[Stable Update] 2023-10-04 - Kernels, Systemd, LibreOffice, NVIDIA, Mesa, GNOME, AMDVLK

This problem is still present. Apparently it’s a kernel 6.5.x issue. Using the LTS kernel 6.1.55 there’s no issue.

Notice for qemu users: QEMU version 8.1.1-1 has a regression

I’m holding off on upgrading my system, but am interested if anyone has already upgraded & encountered this problem. If you’re experiencing issues, feel free to post here

After the reboot my system wouldn’t mount an ntfs drive, dolphin said something about bad parameters or even bad disk.
I booted windows 10 (dual boot) and did a scan on the disk, it said there were errors that had to be corrected, ran it again and found no errors, booted manjaro and the drive was correctly mounted. It was working before the update.

Also I had to rebuild input-remapper from the aur because of python although the error I got from it referenced python3.11 instead of 3.10

After updating and rebooting, my Plasma session was once again corrupted. It did not fully load, and I was left with only one visible panel, which was completely empty and transparent, but without the normal blur — this has been happening quite a few times over the last two months.

Luckily, I didn’t need to restore a backup this time. Logging into a tty, shutting down the GUI session with… :arrow_down:

sudo systemctl stop sddm

… and then deleting the contents of ~/.cache, followed by… :arrow_down:

sudo systemctl start sddm 

… gave me back my trusted desktop environment.

This does however seem more like a regression in Plasma — or perhaps in the KDE Frameworks — introduced with an earlier update. But as I said, it was easy to fix this time. :man_shrugging:

1 Like

I’ve not had any problems with KDE (X11) for many months now, and even back then it was minor issues like the mouse acceleration bug or Kate/Dolphin/Konsole forgetting their window size/position after an update. Maybe it’s because I don’t use themes or any other major customization beyond changing the wallpaper… getting too old to care about that stuff any more. :laughing:

I will have to reinstall the operating system because KDE became extremely slow and no matter how many times I try to reinstall packages, delete caches and so on, nothing works.

I also have 22 new packages of lib32 files, i wonder why mesa-demos/lib32-mesa-demos is installed at all.

Do we maybe need mesa-demos in future? i also checked depencys and no required/optional depencys at all.

As @MrLavender said, maybe its theme or widgets related? Anyways, i use customised Theme also and had no big issue in KDE X11 at all… on both of my systems, AMD only and Intel/Nvidia btw.

3 posts were split to a new topic: Audio device not detected in PulseAudio after update

After this update I was unable to start Gnome Terminal. The reason was a locale error: locale: Cannot set LC_ALL to default locale: No such file or directory. Upon further investigation, I found that for some strange reason the en_GB.UTF-8 UTF-8 line was commented out in the /etc/locale-gen file, and the de_DE.UTF-8 UTF-8 line was uncommented instead. I don’t know German and have no interest in this locale. Fixed this and ran the sudo locale-gen command to regenerate the locales.

2 Likes

Update partially broke notifications on my i3wm install. At least in a sense that numerous my custom scripts have relied on dunst daemon and dunstify command to handle notifications. Notifications from other apps were shown, but were themed differently than before the update.

I was able to determine that xfce4-notifyd service took over the notifications, and only one service may be used to handle the notifications. So the dunst service did not start properly.

I have stopped xfce4-notifyd and started dunst service:

systemctl stop --user xfce4-notifyd.service
systemctl start --user dunst.service
systemctl enable --user dunst.service

My attempt to simply disable xfce4-notifyd service failed (something about service file being a hardlink and/or having no installation instructions). So have I simply removed the package:

pamac remove xfce4-notifyd xfce4-power-manager

xfce4-power-manager has to go as well, because for some reason it explicitly depends on xfce4-notifyd (not on notification-daemon in general). So to remove xfce4 notification daemon, you need to remove xfce4 power manager. I am not using it either way, so this solution seems fine to me.

Hi there,
My USB3 HDD drive can’t be mounted after the update to kernel 6.1.55-1.
The message is:
“wrong fs type, bad option, bad superblock on /dev/sdx, missing codepage or helper program, or other error”

It works perfectly fine on kernel 5.4.257-1 and the previous version of kernel 6.1 though.

Thank you.

1 Like

During finishing the update yesterday evening, my machine automatically restarted (evtl. crashed, I don’t know). I was a bit worried, but after the restart everything seems to work well yesterday evening. Today morning also everything was fine first.

But then, when leaving the computer a few minutes alone and getting back, I noticed that I was logged out from Gnome. I was wondering, because I disabled this functionality.

After logging in again, my desktop settings were gone. All extensions are disabled.
Many programs are not starting anymore now.

When trying to start brave or chromium for example I get the error:

libva error: vaGetDriverNames() failed with unknown libva error

Not sure what to do now.

Seemed to go as smoothly as usual, but as the messages scrolled up my screen I noticed (but didn’t have time to make a note of) something about a syslinux BIOS setup script that I think needed to be run.

Does anyone know what this is and if it’s needed? I’m not even sure what syslinux is.

@thorsten79

1 Like

Sorry, this is absolutely not helping. You did not read my post as well. My kernel is running.

I was able to trigger the problem down to the gnome keyring. Somehow the behaviour changed after the update. I had to disable the keyring again , so that everything works again.

I had the same problem. The second, third and fourth disk, (ntfs) could not be mounted anymore in dolphin. The workaround was to install ntfs-automount (AUR).

I would think this is also related to my post above about mounting an encfs mount on a ntfs drive. Or at least, it would make logical sense the two are related. Basically something needed whitelisting in libfuse around NTFS and it looks like that was merged in so I would think shortly this would get fixed here

Wild guess here, but maybe it had to do with the default password hashing algorithm change? See the ‘2023-10-04’ part in the “Known Issues and Solutions” post right underneath the announcement.

I can imagine a keyring using encryption and probably some hashing as well. If it hashes its data using the default algorithm and after the Manjaro Stable update it then has a different hashing algorithm, then during validation check against the hashed data, it will fail that check (and probably give errors).
But like I said, wild guess…

So sad that still the newest nvidia driver keeps flickering and let us high refresh TFT user’s abandonded in the dust and this since April :face_with_spiral_eyes:

A post was split to a new topic: Issue with crowdsec after this up date