From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: with ECARTIS (v1.0.0; list xfs); Fri, 13 Apr 2007 01:04:32 -0700 (PDT) Received: from pentafluge.infradead.org (pentafluge.infradead.org [213.146.154.40]) by oss.sgi.com (8.12.10/8.12.10/SuSE Linux 0.7) with ESMTP id l3D84RfB010574 for ; Fri, 13 Apr 2007 01:04:29 -0700 Date: Fri, 13 Apr 2007 08:26:47 +0100 From: Christoph Hellwig Subject: Re: [PATCH] xfs_repair - move realtime extent processing to a separate function Message-ID: <20070413072647.GC20326@infradead.org> References: <200704130415.OAA16550@larry.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200704130415.OAA16550@larry.melbourne.sgi.com> Sender: xfs-bounce@oss.sgi.com Errors-to: xfs-bounce@oss.sgi.com List-Id: xfs To: Barry Naujok Cc: xfs@oss.sgi.com, 'xfs-dev' On Fri, Apr 13, 2007 at 02:22:10PM +1000, Barry Naujok wrote: > While changing the process_bmbt_reclist_int() function, I observed a > realtime check inside the block map get/set state loop which is quite > CPU intensive. Upon further investigation, this loop is not used at > all for realtime extents and that the two types of extents are pretty > much processed exclusively. > > So, I simplified the functionality by moving the realtime extent > processing into it's own function and fixing a bug at the same time > when it comes to realtime inodes with attributes (it was comparing > attr extents to the realtime volume bmap instead of the normal bmap). Nice cleanup, looks good.