[CLE266][regression] Jaunty and Intrepid screen xserver won't start

[BadBoy]

Computer: Patriot 2004 (old laptop)
Graphics: VIA Technologies, Inc. VT8623 [Apollo CLE266] integrated CastleRock graphics (Reported by 8.04 LTS)
System: Live CD Jaunty

The system boots just fine. I can see the new Jaunty Boot logo while booting.
Suddenly it disappears and I am at a text console.

Typing startx produces:
CHROME(0): No valid modes found

The display works perfectly under 8.04 LTS. Intrepid and Jaunty are affected by this regression.

[lspci]
00:00.0 Host bridge [0600]: VIA Technologies, Inc. VT8623 [Apollo CLE266] [1106:3123]
     Subsystem: Elitegroup Computer Systems Device [1019:b320]
01:00.0 VGA compatible controller [0300]: VIA Technologies, Inc. VT8623 [Apollo CLE266] integrated CastleRock graphics [1106:3122] (rev 03)
     Subsystem: Elitegroup Computer Systems Device [1019:b320]

[BadBoy (sklep-szybkieczytanie)]

Attaching the xorg.conf from the 8.04 LTS.

[BadBoy (sklep-szybkieczytanie)]

Here is a screenshot of modes detected by 8.04 LTS
This is after typing the "scan" command.

[BadBoy (sklep-szybkieczytanie)]

Here is a screenshot of modes detected by Jaunty Live
This is after typing the "scan" command.

[BadBoy (sklep-szybkieczytanie)]

Overwriting the faulty config from Jaunty with a working config file taken from 8.04 LTS does not solve the problem.

[Martin Olsson (mnemo)]

To automatically attach more detailed logs can you please run this command on the affected computer:
apport-collect 367442

[BadBoy (sklep-szybkieczytanie)]

Architecture: i386
DistroRelease: Ubuntu 9.04
MediaBuild: Ubuntu 9.04 "Jaunty Jackalope" - Release i386 (20090420.1)
Package: xserver-xorg-video-openchrome 1:0.2.903+svn713-1ubuntu1
PackageArchitecture: i386
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
ProcVersion: Linux version 2.6.28-11-generic (buildd@palmer) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #42-Ubuntu SMP Fri Apr 17 01:57:59 UTC 2009
Uname: Linux 2.6.28-11-generic i686
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
Xrandr:

glxinfo:

setxkbmap:

xdpyinfo:

xkbcomp:

[BadBoy (sklep-szybkieczytanie)]

Architecture: i386
DistroRelease: Ubuntu 9.04
MediaBuild: Ubuntu 9.04 "Jaunty Jackalope" - Release i386 (20090420.1)
Package: xserver-xorg-video-openchrome 1:0.2.903+svn713-1ubuntu1
PackageArchitecture: i386
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
ProcVersion: Linux version 2.6.28-11-generic (buildd@palmer) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #42-Ubuntu SMP Fri Apr 17 01:57:59 UTC 2009
Uname: Linux 2.6.28-11-generic i686
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
Xrandr:

glxinfo:

setxkbmap:

xdpyinfo:

xkbcomp:

[BadBoy (sklep-szybkieczytanie)]

Architecture: i386
DistroRelease: Ubuntu 9.04
MediaBuild: Ubuntu 9.04 "Jaunty Jackalope" - Release i386 (20090420.1)
Package: xserver-xorg-video-openchrome 1:0.2.903+svn713-1ubuntu1
PackageArchitecture: i386
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
ProcVersion: Linux version 2.6.28-11-generic (buildd@palmer) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #42-Ubuntu SMP Fri Apr 17 01:57:59 UTC 2009
Uname: Linux 2.6.28-11-generic i686
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
Xrandr:

glxinfo:

setxkbmap:

xdpyinfo:

xkbcomp:

[BadBoy (sklep-szybkieczytanie)]

I had to install python-launchpadlib. In order to do that I have copied the sources.list from my other computer with Jaunty.

1) First apport-collect is done after having downloaded the package
2) Second is after trying to launch startx
3) Third is after having copied the xorg.conf from the 8.04 LTS

Thank you for looking into the problem :)

[Martin Olsson (mnemo)]

I've forwarded your bug reported upstream to the developers of the openchrome driver. Feel free to subcribe to the upstream bug report as well, in case they have additional questions.

