HOWTO: Install Nvidia drviers including DKMS module

[go_beep_yourself]

did any one had a low res problem and fixed it?

with no xorg.conf in karmic it seems to be inpossible to fix this very low res issue, i tried many ways to create one 'xorg.conf' but non worked.
it the fourth time i reinstall karmic trying to get this nvidia working.

You are not the only one. I think it's time lots of people file bug reports at https://bugs.launchpad.net/ubuntu/+s...enter/+filebug
to get the developers' attention. They could try to blame this on nvidia, but I even have low resolution using the nv driver. Ubuntu (Looking in System>Pref>Display) can no longer detect the model of my monitor like it has in the past or with my other monitor (dual monitors), possibly the reason why resolution can't be set higher.

[robertsaron]

This how to is amazing. Parts of it helped. I got to the step where it says to run the script, then had to quit.There are several questions I need help with. First off where to get the script? I did not see a link anywhere to download it.

Second, the path to /etc/default/linux-restricted-modules-common does not exist on my system. I am using the latest version of Kbuntu. So maybe it is different. What is the path then if it is different?

The next thing, only thing the machine boots into is shell. So how to boot into safe mode from shell? I already have the latest driver for nvidia downloaded. When doing the install, Here is an error of comes up.

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 preevents the NVIDIA kernel module from obtaining ownership of the NVIDIA graphics device(s), or NVIDIA GPU installed in this system is not supported by this NVIDIA Linux graphics driver release.

My card is a nvidia FX 5800 Geforce, Kernel version is 2.6.32.22-generic, Nvidia version trying to install: 195.36.24

[uniden9]

It appears nvidia has changed the extract folder structure. The usr/src/nv is now kernel folder in the newer 256.53 driver.

Edit line 40 of installdkmc.sh V4 and change it from:
readonly installersourcedir=usr/src/nv
to
readonly installersourcedir=kernel

With out this change, you will get a error similar to
cp: cannot stat `/tmp/tmp.{Random Text}/installer/usr/src/nv/*': No such file or directory

For information on the extraction change, just run:
./NVIDIA-Linux-x86_64-256.53.run --extract-only --target nvidia_extract
and compare it to a older driver.

[Nitronium]

What a fracking awesome thread. Npl, you're a dude, many thanks. I chose not to run your script, but rather do it manually so that I know exactly what's going on. Love it, cheers man...

[spleach]

Cheers for the script. I also went through each step manually and am using it now with 295.20 drivers.
I did two things differently though:-
1) After extracting, moved the kernel sources from ./NVIDIA-Linux_x86_64-295.20/kernel/ rather than from ./NVIDIA-Linux_x86_64-295.20/usr/src/nv. This is as uniden says above.

2) I didn't have a Makefile in that directory, so a simple `make` doesn't work. Instead, there are three alternatives: makefile, Makefile.kbuild and Makefile.nvidia
So I added a PRE_BUILD command to dkms.conf:-
PRE_BUILD="make -f makefile"

This pre-build step creates a sym-link from Makefile to Makefile.kbuild, allowing the normal make command to work. I'm updating to the latest kernel now... If this works, it'll be the first time in ages(!) I've ever rebooted cleanly after upgrading the kernel. Fingers crossed!