Re: [PATCH] sched/deadline: Make dl-server nohz full aware

[Andrea Righi <arighi@nvidia.com>]

Hi Juri,

On Tue, May 12, 2026 at 11:02:37AM +0200, Juri Lelli wrote:
> The dl_server_timer() causes spurious IPIs on nohz_full cores, breaking
> isolation guarantees. The timer executes on a housekeeping core and
> eventually calls tick_nohz_dep_set_cpu(), sending IPIs to isolated cores
> even when only a single task is running.
> 
> The problem is that dl-servers are not coordinated with nohz_full tick
> state. Timers can fire and send IPIs to otherwise undisturbed cores.
> 
> Fix by managing servers in sched_can_stop_tick():
> 
> - When RT tasks run with CFS/SCX tasks, start the appropriate server
>   and keep the tick running
> - When only RT tasks remain, stop all servers and allow tick to stop
>   (except for >1 RR tasks which need the tick for round-robin)
> - When only CFS/SCX tasks remain, stop all servers before stopping tick
> 
> Introduce dl_servers_stop_all() to reduce duplication and abstract
> server management from core.c. Unify RT handling into one block that
> handles both RR and FIFO cases.
> 
> Fixes: 557a6bfc662c ("sched/fair: Add trivial fair server")
> Reported-by: David Haufe <dhaufe@simplextrading.com>
> Closes: https://lore.kernel.org/lkml/CAKJHwtOw_G67edzuHVtL1xC5Vyt6StcZzihtDd0yaKudW=rwVw@mail.gmail.com
> Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
> ---
> I had to modify my first original attempt at fixing this (please take a
> look at the linked report/discussion) to also take SCX into
> consideration.

As mentioned by Frederic, we don't allow to load BPF schedulers when isolcpus=
is used, so I think we can simplify the sched_can_stop_tick() part.

> 
> FYI, I temporarily pushed the script I'm using to repro and verify the
> fix here
> 
> https://github.com/jlelli/sched-deadline-tests/blob/master/test-dlserver-nohz.sh
> ---
>  kernel/sched/core.c     | 43 +++++++++++++++++++++++--------------------
>  kernel/sched/deadline.c | 14 ++++++++++++++
>  kernel/sched/sched.h    |  1 +
>  3 files changed, 38 insertions(+), 20 deletions(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index b905805bbcbe4..98759255c306b 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -1414,30 +1414,35 @@ static inline bool __need_bw_check(struct rq *rq, struct task_struct *p)
>  
>  bool sched_can_stop_tick(struct rq *rq)
>  {
> -	int fifo_nr_running;
> -
>  	/* Deadline tasks, even if single, need the tick */
>  	if (rq->dl.dl_nr_running)
>  		return false;
>  
>  	/*
> -	 * If there are more than one RR tasks, we need the tick to affect the
> -	 * actual RR behaviour.
> +	 * If there are RT tasks, we may need the tick (for >1 RR tasks),
> +	 * but we must also service lower-priority CFS/SCX tasks via dl-servers.

No need to mention SCX, maybe we can add a note that SCX is incompatible with
isolcpus, so there's no SCX task to run here.

>  	 */
> -	if (rq->rt.rr_nr_running) {
> -		if (rq->rt.rr_nr_running == 1)
> -			return true;
> -		else
> +	if (rq->rt.rt_nr_running) {
> +		if (rq->cfs.h_nr_queued) {
> +			dl_server_start(&rq->fair_server);
> +			return false;
> +		}
> +#ifdef CONFIG_SCHED_CLASS_EXT
> +		if (rq->scx.nr_running) {
> +			dl_server_start(&rq->ext_server);
> +			return false;
> +		}
> +#endif

This #ifdef block can go away.

> +		/*
> +		 * Only RT tasks, no CFS/SCX. Stop servers to prevent spurious

CFS/SCX -> CFS.

