linux-kernel - Re: [PATCH v3 1/3] sched: Don't account execution time for task group

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: <585ca8ea-c371-dca4-52ef-0f2d2e019774@linux.dev>
Date:   Tue, 7 Nov 2023 11:49:15 +0800
From:   Yajun Deng <yajun.deng@...ux.dev>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     mingo@...hat.com, juri.lelli@...hat.com,
        vincent.guittot@...aro.org, dietmar.eggemann@....com,
        rostedt@...dmis.org, bsegall@...gle.com, mgorman@...e.de,
        bristot@...hat.com, vschneid@...hat.com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 1/3] sched: Don't account execution time for task group


On 2023/11/6 20:49, Peter Zijlstra wrote:
> On Tue, Oct 31, 2023 at 08:59:26PM +0800, Yajun Deng wrote:
>> The rt entity can be a task group. We will account execution time for
>> each task. For consistency, we don't need to account execution time for
>> task group.
>>
>> Pass a parameter to update_current_exec_runtime, let the caller decide
>> whether account execution time.
>>
>> diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
>> index 6aaf0a3d6081..79cf80d73822 100644
>> --- a/kernel/sched/rt.c
>> +++ b/kernel/sched/rt.c
>> @@ -1018,7 +1018,8 @@ static void update_curr_rt(struct rq *rq)
>>   
>>   	trace_sched_stat_runtime(curr, delta_exec, 0);
>>   
>> -	update_current_exec_runtime(curr, now, delta_exec);
>> +	update_current_exec_runtime(curr, now, delta_exec,
>> +				    rt_entity_is_task(rt_se));
>>   
>>   	if (!rt_bandwidth_enabled())
>>   		return;
> ok, I think I've managed to confuse myself again.
>
> But at this point rt_se := &rq->curr->rt, which is *always* a task, no?


I think so, but it can be safer to use rt_entity_is_task().