Dear Maintainer,
I'm running Cinnamon 3.0.7-2 everything was fine until a recent update.
Now clicking on certain items on the cinnamon-panel causes a crash and
when running cinnamon --replace from a tty causes a segmentation fault.
This is the output from the command and the backtrace generated by gdb:
------------------------------------------------------------
GNU gdb (Debian 7.11.1-2) 7.11.1
Copyright (C) 2016 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from cinnamon...Reading symbols from /usr/lib/debug/.build-
id/be/08ba4f5f8382c9f3421a2588b412c88503dc03.debug...done.
done.
(gdb) run
Starting program: /usr/bin/cinnamon --replace
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7fffe2c1b700 (LWP 9800)]
[New Thread 0x7fffe241a700 (LWP 9801)]
[New Thread 0x7fffe1908700 (LWP 9802)]
[New Thread 0x7fffe1107700 (LWP 9803)]
[New Thread 0x7fffe0906700 (LWP 9804)]
[New Thread 0x7fffcb0ff700 (LWP 9807)]
Cjs-Message: JS LOG: About to start Cinnamon
(cinnamon:9796): GVFS-RemoteVolumeMonitor-WARNING **: remote volume monitor
with dbus name org.gtk.vfs.GoaVolumeMonitor is not supported
(cinnamon:9796): GVFS-RemoteVolumeMonitor-WARNING **: remote volume monitor
with dbus name org.gtk.vfs.GoaVolumeMonitor is not supported
(cinnamon:9796): GVFS-RemoteVolumeMonitor-WARNING **: remote volume monitor
with dbus name org.gtk.vfs.GoaVolumeMonitor is not supported
St-Message: cogl npot texture sizes SUPPORTED
Cjs-Message: JS LOG: Cinnamon started at Tue Nov 01 2016 12:45:00 GMT-0400
(EDT)
[New Thread 0x7fffc97bf700 (LWP 9809)]
[New Thread 0x7fffc8dee700 (LWP 9810)]
[New Thread 0x7fffbbfff700 (LWP 9811)]
[New Thread 0x7fffbb7fe700 (LWP 9812)]
[New Thread 0x7fffbaffd700 (LWP 9813)]
[New Thread 0x7fffba7fc700 (LWP 9814)]
[New Thread 0x7fffb9ffb700 (LWP 9815)]
[New Thread 0x7fffb97fa700 (LWP 9816)]
openGL version 3.1 detected (GL3 Cogl Driver)
(cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized
20)
(cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized
22)
(cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized
22)
(cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized
22)
(cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized
22)
(cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized
28)
(cinnamon:9796): Cvc-CRITICAL **: gvc_mixer_card_get_index: assertion
'GVC_IS_MIXER_CARD (card)' failed
(cinnamon:9796): Cvc-CRITICAL **: gvc_mixer_card_get_index: assertion
'GVC_IS_MIXER_CARD (card)' failed
(cinnamon:9796): Cvc-CRITICAL **: gvc_mixer_card_get_index: assertion
'GVC_IS_MIXER_CARD (card)' failed
(cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized
20)
Thread 1 "cinnamon" received signal SIGSEGV, Segmentation fault.
