Blackscreen after fresh install, known fix, I just cant seem to repeat it

Im having the same issue as this person, but they seemed to solve it where as I have been having issues or needing more guidance:

What I have done:
Attempted an install on my pc, hardware as follows:
cpu i7-12700k
gpu RTX4090
32gb RAM
installing on external SSD

I installed with Grub since I read it works better as a dual boot with windows
Installer says everything went fine, I restart, load into grub, and then after that I just get the same blackscreen and underscore seen on the linked post.

My attempts to fix it:
boot into the live usb, mount the drive I installed the OS to
chroot into the drive’s root
change user to my user account on the drive (not root, since nvidia-inst was telling me I cant run it as root)
but now that I am using my user account I can run nvidia-inst it just keeps telling me my root password is wrong (im positive Im typing it correctly as it is the same as my user’s password)

[user@EndeavourOS /]$ nvidia-inst
lspci: Unable to load libkmod resources: error -2
2025-05-28 15:19:50: Info: nvidia-inst version 25.5-1
2025-05-28 15:19:50: Info: Command line: nvidia-inst 
2025-05-28 15:19:50: Info: Selected mode: nvidia (Nvidia's open source)
2025-05-28 15:19:52: Info: Installing packages: dkms nvidia-open-dkms nvidia-utils nvidia-settings nvidia-hook
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
COMMANDS TO RUN:
    pacman -Syuq --noconfirm --noprogressbar --needed dkms nvidia-open-dkms nvidia-utils nvidia-settings nvidia-hook
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

==> NOTE: running the commands may take several minutes...

Root Password: 
su: Authentication failure

Someone please help me figure out what Im doing wrong

Have you tried using sudo instead of su ?

If you are in arch-chroot (or root), directly run the command that nvidia-inst showed under
COMMANDS TO RUN:

pacman -Syuq --noconfirm --noprogressbar --needed dkms nvidia-open-dkms nvidia-utils nvidia-settings nvidia-hook

So, I came back to this after dinner, only now when I boot into the live USB and chroot into the os drive, nvidia-inst is telling me “could not find Nvidia GPU”

[user@EndeavourOS /]$ nvidia-inst
pcilib: Cannot open /proc/bus/pci
lspci: Cannot find any working access method.
/usr/bin/nvidia-inst: line 618: /dev/null: Permission denied
2025-05-28 20:18:17: Info: nvidia-inst version 25.5-1
2025-05-28 20:18:17: Info: Command line: nvidia-inst 
2025-05-28 20:18:17: Warning: could not find Nvidia GPU.

And if I run that command it told me to before, it essentially complains a lot about certain paths not existing or being mounted, I can paste the entire output if you think that would help but its long.

Please do. :wink:
Logs usually are very useful.

As the trend with “picking it up the next day” continues, after booting into the live usb and chroot’ing to the OS drive, I run the same command as before

pacman -Syuq --noconfirm --noprogressbar --needed dkms nvidia-open-dkms nvidia-utils nvidia-settings nvidia-hook

and I get this output, its asking for confirmation to delete a corrupted package, but it doesn’t actually wait for a response.

[root@EndeavourOS /]# sudo pacman -Syuq --noconfirm --noprogressbar --needed dkms nvidia-open-dkms nvidia-utils nvidia-settings nvidia-hook
:: Synchronizing package databases...
 endeavouros downloading...
 core downloading...
 extra downloading...
 multilib downloading...
:: Starting full system upgrade...
resolving dependencies...
looking for conflicting packages...
warning: dependency cycle detected:
warning: eglexternalplatform will be installed before its nvidia-utils dependency

Package (10)               New Version   Net Change  Download Size

extra/egl-gbm              1.1.2.1-1       0.03 MiB       0.01 MiB
extra/egl-wayland          4:1.1.19-1      0.10 MiB       0.04 MiB
extra/egl-x11              1.0.1-1         0.14 MiB       0.04 MiB
extra/eglexternalplatform  1.2.1-1         0.02 MiB       0.01 MiB
extra/libxnvctrl           570.153.02-1    0.45 MiB       0.07 MiB
extra/dkms                 3.2.1-1         0.15 MiB       0.05 MiB
endeavouros/nvidia-hook    1.5.2-1         0.00 MiB       0.01 MiB
extra/nvidia-open-dkms     570.153.02-1  100.42 MiB       9.99 MiB
extra/nvidia-settings      570.153.02-1    1.53 MiB       0.76 MiB
extra/nvidia-utils         570.153.02-1  789.54 MiB     269.12 MiB

