From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from sandeen.net ([63.231.237.45]:47310 "EHLO sandeen.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727453AbeGXWva (ORCPT ); Tue, 24 Jul 2018 18:51:30 -0400 Subject: Re: [PATCH 1/2 V3] xfs_repair: notify user if free inodes contain errors From: Eric Sandeen References: <75f628bc-e8e7-b4fd-0bbb-fb3ca0fabac8@sandeen.net> <3a14c161-a476-58a9-1a13-c2225f5bbb51@sandeen.net> Message-ID: <39c6e502-df2f-6b0f-839e-1ee36cd57131@sandeen.net> Date: Tue, 24 Jul 2018 14:43:00 -0700 MIME-Version: 1.0 In-Reply-To: <3a14c161-a476-58a9-1a13-c2225f5bbb51@sandeen.net> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Eric Sandeen , linux-xfs On 7/23/18 9:35 PM, Eric Sandeen wrote: > On 7/23/18 7:34 PM, Eric Sandeen wrote: >> xfs_repair checks allocated but unused (free) inodes in on-disk clusters, >> and up until now silently repairs any errors, and as a result does not >> alter exit status if errors are found. >> >> The in-kernel verifiers will be noisy about these errors and instruct >> the user to run repair, so it's best if repair is explicit about any >> fixes it makes as a result. >> >> To ensure we catch anything the kernel would complain about, re-use >> xfs_dinode_verify to determine whether we must clear a free inode. >> >> Note, however, that the verifier contains only a subset of the checks >> currently in clear_dinode. This should be ok; if it's not, the checks >> should be added to the verifier in any case. >> >> Signed-off-by: Eric Sandeen > Ugh, dammit, pulled the trigger too soon, tests were looking fine until > the last few. Need to look at this more. With the free inode extent validator fix I sent earlier today I think it's safe to resume normal review activities for these 2 patches. :) Thanks, -Eric