linux-kernel - 回复: [PATCH v2] sched/eevdf: Force propagating min_slice of cfs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID:
 <ME0P300MB0414C750ED323B988D8175708E5B2@ME0P300MB0414.AUSP300.PROD.OUTLOOK.COM>
Date: Thu, 14 Nov 2024 13:44:01 +0000
From: 解 咏梅 <xieym_ict@...mail.com>
To: Tianchen Ding <dtcccc@...ux.alibaba.com>
CC: 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>, Valentin Schneider <vschneid@...hat.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject:
 回复: [PATCH v2] sched/eevdf: Force propagating min_slice of cfs_rq when a task changing slice

Sorry I cannot change to pure text mode by my cellphone.

since commit  5e963f2bd4, there's no CFS any more as my understanding.

There's too much change. There's no preempt check in each scheduler tick. Peter introduces a hrtimer to mark the time of lasted slice used up. Preempt check only happens in wake up path.

This is mark for disable preempt
curr->vlag == curr->deadline means no preempt.

Since your fist patch for "Force propagating min_slice of cfs_rq", I read the source and found there's too much new things in   eevdf. 

Regards,
Yongmei.

________________________________________
发件人: Tianchen Ding <dtcccc@...ux.alibaba.com>
发送时间: 2024年11月14日 15:47
收件人: 解 咏梅 <xieym_ict@...mail.com>
抄送: 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>; Valentin Schneider <vschneid@...hat.com>; linux-kernel@...r.kernel.org <linux-kernel@...r.kernel.org>
主题: Re: [PATCH v2] sched/eevdf: Force propagating min_slice of cfs_rq when a task changing slice

On 2024/11/14 15:33, 解 咏梅 wrote:
> delayed dequeue is nessary for eevdf to maintain lag. Paw’s relative vruntime is
> not necessary any more in migration path.
>
>
> it is not a tuning option.
>
> regards,
> Yongmei

I don't know why you so focus on DELAY_DEQUEUE, it is not related to the case I
explained.

The case is about cgroup hierarchy. And the task_A in my case is already blocked
and *out of rq*

I'm talking about its enqueue path when woken up.