Bug#522054: Lockups in intel xorg driver
Hi,
I've also seen lockups in the intel DRI driver since about 1-2 weeks.
The responsible upgrade might have been some time before that since I
(didn't use to) reboot daily but use suspend.
I first thought it were 2.6.29 to blame, but I've now had comparable
lockups with 2.6.26, so that is ruled out.
Today, the lockup occured when switching to a tab in epiphany that had
an embedded gnome-mplayer (i.e. xvideo). I believe I can relate other
crashes to xv as well. Today it was just a plain lockup, with an
completely unusable X server, not doing any screen refresh (I believe
sometimes the mouse still worked, not sure about today).
I then used SysRq+K to kill the processed on the current terminal, which
caused GDM to restart the X server to recover. In other situations I've
had to reboot, and I can also remember the keyboard LEDs flashing (which
indicates a kernel panic).
Anyway, my Xorg.0.old log file shows:
exaCopyDirty: Pending damage region empty!
Backtrace:
0: /usr/bin/X(xorg_backtrace+0x3b) [0x81323fb]
1: /usr/bin/X(xf86SigHandler+0x51) [0x80c95f1]
2: [0xb7f9f400]
3: /usr/lib/libdrm_intel.so.1 [0xb78d4901]
4: /usr/lib/libdrm_intel.so.1 [0xb78d49e2]
5: /usr/lib/libdrm_intel.so.1(drm_intel_bo_unreference+0x16)
[0xb78d33e6]
6: /usr/lib/xorg/modules/drivers//intel_drv.so [0xb79320a7]
7: /usr/bin/X [0x80dd1d8]
8: /usr/lib/xorg/modules/extensions//libextmod.so(XvdiPutImage+0x190)
[0xb7a1d420]
9: /usr/lib/xorg/modules/extensions//libextmod.so [0xb7a20faa]
10: /usr/lib/xorg/modules/extensions//libextmod.so(ProcXvDispatch+0x3a)
[0xb7a2242a]
11: /usr/bin/X(Dispatch+0x33f) [0x808c7ff]
12: /usr/bin/X(main+0x3bd) [0x80719dd]
13: /lib/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7c5b775]
14: /usr/bin/X [0x8070e91]
Fatal server error:
Caught signal 11. Server aborting
[... Unloading modules ...]
(II) AIGLX: Suspending AIGLX clients for VT switch
The last line sounds like a bad idea, since one of the unloading lines
before read (II) Video Bus: Close
I will install libdrm-intel1-dbg now, so maybe I'll be able to provide a
more detailed backtrace sometime. Of course I can't tell whether the
segfault in above logfile was before I had pressed SysRq+K or triggered
by that. However I had the impression that Xv triggered the crash, and
indeed the backtrace contains ProcXvDispatch and XvdiPutImage
Note that I have some packages from experimental installed, so my system
is not a plain "sid".
best regards,
Erich Schubert
--
erich@(vitavonni.de|debian.org) -- GPG Key ID: 4B3A135C (o_
You know we all became mathematicians for the same reason: //\
we were lazy. --- Max Rosenlicht V_/_
Die Mathematik muss man schon deswegen studieren, weil sie
die Gedanken ordnet. --- Michail Wassiljewitsch Lomonossow
Reply to: