[PATCH] sched: deliver sched_switch events to a current task

[PATCH] sched: deliver sched_switch events to a current task

[Andrew Vagin]

Otherwise they can't be filtered for a defined task.

perf record -e sched:sched_switch ./foo

This command doesn't report any event without this patch.

I think it isn't a big problem for security, if someone will know
who will be executed next. By default perf is disabled for non-root
users.

I need this events for profiling sleep times.  sched_switch is used for
getting callchains and sched_stat_* is used for getting time periods.
This events are combined in user space, then it can be analized by
perf tools.

Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Arun Sharma <asharma@fb.com>
Signed-off-by: Andrew Vagin <avagin@openvz.org>
---
 kernel/sched/core.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index d5594a4..c5d6a0f 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1910,12 +1910,12 @@ static inline void
 prepare_task_switch(struct rq *rq, struct task_struct *prev,
 		    struct task_struct *next)
 {
+	trace_sched_switch(prev, next);
 	sched_info_switch(prev, next);
 	perf_event_task_sched_out(prev, next);
 	fire_sched_out_preempt_notifiers(prev, next);
 	prepare_lock_switch(rq, next);
 	prepare_arch_switch(next);
-	trace_sched_switch(prev, next);
 }
 
 /**
-- 
1.7.1

Re: [PATCH] sched: deliver sched_switch events to a current task

[Peter Zijlstra]

On Thu, 2012-07-12 at 14:14 +0400, Andrew Vagin wrote:
> Otherwise they can't be filtered for a defined task.
> 
> perf record -e sched:sched_switch ./foo
> 
> This command doesn't report any event without this patch.
> 
> I think it isn't a big problem for security, if someone will know
> who will be executed next. By default perf is disabled for non-root
> users.
> 
> I need this events for profiling sleep times.  sched_switch is used for
> getting callchains and sched_stat_* is used for getting time periods.
> This events are combined in user space, then it can be analized by
> perf tools. 

Thanks!

[tip:perf/core] sched: Deliver sched_switch events to the current task

[tip-bot for Andrew Vagin]

Commit-ID:  895dd92c032e1604aa0d7afaef7716e058343b67
Gitweb:     http://git.kernel.org/tip/895dd92c032e1604aa0d7afaef7716e058343b67
Author:     Andrew Vagin <avagin@openvz.org>
AuthorDate: Thu, 12 Jul 2012 14:14:29 +0400
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Thu, 26 Jul 2012 12:23:10 +0200

sched: Deliver sched_switch events to the current task

Otherwise they can't be filtered for a defined task:

  perf record -e sched:sched_switch ./foo

This command doesn't report any events without this patch.

I think it isn't a security concern if someone knows who will
be executed next - this can already be observed by polling /proc
state. By default perf is disabled for non-root users in any case.

I need these events for profiling sleep times.  sched_switch is used for
getting callchains and sched_stat_* is used for getting time periods.
These events are combined in user space, then it can be analyzed by
perf tools.

Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Arun Sharma <asharma@fb.com>
Link: http://lkml.kernel.org/r/1342088069-1005148-1-git-send-email-avagin@openvz.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/core.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 468bdd4..ad732b5 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1910,12 +1910,12 @@ static inline void
 prepare_task_switch(struct rq *rq, struct task_struct *prev,
 		    struct task_struct *next)
 {
+	trace_sched_switch(prev, next);
 	sched_info_switch(prev, next);
 	perf_event_task_sched_out(prev, next);
 	fire_sched_out_preempt_notifiers(prev, next);
 	prepare_lock_switch(rq, next);
 	prepare_arch_switch(next);
-	trace_sched_switch(prev, next);
 }
 
 /**