lkml.org 
[lkml]   [2016]   [Apr]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH] sched/cpufreq: don't trigger cpufreq update w/o real rt/deadline tasks running
From
On Thu, Apr 21, 2016 at 2:33 PM, Peter Zijlstra <peterz@infradead.org> wrote:
> On Thu, Apr 21, 2016 at 09:09:43AM +0800, Wanpeng Li wrote:
>> >> Sometimes .update_curr hook is called w/o tasks actually running, it is
>> >> captured by:
>> >>
>> >> u64 delta_exec = rq_clock_task(rq) - curr->se.exec_start;
>> >>
>> >> We should not trigger cpufreq update in this case for rt/deadline
>> >> classes, and this patch fix it.
>
>> I add a print to print when delta_exec is zero for rt class, something
>
> So its zero, so what?
>
>> like below:
>
>> watchdog/5-48 [005] d... 568.449105: update_curr_rt: rt
>> delta_exec is zero
>> watchdog/5-48 [005] d... 568.449111: <stack trace>
>> => put_prev_task_rt
>> => pick_next_task_idle
>
> So we'll go idle, but as of this point we're still running the rt task.

Skipping the update in that case might be the right thing to do, though.

It doesn't matter in 4.6-rc, because the current governors don't use
util/max anyway, so they just get an extra call they can use to
evaluate things.

However, it matters for schedutil, because it will (over)react to the
special util/max combination then. So this looks like a change to
make in 4.7.

\
 
 \ /
  Last update: 2016-04-21 19:41    [W:0.097 / U:0.144 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site