Re: [PATCH v2 2/2] sched_ext: Fix cpu_released while changing sched policy of the running task

['Tejun Heo' <tj@kernel.org>]

Hello,

On Sat, Jun 28, 2025 at 07:20:59AM +0000, liuwenfang wrote:
>  static void switched_from_scx(struct rq *rq, struct task_struct *p)
>  {
> +	switch_class(rq, p, true);
> +
>  	scx_ops_disable_task(p);
>  }

Hmm... but this function can be called when @p is not currently running from
setscheduler() path, and it wouldn't make sense to call switch_class()
during that.

>  static void wakeup_preempt_scx(struct rq *rq, struct task_struct *p,int wake_flags) {}
> -static void switched_to_scx(struct rq *rq, struct task_struct *p) {}
> +static void switched_to_scx(struct rq *rq, struct task_struct *p)
> +{
> +	lockdep_assert_rq_held(rq);
> +
> +	if (static_branch_unlikely(&scx_ops_cpu_preempt) &&
> +	    unlikely(rq->scx.cpu_released)) {
> +		/*
> +		 * If the previous sched_class for the current CPU was not SCX,
> +		 * notify the BPF scheduler that it again has control of the
> +		 * core. This callback complements ->cpu_release(), which is
> +		 * emitted in switch_class().
> +		 */
> +		if (SCX_HAS_OP(cpu_acquire))
> +			SCX_CALL_OP(SCX_KF_REST, cpu_acquire, rq, cpu_of(rq), NULL);
> +		rq->scx.cpu_released = false;
> +	}
> +}

Ditto. This should only apply if @p is current, right?

Thanks.

-- 
tejun