[Q45] 1600x1200 resolution is not available on the 2nd monitor (it was on ubuntu 9.04)

[[e-mail]]

Binary package hint: xserver-xorg-video-intel

After installing ubuntu 9.10 I cannot set the resolution of my 2nd monitor higher then 1280 x 1024. Previously both of my monitors was working on 1600 x 1024. I have an Integrated Intel Graphics Media Accelerator 4500 in a HP Compaq dc7900 Ultra-slim Desktop PC (http://h10010.www1.hp.com/wwpc/us/en/sm/WF06a/12454-12454-64287-321860-3328898-3785404.html)

ProblemType: Bug
Architecture: amd64
Date: Mon Nov 2 10:52:09 2009
DistroRelease: Ubuntu 9.10
MachineType: Hewlett-Packard HP Compaq dc7900 Ultra-Slim Desktop
Package: xserver-xorg-video-intel 2:2.9.0-1ubuntu2
ProcCmdLine: root=UUID=c5c9baa8-4906-479d-8c3d-caa3202d1199 ro quiet splash
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-14.48-generic
RelatedPackageVersions:
 xserver-xorg 1:7.4+3ubuntu7
 libgl1-mesa-glx 7.6.0-1ubuntu4
 libdrm2 2.4.14-1ubuntu1
 xserver-xorg-video-intel 2:2.9.0-1ubuntu2
 xserver-xorg-video-ati 1:6.12.99+git20090929.7968e1fb-0ubuntu1
SourcePackage: xserver-xorg-video-intel
Uname: Linux 2.6.31-14-generic x86_64
dmi.bios.date: 03/05/2009
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: 786G1 v01.16
dmi.board.asset.tag: CZC9363FV9
dmi.board.name: 3033h
dmi.board.vendor: Hewlett-Packard
dmi.chassis.asset.tag: CZC9363FV9
dmi.chassis.type: 15
dmi.chassis.vendor: Hewlett-Packard
dmi.modalias: dmi:bvnHewlett-Packard:bvr786G1v01.16:bd03/05/2009:svnHewlett-Packard:pnHPCompaqdc7900Ultra-SlimDesktop:pvr:rvnHewlett-Packard:rn3033h:rvr:cvnHewlett-Packard:ct15:cvr:
dmi.product.name: HP Compaq dc7900 Ultra-Slim Desktop
dmi.sys.vendor: Hewlett-Packard
fglrx: Not loaded
system:
 distro: Ubuntu
 architecture: x86_64kernel: 2.6.31-14-generic

[Geir Ove Myhr (gomyhr)]

Thank you for reporting this bug and helping to make ubuntu better.

If I understand your log files correctly you have one monitor attached to a VGA port and one (a Iiyama PLE511S) to a HDMI port. The one on HDMI only shows 1280x1024. Right? You say the PLE511S is supposed to have 1600x1200, right?

We would like some more information in order to be able to troubleshoot this problem:
- Does it make a difference if you disconnect the monitor on the VGA port? I.e. so that the PLE 511S is the only monitor attached.
- With only the PLE511S connected, could you run `sudo get-edid | parse-edid` and attach the output as an attachment here (you will need the read-edid package for this).
- Could you start ubuntu with nomodeset on the kernel command line ('Esc' on boot to display the Grub menu, 'e' to edit the command line for the selected line, add nomodeset after quiet splash). This may give you full resolution on both monitors. In any case, please attach /var/log/Xorg.0.log and the output of `xrandr --verbose` here.

Btw, it seems you are not the first to have resolution problems with PLE511S: http://www.nvnews.net/vbulletin/showthread.php?t=80516

[[e-mail] (ubuntu-szantocsalad)]

Hi

Yes you are reading the logs correctly. Actually I have two identical IIyama monitors. One on the VGA the other On the HDMI.I added the nomodeset on the grub menu, but it had no positive effect. I also tried to disable the monitor on VGA, but the 1600x1200 resolution does not become available on the HDMI output. However I attached an HP monitor to the HDMI detected it, so I had one IIyama on the VGA and one HP on the HDMI. At this point 1600x1200 was available on both of my monitors, selected it, then put replaced the HP monitir on HDMI with the IIyama and it still worked. Of course it only works until I detect again the monitors. I send you attached the output when only the IIyama was attached on HDMI.

