From mboxrd@z Thu Jan 1 00:00:00 1970 From: Theodore Tso Subject: Re: [PATCH 0/5] remove superflous ->write_super instances Date: Mon, 27 Apr 2009 12:54:20 -0400 Message-ID: <20090427165420.GC6781@mit.edu> References: <20090427134640.128837000@bombadil.infradead.org> <20090427143221.GA32264@mit.edu> <20090427143600.GA18405@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , linux-fsdevel@vger.kernel.org To: Christoph Hellwig Return-path: Received: from thunk.org ([69.25.196.29]:44606 "EHLO thunker.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753383AbZD0QyZ (ORCPT ); Mon, 27 Apr 2009 12:54:25 -0400 Content-Disposition: inline In-Reply-To: <20090427143600.GA18405@infradead.org> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Mon, Apr 27, 2009 at 10:36:00AM -0400, Christoph Hellwig wrote: > On Mon, Apr 27, 2009 at 10:32:21AM -0400, Theodore Tso wrote: > > On Mon, Apr 27, 2009 at 09:46:40AM -0400, Christoph Hellwig wrote: > > > Remove empty write_super methods and unessecary s_dirt maintainance. > > > > I noticed you didn't supply a patch for ext4, so I'm guessing you > > figured it out, but there is an (outdated) comment saying we should do > > this for ext4; however, given that we now support non-journalled > > operation, we do need to keep write_super() for when we aren't using > > the journal to write the superblock. I'll make a note to myself that > > we need to update the comment in ext4/super.c. > > Exactly, ext4 needs it for non-journaled mode. It might make sense > to just give the two modes different super_operations so that the normal > journaled mode doesn't need a write_super at all. > True. We'll bloat the kernel text space a little by having a separate super_operations but it will allow us to clean up the code; it's probably a worthwhile tradeoff, since we it's not just ext4_write_super(), but also ext4_freeze() and ext4_unfreeze(). - Ted