[Stable Update] 2024-11-30 - Kernels, Plasma, GNOME, COSMIC, LXQT, SYSTEMD

ddns · 1 December 2024 08:31

I only have one issue, and that is that in the Supertuxkart game the music sounds bad.
It only happens with Supertuxkart, the rest of the applications sound good.

Maybe it is necessary to rebuild the package. Pacman says it was created on 01/16/2023.

Photon · 1 December 2024 09:05

That’s possible, but I think that updating the cache has the same effect and doesn’t require creating a symlink that belongs to no package.

menerond · 1 December 2024 09:09

Unfortunately this did not fix it but produced now another error message:

Gdk-Message: 10:05:28.519: Error 71 (Protocol error) dispatching to Wayland display.

I found this thread: [SOLVED] (GNOME 47) GTK 4.16.1 generates Wayland protocol error / [testing] Repo Forum / Arch Linux Forums and indeed setting

GSK_RENDERER=ngl

in

/etc/environment

fixes the issue and dino starts. Although the mouse cursor size is not anymore correctly recognized and it becomes gigantic.

Thanks for all the efforts!

EDIT: Also needed to log out and log in again for it to work. Just in case someone wants to reproduce.

Teo · 1 December 2024 09:14

What exactly do you mean by “updating the cache”? pamac -force-refresh or what? Cause pacman -Syu updates the cache?

Photon · 1 December 2024 09:17

No, sorry for the confusion, I meant running gdk-pixbuf-query-loaders --update-cache. The solution has been added to the list of known problems with this update in this thread’s second post as well.

fonic · 1 December 2024 09:19

After the update, when resuming from sleep/suspend or hibernation, the displays (two in my case) stay switched off.

They can be revived by switching to a different virtual terminal and back, e.g. CTRL+ALT+F3 → displays turn on, CTRL+ALT+F2 → back to KDE Plasma’s lock screen.

What’s interesting about this:
Isn’t the hook in /usr/lib/systemd/system-sleep/nvidia (which calls /usr/bin/nvidia-sleep.sh) supposed to automate that exact same workaround?

I noticed a couple of errors in both dmesg and /var/log/Xorg.0.log that seem to coincide:

[11021.049422] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: Shader Program Header 11 Error
[11021.049517] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: Shader Program Header 18 Error
[11021.049606] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: ESR 0x405840=0xa2040800
[11021.049697] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: ESR 0x405848=0x80000000
[11021.050063] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: ChID 00c0, Class 0000c597, Offset 00000000, Data 00000000
[11026.188365] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: Shader Program Header 18 Error
[11026.188457] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: ESR 0x405840=0x82040000
[11026.188547] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: ESR 0x405848=0x80000000
[11026.188897] NVRM: Xid (PCI:0000:26:00): 13, pid='<unknown>', name=<unknown>, Graphics Exception: ChID 0038, Class 0000c597, Offset 00000000, Data 00000000
<many more of the same nature>

[  9555.449] (EE) NVIDIA(0): The NVIDIA X driver has encountered an error; attempting to
[  9555.470] (II) NVIDIA(0): Error recovery was successful.
[ 11049.636] (EE) NVIDIA(0): The NVIDIA X driver has encountered an error; attempting to
[ 11049.660] (II) NVIDIA(0): Error recovery was successful.

Man, I’m really starting to hate NVIDIA on Linux…

EDIT:
Fixed this by applying a few modfications to hook /usr/lib/systemd/system-sleep/nvidia.

Photon · 1 December 2024 11:14

@Rick and anybody interested: I reported the issue at LightDM, please give it an upvote, if you agree!

github.com/canonical/lightdm

Please improve the error handling (catch possible coredumps)

opened 11:10AM - 01 Dec 24 UTC

Photon89

Hi guys, yesterday I spent almost the whole day trying to figure out what cau…sed a LightDM crash. The crash put the system into an unusable state (a loop of black screens with no TTY access). There are several components to it: 1. When LightDM crashes, it is not obvious that it is LightDM that is crashed. It could be a video driver problem as well. 2. When the system is not accessible, it is hard to track down the issue. 3. LightDM doesn't provide a useful error output, only a rather cryptic stack trace in the journal. In my case, the original issue was that libpixbufloader-svg.so hasn't been found. Instead of dying, LightDM could - Offer limited functionality (without fancy graphics) and/or - Show some error message (graphically or at least in CLI output) which helps to find the issue In the end I could only solve the issue by installing SDDM, logging into Xfce, seeing that other apps crash as well and finding meaningful error messages in their output. Here is what was in the journal: https://pastebin.com/i6QgFkBg As you can see, no mention of the missing .so file at all. This made things really tough to investigate!

