Xorg crashed with SIGSEGV in _swrast_copy_teximage2d [i855GM]

Bug #287485 reported by Eric Drechsel
2
Affects Status Importance Assigned to Milestone
Mesa
Won't Fix
Medium
mesa (Ubuntu)
Invalid
Medium
Unassigned

Bug Description

Binary package hint: xorg

Thought it was https://bugs.launchpad.net/gentoo/+source/mesa/+bug/277709 , but no joy.

To reproduce I start a failsafe console session, run kwin (kwin has compositing enabled, with opengl backend - no crash with xrender backend).

Xorg backtrace (from startx over ssh):

(EE) intel(0): tried to update DSPARB with both planes enabled!
(EE) intel(0): underrun on pipe B!
(EE) intel(0): underrun on pipe B!
(EE) intel(0): underrun on pipe B!

Backtrace:
0: /usr/bin/X11/X(xf86SigHandler+0x79) [0x80c2ff9]
1: [0xb8045400]
2: /usr/lib/dri/i915_dri.so(intelCopyTexImage2D+0x14a) [0xaf67e83a]
3: /usr/lib/dri/i915_dri.so(_mesa_CopyTexImage2D+0x1d4) [0xaf718334]
4: /usr/lib/xorg/modules/extensions//libglx.so [0xb7b351ca]
5: /usr/lib/xorg/modules/extensions//libglx.so [0xb7b5e36f]
6: /usr/lib/xorg/modules/extensions//libglx.so [0xb7b62c0a]
7: /usr/bin/X11/X(Dispatch+0x34f) [0x808c88f]
8: /usr/bin/X11/X(main+0x47d) [0x8071d1d]
9: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7c44685]
10: /usr/bin/X11/X [0x8071101]
Saw signal 11. Server aborting.
xinit: connection to X server lost.

waiting for X server to shut down

gdb backtrace attached.

[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: Matsushita Electric Industrial Co., Ltd. Device [10f7:8338]
00:02.0 VGA compatible controller [0300]: Intel Corporation 82852/855GM Integrated Graphics Device [8086:3582] (rev 02)
     Subsystem: Matsushita Electric Industrial Co., Ltd. Device [10f7:8338]

Revision history for this message
Eric Drechsel (ericdrex) wrote :
Revision history for this message
Chris Coulson (chrisccoulson) wrote :

The crash is happening in mesa, so re-assigning. Could you please try obtaining a backtrace with libgl1-mesa-dri-dbg installed.

Thanks

Changed in xorg:
importance: Undecided → Medium
status: New → Incomplete
Revision history for this message
tschenser (tschenser) wrote :

Eric, it seems that debuggings symbols from the -dbg package are invalid somehow. To get a full symbolic stack trace it seem to be needed to compile mesa on your own. I did this by loading source, diff and dsc of the package, unpack it with dpkg-source -x *.dsc and then apply all the patches under debian/patches. Then you can do ./configure & make & make install as usual with config flags I described in #277709. Good luck.

Revision history for this message
Eric Drechsel (ericdrex) wrote :

it's installed. same difficulty Salze had in https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/277709 ?
i didn't install from deb http://ddebs.ubuntu.com intrepid main universe. Is there any reason to use that instead of the main repo packages?

Revision history for this message
Eric Drechsel (ericdrex) wrote :

ok, built mesa with debug symbols and generated a backtrace. Looks just like yours from 277709. Did I fail to apply the patches to mesa? I did:
apt-get source mesa
dpkg-source -x mesa*.dsc # this command applies all patches via quilt yes?

Revision history for this message
tschenser (tschenser) wrote :

No, orig.source is patched with diff, the result (among other things) is that alle patches are under debian/patches. But they are no applied to source files yet. So you have to do this via patch -p0 < patch.
Your latest gdb-stack is fixed by now but you run into it again, bacause you haven't the patch installed.

Revision history for this message
Eric Drechsel (ericdrex) wrote :

ok, here's a proper backtrace

Revision history for this message
tschenser (tschenser) wrote :

Please attach Xorg.0.log and xorg.conf. Did you enable framebuffer compression? I don't know if its the cause, but I heard this function is buggy.

