You are not logged in.

#1 2025-06-10 20:43:03

Yerik
Member
Registered: 2025-06-10
Posts: 4

[Solved] Cannot properly downgrade Nvidia Drivers

Hello, first of all a disclaimer, this is my first time dealing with nvidia drivers in arch. My other arch install isn't focused on gaming so I've never bothered installing proper drivers (and it's an AMD gpu)

I've installed an RTX 5070 on my computer today (6.15.1arch1-1 kernel) and had to switch to nvidia-open drivers. The latest version of the nvidia-open has problem with a specific game I wish to play so I've decided to downgrade the driver.

I've used `sudo downgrade nvidia-open nvidia-utils` to 570.133.02-1

yerik@main:~ $ pacman -Qs nvidia
local/egl-gbm 1.1.2.1-1
    The GBM EGL external platform library
local/egl-wayland 4:1.1.19-1
    EGLStream-based Wayland external platform
local/egl-x11 1.0.2-1
    NVIDIA XLib and XCB EGL Platform Library
local/libvdpau 1.5-3
    Nvidia VDPAU library
local/nvidia-open 570.153.02-1
    NVIDIA open kernel modules
local/nvidia-utils 570.153.02-1
    NVIDIA drivers utilities

I rebuilt my initramfs by using `mkinitcpio -P`

Then after rebooting I noticed that the nova_core (but not noveau) module was loaded instead of the nvidia modules and drivers. Using nvidia-smi also presented an error (could not communicate with nvidia drivers). I searched on the wiki and decided to add the nvidia modules to my mkinitcpio.conf to force them being loaded early. mkinitcpio threw an error that it couldn't find nvidia modules.

I searched a bit more and found out that the driver version is tied to the kernel version. Since this specific nvidia-open version was released on may 29, I downgraded my kernel and the linux-headers to the same version that was released on may 29, which is 6.14.9-arch1-1.

After downgrading my kernel, it is still loading the nova module instead of the nvidia. And rebuilding the initramfs gives the same "module not found" error for nvidia modules.

I'm unsure how to proceed or how to guarantee I got the right kernel version for the nvidia driver. What puzzles me is that I saw on proton db and checked with a few friends that are running the 570.153.02-1 driver on the 6.14.10 kernel (which was the latest until last week) without issues.

Also for reference if I do a full system upgrade `sudo pacman -Syu` the latest driver and latest kernel work just fine. I only wish to downgrade to play one game (and because I thought it would be easy).

---

Solution: Instead of using the nvidia-open driver, I switched to the nvidia-open-dkms version, which compiled the kernel modules against my specific kernel version. Downgrading the driver also made Clair Obscur work again.

Last edited by Yerik (2025-06-11 23:04:46)

Offline

#2 2025-06-10 21:30:27

mackin_cheese
Member
Registered: 2025-01-07
Posts: 477

Re: [Solved] Cannot properly downgrade Nvidia Drivers

Alright, this seems to be an XY problem, let's start at that specific game, what is the specific game you speak of and what is the exact errors you get when running the game with the latest nvidia drivers

Offline

#3 2025-06-10 21:38:11

seth
Member
From: Won't reply 2 private help req
Registered: 2012-09-03
Posts: 75,278

Re: [Solved] Cannot properly downgrade Nvidia Drivers

https://gitlab.archlinux.org/archlinux/ … /issues/15
blacklist nova_core

After downgrading my kernel, it is still loading the nova module instead of the nvidia.

very unlikely, the module only shows up w/ 6.15

That aside you want to address the actual problem as suggested by mackin_cheese (though there's a truckload of issues w/ 575xx reported already)

Online

#4 2025-06-10 21:49:22

Yerik
Member
Registered: 2025-06-10
Posts: 4

Re: [Solved] Cannot properly downgrade Nvidia Drivers

mackin_cheese wrote:

Alright, this seems to be an XY problem, let's start at that specific game, what is the specific game you speak of and what is the exact errors you get when running the game with the latest nvidia drivers

I'm attempting to play Clair Obscur,

The error is an assertion failure on the vulkan pipeline "Expression !status && vkCreateComputePipelines".

and on protondb (https://www.protondb.com/app/1903340) I found a few complaints about the game working previously but not with the current drivers (and with 40XX and 50XX RTX GPUs), which Is what set me on this path.


seth wrote:

very unlikely, the module only shows up w/ 6.15

That aside you want to address the actual problem as suggested by mackin_cheese (though there's a truckload of issues w/ 575xx reported already)

It could be that on my notes I made a mistake on the environment I was testing at the moment, but the nvidia driver didn't load for sure on the 6.9: The screen was low res (I couldn't set the correct resolution) and nvidia-smi couldn't communicate with the drivers.

Besides being an XY problem or not, I'm also interested in why I couldn't downgrade the drivers properly as that might be useful in the future, considering the hardware is still pretty new and it's drivers could be error prone

Offline

#5 2025-06-10 21:55:44

seth
Member
From: Won't reply 2 private help req
Registered: 2012-09-03
Posts: 75,278

Re: [Solved] Cannot properly downgrade Nvidia Drivers

I'm also interested in why I couldn't downgrade the drivers properly

We'll have to see data from the actually failing system, what I can tell you right away is about the nova_core collision in 6.15, just blacklist that.