[Martin Olsson (mnemo)]

Dear BadBoy,

Upstream developers have indicated that this upstream commit might fix your issue:
http://www.openchrome.org/trac/changeset/738

It would be great to hear if this change works on your machine.

---

To test the patch you need to do something like:
1. Create an empty directory
2. run "apt-get source xserver-xorg-video-openchrome"
3. run "apt-get build-dep xserver-xorg-video-openchrome"
4. cd into the actual package directory
5. apply the patch / do the changes to the source code
6. run "debuild -us -uc -b" to compile the driver and build a .DEB
7. do "cd .." and then do "sudo dpkg -i *.deb" to install the updated driver

Also, to revert back to the original jaunty driver you can probably do:
sudo apt-get install --reinstall xserver-xorg-video-openchrome/jaunty

[Bryce Harrington (bryce)]

Patch looks pretty sane; this is probably going to be acceptable as an SRU candidate.
Thanks for chasing the patch down, mnemo.

[BadBoy (sklep-szybkieczytanie)]

Hi Martin,

I've been running the Live CD Jaunty only.
Is it possible to test your solution on the Live session?

[Bryce Harrington (bryce)]

I've added an openchrome package with this patch here:
https://edge.launchpad.net/~bryceharrington/+archive/ppa

[Martin Olsson (mnemo)]

@BadBoy, since Bryce created a package with the patch it's now sufficient if you download and install the updated DEBs.

Basically you can do:
wget http://ppa.launchpad.net/bryceharrington/ppa/ubuntu/pool/main/x/xserver-xorg-video-openchrome/xserver-xorg-video-openchrome_0.2.903+svn713-1ubuntu1.1~bug367442~1_i386.deb
wget http://ppa.launchpad.net/bryceharrington/ppa/ubuntu/pool/main/x/xserver-xorg-video-openchrome/xserver-xorg-video-via_0.2.903+svn713-1ubuntu1.1~bug367442~1_i386.deb

Followed by:
sudo dpkg -i xserver-xorg-video-openchrome_0.2.903+svn713-1ubuntu1.1~bug367442~1_i386.deb
sudo dpkg -i xserver-xorg-video-via_0.2.903+svn713-1ubuntu1.1~bug367442~1_i386.deb

Once you have done this, try to start X again and finally re-run the log file collection command so we can see how it went:
apport-collect 367442

It should be possible to do all these steps straight from the live CD and if you're lucky and the upstream dev was right then you might be able to run the apport-collect command from a terminal inside a a running X.org

[BadBoy (sklep-szybkieczytanie)]

I was too happy. And too early.

Seems that I am able to launch GDM only by:

1) $ sudo startx
2) Black screen nothing happens
---> apport-collect
3) Pressing Ctrl+Alt+F2
4) $ sudo /etc/init.d/gdm restart
5) GDM finally loads
---> apport-collect

[Martin Olsson (mnemo)]

Ok just to summarize...

First you run startx and then you got:
http://launchpadlibrarian.net/26065697/XorgLog.txt
http://launchpadlibrarian.net/26065698/XorgLogOld.txt

Then you VT-switch and restart gdm, after which you got:
http://launchpadlibrarian.net/26065777/XorgLog.txt
http://launchpadlibrarian.net/26065790/XorgLogOld.txt

[Martin Olsson (mnemo)]

Honestly, from that first xorg.log it looks like stuff it working. There is nothing printed into "dmesg" when you're doing these attempts is it?

To debug further, can you please throw in these options on by one into the device section of your xorg.conf and see if you can find a config that boots properly:
Option "AccelMethod" "XAA"
Option "NoAccel" "true"
Option "ExaNoComposite" "true"
(if you want to read about these options you can open "man openchrome")

For example to test with noaccel option you'd use something roughly equivalent to this /etc/X11/xorg.conf:

Section "Device"
        Identifier "Configured Video Device"
        Option "NoAccel" "true"
EndSection

Section "Monitor"
        Identifier "Configured Monitor"
EndSection

Section "Screen"
        Identifier "Default Screen"
        Monitor "Configured Monitor"
        Device "Configured Video Device"
EndSection

[BadBoy (sklep-szybkieczytanie)]

Yes, the order of those dumps is as indicated.
The first dump is taken before I see GDM
The second dump is taken after I log on successfully to GDM

Going to try you xorg.conf sugestion.

