Messages in this thread | | | Date | Tue, 5 Apr 2016 19:17:39 +0200 | From | luca abeni <> | Subject | Re: [RFC v2 3/7] Improve the tracking of active utilisation |
| |
On Tue, 5 Apr 2016 16:48:03 +0200 Peter Zijlstra <peterz@infradead.org> wrote:
> On Fri, Apr 01, 2016 at 05:12:29PM +0200, Luca Abeni wrote: > > + /* > > + * We cannot use inactive_task_timer() to invoke sub_running_bw() > > + * at the 0-lag time, because the task could have been migrated > > + * while SCHED_OTHER in the meanwhile. > > + */ > > + if (hrtimer_active(&p->dl.inactive_timer) && > > + !hrtimer_callback_running(&p->dl.inactive_timer)) > > sub_running_bw(&p->dl, &rq->dl); > > hrtimer_is_queued() ?
Uhm... I do not remember why I used this condition, but hrtimer_is_queued() should be the right thing to be used, yes... I'll update the patch.
Thanks, Luca
| |