Re: [RFC PATCH v2 13/25] sched/rt: Add HCBS related checks and operations for rt tasks

[Juri Lelli <juri.lelli@redhat.com>]

Hi!

On 31/07/25 12:55, Yuri Andriaccio wrote:
> From: luca abeni <luca.abeni@santannapisa.it>
> 
> Add checks wheter a task belongs to the root cgroup or a rt-cgroup, since HCBS
> reuses the rt classes' scheduler, and operate accordingly where needed.
> 
> Co-developed-by: Alessio Balsini <a.balsini@sssup.it>
> Signed-off-by: Alessio Balsini <a.balsini@sssup.it>
> Co-developed-by: Andrea Parri <parri.andrea@gmail.com>
> Signed-off-by: Andrea Parri <parri.andrea@gmail.com>
> Co-developed-by: Yuri Andriaccio <yurand2000@gmail.com>
> Signed-off-by: Yuri Andriaccio <yurand2000@gmail.com>
> Signed-off-by: luca abeni <luca.abeni@santannapisa.it>
> ---
>  kernel/sched/core.c     |   3 +
>  kernel/sched/deadline.c |  16 ++++-
>  kernel/sched/rt.c       | 147 +++++++++++++++++++++++++++++++++++++---
>  kernel/sched/sched.h    |   6 +-
>  kernel/sched/syscalls.c |  13 ++++
>  5 files changed, 171 insertions(+), 14 deletions(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 3a69cb906c3..6173684a02b 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -2196,6 +2196,9 @@ void wakeup_preempt(struct rq *rq, struct task_struct *p, int flags)
>  {
>  	struct task_struct *donor = rq->donor;
>  
> +	if (is_dl_group(rt_rq_of_se(&p->rt)) && task_has_rt_policy(p))
> +		resched_curr(rq);

Why this unconditional resched for tasks in groups?

...

> @@ -715,6 +744,14 @@ enqueue_task_rt(struct rq *rq, struct task_struct *p, int flags)
>  	check_schedstat_required();
>  	update_stats_wait_start_rt(rt_rq_of_se(rt_se), rt_se);
>  
> +	/* Task arriving in an idle group of tasks. */
> +	if (IS_ENABLED(CONFIG_RT_GROUP_SCHED) &&
> +	    is_dl_group(rt_rq) && rt_rq->rt_nr_running == 0) {
> +		struct sched_dl_entity *dl_se = dl_group_of(rt_rq);
> +
> +		dl_server_start(dl_se);
> +	}
> +
>  	enqueue_rt_entity(rt_se, flags);

Is it OK to start the server before the first task is enqueued in an
idle group?

...

> @@ -891,6 +936,34 @@ static void wakeup_preempt_rt(struct rq *rq, struct task_struct *p, int flags)
>  {
>  	struct task_struct *donor = rq->donor;
>  
> +	if (!rt_group_sched_enabled())
> +		goto no_group_sched;
> +

I think the below deserves a comment detailing the rules of preemption
(inside/outside groups, etc.).

> +	if (is_dl_group(rt_rq_of_se(&p->rt)) &&
> +	    is_dl_group(rt_rq_of_se(&rq->curr->rt))) {
> +		struct sched_dl_entity *dl_se, *curr_dl_se;
> +
> +		dl_se = dl_group_of(rt_rq_of_se(&p->rt));
> +		curr_dl_se = dl_group_of(rt_rq_of_se(&rq->curr->rt));
> +
> +		if (dl_entity_preempt(dl_se, curr_dl_se)) {
> +			resched_curr(rq);
> +			return;
> +		} else if (!dl_entity_preempt(curr_dl_se, dl_se)) {

Isn't this condition implied by the above?

...

> +#ifdef CONFIG_RT_GROUP_SCHED
> +static int group_push_rt_task(struct rt_rq *rt_rq)
> +{
> +	struct rq *rq = rq_of_rt_rq(rt_rq);
> +
> +	if (is_dl_group(rt_rq))
> +		return 0;
> +
> +	return push_rt_task(rq, false);
> +}
> +
> +static void group_push_rt_tasks(struct rt_rq *rt_rq)
> +{
> +	while (group_push_rt_task(rt_rq))
> +		;
> +}
> +#else
> +static void group_push_rt_tasks(struct rt_rq *rt_rq)
> +{
> +	push_rt_tasks(rq_of_rt_rq(rt_rq));
> +}
> +#endif
> +

Hummm, maybe too much for a single patch? I am a little lost at this
point. :))

Thanks,
Juri