kernel panic when using Bumblebee

Hello,

I’m having an issue where the nvidia kernel module seems to sometimes crash when I run a program using optirun. It never happens just after boot, it only happens when my computer has been on for a few days. When the crash occurs, I’m given a kernel panic screen (see below).

After the crash, my desktop works fine (I switch to it using ctrl-alt-F7) under my intel CPU. However, I can’t run optimus programs anymore (until I reboot):

$ optirun -v --debug glxgears
[63621.164720] [INFO]Configured driver: nvidia
… nothing happens

dmesg - output of the kernel panic:
http://pastebin.com/6KViu1dt

Xorg.8.log - the xorg log for bumblebee (this is actually from running bumblebee again after the crash):
http://pastebin.com/jqqcg9ep

bumblebee.conf - Bumblebee config file:
http://pastebin.com/c1KDgJDH

xorg.conf.nvidia - Xorg file for nvidia:
http://pastebin.com/Kw6LAu5n

This issue was reported to Bumblebee a few months ago, and they concluded that it was a nvidia driver issue:

My laptop:
Sager NP9170 (Clevo P170EM)
Intel i7-3720QM
GeForce GTX 670M 1.5GB
linux distro: Arch Linux
desktop: KDE
Linux 3.8.8 (64 bit)
xorg server 1.14.1
nvidia driver 313.30
bumblebee 3.2
primus 20130425 (2013-04-25)

I have had this issue ever since I got this laptop nearly a year ago. However, I wasn’t using my nvidia card much until recently. Any ideas?

thanks for the help!
Kevin

To clarify, this actually looks like a different issue: you have a different exception (division by zero (?), not a null pointer dereference), and in that bug the problem was always reproducible while you say you only see this after a few days’ uptime (do you keep your laptop on, or is that after a few suspend-resume cycles?).

In both cases, Bumblebee developers can’t do anything about it, even if they wanted to, because the exception is in closed-source code, and they don’t have the affected hardware, and in any case I think module faults like those should be looked at by nVidia.

Kevin, you probably should provide nvidia-bug-report archive produced after the failure.

Ok. I’ll make a nvidia-bug-report archive next time.

The crash occurs after a few suspend-resume cycles.