Most recent update makes nvidia drivers fail (580xx, in my case)

[cefk]

Up to today’s kernel update, everything was working fine.

Look at this, to get some details:

5606d4e

This is the output, when I do a sudo pacman -S nvidia-580xx-dkms:

Warnung: nvidia-580xx-dkms-580.159.03-2 ist aktuell – Reinstalliere
Abhängigkeiten werden aufgelöst …
Nach in Konflikt stehenden Paketen wird gesucht …

Paket (1)                  Alte Version  Neue Version  Netto-Veränderung

cachyos/nvidia-580xx-dkms  580.159.03-2  580.159.03-2           0,00 MiB

Gesamtgröße der installierten Pakete:  117,90 MiB
Größendifferenz der Aktualisierung:      0,00 MiB

:: Installation fortsetzen? [J/n]
(1/1) Schlüssel im Schlüsselbund werden geprüft             [--------------------------------] 100%
(1/1) Paket-Integrität wird überprüft                       [--------------------------------] 100%
(1/1) Paket-Dateien werden geladen                          [--------------------------------] 100%
(1/1) Auf Dateikonflikte wird geprüft                       [--------------------------------] 100%
:: Pre-transaction-Hooks werden gestartet …
(1/1) Remove upgraded DKMS modules
==> dkms remove nvidia/580.159.03
:: Paketänderungen werden verarbeitet …
(1/1) Reinstalliert wird nvidia-580xx-dkms                  [--------------------------------] 100%
:: Post-transaction-Hooks werden gestartet …
(1/5) Arming ConditionNeedsUpdate…
(2/5) Install DKMS modules
==> dkms install --no-depmod nvidia/580.159.03 -k 6.18.28-1-cachyos-lts

Error! Bad return status for module build on kernel: 6.18.28-1-cachyos-lts (x86_64)
Consult /var/lib/dkms/nvidia/580.159.03/build/make.log for more information.
==> WARNING: `dkms install --no-depmod nvidia/580.159.03 -k 6.18.28-1-cachyos-lts’ exited 10
==> dkms install --no-depmod nvidia/580.159.03 -k 7.0.5-1-cachyos-bore-lto

Error! Bad return status for module build on kernel: 7.0.5-1-cachyos-bore-lto (x86_64)
Consult /var/lib/dkms/nvidia/580.159.03/build/make.log for more information.
==> WARNING: `dkms install --no-depmod nvidia/580.159.03 -k 7.0.5-1-cachyos-bore-lto’ exited 10
(3/5) Updating linux initcpios…
==> Building image from preset: /etc/mkinitcpio.d/linux-cachyos-bore-lto.preset: ‘default’
==> Using default configuration file: ‘/etc/mkinitcpio.conf’
 → -k /boot/vmlinuz-linux-cachyos-bore-lto -g /boot/initramfs-linux-cachyos-bore-lto.img
==> Using drop-in configuration file: ‘10-chwd.conf’
==> Using drop-in configuration file: ‘10-chwd-kms.conf’
==> Starting build: ‘7.0.5-1-cachyos-bore-lto’
 → Running build hook: [base]
 → Running build hook: [systemd]
 → Running build hook: [autodetect]
 → Running build hook: [microcode]
 → Running build hook: [modconf]
 → Running build hook: [block]
 → Running build hook: [keyboard]
 → Running build hook: [sd-vconsole]
 → Running build hook: [plymouth]
 → 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-cachyos-bore-lto.img’
==> WARNING: errors were encountered during the build. The image may not be complete.
==> Building image from preset: /etc/mkinitcpio.d/linux-cachyos-lts.preset: ‘default’
==> Using default configuration file: ‘/etc/mkinitcpio.conf’
 → -k /boot/vmlinuz-linux-cachyos-lts -g /boot/initramfs-linux-cachyos-lts.img
==> Using drop-in configuration file: ‘10-chwd.conf’
==> Using drop-in configuration file: ‘10-chwd-kms.conf’
==> Starting build: ‘6.18.28-1-cachyos-lts’
 → Running build hook: [base]
 → Running build hook: [systemd]
 → Running build hook: [autodetect]
 → Running build hook: [microcode]
 → Running build hook: [modconf]
 → Running build hook: [block]
 → Running build hook: [keyboard]
 → Running build hook: [sd-vconsole]
 → Running build hook: [plymouth]
 → 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-cachyos-lts.img’
==> WARNING: errors were encountered during the build. The image may not be complete.
Fehler: Befehl konnte nicht korrekt ausgeführt werden
(4/5) Check if user should be informed about rebooting after certain system package upgrades.
==> INFO: Reboot is recommended due to the upgrade of core system package(s).
(5/5) Checking which packages need to be rebuilt

Downgrading to the older kernels isn’t an option, for me (I always clear the cache), but it’s also not a great idea, imho, because the recent update likely fixes a vulnerability, that dropped, yesterday. And I don’t enjoy being vulnerable.

