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, 9 Sep 2002 17:51:33 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Mon, 9 Sep 2002 17:51:33 -0400 Received: from ute.rlxtechnologies.com ([65.66.195.252]:45067 "EHLO mail.rocketlogix.com") by vger.kernel.org with ESMTP id ; Mon, 9 Sep 2002 17:51:32 -0400 Subject: Re: oops on 2.4.20-pre5-ac2 From: Dan Eaton To: Alan Cox Cc: linux-kernel@vger.kernel.org In-Reply-To: <1031603294.29715.20.camel@irongate.swansea.linux.org.uk> References: <1031596139.25426.184.camel@dan> <1031603294.29715.20.camel@irongate.swansea.linux.org.uk> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 09 Sep 2002 16:51:05 -0500 Message-Id: <1031608265.25426.245.camel@dan> Mime-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 2002-09-09 at 15:28, Alan Cox wrote: > Thats a very very strange machine 8). Your diagnosis is correct. We are > assuming conventional PC north bridge behaviour. I take it this is your > blade systems. Your diagnosis is correct. This is one of our blade servers. :) > > As far as I can make out from the ALi docs you can't get an ALi north > bridge at anywhere but 0:0.0 in which case making the check > > if(north && north->vendor == ...) > > should do the trick. Our north bridge is a Micron north bridge (see lspci output below). As you can see it is not in the "expected" pci bus location. The fix you listed above did the trick as you expected. The server booted fine, but (there's always a but).....the two drives came up in PIO mode and could not even be forced (hdparm -d1 /dev/hd?) into DMA mode. Looking at dmesg, I got the following message associated with IDE configuration: ALi15x3: simplex device with no drives: DMA disabled The __init ide_get_or_set_dma_base() function in drivers/ide/setup-pci.c is generating this error. I noticed in the source code a "FIXME" comment that seemed to indicate the problem I was having is a potential shortcoming in the current implementation. I added the PCI_DEVICE_ID_AL_M5229 (my IDE controller) as a case to the switch construct and this seemed to fix me completely up. Both drives came up in udma4 mode (as expected). My contribution to the "FIXME" effort would be to suggest that this mod be maded to this function (but I'm no ide expert). Thanks for your help, Dan lscpi output: 00:18.0 Host bridge: Crucial Technology: Unknown device 3320 (rev 03) Subsystem: Micron: Unknown device 1042 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- Status: Cap- 66Mhz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- SERR-