wine-git/google earth crashes with DRM_I830_CMDBUFFER: -22 on Thinkpad X60

I've posted a .deb of the latest intel driver git HEAD here if you would like to test it:
http://people.ubuntu.com/~bryce/Testing/intel/xserver-xorg-video-intel_2.1.1~git20071004-0ubuntu1_i386.deb

Since it's straight out of git, it may include new issues. Direct rendering often stops working, for instance. However it would be interesting to hear if it addresses this issue or not.

Bryce, that .deb you gave me solves the problem. Thanks!

... but it's unstable. I can crash X by doing
  svn checkout http://win16test.googlecode.com/svn/trunk/ win16test
  cd win16test/src
  sh install_openwatcom
OpenWatcom starts installing, probably gets most of the way
through, but then the X server crashes, and I'm logged out.
This wasn't happening with the old xserver package.

Sadly, the problem reappeared when I updated to the official
beta version (2.1.1-0ubuntu6). I just checked the Debian bug
report again, and I see the fix was a change to the Mesa
package. Gutsy's Mesa package does seem to have the fix:

+ * Build with -fno-strict-aliasing to fix misbuild of i965_dri.so
+ (closes: #394311). Thanks to Michel Dänzer for the fix, and to Ryan
+ Richter for the report and testing.

cf. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=394311

but maybe we missed something. The package you linked
to directly -- was it compiled differently, perhaps in a way that
turned off the -fstrict-aliasing optimization?

Hi Dan,

The package I built was compiled on my own hardware but should be following all the same rules as the official package. I don't have the log so can't verify whether -fstrict-aliasing was turned on or off, however I did not intentionally change that.

Btw, I'm hoping to put out an update of -intel before too long, once a sufficient number of critical fixes are gathered and tested. If we can get this one narrowed down, I'd like to shoot for having it included, so if you make progress with it let me know.

Not sure what I can do to help. I suppose if you
gave me a shell script that grabbed the source
package and built the binary, I could try fiddling
with compiler options...

I am experiencing the same problem with the same video hardware reported by dankegel. For me, to reproduce:

- Start Blender from the command line
- Switch workspace by pressing ctrl-alt-right arrow.
- Switch workspace back

Once I've switched back to the orig. workspace in which I started Blender, Blender has crashed and in the terminal session I have the message:
DRM_I830_CMDBUFFER: -22

Bryce: I've tried your replacement xserver-xorg-video-intel package that you linked to above and which was compiled from the git head as it was a few weeks ago. I also tried another more recent package of yours at http://people.ubuntu.com/~bryce/Testing/intel-Gutsy/xserver-xorg-video-intel_2.1.1-0ubuntu10~bwh1_i386.deb. On both occasions after reinstalling I restarted the local X server. However, neither of the replacement packages resolved this issue.

Let me know if you would like me to test new packages, etc.

Of course, the above mentioned problem goes away in Gutsy if I simply turn off desktop visual effects.

I just tried Maya PLE 8.5 in Wine, and gosh darn it, that
too aborts with
DRM_I830_CMDBUFFER: -22

My Gutsy install with Wine 0.9.49 and Google SketchUp 6 crashes on startup with this same error. Visual effects on or off.

On Nov 13, 2007 1:37 PM, micah geisel <email address hidden> wrote:
> My Gutsy install with Wine 0.9.49 and Google SketchUp 6 crashes on
> startup with this same error. Visual effects on or off.

Yes, in my tests, it doesn't matter if visual effects are on or off.

That and the fact that I get no sound
make Gutsy kind of unusable for some things.

I have the exact same problem using SketchUp 6, wine0.9.49 and Feisty FujitsuP1610.

Would be happy to send more info or test any proposed solutions...

Update/Additional Info:
P1610 has same graphics hardware as Thinkpad X60 mentioned above:
00:02.1 Display controller: Intel Corporation Mobile 945GM/GMS/940GML Express Integrated Graphics Controller (rev 03)

I also performed a dist upgrade to gusty, but still same problem. Googling around, the only info I could find on this error was about making sure to complie mesa with -fno-strict-aliasing option. I recompiled and installed libgl1-mesa-dri with this compile option, but still does not fix the problem.

Also, I have another desktop machine that has an Intel865 graphics adapter with Gusty installed on it - same problem.

I suspect that the newer -intel driver uploaded for Hardy may resolve this issue - can you guys test Hardy alpha?

Yep, the problem I described above (Blender crashing with DRM_I830_CMDBUFFER: -22 message) seems to be resolved with Hardy Alpha. Thanks very much. However, unlike everyone else my problem didn't involve Wine so would be good if some others could test their issues with Hardy.

Wine, and glxgears, now crash on exit with hardy beta.

$ gdb glxgears
gdb) run
--- run and click the go-away box in the upper right corner of the glxgears window ---
XIO: fatal IO error 104 (Connection reset by peer) on X server ":0.0"
      after 2974 requests (243 known processed) with 0 events remaining.

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb7c10af0 (LWP 23806)]
0xb7f3ebad in ?? () from /lib/ld-linux.so.2
(gdb) bt
#0 0xb7f3ebad in ?? () from /lib/ld-linux.so.2
#1 0xb7f2aaac in ?? ()
#2 0xb7f46ff4 in ?? () from /lib/ld-linux.so.2
#3 0xbf9b8a30 in ?? ()
#4 0x08053280 in ?? ()
#5 0xbf9b8a64 in ?? ()
#6 0xb7f35059 in ?? () from /lib/ld-linux.so.2
#7 0xb7c19cb4 in ?? () from /lib/tls/i686/cmov/libdl.so.2
#8 0x00000000 in ?? ()

See also bug 182262.

Hi all:
    I updated mesa from 7.0.2 to 7.0.3 and the problem is solved. Hope this helps.

These websites may help:
http://www.mesa3d.org/
http://dri.freedesktop.org/wiki/Building#head-b3fb665c9f24b4a32424b78428c082e830832250

Looks like this has been fixed some time ago. If you can reproduce it with 9.04beta or newer, please reopen.