From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phillip Susi Subject: Re: mkfs.ext4 vs. e2fsck discard oddities Date: Fri, 09 Mar 2012 10:14:18 -0500 Message-ID: <4F5A1E4A.9060804@ubuntu.com> References: <4F4D1020.5060204@redhat.com> <34CA4C93-194F-45F3-9A94-E181A7DFEBCE@mit.edu> <20120301143834.GA30578@thunk.org> <4F58E2D7.205@ubuntu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Ted Ts'o , Eric Sandeen , ext4 development To: Lukas Czerner Return-path: Received: from cdptpa-omtalb.mail.rr.com ([75.180.132.120]:4337 "EHLO cdptpa-omtalb.mail.rr.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754380Ab2CIPOV (ORCPT ); Fri, 9 Mar 2012 10:14:21 -0500 In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On 3/9/2012 3:59 AM, Lukas Czerner wrote: > Why would we try to check UNINIT groups with valid descriptor checksums > ? I think that this problem will be solved with BLOCK_DISCARDED flag as > we discussed with Ted in another thread. No need to have yet another > option so it is win-win :) Because not skipping a specific action on an uninitialized group ( discard ) is a specific case of the more general form of not skipping uninitialized groups. I thought that it might sometimes be useful to actually verify the group is correct instead of trusting the uninit flag, especially if you are about to discard it. Also any other things that are added in the future and skipped for uninit groups would not need yet another flag to specifically not skip that action, since it will be covered by the more general flag already. Also the way the code was structured it looked like it would be much simpler to bypass the skip and do the full check of the uninit group than to modify it to discard the group even though checking it was skipped.