From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: [e2fsprogs] Implement resync of declared blocks for software RAID Date: Tue, 08 Dec 2009 21:26:35 +0100 Message-ID: <87bpi9xldw.fsf@basil.nowhere.org> References: <20091208183046.GH4508@clouds> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-ext4@vger.kernel.org To: Jody McIntyre Return-path: Received: from one.firstfloor.org ([213.235.205.2]:53102 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S965956AbZLHU0c (ORCPT ); Tue, 8 Dec 2009 15:26:32 -0500 In-Reply-To: <20091208183046.GH4508@clouds> (Jody McIntyre's message of "Tue, 08 Dec 2009 13:30:48 -0500") Sender: linux-ext4-owner@vger.kernel.org List-ID: Jody McIntyre writes: > This patch resyncs declared blocks on journal recovery. This must be done > as part of journal replay for filesystems with > JFS_FEATURE_INCOMPAT_DECLARE_BLOCKS; we have previously guaranteed to MD > that we will resync any blocks that may have been undergoing writes at > the time of a system crash. > > The SET_RESYNC_ALL and CLEAR_RESYNC_ALL ioctls are used to instruct MD > to resync all blocks being read and written. > > This patch is UNTESTED and is being sent only because I am no longer > working on declared mode. Is this the only e2fsprogs patch needed for this feature, apart from the kernel patches you posted? Too bad you're not working on it anymore, I definitely liked the idea. -Andi -- ak@linux.intel.com -- Speaking for myself only.