0x00007fffeb003010 in ?? () from /usr/lib/x86_64-linux-gnu/libnvidia-
glcore.so.367.44
(gdb) bt
#0 0x00007fffeb003010 in () at /usr/lib/x86_64-linux-gnu/libnvidia-
glcore.so.367.44
#1 0x00007fffeaf1a311 in () at /usr/lib/x86_64-linux-gnu/libnvidia-
glcore.so.367.44
#2 0x00007fffeaf24136 in () at /usr/lib/x86_64-linux-gnu/libnvidia-
glcore.so.367.44
#3 0x00007fffeab8c799 in () at /usr/lib/x86_64-linux-gnu/libnvidia-
glcore.so.367.44
#4 0x00007ffff78c1ed7 in meta_sync_ring_insert_wait (self=0x555558915370) at
compositor/meta-sync-ring.c:285
#5 0x00007ffff78c1ed7 in meta_sync_ring_insert_wait () at compositor/meta-
sync-ring.c:593
#6 0x00007ffff78bac35 in meta_pre_paint_func (data=0x555555ec26d0) at
compositor/compositor.c:1394
#7 0x00007ffff7169574 in _clutter_run_repaint_functions
(flags=flags@entry=CLUTTER_REPAINT_FLAGS_PRE_PAINT) at clutter-main.c:3449
#8 0x00007ffff716a317 in clutter_clock_dispatch (master_clock=0x555556936040
[ClutterMasterClockDefault], stages=0x555557e8c060)
at clutter-master-clock-default.c:437
#9 0x00007ffff716a317 in clutter_clock_dispatch (source=<optimized out>,
callback=<optimized out>, user_data=<optimized out>)
at clutter-master-clock-default.c:567
#10 0x00007ffff64dc7d7 in g_main_context_dispatch (context=0x5555557bff40) at
././glib/gmain.c:3203
#11 0x00007ffff64dc7d7 in g_main_context_dispatch
(context=context@entry=0x5555557bff40) at ././glib/gmain.c:3856
#12 0x00007ffff64dca40 in g_main_context_iterate (context=0x5555557bff40,
block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>)
at ././glib/gmain.c:3929
#13 0x00007ffff64dcd62 in g_main_loop_run (loop=0x5555557cb400) at
././glib/gmain.c:4125
#14 0x00007ffff78dcc3c in meta_run () at core/main.c:542
#15 0x0000555555555830 in main (argc=<optimized out>, argv=<optimized out>) at
main.c:329
(gdb) quit
A debugging session is active.
Inferior 1 [process 9796] will be killed.
Quit anyway? (y or n) y
------------------------------------------------------------
¡Hola William! El 2016-11-01 a las 12:56 -0400, William L. DeRieux IV escribió: Can you try to narrow down the list of upgraded packaged that caused this? I haven't seen these warnings before. I guess that something broke within the nvidia driver. Also the backtrace points to the nvidia glcore libs. Mmmh, this is the kernel from stable running in stretch, that's not really expected, you might want to switch to the kernel in stretch. Anyway, assuming that the problem is in the nvidia driver, if the driver was updated, the kernel module needs to be reloaded, which means that you need to reboot the machine. If you haven't done so, please check if the problem solves by itself after a reboot. If that's not good, there is newer nvidia driver in sid, you might might want to try it. Also, it might be a good idea to test downgrading the nvidia drivers to the last known to work version, you can use http://snapshots.debian.org to search for this. We might need to reassign this issue to the nvidia developers depending on your tests results. Happy hacking,
On Wed, 2 Nov 2016 11:42:54 +0100 Maximiliano Curia wrote: > Control: tag -1 + moreinfo > > ¡Hola William! > > El 2016-11-01 a las 12:56 -0400, William L. DeRieux IV escribió: > > Package: cinnamon > > Version: 3.0.7-2 > > Severity: important > > > I'm running Cinnamon 3.0.7-2 everything was fine until a recent update. > > Now clicking on certain items on the cinnamon-panel causes a crash and > > when running cinnamon --replace from a tty causes a segmentation fault. > > Can you try to narrow down the list of upgraded packaged that caused this? > > > This is the output from the command and the backtrace generated by gdb: > > > (cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized > > 20) > > > > (cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized > > 22) > > > > (cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized > > 22) > > > > (cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized > > 22) > > > > (cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized > > 22) > > > > (cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized > > 28) > > > > (cinnamon:9796): Cvc-CRITICAL **: gvc_mixer_card_get_index: assertion > > 'GVC_IS_MIXER_CARD (card)' failed > > > > (cinnamon:9796): Cvc-CRITICAL **: gvc_mixer_card_get_index: assertion > > 'GVC_IS_MIXER_CARD (card)' failed > > > > (cinnamon:9796): Cvc-CRITICAL **: gvc_mixer_card_get_index: assertion > > 'GVC_IS_MIXER_CARD (card)' failed > > > > (cinnamon:9796): St-WARNING **: Failed to allocate offscreen for texture (sized > > 20) > > I haven't seen these warnings before. I guess that something broke within the > nvidia driver > > > Thread 1 "cinnamon" received signal SIGSEGV, Segmentation fault. > > 0x00007fffeb003010 in ?? () from /usr/lib/x86_64-linux-gnu/libnvidia- > > glcore.so.367.44 > > (gdb) bt > > #0 0x00007fffeb003010 in () at /usr/lib/x86_64-linux-gnu/libnvidia- > > glcore.so.367.44 > > #1 0x00007fffeaf1a311 in () at /usr/lib/x86_64-linux-gnu/libnvidia- > > glcore.so.367.44 > > #2 0x00007fffeaf24136 in () at /usr/lib/x86_64-linux-gnu/libnvidia- > > glcore.so.367.44 I tried using kernel 4.7 instead of 3.16, which made no difference. I also tried several version(s) of the nvidia-driver 367.44-3 (stretch), 367.57-1 (sid), and 370.28-1 (experimental). In all cases gdb generated the same backtrace. I don't think this is an issue with the nvidia-driver, as a lot more should have been broken by it.
