From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+w=401wt.eu-S1763517AbXGUTLJ@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1763517AbXGUTLJ (ORCPT <rfc822;w@1wt.eu>);
	Sat, 21 Jul 2007 15:11:09 -0400
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754881AbXGUTK4
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Sat, 21 Jul 2007 15:10:56 -0400
Received: from mail.screens.ru ([213.234.233.54]:43955 "EHLO mail.screens.ru"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1753324AbXGUTKz (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Sat, 21 Jul 2007 15:10:55 -0400
Date: Sat, 21 Jul 2007 23:11:04 +0400
From: Oleg Nesterov <oleg@tv-sign.ru>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>, Ingo Molnar <mingo@elte.hu>,
       linux-kernel@vger.kernel.org
Subject: Re: [PATCH] fix theoretical ccids_{read,write}_lock() race
Message-ID: <20070721191104.GA83@tv-sign.ru>
References: <20070721115712.GA871@tv-sign.ru> <20070721123159.GB1769@elte.hu> <20070721141814.GA1013@tv-sign.ru> <20070721150206.GA1083@tv-sign.ru> <20070721120235.7505338a.akpm@linux-foundation.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20070721120235.7505338a.akpm@linux-foundation.org>
User-Agent: Mutt/1.5.11
Sender: linux-kernel-owner@vger.kernel.org
X-Mailing-List: linux-kernel@vger.kernel.org

On 07/21, Andrew Morton wrote:
>
> On Sat, 21 Jul 2007 19:02:06 +0400 Oleg Nesterov <oleg@tv-sign.ru> wrote:
> 
> > Make sure that spin_unlock_wait() is properly ordered wrt atomic_inc().
> > 
> > Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
> > 
> > --- t/net/dccp/ccid.c~ccid	2006-12-18 18:17:31.000000000 +0300
> > +++ t/net/dccp/ccid.c	2007-07-21 18:29:21.000000000 +0400
> > @@ -40,6 +40,7 @@ static inline void ccids_write_unlock(vo
> >  static inline void ccids_read_lock(void)
> >  {
> >  	atomic_inc(&ccids_lockct);
> > +	smp_mb__after_atomic_inc();
> >  	spin_unlock_wait(&ccids_lock);
> >  }
> >  
> 
> Why not just use standard rwlocks in there?
> 
> (This is probably an FAQ, but it should be).

Perhaps because read_lock() doesn't allow to sleep?

Oleg.