Re: [patch] CFS scheduler, -v15

[Srivatsa Vaddagiri <vatsa@in.ibm.com>]

On Wed, Jun 06, 2007 at 01:19:01PM +0200, Dmitry Adamushko wrote:
> >Yes this is the approach I prefer, because we burden the fast/normal
> >path less that way (RT->NORMAL transition is not common).
> 
> I don't think that rt_sched_class :: dequeue_task_rt() is in any of
> such "fast pathes" that we should really care about an additional
> math. operation.
> 
> If this approach is ok, logically-wise (no side effects from a short
> 'delta_exec', esp. on RT -> NORMAL).. I think it's better as it keeps
> the 'sched_class' interface simpler.

I can see your point, given that we just update exec_start in
dequeue_task_rt(). However the situation is slightly more complex in my
patch stack, as I need to update other things during group change. Perhaps
we can postpone this discussion until I post that patch stack.

> >That's why I
> >was considering a set_curr_task() method in fair_sched_class which will
> >be invoked in __setscheduler() if the new policy of currently running
> >task happens to be SCHED_NORMAL/BATCH. Alternately if the new policy of
> >currently running task happens to be SCHED_FIFO (and its old policy was
> >SCHED_NORMAL) we need to invoke put_prev_task() method (so that
> >fair_clock etc is updated based on outgoing task's execution time in
> >SCHED_NORMAL class).
> 
> rt_sched_class :: put_prev_task() from __setscheduler() ?

No, fair_sched_class :: put_prev_task() if we are transitioning from
NORMAL->RT. That will update the fair_clock based on execution time
of current task in fair_sched class?

> But it's not
> supposed to be called from here, logically-wise. You just rely on its
> current behavior (which is only about updating 'exec_start' and
> 'exec_sum') -- that's just bad. Maybe I misunderatood your intention
> though..

-- 
Regards,
vatsa