[i965GM] Xorg starts to black screen with some corruption when using driver "VESA" (KMS bug)

[Geir Ove Myhr]

Binary package hint: xserver-xorg-video-vesa

When adding the line
  Driver "VESA"
to the "Device" section of xorg.conf, xorg starts up with only a black screen with a few lines of scattered white dots on the upper part. It starts normally when I use UMS instead of KMS (i915.modeset=0 on boot).

This is a Lenovo X61 Tablet which has intel 965GM graphics.

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile PM965/GM965/GL960 Memory Controller Hub [8086:2a00] (rev 0c)
 Subsystem: Lenovo Device [17aa:20b3]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller [8086:2a02] (rev 0c)
 Subsystem: Lenovo Device [17aa:20b5]

ProblemType: Bug
Architecture: i386
Date: Sat Jul 18 16:56:13 2009
DistroRelease: Ubuntu 9.10
MachineType: LENOVO 7764CTO
Package: xserver-xorg-video-vesa 1:2.2.0-1
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.31-3-generic root=UUID=938751f9-b020-46b7-a6d6-d4ec43b0c40a ro single
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-3.19-generic
RelatedPackageVersions:
 xserver-xorg 1:7.4+3ubuntu5
 libgl1-mesa-glx 7.5~rc4-1ubuntu3
 libdrm2 2.4.11-1ubuntu1
 xserver-xorg-video-intel 2:2.7.99.901+git20090702.74227141-0ubuntu1
 xserver-xorg-video-ati 1:6.12.99+git20090629.f39cafc5-0ubuntu1
SourcePackage: xserver-xorg-video-vesa
Uname: Linux 2.6.31-3-generic i686
dmi.bios.date: 07/29/2008
dmi.bios.vendor: LENOVO
dmi.bios.version: 7SET34WW (1.20 )
dmi.board.name: 7764CTO
dmi.board.vendor: LENOVO
dmi.board.version: Not Available
dmi.chassis.asset.tag: No Asset Information
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: Not Available
dmi.modalias: dmi:bvnLENOVO:bvr7SET34WW(1.20):bd07/29/2008:svnLENOVO:pn7764CTO:pvrThinkPadX61Tablet:rvnLENOVO:rn7764CTO:rvrNotAvailable:cvnLENOVO:ct10:cvrNotAvailable:
dmi.product.name: 7764CTO
dmi.product.version: ThinkPad X61 Tablet
dmi.sys.vendor: LENOVO
fglrx: Not loaded
system:
 distro: Ubuntu
 architecture: i686kernel: 2.6.31-3-generic

[Geir Ove Myhr (gomyhr)]

Nothing is added to dmesg output after the crash, but I attach it anyway.

I will try to get a full backtrace. I was first trying to make it
generate a core file that I could debug later, but it didn't work.
Will try some of the other options on
http://wiki.ubuntu.com/X/Backtracing (if the core file method would
have worked I wouldn't have needed to connect the computer to wired
ethernet and borrow my wife's computer to ssh in).

[Geir Ove Myhr (gomyhr)]

It seems that the crash initially reported doesn't happen every time
and is a side issue (I'll update the title and description to remove
the crash). The real issue is that the screen is black with some
random white dots scattered on the top of the monitor (picture
attached). It seems that xorg is running fine without anything
changing on the screen. CapsLock still turns on and off the LED.

I started a gdb session from a remote computer anticipating to do a
full backtrace after the crash. Since the crash didn't happen I just
took some backtraces and went on. I switched VT a few times, and that
resulted in a SIGUSR1 and I had to tell gdb to continue. Attaching the
gdb session log.

Since this seems related to KMS (which a presume is part of DRM) I
turned on DRM debugging by
echo 1 | sudo tee /sys/module/drm/parameters/debug
This results in a lot of extra information in the dmesg output.
Attaching that as well.

Anything else I should do?

[Geir Ove Myhr (gomyhr)]

Here is Xorg.0.log from when it is not crashing, just showing the black screen with corruption.

[Geir Ove Myhr (gomyhr)]

Still present on daily-live from 20090807 with linux 2.6.31-5. Tested with LiveUSB image this time.

[Geir Ove Myhr (gomyhr)]

Still there on the released Karmic. I also tried Lucid LiveUSB image in case something in 2.6.32-rcX would have fixed this, but it is still the same. I don't get crashs anymore. With LiveUSB the gdm startup sound plays, and I'm able to blindly change VT, log in, and reboot with `sudo reboot`.

The VTs look different from VT7 which has the X-session. They have just a little bar in the upper left corner. The corrupted screen does not change when I log in and enter commands.

[Geir Ove Myhr (gomyhr)]

I have previously reported this as an ubuntu bug at https://bugs.launchpad.net/bugs/401214, but this is with fresh logs and updated software.

When adding the line
  Driver "VESA"
to the "Device" section of xorg.conf, xorg starts up with only a black screen with a few lines of scattered white dots on the upper part. It starts normally when I use UMS instead of KMS (nomodeset on boot).

Switching to a VT with Ctrl-Alt-F1 removes the corruption and shows a blank screen. Switching back to VT7 gives the same corruption back. Both xorg and the VTs seem to be fully functional except for the display, I can log in and run commands blindly.

Xorg.0.log is almost identical in the KMS and UMS cases. Only difference that I don't know if is trivial is:

1750c1750
< (II) VESA(0): virtual address = 0x7f9d54b7a000,
---
> (II) VESA(0): virtual address = 0x7fdcc4cd1000,

There is a picture of the corruption at
http://launchpadlibrarian.net/29238382/img_1503.jpg

I have had this problem since the first time I tried to use the VESA driver with KMS (spring/summer 2009). The logs are for drm-intel-next kernel from 2009-12-22, xserver 1.7.3.901 (1.7.4 RC 1), and vesa driver 1:2.2.1-2 on Ubuntu Lucid (10.04) development version.

This is a Lenovo X61 Tablet which has intel 965GM graphics.

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile PM965/GM965/GL960 Memory Controller Hub [8086:2a00] (rev 0c)
 Subsystem: Lenovo Device [17aa:20b3]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller [8086:2a02] (rev 0c)
 Subsystem: Lenovo Device [17aa:20b5]

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32418)
Xorg.0.log with KMS

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32419)
Xorg.0.log with UMS

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32420)
Xorg.0.log with KMS (timestamps removed)

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32421)
Xorg.0.log with UMS (timestamps removed)

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32422)
Output of dmesg with KMS (drm.debug=0x0f)

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32423)
Output of dmesg with UMS (drm.debug=0x0f)

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32424)
Output of intel_reg_dumper with KMS

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=32425)
Output of intel_reg_dumper with UMS

[Jesse Barnes (jbarnes-virtuousgeek)]

Sorry, using the VESA driver on top of KMS isn't supported. They drive the GPU in totally different ways and aren't compatible.

[Geir Ove Myhr (gomyhr)]

According to upstream, the vesa is not expected to work without nomodeset, so I'm closing this bug.