From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Mon, 18 Nov 2002 16:58:11 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Mon, 18 Nov 2002 16:57:24 -0500 Received: from parcelfarce.linux.theplanet.co.uk ([195.92.249.252]:12813 "EHLO www.linux.org.uk") by vger.kernel.org with ESMTP id ; Mon, 18 Nov 2002 16:55:34 -0500 Message-ID: <3DD96358.6050701@pobox.com> Date: Mon, 18 Nov 2002 17:02:00 -0500 From: Jeff Garzik User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2b) Gecko/20021018 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Peter Chubb CC: linux-kernel@vger.kernel.org, Alan Cox Subject: Re: [PATCH] Tulip driver fails to link (2.5.48) References: <15833.24514.280850.81268@wombat.chubb.wattle.id.au> In-Reply-To: <15833.24514.280850.81268@wombat.chubb.wattle.id.au> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Peter Chubb wrote: > Hi, > In the tulip driver, the .remove entry of struct pci_driveris > declared __devexit_p but the function is declared __exit. This > causes a relocation error when building as a built-in (not a module), > when CONFIG_HOTPLUG is defined. Yes... alan snuck that __devexit_p usage in there ;-) We need to create __exit_p as the #warning indicates, because this driver does not support CONFIG_HOTPLUG intentionally. For now, I think it is better to add #ifdef MODULE to the driver in struct pci_driver definition, because using "__devexit" as in your patch is incorrect. So if you wanted to submit the patch changing __devexit_p to #ifdef MODULE, that could probably be applied... Jeff