From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx0a-00082601.pphosted.com ([67.231.145.42]:48099 "EHLO mx0a-00082601.pphosted.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750914AbaACRaZ convert rfc822-to-8bit (ORCPT ); Fri, 3 Jan 2014 12:30:25 -0500 From: Chris Mason To: "dsterba@suse.cz" CC: "wangsl.fnst@cn.fujitsu.com" , "linux-btrfs@vger.kernel.org" , CML , "sbehrens@giantdisaster.de" Subject: Re: [PATCH] Btrfs: only fua the first superblock when writting supers Date: Fri, 3 Jan 2014 17:30:05 +0000 Message-ID: <1388770244.29987.0.camel@ret.masoncoding.com> References: <1388744577-13968-1-git-send-email-wangsl.fnst@cn.fujitsu.com> <20140103170350.GP6498@twin.jikos.cz> In-Reply-To: <20140103170350.GP6498@twin.jikos.cz> Content-Type: text/plain; charset="utf-7" MIME-Version: 1.0 Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Fri, 2014-01-03 at 18:03 +-0100, David Sterba wrote: +AD4- On Fri, Jan 03, 2014 at 06:22:57PM +-0800, Wang Shilong wrote: +AD4- +AD4- We only intent to fua the first superblock in every device from +AD4- +AD4- comments, fix it. +AD4- +AD4- Good catch, this could gain some speedup when there are up to 2 less +AD4- flushes. +AD4- +AD4- There's one thing that's a different from currnet behaviour: +AD4- Without this patch, all the superblocks are written with FUA, now only +AD4- the first one, so my question is what if the first fails and the others +AD4- succeed but do not get flushed immediatelly? +AD4- +AD4- This is more of a theoretical scenario, and if the 1st superblock write +AD4- fails more serious problems can be expected. But let's say the write +AD4- error of 1st is transient, do you or others think that it's reasonable +AD4- to try to write all the remainig sb's with FUA? Not a bad idea, if we get a failure on the first SB, fua the others? I think it does make sense to do the others non-fua, just because they only get used in emergencies anyway. -chris