btw I retried installing nvidia-all on the latest updates. Got this
Error! Bad return status for module build on kernel: 6.13.4-zen1-1.1-zen (x86_64)
Consult /var/lib/dkms/nvidia/565.77/build/make.log for more information.
This driver is working fine booting in LTS though.
The previous setup with repo drivers in Wayland was just too unstable… all kinds of issues
This is old dkms stuff. You have installed 570.xxx driver. Ignore and delete all
“old” nvidia driver there.
Well, he don´t have a config/profile file. ( /home/user/.config/autorandr/) Reinstall autorandr → reboot can fix this or he make a own profile (see readme file) + all monitors must be connected.
But, (my mind) only inside x11 session then switch to wayland.
Of course, but acer laptops and support for linux and bios updates are a pain.
I reinstalled autorandr.
nvidia-all just updated to support 570 drivers; with that, it installs without error.
However, when I enter X11, the UI is super laggy and unusable (no freeze). When I enter Wayland, the HDMI output stops after being available for the login screen.
linux-lts + nvidia-all + 565 + X11 seems to be the only combination that works decent currently.
I don’t have it on my daily driver (Acer laptop/Intel/nVidia/multimonitor setup via HDMI), nor on my other machines (AMD/AMD). But I suspect that he has a profile that is faulty, that would explain the error message.
@Hanuman, if you have anything with autorandr in /home/user/.config/, then make a backup of it and then delete the autorandr profile. Then reboot and check if you can log into an x11 session without issues.
Or you can also create a new user as a test and check whether the same issues occur there.
There is no autorandr file there.
I doubt creating a new user would help; considering that even live boot has issues. I reverted to nvidia-all drivers for now; will try creating new user next time that I re-install the repo drivers.
Creating a new user does not help. Still freezes.
Buying a new laptop won’t be before summer either.
Also to note: when I tried Fedora, there were also GPU glitches with Wayland; and it didn’t allow me to revert back to X11.
okay another solution i found that works for me in the New Mokka version
is downloading the open source drivers apart of nvidia-all drivers. The following command needs to be repeated each time an update ruins it though, but works otherwise on my machine, also gaming.
sudo pacman -Sy nvidia-open-dkms egl-wayland lib32-nvidia-utils lib32-opencl-nvidia nvidia-settings opencl-nvidia nvidia-utils
if this did not work you may need to do this step as well, i did, but do not know if it is needed to fix the issue.
#open with you favourite text editor
/etc/dracut.conf.d/cmdline.conf
#add line
force_drivers+=" nvidia nvidia_modeset nvidia_uvm nvidia_drm "
#save as cmdline.conf and rebuild kernels with command below
sudo dracut-rebuild
See how you go.
If I install those, it replaces the other drivers. Can’t have both installed together.
When installing with nvidia-all v565, if I select YES to support older hardware, then it’s more stable with LTS kernel. Still not perfectly stable but a lot more stable.
Sure it does, but then it works, for me at least. (though for some reason i had to try twice before it compiled the dkms modules of all three kernels right, the second time around.)
If it does not work for you, then you can always reinstall the nvidia-all drivers, or even the original drivers. It is simply another option to try out. i have the Main, Zen and Lts kernel working on the above open source drivers. And everything has been working great!
In what order are you doing it to have both installed?
If I try your command, it gives this
:: Synchronizing package databases...
garuda is up to date
core-x86-64-v3 is up to date
core 116.2 KiB 106 KiB/s 00:01 [------------------------------------------] 100%
extra-x86-64-v3 1098.8 KiB 70.9 KiB/s 00:16 [------------------------------------------] 100%
extra 7.7 MiB 1279 KiB/s 00:06 [------------------------------------------] 100%
multilib-x86-64-v3 59.9 KiB 18.1 KiB/s 00:03 [------------------------------------------] 100%
multilib 131.7 KiB 83.0 KiB/s 00:02 [------------------------------------------] 100%
chaotic-aur 638.9 KiB 115 KiB/s 00:06 [------------------------------------------] 100%
error: failed retrieving file 'multilib-x86-64-v3.db' from mirror.sunred.org : Operation too slow. Less than 1 bytes/sec transferred the last 10 seconds
resolving dependencies...
looking for conflicting packages...
