[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ff47f735-241d-4dc8-90b5-0b57f7148128@redhat.com>
Date: Fri, 8 Nov 2024 13:47:35 -0500
From: Waiman Long <llong@...hat.com>
To: Juri Lelli <juri.lelli@...hat.com>, Waiman Long <llong@...hat.com>
Cc: Ingo Molnar <mingo@...hat.com>, Peter Zijlstra <peterz@...radead.org>,
Vincent Guittot <vincent.guittot@...aro.org>,
Dietmar Eggemann <dietmar.eggemann@....com>,
Steven Rostedt <rostedt@...dmis.org>, Ben Segall <bsegall@...gle.com>,
Mel Gorman <mgorman@...e.de>, Valentin Schneider <vschneid@...hat.com>,
linux-kernel@...r.kernel.org, Phil Auld <pauld@...hat.com>,
Joel Fernandes <joel@...lfernandes.org>
Subject: Re: [PATCH] sched/deadline: Skip overflow check if 0 capacity
On 11/8/24 11:31 AM, Juri Lelli wrote:
> That said though, I believe I just found an additional issue. With the
> above the system doesn't crash (it did w/o it), but happily moves
> DEADLINE tasks out of a domain with a single CPU going offline. Last
> time I looked at this we were properly checking and failing the hotplug
> operation, but it was indeed a while ago, so not sure yet what changed.
> More staring.
>
> Oh, so broken, yay. 🙂
Is it the case that the null total_bw bug let the cpu offline succeeds
in a 1-cpu partition with a DL task and cause it to crash? I am also OK
if you adds the check for the presence of a DL task in the cpu and fails
the offline check in this case.
Thanks,
Longman
Powered by blists - more mailing lists