Created attachment 14565 [details]
files list for nvidia-current-550.78

files list for nvidia-current-550.78

Created attachment 14566 [details]
files list for nvidia470-470.239.06-2.mga9.nonfree

files list for nvidia470-470.239.06-2.mga9.nonfree

I have already tested both OK in Bug 32579, on my workstation "svarten".

$ inxi -SMCG
System:
  Host: svarten.tribun Kernel: 6.6.28-desktop-1.mga9 arch: x86_64 bits: 64
    Desktop: KDE Plasma v: 5.27.10 Distro: Mageia 9
Machine:
  Type: Desktop Mobo: ASRock model: P55 Pro serial: <superuser required>
    BIOS: American Megatrends v: P2.60 date: 08/20/2010
CPU:
  Info: dual core model: Intel Core i7 870 bits: 64 type: MT MCP cache:
    L2: 512 KiB
  Speed (MHz): avg: 1205 min/max: 1200/2934 cores: 1: 1205 2: 1205 3: 1205
    4: 1205
Graphics:
  Device-1: NVIDIA GM107 [GeForce GTX 750] driver: nvidia v: 550.78
  Display: x11 server: X.org v: 1.21.1.8 with: Xwayland v: 22.1.9 driver: X:
    loaded: nvidia,v4l gpu: nvidia,nvidia-nvswitch resolution: 3840x2160~60Hz
  API: OpenGL v: 4.6.0 NVIDIA 550.78 renderer: NVIDIA GeForce GTX
    750/PCIe/SSE2

CC: (none) => fri
Assignee: bugsquad => qa-bugs
Blocks: (none) => 32579
Summary: Update request: nvidia-current-550.78-1.mga9.nonfree => Update request: nvidia-current-550.78-1.mga9.nonfree & nvidia470-470.239.06-2.mga9.nonfree

MGA9-64 Plasma, i5-7500, nvidia Quadro K620. My test install has both server and desktop kernels installed.

I started in the server kernel, using the nvidia-current driver. I downloaded b
both lists with qarepo, and installed updates, including nvidia-current. The driver was built successfully. I rebooted, tried a few things, with no issues to report.

I booted into the desktop kernel, and in the process the driver was built for this kernel. Again I tried a few things, with no issues noted. Then I used XFdrake to switch to the 470 driver still in the repo downloaded by qarepo. I rebooted into the desktop kernel, tried a few things, again with no issues.

I booted into the server kernel, and the 470 driver was built and installed in the process. I tried a few things, with no issues noted. Finally, I used XFdrake to switch back to nvidia-current once more, rebooted, tried a few things, with no issues to report.

Looks OK on this hardware.

CC: (none) => andrewsfarm

I've had less luck.  MCC wouldn't allow me to install the relevant nvidia-current for 550.  No reason given.
I suspect I'm just missing packages it wouldn't allow me to install.
As a consequence:
Nvidia Display Settings cannot see me 1080Ti
cuda-z cannot see the card either.

CC: (none) => tablackwell

@Tony: so what packages do you have?
  On my system:
$ rpm -qa | grep nvidia-current
nvidia-current-doc-html-550.78-1.mga9.nonfree
nvidia-current-utils-550.78-1.mga9.nonfree
dkms-nvidia-current-550.78-1.mga9.nonfree
x11-driver-video-nvidia-current-550.78-1.mga9.nonfree
nvidia-current-cuda-opencl-550.78-1.mga9.nonfree

Do urpmi work better to install the packages?

Little nvidia trouble in past, but got to get around this:

OK, deleted everything nvidia via MCC, rebooted.

