[i945GME] intel dual screen not working on karmic

[alain57]

Hello,
on ubuntu 9.04 dual screen worked, but was buggy (the external screen displayed only half (vertical split), but the fun was, that putting a window in full screen on that external display worked... only window mode had a bug :(

but now i tried my luck on 9.10, and ... it's no more working

my netbook (with intel GMA 950) has a resolution of 1024x600
and the external monitor (a dell one) has a resolution of 1680x1050

so with the System => Preference => display tool (i guess the name, my ubuntu is French) i move the screens side by side with clone function disabled.
and then GDM as X won't start

here are the changes that the display tool made in my xorg.cong

SubSection "Display"
               Virtual 2704 1050
EndSubSection

when i start x with startx i receive a message that says "could not add fb"

so i did a dmesg | grep drm

here is the output :
[ 1.695113] [drm] Initialized drm 1.1.0 20060810
[ 2.062285] fb0: inteldrmfb frame buffer device
[ 2.062306] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
[ 2.136914] [drm] DAC-6: set mode 1680x1050 17
[ 2.448490] [drm] LVDS-8: set mode 1024x600 19
[ 12.974479] [drm:drm_mode_addfb] *ERROR* mode new framebuffer width not within limits
[ 14.069862] [drm:drm_mode_addfb] *ERROR* mode new framebuffer width not within limits
[ 15.241898] [drm:drm_mode_addfb] *ERROR* mode new framebuffer width not within limits
[ 16.245814] [drm:drm_mode_addfb] *ERROR* mode new framebuffer width not within limits
[ 17.061783] [drm:drm_mode_addfb] *ERROR* mode new framebuffer width not within limits
[ 18.106475] [drm:drm_mode_addfb] *ERROR* mode new framebuffer width not within limits

i can make further tests, so if you have any idea or tips, i'll be happy to help you guys resolve this bug...

[Philip Muškovac (yofel)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. This bug did not have a package associated with it, which is important for ensuring that it gets looked at by the proper developers. You can learn more about finding the right package at https://wiki.ubuntu.com/Bugs/FindRightPackage. I have classified this bug as a bug in xserver-xorg-video-intel.
For future reference you might be interested to know that a lot of applications have bug reporting functionality built in to them. This can be accessed via the Report a Problem option in the Help menu for the application with which you are having an issue. You can learn more about this feature at https://wiki.ubuntu.com/ReportingBugs.

[Geir Ove Myhr (gomyhr)]

We would like some more information from you.

1. Could you attach the output of `lspci -vvnn` as an attachment to this bug report? That helps us identify your hardware exactly.
2. Could you remove the file $HOME/.config/monitors.xml. This is the file that stores the monitors-side-by-side information, but we want it out of the way for troubleshooting purposes.
3. Could you attach the full output of dmesg?
4. Does the problem happen even without the external monitor connected?
5. Does the problem go away if you change 2704 in the virtual line to 2048, but come back for 2049?

[alain57 (alain57)]

here are my dmesg output with
SubSection "Display"
               Virtual 2704 1050
EndSubSection

1-screen is booting, without my external monitor connected
2-screen is booting with my external monitor connected

in both case, X won't run as long as i not modify the xorg.conf file and remove the subSection part.

[alain57 (alain57)]

i did some tests
funnilly when doing dual screen with both screen at 800x600 (not clone mode) it worked
i modify the screen size on my netbook at 1024x600 and keep a resolution of 800x600 on the external screen and it worked to

then i had the idea to increase the external screen's resolution to 1024x768
and it bugs ... only the cursor was visible, both screen were black (no background, no icon, no window....)

when i increase the resolution even more, then X won't start, maybe because of the 2048x2048 limitation

what is funny, is that this limitation did not exist on windows or on mac os (yes i tried hackintosh)

[alain57 (alain57)]

when the screen are black, i also notice that the cursor change... for exemple when i move over a textzone the pointer change to text pointer (but like i said, i see nothing :( )

i did other tests with the screen position utility
instead having both screen side by side, i put one on the top and the other at the bottom
in this configuration, my netbook worked well with 1024x600 and the external monitor with 1024x768

so dual screen only work if the width is the same when there are in top/bottom position
and dual screen also work if the height is the same when there are side by side

[alain57 (alain57)]

i just tested a new configuration

netbook with 1024x600 (on top)
external with 120x1024 (on bottom)

and so the dual screen work :)
when i put them side by side, the utility ask for a virtual resolution (in the previous cases, it did not ask and did not add modifications to xorg.conf..)

i will do further test with higher resolutions, but i'm at home, and my external monitor only goes up to 1280x1024, whereas the external monitor i use at work does better ...

