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, 27 May 2002 04:11:45 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Mon, 27 May 2002 04:11:44 -0400 Received: from mail0.epfl.ch ([128.178.50.57]:23827 "HELO mail0.epfl.ch") by vger.kernel.org with SMTP id ; Mon, 27 May 2002 04:11:43 -0400 Message-ID: <3CF1EA3F.4070608@epfl.ch> Date: Mon, 27 May 2002 10:11:43 +0200 From: Nicolas Aspert Organization: LTS-DE-EPFL User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0rc3) Gecko/20020523 X-Accept-Language: en-us, ja MIME-Version: 1.0 To: Alan Cox CC: Alessandro Morelli , linux-kernel@vger.kernel.org Subject: Re: PROBLEM: memory corruption with i815 chipset variant In-Reply-To: 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 Alan Cox wrote: > > That means its actually using the same GART code as the 440BX and friends > if I remember rightly (the i815 special stuff is for on board video) Alessandro reported the problem to me also. I went through the i815 specs and found 2 'strange' things (maybe they are not but...) 1) No 'ERRSTS' register (well... a bus that does no error should be a feature ;-) 2) The ATTBASE register to which the *_configure functions write is different from other Intel chipsets. In the i815, the ATT base adress should be written between bits 12 and 28, whereas in all other Intel chipsets, it should be written between bits 12 and 31 (don't ask me why Intel feels like changing the adresses/specs for registers at each new chipsets....) . Alan, do you think this could cause all those troubles ? > >>Without agpgart module, kernel seems stable. A naive (totally naive, >>I admit it) interpretation suggests a problem in setting the AGP aperture. > > > Does the ram survive memtest86 overnight with no errors logged if you boot > memtest86 and just leave it ? From what Alessandro reported, it seems clear that the 'insmod agpgart' triggers the mayhem, including memtest failures. Best regards Nicolas. -- Nicolas Aspert Signal Processing Institute (ITS) Swiss Federal Institute of Technology (EPFL)