsigned long thresh)
{
wq_watchdog_thresh = 0;
del_timer_sync(&wq_watchdog_timer);
if (thresh) {
wq_watchdog_thresh = thresh;
wq_watchdog_reset_touched();
mod_timer(&wq_watchdog_timer, jiffies + thresh * HZ);
}
}
static int wq_watchdog_param_set_thresh(const char *val,
const struct kernel_param *kp)
{
unsigned long thresh;
int ret;
ret = kstrtoul(val, 0, &thresh);
if (ret)
return ret;
if (system_wq)
wq_watchdog_set_thresh(thresh);
else
wq_watchdog_thresh = thresh;
return 0;
}
static const struct kernel_param_ops wq_watchdog_thresh_ops = {
.set = wq_watchdog_param_set_thresh,
.get = param_get_ulong,
};
module_param_cb(watchdog_thresh, &wq_watchdog_thresh_ops, &wq_watchdog_thresh,
0644);
static void wq_watchdog_init(void)
{
wq_watchdog_set_thresh(wq_watchdog_thresh);
}
#else /* CONFIG_WQ_WATCHDOG */
static inline void wq_watchdog_init(void) { }
#endif /* CONFIG_WQ_WATCHDOG */
static void __init wq_numa_init(void) static void __init wq_numa_init(void)
{ {
cpumask_var_t *tbl; cpumask_var_t *tbl;
...@@ -5325,6 +5490,9 @@ static int __init init_workqueues(void) ...@@ -5325,6 +5490,9 @@ static int __init init_workqueues(void)
!system_unbound_wq || !system_freezable_wq || !system_unbound_wq || !system_freezable_wq ||
!system_power_efficient_wq || !system_power_efficient_wq ||
!system_freezable_power_efficient_wq); !system_freezable_power_efficient_wq);
wq_watchdog_init();
return 0; return 0;
} }
early_initcall(init_workqueues); early_initcall(init_workqueues);
...@@ -812,6 +812,17 @@ config BOOTPARAM_HUNG_TASK_PANIC_VALUE ...@@ -812,6 +812,17 @@ config BOOTPARAM_HUNG_TASK_PANIC_VALUE
default 0 if !BOOTPARAM_HUNG_TASK_PANIC default 0 if !BOOTPARAM_HUNG_TASK_PANIC
default 1 if BOOTPARAM_HUNG_TASK_PANIC default 1 if BOOTPARAM_HUNG_TASK_PANIC
config WQ_WATCHDOG
bool "Detect Workqueue Stalls"
depends on DEBUG_KERNEL
help
Say Y here to enable stall detection on workqueues. If a
worker pool doesn't make forward progress on a pending work
item for over a given amount of time, 30s by default, a
warning message is printed along with dump of workqueue
state. This can be configured through kernel parameter
"workqueue.watchdog_thresh" and its sysfs counterpart.
endmenu # "Debug lockups and hangs" endmenu # "Debug lockups and hangs"
config PANIC_ON_OOPS config PANIC_ON_OOPS
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment