nvidia-304 fails to install properly causing unity-control-center to crash

[Erick Brunzell]

I get this crash trying to open System Settings in Xenial. The System Settings UI fails to open and apport takes over.

ProblemType: Crash
DistroRelease: Ubuntu 16.04
Package: unity-control-center 15.04.0+16.04.20160119-0ubuntu1
ProcVersionSignature: Ubuntu 4.4.0-2.16-generic 4.4.0
Uname: Linux 4.4.0-2-generic x86_64
ApportVersion: 2.19.4-0ubuntu1
Architecture: amd64
CrashCounter: 1
CurrentDesktop: GNOME-Flashback:Unity
Date: Fri Jan 29 07:27:47 2016
ExecutablePath: /usr/bin/unity-control-center
InstallationDate: Installed on 2015-11-10 (80 days ago)
InstallationMedia: Ubuntu 16.04 LTS "Xenial Xerus" - Alpha amd64 (20151109)
ProcCmdline: unity-control-center --overview
ProcEnviron:
 LANGUAGE=en_US
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SegvAnalysis:
 Segfault happened at: 0x7fc5ad919809: mov 0xd8(%rdi),%rbx
 PC (0x7fc5ad919809) ok
 source "0xd8(%rdi)" (0x000000d8) not located in a known VMA region (needed readable region)!
 destination "%rbx" ok
SegvReason: reading NULL VMA
Signal: 11
SourcePackage: unity-control-center
StacktraceTop:
 ?? () from /usr/lib/x86_64-linux-gnu/libcogl.so.20
 ?? () from /usr/lib/x86_64-linux-gnu/libclutter-1.0.so.0
 ?? () from /usr/lib/x86_64-linux-gnu/libgdk-3.so.0
 ?? () from /usr/lib/x86_64-linux-gnu/libgdk-3.so.0
 gdk_display_get_event () from /usr/lib/x86_64-linux-gnu/libgdk-3.so.0
Title: unity-control-center crashed with SIGSEGV in gdk_display_get_event()
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip lpadmin plugdev sambashare sudo

[Erick Brunzell (lbsolost)]

This only occurs if using the nvidia-304 graphics driver. It does not occur if I use the nouveau driver. The nvidia driver also fails to provide the proper screen resolution.

[Kev Bowring (flocculant)]

does this occur when NOT using CurrentDesktop: GNOME-Flashback:Unity

i.e. does at some point trying to use old hardware and forcing unity on it fail? if that's the case - then wrong DM perhaps. eg not a bug.

[Erick Brunzell (lbsolost)]

I'll try Unity after switching back to the nvidia driver. I've had to use flashback w/metacity with the nouveau driver because the Unity DE freezes within just a minute or two while using the nouveau driver.

[Erick Brunzell (lbsolost)]

The default Unity DE fails to load altogether with the nvidia driver, it just bounces right back to the login screen. Unity does load with the nouveau driver activated but it slows to a crawl and eventually just freezes. But that hardware has never been able to run Unity w/nouveau (llvmpipe / Gallium 0.4 on NV4C).

[Apport retracing service (apport)]

StacktraceTop:
 _cogl_renderer_handle_native_event () from /tmp/apport_sandbox_8UteiF/usr/lib/x86_64-linux-gnu/libcogl.so.20
 cogl_gdk_filter () from /tmp/apport_sandbox_8UteiF/usr/lib/x86_64-linux-gnu/libclutter-1.0.so.0
 gdk_event_apply_filters (xevent=xevent@entry=0x7ffce7c84d00, event=event@entry=0xaf3170, window=window@entry=0x0) at /build/gtk+3.0-vke6KI/gtk+3.0-3.18.6/./gdk/x11/gdkeventsource.c:81
 gdk_event_source_translate_event (xevent=0x7ffce7c84d00, event_source=0x6d4df0) at /build/gtk+3.0-vke6KI/gtk+3.0-3.18.6/./gdk/x11/gdkeventsource.c:195
 _gdk_x11_display_queue_events (display=0x68d020) at /build/gtk+3.0-vke6KI/gtk+3.0-3.18.6/./gdk/x11/gdkeventsource.c:338

[Erick Brunzell (lbsolost)]

I believe this is just a side-effect of the nvidia driver not installing properly with the 4.2, 4.3, or 4.4 series kernels. I found this in JournalErrors:

Jan 29 07:26:56 username-AMD-desktop kernel: nvidia: module license 'NVIDIA' taints kernel.
Jan 29 07:26:56 username-AMD-desktop kernel: Disabling lock debugging due to kernel taint
Jan 29 07:26:56 username-AMD-desktop kernel: nvidia: Unknown symbol mtrr_del (err 0)
Jan 29 07:26:56 username-AMD-desktop kernel: nvidia: Unknown symbol mtrr_add (err 0)

So currently it affects Wily, Xenial, and the Trusty dailies w/Wily kernel. I'm not tech savvy enough to know which bug(s) may be the actual culprit, but I think either bug 1477593 or 1498146. I realize those are for different nvidia driver versions but there are duplicates that point to nvidia-304.

[Erick Brunzell (lbsolost)]

I changed the title to more properly represent the underlying problem. As mentioned above I suspect this is a duplicate of (or related to) either bug 1477593 or bug 1498146. In fact one or the other of those might be a duplicate of the other. Currently the Wily HWE stack seems to be working OK in 14.04.4 but I'll have to verify that after trusty-proposed is switched off and the 14.04.4 release candidates build. I'll then follow up testing the current state of Xenial and Wily ASAP.

[Erick Brunzell (lbsolost)]

This has no longer been an issue since 16.04 final was released so marking it invalid.

[andyczerwonka (andy-czerwonka)]

I am getting this error when attaching an external monitor to my mini displayport. I am running the latest nvidia driver and 16.04. I basically can't use my external monitor unless I run in Power Save Intel mode.

[Erick Brunzell (lbsolost)]

@ andyczerwonka, you need to file a new bug report. Since you mention both nVidia and Intel I assume you have hybrid graphics which presents a whole different situation.