From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267263AbUHSS4a (ORCPT ); Thu, 19 Aug 2004 14:56:30 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267258AbUHSS43 (ORCPT ); Thu, 19 Aug 2004 14:56:29 -0400 Received: from the-village.bc.nu ([81.2.110.252]:1925 "EHLO localhost.localdomain") by vger.kernel.org with ESMTP id S267269AbUHSS4X (ORCPT ); Thu, 19 Aug 2004 14:56:23 -0400 Subject: Re: 2.6.8.1-mm2 (nvidia breakage) From: Alan Cox To: Bjorn Helgaas Cc: Michael Geithe , Linux Kernel Mailing List In-Reply-To: <200408191051.11891.bjorn.helgaas@hp.com> References: <20040819092654.27bb9adf.akpm@osdl.org> <200408191051.11891.bjorn.helgaas@hp.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1092938035.28370.13.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.6 (1.4.6-2) Date: Thu, 19 Aug 2004 18:53:56 +0100 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Iau, 2004-08-19 at 17:51, Bjorn Helgaas wrote: > I assume this problem is with the Nvidia binary-only driver? My guess > is that the driver doesn't call pci_enable_device() before using > pci_dev->irq. I don't have source for that driver, so I can't verify > this. An obvious test would be for someone with an Nvidia to write a little module that does nothing but pci_enable_device it. Load and unload that then see what happens