From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ipmail06.adl2.internode.on.net ([150.101.137.129]:49794 "EHLO ipmail06.adl2.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933267AbcJZVPo (ORCPT ); Wed, 26 Oct 2016 17:15:44 -0400 Date: Thu, 27 Oct 2016 08:15:40 +1100 From: Dave Chinner Subject: Re: [PATCH 28/39] xfs_repair: handle multiple owners of data blocks Message-ID: <20161026211540.GL23194@dastard> References: <147743661772.11035.560864407573832590.stgit@birch.djwong.org> <147743679506.11035.9193078348507369875.stgit@birch.djwong.org> <20161026105758.GC8254@infradead.org> <20161026171553.GC26572@birch.djwong.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20161026171553.GC26572@birch.djwong.org> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: "Darrick J. Wong" Cc: Christoph Hellwig , linux-xfs@vger.kernel.org On Wed, Oct 26, 2016 at 10:15:53AM -0700, Darrick J. Wong wrote: > On Wed, Oct 26, 2016 at 03:57:58AM -0700, Christoph Hellwig wrote: > > > --- a/repair/dinode.c > > > +++ b/repair/dinode.c > > > @@ -722,6 +722,9 @@ _("Fatal error: inode %" PRIu64 " - blkmap_set_ext(): %s\n" > > > * checking each entry without setting the > > > * block bitmap > > > */ > > > + if (type == XR_INO_DATA && > > > + xfs_sb_version_hasreflink(&mp->m_sb)) > > > + goto skip_dup; > > > if (search_dup_extent(agno, agbno, ebno)) { > > > do_warn( > > > _("%s fork in ino %" PRIu64 " claims dup extent, " > > > @@ -731,6 +734,7 @@ _("%s fork in ino %" PRIu64 " claims dup extent, " > > > irec.br_blockcount); > > > goto done; > > > } > > > +skip_dup: > > > > For some weird reason this goto makes me sad. I'm getting into > > major nitpick terretory here, but can we avoid it? Either duplicate > > the *tot increment or find some funky condition? > > > > Otherwise this looks fine. > > Ugh, unnecessary goto. Sad! ;) > > But in all seriousness an 'else if' could do the same work, so: > > if (type == XR_INO_DATA && > xfs_sb_version_hasreflink(&mp->m_sb)) > ; /* avoid the dup extent check below */ > else if (search_dup_extent(agno, agbno, ebno)) { > do_warn(...); > } > *tot += irec.br_blockcount; Even that is unnecessary, yes? if ((type != XR_INO_DATA || !xfs_sb_version_hasreflink(&mp->m_sb)) && search_dup_extent(agno, agbno, ebno) { do_warn(...); } Cheers, Dave. -- Dave Chinner david@fromorbit.com