> +		 * wakeups. Tick can stop for single RR or any FIFO, but must
> +		 * run for multiple RR (round-robin behavior).
> +		 */
> +		dl_servers_stop_all(rq);
> +		if (rq->rt.rr_nr_running > 1)
>  			return false;
> -	}
> -
> -	/*
> -	 * If there's no RR tasks, but FIFO tasks, we can skip the tick, no
> -	 * forced preemption between FIFO tasks.
> -	 */
> -	fifo_nr_running = rq->rt.rt_nr_running - rq->rt.rr_nr_running;
> -	if (fifo_nr_running)
>  		return true;
> +	}
>  
>  	/*
>  	 * If there are no DL,RR/FIFO tasks, there must only be CFS or SCX tasks
> @@ -1462,6 +1467,7 @@ bool sched_can_stop_tick(struct rq *rq)
>  			return false;
>  	}
>  
> +	dl_servers_stop_all(rq);
>  	return true;
>  }
>  #endif /* CONFIG_NO_HZ_FULL */
> @@ -8810,10 +8816,7 @@ int sched_cpu_dying(unsigned int cpu)
>  		WARN(true, "Dying CPU not properly vacated!");
>  		dump_rq_tasks(rq, KERN_WARNING);
>  	}
> -	dl_server_stop(&rq->fair_server);
> -#ifdef CONFIG_SCHED_CLASS_EXT
> -	dl_server_stop(&rq->ext_server);
> -#endif
> +	dl_servers_stop_all(rq);
>  	rq_unlock_irqrestore(rq, &rf);
>  
>  	calc_load_migrate(rq);
> diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
> index edca7849b165d..c2b3d6bbe4828 100644
> --- a/kernel/sched/deadline.c
> +++ b/kernel/sched/deadline.c
> @@ -1826,6 +1826,20 @@ void dl_server_stop(struct sched_dl_entity *dl_se)
>  	dl_se->dl_server_active = 0;
>  }
>  
> +/*
> + * Stop all dl-servers on this runqueue. Called when transitioning to a state
> + * where the tick can be stopped (e.g., single RR/FIFO task, or no RT tasks).
> + * This ensures server timers are disarmed and won't cause spurious wakeups on
> + * nohz_full isolated cores.
> + */
> +void dl_servers_stop_all(struct rq *rq)
> +{
> +	dl_server_stop(&rq->fair_server);
> +#ifdef CONFIG_SCHED_CLASS_EXT
> +	dl_server_stop(&rq->ext_server);
> +#endif
> +}

And I think the dl_servers_stop_all() helper still makes sense, stopping the
ext_server is still needed in sched_cpu_dying() and calling dl_server_stop() on
an already-inactive server is harmless in the no-RT path.

Thanks,
-Andrea

> +
>  void dl_server_init(struct sched_dl_entity *dl_se, struct rq *rq,
>  		    dl_server_pick_f pick_task)
>  {
> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> index 9f63b15d309d1..26cf1d14efde5 100644
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -412,6 +412,7 @@ extern void dl_server_update_idle(struct sched_dl_entity *dl_se, s64 delta_exec)
>  extern void dl_server_update(struct sched_dl_entity *dl_se, s64 delta_exec);
>  extern void dl_server_start(struct sched_dl_entity *dl_se);
>  extern void dl_server_stop(struct sched_dl_entity *dl_se);
> +extern void dl_servers_stop_all(struct rq *rq);
>  extern void dl_server_init(struct sched_dl_entity *dl_se, struct rq *rq,
>  		    dl_server_pick_f pick_task);
>  extern void sched_init_dl_servers(void);
> 
> ---
> base-commit: 4ac4d6549a6563878d7c19c154e017f6cb7114d3
> change-id: 20260512-upstream-fix-dlserver-nohzfull-b4-b745e2a967ed
> 
> Best regards,
> --  
> Juri Lelli <juri.lelli@redhat.com>
>