From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932326AbZHUPsa (ORCPT ); Fri, 21 Aug 2009 11:48:30 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932253AbZHUPs3 (ORCPT ); Fri, 21 Aug 2009 11:48:29 -0400 Received: from cantor.suse.de ([195.135.220.2]:47908 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932196AbZHUPs3 (ORCPT ); Fri, 21 Aug 2009 11:48:29 -0400 Date: Fri, 21 Aug 2009 17:48:29 +0200 From: Jan Kara To: Christoph Hellwig Cc: Jan Kara , LKML , Evgeniy Polyakov , ocfs2-devel@oss.oracle.com, Joel Becker , Felix Blyakher , xfs@oss.sgi.com, Anton Altaparmakov , linux-ntfs-dev@lists.sourceforge.net, OGAWA Hirofumi , linux-ext4@vger.kernel.org, tytso@mit.edu Subject: Re: [PATCH 07/17] vfs: Introduce new helpers for syncing after writing to O_SYNC file or IS_SYNC inode Message-ID: <20090821154829.GE3007@duck.novell.com> References: <1250697884-22288-1-git-send-email-jack@suse.cz> <1250697884-22288-8-git-send-email-jack@suse.cz> <20090819162638.GE6150@infradead.org> <20090820121531.GC16486@duck.novell.com> <20090820162729.GA24659@infradead.org> <20090821152339.GD3007@duck.novell.com> <20090821153235.GA5874@infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090821153235.GA5874@infradead.org> User-Agent: Mutt/1.5.17 (2007-11-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri 21-08-09 11:32:35, Christoph Hellwig wrote: > On Fri, Aug 21, 2009 at 05:23:39PM +0200, Jan Kara wrote: > > Hmm, I can imagine it would buy us something in two cases (but looking at > > the code, neither is implemented in such a way that it would really help > > us in any way): > > 1) when an inode and it's data are stored in one block (e.g. OCFS2 or UDF) do > > this. > > > In the first case we would wait for block with data to be written only to > > submit it again because inode was still dirty. > > But we would not actually store in the pagecache or at least never > writeback it via the pacache in that case, but always just write it as > part of the inode. Yes, probably. Actually, that's what UDF does. It's just that then the sequence fdatawrite(), fdatawait() does not really send the data to disk. But that probably does not matter. Honza -- Jan Kara SUSE Labs, CR