From your post I’ve learned about those meta packages, seems they are not really necessary, I don’t have them installed.
Those meta packages get automatically installed as soon as you have an EOL kernel and we dropped it. For now it installs linux610 and will switch to linux611 when we drop 6.10 series.
4 Likes
LANG=en; sudo mhwd-kernel -li
[sudo] password for andreas:
Currently running: 6.11.2-4-MANJARO (linux611)
The following kernels are installed in your system:
* linux515
* linux61
* linux610
* linux611
* linux612
I only tried to remove linux610 ahead of time 
(linux66 LTS did not work with my dual seat setup)
Currently you can’t but you can leave a comment like I did here: Find a proper solution to automate EOL kernel removal via ALPM (#13) · Issues · Release plan / Packaging · GitLab. Also you missed to remove 6.9 ahead of time, otherwise you wouldn’t have that meta package installed to begin with …
Yes that is true.
But there were 4 other kernels installed at this time so i saw no harm in removing it later
Should i move my posts ?
you need to active thes services ( only for Wayland or Gnome )
Enabling nvidia-resume...
Created symlink '/etc/systemd/system/systemd-suspend.service.wants/nvidia-resume.service' → '/usr/lib/systemd/system/nvidia-resume.service'.
Created symlink '/etc/systemd/system/systemd-hibernate.service.wants/nvidia-resume.service' → '/usr/lib/systemd/system/nvidia-resume.service'.
Enabling nvidia-hibernate...
Created symlink '/etc/systemd/system/systemd-hibernate.service.wants/nvidia-hibernate.service' → '/usr/lib/systemd/system/nvidia-hibernate.service'.
Enabling nvidia-suspend...
Created symlink '/etc/systemd/system/systemd-suspend.service.wants/nvidia-suspend.service' → '/usr/lib/systemd/system/nvidia-suspend.service'.also in case wayland & nvidia driver you have to add
‘nvidia_drm.fbdev=1’
on boot kernel ( KDE and Wayland only ) if you have a black screen
Those services are already activated, and you can’t use nvidia_drm.fbdev=1 with Kernel 6.11, which I think is part of the problem.
https://forums.developer.nvidia.com/t/drm-fbdev-wayland-presentation-support-with-linux-kernel-6-11-and-above/307920
Although it seems Manjaro has the most recent patch in nvidia-utils here: https://gitlab.manjaro.org/packages/extra/nvidia-utils/-/blob/master/6.11-fbdev.patch?ref_type=heads
sddm login problems/weirdness with multiple monitors on plasma:
I think part of the nvidia suspend/resume problem may be the drm differences from running:
$ cat /sys/class/graphics/fb0/name
Kernel 6.6: nvidia-drmdrmfb
Kernel 6.11: simpledrmdrmfb
For amdgpu users experiencing hangs or freezes, especially after suspend, beginning with kernel 6.11+;
This boot option ( amdgpu.dcdebugmask=0x200 ) seems to work so far.
The option disables PSR-SU, sometimes called PSR 2.0.
I have not experienced a hang after sleep since applying.
1 Like
Not anymore, it’s now default.
@Yochanan if it’s enabled, shouldn’t cat /sys/class/graphics/fb0/name show nvidia-drmdrmfb and not simpledrmdrmfb?
I’m trying to understand if this indeed broken on my system when using kernel 6.11.
Apparently. It does for me.
Mind you, I’m using nvidia-open-dkms, however I don’t see how that would make a difference.
So the open source driver, e.g. sudo mhwd -i pci video-linux?
Is there anything I have to run to make sure my driver/kernel are indeed using the latest patches etc?
with Xfce - X11 and nvidia 560.35.03-10
and no option ‘nvidia_drm.fbdev=1’
cat /sys/class/graphics/fb0/name
EFI VGAthis not good , i use a DVI link
No. That will install nouveau, the reverse engineered open source driver. We don’t yet have support in MHWD for the open kernel module. One must manually install nvidia-open-dkms. It’s only available on the unstable branch with the 560.XX series (New Feature Branch). Our stable and testing branches are currently using the 550.XX series (Production Branch).
I’m on unstable, I’ll give it a go.
Edit: that didn’t work out well. My two other monitors didn’t come on and performance was really bad, rolled back.
Oh well 
Please create a new Graphics & Display for further discussion and troubleshooting.
Also, please test nvidia-dkms vs. linux611-nvidia to see if you notice any difference.
1 Like
Hi,
I’m back on unstable with XFCE, linux611 and linux611-nvidia 560.35.03-16.
Resume from suspend work if :
systemctl enable nvidia-suspend.service
systemctl enable nvidia-resume.service
Reboot and test → OK.
However, if the user session was opened and locked when suspend, after resume on lightdm-gtk-greeter when I unlock the user session, I have a black screen with mouse cursor.
Then it need ctrl+alt+F2 and login to CLI to do systemctl restart lightdm.service.
Then it’s possible to open the user session but work opened before suspend is gone because of lightdm restart.
The only workaround I found is to proper close the user session, then suspend. When resume login work as expected.
Wish You well
Guys, what I currently don’t like about Nvidia drivers is the following:
- you need to know if you have to use the open kernel one or not
- you might need to enable systemd services to make it work (this is now done via alpm postinstall scripts)
- you have to know if you enable or disable the GSP firmware
I have a fairly new 4060 mobile GPU and that driver drives me nuts. 550 is the last proper old style driver Nvidia will release. It already has linux 6.12 support, what the newer drivers don’t have. If you look how hard Arch is trying to get this driver working.
So we should stop expecting that the user is willing to tweak and setup the graphics drivers on his own. That we did in the past for the user via mhwd and other tools. Now I see the Arch-way trend more popping up, which is not what we originally aimed for.
Some users might get it going and done when deep diving into that topic, but that are only few. Sure we give some hints to do this or that, but if you simply update your system or install it fresh you expect it to simple working without the need of using the terminal.
So for now all users might still use the proprietary drivers. Nvidia recommends the open-kernel ones for a lot of their newer hardware. So if the proprietary drivers stop supporting those newer hardware it is harder for us getting those adopted to the new ones. There are no compiled kernel modules for the open-kernel versions yet.
To quote some of that article:
As mentioned earlier, if you’re unsure which driver to pick for the GPUs in your system, NVIDIA created a helper script to guide you through the selection process.
To use it, first install the
nvidia-driver-assistantpackage with your package manager, then run the script:
$ nvidia-driver-assistant
I mentioned it before when this article hit the web. nvidia-driver-assistant seems to be part of some package we don’t ship yet or is somewhere in the installer run file or only distributed for Ubuntu and Suse.
I see unstable branch as a stage branch for experiments and builds, which now get widely used by our users too. Those who want the latest packages. However, we have to dig into that Nvidia driver problem more deeply and together who are willing to test, willing to break their systems on purpose to get it done properly.
In the end a simple sudo pacman -Syu should be enough to get newer drivers properly working. No hints, no expectations the user may administrate his machine - anything.
I see a direction to get there, however it might be still a lot of work. Nvidia feature drivers are not so great in sense of stability. Sure Nvidia is now more open and discuss those issue now more openly: 560 release feedback & discussion - Linux - NVIDIA Developer Forums
So as long as we don’t do our proper homework as distro maintainers I won’t ship newer Nvidia drivers on stable nor testing branches!
10 Likes