linux-kernel - Re: [PATCH] sched/fair: Track current se's EEVDF parameters

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: <7537e534-9616-4f45-9974-399453aee276@gmail.com>
Date:   Thu, 2 Nov 2023 21:31:15 +0800
From:   Yiwei Lin <s921975628@...il.com>
To:     Abel Wu <wuyun.abel@...edance.com>, mingo@...hat.com,
        peterz@...radead.org
Cc:     vincent.guittot@...aro.org, dietmar.eggemann@....com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] sched/fair: Track current se's EEVDF parameters


On 11/2/23 14:27, Abel Wu wrote:
> On 11/2/23 1:44 AM, s921975628@...il.com Wrote:
>> From: Yiwei Lin <s921975628@...il.com>
>>
>> After dequeuing the current-picked scheduling entity with
>> `__dequeue_entity`, its contribution to the EEVDF parameters
>> cfs_rq->avg_vruntime and cfs_rq->avg_load are also removed.
>> Because these should in fact be considered for the EEVDF algorithm,
>> we took curr as the special case and inserted back the contributions
>> when requests for cfs_rq->avg_vruntime and cfs_rq->avg_load.
>
> Being 'curr' means its vruntime is increasing, so does its
> contribution to avg_vruntime. And you failed to explain the
> most important part that how to commit its contribution to
> avg_vruntime (specifically in update_curr()).
>
> Regards,
>     Abel

Thanks for pointing out the missing part.  I should remember to commit back
the latest contribution at `update_curr()`. In other words, adding
delta * scale_load_down(curr->load.weight) to avg_vruntime.

Thanks,
     Yiwei Lin