NVIDIA 384.59 Linux Driver Released Along With 375.82

[Azrael5]

Ok I get this error information on the latest driver dowloaded frome the NVIDIA site:
ERROR: An NVIDIA kernel module 'nvidia-drm' appears to already be loaded in your kernel. This may be because it is in use (for example, by
an X server, a CUDA program, or the NVIDIA Persistence Daemon), but this may also happen if your kernel was configured without
support for module unloading. Please be sure to exit any programs that may be using the GPU(s) before attempting to upgrade your
driver. If no GPU-based programs are running, you know that your kernel supports module unloading, and you still receive this
message, then an error may have occured that has corrupted an NVIDIA kernel module's usage count, for which the simplest remedy is
to reboot your computer.

What is it meaning?

[bug77]

after having had problems with ppa, I install the first time from the distro so all is in place but later updates always from Nvidia. Never had problems like this.

PPAs are a bit risky, but Ubuntu is pretty good at providing recovery options.

Relax, guys, I'm one of the maintainers of the graphics-drivers PPA and we are definitely working on this. The new version should be available in the main PPA a bit later today.

But please, please don't install directly using the .run packages. These run the risk of putting files in the wrong place or overwriting system files and can cause major breakage if certain other packages (such as the kernel or Mesa, at least) are updated. The .deb packages we provide ensure that all files get put in the proper places and are managed correctly during updates. If you have any issues with our packages, please don't hesitate to contact us on Launchpad and describe the issue.

Do not mind debianxfce, he has a unique way of looking at Linux.

[Guest]

Ok I get this error information on the latest driver dowloaded frome the NVIDIA site:
ERROR: An NVIDIA kernel module 'nvidia-drm' appears to already be loaded in your kernel. This may be because it is in use (for example, by
an X server, a CUDA program, or the NVIDIA Persistence Daemon), but this may also happen if your kernel was configured without
support for module unloading. Please be sure to exit any programs that may be using the GPU(s) before attempting to upgrade your
driver. If no GPU-based programs are running, you know that your kernel supports module unloading, and you still receive this
message, then an error may have occured that has corrupted an NVIDIA kernel module's usage count, for which the simplest remedy is
to reboot your computer.

What is it meaning?

Try existing Xorg, close all command line daemons that may be using CUDA.
Then 'ps -Af | grep nvidia'. If the nvidia-persistence daemon is running, kill it.
Then 'modprobe -r nvidia-drm'.
Check 'lsmod | grep nvidia' output after that. if you see any nvidia modules loaded 'rmmod them'.
Then try installing the nvidia driver again.

[Azrael5]

Try existing Xorg, close all command line daemons that may be using CUDA.
Then 'ps -Af | grep nvidia'. If the nvidia-persistence daemon is running, kill it.
Then 'modprobe -r nvidia-drm'.
Check 'lsmod | grep nvidia' output after that. if you see any nvidia modules loaded 'rmmod them'.
Then try installing the nvidia driver again.

How to recognize them? May I update by ubuntu recovery section? (I'm using Kubuntu). I've switched to Mesa driver but other kind of problem occurs. this is one of the kind of problem the final user detests.

[mike44]

sudo service lightdm stop
or alike. google is your friend.

With nvidia download you could always use uninstall in case of problems but I've never needed such. What about the Ubuntu effort to make install easier?

[Guest]

How to recognize them? May I update by ubuntu recovery section? (I'm using Kubuntu). I've switched to Mesa driver but other kind of problem occurs. this is one of the kind of problem the final user detests.

If you mean chrooting from kubuntu live, then yes.

but while running ubuntu, you can do the steps above.
nvidia persistence daemon is called 'nvidia-persistenced'. you can 'killall nvidia-persistenced'.

the kernel modules have the nvidia prefix.

lsmod | grep nvidia
nvidia_drm             45056  1
nvidia_modeset        835584  7 nvidia_drm
nvidia              12976128  164 nvidia_modeset
drm_kms_helper        126976  1 nvidia_drm
drm                   294912  4 nvidia_drm,drm_kms_helper

If lsmod | grep nvidia returns nothing, the nvidia kernel module is not loaded (it can be installed on disk but not loaded).

running 'modprobe -r nvidia-drm' unloads them recursively. it also unloads drm and drm_kms_helper.
Then update your drivers.
modprobe nvidia-drm and start xorg again.

That said, the upgrade process is easier if you use prepackaged .debs from a ppa.

Two things are prevent the nvidia kernel module from being unloaded:
1) Xorg. You can exist this by exiting your display manager (gdm, kdm, lightdm, etc...).
2) running a CUDA application (if you don't know what this, then chances are you are not running one).

[pinguinpc]

In this nvidia driver wine staging still working and manual installing works

[TheLexMachine]

What's going on with the Ubuntu GPU PPA? I just gave it a look and there's nothing there except a bunch of empty directories and odd files. Looks like someone cleared it out a few days ago, based on the modification dates I see.

[timofonic]

Any hopes to have vGPU features for consumer cards? Intel is doing it with GVT. I have a laptop with GTX960M.

[Guest]

Debian like ubuntu has old nvidia drivers, so do not use them when there is newer available at nvidia site. Be warned, you can have a non booting system when removing debian packaged nvidia drivers.

That's more or less why I have stuck with Arch Linux. I can just edit the PKGBUILD, change version, modify any resultant inconsistencies, and install the different version. At the moment, I don't have time to learn how to do so on .RPM or .DEB distribution.