Dell C400 (i830) freezes in Hardy - Needs PipeA Quirk

Bug #267570 reported by Nick Rhodes on 2008-09-07
Affects Status Importance Assigned to Milestone
xserver-xorg-video-intel (Ubuntu)
Bryce Harrington
Declined for Intrepid by Bryce Harrington

Bug Description

Binary package hint: xserver-xorg-video-intel

Like various other bug reports and on the forums, I am suffering random freezes (machine 100% unresponsive, caps lock, tty, ssh etc) which occur between 5 minutes and 1 hour usually, I have not yet reached 2 hours uptime with Hardy vanilla install.

I originally suspected a Kernel issue, but after trying other Kernels (intrepid ones and hardy/realtime) I moved onto thinking its an issue with X.

So when I disabled to intel driver and using the VESA driver, I managed an uptime of 5 hours (and counting).

I can try the Xserver-xorg package from Gutsy to see if I get freezes with that (I know that I suffer from freezing when I shut my lid with Gutsy on my C400).

This error occurs on Hardy liveCD, Hardy installed with no updates (xserver-xorg-video 2.2.1-1ubuntu12) and with all updates installed (xserver-xorg-video 2.2.1-1ubuntu13.6).

Machine is a Dell C400 Laptop, with the 830 gfx chip.

Cheers, Nick

description: updated

Please include your /var/log/Xorg.0.log.

Offhand, this sounds like the pipe-A problem. Try setting this option in your xorg.conf:

  Section "Device"
      Option "ForceEnablePipeA" "true"

If that solves the issue let me know and I can put in a quirk for it.

Changed in xserver-xorg-video-intel:
status: New → Incomplete
Nick Rhodes (ngrhodes) wrote :

I have just added the option above, will keep you post if it makes a difference.

xorg log should be attached.

Nick Rhodes (ngrhodes) wrote :

My xorg.conf for reference.

Nick Rhodes (ngrhodes) wrote :

Progress report.

Machine is still locking - but FAR less frequently than before, I would say that the pipe-A has greatly reduced the frequency.

It has frozen 3 times in the past 2 days.

First time was lid closing and reopening the lid caused a freeze approx 1 or 2 seconds after the screen came back on.

Second time was about 10 minutes after the first time.

Third time I had been using the machine for approx 1 hour (from cold boot) and it froze.

I left the laptop on overnight and it has not frozen, but I noted the screen had not gone blank and the machine was very hot (like no power saving had occurred) - I don't know if this is a xorg issue and/or power management ?

Nick Rhodes (ngrhodes) wrote :

I have been the latest kernel out of hardy-proposed and the machine has been running for 3 hours 45 minutes, most of which has been idle - and no problems, scree blanked and machine cool to touch.
Opening and closing the lid works fine.

As a final test, tomorrow, I shall leave the current kernel on and remove the pipe-A option and see if the freezes reoccur as a confirmation.

Nick Rhodes (ngrhodes) wrote :

After 11 hours of uptime with the proposed Kernel my laptop froze again :(
Now running the 24-19 kernel again.

Nick Rhodes (ngrhodes) wrote :

Had 3 freezes today within an hour, still running the pipeA option.

So I do not think this has fixed the issue.

Nick Rhodes (ngrhodes) wrote :

I dont know if this information will help but:

I have temporarily installed Debian Testing - Lenny.

Fully updated (as of yesterday) (xserver-xorg-video-intel (2:2.3.2-2+lenny4), xserver-xorg-core (2:1.4.2-5) , linux-image-2.6.26-1-686 (2.6.26-4) ) and it froze quickly (within the hour but I was not at the machine).

Adding the pipe-A option and vga=733 to grub and the machine has been running 24hours 100% stable (including closing lid a few times), though I am suffering occasional screen artifacts which go away after a screen refresh (eg changing to a tty and back)

These are the only changes made to Debian and Ubuntu. Both running 100% default desktop/laptop configurations.

Bryce Harrington (bryce) wrote :

Can you test if both the pipe-A option and the vga=733 setting are required for the stability, or if one or the other does it?

Elod VALKAI (elod) wrote :

It works without Pipe-A option after updating to xorg intel driver 2:2.4.1-1ubuntu6.

