Xorg crashed with SIGSEGV in BlockHandler() - Regression since 290.10

[Roger]

Happened on a fully up-to-date

Precise. I've had all sorts of problems with session crashes since nvidia-current moved to 295.20 - first on Oneiric and now on Precise. More often than not, Apport wants to report a bug on some other program that segfaults. But, looking at the logs, these segfaults always follow on segfaults in/problems with nvidia/Xorg. So I suspect nvidia-current version 295.20 may be the root cause, even though I'm using Unity and not Gnome Shell, with its well documented troubles with Nvidia.

ProblemType: Crash
DistroRelease: Ubuntu 12.04
Package: xserver-xorg-core 2:1.11.4-0ubuntu6
ProcVersionSignature: Ubuntu 3.2.0-18.29-generic 3.2.9
Uname: Linux 3.2.0-18-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 1.94.1-0ubuntu2
Architecture: amd64
Date: Thu Mar 15 09:06:47 2012
DistroCodename: precise
DistroVariant: ubuntu
ExecutablePath: /usr/bin/Xorg
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Alpha amd64 (20120314)
ProcCmdline: /usr/bin/X :0 -auth /var/run/lightdm/root/:0 -nolisten tcp vt7 -novtswitch
ProcEnviron:

SegvAnalysis:
 Segfault happened at: 0x7fd26466b619: mov 0x68(%rax),%rdx
 PC (0x7fd26466b619) ok
 source "0x68(%rax)" (0x00000067) not located in a known VMA region (needed readable region)!
 destination "%rdx" ok
SegvReason: reading NULL VMA
Signal: 11
SourcePackage: xorg-server
StacktraceTop:
 ?? () from /usr/lib/x86_64-linux-gnu/xorg/extra-modules/nvidia_drv.so
 BlockHandler ()
 WaitForSomething ()
 ?? ()
 ?? ()
Title: Xorg crashed with SIGSEGV in BlockHandler()
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups:

[Apport retracing service (apport)]

Stacktrace:
 #0 0x00007fd26466b619 in ?? ()
 No symbol table info available.
 #1 0x0000000000000000 in ?? ()
 No symbol table info available.
StacktraceTop:
 ?? ()
 ?? ()

[Roger (r-wiberg)]

I can now report that the 290.10 drivers from Nvidia work flawlessly for me in Precise. I'm using an Nvidiea 9800 GT card and an Asus P6T DELUXE V2 motherbord on a desktop computer. No more crashes.

If the 295.20 drivers don't work with search in gnome-shell, regardless of system it seems, and cause session crashes in unity, at least on some systems, maybe it would be a good idea to provide the tried and true 290.10 drivers in Precise, at least as an option?

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Bryce Harrington (bryce)]

Roger,

There is a new 295.33 driver in Ubuntu now, which is said to resolve a number of crash issues. I can't say for sure from release notes whether it resolves your crash bug, but I think it would be worth re-testing. Update to the latest software, and see whether or not the crash can be reproduced.

[Roger (r-wiberg)]

Right, I uninstalled the 290.10 kernel module and installed nvidia-current 295.33 from the repos. Seems to be an improvement but after only a minute or so I got

Apr 9 09:17:18 blenheim gnome-session[2417]: Gdk-WARNING: gnome-session: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.#012

when scrolling in a LibreOffice document. It's also happend when using VirtualBox (both 4.1.10 and 4.1.12). But Nvidia's 290.10 version continues to work flawlessly. I'll attach a copy of the nvidia-bug-report as well.

[Roger (r-wiberg)]

And here's the attachment...