[SOLVED] Cannot get past login screen;login loop - nvidia?
Posted: Wed Apr 10, 2013 4:24 am
My system has been running great for several months now but the other day I installed the latest CUDA SDK and Samples but unfortunately I selected 'yes' when it asked if I wanted the video card driver updated also. At the end of the install there was a warning about a version mismatch between the NVIDIA driver and API but it wasn't clear if it installed in spite of this and everything appeared to be running correctly. The computer ran fine for 2 days afterwards (including several reboots).
Today when I tried to login the desktop showed briefly (ie my wallpaper and the menu bar, which appeared as the default rather than my usual theme), then the screen went black, then the shell login (ie console) showed briefly, and then I ended up back at the login screen.
The resolution of the screen is the same as always so at first I didn't suspect the video card but given that's the only thing that's been changed recently it was the natural suspect. After booting into recovery mode and typing inxi -Fxz into the console I get the following regarding the video card (my typing; no c&p available obv):
Graphics: Card: Failed to Detect Video Card! X-Vendor: N/A drivers: (unloaded: fbdev, vesa, nouveau) FAILED: nvidia tty size: 80x25 Advanced Data: N/A for root out of X
There is nothing else interesting in the output but the culprit is pretty clear. So far the closest topic I have found is: http://forums.linuxmint.com/viewtopic.php?f=46&t=125427 but there's not really any help there.
So the question is: What do I do now? Probably I just need to reinstall the proprietary nvidia drivers to overwrite whatever damage I did with the CUDA package and maybe update X, but how do I do those things?
Video card is an nvidia GT 520 and I'm running Mint 14 Nadia Cinnamon 64bit and I had been successfully using the proprietary drives from nvidia-current package before this.
Thanks for the help!
UPDATE 1:
Should also mention: When I add the nomodeset xforcevesa boot parameters and reboot the loop is interrupted at the console login (ie just before returning me to the login screen) and I get something like
UPDATE 2:
I purged and reinstalled the nvidia drivers with
and restarted. Now X (X.Org X Server 1.13.0) fails and I get to see the failure message where it once again tells me that the NVIDIA kernel module has version 304.54 but this NVIDIA driver component has version 304.43. Seems like I'm getting somewhere, now I need to figure out how to roll back the one or upgrade the other.
When I check the /var/log/Xorg.0.log output the fatal error I get is 'no screens found' and scrolling through the text I see that it claims 'Screen(s) found, but none have a usable configuration'. All little further up I see a line relating to Update 1, it reads 'open /dev/fb0: No such file or directory', there are similar errors relating to /dev/dri/card0. Further up are some warning about modules not existing.
Probably it would be useful to just post the whole log file, but I'm not sure how to get it off that machine and to my laptop so I can post it here, so if you want to see the log file please also tell me how I can get it to you.
Today when I tried to login the desktop showed briefly (ie my wallpaper and the menu bar, which appeared as the default rather than my usual theme), then the screen went black, then the shell login (ie console) showed briefly, and then I ended up back at the login screen.
The resolution of the screen is the same as always so at first I didn't suspect the video card but given that's the only thing that's been changed recently it was the natural suspect. After booting into recovery mode and typing inxi -Fxz into the console I get the following regarding the video card (my typing; no c&p available obv):
Graphics: Card: Failed to Detect Video Card! X-Vendor: N/A drivers: (unloaded: fbdev, vesa, nouveau) FAILED: nvidia tty size: 80x25 Advanced Data: N/A for root out of X
There is nothing else interesting in the output but the culprit is pretty clear. So far the closest topic I have found is: http://forums.linuxmint.com/viewtopic.php?f=46&t=125427 but there's not really any help there.
So the question is: What do I do now? Probably I just need to reinstall the proprietary nvidia drivers to overwrite whatever damage I did with the CUDA package and maybe update X, but how do I do those things?
Video card is an nvidia GT 520 and I'm running Mint 14 Nadia Cinnamon 64bit and I had been successfully using the proprietary drives from nvidia-current package before this.
Thanks for the help!
UPDATE 1:
Should also mention: When I add the nomodeset xforcevesa boot parameters and reboot the loop is interrupted at the console login (ie just before returning me to the login screen) and I get something like
Code: Select all
Buffer I/O error on device fd0, logical block 0
end_request: I/O error, dev fd0, sector 0
I purged and reinstalled the nvidia drivers with
Code: Select all
sudo apt-get purge nvidia-current
sudo apt-get install nvidia-current
When I check the /var/log/Xorg.0.log output the fatal error I get is 'no screens found' and scrolling through the text I see that it claims 'Screen(s) found, but none have a usable configuration'. All little further up I see a line relating to Update 1, it reads 'open /dev/fb0: No such file or directory', there are similar errors relating to /dev/dri/card0. Further up are some warning about modules not existing.
Probably it would be useful to just post the whole log file, but I'm not sure how to get it off that machine and to my laptop so I can post it here, so if you want to see the log file please also tell me how I can get it to you.