From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751432AbWH1IPJ (ORCPT ); Mon, 28 Aug 2006 04:15:09 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751428AbWH1IPJ (ORCPT ); Mon, 28 Aug 2006 04:15:09 -0400 Received: from pentafluge.infradead.org ([213.146.154.40]:17116 "EHLO pentafluge.infradead.org") by vger.kernel.org with ESMTP id S1751421AbWH1IPI (ORCPT ); Mon, 28 Aug 2006 04:15:08 -0400 Subject: Re: [2.6.17.11] strange pcie errors/warnings on Abit KN9-SLI mainboard From: Arjan van de Ven To: Jan De Luyck Cc: linux-kernel@vger.kernel.org In-Reply-To: <200608280755.56015.ml_linuxkernel_20060528@kcore.org> References: <200608280755.56015.ml_linuxkernel_20060528@kcore.org> Content-Type: text/plain Organization: Intel International BV Date: Mon, 28 Aug 2006 10:14:57 +0200 Message-Id: <1156752897.3034.163.camel@laptopd505.fenrus.org> Mime-Version: 1.0 X-Mailer: Evolution 2.2.3 (2.2.3-2.fc4) Content-Transfer-Encoding: 7bit X-SRS-Rewrite: SMTP reverse-path rewritten from by pentafluge.infradead.org See http://www.infradead.org/rpr.html Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 2006-08-28 at 07:55 +0200, Jan De Luyck wrote: > Hello, > > (running 2.6.17.11 vanilla on Debian SID) > > I recently acquired a new pc, with an ABIT KN9-SLI mainboard, using an AMD64x2 > AM2 processor. > > System boots fine, but I have some messages/errors in the dmesg that I'm worried > about. Googling around for them didn't really show up much. > > First: > > Aug 27 15:35:04 whocares kernel: PCI-DMA: Disabling IOMMU. > > IOMMU is (as far as I can see) enabled: > whocares:/var/log# cat /usr/src/build/linux-2.6/.config | grep IOMMU > CONFIG_GART_IOMMU=y > > I can't really determine if this is normal. According to the code it seems that > this is disabled by default when you don't have AGP? (I'm not a kernel-coder, so > I may be very wrong on this) Hi, unless you have >= 4Gb of ram you don't need an IOMMU (in fact using it would only cause extra overhead)... so the kernel will not use it in that case. Greetings, Arjan van de Ven