linux-kernel - Re: [PATCH] sched: put rq's sched_avg under CONFIG_FAIR_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] [thread-next>] [day] [month] [year] [list]

Message-ID: <xm26lhwzhsqh.fsf@sword-of-the-dawn.mtv.corp.google.com>
Date:	Tue, 25 Feb 2014 11:55:50 -0800
From:	bsegall@...gle.com
To:	Dietmar Eggemann <dietmar.eggemann@....com>
Cc:	Srikar Dronamraju <srikar@...ux.vnet.ibm.com>,
	Peter Zijlstra <peterz@...radead.org>,
	"linux-kernel\@vger.kernel.org" <linux-kernel@...r.kernel.org>,
	pjt@...gle.com
Subject: Re: [PATCH] sched: put rq's sched_avg under CONFIG_FAIR_GROUP_SCHED

Dietmar Eggemann <dietmar.eggemann@....com> writes:

> On 25/02/14 13:16, Srikar Dronamraju wrote:
>>> The struct sched_avg of struct rq is only used in case group
>>> scheduling is enabled inside __update_tg_runnable_avg() to update
>>> per-cpu representation of a task group.  I.e. that there is no need to
>>> maintain the runnable avg of a rq in the !CONFIG_FAIR_GROUP_SCHED case.
>>>
>>> This patch guards struct sched_avg of struct rq and
>>> update_rq_runnable_avg() with CONFIG_FAIR_GROUP_SCHED.
>>>

Reviewed-by: Ben Segall <bsegall@...gle.com>

If we ever decide to use runnable_avg_sum/period in balance or
something, it's trivial enough to move it back, and there is no point in
calculating unusued stats until then.

>> 
>> While this patch looks good, I see fields in sched_avg viz decay_count,
>> last_runnable_update, load_avg_contrib  only relevant to sched_entity.
>> i.e they don't seem to be updated or used for rq->avg. Should we look at
>> splitting sched_avg so that rq->avg doesn't have unwanted fields?
>
> Yes, AFAICS at least load_avg_contrib and decay_count are only relevant
> for struct sched_entity whereas last_runnable_update is used in
> __update_entity_runnable_avg() itself.
>
> By having struct sched_avg embedded into struct sched_entity and struct
> rq, __update_entity_runnable_avg() and __update_tg_runnable_avg() can be
> used on both and moreover, all current members of struct sched_avg
> belong logically together.
>
> With this patch I was more interested in the fact that we do not have to
> maintain the load avg for struct rq in a !CONFIG_FAIR_GROUP_SCHED system.
>

Yeah, last_runnable_update and runnable_avg_sum/period are used,
decay_count and load_avg_contrib could be moved to the se just fine, and
won't cause any problems.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/