From: Ying Han Subject: Re: [PATCH] ext2: Fix data corruption for racing writes Date: Fri, 10 Apr 2009 11:10:41 -0700 Message-ID: <604427e00904101110j68e0bd31t885e5d60b8842d80@mail.gmail.com> References: <1238715399-22172-1-git-send-email-jack@suse.cz> <20090406102329.GD31189@duck.suse.cz> <20090409133030.a7ea94c7.akpm@linux-foundation.org> <604427e00904091359t3e068ffm1868d164824fb575@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Jan Kara , yinghan@gmail.com, linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org, nickpiggin@yahoo.com.au To: Andrew Morton Return-path: Received: from smtp-out.google.com ([216.239.33.17]:47789 "EHLO smtp-out.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1760084AbZDJSKp (ORCPT ); Fri, 10 Apr 2009 14:10:45 -0400 In-Reply-To: <604427e00904091359t3e068ffm1868d164824fb575@mail.gmail.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu, Apr 9, 2009 at 1:59 PM, Ying Han wrote: > On Thu, Apr 9, 2009 at 1:30 PM, Andrew Morton wrote: >> On Mon, 6 Apr 2009 12:23:29 +0200 >> Jan Kara wrote: >> >>> > >>> > I tried this patch and seems i got deadlock on the truncate_mutex. >>> > Here is the message after enabling lockdep. I pasted the same message >>> > on the origianal thread. >>> >>> ... >>> >>> I don't think this is a deadlock (or is the machine hung?). The thread >>> was just waiting for a long time. I'd think that you'll occasionally get >>> exactly the same message even without my patch if you stress the machine >>> like you do. >>> >> >> Well, it's easy to tell the difference >> >> deadlock: system never recovers >> long-sucky-delay: system eventually recovers. >> >> Which was it?? > Guess i have to retest it. I didn't wait long enough to see what > happened on the machine. However, i do see the machine got rebooted > later, but i am not sure what is the reason it is got reboot. > > I will rerun the patch and keep an eye on it this time. Ok, i rerun the patch and i don't see the message poping this time and the machine is up healthy. > > --Ying >> >