Ugh.. im getting so pissed trying to setup Nvidia drivers for my GeForce2 MX, things that are happening are making no sense!

I downloaded tarballs of both kernel and GLX packages, and following the README installed them with "make install" and this inserted a NVdriver into my running kernel. I modified XF86Config and changed to nvidia drivers and added glx to load. I did startx, and after 5 seconds of wild flickering the nvidia logo was displayed and hanged my system - had to reboot. Then i remembered i have rivafb running so i recompiled my kernel (2.4.17) excluding rivafb. i tried startx then and after another couple of seconds of wild flickering i got a couple of interesting dark blue lines, going vertically i beleive that hanged my system. After another reboot i took a look at XFree log and all that was in it was a bunch of meaningless symbols (like when you open a binary file with less) and at the end about 50 names of different libraries, gnome, KDE and the like, nothing that can help me or that makes any sense. Then i tried another make install for nvidia kernel and GLX since i recompiled the kernel, and GLX went smoothly while kernel refused to install with:

os-interface.d:1: warning: NUL character seen; rest of line ignored

os-interface.d:131: warning: NUL character seen; rest of line ignored

os-interface.d:132: warning: NUL character seen; rest of line ignored

os-interface.d:133: warning: NUL character seen; rest of line ignored

os-interface.d:134: warning: NUL character seen; rest of line ignored

os-interface.d:135: warning: NUL character seen; rest of line ignored

os-interface.d:136: warning: NUL character seen; rest of line ignored

os-interface.d:137: warning: NUL character seen; rest of line ignored

os-interface.d:137: *** missing separator. Stop.

which again makes no sense to me...

Then i restored my old XF86Config and came here for help, cuz i have no idea whats wrong now...



Please if you have any suggestion tell me, im running a 2.4.17 kernel..

Ahhh!

Thank you in advance!

-NsKl

OK, first, if you have an athlon and you compiled in K7 support you'll have to add "mem=nopentium" to the append line in your lilo.conf, second, you could have some mesaGL libs left from the slack install that maybe conflicting with the nvidia GLX libs, download and run nv_check.sh to check the install.

Ok, thanks Aussie, i do have a Athlon, and i didnt find that piece of information in the readme...
I'll try those things as soon as i get home, thanks for help.

-Nskl

I added append="nopentium" to the end of lilo.conf and that got me a step further.


Now when i startx i get the nvidia logo, and then i get a checker X screen (real eyesore) and it hangs just before it should load my WM.


I couldnt get nv_check.sh to work. When i click on your link it loads a page with a what looks like a bash script. I saved it as such and made it executable but it doesnt work (./nvchooser.sh: line 35: syntax error near unexpected token `echo'


./nvchooser.sh: line 35: `fi echo "checking /usr/X11R6/etc/X11/XF86Config..."')


How am i supposed to get that script to work? (i have a feeling its something really lame i should know after a year of Linux... )

What do you think is causing X to hang? I disabled rivafb so it looks like conflicting libraries, although i dont know of any GL libraries Slack installs by default...



Thanks for all your support



-Nskl

Download it with a shift click and run it with "sh nv_check.sh"

i added the append="mem=nopentium" line after my kernel entry in lilo.conf, ran the nv checker succesfully and i still get lockups when i startx (the nVIDIA logo appears and hangs the machine). The NVdriver is modprobed and shows up in lsmod. nv checker reports everything is how it should be.
What else could be possibly causing these hangs? Is there anything special i might need in the kernel? I'm stuck again, have no idea what might be wrong...
Any help greatly appreciated!
Thanks to everyone in advance

-NsKl

Ok, next thing that could be doing this is agpgart, if its compiled into the kernel it can cause problems it should be a module. Have a look at the agp section of the nvidia readme for agp options.

Ok, i recompiled the kernel making agpgart a module. I tried starting X without agpgart loaded, and i got the black screen with vertical blue lines like before, and of course the whole system hanged. In the how-to i looked through the AGP section and it said that agp will initialize with NVidia agp module, then search for agpgart and iif not present it will use no agp. So i guess this is not a AGP problem, I'll play a little bit with AGP options, but i dont think it will solve anything, and rebboting the system every 5 minutes is a big pain in the @$$, god bless ReiserFS! at least i dont need to wait for ext2 check at every boot....

Is there anything else that comes to your mind and could cause this kind of hangs? Maybe some BIOS settings? Im using a A7V266-e main board (athlon processor) and i already included append="mem=nopentium" to lilo.cong at the end of the last Kernel entry. Any other things to check?
Thanks for all the help again.

-NSKL

The only other thing I can think of is perhaps a problem with the video card, do you have another one you can use to check?

Ahh, finally got it to work! I was playing with AGP options and added Option "NVagp" "0" to disable AGP and not it works like a charm, 3 hours of soldier of fortune without any instability problems. Just one more question tho. Setting AGP to 0 (disabling it) did i slow down any rendering / am i loosing any performance this way?
Thanks for all the help again Aussie!

-NSKL

You'll be loosing a small amount of performance, but you won't notice it while playing, only with a benchmark. Glad you got things going :-)