Re: [PATCH] sched/fair: fix runnable_avg for throttled cfs

[Vincent Guittot]

On Thu, 27 Feb 2020 at 16:34, Phil Auld <pauld@xxxxxxxxxx> wrote:
>
> On Thu, Feb 27, 2020 at 03:58:02PM +0100 Vincent Guittot wrote:
> > On Thu, 27 Feb 2020 at 14:10, Vincent Guittot
> > <vincent.guittot@xxxxxxxxxx> wrote:
> > >
> > > On Thu, 27 Feb 2020 at 12:20, Dietmar Eggemann <dietmar.eggemann@xxxxxxx> wrote:
> > > >
> > > > On 26.02.20 21:01, Vincent Guittot wrote:
> > > > > On Wed, 26 Feb 2020 at 20:04, <bsegall@xxxxxxxxxx> wrote:
> > > > >>
> > > > >> Vincent Guittot <vincent.guittot@xxxxxxxxxx> writes:
> > > > >>
> > > > >>> When a cfs_rq is throttled, its group entity is dequeued and its running
> > > > >>> tasks are removed. We must update runnable_avg with current h_nr_running
> > > > >>> and update group_se->runnable_weight with new h_nr_running at each level
> > > >
> > > >                                               ^^^
> > > >
> > > > Shouldn't this be 'current' rather 'new' h_nr_running for
> > > > group_se->runnable_weight? IMHO, you want to cache the current value
> > > > before you add/subtract task_delta.
> > >
> > > hmm... it can't be current in both places. In my explanation,
> > > "current" means the current situation when we started to throttle cfs
> > > and "new" means the new situation after we finished to throttle the
> > > cfs. I should probably use old and new to prevent any
> > > misunderstanding.
> >
> > I'm about to send a new version to fix some minor changes: The if
> > statement should have some  { }   as there are some on the else part
> >
> > Would it be better for you if i use old and new instead of current and
> > new in the commit message ?
> >
>
> Seems better to me. You could also consider "the old" and "the new".

ok, will do
>
> Cheers,
> Phil
>
> > >
> > > That being said, we need to update runnable_avg with the old
> > > h_nr_running: the one before we started to throttle the cfs which is
> > > the value saved in group_se->runnable_weight. Once we have updated
> > > runnable_avg, we save the new h_nr_running in
> > > group_se->runnable_weight that will be used for next updates.
> > >
> > > >
> > > > >>> of the hierarchy.
> > > > >>
> > > > >> You'll also need to do this for task enqueue/dequeue inside of a
> > > > >> throttled hierarchy, I'm pretty sure.
> > > > >
> > > > > AFAICT, this is already done with patch "sched/pelt: Add a new
> > > > > runnable average signal" when task is enqueued/dequeued inside a
> > > > > throttled hierarchy
> > > > >
> > > > >>
> > > > >>>
> > > > >>> Fixes: 9f68395333ad ("sched/pelt: Add a new runnable average signal")
> > > > >>> Signed-off-by: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
> > > > >>> ---
> > > > >>> This patch applies on top of tip/sched/core
> > > > >>>
> > > > >>>  kernel/sched/fair.c | 10 ++++++++++
> > > > >>>  1 file changed, 10 insertions(+)
> > > > >>>
> > > > >>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > > > >>> index fcc968669aea..6d46974e9be7 100644
> > > > >>> --- a/kernel/sched/fair.c
> > > > >>> +++ b/kernel/sched/fair.c
> > > > >>> @@ -4703,6 +4703,11 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
> > > > >>>
> > > > >>>               if (dequeue)
> > > > >>>                       dequeue_entity(qcfs_rq, se, DEQUEUE_SLEEP);
> > > > >>> +             else {
> > > > >>> +                     update_load_avg(qcfs_rq, se, 0);
> > > > >>> +                     se_update_runnable(se);
> > > > >>> +             }
> > > > >>> +
> > > > >>>               qcfs_rq->h_nr_running -= task_delta;
> > > > >>>               qcfs_rq->idle_h_nr_running -= idle_task_delta;
> > > > >>>
> > > > >>> @@ -4772,6 +4777,11 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
> > > > >>>               cfs_rq = cfs_rq_of(se);
> > > > >>>               if (enqueue)
> > > > >>>                       enqueue_entity(cfs_rq, se, ENQUEUE_WAKEUP);
> > > > >>> +             else {
> > > > >>> +                     update_load_avg(cfs_rq, se, 0);
> > > > >>
> > > > >>
> > > > >>> +                     se_update_runnable(se);
> > > > >>> +             }
> > > > >>> +
> > > > >>>               cfs_rq->h_nr_running += task_delta;
> > > > >>>               cfs_rq->idle_h_nr_running += idle_task_delta;
> >
>
> --
>