[BadBoy (sklep-szybkieczytanie)]

I have just finished the tests:

1) Option "NoAccel" "true"
System boots OK - no problems found.

2) Option "AccelMethod" "XAA"
System freezes as previously on "waiting" mouse cursor

3) Option "ExaNoComposite" "true"
System freezes a bit later with "arrow" mouse cursor

Seems that your NoAccell sugestion fixes all the issues! :)

[Martin Olsson (mnemo)]

ok so we're probably dealing with two different bugs here. Seems to me that we should merge the braces patch to karmic right away at least.

[BadBoy (sklep-szybkieczytanie)]

OK. First of all I am really amazed that you helped me so quickly!
Thank you!

Second of all, there is no acceleration in this notebook. No, I don't think so.
It's taking 100% CPU in order to move a window.

Still it's waaay faster on Jaunty than on 8.04 LTS.
I am preparing to upgrade to Jaunty now :)

[Bartosz Kosiorek (gang65)]

I think this bug was fixed by changeset 738. http://www.openchrome.org/trac/changeset/738.

Please build, install the latest version of openchrome.
The instructions how to do it is there:
https://help.ubuntu.com/community/OpenChrome

[Robert Moerland (veel-mail)]

Confirming that the 738 releases fixes the 'No valid modes found' error for me.

I had exactly the same problem, I was dropped to a text-only login screen after a successful boot. I have followed the guide at https://help.ubuntu.com/community/OpenChrome to build and install the 738 release. Rebooted the machine and immediately got a graphical login screen.

Thanks for the quick fix.

[Bartosz Kosiorek (gang65)]

This bug exist only in Ubuntu Jaunty and Karmic (xserver-xorg-video-openchrome 1:0.2.903+svn713-1ubuntu1).
It doesn't exist in Ubuntu Hardy and Ubuntu Intrepid

[Bartosz Kosiorek (gang65)]

This bug has been already fixed in the Openchrome 0.2.903+svn741-1 (Ubuntu Karmic)

[Bartosz Kosiorek (gang65)]

This bug is already fixed in the Karmic, in package 0.2.903+svn741-1

[BadBoy (sklep-szybkieczytanie)]

I am sorry to inform you that this bug still affects the following video card in Karmic:

01:00.0 VGA compatible controller [0300]: VIA Technologies, Inc. VT8623 [Apollo CLE266] integrated CastleRock graphics [1106:3122] (rev 03)

I think that the following solution might work:

Section "Device"
        Identifier "Configured Video Device"
        Option "NoAccel" "true"
EndSection

[Bartosz Kosiorek (gang65)]

Hello BadBoy.
What is native resolution of your LCD?

Please build/install the latest openchrome driver and enable debug mode.
The detail build instruction for Ubuntu/Debian is available at:
https://help.ubuntu.com/community/OpenChrome

Attach your /var/log/Xorg.0.log file.

[BadBoy (sklep-szybkieczytanie)]

The native resolution is 1024x768.

[BadBoy (sklep-szybkieczytanie)]

I am unable to load the system.
I don't even see the black and white ring...

I tried updating it from the command line. No luck.

[BadBoy (sklep-szybkieczytanie)]

Downgrading the driver to the one you sent me the last time (which worked in Jaunty) did not help.

[BadBoy (sklep-szybkieczytanie)]

I bet that all you have to do is make the noaccel option set to true by default for this graphic card. This way I would be able to install the system without any trouble.

[Bartosz Kosiorek (gang65)]

Attach your /var/log/Xorg.0.log file.
If you can from after reproducing this issue (or /var/log/Xorg.log.old).

[Bartosz Kosiorek (gang65)]

Try this patch with the latest openchrome revision.
Attach your Xorg.0.log file.

[BadBoy (sklep-szybkieczytanie)]

Bartosz, I am doing everything I can to install this new system. Trying to upgrade Jaunty.
I am unable to update nor do anything because the system freezes before the black-and-white ring.
The alternate Ubuntu installer is lame and does not install proper Polish version of the system.

