Re: [PATCH v2 sched_ext/for-7.3 24/36] sched_ext: Add reject DSQ for cap-rejected dispatches

[Tejun Heo <tj@kernel.org>]

On Mon, 06 Jul 2026 02:08:49 +0000, sashiko-bot wrote:
> - [Critical] Kernel-forced bypass to the local DSQ on offline CPUs interacts with capability rejection to create an unrecoverable infinite scheduling loop.
> - [High] Unbounded task re-rejection in `scx_reenq_reject()` creates a hardirq storm that can starve the stall watchdog and deadlock the system.
> - [Medium] Missing `p->migration_pending` bypass in `scx_local_or_reject_dsq()` leads to spurious rejections and stranded migrating tasks.

24-1 (offline) and 24-3 (migration_pending): confirmed, fixed for v3.
scx_local_or_reject_dsq() now force-admits to the local DSQ when the rq is
offline or the task is migration-disabled/migration_pending, so there's no
reject-DSQ bounce or WARN.

24-2 (unbounded re-reject storm): false positive. runnable_at is preserved
across re-enqueue, so the task ages and the stall watchdog fires and bypasses,
breaking the loop. The re-enqueue is bounded and the storm is intentional.