¡Hola William! El 2016-11-02 a las 17:32 -0400, William L. DeRieux IV escribió: The list of upgraded packages might be useful. You can obtain this information from the /var/log/dpkg.log* files. Using snapshots.debian.org, you could downgrade the cinnamon packages and test this again to check if any of the recent changes in cinnamon causes the issue. Happy hacking,
On Thu, 3 Nov 2016 11:06:23 +0100 Maximiliano Curia wrote: > ¡Hola William! > > El 2016-11-02 a las 17:32 -0400, William L. DeRieux IV escribió: > >>> I'm running Cinnamon 3.0.7-2 everything was fine until a recent update. > >>> Now clicking on certain items on the cinnamon-panel causes a crash and > >>> when running cinnamon --replace from a tty causes a segmentation fault. > > >> Can you try to narrow down the list of upgraded packaged that caused this? > > The list of upgraded packages might be useful. You can obtain this information > from the /var/log/dpkg.log* files. > > Using snapshots.debian.org, you could downgrade the cinnamon packages and test > this again to check if any of the recent changes in cinnamon causes the issue. > > Happy hacking, > -- > "Inside every large problem is a small problem struggling to get out." > -- Hoare's Law of Large Problems > Saludos /\/\ /\ >< `/ Ok, here is an update to this issue. The initial issue that I had, that prompted this, was that the cinnamon-panel had become non-responsive and I could not launch any apps, etc. I also rebooted to ensure that is wasn't something that would have been fixed by a reboot. Adfter rebooting -- the cinnamon panel was still non-responsive -- to a tty and ran *cinnamon --replace* which crashed (as reported). I switched to the mate desktop to get around the issue -- at least for the time being. The other day I switched back to cinnamon to re-test this and the cinnamon-panel was functioning correctly. Oddly, enough running *cinnamon --replace* from a tty still crashes, but I was able to run it, from a terminal, within cinnamon without issues. PS: I can't remember what date that I first experienced this -- but it was a few days ago -- I switched to MATE to get around it temporarily and I reported the issue a few days later. I guess the point is that cinnamon seems to be functioning correctly -- even though *cinnamon --replace* can't seem to executed from a tty (but does not fail if run from within cinnamon). I'm not sure if this may now be expected behavior...
Hi William, hi everyone, I'm closing this bug as obsolete and not reproducible. The reported crash was entirely inside libnvidia-glcore.so.367.44, reached from muffin's GLX_EXT_x11_sync_object path, with no frame in Debian-built code. In your last message you already reported that the original problem had gone away, and that "cinnamon --replace" only failed when started from another tty, where the VT owning the X session is not active and the NVIDIA GL context can fail to allocate buffers and fences; restarting the window manager is meant to be done from within the running session. The information requested at the time (upgraded packages from /var/log/dpkg.log*, or a downgrade test through snapshot.debian.org) never arrived, so the regression was never identified. Almost ten years later none of this can be acted upon: cinnamon 3.0.7 is long gone (unstable ships 6.6.9), the nvidia-graphics-drivers 367.x series is no longer in the archive, and the crashing code has been rewritten — current muffin disables the sync ring and falls back to XSync() when it fails, instead of keeping to use a broken sync object. If anyone still sees a comparable crash with Cinnamon 6.6.x, please open a new report with a fresh backtrace and the details of the graphics driver in use. Thanks for the original report, and sorry for the very long silence.