Originally Posted by ledoc
This could be the time-out of gdm(/kdm/xdm). It attempts to restart X if the server didn't come up within N seconds. Look for some gdmrc config and increase that time-out.
That indeed solves X trying multiple times. It now starts on it's first try! Thanks!
Originally Posted by danix
An nvidia-bug-report.log would be useful. You can generate this by running "nvidia-bug-report.sh" as root.
If you could create one for the failing 195.36.XX driver, and another one for the 190.53 driver, that would be helpful as well.
I attached a report log for 195.36.24, I'd have to downgrade xorg to 1.7 to create one for 190.53, since this one doesn't support xorg 1.8
I've read this: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=563624
Is there something that slows down the loading of the GLX module in the 195.36.xx versions of the Linux nvidia driver? It was pretty fast in 190.50.xx.