Re: [patch] voluntary-preempt-2.6.9-rc1-bk4-Q9

[Ingo Molnar]

* Mark_H_Johnson@xxxxxxxxxxxx <Mark_H_Johnson@xxxxxxxxxxxx> wrote:

> 00000002 0.003ms (+0.000ms): dummy_switch_tasks (schedule)
> 00000002 0.003ms (+0.000ms): schedule (worker_thread)
> 00000002 0.003ms (+0.000ms): schedule (worker_thread)
> 00000002 0.004ms (+0.000ms): schedule (worker_thread)
> 00000002 0.004ms (+0.000ms): schedule (worker_thread)
> 00000002 0.004ms (+0.000ms): schedule (worker_thread)
> 00000002 0.004ms (+0.274ms): schedule (worker_thread)
> 04000002 0.279ms (+0.000ms): __switch_to (schedule)

a quick suggestion: could you add this near the top of sched.c (below
the #include lines):

 #define static

this will turn off all inlining and makes the scheduler-internal
functions visible. If there's any scheduler-internal overhead we should
see it. Maybe this is the CR3 flush (switch_mm()) - but 274 usecs is
still excessive ...

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/