Online

#6 2025-06-11 21:37:16

Yerik
Member
Registered: 2025-06-10
Posts: 4

Re: [Solved] Cannot properly downgrade Nvidia Drivers

seth wrote:

I'm also interested in why I couldn't downgrade the drivers properly

We'll have to see data from the actually failing system, what I can tell you right away is about the nova_core collision in 6.15, just blacklist that.

Yesterday I did a full system update because I had some important things to do, but today when reproducing the issue to gather system logs I found something interesting, after I downgraded the driver, initcpio couldn't find the nvidia modules.

I suppose this is the main issue and why it was loading nova_core (which I also blacklisted now)

here are the commands I ran:

yerik@main:~ $ sudo pacman -Syu
:: Synchronizing package databases...
 core is up to date
 extra is up to date
 multilib is up to date
:: Starting full system upgrade...
 there is nothing to do

yerik@main:~ $ uname -r
6.15.1-arch1-2

yerik@main:~ $ pacman -Qs nvidia
local/egl-gbm 1.1.2.1-1
    The GBM EGL external platform library
local/egl-wayland 4:1.1.19-1
    EGLStream-based Wayland external platform
local/egl-x11 1.0.2-1
    NVIDIA XLib and XCB EGL Platform Library
local/lib32-nvidia-utils 575.57.08-1
    NVIDIA drivers utilities (32-bit)
local/libvdpau 1.5-3
    Nvidia VDPAU library
local/nvidia-open 575.57.08-5
    NVIDIA open kernel modules
local/nvidia-utils 575.57.08-3
    NVIDIA drivers utilities

yerik@main:~ $ sudo downgrade nvidia-open nvidia-utils lib32-nvidia-utils
loading packages...
warning: downgrading package lib32-nvidia-utils (575.57.08-1 => 570.153.02-1)
warning: downgrading package nvidia-open (575.57.08-5 => 570.153.02-1)
warning: downgrading package nvidia-utils (575.57.08-3 => 570.153.02-1)
resolving dependencies...
looking for conflicting packages...

Packages (3) lib32-nvidia-utils-570.153.02-1  nvidia-open-570.153.02-1  nvidia-utils-570.153.02-1

Total Installed Size:  1043.13 MiB
Net Upgrade Size:      -105.46 MiB

:: Proceed with installation? [Y/n] 
(3/3) checking keys in keyring                                                  [##############################################] 100%
(3/3) checking package integrity                                                [##############################################] 100%
(3/3) loading package files                                                     [##############################################] 100%
(3/3) checking for file conflicts                                               [##############################################] 100%
(3/3) checking available disk space                                             [##############################################] 100%
:: Processing package changes...
(1/3) downgrading nvidia-utils                                                  [##############################################] 100%
(2/3) downgrading lib32-nvidia-utils                                            [##############################################] 100%
(3/3) downgrading nvidia-open                                                   [##############################################] 100%
:: Running post-transaction hooks...
(1/8) Creating system user accounts...
(2/8) Reloading system manager configuration...
(3/8) Restarting marked services...
(4/8) Reloading device manager configuration...
(5/8) Arming ConditionNeedsUpdate...
(6/8) Updating module dependencies...
(7/8) Updating linux initcpios...
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -g /boot/initramfs-linux.img
==> Starting build: '6.15.2-arch1-1'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [microcode]
  -> Running build hook: [modconf]
  -> Running build hook: [kms]
  -> Running build hook: [keyboard]
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> ERROR: module not found: 'nvidia'
==> ERROR: module not found: 'nvidia_modeset'
==> ERROR: module not found: 'nvidia_uvm'
==> ERROR: module not found: 'nvidia_drm'
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux.img'
==> WARNING: errors were encountered during the build. The image may not be complete.
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: '6.15.2-arch1-1'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [microcode]
  -> Running build hook: [modconf]
  -> Running build hook: [kms]
==> WARNING: Possibly missing firmware for module: 'ast'
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: 'xhci_pci_renesas'
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
==> WARNING: Possibly missing firmware for module: 'wd719x'
==> WARNING: Possibly missing firmware for module: 'qed'
==> WARNING: Possibly missing firmware for module: 'qla2xxx'
==> WARNING: Possibly missing firmware for module: 'qla1280'
==> WARNING: Possibly missing firmware for module: 'bfa'
==> WARNING: Possibly missing firmware for module: 'aic94xx'
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> ERROR: module not found: 'nvidia'
==> ERROR: module not found: 'nvidia_modeset'
==> ERROR: module not found: 'nvidia_uvm'
==> ERROR: module not found: 'nvidia_drm'
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux-fallback.img'
==> WARNING: errors were encountered during the build. The image may not be complete.
error: command failed to execute correctly
(8/8) Reloading system bus configuration...
add lib32-nvidia-utils to IgnorePkg? [y/N] 
add nvidia-open to IgnorePkg? [y/N] 
add nvidia-utils to IgnorePkg? [y/N]

yerik@main:~ $ pacman -Qs nvidia
local/egl-gbm 1.1.2.1-1
    The GBM EGL external platform library
local/egl-wayland 4:1.1.19-1
    EGLStream-based Wayland external platform
local/egl-x11 1.0.2-1
    NVIDIA XLib and XCB EGL Platform Library
local/lib32-nvidia-utils 570.153.02-1
    NVIDIA drivers utilities (32-bit)
local/libvdpau 1.5-3
    Nvidia VDPAU library
local/nvidia-open 570.153.02-1
    NVIDIA open kernel modules
local/nvidia-utils 570.153.02-1
    NVIDIA drivers utilities

the important part:

