From: Dave Chinner Subject: Re: [Ext4 punch hole 1/5] Ext4 Punch Hole Support: Convert Blocks to Uninit Exts Date: Thu, 3 Mar 2011 07:23:13 +1100 Message-ID: <20110302202313.GA15097@dastard> References: <4D6C6318.2010105@linux.vnet.ibm.com> <02A57041-5FC1-419D-89D2-47D541616DD4@dilger.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Allison Henderson , linux-ext4@vger.kernel.org To: Andreas Dilger Return-path: Received: from ipmail07.adl2.internode.on.net ([150.101.137.131]:55344 "EHLO ipmail07.adl2.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754997Ab1CBUXQ (ORCPT ); Wed, 2 Mar 2011 15:23:16 -0500 Content-Disposition: inline In-Reply-To: <02A57041-5FC1-419D-89D2-47D541616DD4@dilger.ca> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Mon, Feb 28, 2011 at 11:42:48PM -0700, Andreas Dilger wrote: > On 2011-02-28, at 8:08 PM, Allison Henderson wrote: > > This first patch adds a function to convert a range of blocks > > to an uninitialized extent. This function will > > be used to first convert the blocks to extents before > > punching them out. > > This was proposed as a separate function for FALLOCATE by Dave > Chinner (based on XFS_IOC_ZERO_RANGE), so this is useful as a > standalone function. XFS_IOC_ZERO_RANGE converts a range to unwritten extents, not uninitialised extents. An uninitialised extent is one that is allocated but had not data written to it (i.e. contains stale data), while an unwritten/preallocated extent is guaranteed to contain zeros. This may be just a terminology issue, but we should try to use the same jargon across all filesystems... Cheers, Dave. -- Dave Chinner david@fromorbit.com