if it work with a top/bottom configuration, i will use it as long as this bug is not corrected, but it's a bad usability

ps : it would be cool to have the screen utility work like on mac.
for exemple putting a high virtual resolution in xorg.conf so user can put there screens where they want, without needing to logout to modify xorg

[alain57 (alain57)]

i forgot to answer your last question

X run with
SubSection "Display"
               Virtual 2048 2048
EndSubSection

but it will not run if i increase the value at 2049

[Geir Ove Myhr (gomyhr)]

I didn't think about it at first, but this looks like it could be this upstream bug:
http://bugs.freedesktop.org/show_bug.cgi?id=21190

This was marked as fixed today. It has been committed to the development version of the driver (http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=d655a3ff423e69c19a5dc07140cbf3caaa32cb86), but there is a kernel fix which hasn't landed yet. There are two ubuntu bugs which has marked that bug as its upstream, bug 391418 and bug 351756. I suspect that once the fix for those bugs in ubuntu, this will be fixed as well.

[alain57 (alain57)]

yes that is exactly the same problem.
are there any ppa kernel or intel driver for karmic i can test ?

[Geir Ove Myhr (gomyhr)]

It looks like xserver-xorg-video-intel from xorg-edgers PPA (https://launchpad.net/~xorg-edgers/+archive/ppa) has the fix included. However, my guess is that it will only be fixed with KMS disabled (see https://lists.ubuntu.com/archives/ubuntu-x/2009-June/000578.html for how). The corresponding kernel patch hasn't been commited yet (we should be able to see it in this log when it is: http://git.kernel.org/?p=linux/kernel/git/anholt/drm-intel.git;a=summary) and from there until it is in the mainline kernel and built for ubuntu, I don't know how long it takes. I think the kernel patch is necessary to fix it for KMS.

[alain57 (alain57)]

i disabled KMS because of a brightness issue (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/388216)

i will try the ppa repository

[Bryce Harrington (bryce)]

Geir, if it does turn out that the kernel patch is needed to solve this, then please reassign the bug to linux tagged xorg-needs-kernel-fix, and attach the patch to this bug report.

[Geir Ove Myhr (gomyhr)]

It looks like the kernel patch has landed in Eric Anholt's drm-next respository: http://git.kernel.org/?p=linux/kernel/git/anholt/drm-intel.git;a=commit;h=5e4d6fa72619aeea271d2ad704757717b06e291a

alain57, it would be nice if you could
0. (optional, but nice) enable KMS to verify that your problem also occurs with KMS
1. with KMS enabled and the xorg-edgers PPA driver, check if it is works with the kernel you have
2. with KMS disabled, verify that the PPA version of the driver fixes this bug for you

[alain57 (alain57)]

xorg-edgers did not fix anything for me :(

[alain57 (alain57)]

there are definitively some improvement on the normal xorg intel driver (or gnome-display-properties)

i had the same problem on my other notebook (a 15" acer with intel graphics)
i just test this notebook again at work, and now the dual screen work !

i didn't use xorg-edgers on this one.

but i guess that there is still a problem on my 10", because the max resolution is 2048x2048 (funnily windows and macos don't have that limitation) whereas it is 4096x4096 on my 15"

a possible solution could be to increase the maximum resolution for the GMA 950 like it is on the X3100

here the output of xrandr on my 15" (i don't have the 10" right now)

Screen 0: minimum 320 x 200, current 2960 x 1050, maximum 4096 x 4096
VGA connected 1680x1050+1280+0 (normal left inverted right x axis y axis) 434mm x 270mm
   1680x1050 60.0*+
   1280x1024 75.0 60.0
   1152x864 75.0
   1024x768 75.0 60.0
   800x600 75.0 60.3
   640x480 75.0 59.9
   720x400 70.1
LVDS connected 1280x800+0+0 (normal left inverted right x axis y axis) 331mm x 207mm
   1280x800 60.0*+
   1024x768 85.0 75.0 70.1 60.0
   832x624 74.6
   800x600 85.1 72.2 75.0 60.3 56.2
   640x480 85.0 72.8 75.0 59.9
   720x400 85.0
   640x400 85.1
   640x350 85.1
TV disconnected (normal left inverted right x axis y axis)

[Bryce Harrington (bryce)]

Yeah, dual-head was being limited to a max of 2048 height or width on 945, but that has been fixed. It's a kernel issue really, and if the patch hasn't already been included in karmic's kernel, it soon will be.

[alain57 (alain57)]

thx for the news :)
i hoppe it will be include soon :)

[dienarr (dasnarr)]

this is still not working

[Geir Ove Myhr (gomyhr)]

dienarr, please report your problem as a separate bug report with the `ubuntu-bug` command. Since alain57 hasn't repoened this bug I assume his problem has been fixed.