DMA bug with ICH7-M

Bug #897777 reported by Ben Whitten
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Fix Released
High
Unassigned

Bug Description

I am using a Unicorn computers ENDAT-2794A motherboard http://www.unicorn-computer.com.tw/products/itx/2794a.html
With the stock latest release of ubuntu 11.10.
The boot process is very slow and jams for a period of time, when it is booted I cannot install to the harddrive as it reports an IO error.
Looking at dmesg | grep ata I see that there are lots of 'failed command READ DMA EXT' errors.
After some googling I have found that there was a previous bug with the ata_piix driver (the driver in use) with the ICH7, where the UDMA rate was set to 133 but the max supported by hardware is 100. However, in the dmesg I see that it is backing right down to 33, I presume that this point is where the system becomes responsive again.

Best Regards,
Ben

ProblemType: Bug
DistroRelease: Ubuntu 11.10
Package: linux-image-3.0.0-12-generic 3.0.0-12.20
ProcVersionSignature: Ubuntu 3.0.0-12.20-generic 3.0.4
Uname: Linux 3.0.0-12-generic i686
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.24.
ApportVersion: 1.23-0ubuntu3
Architecture: i386
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: ubuntu 4193 F.... pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'Intel'/'HDA Intel at 0xfdff8000 irq 44'
   Mixer name : 'Realtek ALC888'
   Components : 'HDA:10ec0888,10ec0888,00100001'
   Controls : 40
   Simple ctrls : 22
CasperVersion: 1.287
Date: Tue Nov 29 11:49:55 2011
IwConfig:
 lo no wireless extensions.

 eth0 no wireless extensions.

 eth1 no wireless extensions.
LiveMediaBuild: Ubuntu 11.10 "Oneiric Ocelot" - Release i386 (20111012)
ProcEnviron:
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcKernelCmdLine: noprompt cdrom-detect/try-usb=true persistent file=/cdrom/preseed/hostname.seed boot=casper initrd=/casper/initrd.lz quiet splash --
RelatedPackageVersions:
 linux-restricted-modules-3.0.0-12-generic N/A
 linux-backports-modules-3.0.0-12-generic N/A
 linux-firmware 1.60
RfKill:

SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 06/10/2009
dmi.bios.vendor: Phoenix Technologies, LTD
dmi.bios.version: 6.00 PG
dmi.board.name: 945GSE
dmi.chassis.type: 3
dmi.modalias: dmi:bvnPhoenixTechnologies,LTD:bvr6.00PG:bd06/10/2009:svn:pn:pvr:rvn:rn945GSE:rvr:cvn:ct3:cvr:

Revision history for this message
Ben Whitten (benwhitten) wrote :
Brad Figg (brad-figg)
Changed in linux (Ubuntu):
status: New → Confirmed
Revision history for this message
Brad Figg (brad-figg) wrote :

@Ben,

Is this the first time that you've tried to install Ubuntu onto this system? If it was running a previous version, can you tell us which version was successfully running?

Brad Figg (brad-figg)
Changed in linux (Ubuntu):
importance: Undecided → High
Revision history for this message
Ben Whitten (benwhitten) wrote :

@Brad
This is the first time I have tried personally, however this is an industrial grade board and the manufacturers state linux runs on it. I have contacted them to confirm what version of the kernel they can run on it without issue to try and narrow it down.

Revision history for this message
Brad Figg (brad-figg) wrote :

@Ben,

If you want to try to narrow this down, you could download a daily ISO of the current development series [1] and try that. You could also pull down the Natty release[2] and try that as well. I'd suggest the current development series first if possible.

[1] http://cdimage.ubuntu.com/cdimage/daily-live/current/
[2] http://cdimage.ubuntu.com/cdimage/releases/11.04/release/

Revision history for this message
Ben Whitten (benwhitten) wrote :

@Brad

I have tried the snapshot dated yesterday (29th November) and it exhibits the same problem.

Unicorn computers have confirmed that it used to work with older kernels it seems, I have not had a chance to verify this. Here is their list of examples;
NAME KERNEL
Debian 4.0 2.6.18-6-686
FreeBSD 6.4
FreeBSD 7.1
Fedora Core 8 2.6.23.1-42.fc8
Fedora Core 9 2.6.25-14.fc9.i686
OpenSUSE 10.3 2.6.22.5-31-default
OpenSUSE 11.0
Arch Linux 0.72
Mandrake 10.0
Ubuntu 8.04 2.6.24-16-genernic

I shall now start going back and seeing where the break appears.
Also attached is a picture of the default BIOS options for the IDE controller on this board, in-case there is an option in here.

Revision history for this message
Brad Figg (brad-figg) wrote :

@ben,

Those are some very old kernels. Maybe the safest one to start with would be the Lucid (10.04) release.

Revision history for this message
Ben Whitten (benwhitten) wrote :

@Brad

That they are, I have run up a USB image of 10.04 LTS and it cant install to the harddrive.
However, it is not the complaining about the DMA in dmesg but it does stay on UDMA/133 which is an old bug and out of spec for ICH7, fixed in ata_piix since then.
I have attached dmesg from 10.04 for completeness.

Revision history for this message
Ben Whitten (benwhitten) wrote :

I'm not sure where to look to help debugging this issue, shall I take it to the linux-ide mailing list. See if they have any suggestions?

Revision history for this message
Brad Figg (brad-figg) wrote :

@ben

Yes, please do.

Revision history for this message
Ben Whitten (benwhitten) wrote :

Attached is the install log from Ubuntu 8.04.4 live USB and shows the same problem. I'm not sure how Unicorn verified that it works, they are also reporting 11.10 works.
To test if it was a individual hardware fault I have used a different board for this test but it acts exactly the same.

Revision history for this message
Ben Whitten (benwhitten) wrote :

This bug can be closed as the issue was fixed with a BIOS update from Unicorn.

Thanks for your help.

description: updated
Changed in linux (Ubuntu):
status: Confirmed → Fix Released
Revision history for this message
Benoit Calvez (benoit-calvez) wrote :

I've got an answer from Unicorn.

If you've got the issue,you have to switch the mode into “Enhanced Mode” of On-Chip Serial ATA to solve the problem.
See the attached bios screenshot.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.