From: Andreas Dilger Subject: Re: [PATCH] ext4: Enable delalloc and mballoc by default. Date: Wed, 24 Oct 2007 15:06:39 -0600 Message-ID: <20071024210639.GC26078@webber.adilger.int> References: <1193244441-10149-1-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1193244441-10149-2-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1193244441-10149-3-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1193244441-10149-4-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1193244441-10149-5-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1193244441-10149-6-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <471F7F41.6080800@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "Aneesh Kumar K.V" , linux-ext4@vger.kernel.org To: Eric Sandeen Return-path: Received: from mail.clusterfs.com ([74.0.229.162]:60899 "EHLO mail.clusterfs.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758207AbXJXVGl (ORCPT ); Wed, 24 Oct 2007 17:06:41 -0400 Content-Disposition: inline In-Reply-To: <471F7F41.6080800@redhat.com> Sender: linux-ext4-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Oct 24, 2007 12:22 -0500, Eric Sandeen wrote: > Aneesh Kumar K.V wrote: > > @@ -1279,6 +1280,9 @@ clear_qf_name: > > case Opt_delalloc: > > set_opt (sbi->s_mount_opt, DELALLOC); > > break; > > If delalloc, mballoc, extents are the new defaults, is there a reason to > keep them as options? When would you need to specify -o extents, now, > for example? (though my brain is fuzzy today, maybe I'm missing > something) If this were not a filesystem ending in "dev" I could see > keeping it for compatibility with existing fstabs.... It is useful to be able to mount w/o extents/delalloc/mballoc for perf testing and functional testing of the block-mapped file path in ext4. Also, some users might want the ability to use features of ext4 w/o the incompatibility of extents. > > set_opt(sbi->s_mount_opt, EXTENTS); > > + set_opt(sbi->s_mount_opt, DELALLOC); > > + set_opt(sbi->s_mount_opt, MBALLOC); I think the other thing to do is enable the INCOMPAT_EXTENTS flag in mkfs.ext4 by default, so that extents is enabled/disabled in the same manner as other ext* features. We can remove the above once we have an e2fsprogs that specifically sets all of the ext4 features (large inodes, etc) for ext4 filesystems. Cheers, Andreas -- Andreas Dilger Sr. Software Engineer, Lustre Group Sun Microsystems of Canada, Inc.