Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:7 as permitted sender) client-ip=2620:137:e000::3:7;
Message-ID: <446aeb22-55c4-4ef6-81cc-1d1d994d5268@gmail.com>
Date:   Thu, 16 Nov 2023 14:51:29 +0800
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH 1/4] sched/eevdf: Fix vruntime adjustment on reweight
Content-Language: en-US
To:     Abel Wu <wuyun.abel@bytedance.com>
Cc:     Barry Song <21cnbao@gmail.com>,
        Benjamin Segall <bsegall@google.com>,
        Chen Yu <yu.c.chen@intel.com>,
        Daniel Jordan <daniel.m.jordan@oracle.com>,
        "Gautham R . Shenoy" <gautham.shenoy@amd.com>,
        Joel Fernandes <joel@joelfernandes.org>,
        K Prateek Nayak <kprateek.nayak@amd.com>,
        Mike Galbraith <efault@gmx.de>,
        Qais Yousef <qyousef@layalina.io>,
        Tim Chen <tim.c.chen@linux.intel.com>,
        Yicong Yang <yangyicong@huawei.com>,
        Youssef Esmat <youssefesmat@chromium.org>,
        linux-kernel@vger.kernel.org,
        Peter Zijlstra <peterz@infradead.org>,
        Ingo Molnar <mingo@kernel.org>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        Valentin Schneider <valentin.schneider@arm.com>
References: <20231107090510.71322-1-wuyun.abel@bytedance.com>
 <20231107090510.71322-2-wuyun.abel@bytedance.com>
 <da42f1d7-adfe-485e-987e-3e8dae78b4c2@gmail.com>
 <23cebdf8-1a51-481e-a8c2-e17021099e05@bytedance.com>
 <19c6accf-1cc0-489d-9c82-e0ddf982b098@bytedance.com>
From:   Yiwei Lin <s921975628@gmail.com>
In-Reply-To: <19c6accf-1cc0-489d-9c82-e0ddf982b098@bytedance.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 11/16/23 13:07, Abel Wu wrote:
> On 11/16/23 12:48 PM, Abel Wu Wrote:
>> On 11/15/23 11:36 PM, Yiwei Lin Wrote:
>>>
>>>> @@ -3712,8 +3811,17 @@ static void reweight_entity(struct cfs_rq 
>>>> *cfs_rq, struct sched_entity *se,
>>>>       enqueue_load_avg(cfs_rq, se);
>>>>       if (se->on_rq) {
>>>>           update_load_add(&cfs_rq->load, se->load.weight);
>>>> -        if (cfs_rq->curr != se)
>>>> -            avg_vruntime_add(cfs_rq, se);
>>>> +        if (!curr) {
>>>> +            /*
>>>> +             * The entity's vruntime has been adjusted, so let's 
>>>> check
>>>> +             * whether the rq-wide min_vruntime needs updated too. 
>>>> Since
>>>> +             * the calculations above require stable min_vruntime 
>>>> rather
>>>> +             * than up-to-date one, we do the update at the end of 
>>>> the
>>>> +             * reweight process.
>>>> +             */
>>>> +            __enqueue_entity(cfs_rq, se);
>>>> +            update_min_vruntime(cfs_rq);
>>>> +        }
>>>>       }
>>>>   }
>>> Sorry if I am asking stupid question...... It looks like 
>>> reweight_entity() may have chance to change the weight of 
>>> cfs_rq->curr entity, but we'll never update_min_vruntime() when 
>>> reweighting it. Is there any reason that we can skip the 
>>> update_min_vruntime() for this case?
>>
>> No, you are right!
>
> I was intended to update_min_vruntime() if se->on_rq and no matter
> it is curr or not, just as you suggested. But after a second thought
> I wonder if it is necessary to update *NOW*, since we will always
> update_curr() before making any change to cfs_rq. Thoughts?
I lost the fact that we'll update_min_vruntime() every time we 
update_curr(). Because of this fact, we can indeed wait until we need 
the correct min_vruntime and update_min_vruntime() then. The only 
consideration that I came up with is that the sched_debug may not be 
able to reflect the accurate min_vruntime in time. But this may not be a 
big problem.

Further, I have another advanced thought we can remove the 
update_min_vruntime() here in the reweight_entity() directly to save 
more time. The reason that I think this is because min_vruntime is not 
for normalization of vruntime as before which is required on CFS, so we 
will always update_curr() for the latest min_vruntime before using it. 
Also, the update_min_vruntime() in dequeue_entity() may also be removed 
as the reason, i.e. just do update_min_vruntime() in update_curr() to 
simplify. What do you think?

Thanks,
Yiwei Lin