From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266291AbUGOTZ5 (ORCPT ); Thu, 15 Jul 2004 15:25:57 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266292AbUGOTZ5 (ORCPT ); Thu, 15 Jul 2004 15:25:57 -0400 Received: from out002pub.verizon.net ([206.46.170.141]:22694 "EHLO out002.verizon.net") by vger.kernel.org with ESMTP id S266291AbUGOTZz (ORCPT ); Thu, 15 Jul 2004 15:25:55 -0400 Message-ID: <40F693EE.5D65D02F@verizon.net> Date: Thu, 15 Jul 2004 20:25:50 +0100 From: "Max T. Woodbury" X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: PCI configuration changes getting lost. Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Authentication-Info: Submitted using SMTP AUTH at out002.verizon.net from [4.15.37.40] at Thu, 15 Jul 2004 14:25:54 -0500 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I'm having trouble with an old Thinkpad 760 ED. The BIOS puts junk (0xFF) in the INTERRUPT_LINE register of the PCMCIA configuration registers. I can put the correct value (0x05) in with 'setpci', but the information is not propagated to the /proc space. The information is not preserved after an 'init 6'. (There is at least one other BIOS induced problem with this machine and PCI configuration registers but it is a bit less serious. I do have the latest BIOS.) What would be involved in updating the /proc information when 'setpci' is used to change the configuration? Does the other pci 'fixer' 'pcitweak' do a better job? [Don't remove PCI from the subject line. I'm using it to keep track of interesting things in this otherwise very noisy mailing list.] max@mtew.isa-geek.net