Re: [External] Re: [RFC PATCH 5/7] sched/fair: Take care of group/affinity/sched_class change for throttled task

From: Aaron Lu
Date: Fri Mar 14 2025 - 07:40:25 EST


On Fri, Mar 14, 2025 at 10:21:15AM +0530, K Prateek Nayak wrote:
> Hello Aaron,
>
> On 3/13/2025 12:51 PM, Aaron Lu wrote:
> [..snip..]
>
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -5876,8 +5876,8 @@ static void throttle_cfs_rq_work(struct
> > callback_head *work)
> >
> > update_rq_clock(rq);
> > WARN_ON_ONCE(!list_empty(&p->throttle_node));
> > - list_add(&p->throttle_node, &cfs_rq->throttled_limbo_list);
> > dequeue_task_fair(rq, p, DEQUEUE_SLEEP | DEQUEUE_SPECIAL);
> > + list_add(&p->throttle_node, &cfs_rq->throttled_limbo_list);
> > resched_curr(rq);
>
> nit. Perhaps this bit can be moved to Patch 2 to consolidate all
> changes in throttle_cfs_rq_work()

No problem.

I placed it here to better illustrate why list_add() has to be done
after dequeue_task_fair().

Thanks,
Aaron

> >
> > out_unlock:
>
> [..snip..]
>
> --
> Thanks and Regards,
> Prateek
>