From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752110AbZHXJ3J (ORCPT ); Mon, 24 Aug 2009 05:29:09 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752069AbZHXJ3G (ORCPT ); Mon, 24 Aug 2009 05:29:06 -0400 Received: from cantor2.suse.de ([195.135.220.15]:39656 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751309AbZHXJ3F (ORCPT ); Mon, 24 Aug 2009 05:29:05 -0400 Date: Mon, 24 Aug 2009 11:29:05 +0200 From: Jan Kara To: Jamie Lokier Cc: Jan Kara , LKML , hch@lst.de, linux-fsdevel@vger.kernel.org Subject: Re: [PATCH 0/17] Make O_SYNC handling use standard syncing path (Version 2) Message-ID: <20090824092905.GA10080@duck.novell.com> References: <1250875447-15622-1-git-send-email-jack@suse.cz> <20090822162711.GA7542@shareable.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090822162711.GA7542@shareable.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 Sat 22-08-09 17:27:11, Jamie Lokier wrote: > Jan Kara wrote: > > The patch set unifines O_SYNC handling with standard fsync() path. After this, > > we have just one place forcing a single file to disk so filesystems like ext3 / > > ext4 don't have to force a transaction commit in ext?_file_write for O_SYNC > > files / IS_SYNC inodes. The code is also cleaner this way (actually about 150 > > lines shorter), we don't sync the inode several times as it happened previously > > etc. > > Afaik, O_SYNC requires just the written data to be committed to disk, > but fsync() requires all dirty data for the file (including written by > other processes / descriptors) to be committed to disk. > > So doing the equivalent of fsync() after write might be the wrong > thing to do. It's OK, we essentially end up doing fdatasync() on the range of the file where the write happened. So no unnecessary writing happens. Honza -- Jan Kara SUSE Labs, CR