linux-kernel - Re: [PATCH v4 4/6] sched: Remove double_rq_lock() from __migrate

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: <20140812082120.GO9918@twins.programming.kicks-ass.net>
Date:	Tue, 12 Aug 2014 10:21:20 +0200
From:	Peter Zijlstra <peterz@...radead.org>
To:	Kirill Tkhai <ktkhai@...allels.com>
Cc:	linux-kernel@...r.kernel.org, pjt@...gle.com, oleg@...hat.com,
	rostedt@...dmis.org, umgwanakikbuti@...il.com, tkhai@...dex.ru,
	tim.c.chen@...ux.intel.com, mingo@...nel.org,
	nicolas.pitre@...aro.org
Subject: Re: [PATCH v4 4/6] sched: Remove double_rq_lock() from
 __migrate_task()

On Wed, Aug 06, 2014 at 12:06:27PM +0400, Kirill Tkhai wrote:
> 
> Let's use ONRQ_MIGRATING instead.

I feel there should be a little more in the changelog; how about
something like:

Avoid double_rq_lock() and use ONRQ_MIGRATING for __migrate_task(). The
advantage is (obviously) not holding two 'rq->lock's at the same time
and thereby increasing parallelism.

The important point to note is that because we acquire dst->lock
immediately after releasing src->lock the potential wait time of
task_rq_lock() callers on ONRQ_MIGRATING is not longer than it would
have been in the double rq lock scenario.

> Signed-off-by: Kirill Tkhai <ktkhai@...allels.com>

Content of type "application/pgp-signature" skipped