lkml.org 
[lkml]   [2016]   [Oct]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [RFC PATCH 2/3] sched: Introduce Window-Assisted CPU utilization Tracking
On 2016-10-28 00:43, Peter Zijlstra wrote:
> On Fri, Oct 28, 2016 at 12:10:41AM -0700, Vikram Mulukutla wrote:
>> +u64 walt_ktime_clock(void)
>> +{
>> + if (unlikely(walt_ktime_suspended))
>> + return ktime_to_ns(ktime_last);
>> + return ktime_get_ns();
>> +}
>
>> +static int walt_suspend(void)
>> +{
>> + ktime_last = ktime_get();
>> + walt_ktime_suspended = true;
>> + return 0;
>> +}
>
> No, ktime_get() will not be used in the scheduler. Imagine the joy if
> that thing ends up being the HPET.

Agreed, this is an artifact from the full implementation that feeds into
the
interactive governor, and thus both needed to use the same time source.
It
shall go away.

\
 
 \ /
  Last update: 2016-10-28 09:59    [W:0.153 / U:0.336 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site