Bartosz. NIE DA SIE TEGO ZROBIC. :)

[BadBoy (sklep-szybkieczytanie)]

CONFIRMED.

Actually I managed to upgrade the distro from Jaunty to Karmic.
So the "noaccel" option was left at custom value set to true.

The system launched and is running.

[BadBoy (sklep-szybkieczytanie)]

> Please build/install the latest openchrome driver and enable debug mode.
> The detail build instruction for Ubuntu/Debian is available at:
> https://help.ubuntu.com/community/OpenChrome

Having done that, the following has changed:

1) It is possible to boot and actually see the desktop (no freezes) without the "noaccel" option! Bravo!
2) The screen updates in a strange way. When you scroll pages only portions of information are updated. When you drag a window, only some of it is dragged and some is left at a previous position

Setting "noaccel" option to true fixes the problem described above (2)

[Bartosz Kosiorek (gang65)]

Attach your /var/log/Xorg.0.log file (without noaccell option), with enabled debug mode, after reproducing the scroll bug.

[BadBoy (sklep-szybkieczytanie)]

How do I enable the debug mode?

[Bartosz Kosiorek (gang65)]

This bug was fixed in openchrome svn r805
http://www.openchrome.org/trac/changeset/805

Check the latest openchrome revision.
How to enable debug mode was described at:
https://help.ubuntu.com/community/OpenChrome#Problems and solutions

[Bartosz Kosiorek (gang65)]

Hi.
Please check the latest openchrome revision.

Try also add following option to xorg.conf file:

       Option "modeSwitch" "new"

[BadBoy (sklep-szybkieczytanie)]

Tried to install Ubuntu using the alternate CD:

I installed the base system and the ubuntu-desktop using aptitude. But I was unable to build a debug version of Openchrome driver (SVN). When trying to "configure" the package I get the error saying that Xorg Server is not installed amongst other three packages the names of which I cannot recall right now. Four packages missing altogether that were not available in the repository. Building Openchrome driver failed.

The Live CD 9.10 installation was a real nightmare, too.

I used a "safe graphics" installation only to see that my graphics card was blinking the whole screen not able to load the X Server (the graphics). I fixed this by issuing the following command:

# /etc/init.d/gdm stop

Then, as always, I added the "noaccel" option to the xorg.conf file and everything went smooth after all.

I hope the above information helps in finding the bug.

NOW, let me try the "modeSwitch" "new" :)

[BadBoy (sklep-szybkieczytanie)]

OK. Built it successfully. And checked it!

1) Commented out the "no accel" option
2) Added modeSwitch new option

Still there is the problem with disappearing windows. When I drag a window by clicking and holding the title bar in the right-bottom direction it disappears looking like being eaten from above and left side of the window. After several tries the system finally freezes, although it's possible to access it by SSH.

Not everything is refreshed. Like the notification area in the upper-right corner. There is a half a Volume Indicator icon in place of Ubuntu One icon. The login screen was distorted and looked like a puzzle, although most part of it was visible.

But I was able to scroll the screen when using Chromium Browser. It looked perfectly!

Attaching my xorg.logs

[BadBoy (sklep-szybkieczytanie)]

Xorg.0.log.old

[BadBoy (sklep-szybkieczytanie)]

Xorg.1.log

[Bartosz Kosiorek (gang65)]

Hello.
Try to disable DRI, by adding option to your xorg.conf file.

Is with diabled DRI and modeSwitch new option is working correctly?

[BadBoy (sklep-szybkieczytanie)]

How do I disable DRI?

[Bartosz Kosiorek (gang65)]

Edit /etc/X11/xorg.conf and disable DRI in the "Module" section:

    Section "Module"
       ...
       Disable "dri"
       ...
    EndSection

[BadBoy (sklep-szybkieczytanie)]

1) Commented out the "no accel" option
2) Added modeSwitch new option
3) Disabled DRI

All works like a charm. Bartosz, you are brilliant :)
Are you going to make these options default in future releases?

[Timo Aaltonen (tjaalton)]

fixed in lucid.

[Timo Aaltonen (tjaalton)]

actually, won't close it before it's confirmed to work on lucid

[Timo Aaltonen (tjaalton)]

enough waiting, closing as fixed.