Regards

Peter

[Geir Ove Myhr (gomyhr)]

So you have the IIyama monitor attached to the HDMI output of the computer, but is it connected to a HDMI or a DVI input on the monitor? The EDID listed in Xorg.0.log is only 128 bytes long, but as far as I know, HDMI devices are required to have extension block making the total block size bigger. Does `sudo get-edid >hdmi-edid.bin` make only a 128 byte file, or could it be that X doesn't read all the blocks?

[Geir Ove Myhr (gomyhr)]

Btw, the HDMI specification version 1.3c is available at http://www.hdmi.org/download/HDMISpecification13a.pdf . The relevant section is 8.3.

[[e-mail] (ubuntu-szantocsalad)]

This is a 5 years old monitor with DVI and D-sub inputs. The HDMI output of the computer is connected to the DVI input of the monitor (and the d-sub to the d-sub on the other monitor)

[Geir Ove Myhr (gomyhr)]

Good, then the size of the EDID block makes sense. Not sure if there is something else which is non-standard with the EDID information. I guess it has to be something, since the 1600x1200 mode is not listed in Xorg.0.log after the EDID for HDMI1.

Since it does not make a difference for the resolution and we want to keep the test case as simple as possible, we would like to have only the HDMI/DVI connected when troubleshooting this. Also please move away /etc/X11/xorg.conf so that we have a standard configuration (it should not make any difference for the behaviour). Could you provide `xrandr --verbose`, `dmesg` and Xorg.0.log for that case?

We would also like the same logs for the UMS case (i.e. nomodeset boot option). Even though it does not make any difference in behaviour, there may be differences in the logs.

If you have access to a computer (with intel graphics chipset) with DVI output it would also be a useful data point to know what happens if you connect the monitor to that. You don't need to install ubuntu on that computer, a LiveCD/LiveUSB test would do. Still, the logs for that case would be interesting (can be copied off the LiveCD session).

[[e-mail] (ubuntu-szantocsalad)]

Hi

I removed xorg.conf and I created the three files for you. These are without setting the nomodeset because ubuntu wont start up when setting nomodeset (Last time I added nomodeset to the menu, but pressed escape to exit so the modifications were lost). If it helps I can boot my machine with an Ubuntu 9.04 live CD where the 1600x1200 resolution is available and send you log files.

P

[Geir Ove Myhr (gomyhr)]

> If it helps I can boot my
> machine with an Ubuntu 9.04 live CD where the 1600x1200 resolution is
> available and send you log files.

That would be nice. The driver included in Jaunty is not as verbose by
default as we would like, though. To make it include the mode-setting
decisions in the log file, edit /etc/X11/xorg.conf (which should be
there on a Jaunty LiveCD IIRC, even though it is not present by
default in Karmic) and to the "Device" section the line
  Option "ModeDebug" "true"
Then log out. Xorg will then be restarted and a new Xorg.0.log with
the debug information generated. You will be automatically logged in
again after a little while. You can then get the log file and `xrandr
--verbose`.

With this information I think we will have all we need to pass it on
to the intel developers.

[Geir Ove Myhr (gomyhr)]

Forwarding a bug report from ubuntu user <email address hidden>:
  https://bugs.launchpad.net/bugs/470422

[Problem]
Two identical IIyama PLE511S monitors are connected to the computer with Q45 chipset. One is connected to the VGA (D-sub) output and uses the D-sub input on the monitor. The other is using the HDMI output on the computer and DVI input on the monitor. The native resolution of the monitors are 1600x1200, but this seems to only be available on the VGA connected one. The computer reads the EDID just fine from both monitors and they both look sane at first sight. parse-edid gives a mode with 1600x1200 resolution. This is with KMS. The user reports unrelated problems that makes it impossible to test with UMS.

