Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 11 Dec 2002 03:48:10 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 11 Dec 2002 03:48:10 -0500 Received: from krusty.dt.E-Technik.Uni-Dortmund.DE ([129.217.163.1]:3090 "EHLO mail.dt.e-technik.uni-dortmund.de") by vger.kernel.org with ESMTP id ; Wed, 11 Dec 2002 03:48:09 -0500 Date: Wed, 11 Dec 2002 09:55:53 +0100 From: Matthias Andree To: lkml Cc: marcelo , Andrew Morton Subject: Re: Linux 2.4.21-pre1 Message-ID: <20021211085553.GB8740@merlin.emma.line.org> Mail-Followup-To: lkml , marcelo , Andrew Morton References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1906 Lines: 42 On Tue, 10 Dec 2002, Marcelo Tosatti wrote: > So here goes the first pre of 2.4.21 including the new IDE code merged > from Alan's tree. > > Test it carefully, since the new IDE code is not yet fully tested. > > Do not use it with critical data. Oh, speaking of "carefully", could Andrew Morton's mount -o dirsync feature patch for 2.4.21 please be merged early in v2.4.21-pre? User space (util-linux, e2fsprogs' lsattr/chattr) has been supporting it for a while, just the kernel lacked it. This -o dirsync feature is necessary to give applications that are either unaware of Linux' default -o async or that do not want to cater for deliberate incompatibilities a big performance boost: these applications (Postfix on ext2fs but not on 2.4 ext3fs or reiserfs, qmail on any Linux fs, probably a lot others I haven't even looked at) traditionally *require* chattr +S or mount -o sync to work reliably, after the merge, they could instead go with mount -o dirsync or chattr +D, which improves write speed quite a bit. Not applying the dirsync patch means that a stupid BSD ffs in default configuration will beat Linux (that needs -o sync) hands-down without any tuning the admin could do. One might also consider making -o dirsync the default while one's there, to give users a default-reliable setting, but this is going to rouse the performance fetishists who don't care about data persistence and who don't care to collect their new files from lost+found on ext2fs but rather cheat benchmarks. I can elaborate a bit more, but this has been discussed multiple times and after all, the code is there. > Summary of changes from v2.4.20 to v2.4.21-pre1 ... - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/