From: Christoph Hellwig Subject: Re: [RFC] JBD ordered mode rewrite Date: Sat, 8 Mar 2008 07:14:20 -0500 Message-ID: <20080308121420.GA17812@infradead.org> References: <20080306174209.GA14193@duck.suse.cz> <20080307235210.GM1881@webber.adilger.int> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org To: Andreas Dilger Return-path: Received: from bombadil.infradead.org ([18.85.46.34]:34561 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751065AbYCHMOW (ORCPT ); Sat, 8 Mar 2008 07:14:22 -0500 Content-Disposition: inline In-Reply-To: <20080307235210.GM1881@webber.adilger.int> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Fri, Mar 07, 2008 at 04:52:10PM -0700, Andreas Dilger wrote: > I'm looking at what implications this has for delayed allocation in ext4, > because the vast majority of file data will be unmapped in that case > and a journal commit in ordered mode will no longer cause the data to > be flushed to disk. The buffers shouldn't be unmapped. They are accounted for and doing the delalloc conversion is easier than really allocating blocks for truely unmapped blocks. You should probably reuse BH_Delay for that as it has all the right handling in buffer.c in place due to XFS. Also on any filesystem with ->page_mkwrite implemented unmapped buffers should be entirely gone because we now have the proper early reservation / allocation infrastructure in place.