Next, Tried to install from the list:

Selected dkms-nvidia-current-550.78-1.mga9 - OK

then tried to select nvidia-current-all:
Sorry, the following package cannot be selected:

- nvidia-current-all-550.78-1.mga9.nonfree.x86_64 (due to conflicts with nvidia-current-lib32-550.78-1.mga9.nonfree.x86_64)

Further down list, tried to select nvidia-current-lib32:
Sorry, the following package cannot be selected:

- nvidia-current-lib32-550.78-1.mga9.nonfree.x86_64 (due to unsatisfied libgl1)

Installed the rest from the list, rebooted:

cuda-z had disappeared - reinstalled it:

$ cuda-z
libpng warning: iCCP: known incorrect sRGB profile
libpng warning: iCCP: known incorrect sRGB profile
libpng warning: i

Tried to run nvidia-settings:
$ nvidia-settings

ERROR: NVIDIA driver is not loaded


(nvidia-settings:4086): GLib-GObject-CRITICAL **: 16:22:21.878: g_object_unref: assertion 'G_IS_OBJECT (object)' failed

** (nvidia-settings:4086): CRITICAL **: 16:22:21.879: ctk_powermode_new: assertion '(ctrl_target != NULL) && (ctrl_target->h != NULL)' failed

Rebooted again.
No msg re building nvidia driver.
logged in as root, ran XFdrake.  No popup re using the proprietary driver.
Still no nvidia driver loaded.

re comment 6

$ rpm -qa | grep nvidia-current
nvidia-current-doc-html-550.78-1.mga9.nonfree
nvidia-current-utils-550.78-1.mga9.nonfree
dkms-nvidia-current-550.78-1.mga9.nonfree
nvidia-current-cuda-opencl-550.78-1.mga9.nonfree
nvidia-current-devel-550.78-1.mga9.nonfree
x11-driver-video-nvidia-current-550.78-1.mga9.nonfree

only difference seems to be that I also selected devel, but wouldn't have expected that to be causing the issue?

Try running drakx11, select the driver, etc, reboot.

drakx11 from root text login didn't give me any save option, just exit (because no change?)
reboot: built nothing, nvidia driver not loaded.

This is a test system.  I'm tempted to save what little I need on another partition and reinstall M9, but happy to hear other options!

In all steps below it is good to launch the tools from a terminal to see output.
For same reason keep journalctl -f running in another terminal.

1) run drakx11 to select nouveau or xorg modesetting, reboot
It should now use a free driver.

2) cleaning: remove all nvidia packages
verify you have the correct kernel-*-devel for your kernel

3) run drakx11 to select nvidia "Driver: Latest Legacy" that is nvidia470
then urpmi or drakrpm to install corresponding nvidia470-cuda-opencl
(this is not installed by drakx11, but other parts are)

4) reboot. do 470 work?

5) repeat 2 but select "Driver: Production) that is 550, and corresponding cuda-opencl.

6) reboot. do 550 work?

Typo: in 5) repeat *3* not 2.

nvidia-current-lib32 is for installing the 32bit nvidia libraries in 64bit environment, which are useful for 32bit gaming for instance (steam, wine, etc.), but to use them you need the mga 32bit media sources enabled (which apparently isn't), that's why it returns the error "unsatisfied libgl1" (which is a 32bit library from 32bit libglvnd and partially requires a 32bit mesa).

nvidia-current-all is for installing the "full option" nvidia drivers (including cuda and 32bit compatibility).

For the rest, what /usr/sbin/update-alternatives --display gl_conf shows?

OK, wiped the partition and reinstalled M9 Classical.
Enabled repository (aarnet.edu.au) then enabled core-updates-testing and nonfree updates testing.

won't let me install nvidia-current-all or nvidia-current-lib32.  Installed the rest.  Rebooted.  No further msg re building nvidia but nvidia settings shows driver version 550.78 and after installing cuda-z it works.

Noted comment 14.  Enabled Core 32bit Release (distrib 31) - it wasn't enabled.
MCC now allows selecting nvidia-current-all and installing it.

Can't explain all the unexpected difficulties I experienced above, but on a clean system everything 'just works' as expected.

No residual problems at my end.  Sorry for this storm in a teacup.
OK to go from my point of view (and all the others who have looked at it I think...)

Good, thanks.

Whiteboard: (none) => MGA9-64-OK
Keywords: (none) => validated_update
CC: (none) => sysadmin-bugs

An update for this issue has been pushed to the Mageia Updates repository.

https://advisories.mageia.org/MGAA-2024-0145.html

Resolution: (none) => FIXED
Status: NEW => RESOLVED