From: Jan Kara Subject: Re: [PATCH 2/2] ext2: Add blk_issue_flush() to syncing paths Date: Wed, 14 Jan 2009 18:40:23 +0100 Message-ID: <20090114174023.GK19950@duck.suse.cz> References: <1231945948-23676-1-git-send-email-jack@suse.cz> <1231945948-23676-2-git-send-email-jack@suse.cz> <496E2276.8050309@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-ext4@vger.kernel.org, Andrew Morton , linux-kernel@vger.kernel.org, pavel@suse.cz To: Eric Sandeen Return-path: Received: from styx.suse.cz ([82.119.242.94]:46749 "EHLO mail.suse.cz" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1756759AbZANRkY (ORCPT ); Wed, 14 Jan 2009 12:40:24 -0500 Content-Disposition: inline In-Reply-To: <496E2276.8050309@redhat.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Wed 14-01-09 11:35:50, Eric Sandeen wrote: > Jan Kara wrote: > > To be really safe that the data hit the platter, we should also flush drive's > > writeback caches on fsync and for O_SYNC files or O_DIRSYNC inodes. > > Seems sane, but aren't we getting really divergent behavior here between > ext2, ext3, and ext4 w.r.t. drive cache flushing for sync paths? Well, but ext3/4 should do a barrier on a transaction commit (if the user really cares about data integrity) and hence it implicitely does the same. Honza -- Jan Kara SUSE Labs, CR