From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1765336AbXLMHX5 (ORCPT ); Thu, 13 Dec 2007 02:23:57 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1764218AbXLMHN3 (ORCPT ); Thu, 13 Dec 2007 02:13:29 -0500 Received: from pentafluge.infradead.org ([213.146.154.40]:46627 "EHLO pentafluge.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1762172AbXLMHN2 (ORCPT ); Thu, 13 Dec 2007 02:13:28 -0500 Date: Wed, 12 Dec 2007 23:10:22 -0800 From: Greg KH To: Stephen Rothwell Cc: michael@ellerman.id.au, linuxppc-dev@ozlabs.org, "Kyle A. Lucke" , paulus@samba.org, linux-kernel@vger.kernel.org, David Gibson Subject: Re: [PATCH] Introduce driver_create/remove_dir Message-ID: <20071213071022.GB13764@kroah.com> References: <20071205093054.GA23229@kroah.com> <1196853031.6759.7.camel@concordia> <20071205214103.GA7074@kroah.com> <1196912898.14754.13.camel@concordia> <20071212105633.725496cb.sfr@canb.auug.org.au> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20071212105633.725496cb.sfr@canb.auug.org.au> User-Agent: Mutt/1.5.16 (2007-06-09) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Dec 12, 2007 at 10:56:33AM +1100, Stephen Rothwell wrote: > > Signed-off-by: Stephen Rothwell > --- > drivers/base/driver.c | 24 ++++++++++++++++++++++++ > drivers/net/iseries_veth.c | 15 +++++++-------- > include/linux/device.h | 3 +++ > 3 files changed, 34 insertions(+), 8 deletions(-) > > Greg, does this look like a reasonable solution to iseries_veth accessing > the "private" kobject in struct device_driver? This version is against > maimline, but the stuff you have in mm would just need to update > driver_create_dir ... Hm, we just want to be able to get to the kobject somehow here. Not create a new api that doesn't match up. I'll think about it and figure something that matches the other portions of the api. > Also something along the lines of device_add_dir() might be good if you > want to hide the kobject in struct device as well. As devices _should_ always be dynamic, hopefully I'll not have to do that. But knowing some of the platform devices, I'm afraid I'll have to do that split :) thanks, greg k-h