[larsallan4k]

I got the seemingly exact same issue.. but I have an older pascal nvidia MX150 gpu. And now I am writing this from the live disk install… and cannot install cachyos.

This is the error I get when the install was almost done:
Kommandoen mkinitcpio blev færdig med afslutningskoden 2.

Output:

==> Building image from preset: /etc/mkinitcpio.d/linux-cachyos-lts.preset: ‘default’

==> Using default configuration file: ‘/etc/mkinitcpio.conf’

-> -k /boot/vmlinuz-linux-cachyos-lts -g /boot/initramfs-linux-cachyos-lts.img

==> Using drop-in configuration file: ‘10-chwd.conf’

==> Using drop-in configuration file: ‘10-chwd-kms.conf’

==> Starting build: ‘6.18.28-1-cachyos-lts’

-> Running build hook: [base]

-> Running build hook: [systemd]

-> Running build hook: [autodetect]

-> Running build hook: [microcode]

-> Running build hook: [modconf]

-> Running build hook: [block]

-> Running build hook: [keyboard]

-> Running build hook: [sd-vconsole]

-> Running build hook: [plymouth]

-> Running build hook: [filesystems]

==> 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-cachyos-lts.img’

==> WARNING: errors were encountered during the build. The image may not be complete.

==> Building image from preset: /etc/mkinitcpio.d/linux-cachyos.preset: ‘default’

==> Using default configuration file: ‘/etc/mkinitcpio.conf’

-> -k /boot/vmlinuz-linux-cachyos -g /boot/initramfs-linux-cachyos.img

==> Using drop-in configuration file: ‘10-chwd.conf’

==> Using drop-in configuration file: ‘10-chwd-kms.conf’

==> Starting build: ‘7.0.5-1-cachyos’

-> Running build hook: [base]

-> Running build hook: [systemd]

-> Running build hook: [autodetect]

-> Running build hook: [microcode]

-> Running build hook: [modconf]

-> Running build hook: [block]

-> Running build hook: [keyboard]

-> Running build hook: [sd-vconsole]

-> Running build hook: [plymouth]

-> Running build hook: [filesystems]

==> 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-cachyos.img’

==> WARNING: errors were encountered during the build. The image may not be complete.

[Oli]

Sounds similar to this thread.

[cefk]

@cefk

Downgrading to the older kernels isn’t an option, for me (I always clear the cache), but it’s also not a great idea, imho, because the recent update likely fixes a vulnerability, that dropped, yesterday. And I don’t enjoy being vulnerable.

The reason may be the same, but I don’t think, the proposed solution is safe (see my quote from above).

I’d rather wait for a new kernel-update, than risk a downgrade to a highly likely vulnerable kernel.

[Oli]

I just updated to 7.0.5-2 and there was no problem. That’s why I waited. BTW downgrading is not a problem because there is a temporary mitigation for the Dirty Frag vulnerability:

New Root Kernel Bug - Dirty Frag

Even the mainline Linux kernel itself doesn’t appear to have any patches, as one colleague of mine reported a successful trigger of the exploit on a CachyOS machine running kernel 7.0.3-1-cachyos, and also on an updated Arch box source: Devastating 'Dirty Frag' exploit leaks out, gives immediate root access on most Linux machines since 2017, no patches available, no warning given — Copy Fail-like vulnerability had its embargo broken | Tom's Hardware

[cefk]

Thanks for the info

I’m on the bore-lto kernel (latest is still 7.0.5-1).

But, I just remembered, that I also have the lts-kernel installed as fallback.
And the nvidia-580xx-dkms don’t install correctly (don’t work) for both kernels, at the moment.

If it were specific to the 7.0.5-1 kernels, I’d assume, the lts-version should’ve worked, am I correct?
Or am I missing something?

[Oli]

@cefk

If it were specific to the 7.0.5-1 kernels, I’d assume, the lts-version should’ve worked, am I correct?

Partly, because it doesn’t mean other versions won’t work either. I updated from 6.19.x-x to 7.0.5-2 and both versions worked. You could just install 7.0.5-2 and use it until bore-lto is updated.

[cefk]

At first, I forgot to install the headers, but when I did, it worked

❯ uname -a
Linux d1 7.0.5-2-cachyos #1 SMP PREEMPT_DYNAMIC Sat, 09 May 2026 11:29:28 +0000 x86_64 GNU/Linux

❯ nvidia-smi
Sun May 10 09:10:15 2026
±----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.159.03 Driver Version: 580.159.03 CUDA Version: 13.0 |
±----------------------------------------±-----------------------±---------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce GTX 1060 6GB Off | 00000000:01:00.0 On | N/A |
| 33% 37C P8 11W / 120W | 368MiB / 6144MiB | 0% Default |
| | | N/A |
±----------------------------------------±-----------------------±---------------------+

Thank’s a lot for your help, Oli