:: nvidia-open-dkms-570.124.04-1.1 and nvidia-dkms-tkg-565.77-263 are in conflict (NVIDIA-MODULE). Remove nvidia-dkms-tkg? [y/N] y
:: lib32-nvidia-utils-570.124.04-1.1 and lib32-nvidia-utils-tkg-565.77-263 are in conflict (lib32-nvidia-libgl). Remove lib32-nvidia-utils-tkg? [y/N] y
:: nvidia-utils-570.124.04-1.1 and nvidia-utils-tkg-565.77-263 are in conflict (nvidia-libgl). Remove nvidia-utils-tkg? [y/N] y
:: lib32-opencl-nvidia-570.124.04-1.1 and lib32-opencl-nvidia-tkg-565.77-263 are in conflict. Remove lib32-opencl-nvidia-tkg? [y/N] y
:: egl-wayland-4:1.1.18-1.1 and nvidia-egl-wayland-tkg-565.77-263 are in conflict. Remove nvidia-egl-wayland-tkg? [y/N] y
:: nvidia-settings-570.124.04-1.1 and nvidia-settings-tkg-565.77-263 are in conflict. Remove nvidia-settings-tkg? [y/N] y
:: opencl-nvidia-570.124.04-1.1 and opencl-nvidia-tkg-565.77-263 are in conflict. Remove opencl-nvidia-tkg? [y/N] y
warning: dependency cycle detected:
warning: nvidia-utils will be installed before its egl-wayland dependency
Package (16) Old Version New Version Net Change Download Size
extra-x86-64-v3/egl-x11 1.0.0-1.1 0.15 MiB
lib32-nvidia-utils-tkg 565.77-263 -224.01 MiB
lib32-opencl-nvidia-tkg 565.77-263 -19.80 MiB
extra-x86-64-v3/libxnvctrl 570.124.04-1.1 0.45 MiB 0.07 MiB
nvidia-dkms-tkg 565.77-263 -85.14 MiB
nvidia-egl-wayland-tkg 565.77-263 -0.15 MiB
nvidia-settings-tkg 565.77-263 -1.72 MiB
nvidia-utils-tkg 565.77-263 -689.09 MiB
opencl-nvidia-tkg 565.77-263 -42.03 MiB
extra-x86-64-v3/egl-wayland 4:1.1.18-1.1 0.10 MiB 0.04 MiB
multilib-x86-64-v3/lib32-nvidia-utils 570.124.04-1.1 246.20 MiB 49.54 MiB
multilib-x86-64-v3/lib32-opencl-nvidia 570.124.04-1.1 20.20 MiB 2.66 MiB
extra-x86-64-v3/nvidia-open-dkms 570.124.04-1.1 100.28 MiB 9.97 MiB
extra-x86-64-v3/nvidia-settings 570.124.04-1.1 1.55 MiB 0.76 MiB
extra-x86-64-v3/nvidia-utils 570.124.04-1.1 789.03 MiB 268.93 MiB
extra-x86-64-v3/opencl-nvidia 570.124.04-1.1 62.54 MiB 5.28 MiB
Total Download Size: 337.24 MiB
Total Installed Size: 1220.50 MiB
Net Upgrade Size: 158.56 MiB
:: Proceed with installation? [Y/n]
yes i delete nvidia-utils, install the drivers, and then reinstall nvidia-utils
all fixed, for me at least.
see how you go. If it stuffs up redo your nvidia-all driver that worked, but i think it will work, for i had the same issues as you. More often than not by simply removing the offending software, installing what you want to install, and then reinstall what you had to uninstall, will fix the problem. However sometimes more than one app needs to go, then it can become more problematic, to be able to delete offending software.
(You may need to have to install twice, before it will compile the dkms of the zen kernel rightly. NMainn-line, and Lts do fine first try. But for some reason, at times, it hangs the first round. Simply kill the pacman lock, through garuda-rani, maintenance page, that is if the zen’ kernel dkms hangs, and then re-run the driver command, that fixed it both times it happened with me. (i did a reinstall because i tried out another distro, and need the drive Mokka was on, but still also wanted to have a closer look at Mokka yet, and i had this issue twice, after an update.)
EDIT:
Not sure if you tried this line below, if you did you do not need this line, or at least i do not, for it throws up an compile error i saw. Maybe check it while compiling if you did at it.