Total Download Size:   280.09 MiB
Total Installed Size:  892.37 MiB

:: Proceed with installation? [Y/n] 
:: Retrieving packages...
 nvidia-utils-570.153.02-1-x86_64 downloading...
 nvidia-open-dkms-570.153.02-1-x86_64 downloading...
 nvidia-settings-570.153.02-1-x86_64 downloading...
 libxnvctrl-570.153.02-1-x86_64 downloading...
 dkms-3.2.1-1-any downloading...
 egl-x11-1.0.1-1-x86_64 downloading...
 egl-wayland-4:1.1.19-1-x86_64 downloading...
 nvidia-hook-1.5.2-1-any downloading...
 egl-gbm-1.1.2.1-1-x86_64 downloading...
 eglexternalplatform-1.2.1-1-any downloading...
checking keyring...
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
checking package integrity...
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: GPGME error: Invalid crypto engine
error: dkms: missing required signature
:: File /var/cache/pacman/pkg/dkms-3.2.1-1-any.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: nvidia-open-dkms: missing required signature
:: File /var/cache/pacman/pkg/nvidia-open-dkms-570.153.02-1-x86_64.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: eglexternalplatform: missing required signature
:: File /var/cache/pacman/pkg/eglexternalplatform-1.2.1-1-any.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: egl-wayland: missing required signature
:: File /var/cache/pacman/pkg/egl-wayland-4:1.1.19-1-x86_64.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: egl-gbm: missing required signature
:: File /var/cache/pacman/pkg/egl-gbm-1.1.2.1-1-x86_64.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: egl-x11: missing required signature
:: File /var/cache/pacman/pkg/egl-x11-1.0.1-1-x86_64.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: nvidia-utils: missing required signature
:: File /var/cache/pacman/pkg/nvidia-utils-570.153.02-1-x86_64.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: libxnvctrl: missing required signature
:: File /var/cache/pacman/pkg/libxnvctrl-570.153.02-1-x86_64.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: nvidia-settings: missing required signature
:: File /var/cache/pacman/pkg/nvidia-settings-570.153.02-1-x86_64.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: nvidia-hook: missing required signature
:: File /var/cache/pacman/pkg/nvidia-hook-1.5.2-1-any.pkg.tar.zst is corrupted (invalid or corrupted package (PGP signature)).
Do you want to delete it? [Y/n] 
error: failed to commit transaction (invalid or corrupted package (PGP signature))
Errors occurred, no packages were upgraded.
[root@EndeavourOS /]#

(same output if i run with sudo)

running just nvidia-inst as my user is still giving the “no nvidia gpu found” error though

[user@EndeavourOS /]$ nvidia-inst
pcilib: Cannot open /proc/bus/pci
lspci: Cannot find any working access method.
/usr/bin/nvidia-inst: line 618: /dev/null: Permission denied
2025-05-29 08:32:23: Info: nvidia-inst version 25.5-1
2025-05-29 08:32:23: Info: Command line: nvidia-inst 
2025-05-29 08:32:23: Warning: could not find Nvidia GPU.

Bumping, I tried doing a fresh install just in case i might have done something wrong, but still the same issue with the blackscreen and “could not find Nvidia GPU” from nvidia-inst

What I am doing:
Boot to live usb, start install, select my options, time, keyboard, etc, select KDE desktop, leave all packages default, wipe my entire (external) ssd to make room for a boot partition, swap, and primary endervour OS partition (installer is doing this part itself) I then install and boot to the drive, after it finishes booting, black screen with a non-blinking underscore.

I boot back into the live USB, mount OS drive, chroot to the drive, swap user from root to my actual user on the OS drive, run nvidia-inst and get the same errors:

[user@EndeavourOS /]$ nvidia-inst
pcilib: Cannot open /proc/bus/pci
lspci: Cannot find any working access method.
/usr/bin/nvidia-inst: line 618: /dev/null: Permission denied
2025-05-29 08:32:23: Info: nvidia-inst version 25.5-1
2025-05-29 08:32:23: Info: Command line: nvidia-inst 
2025-05-29 08:32:23: Warning: could not find Nvidia GPU.

I really dont know where to go from here

Figured it out.

Did a fresh install, chose not to install a desktop env.

Booted to terminal after installing, nvidia-inst worked perfectly, then I went about manually installing the desktop environment and it’s up and running. In case anyone else is stuck on this in the future.

3 Likes

Nice work!

BTW, which boot menu entry did you choose when you started to install?

Grub, sorry, forgot to include that. I dont have a compelling reason why other than its what I know from other distros, and system.d kept failing from some “read only efi” error

No, I didn’t mean which bootloader (grub or systemd-boot) you selected.

I meant which menu entry did you select (the very first “page”) right after you booted the USB installer? The Default or NVIDIA or something else?

Oh, sorry, when I tried both Nvidia and default, both resulted in the same black screen issue, but on this specific run of installation where i chose no desktop env, I chose the NVidia installation just because i figured it was probably installing other stuff i might need for nvidia cards

There’s this issue with the mesa package (discussed here also).
Maybe that was the culprit for your system too?

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.