Re: [PATCH] sched/fair: Remove idle_balance() declaration in sched.h

[Peter Zijlstra]

On Tue, Feb 11, 2014 at 04:33:40PM +0100, Daniel Lezcano wrote:
> On 02/11/2014 04:11 PM, Peter Zijlstra wrote:
> >On Tue, Feb 11, 2014 at 04:01:04PM +0100, Daniel Lezcano wrote:
> >>The idle_balance() function is called within a #ifdef CONFIG_SMP section.
> >>
> >>Remove its declaration in sched.h for !CONFIG_SMP because it is pointless.
> >>
> >>Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
> >
> >The below again makes a horrible mess of idle_balance() -- which you
> >tried to clean up.. but it does rid us of some #ifdef goo.
> >
> >Hmmm?
> 
> Yes, it sounds ok.
> 
> Why is idle_enter_fair() called unconditionally in idle_balance() ?
> Isn't the call in pick_next_task_idle enough ?  Shouldn't be called when we
> will go to idle effectively ?
> 
> If I am not wrong idle_enter_fair() is called from idle_balance() but a task
> may be pulled, so we the next task won't be the idle task and idle_exit_fair
> won't be called at put_prev_task.
> 
> May be I missed this change which was done at purpose in the previous
> patchset you sent...

lkml.kernel.org/r/CAKfTPtAMmhq0tBpu+pk_uzwrfUM-6Fz4Do1yrzc4NRQS4cZ_-A@xxxxxxxxxxxxxx

I seem to be terminally confused on the subject, but Vincent explains it
there I think.
--
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/