Re: [PATCH] sched: sync with the cfs_rq when changing sched class

[T. Zhou]

On Sat, Aug 15, 2015 at 01:24:12PM +0900, Byungchul Park wrote:
> On Fri, Aug 14, 2015 at 08:59:02PM +0800, T. Zhou wrote:
> > Hi,
> > 
> > On Thu, Aug 13, 2015 at 02:55:55PM +0900, byungchul.park@xxxxxxx wrote:
> > > +static void attach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se)
> > > +{
> > > +	se->avg.last_update_time = cfs_rq->avg.last_update_time;
> > > +	cfs_rq->avg.load_avg += se->avg.load_avg;
> > > +	cfs_rq->avg.load_sum += se->avg.load_sum;
> > > +	cfs_rq->avg.util_avg += se->avg.util_avg;
> > > +	cfs_rq->avg.util_sum += se->avg.util_sum;
> > > +}
> > 
> > I see this function is used in enqueue_entity_load_avg() also.
> > In tip tree code, enqueue_entity_load_avg() uses cfs_rq_clock_task()
> > as se->last_update_time in migration condition.
> > Here, you use cfs_rq->avg.last_update_time as se->last_update_time.
> > If se->last_update_time is different, next decay may be different too.
> > Just from inspecting code, maybe some reasonable there?
> 
> hello zhou,
> 
> update_cfs_rq_load_avg() would update cfs_rq->avg.last_update_time to now,
> which is returned from cfs_rq_clock_task(). :)
> 
> thanks,
> byungchul
> 

Hi Byungchul,

yes, you are right. se and cfs_rq update load avg at same time. :)

thanks
-- 
Tao
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/