[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+HDTgSnab1eTWkFX-E9Yjvd8R3Apw4PyqbBHnYYg246Va1zYA@mail.gmail.com>
Date: Fri, 6 Oct 2023 17:20:41 -0400
From: Vineeth Pillai <vineethrp@...gle.com>
To: Shrikanth Hegde <sshegde@...ux.vnet.ibm.com>
Cc: "Joel Fernandes (Google)" <joel@...lfernandes.org>,
Suleiman Souhlal <suleiman@...gle.com>,
Hsin Yi <hsinyi@...gle.com>,
Frederic Weisbecker <frederic@...nel.org>,
"Paul E . McKenney" <paulmck@...nel.org>,
linux-kernel@...r.kernel.org, Ingo Molnar <mingo@...hat.com>,
Peter Zijlstra <peterz@...radead.org>,
Juri Lelli <juri.lelli@...hat.com>,
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>,
Daniel Bristot de Oliveira <bristot@...hat.com>,
Valentin Schneider <vschneid@...hat.com>
Subject: Re: [PATCH RFC] sched/fair: Avoid unnecessary IPIs for ILB
Hi Shrikanth,
> Hi Joel/Vineeth.
>
> Its an interesting patch.
>
> Gave this patch a try on powerpc system with 96 CPU. (12 cores SMT8)
> Was debugging an issue where ILB count goes up significantly at a specific
> busy CPU count. Haven't yet found out why. Its WIP. Was curious if this patch
> would address that issue.
>
Increased SCHED_SOFTIRQ was another problem that we noticed and we
have a separate fix for that in the chromeos tree. We were planning to
send the patch after a bit more testing. You could try the patch and
see if it solves the problem that you are seeing:
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/918f229ba2ff720f5dfae4768469acbb6ea39fe2
Thanks
Powered by blists - more mailing lists