From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932419AbVKGDd0 (ORCPT ); Sun, 6 Nov 2005 22:33:26 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932428AbVKGDd0 (ORCPT ); Sun, 6 Nov 2005 22:33:26 -0500 Received: from verein.lst.de ([213.95.11.210]:24286 "EHLO mail.lst.de") by vger.kernel.org with ESMTP id S932419AbVKGDdZ (ORCPT ); Sun, 6 Nov 2005 22:33:25 -0500 Date: Mon, 7 Nov 2005 04:33:22 +0100 From: Christoph Hellwig To: Arnd Bergmann Cc: linux-kernel@vger.kernel.org, Christoph Hellwig , axboe@suse.de Subject: Re: [PATCH 13/25] loop: move ioctl32 code to loop.c Message-ID: <20051107033322.GA15864@lst.de> References: <20051105162650.620266000@b551138y.boeblingen.de.ibm.com> <20051105162715.758021000@b551138y.boeblingen.de.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20051105162715.758021000@b551138y.boeblingen.de.ibm.com> User-Agent: Mutt/1.3.28i X-Spam-Score: -4.901 () BAYES_00 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Nov 05, 2005 at 05:27:03PM +0100, Arnd Bergmann wrote: > The loop device driver is the only user of its ioctl > commands, so we can move the conversion handlers > for 32 bit emulation into the driver itself. > > This patch just moves over the function, it would > probably be a good idea to get rid of get_fs/set_fs > calls here by integrating better with the driver. Yes. just moving over is pointless. for things like loop where there's just one driver implementing the api moving it there does make sense, but please bring the compat layer up to standards first.