hung_task debugging: Add tracepoint to report the hang

Currently check_hung_task() prints a warning if it detects the
problem, but it is not convenient to watch the system logs if
user-space wants to be notified about the hang.

Add the new trace_sched_process_hang() into check_hung_task(),
this way a user-space monitor can easily wait for the hang and
potentially resolve a problem.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Dave Sullivan <dsulliva@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/20131019161828.GA7439@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This commit is contained in:
Oleg Nesterov 2013-10-19 18:18:28 +02:00 committed by Ingo Molnar
parent 4c4e4f6136
commit 6a716c90a5
2 changed files with 23 additions and 0 deletions

View File

@ -424,6 +424,25 @@ TRACE_EVENT(sched_pi_setprio,
__entry->oldprio, __entry->newprio)
);
#ifdef CONFIG_DETECT_HUNG_TASK
TRACE_EVENT(sched_process_hang,
TP_PROTO(struct task_struct *tsk),
TP_ARGS(tsk),
TP_STRUCT__entry(
__array( char, comm, TASK_COMM_LEN )
__field( pid_t, pid )
),
TP_fast_assign(
memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN);
__entry->pid = tsk->pid;
),
TP_printk("comm=%s pid=%d", __entry->comm, __entry->pid)
);
#endif /* CONFIG_DETECT_HUNG_TASK */
#endif /* _TRACE_SCHED_H */
/* This part must be outside protection */

View File

@ -16,6 +16,7 @@
#include <linux/export.h>
#include <linux/sysctl.h>
#include <linux/utsname.h>
#include <trace/events/sched.h>
/*
* The number of tasks checked:
@ -92,6 +93,9 @@ static void check_hung_task(struct task_struct *t, unsigned long timeout)
t->last_switch_count = switch_count;
return;
}
trace_sched_process_hang(t);
if (!sysctl_hung_task_warnings)
return;
sysctl_hung_task_warnings--;