From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933464AbXCONHM (ORCPT ); Thu, 15 Mar 2007 09:07:12 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S933537AbXCONHM (ORCPT ); Thu, 15 Mar 2007 09:07:12 -0400 Received: from moutng.kundenserver.de ([212.227.126.187]:53084 "EHLO moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933459AbXCONHL (ORCPT ); Thu, 15 Mar 2007 09:07:11 -0400 Subject: Re: [PATCH] Remove "decl_subsys_name" macro and single usage of it. From: Kay Sievers To: "Robert P. J. Day" Cc: Linux Kernel Mailing List , Greg KH In-Reply-To: References: <3ae72650703150322h10882044nbcfdddd4187d31ce@mail.gmail.com> Content-Type: text/plain Date: Thu, 15 Mar 2007 14:08:07 +0100 Message-Id: <1173964087.5099.14.camel@pim.off.vrfy.org> Mime-Version: 1.0 X-Mailer: Evolution 2.8.2 Content-Transfer-Encoding: 7bit X-Provags-ID: V01U2FsdGVkX1+c2hxKyGZpqfpzGw4soEFMzJYkkDsxw5cZWiW uWXEqmUyYCNa+KJ5aKyVWVhClQ0K0FjFi+SofvZ+ZYcHYOUHrN MYCWrzzAj5spysB9UQvkQ== Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 2007-03-15 at 08:52 -0400, Robert P. J. Day wrote: > On Thu, 15 Mar 2007, Kay Sievers wrote: > > > On 3/1/07, Robert P. J. Day wrote: > > > > > > Remove the macro "decl_subsys_name" which can be used to declare a > > > sysfs subsystem, along with the single invocation of it in the source > > > tree, since there appears to be little value in creating a subsystem > > > whose subsystem name differs from its structure name. Everyone else > > > just uses "decl_subsys". > > > > Sometimes you want shorter variable names as the created objects in > > sysfs, because the 80 columns limit of the kernel source makes long > > variable names pretty inconvenient. > > And all these decl_* macros, obviously can't create any names that > > contain '-', which is pretty annoying. There is also a user in the > > experimental patches in Greg's tree. Please leave it there for now. > > no problem but that patch *was* based on a suggestion by greg in the > first place. i'm guessing he just forgot. :-) A cleanup of the driver-core source is long overdue, that's why we like everything that goes away here, that isn't really needed. :) In the end, we should probably just get rid of the whole "struct subsystem". We can move the lock to the kset and ditch the whole weird idea of a "subsystem", which is nothing but a collection of ksets, which the object model can handle without "struct subsystem" just fine. Thanks, Kay