Re: [PATCH 14/17] sched_ext: Delegate proxy donor admission to BPF schedulers

[Tejun Heo]

Hello,

On Mon, Aug 17, 2026 at 09:02:26PM +0200, Andrea Righi wrote:
...
> So there are 3 cases:
>  - task_proxy_move_active() prevents moving a donor that is currently running or
>    donating,
>  - task_cpu(D) != D->wake_cpu prevents moving an inactive donor whose scheduling
>    context has already been moved to its owner's CPU by proxy exec,
>  - an inactive donor that has not yet been proxy-migrated remains under BPF
>    placement control.

I think this mostly stems from my not understanding the full picture, so
while proxy execution is in progress - the donor's context is running the
owner - the execution cannot leave the owner's last CPU? That *seems* to be
the behavior assumed by the core code. It's a bit odd that the owner becomes
essentially CPU-pinned.

Thanks.

-- 
tejun