linux-kernel - Re: [PATCH] workqueue: remove the unneeded cpu_relax() in __queue

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: <20140526105443.GA5555@mtj.dyndns.org>
Date:	Mon, 26 May 2014 06:54:43 -0400
From:	Tejun Heo <tj@...nel.org>
To:	Lai Jiangshan <laijs@...fujitsu.com>
Cc:	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] workqueue: remove the unneeded cpu_relax() in
 __queue_work()

On Mon, May 26, 2014 at 01:27:55PM +0800, Lai Jiangshan wrote:
> changing pwq:
> 	install pwq
> 	lock(pool->lock)
> 	put_pwq();
> 	unlock(pool->lock)
> 
> __queue_work():
> 	lock(pool->lock)
> 	test ref and find it zero;
> 	see the installation here;
> 	it is guaranteed to get the installed pwq on the immediate next try.
> 	unlock()
> 	retry.

The fact that pool->lock locking happens to provide enough barrier for
the above to work is an accidental implementation detail.  We
theoretically can move refcnting out of pool->lock.  Nothing
semantically guarantees that barrier to be there to interlock pwq
qinstallation and the last put.  Removing that cpu_relax() doesn't buy
us *ANYTHING* and removing that with rationale of making it go faster
would easily win pointless micro optimization award of the year.  Just
let it go.

-- 
tejun
--
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/