lkml.org 
[lkml]   [2013]   [Jun]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: deadlock in scheduler enabling HRTICK feature
On 6/26/13 1:05 AM, Peter Zijlstra wrote:
>> What is the expectation that the feature provides? not a whole lot of
>> documentation on it. I walked down the path wondering if it solved an odd
>> problem we are seeing with the CFS in 2.6.27 kernel.
>
> Its supposed to use hrtimers for slice expiry instead of the regular tick.

So theoretically CPU bound tasks would get preempted sooner? That was my
guess/hope anyways.

>
> IIRC it did work at some point but bitrotted a bit since. The good news is that
> the deadline scheduler wants to use it and I'll probably have to fix it up
> then.

Hmmm.... meaning I should not be expecting anything in the next couple
of months? Any gut opinions on how to approach the nested problem - at
least a quick hack for me to see if this option has any impact on our
problem? eg., a CPU variable noting we already have the runqueue lock
and no need to grab it a second time.

David


\
 
 \ /
  Last update: 2013-06-26 19:21    [W:0.065 / U:0.356 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site