[i855] [x-updates] X fails to start with "Cannot run in framebuffer mode" error

[Paganini]

Binary package hint: xserver-xorg-video-intel

After updating to the most recent xserver-xorg-video-intel found in the x-swat PPA for Jaunty (uploaded on the 19th of October I think) X fails to start. The log contains this error:

(EE) Failed to load module "i810" (module does not exist, 0)

but this message is harmless and can safely be ignored. The real error is this one:

"Cannot run in framebuffer mode"

This is in fact the same as bug #383407 in Karmic.

WORKAROUND:
Create a minimal xorg.conf, for instance like this:

Section "Device"
    Identifier "my-intel-card"
    Driver "intel"
EndSection

I'm on an IBM Thinkpad X40 laptop with an Intel 82852/855GM Integrated Graphics Device [8086:3582] (rev 02)

ProblemType: Bug
Architecture: i386
DistroRelease: Ubuntu 9.04
NonfreeKernelModules: ath_hal
Package: xserver-xorg-video-intel 2:2.6.3-0ubuntu9.3
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
ProcVersion: Linux version 2.6.28-16-generic (buildd@vernadsky) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #55-Ubuntu SMP Tue Oct 20 19:48:24 UTC 2009
SourcePackage: xserver-xorg-video-intel
Uname: Linux 2.6.28-16-generic i686

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation 82852/82855 GM/GME/PM/GMV Processor to I/O Controller [8086:3580] (rev 02)
     Subsystem: IBM Device [1014:055c]
00:02.0 VGA compatible controller [0300]: Intel Corporation 82852/855GM Integrated Graphics Device [8086:3582] (rev 02)
     Subsystem: IBM Device [1014:0557]

[Paganini (nebanks)]

Also note that the files that ubuntu-bug automatically attached may have misleading information - in order to get X to start so I could file the bug I had to revert to the xserver-xorg-video-intel package found in the main Jaunty repositories.

Xorg.0.log.borked is the log from the failed attempt, which I made without an xorg.conf.

[Tormod Volden (tormodvolden)]

Please also attach your dmesg output. What happens (logs) if you have a minimal xorg.conf which specifies the "intel" driver?

[Paganini (nebanks)]

Here is my dmesg output

[Paganini (nebanks)]

With a minimal xorg.conf specifying the intel driver X *does* in fact start (woohoo!) - unfortunatetly, my mouse pointer is invisible (darn!).

[Paganini (nebanks)]

I merged my minimal xorg.conf with my previous xorg.conf. Everything appears to be working smoothly now, although I still get unwatchably slow framerates in .MKV video. (Trying to watch .MKVs is the reason started using the x-swat drivers in the first place.)

The weird thing is, my mouse pointer is invisible until I lock and unlock the screen. Then it returns. No idea what that's all about.

[Paganini (nebanks)]

And here's dmesg output from the boot with the working (but invisible pointer) X

[Paganini (nebanks)]

And here's Xorg.0.log from the same boot

[Paganini (nebanks)]

The invisible pointer issue seems to be setting related. Some setting associated with my account doesn't like the new driver configuration. I created a new account for testing and the mouse pointer works fine there. Anyone have an idea what setting might be messing things up?

[Tormod Volden (tormodvolden)]

In Jaunty, we had 116_8xx_disable_dri.patch which turned off DRI for i8xx cards:
"because it fails to run without freezing on i810 and i865G chips. (See LP 304871)"

The x-updates (and Karmic) packages have reenabled DRI, that's why you might get trouble.

Can you please confirm that you are running without DRI with the Jaunty package, and that you can run the x-updates package fine with DRI disabled? Use this in the Device section of your xorg.conf: Option "DRI" "off"

[Paganini (nebanks)]

Hi Tormod,

I updgraded to Karmic yesterday, so this will be more difficult for me to test. I will have more resources when I get home next week, so I will look into it then.

Just a side note, Karmic has the same issue with the mouse pointer that the x-updates package did in Jaunty - it's invisible when I first boot up, but appears after I lock and unlock the screen.

[Michael Langerhorst (m-langerhorst)]

@Tormod Volden:
I reinstalled 9.04 in the meantime (had an image). When I looked up my xorg.conf, I didn't find any option with DRI. You find my xorg as an attachment. Is there a realistic chance that intel will patch this driver within a couple of weeks? As far as I know the buggy intel driver is quite a while well known.
I wasn't able to install your patch - I would need a better description.

[Tormod Volden (tormodvolden)]

Paganini, the official Karmic version is the same as in x-updates/Jaunty so I expect the same issue there, unless some kernel changes have fixed it.

Michael, there is no such option in xorg.conf by default, you will have to add it. In Jaunty it was disabled in the driver, not in the xorg.conf.

[Albert Damen (albrt)]

Log message:

Fatal server error:
Cannot run in framebuffer mode. Please specify busIDs for all framebuffer devices

is the same we have seen in Karmic in bug 383407. It was fixed in the xserver:

xorg-server (2:1.6.1.901-2ubuntu2) karmic; urgency=low

  * Add xserver-1.5.0-bad-fbdev-thats-mine.patch - If no xorg.conf is
    specified, framebuffer device can erroneously grab the PCI. Make
    it fail instead in this case.
    (LP: #383407)

[Tormod Volden (tormodvolden)]

Thanks Albert! So we could make a patched xorg-server for x-updates/jaunty but on the other hand I feel like end-of-line'ing it. For now I think a minimal xorg.conf specifying "intel" is a good enough workaround. I will leave the bug report open so people can find it, but from my side it is a "wontfix".

Paganini, please make sure your pointer issue is reported in Karmic.

[JC Hulce (soaringsky)]

Thank you for taking the time to report this bug. This issue has been fixed in newer versions of Ubuntu, and Jaunty is EOL, so I am closing this bug task.