The screen corruption issue is not fixed, though, and it's not minor. Characters get messed up, simply missing, or occasionally blinking. Switching to text and back fixes it, for a couple of seconds.

Elod VALKAI (elod) wrote :

I forgot I'm using Intrepid with 2.6.27-3.4.

Elod VALKAI (elod) wrote :

Some option from the folowing seems to fix screen corruption:

        Option "AccelMethod" "exa"
        Option "MigrationHeuristic" "greedy"
        Option "ExaNoComposite" "true"

Nick Rhodes (ngrhodes) wrote :


Many thanks, those settings worked for my Debian install.


Removing the pipe-A option caused my machine to lockup when I closed the screen.
I have yet to test for general stability or with the pipe-A option back in, but the vga=773 option removed.

Nick Rhodes (ngrhodes) wrote :

With only vga=773 I got 2 freezes using the machine, about 1 hour uptime each time.
Now trying pipeA only.

Nick Rhodes (ngrhodes) wrote :

I now have 10 hours of uptime, lid closing is working reliably.

Currently running pipe-A and vga=773

The fix - disable compiz.

I shall report back tomorrow if machine is still running.

Then I can look at disabling pipe-A and vga=773 and see if machine stays stable.

Nick Rhodes (ngrhodes) wrote :

Compiz disabled, vga=773 enabled, pipeA enabled: managed over 48hours uptime.
Compiz disabled, vga=773 enabled, pipeA removed: machine froze twice, once whilst not in use, once with screen closed - pipeA required for correct function of powersaving features ?

Bryce Harrington (bryce) on 2008-10-08
Changed in xserver-xorg-video-intel:
importance: Undecided → High
status: Incomplete → Triaged
Bryce Harrington (bryce) wrote :

Hi Nick,

Checking in to see how things are sitting? Has the no-compiz/vga=773/pipeA combo eliminated the issues?

Also, did you run any tests with pipeA enabled but no vga=773 specified? I'm wondering if the latter makes any difference.

Changed in xserver-xorg-video-intel:
status: Triaged → Incomplete
Nick Rhodes (ngrhodes) wrote :


Yes, I have tested vga=773 and it appears to made no difference to system stability (just means I can see the splash screens as the OS loads).

The no-compiz/pipeA combination is still 100% stable.

Cheers, Nick

Bryce Harrington (bryce) on 2008-10-28
Changed in xserver-xorg-video-intel:
assignee: nobody → bryceharrington
Bryce Harrington (bryce) wrote :

I've attached a patch to bug #288859 that should fix this one too.

Bryce Harrington (bryce) wrote :

Already fixed in Intrepid/Jaunty. As mentioned above, fix is committed to Hardy for SRU.

Changed in xserver-xorg-video-intel:
status: Incomplete → Fix Released
status: New → Fix Committed
Martin Pitt (pitti) wrote :

Accepted into intrepid-proposed, please test and give feedback here. Please see for documentation how to enable and use -proposed. Thank you in advance!

Philip Wyett (philwyett) wrote :

Any updates available from those affected regards the package with patch in hardy-proposed as it it 49 days in proposed and really should be now pushed to updates if there are no objections.



Martin Pitt (pitti) wrote :

Philip, did you test the package from -proposed, and did it work for you?

Launchpad Janitor (janitor) wrote :

This bug was fixed in the package xserver-xorg-video-intel - 2:2.2.1-1ubuntu13.8

xserver-xorg-video-intel (2:2.2.1-1ubuntu13.8) hardy-proposed; urgency=low

  * 25_8xx_pipea_quirk.patch: Use Pipe-A quirk for all 855 and earlier
    chips. In Hardy we've been patching 8xx chipsets individually up
    until now, but upstream investigated and found this issue affects all
    of the early 8xx chips. This patch was introduced upstream about 6
    months ago and has been shipping in Intrepid. It safely solves a wide
    range of xorg-hang-on-lid-close issues.
    (LP: #288859, #267570)

 -- Bryce Harrington <email address hidden> Mon, 03 Nov 2008 17:14:46 -0800

Changed in xserver-xorg-video-intel:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers