How do I get the 440.64 NVIDIA driver?

[steve7233]

I looked up the latest driver for the 10 series on NVIDIA's web site.

Version 440.64
Release Date Fri Feb 28, 2020
Operating System Linux 64-bit
Language English (US)
File Size 134.76 MB

I would rather have the package for Eoan Ermine than risk running the generic version from NVIDIA.

I believe there is another repo that has more updated packages than the main ubuntu ones. I just forgot where it lives.

[steve7233]

I found it: https://launchpad.net/~graphics-driv...ive/ubuntu/ppa Unfortunately they don't have all the 440 packages in the repo yet. The best I could get was the 435.21 driver.

[oshunluvr]

At here with KDE Neon 18.04 LTS:

stuart@office:~$ dkms status
nvidia, 440.64, 5.0.0-37-generic, x86_64: installed
nvidia, 440.64, 5.3.0-42-generic, x86_64: installed

Using this PPA: https://launchpad.net/~graphics-driv...ive/ubuntu/ppa




Why do you think something is missing?

[steve7233]

At here with KDE Neon 18.04 LTS:

stuart@office:~$ dkms status
nvidia, 440.64, 5.0.0-37-generic, x86_64: installed
nvidia, 440.64, 5.3.0-42-generic, x86_64: installed

Using this PPA: https://launchpad.net/~graphics-driv...ive/ubuntu/ppa

[ATTACH=CONFIG]8721[/ATTACH]

Why do you think something is missing?

That is weird. On my system, the same repo only goes up to NVIDIA 435 It doesn't show any of 440. How can this be since we are accessing the same repo? I am using muon what did you use to get the list?

[oshunluvr]

sudo apt update

[steve7233]

I found the 440 driver. It must have changed something when I shut it down for the night. I tried upgraded to the 440 driver. I rebooted. Black screen. I power reset. I tried every recovery mode and kernal all broken. I tryed switching to a tty and got no response. So I grabbed my cell and posted. Do I need to reinstall.

Emergency mode on all recovery mode kernals is broken. Also my phone is at 12% power. I'll have recharge it before I can respond again. It's a fast USB 3 charger so about an 1 HR and a half.

[steve7233]

After about the 16th try emergency mode finally worked. I had to do a google search because I forgot how the ubuntu auto driver install is syntax. I now have the newest NVIDIA driver from Ubuntu. I guess muon is bugged and you have to use apt manually to get the NVIDIA driver installed properly. I never encountered a hosed system because of mis-install by muon but I guess anything is possible. Hopefully, the version of muon in 20.04 isn't bugged.

[oshunluvr]

IMO I would never attempt a kernel module install from Muon or any other GUI package manager - at least not until we get a good one. If you want to install a new nvidia driver (assuming it's in a PPA), open Konsole. Then type:

apt-cache search nvidia-driver

This will list all available drivers - you're looking for "nvidia-driver-..." with the version at the end. So 440 is "nvidia-driver-440". This is a meta package which means it should include all you need when you install it. Before installing, enter:

dkms status

This will list all installed kernel modules that use dkms like nvidia. It should list all your installed kernels that have the driver installed to them. Mine currently shows:

[FONT=monospace][COLOR=#54FF54][B]stuart@office[/B][/COLOR][COLOR=#000000]:[/COLOR][COLOR=#54FFFF][B]~[/B][/COLOR][COLOR=#000000]$ dkms status                          [/COLOR]
nvidia, 440.64, 5.0.0-37-generic, x86_64: installed
nvidia, 440.64, 5.3.0-45-generic, x86_64: installed[/FONT]

Then install the new nvidia driver and check back to be sure it installed by repeating the above dkms status command. Then reboot when it looks OK.

