From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ipmail07.adl2.internode.on.net ([150.101.137.131]:34134 "EHLO ipmail07.adl2.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750832AbcJZD0k (ORCPT ); Tue, 25 Oct 2016 23:26:40 -0400 Date: Wed, 26 Oct 2016 14:26:38 +1100 From: Dave Chinner Subject: Re: [PATCH 10/39] xfs_db: add support for checking the refcount btree Message-ID: <20161026032638.GI23194@dastard> References: <147743661772.11035.560864407573832590.stgit@birch.djwong.org> <147743667991.11035.9533585332530820648.stgit@birch.djwong.org> <20161026004920.GD23194@dastard> <20161026011308.GC23568@birch.djwong.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20161026011308.GC23568@birch.djwong.org> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: "Darrick J. Wong" Cc: linux-xfs@vger.kernel.org On Tue, Oct 25, 2016 at 06:13:09PM -0700, Darrick J. Wong wrote: > On Wed, Oct 26, 2016 at 11:49:20AM +1100, Dave Chinner wrote: > > On Tue, Oct 25, 2016 at 04:04:39PM -0700, Darrick J. Wong wrote: > > > Do some basic checks of the refcount btree. xfs_repair will have to > > > check that the reference counts match the various bmbt mappings. > > > > > > Signed-off-by: Darrick J. Wong > > ..... > > > @@ -1561,10 +1597,15 @@ check_set_dbmap( > > > agbno, agbno + len - 1, c_agno, c_agbno); > > > return; > > > } > > > - check_dbmap(agno, agbno, len, type1); > > > + check_dbmap(agno, agbno, len, type1, is_reflink(type2)); > > > mayprint = verbose | blist_size; > > > for (i = 0, p = &dbmap[agno][agbno]; i < len; i++, p++) { > > > - *p = (char)type2; > > > + if (*p == DBM_RLDATA && type2 == DBM_DATA) > > > + ; /* do nothing */ > > > + if (*p == DBM_DATA && type2 == DBM_DATA) > > > + *p = (char)DBM_RLDATA; > > > + else > > > + *p = (char)type2; > > > > What this /* do nothing */ case for? > > Handles the case that a data block with multiple owners encounters > another owner. The second case in the block handles the case that > a data block with a single owner encounters a second owner. > > (Assuming you're prodding me to add a comment, I'll go do that.) Ah, no, I'm asking why you added dead code: if (foo) ; if (bar) .... else .... the "if (foo) ;" case is dead code - it doesn't need to exist. Did you mean this: if (foo) ; else if (bar) .... else .... Cheers, Dave. -- Dave Chinner david@fromorbit.com