Re: [PATCH resend 1/2] sched: psi: loosen clock sync between scheduler and aggregator

[Johannes Weiner]

On Tue, Feb 03, 2026 at 10:38:09AM -0500, Johannes Weiner wrote:
> Yes. The backward motion is what caused the multi-second errors that
> immediately shift even the long-term averages. Those are blatantly
> noticable and cause practical issues. And notably, these were the ONLY
> issue anyone reported against 3840cbe24cf0 in three years.

Grr.

"[...] reported against df77430639c9 ("psi: Reduce calls to
ched_clock() in psi") in three years", of course.