I find it a bit strange that the HDMI-DVI connected monitor is reported as HDMI instead of DVI, but that may be normal [1].

The problem persists if the VGA connected monitor is disconnected. Under Ubuntu 9.04 (with UMS and intel driver 2.6.3 + some patches) the monitor was assigned a resolution of 1600x1200.

[1]: http://www.hdmi.org/download/HDMISpecification13a.pdf section 8.3.3 says that when there are no EDID extension blocks it should be treated as a DVI.

[Original]

Binary package hint: xserver-xorg-video-intel

After installing ubuntu 9.10 I cannot set the resolution of my 2nd monitor higher then 1280 x 1024. Previously both of my monitors was working on 1600 x 1024. I have an Integrated Intel Graphics Media Accelerator 4500 in a HP Compaq dc7900 Ultra-slim Desktop PC (http://h10010.www1.hp.com/wwpc/us/en/sm/WF06a/12454-12454-64287-321860-3328898-3785404.html)

This is a 5 years old monitor with DVI and D-sub inputs. The HDMI output of the computer is connected to the DVI input of the monitor (and the d-sub to the d-sub on the other monitor)

ProblemType: Bug
Architecture: amd64
Date: Mon Nov 2 10:52:09 2009
DistroRelease: Ubuntu 9.10
MachineType: Hewlett-Packard HP Compaq dc7900 Ultra-Slim Desktop
Package: xserver-xorg-video-intel 2:2.9.0-1ubuntu2
ProcCmdLine: root=UUID=c5c9baa8-4906-479d-8c3d-caa3202d1199 ro quiet splash
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-14.48-generic
RelatedPackageVersions:
 xserver-xorg 1:7.4+3ubuntu7
 libgl1-mesa-glx 7.6.0-1ubuntu4
 libdrm2 2.4.14-1ubuntu1
 xserver-xorg-video-intel 2:2.9.0-1ubuntu2
 xserver-xorg-video-ati 1:6.12.99+git20090929.7968e1fb-0ubuntu1
SourcePackage: xserver-xorg-video-intel
Uname: Linux 2.6.31-14-generic x86_64
dmi.bios.date: 03/05/2009
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: 786G1 v01.16
dmi.board.asset.tag: CZC9363FV9
dmi.board.name: 3033h
dmi.board.vendor: Hewlett-Packard
dmi.chassis.asset.tag: CZC9363FV9
dmi.chassis.type: 15
dmi.chassis.vendor: Hewlett-Packard
dmi.modalias: dmi:bvnHewlett-Packard:bvr786G1v01.16:bd03/05/2009:svnHewlett-Packard:pnHPCompaqdc7900Ultra-SlimDesktop:pvr:rvnHewlett-Packard:rn3033h:rvr:cvnHewlett-Packard:ct15:cvr:
dmi.product.name: HP Compaq dc7900 Ultra-Slim Desktop
dmi.sys.vendor: Hewlett-Packard
fglrx: Not loaded
system:
 distro: Ubuntu
 architecture: x86_64kernel: 2.6.31-14-generic

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=30958)
Xorg.0.log with only the HDMI/DVI connected

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=30959)
dmesg with only the HDMI/DVI connected

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=30960)
xrandr --verbose with only HDMI/DVI connected

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=30961)
Xorg.0.log from Ubuntu 9.04 (with both VGA and HDMI/DVI connected)

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=30962)
xrandr --verbose from Ubuntu 9.04 (with both VGA and HDMI/DVI)

[[e-mail] (ubuntu-szantocsalad)]

I generated both files under Ubuntu 9.04 . Let me know if you need anything else.

