Nvidia module udev issue with 331.49 driver?

I’m not even sure if this makes sense or if it’s just a coincidence relating to something else, but for about the last couple of weeks my system has been hanging at boot while udev is syncing devices.

I use Slackware 14.1, I have a Zotac GTS 450 Amp Edition gpu. While at boot udevadm has randomly been hanging the last couple of weeks, like probably for a third of my boots and I have to do a hard power off and restart until it boots without hanging. I turned on info/debug mode for udev for more verbose info and udev was hanging while it was syncing devices, that was the most info I was able to get from my logs. The weird thing is, the problem seemed to coincide around the time I installed the 331.49 driver. Not sure if it was a coincidence or what, but I rolled back to 331.38 driver and it hasn’t happened again for the last couple of days.

Could this even possibly be related? Thanks.
nvidia-bug-report.log.gz (85.3 KB)

Please run nvidia-bug-report.sh and attach the resulting nvidia-bug-report.log.gz file.

Ok I added it. Same problem was happening with 331.38 so I went back to 331.49. If it’s not related to the module, I’m pretty sure a hardware issue related to the graphics card itself.

Was this log generated after you had rebooted? There’s a kernel bug that triggers a BUG() call occasionally when the NVIDIA module is being loaded at boot. If you check your system log, do you see messages along the lines of, “kernel BUG at drivers/cpufreq/cpufreq.c:1456!”?

Yep I have to reboot, udev will timeout for about 120 seconds but my hardware will never become active if I wait for it. I have never been able to find anything in my system logs that indicates an error. Even with udev logging on, the only relevant message is “message (SYNC) received” and that’s where it hangs into a timeout.

Also I only started getting udev sync problems when I started using the 331 drivers. The first month and a half I was using Slackware 14.1 this never happened once and I haven’t added any other hardware or new kernel modules at boot. Again it could be a coincidence and related to something else but the timing of it makes me suspect either the nvidia module or my graphics card.

Ok I think it’s definitely the 331 nvidia module that is causing it. If I remove the .conf out of /etc/X11/xorg.conf.d, I could reboot 10 times in a row without issues. What I don’t get is there’s an nvidia module that still loads itself regardless of what I have set in xorg.conf.d. X won’t start if I remove the .conf file, but I still see an nvidia module in my module list. Can someone explain?

See if this gets a match - it might be set to be auto-loaded:

find /etc/mod* | grep nvidia

No. I have a blacklist for nouveau in /etc/modprobe.d, that’s it. I’ve only used Nvidia’s own binary installers, so all its behavior is on its own doing.