sched/rt: Do not account zero delta_exec in update_curr_rt()
authorKirill Tkhai <tkhai@yandex.ru>
Wed, 30 Jan 2013 12:50:36 +0000 (16:50 +0400)
committerIngo Molnar <mingo@kernel.org>
Thu, 31 Jan 2013 09:31:13 +0000 (10:31 +0100)
There are several places of consecutive calls of
dequeue_task_rt() and put_prev_task_rt() in the scheduler.
For example, function rt_mutex_setprio() does it.

The both calls lead to update_curr_rt(), the second of it
receives zeroed delta_exec. The only effective action in this
case is call of sched_rt_avg_update(), which can change
rq->age_stamp and rq->rt_avg. But it is possible in case of
""floating"" rq->clock. This fact is not reasonable to be
accounted. Another actions do nothing.

Signed-off-by: Kirill V Tkhai <tkhai@yandex.ru>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
CC: linux-rt-users <linux-rt-users@vger.kernel.org>
Link: http://lkml.kernel.org/r/931541359550236@web1g.yandex.ru
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/sched/rt.c

index 2f69ca9..94abca4 100644 (file)
@@ -925,8 +925,8 @@ static void update_curr_rt(struct rq *rq)
                return;
 
        delta_exec = rq->clock_task - curr->se.exec_start;
-       if (unlikely((s64)delta_exec < 0))
-               delta_exec = 0;
+       if (unlikely((s64)delta_exec <= 0))
+               return;
 
        schedstat_set(curr->se.statistics.exec_max,
                      max(curr->se.statistics.exec_max, delta_exec));