callegar · 1 December 2024 11:30

Tragic update for the Zephyrus ROG G14 2022.

supergfxctl is now broken and cannot find dGPU. This happens with all the testable kernels (6.6.x, 6.11.x, 6.12.1). Apparently the dGPU works: if I attach an external screen to the HDMI port that is connected only via the dGPU and I do DRI_PRIME=1 glxinfo the dGPU starts and the external screen is switched on.
most serious issue. All the recent kernels are broken on this laptop and cannot resume from hibernation (they freeze). I already experienced this with the past stable update, where I had to stay on 6.10 because 6.11 could not resume from hibernation. Now there is no 6.10. So no hibernation. Very unfortunate on this laptop that only sleeps to idle and where the battery in suspend mode does not last more than a few hours.

Pawef · 1 December 2024 12:23

Can i use for fresh installation system?

nikgnomic · 1 December 2024 17:41

2 posts were split to a new topic: After updating and rebooting, the NVIDIA driver failed to load

Riquez · 1 December 2024 12:47

EDIT: Problem also exists on another distro, so it must be server or network related.

Im having trouble copying files over home network after the update.
(Gnome) I get very slow transfer speed 200kb & then it times-out. Cannot get anything above 100Mb to complete.

The set up is Manjaro Gnome laptop. wi-fi. copy files to a Raspberry Pi (raspbian) with a 1TB USB drive. via SMB

I use it daily, so the change to this normally reliable process was noticed immediately after the update. Everything was rebooted etc.
Havent digged deeper yet

acca85 · 1 December 2024 13:14

Pacman crashed while updating, I had no activity on the pc so I rebooted, I had to recover the system (thanks to the guys in the forum).
Seems like something got updated, something not, Firefox says it’s at Version 133 but now a few applications don’t start and who knows what else.
If I try to install the update again I get

there is nothing to do

How can I force a reinstall of everything that was supposed to get updated?

GkhnG · 1 December 2024 13:18

Replace git packages with their normal counterparts as much as possible. This is how I solved the same problem

Porru · 1 December 2024 13:32

I’m facing this conflict:

:: manjaro-gnome-extension-settings-20240920-2 and gnome-shell-extension-custom-accent-colors-11-2 are in conflict. Remove gnome-shell-extension-custom-accent-colors? [b/E] b
errorea:transakzioa prestatzean huts egin du (ezin dira menpekotasunak bete)
:: installing libjxl (0.11.1-1) breaks dependency 'libjxl.so=0.10-64' required by ffmpeg6.1
:: installing rubberband (4.0.0-1) breaks dependency 'librubberband.so=2-64' required by ffmpeg6.1
:: installing x265 (4.0-1) breaks dependency 'libx265.so=209-64' required by ffmpeg6.1

Trying to remove any of them won’t work due to dependencies.

philm · 1 December 2024 14:21

You can consult the /var/log/pacman.log to see which packages to updated, create a list out of it and reinstall those. Another option would be: Mapare - manjaro package restore

Rick · 1 December 2024 14:22

Thanks for the tip! I will do it.

philm · 1 December 2024 14:23

You can remove gnome-shell-extension-custom-accent-colors as GNOME 47 is supporting accent colors now. ffmpeg6.1 is an AUR package. You may want to remove it before you update your system. Install it later if needed, again.

nd3i · 1 December 2024 14:32

I’m seeing the same thing: conflict building v86d, then gibberish output. How can I check (& resolve) the conflict?

FreddyX · 1 December 2024 14:38

Thanks for the info.

btw, same problem on TuxedoOS (plasma 6.23, kernel 6.11). I now use Intel with Wayland and Nvidia with X11. That seems to me to be the best solution at the moment.

philm · 1 December 2024 14:42

@neoxpert, @vostmarhk, @purplemunchies, @Marsa, @FreddyX, @fonic, @Q.Singularity

State which exact nvidia packages you have installed via pacman -Q | grep nvidia and let me know if the latest packages work for you now on 6.12 kernel series.