In my case , I actually have three kernels installed at the moment but the nvidia driver is only installed into the latest version of each series:

[FONT=monospace][COLOR=#54FF54][B]stuart@office[/B][/COLOR][COLOR=#000000]:[/COLOR][COLOR=#54FFFF][B]~[/B][/COLOR][COLOR=#000000]$ dpkg -l |grep linux-image-5 | grep ii[/COLOR]
[COLOR=#FF5454][B]ii[/B][/COLOR][COLOR=#000000]  linux-image-5.0.0-36-generic                      5.0.0-36.39~18.04.1                              amd64        Signed kernel image generic[/COLOR]
[COLOR=#FF5454][B]ii[/B][/COLOR][COLOR=#000000]  linux-image-5.0.0-37-generic                      5.0.0-37.40~18.04.1                              amd64        Signed kernel image generic[/COLOR]
[COLOR=#FF5454][B]ii[/B][/COLOR][COLOR=#000000]  linux-image-5.3.0-45-generic                      5.3.0-45.37~18.04.1                              amd64        Signed kernel image generic[/COLOR][/FONT]

So if I booted to the -36 kernel I would not get my nvidia driver, it would use noveau instead. So I can fix that:

[FONT=monospace][COLOR=#54FF54][B]stuart@office[/B][/COLOR][COLOR=#000000]:[/COLOR][COLOR=#54FFFF][B]~[/B][/COLOR][COLOR=#000000]$ sudo dkms install -m nvidia/440.64  -k 5.0.0-36-generic[/COLOR][/FONT][FONT=monospace][COLOR=#000000]
[/COLOR][/FONT][FONT=monospace]
Kernel preparation unnecessary for this kernel.  Skipping...
applying patch disable_fstack-clash-protection_fcf-protection.patch...patching file Kbuild
Hunk #1 succeeded at 79 (offset 8 lines).

applying patch buildfix_kernel_5.6.patch...patching file common/inc/nv-procfs.h
patching file common/inc/nv-time.h
patching file conftest.sh
Hunk #1 succeeded at 1267 (offset -3 lines).
Hunk #2 succeeded at 2936 (offset -3 lines).
patching file nvidia-modeset/nvidia-modeset.Kbuild
patching file nvidia-uvm/nvidia-uvm.Kbuild
patching file nvidia/linux_nvswitch.c
patching file nvidia/nv-procfs.c
patching file nvidia/nvidia.Kbuild
patching file nvidia/nvlink_linux.c


Building module:
cleaning build area...
unset ARCH; [ ! -h /usr/bin/cc ] && export CC=/usr/bin/gcc; env NV_VERBOSE=1 'make' -j8 NV_EXCLUDE_BUILD_MODULES='' KERNEL_UNAME=5.0.0-36-generic IGNORE_XEN_PRESENCE=1 IGNORE_CC_MISMATCH=
1 SYSSRC=/lib/modules/5.0.0-36-generic/build LD=/usr/bin/ld.bfd modules.......
cleaning build area...

DKMS: build completed.

nvidia.ko:
Running module version sanity check.
 - Original module
   - No original module exists within this kernel
 - Installation
   - Installing to /lib/modules/5.0.0-36-generic/kernel/drivers/char/drm/

nvidia-modeset.ko:
Running module version sanity check.
 - Original module
   - No original module exists within this kernel
 - Installation
   - Installing to /lib/modules/5.0.0-36-generic/kernel/drivers/char/drm/

nvidia-drm.ko:
Running module version sanity check.
 - Original module
   - No original module exists within this kernel
 - Installation
   - Installing to /lib/modules/5.0.0-36-generic/kernel/drivers/char/drm/

nvidia-uvm.ko:
Running module version sanity check.
 - Original module
   - No original module exists within this kernel
 - Installation
   - Installing to /lib/modules/5.0.0-36-generic/kernel/drivers/char/drm/

depmod...

DKMS: install completed.[/FONT]

and now:

[FONT=monospace][COLOR=#54FF54][B]stuart@office[/B][/COLOR][COLOR=#000000]:[/COLOR][COLOR=#54FFFF][B]~[/B][/COLOR][COLOR=#000000]$ dkms status[/COLOR]
nvidia, 440.64, 5.0.0-36-generic, x86_64: installed
nvidia, 440.64, 5.0.0-37-generic, x86_64: installed
nvidia, 440.64, 5.3.0-45-generic, x86_64: installed[/FONT]

Good to go. Now I can reboot using the -36 kernel and nvidia will load. When you install a new nvidia driver, you should see the above output for each kernel (the latest of the same version) and if you don't - don't reboot until you figure out why.