Re: Question about a8ea6fc9b089 ("sched: Stop PF_NO_SETAFFINITY from being inherited by various init system threads")

[Valentin Schneider]

On 10/06/21 10:04, Paul E. McKenney wrote:

Hi,
> Hello, Frederic,
>
> This commit works well, but has the unfortunate side-effect of making
> smp_processor_id() complain when used in a preemptible region even
> though the kthread has been pinned onto a single CPU by a call to
> set_cpus_allowed_ptr().  (Which did return success.)
>

On which tree are you encountering this?

Looking at check_preemption_disabled() and CPU affinity, v5.13-rc5 has:

        /*
         * Kernel threads bound to a single CPU can safely use
         * smp_processor_id():
         */
        if (current->nr_cpus_allowed == 1)
                goto out;

tip/sched/core additionally hinges that on PF_NO_SETAFFINITY:

  570a752b7a9b ("lib/smp_processor_id: Use is_percpu_thread() instead of nr_cpus_allowed")

The former shouldn't be affected by Frederic's patch, and the latter should
only cause warnings if the pinned task isn't a "proper" kthread (thus
doesn't have PF_NO_SETAFFINITY)... Exceptions that come to mind are things
like UMH which doesn't use kthread_create().

> This isn't a big deal -- I can easily switch to raw_smp_processor_id(),
> which is arguably a better choice anyway because it prevents the
> complaints from flooding out any real warnings due to error returns
> from set_cpus_allowed_ptr() or something else unpinning the kthread.
> Which I am in the process of doing:
>
> 516e52e9f5ec ("scftorture: Avoid excess warnings")
> 475d6d49f21d ("refscale: Avoid excess warnings in ref_scale_reader()")
>
> But I figured that I should check to see if this change was in fact
> intentional.
>
>                                                       Thanx, Paul