Re: [PATCH] sched: wake-affine throttle

[Michael Wang <wangyun@linux.vnet.ibm.com>]

On 04/22/2013 06:35 PM, Ingo Molnar wrote:
> 
> * Peter Zijlstra <peterz@infradead.org> wrote:
> 
>> OK,.. Ingo said that pipe-test was the original motivation for 
>> wake_affine() and since that's currently broken to pieces due to 
>> select_idle_sibling() is there still a benefit to having it at all?
>>
>> Can anybody find any significant regression when simply killing 
>> wake_affine()?
> 
> I'd suggest doing a patch that does:
> 
>    s/SD_WAKE_AFFINE/0*SD_WAKE_AFFINE

But by doing this, we won't be able to find 'affine_sd' any more, that
will also skip the select_idle_sibling() logical (in current code),
isn't it?

If we really want to kill the stuff (I prefer not...), I suggest we
forbidden the wake-affine by throttle it with an incredible interval,
that's also easily to be reverted :)

Regards,
Michael Wang

> 
> in all the relevant toplogy.h files, but otherwise keep the logic in 
> place. That way it's easy to revert.
> 
> Thanks,
> 
> 	Ingo
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>