From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932688AbXAJCyN (ORCPT ); Tue, 9 Jan 2007 21:54:13 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932690AbXAJCyN (ORCPT ); Tue, 9 Jan 2007 21:54:13 -0500 Received: from shawidc-mo1.cg.shawcable.net ([24.71.223.10]:16037 "EHLO pd3mo2so.prod.shaw.ca" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932688AbXAJCyM (ORCPT ); Tue, 9 Jan 2007 21:54:12 -0500 Date: Tue, 09 Jan 2007 20:53:57 -0600 From: Robert Hancock Subject: Re: PCI BIOS Bug messages In-reply-to: To: Vasudevan S Cc: linux-kernel Message-id: <45A45545.2040002@shaw.ca> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7bit References: User-Agent: Thunderbird 1.5.0.9 (Windows/20061207) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Vasudevan S wrote: > I run Fedora Core 6 on the 'compaq nc6320' laptop. I am using the > '2.6.19.1' kernel. > > While booting the kernel, I noticed the following error message: > > PCI: BIOS Bug: MCFG area at f8000000 is not E820-reserved > PCI: Not using MMCONFIG. > > After some search, I commented out the 'e820_all_mapped()' check in > the 'pci_mmcfg_init()' function. I no longer see this message and MMCONFIG > method seems to be used now. > > Is this the right thing to do? Maybe on that box you can get away with it, but I believe the purpose of the check is to weed out BIOSes that have totally broken MCFG tables. I think there is some work going on to be able to check and/or bash the chipset registers manually on various chipsets to reduce our reliance on the BIOS getting this right.. -- Robert Hancock Saskatoon, SK, Canada To email, remove "nospam" from hancockr@nospamshaw.ca Home Page: http://www.roberthancock.com/