Hi, I'm yet another person who can't get opengl working under linux :)

I’ve got an nvidia tnt1, redhat 6.2 with X4.01 running. I seem to have gotten the X kernal driver up and running just fine. Got NVIDIA_kernel-0.9-5.tar.gz ran the install script changed one line in my XF86Config and off I went. The GLX driver is another matter entirely. I blew out the links to various libGL.so.*'s as per the faqs request and ran a make; make install (they seem to have done the same thing) and added Load “glx” to my XF86Config, but quake2 just crashes with a segmentation fault. What’s more an app I got with the SDL graphics library called testgl reports I’ve got no Opengl support. Maybe I blew out something I shouldn’t have, but I’m pretty sure I didn’t do anything that wasn’t in the FAQ from nvidia, and if I did I’d have expected the GLX install script to put it back (or at least I’d expect and error about a missing file, not just a segmentation fault). I know the TNT_USERS_README mentions a problem with memory and mentions os-registry.c. But that file isn’t in the GLX tar, was it in the kernel driver tar (which I’ve since deleted). Any help would be appreciated, since I’m way over my head here. Thanks in advance :-).

I’ve seen your question. Have you entirely deleted the mesa drivers? If you did you must reinstall the and relink those files that have the same function(e.g. :ln libGLU.so libMesaGLU.so).

Did you forgot to download the kernel drivers? Then NOTHING will run!

I’m in the same boat - Trying to get accelerated drivers running under redhat 7.0. I’ve got both the kernel and the GLX driver running (I think!) but know of no way to test this out.

What other documentation is there on this? I’ve read through the NVidia instructions - which were next-to useless, and the NVidia-linux mini-howto (http://www.alainsplace.f2s.com/nvidia/nvidia-linux-mini-howto.html)

and I’m still rather clueless!

Hi

Some things to check for if acceleration is missing with Nvidia boards:

o The kernel module must be loaded successfully. Use /sbin/lsmod, on my box it gives

NVdriver 529596 12 (autoclean)
agpgart 18600 2 (autoclean) [NVdriver]

o The XFree glx module is loaded. Look in /var/log/XFree86.0.log

Again from my box

(II) LoadModule: “glx”
(II) Loading /usr/X11R6/lib/modules/extensions/libglx.so
(II) Module glx: vendor=“NVIDIA Corporation”
compiled for 4.0.1, module version = 1.0.5
Module class: XFree86 Server Extension
ABI class: XFree86 Server Extension, version 0.1
(II) Loading extension GLX

which shows that the Nvidia version is used.

o Your Xfree conf file should contain an empty dri section:

Section “DRI”
EndSection

o The first libGL.so on your system is the one from Nvidia

Hope it helps,
Hans Palbøl

Originally posted by Enrico:
[b]I’m in the same boat - Trying to get accelerated drivers running under redhat 7.0. I’ve got both the kernel and the GLX driver running (I think!) but know of no way to test this out.

What other documentation is there on this? I’ve read through the NVidia instructions - which were next-to useless, and the NVidia-linux mini-howto (http://www.alainsplace.f2s.com/nvidia/nvidia-linux-mini-howto.html)

and I’m still rather clueless! [/b]

Hi

Some things to check for if acceleration is missing with Nvidia boards:

o The kernel module must be loaded successfully. Use /sbin/lsmod, on my box it gives

NVdriver 529596 12 (autoclean)
agpgart 18600 2 (autoclean) [NVdriver]

o The XFree glx module is loaded. Look in /var/log/XFree86.0.log

Again from my box

(II) LoadModule: “glx”
(II) Loading /usr/X11R6/lib/modules/extensions/libglx.so
(II) Module glx: vendor=“NVIDIA Corporation”
compiled for 4.0.1, module version = 1.0.5
Module class: XFree86 Server Extension
ABI class: XFree86 Server Extension, version 0.1
(II) Loading extension GLX

which shows that the Nvidia version is used.

o Your Xfree conf file should contain an empty dri section:

Section “DRI”
EndSection

o The first libGL.so on your system is the one from Nvidia

Hope it helps,
Hans Palbøl

Originally posted by Enrico:
[b]I’m in the same boat - Trying to get accelerated drivers running under redhat 7.0. I’ve got both the kernel and the GLX driver running (I think!) but know of no way to test this out.

What other documentation is there on this? I’ve read through the NVidia instructions - which were next-to useless, and the NVidia-linux mini-howto (http://www.alainsplace.f2s.com/nvidia/nvidia-linux-mini-howto.html)

and I’m still rather clueless! [/b]

Thanks. I’ve got the kernel driver working great, and even had the GLX module loading without any complaints in my XFree86.0.log, but I didn’t have the DRI module. I’ll go try that now, thanks again.

Jeremy Gregorio
jgreg@azstarnet.com

I was able to install and get running the latest NVIDIA drivers install and running
under REDHAT 7.0 with Xfree 4.0.1 and NVIDIA’s smp kernel and gl drivers
I install from the rpm’s and had trouble getting the the kernel driver to load so
I rebuilt it from the source rpm just like nvidia’s instructions said.
But quake3 keep crashing so just like on a post that I found out there that talked
about the libGL* lib files so I renamed them and reinstalled the nvidia glx rpm
and tryed quake3 again and found that I needed libglu* so I renamed it back
and have quake3 running.
Have not checked the frame rate. But it is fast enough for me.

The link below has all the info you need to know if you print out THE FAQ file
and keep trying the things on the list you will get it to work.

If when quake exit’s if you shut down xwindows you can see some of the messages from the startup of xwindows and quake and the error when quake stops. Also look at the log file for xwindows.
http://www.nvidia.com/Products/Drivers.nsf/Linux.html

This topic was automatically closed 183 days after the last reply. New replies are no longer allowed.