Latest NVidia Driver Issue
Dave
dcbdbis at comcast.net
Sun Oct 12 03:06:50 UTC 2014
FreeBSD 10.0 Release. AMD64
Source and ports up to date as of 11-Oct-2014
The ports tree recently updated the NVIdia driver to 331.67.
Since then when I attempt to boot my system with the
"nvidia_load="YES", the loader traps and dies, and puts a CPU register
dump on the screen.at the line "nvidia_load="YES", and dumps in screen a
CPU register state then dies.
Things used to work peechee until the update....Now I have to rem the
line in the boot/loader.conf file....and manually kldload the module
after boot up. Everything works peechee from that moment on.
Upon manually kldloading the moduke...no errors.
Linux is built into my custom kernel.
All things were built with the default compile time flags No exceptions
taken.
I don't even know where to begin looking for this problem. And web
crawling doesn't return any search results.
As my system is backed up...I hard-wiped my install and installed 10.1
RC-1....Issue shows up there to with all at defaults.
So I re-hard-wiped my system and did a clean install of 10.0
release.....Issue remains regardless if I install the NVidia modules
from the binary repos....or build it from ports. And the issue remains
regardless if I use the default shipping kernel and load linux in
loader.conf......or use my custom kernel where linux is built in.
Then I tried a different system.....same results. It seems to be
pointing to a faulty nvidia-driver port and build.
Does anyone have any idea what may be going on?
Should I download the latest driver directly from NVidia and install it?
Remember, this is a system which has been working flawlessly...then the
nvidia update, and the broken loader where it displays a CPU register
dump then crashes.
Sincerely and respectfully,
Dave
More information about the freebsd-questions
mailing list