[PlanetCCRMA] fc7 nvidia, madwifi; ntfs

Tracey Hytry shakti at bayarea.net
Thu Feb 7 04:32:01 2008

pyrael <admin.covenantofblood@gmail.com> wrote:

> What card(s) does this work with? I have a GeForce 7200GS and the proprietaty
> and livna drivers won't work at all with the ccrma kernel. I was thinking it
> might be something as silly as the driver from nvidia's website reading the
> kernel name and not knowing what to do with it. As with you, the livna
> driver works flawlessly on my stock FC7 kernel. 

If I understand it correctly;  livna has prebuilt "kmod" kernel drivers for recent fedora kernels.
I've never seen ones that are built for  the ccrma realtime kernels on livna.

I don't know if this would work, and if someone was so brave to try it would be pretty cool:
- Find out what version of the nvidia driver from livna you are using, and download that version from nvidia.
- Try telling it to build a kernel module for the rt kernel.
  ( I do something like 
       sh NVIDIA-Linux-x86_64-169.09-pkg2.run -a -N -K --kernel-name=
             as root, but I've only done this after installing the whole nvidia package and needed to build modules for
             all of the other kernels on fedora7 here)

It's quite possible that the nvidia package from nvidia will refuse to build the kernel module because the glx libraries from livna are different then nvidia's,  or nvidia leaves a marker somewhere telling it which version of itself is installed?  This I really don't know?

For myself, I do the whole nvidia build from the downloaded driver package from the nvidia site.  Then I build the modules for the other currently non-running kernels as shown above.  Every time there's a new xorg update I check and see if the xorg update has written over the glx-core libraries and reinstall the nvidia driver from scratch again, along with all of the kernel modules for the other kernels.  At first this was a real pain, but now it goes pretty fast.

I'm using a GeForce 7600GS at this time.