From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932166AbWGFCm1 (ORCPT ); Wed, 5 Jul 2006 22:42:27 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751053AbWGFCm1 (ORCPT ); Wed, 5 Jul 2006 22:42:27 -0400 Received: from pop5-1.us4.outblaze.com ([205.158.62.125]:19944 "HELO pop5-1.us4.outblaze.com") by vger.kernel.org with SMTP id S1751138AbWGFCm1 (ORCPT ); Wed, 5 Jul 2006 22:42:27 -0400 From: Nigel Cunningham To: Bill Davidsen Subject: Re: ext4 features Date: Thu, 6 Jul 2006 12:42:19 +1000 User-Agent: KMail/1.9.1 Cc: "J. Bruce Fields" , Theodore Tso , Thomas Glanzmann , LKML References: <20060701163301.GB24570@cip.informatik.uni-erlangen.de> <20060705214133.GA28487@fieldses.org> <44AC7647.2080005@tmr.com> In-Reply-To: <44AC7647.2080005@tmr.com> MIME-Version: 1.0 Content-Type: multipart/signed; boundary="nextPart1848871.GN2QJfhoY2"; protocol="application/pgp-signature"; micalg=pgp-sha1 Content-Transfer-Encoding: 7bit Message-Id: <200607061242.23811.ncunningham@linuxmail.org> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org --nextPart1848871.GN2QJfhoY2 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Hi. There are so many points in this conversation where I could jump in and mak= e=20 the comment I want to provide (below). Sorry if I haven't picked the best=20 one. On Thursday 06 July 2006 12:32, Bill Davidsen wrote: > No comment, I would have to see a state table to be sure I saw the races > or that there were none. With a single writer and a sinple dirty bit > there is no issue, it behaves like an elevator, more or less. With > multiple writers I bet changes are written in the order submitted rather > than the order done, but multiple writers without locks are a train > wreck waiting to happen anyway. One application I can see for this careful checking is checkpointing. IIRC,= =20 Linus recently said he'd like to see suspending to disk treated as a specia= l=20 case of checkpointing, and I can see good sense in that. But the support is= =20 just not there at the moment. An important part of implementing that would = be=20 having a filesystem where we could know exactly what the state of the=20 filesystem was at the last checkpoint, and roll back to it if necessary. Of course this would need to be tied to tracking changes in memory and to=20 writing the memory state to storage, but they're separate problems. Ext3 has a history of being the best filesystem to use in developing and=20 testing suspend to disk. It would be great if ext4 was the basis for=20 implementing serious checkpointing support. Regards, Nigel =2D-=20 Nigel, Michelle and Alisdair Cunningham 5 Mitchell Street Cobden 3266 Victoria, Australia --nextPart1848871.GN2QJfhoY2 Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQBErHiPN0y+n1M3mo0RAomGAKC4Dcn2BQmYDpersPj49g+QW3TZdACgjzRt PEIWEuQ9ULJZYoltN6zXaRk= =G+ZN -----END PGP SIGNATURE----- --nextPart1848871.GN2QJfhoY2--