==> ERROR: module not found: 'nvidia'
==> ERROR: module not found: 'nvidia_modeset'
==> ERROR: module not found: 'nvidia_uvm'
==> ERROR: module not found: 'nvidia_drm'

For completeness sake, here's a full system update after I did the downgrade. It doesn't have the same "module not found" error

yerik@main:~ $ sudo downgrade nvsudo pacman -Syu
[sudo] password for yerik: 
:: Synchronizing package databases...
 core is up to date
 extra                                                7.8 MiB  71.8 MiB/s 00:00 [##############################################] 100%
 multilib is up to date
:: Starting full system upgrade...
resolving dependencies...
looking for conflicting packages...

Packages (3) lib32-nvidia-utils-575.57.08-1  nvidia-open-575.57.08-5  nvidia-utils-575.57.08-3

Total Installed Size:  1148.59 MiB
Net Upgrade Size:       105.46 MiB

:: Proceed with installation? [Y/n] 
(3/3) checking keys in keyring                                                  [##############################################] 100%
(3/3) checking package integrity                                                [##############################################] 100%
(3/3) loading package files                                                     [##############################################] 100%
(3/3) checking for file conflicts                                               [##############################################] 100%
(3/3) checking available disk space                                             [##############################################] 100%
:: Processing package changes...
(1/3) upgrading nvidia-utils                                                    [##############################################] 100%
(2/3) upgrading lib32-nvidia-utils                                              [##############################################] 100%
(3/3) upgrading nvidia-open                                                     [##############################################] 100%
:: Running post-transaction hooks...
(1/8) Creating system user accounts...
(2/8) Reloading system manager configuration...
(3/8) Restarting marked services...
(4/8) Reloading device manager configuration...
(5/8) Arming ConditionNeedsUpdate...
(6/8) Updating module dependencies...
(7/8) Updating linux initcpios...
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -g /boot/initramfs-linux.img
==> Starting build: '6.15.2-arch1-1'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [microcode]
  -> Running build hook: [modconf]
  -> Running build hook: [kms]
  -> Running build hook: [keyboard]
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux.img'
  -> Early uncompressed CPIO image generation successful
==> Initcpio image generation successful
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: '6.15.2-arch1-1'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [microcode]
  -> Running build hook: [modconf]
  -> Running build hook: [kms]
==> WARNING: Possibly missing firmware for module: 'ast'
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: 'xhci_pci_renesas'
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
==> WARNING: Possibly missing firmware for module: 'wd719x'
==> WARNING: Possibly missing firmware for module: 'qed'
==> WARNING: Possibly missing firmware for module: 'qla2xxx'
==> WARNING: Possibly missing firmware for module: 'qla1280'
==> WARNING: Possibly missing firmware for module: 'bfa'
==> WARNING: Possibly missing firmware for module: 'aic94xx'
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux-fallback.img'
  -> Early uncompressed CPIO image generation successful
==> Initcpio image generation successful
(8/8) Reloading system bus configuration...

Last edited by Yerik (2025-06-11 21:42:00)

Offline

#7 2025-06-11 22:36:34

Scimmia
Fellow
Registered: 2012-09-01
Posts: 13,725

Re: [Solved] Cannot properly downgrade Nvidia Drivers

You're using the binary module package, which are always built against a specific kernel. You can't just downgrade them alone. Either downgrade the kernel to match, or use DKMS.

Offline

#8 2025-06-11 23:02:15

Yerik
Member
Registered: 2025-06-10
Posts: 4

Re: [Solved] Cannot properly downgrade Nvidia Drivers

Scimmia wrote:

You're using the binary module package, which are always built against a specific kernel. You can't just downgrade them alone. Either downgrade the kernel to match, or use DKMS.

using the DKMS version worked like a charm. I was able to downgrade properly, thank you! For a final question, how am I supposed to know which driver version matches which kernel version? Ofc doing a full system install will always match the version, but in this case of wanting a custom version is there anywhere I can look to see that version XXX.XXX.XXX-X matches kernel Y.YY.Y-Y (though I suppose it's just easier to use dkms anyway) ?

Offline

#9 2025-06-12 01:53:23

Scimmia
Fellow
Registered: 2012-09-01
Posts: 13,725

Re: [Solved] Cannot properly downgrade Nvidia Drivers

Offline

Board footer

Powered by FluxBB