Looks like http://bugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=495943

Revision history for this message
Chris Coulson (chrisccoulson) wrote :

In addition to what tschenser asked for, please also provide the output of "sudo lspci -vvnn" as an attachment. Once you've provided all this information, I'll send the report upstream.

Thanks

Revision history for this message
Eric Drechsel (ericdrex) wrote :
Revision history for this message
Eric Drechsel (ericdrex) wrote :
Revision history for this message
Eric Drechsel (ericdrex) wrote :
Revision history for this message
Eric Drechsel (ericdrex) wrote :

glxinfo says:
  Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable

Revision history for this message
Chris Coulson (chrisccoulson) wrote :

Thanks for the information. I'll send this upstream shortly (https://bugs.freedesktop.org isn't working at the moment).

Changed in mesa:
status: Incomplete → Triaged
Revision history for this message
In , Chris Coulson (chrisccoulson) wrote :

Created an attachment (id=19830)
Backtrace

This was originally reported at https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/287485 by Eric Drechsel:

"To reproduce I start a failsafe console session, run kwin (kwin has compositing enabled, with opengl backend - no crash with xrender backend).

Xorg backtrace (from startx over ssh):

(EE) intel(0): tried to update DSPARB with both planes enabled!
(EE) intel(0): underrun on pipe B!
(EE) intel(0): underrun on pipe B!
(EE) intel(0): underrun on pipe B!

Backtrace:
0: /usr/bin/X11/X(xf86SigHandler+0x79) [0x80c2ff9]
1: [0xb8045400]
2: /usr/lib/dri/i915_dri.so(intelCopyTexImage2D+0x14a) [0xaf67e83a]
3: /usr/lib/dri/i915_dri.so(_mesa_CopyTexImage2D+0x1d4) [0xaf718334]
4: /usr/lib/xorg/modules/extensions//libglx.so [0xb7b351ca]
5: /usr/lib/xorg/modules/extensions//libglx.so [0xb7b5e36f]
6: /usr/lib/xorg/modules/extensions//libglx.so [0xb7b62c0a]
7: /usr/bin/X11/X(Dispatch+0x34f) [0x808c88f]
8: /usr/bin/X11/X(main+0x47d) [0x8071d1d]
9: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7c44685]
10: /usr/bin/X11/X [0x8071101]
Saw signal 11. Server aborting.
xinit: connection to X server lost.

waiting for X server to shut down"

This is with Mesa 7.2. I will attach the users Xorg.0.log.old, xorg.conf, backtrace and output of lspci -vvnn

Revision history for this message
In , Chris Coulson (chrisccoulson) wrote :

Created an attachment (id=19831)
Output of lspci -vvnn

Revision history for this message
In , Chris Coulson (chrisccoulson) wrote :

Created an attachment (id=19832)
xorg.conf

Revision history for this message
In , Chris Coulson (chrisccoulson) wrote :

Created an attachment (id=19833)
Xorg.0.log.old

Revision history for this message
Chris Coulson (chrisccoulson) wrote :

I've reported it upstream now.

Thanks

Changed in mesa:
importance: Undecided → Unknown
status: New → Unknown
Changed in mesa:
status: Unknown → Confirmed
Revision history for this message
In , Eric Drechsel (ericdrex) wrote :

Hi, I'm the original reporter. It seems that there was a wrong setting in some kwin related config file, deleting the several kwin configs eliminates the crash. If someone is interested in the particular setting that causes the crash, I can find it by experimentation but otherwise this bug can be closed. Thanks!

Revision history for this message
Eric Drechsel (ericdrex) wrote :

I revisited this issue tonight and after an unnecessary excursion into the land of bleeding xorg, discovered that other user accounts on the same machine don't have the problem. deleted all kwin related conf files in ~/.kde/share/config/ fixes and now no crash.

Tried to comment same on upstream bug but can't login to fd.org bugzilla.

Changed in mesa:
status: Confirmed → Invalid
Bryce Harrington (bryce)
description: updated
Revision history for this message
Bryce Harrington (bryce) wrote :

Thanks, closing as a config issue

Changed in mesa (Ubuntu):
status: Triaged → Invalid
Changed in mesa:
importance: Unknown → Medium
status: Invalid → Won't Fix
Changed in mesa:
importance: Medium → Unknown
Changed in mesa:
importance: Unknown → Medium
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.