This is ridiculous - I have been spending over eight hours yesterday and today trying to get an NVIDIA 9800GT installed on openSUSE 11.
Nothing in the usual instructions works even close.
Here is the tale of woe:
Tried One-Click Install:
Downloads, claims installation successful.
Do “sax -r”, I get this:
Current configuration will not be read in.
access to your display has been granted
Then I get a console screen with a ton of error messages
Then I get this helpful response:
ups lost card during probing…abort
Something went wrong while X was called with -probeonly
try to call sax2 -p and select a single device.
So I tried sax2 -p and I get a list that shows the onboard 8300 and the framebuffer device - not 9800GT at all.
I try the sax2 -r 0=NVIDIA, get the same thing as above.
I run the configure Nvidia Server settings, I get:
You do not appear to be using the Nvidia driver
and tells me to run the nvidia config thingy
Now I can’t even quit out of the Nvidia server settings applet! Brilliant programming!
So I try it the old way, download the current kernel source (after one bad start getting the wrong kernel source).
Switch to /usr/src/kernel-2.6.25.5-1.1, run makeoldconfig
Download the nvidia install script NVIDIA-Linux-x86_64-177.82-pkg2.run
Try to run it, I get:
First time: /usr/src/linux/include/linux/version.h does not exist because I hadn’t installed the source the first time
Second time after installing Base System, C/C++ Development, and Kernel Development patterns, I get:
Unable to load kernel module ‘nvidia.ko’ This usually happens when the kernel module is built against the wrong or improperly configured kernel source, with a version of gcc that differs from the one used to build the target kernel, or if a driver such as rivafb/nvidiafb is present and prevents the kernel modiule from obtaining ownership of the NVIDIA grahics device.
See log entries ‘kernel module load error’ and ‘kernel messages’ at end of /var/log/nvidia-installer.log
Which in turn says some gibberish like:
no version for struct_module’ found: kernel tainted
module license ‘NVIDIA’ taints kernel
unknown symbol i2c_del_adapter
unknown sumbol i2c_add_adapter
Folks, this is ridiculous. I am not an expert at compiling the kernel. I don’t even know what “configuring your kernel sources” actually means in terms of steps to be done. Google was no help. I found one place that said do the following:
make oldconfig
make modules_prepare
make modules
When I did that, it took forever to do step three. I did only step one on the current kernel sources, do I need to do all three?
I need a step by step on how to get an ASUS 9800GT video card installed on openSUSE 11 which actually explains how to do it, including all the steps to insure that the right kernel is present, configured properly, etc.
None of the existing HOWTO’s do this - they assume you know all about kernel compiling, etc.
I never had to do all this when I installed a NVIDIA card in my other openSUSE 10.3 machine! What has changed? Why all these errors?
Nobody except a hardcore Linux person could possibly install a video card with this sort of situation. I’ve used Linux for several years now as my primary desktop OS, and I can usually figure out what is going on, but this has me stumped.
There has to be an easier way to install a video card than this!
Oh, hardware: I’m running 64-bit openSUSE 11.0 on a AMD Quad-Core 9650 AM2+ CPU, ASUS MB M3N78, 4GB RAM, 1TB HD. This is a brand new install of OpenSUSE yesterday on a new machine.
By the way, the first install failed miserably. I forgot to hook up the network cable and when the install got to the test network part, of course it failed. It reported the installation failed, but when I rebooted everything was fine. Except after hours of installing software, I rebooted and it REPEATED that the installation had failed, and I had to reconfigure everything. Then it did that again, so I wiped that install and reinstalled with the network hooked up and everything seems fine.
This has been a two-day install so far, with this NVIDIA nonsense.