From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756887Ab0KLKab (ORCPT ); Fri, 12 Nov 2010 05:30:31 -0500 Received: from claranet-outbound-smtp05.uk.clara.net ([195.8.89.38]:32974 "EHLO claranet-outbound-smtp05.uk.clara.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756835Ab0KLKa3 (ORCPT ); Fri, 12 Nov 2010 05:30:29 -0500 Message-ID: <4CDD1737.6010104@sciolus.org> Date: Fri, 12 Nov 2010 10:30:15 +0000 From: "R.M. Thomas" User-Agent: Mozilla-Thunderbird 2.0.0.24 (X11/20100329) MIME-Version: 1.0 To: Randy Dunlap CC: Stephen Rothwell , linux-next@vger.kernel.org, LKML , gregkh@suse.de Subject: Re: [PATCH -next] staging/easycap: make module params private/static, fix build References: <20101111125655.15fa7188.sfr@canb.auug.org.au> <20101111104422.28594510.randy.dunlap@oracle.com> <4CDC67FF.3010006@sciolus.org> <4CDC6BA4.3060307@oracle.com> In-Reply-To: <4CDC6BA4.3060307@oracle.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Randy Dunlap wrote: > On 11/11/10 14:02, R.M. Thomas wrote: >> Randy Dunlap wrote: >>> The easycap driver has module parameters (bars, gain, & debug) >>> with global scope that intrude on the kernel namespace and cause >>> build problems. Change the names of them to be driver-specific >>> and make 2 of them static. >> I did do test builds of the driver in-tree prior to submitting the >> recent set of patches, but did not see any warnings so did not spot >> this mistake. There's presumably something wrong with the way I'm >> building the kernel. > > This build problem shows up when this driver is built into the kernel > image, not built as a loadable module. Did you only build as a loadable > module, maybe? and what CPU architecture did you build for? That could > also matter. Yes, that's it. I do development of the driver on a machine running Debian stable with 2.6.26. I detect compilation and house-style errors by copying the source into a clone of the linux-next tree and running make and checkpatch.pl, but obviously I don't install. To test the run-time behaviour of the driver I copy the driver source to a separate (sacrificial) machine which is actually running linux-next and build the driver there out-of-tree as a loadable module, complete with the installation step. One machine is x86-64, the other is x86. I'll change my way of working. Mike