open with you favourite text editor
/etc/dracut.conf.d/cmdline.conf
#add line
force_drivers+=" nvidia nvidia_modeset nvidia_uvm nvidia_drm "
#save as cmdline.conf and rebuild kernels with command below
sudo dracut-rebuild
Just to be clear, you run that command
sudo pacman -Sy nvidia-open-dkms egl-wayland lib32-nvidia-utils lib32-opencl-nvidia nvidia-settings opencl-nvidia nvidia-utils
And then re-install nvidia-all v570? (and not 565 as I currently have?) in which case the issue I have is freeze after login
Your sequence of commands gives this when running nvidia-all
Package (12) Old Version New Version Net Change
lib32-nvidia-utils 570.124.04-1.1 -246.20 MiB
lib32-opencl-nvidia 570.124.04-1.1 -20.20 MiB
nvidia-open-dkms 570.124.04-1.1 -100.28 MiB
nvidia-settings 570.124.04-1.1 -1.55 MiB
nvidia-utils 570.124.04-1.1 -789.03 MiB
opencl-nvidia 570.124.04-1.1 -62.54 MiB
lib32-nvidia-utils-tkg 570.124.04-263 246.30 MiB
lib32-opencl-nvidia-tkg 570.124.04-263 20.20 MiB
nvidia-dkms-tkg 570.124.04-263 103.76 MiB
nvidia-settings-tkg 570.124.04-263 1.72 MiB
nvidia-utils-tkg 570.124.04-263 789.15 MiB
opencl-nvidia-tkg 570.124.04-263 62.54 MiB
Total Installed Size: 1223.67 MiB
Net Upgrade Size: 3.88 MiB
:: Proceed with installation? [Y/n]
(6/6) checking keys in keyring [--------------------------------------------------------------------] 100%
(6/6) checking package integrity [--------------------------------------------------------------------] 100%
(6/6) loading package files [--------------------------------------------------------------------] 100%
(6/6) checking for file conflicts [--------------------------------------------------------------------] 100%
error: failed to commit transaction (conflicting files)
nvidia-utils-tkg: /usr/lib/libnvidia-egl-xcb.so exists in filesystem (owned by egl-x11)
nvidia-utils-tkg: /usr/lib/libnvidia-egl-xcb.so.1 exists in filesystem (owned by egl-x11)
nvidia-utils-tkg: /usr/lib/libnvidia-egl-xcb.so.1.0.0 exists in filesystem (owned by egl-x11)
nvidia-utils-tkg: /usr/lib/libnvidia-egl-xlib.so exists in filesystem (owned by egl-x11)
nvidia-utils-tkg: /usr/lib/libnvidia-egl-xlib.so.1 exists in filesystem (owned by egl-x11)
nvidia-utils-tkg: /usr/lib/libnvidia-egl-xlib.so.1.0.0 exists in filesystem (owned by egl-x11)
nvidia-utils-tkg: /usr/share/egl/egl_external_platform.d/20_nvidia_xcb.json exists in filesystem (owned by egl-x11)
nvidia-utils-tkg: /usr/share/egl/egl_external_platform.d/20_nvidia_xlib.json exists in filesystem (owned by egl-x11)
Errors occurred, no packages were upgraded.
==> WARNING: Failed to install built package(s).
-> exit cleanup done
So I run this command
pacman -Rcns nvidia-utils
Tried installing with “support for older hardware”.
Setup completed with this error
(12/15) Update Nvidia modules in initcpio
dracut[I]: Executing: /usr/bin/dracut --kver=6.12.18-1.1-lts --force
dracut[F]: Can't write to /boot/efi/6b052c11c34a4b0a81ea771114698838/6.12.18-1.1-lts: Directory /boot/efi/6b052c11c34a4b0a81ea771114698838/6.12.18-1.1-lts does not exist or is not accessible.
dracut[I]: Executing: /usr/bin/dracut --kver=6.13.6-zen1-1.1-zen --force
dracut[F]: Can't write to /boot/efi/6b052c11c34a4b0a81ea771114698838/6.13.6-zen1-1.1-zen: Directory /boot/efi/6b052c11c34a4b0a81ea771114698838/6.13.6-zen1-1.1-zen does not exist or is not accessible.
error: command failed to execute correctly
In spite of the error, it still booted with HDMI working.
Login did not cause a freeze this time, but the UI was extremely laggy. Since it’s not a full freeze, the log might contain something this time.
No i think you miss understood me. i run
```
sudo pacman -R nvidia-utils
#and you should also do
sudo pacman -R egl-x11
#and then
sudo pacman -Sy nvidia-open-dkms egl-wayland lib32-nvidia-utils lib32-opencl-nvidia nvidia-settings opencl-nvidia nvidia-utils
sudo pacman -S nvidia-utils egl-x11
#and that should be it!
However if it does not work, eg, it does not complete installing the drivers without fatal errors, then reinstall the nvidia-all driver that worked on at least on one of your kernels.
#Open source will not install drivers for nvidia cards that are too old. i think, just by memory, from GTX 1060 upwards, but i could have this wrong.
The journal is clean, no nvidia/nvrm/drm errors, nothing else that would worry me.
Maybe this has a completely different cause. Open htop and check CPU/MEM usage and under I/O tab DISK READ/WRITE.
The only candidate I notice in the journal is baloo.
If you have “File indexing” enabled, disable this (System Settings → Search → File Search) or via terminal:
balooctl6 disable
You can also try to use the default Breeze theme and check if you also have lags there. Maybe it’s just a theme/widget issue causing the lags.
When I tried Fedora, I had that same extreme lag issue with X11; it only worked with Wayland, but wasn’t stable either.
What happens if you disconnect the external monitor/TV (is a known issue with nvidia + HDMI), does it get better then?
Check and test this anyway: