From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1422801AbWLPXZc (ORCPT ); Sat, 16 Dec 2006 18:25:32 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1422802AbWLPXZc (ORCPT ); Sat, 16 Dec 2006 18:25:32 -0500 Received: from extu-mxob-1.symantec.com ([216.10.194.28]:39881 "EHLO extu-mxob-1.symantec.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1422801AbWLPXZb (ORCPT ); Sat, 16 Dec 2006 18:25:31 -0500 X-Greylist: delayed 1043 seconds by postgrey-1.27 at vger.kernel.org; Sat, 16 Dec 2006 18:25:31 EST X-AuditID: d80ac21c-98b9cbb0000069a0-88-45847c58cba1 Date: Sat, 16 Dec 2006 23:08:24 +0000 (GMT) From: Hugh Dickins X-X-Sender: hugh@blonde.wat.veritas.com To: Peter Zijlstra cc: Martin Michlmayr , Marc Haber , Jan Kara , linux-mm@kvack.org, linux-kernel@vger.kernel.org Subject: Re: 2.6.19 file content corruption on ext3 In-Reply-To: <1166304581.10372.18.camel@twins> Message-ID: References: <20061207155740.GC1434@torres.l21.ma.zugschlus.de> <4578465D.7030104@cfl.rr.com> <20061209092639.GA15443@torres.l21.ma.zugschlus.de> <20061216184310.GA891@unjust.cyrius.com> <1166304581.10372.18.camel@twins> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-OriginalArrivalTime: 16 Dec 2006 23:08:07.0802 (UTC) FILETIME=[0C7A05A0:01C72167] X-Brightmail-Tracker: AAAAAA== Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sat, 16 Dec 2006, Peter Zijlstra wrote: > Moving the cleaning of the page out from under the private_lock opened > up a window where newly attached buffer might still see the page dirty > status and were thus marked (incorrectly) dirty themselves; resulting in > filesystem data corruption. I'm not going to pretend to understand the buffers issues here: people thought that change was safe originally, and I can't say it's not - it just stood out as a potentially weakening change. The patch you propose certainly looks like a good way out, if that moved unlock really is a problem: your patch is very well worth trying by those people seeing their corruption problems, let's wait to hear their feedback. Thanks! Hugh