g'day

ive used the nvidia proprietry drivers from nvidia.com on FC4 without a hitch before but i just couldnt seem to get it running on FC5. eventually i had to resort to installing the nvidia drivers from livna.

yesterday i realized a handicap of these drivers. they dont support proper GLX and as a result i cannot play opengl games like quake4.i tried installing nvidias proper drivers but they refused to load as te X-server had no clue what or where GLX was. i even uninstalled the livna drivers and modules and then tried installing the nvidia.com ones but to no effect. eventually i had to return back to the livna ones and here i am unable to play any games now. i really have to get them up and running as i have to do testing for a opengl game for the WineHQ database.

any help or insight would be appreciated.

Have you read this? It doesn't tell you how to use the proprietary drivers, but it explains why they're not recommended.

mate, it doesnt matter if Red Hat says they are risky. until they themselves can provide drivers that can allow a person to make proper use of the hardware who am i to argue with Nvidia if they release stuff that aint upto red hat's standards. either way the same package did work for me while i was using FC4 so its my interpration that Red Hat might have screwed up here with FC5.

Your problem is probably caused by the nvidia.com installer, which replaces or conflicts with stuff in Fedora Core. I don't know how to fix this. Livna's packages work well, but your problem is that other stuff was messed up by nvidia.com's installer.

Hy there...
I have the same problems with nvidia and i found some solutions on this link:
http://www.softwarecity.org/forum/vi...php?pid=22#p22

Gentoo makes using nVidia software very easy. It does it right by placing the libraries in a certain directory and using a utility to switch from X11 libraries to nVidia libraries. This is alright to do but a game like UT2004 did not like this setup. I had to do LD_PRELOAD="/usr/lib/opengl/nvidia/lib/libGL.so.1" ut2004 in order to load up UT2004. To make this easy, I used an alias like ut2004='LD_PRELOAD="/usr/lib/opengl/nvidia/lib/libGL.so.1" ut2004' to run UT2004.

I suggest running the nvidia installer but force to only compile the modules. Then on the next step, you have to run the installer again but tell it to only decompress it self, so you can copy the libraries to a certain directory that you created. Third move the X11 libraries to a certain directory to get ready to do some linking. Finally link the nvidia libraries to the directory that the X11 libraries were in before you moved them.

A helpful link to keep both DRI (Mesa3D) and nVidia GLX.
http://www.sabi.co.uk/Notes/linuxNVIDIA.html

well, it seems the problem was due to the Livna installer after all. i installed the nvidia.com drivers and after doing startx booted into gnome where i was able to play Quake4 and run glxgears. this worked previously as well except for one small thing. the Livna installation includes a script that modifies the xorg.conf file upon bootup depending on if it can locate the livna nvidia drivers.

last time when i installed the nvidia.com drivers and rebooted this script messed up my xorg.conf file and eventually i had no choice but to totally start fresh. now thankfully it wont present any problems.

i must point out that i cannot verify this presently as i was unable to reboot. there seems to be a hard disk problem and linux gets stuck at starting Udev. will do a format/fresh install this weekend and see how it goes. my main priority is to get FC5 up and running with the ability to play games and to run XGL. uptil now i have been running XGL smoothly using the livna drivers but now i am hoping that it will run in the same way with the Nvidia.com drivers.

well, i got the proprietry drivers installed. apparently the SEinux was the one preventing it from loading up the nvidia and glx modules properly during startx. i found a workaround and finally managed to get xgl running even, its all here if anyones interested.

I'm not sure what makes red hat, who write operating systems, think they know better than nvidia who actually manufacture the video cards themselves, but I'd suggest that if you're using an nvidia card, you go to nvidia for the driver.

That's what I did and I've had no problems whatsoever as long as you follow the instructions. Read 'em first before installing

i couldnt agree with you more.the idea behind open source should be to be open to all inputs, not say that these inputs should not be trusted ( even if their closed source ).