From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755747AbXFYPk7 (ORCPT ); Mon, 25 Jun 2007 11:40:59 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752246AbXFYPkw (ORCPT ); Mon, 25 Jun 2007 11:40:52 -0400 Received: from lixom.net ([66.141.50.11]:56282 "EHLO mail.lixom.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751985AbXFYPkw (ORCPT ); Mon, 25 Jun 2007 11:40:52 -0400 Date: Mon, 25 Jun 2007 10:50:08 -0500 To: Christoph Hellwig , linux-kernel@vger.kernel.org, linux-pcmcia@lists.infradead.org, linuxppc-dev@ozlabs.org, Russell King Subject: Re: [PATCH] pcmcia: CompactFlash driver for PA Semi Electra boards Message-ID: <20070625155008.GA5217@lixom.net> References: <20070625010311.GA31355@lixom.net> <20070625055636.GB1057@infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20070625055636.GB1057@infradead.org> User-Agent: Mutt/1.5.13 (2006-08-11) From: olof@lixom.net (Olof Johansson) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Jun 25, 2007 at 06:56:36AM +0100, Christoph Hellwig wrote: > > +static int bus_notify(struct notifier_block *nb, unsigned long action, > > + void *data) > > +{ > > + struct device *dev = data; > > + > > + printk("bus notify called\n"); > > + > > + /* We are only intereted in device addition */ > > + if (action != BUS_NOTIFY_ADD_DEVICE) > > + return 0; > > + > > + /* We use the direct ops for localbus */ > > + dev->archdata.dma_ops = &dma_direct_ops; > > + > > + return 0; > > +} > > Sorry for not coming back to you after the last posting of the patch, > but I still really this bit of the code. We don't set dma ops from > a driver anywhere else in the tree, so I'd really prefer if you could > handle this in architecture code somewhere. Especially as dma_direct_ops > and the pcmcia_bug_type shouldn't really be something exported to users. > > Also this code is unlogic if not buggy. Just because you have one > electra bridge it doesn't mean all pcmcia is driven by it. In fact I'm > pretty sure there are cardbus/pcmcia bridges that can be plugged into > plain pci slots. Yeah, all good points. Also, I obviously missed to take out the debug printk. I'll post an updated patch separately. Thanks! -Olof