[Geir Ove Myhr (gomyhr)]

I have reported this bug upstream at https://bugs.freedesktop.org/show_bug.cgi?id=24916 . Please subscribe to the upstream bug report by registering at bugs.freedesktop.org and adding yourself in the CC field. Then you will be able to answer questions from the upstream developers.

[Gordon Jin (gordon-jin)]

Increasing priority as this is regression from user's point of view, though maybe it's caused by UMS->KMS transition.

Let's make it simpler: this problem exists even only with the one monitor with HDMI->DVI connection, right? If so, let's don't say it's 2-monitor issue.

[Geir Ove Myhr (gomyhr)]

Yes, the problem exists even if only the one monitor with HDMI->DVI is connected. Changed titile accordingly. For Ubuntu 9.04 (-intel 2.6.3) I only got logs with both connected, but I hope that is fine.

[[e-mail] (ubuntu-szantocsalad)]

thanks, I registered there too

[Carl Worth (cworth)]

Retitling to indicate the driver versions that were reported as working and
not working.

Also assigning to <email address hidden> who has some experience with bugs in
this area of the driver.

-Carl

[yakuizhao (yakui-zhao)]

From the log it seems that 2.6.31.xx kernel is used.
Will you please try the latest kernel(2.6.32-rc6) and see whether the issue still exists?

In fact this issue can be resolved by the following commit:
   >commit 5c61259e6c7290082108e93815f7f72f27da14f4
Author: Zhao Yakui <email address hidden>
Date: Mon Jun 22 13:17:10 2009 +0800

    drm/mode: get the modeline for standard timing in EDID by using CVT/GTF

And the above commit is already shipped in current linus git tree.

So this bug will be marked as resolved.

Thanks.

[Geir Ove Myhr (gomyhr)]

This bug has been marked as fixed upstream. The patch that is supposed to fix it is included in kernel 2.6.32-rc6. You can test this kernel by installing the package http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.32-rc6/linux-image-2.6.32-020632rc6-generic_2.6.32-020632rc6_amd64.deb.

[[e-mail] (ubuntu-szantocsalad)]

I downloaded the new kernel, and I can set the display 1600x1200 I was happy and excited... then I tried to install the free VMWare Player. It first complained about the missing kernel headers for 2.6.32-020632rc6 which I donwloaded from the same location as the kernel. Then VMWare appeared in the start menu, but at first start it wanted to compile further components which failed. I uninstalled the new kernel re started the VMWare installation and it worked fine. Given that the kernel is an RC is it possible that it is not correct?

[Geir Ove Myhr (gomyhr)]

Your test confirms that this particular problem is fixed upstream,
most likely by the patch mentioned in the upstream bug report. See
https://wiki.ubuntu.com/KernelTeam/MainlineBuilds for more information
about the mainline kernel builds.

I don't know enough about VMWare or the kernel to tell if the VMWare
problems is a bug or simply due to something ubuntu-specific which is
not included in the mainline build. You may try Lucid, which currently
has a kernel based on 2.6.32-rc6 (see
http://kernel.ubuntu.com/~kernel-ppa/info/kernel-version-map.html). It
is very early in the Lucid release cycle, so don't expect it to be
stable, but if it starts all right for you and the problem is there as
well, please file a bug report using `ubuntu-bug linux`.

You may also have some luck trying VMWare forums with the specific
error messages you get during compilation. They may tell you what the
problem is.

[[e-mail] (ubuntu-szantocsalad)]

Thanks for helping, I am over all very impressed by the level of support I received. Given that the patch works, but I cannot apply because it causes other problems. Do you know if there is any workaround? If I attach an HP monitor to my PC detect it then plug back the IIyama monitor it works fine, the only problem is that if I restart the computer I loose the settings. Can I somehow save the settings when the HP monitor is attached and restore it later?

[Geir Ove Myhr (gomyhr)]

This bug should be fixed in Lucid and should probably have been closed a while ago. Closing it now.