From mboxrd@z Thu Jan 1 00:00:00 1970 From: Robert Hancock Subject: Re: 2.6.28.7 boot problems Date: Sun, 01 Mar 2009 13:22:38 -0600 Message-ID: <49AAE07E.5000806@gmail.com> References: <40f01080902251346v5f73ed43nacc07bf65c98a727@mail.gmail.com> <49A8AB75.9020006@gmail.com> <40f01080902280827taf007cq613609fb42b9b24@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mail-gx0-f174.google.com ([209.85.217.174]:55119 "EHLO mail-gx0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754505AbZCATWm (ORCPT ); Sun, 1 Mar 2009 14:22:42 -0500 Received: by gxk22 with SMTP id 22so4115912gxk.13 for ; Sun, 01 Mar 2009 11:22:40 -0800 (PST) In-Reply-To: <40f01080902280827taf007cq613609fb42b9b24@mail.gmail.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Andrew Schepler Cc: linux-ide@vger.kernel.org Andrew Schepler wrote: > Robert Hancock wrote: >> Sounds like you don't have the chipset-specific IDE driver enabled for your >> hardware (I don't see any of them enabled in this list). I think that >> IDE_GENERIC used to drive the controller in this case, but not anymore, as >> it can only do so in a crappy, non-DMA way. > > Thanks Robert. > > After more research, I determined my motherboard is a Gigabyte > Technology GA-K8U-939. Its docs say it has a ULi M1689 chipset, which > supports Dual Serial ATA. So I tried enabling ATA, ATA_SFF, and > SATA_ULI. The kernel printed out a couple new lines about sata0 and > sata1, but didn't find any hd[a-d] or otherwise get any farther. > > Then I tried just turning on as many chipsets as I could (skipping a > few whose help pages had dangerous-sounding warnings). Using this > kernel, the alim15x3 driver (CONFIG_BLK_DEV_ALI15X3) did manage to > find hda and finish booting. But I got several messages from the > kernel looking like: > > hda: dma_intr: status=0x51 { DriveReady SeekComplete Error } > hda: dma_intr: error=0x84 { DriveStatusError BadCRC } > ide: failed opcode was: unknown It looks like you're getting some CRC errors - could be you have a bad IDE cable, or some other hardware problem, or (I would guess more likely) the driver is picking the wrong UDMA mode or setting up the controller improperly in some other way. As I mention below, you'll likely want to try the libata driver instead. > > Next I tried out Debian's > linux-image-2.6.28-1-amd64_2.6.28-1_i386.deb. It also loaded hda > using the alim15x3 driver, and gave the same messages as above. > > Do those messages suggest the wrong driver is being used? That > there's a (not-yet critical) hardware problem on my hard disk, which > older kernels don't notice? > > Now I'm going to go try some e2fsck -cc /dev/hda3 and/or badblocks -n > /dev/hda[12] to see if any disk problems turn up that way. > > Attached is my dmesg from Debian's kernel. If you're using a modern distribution, I would say you should try disabling the old IDE drivers under CONFIG_IDE entirely, and try the libata driver instead. This will mean all your drives will show up as sdX instead of hdX, which may mean you need to make adjustments to your boot setup (mounting by label or by UUID, possibly) so that you can still mount the root file system.