Thanks for the insight, poltsy.
Yea, Leigh that sentence reads poorly doesn't it? True, 4x and 8x refer to the AGP port speed. However from what I read there is a correlation between a card's port speed and its voltage for example here
. My intent was to determine compatibility. Maybe I should simply ask, "Is a Quadro FX 4000 compatible with a 1.5 volt AGP 4x slot?"
I don't want to buy a card and find out it won't work with my system.
For now, I've settled on option 4 "Wait until nVidia driver software fixes catch up with my issue.". My issue might be solved already. I noticed there is a 96.43.16 (legacy, prerelease) for Linux x86
. I had a working system with kernel-PAE-188.8.131.52-166.fc12.i686 until performed a "yum update" followed by a "yum erase kernel-PAE-184.108.40.206-166.fc12.i686" (Oops!) and forgot about the nvidia driver dependancy on the kernel.
The release note indicates "Improved compatibility with recent Linux kernels" so I am hopeful.
I'm anxious to try this driver version. I downloaded and I have attempted to use script, NVIDIA-Linux-x86-96.43.16-pkg1.run. Unfortunately I haven't been able to get past this error shown in nvidia-installer.log (attached):
ERROR: Unable to load the kernel module 'nvidia.ko'. This happens most
frequently when this kernel module was built against the wrong or
improperly configured kernel sources, 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 NVIDIA kernel module from
obtaining ownership of the NVIDIA graphics device(s).
I am pretty sure I have the right kernel packages
[dabruhn@precision ~]$ rpm -qa|grep kernel
I can tell which gcc version and package I have installed:
[dabruhn@precision ~]$ cat /proc/version
Linux version 220.127.116.11-174.2.3.fc12.i686.PAE (firstname.lastname@example.org) (gcc version 4.4.2 20091222 (Red Hat 4.4.2-20) (GCC) ) #1 SMP Mon Jan 18 20:06:44 UTC 2010
[dabruhn@precision ~]$ rpm -qa|grep gcc
I thought maybe the nouveau is getting in the way of the install script so I also tried the script after running these commands and received the same error message:
[dabruhn@precision nVidia]$ su -
[root@precision ~]# sed -i '/root=/s|$| rdblacklist=nouveau|' /boot/grub/grub.conf
[root@precision ~]# mv /boot/initramfs-$(uname -r).img /boot/initramfs-$(uname -r)-nouveau.img
[root@precision ~]# dracut /boot/initramfs-$(uname -r).img $(uname -r)
[root@precision ~]# sed -i '/root=/s|$| vmalloc=256m|' /boot/grub/grub.conf
[root@precision ~]# getsebool -a |grep allow_execstack
allow_execstack --> on
[root@precision ~]# telinit 3
I think I have the, "I cannot build the NVIDIA kernel module, or, I can build the NVIDIA kernel module, but modprobe/insmod fails to load the module into my kernel. What is wrong?" error shown in the install guide
[root@precision ~]# locate kernel-headers
I was expecting but did not see a *.h file related to the kernel in directory, /usr/src/kernels/18.104.22.168-174.2.3.fc12.i686.PAE/include/linux
I think I'm missing a source kernel header package.
[root@precision ~]# yum install kernel-source
Loaded plugins: fastestmirror, presto, refresh-packagekit
Loading mirror speeds from cached hostfile
* fedora: mirrors.tummy.com
* rpmfusion-free: mirrors.tummy.com
* rpmfusion-free-updates: mirrors.tummy.com
* rpmfusion-nonfree: mirrors.tummy.com
* rpmfusion-nonfree-updates: mirrors.tummy.com
* updates: mirrors.tummy.com
Setting up Install Process
No package kernel-source available.
Nothing to do
What can I try next to make the install script work?