Re: [PATCH v3 7/7] sched/eevdf: Move to a single runqueue

[Mike Galbraith]

On Mon, 2026-11-02 at 15:58 +0800, Chengfeng Lin wrote:
> 
> If you make further changes in this area, I'd be happy to rerun the
> context-switch benchmark across the full range of worker counts and
> share the results.

I'm sure this little regression will get fixed up as the dust settles.
My role was the easy one, that of hound.. all you need is spare cycles.

The good news from hunting excursion is that what the reporter's probe
did was to validate the DELAY_DEQUEUE mechanism as being a genuine QOS
improvement.  The reported QOS regression was per measurement the
DELAY_DEQUEUE progression going missing when all cycles expended
between the most recent inflection point event and pulling of payout
lever were banked nanoseconds before the real deal inflection point,
the preemption decision.

	-Mike