[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <10df58c2-cb24-4be8-846f-3e708ff04dc9@amd.com>
Date: Fri, 30 May 2025 11:06:52 +0530
From: K Prateek Nayak <kprateek.nayak@....com>
To: Aaron Lu <ziqianlu@...edance.com>, Peter Zijlstra <peterz@...radead.org>
Cc: Valentin Schneider <vschneid@...hat.com>, Ben Segall
<bsegall@...gle.com>, Josh Don <joshdon@...gle.com>,
Ingo Molnar <mingo@...hat.com>, Vincent Guittot
<vincent.guittot@...aro.org>, Xi Wang <xii@...gle.com>,
linux-kernel@...r.kernel.org, Juri Lelli <juri.lelli@...hat.com>,
Dietmar Eggemann <dietmar.eggemann@....com>,
Steven Rostedt <rostedt@...dmis.org>, Mel Gorman <mgorman@...e.de>,
Chengming Zhou <chengming.zhou@...ux.dev>,
Chuyi Zhou <zhouchuyi@...edance.com>, Jan Kiszka <jan.kiszka@...mens.com>,
Florian Bezdeka <florian.bezdeka@...mens.com>
Subject: Re: [PATCH 2/7] sched/fair: prepare throttle path for task based
throttle
Hello Aaron,
On 5/29/2025 5:21 PM, Aaron Lu wrote:
> A side note, now that check_cfs_rq_runtime() only marks cfs_rq's
> throttle status and returns a signal, it no longer does dequeuing
> stuffs, I suppose there is no need to call it in put_prev_entity()?
But perhaps it is needed to queue the bandwidth timer via
__assign_cfs_rq_runtime() if the cfs_rq had built up slack when the last
task was dequeued?
Otherwise, the throttle will only be noted at the time of pick and the
timer will only start then but check_cfs_rq_runtime() in
put_prev_entity() and check_enqueue_throttle() could have already
spotted a slack and queued the timer which can ensure bandwidth is
available sooner and reduce latency.
Let me know if I'm terribly mistaken :)
> Because that signal is now only useful in pick time and we always run
> check_cfs_rq_runtime() on every cfs_rq encountered during pick.>
> Also, check_enqueue_throttle() doesn't look useful either because
> enqueued task will go through pick and we will add a throttle work to it
> if needed. I removed these stuffs and run some tests, didn't notice
> anything wrong yet but perhaps I missed something, comments?
>
> Best regards,
> Aaron
--
Thanks and Regards,
Prateek
Powered by blists - more mailing lists