Re: [PATCH -v2 10/17] sched: Fix migrate_disable() vs set_cpus_allowed_ptr()

[Valentin Schneider]

On 12/10/20 19:25, Tao Zhou wrote:
> Hi Peter,
> @@ -1989,7 +1994,24 @@ static int affine_move_task(struct rq *rq, struct rq_flags *rf,
>         };
>         bool complete = false;
>
> -       /* Can the task run on the task's current CPU? If so, we're done */
> +       /*
> +        * Can the task run on the task's current CPU ? If so,  we're done.
> +        * One scenario can happen here. Consider task T running on CPU P0:
> +        *
> +        *   P0                     P1                    p2

That naming convention weirds me out, and I'm the one who wrote that...

How about:

  https://gitlab.arm.com/linux-arm/linux-vs/-/commit/b30f50c7b2dd1ef70598bbe1a2a90934d894b861

I also have an extra goodie:

  https://gitlab.arm.com/linux-arm/linux-vs/-/commits/mainline/review/migrate_disable_peterz_v2_extras/

The whole thing can be cloned via
git@xxxxxxxxxxxxxxxxxx:linux-arm/linux-vs.git -b mainline/review/migrate_disable_peterz_v2_extras

(I can send those as replies to patch 0 if that helps)