You may need to rebuild any Python packages you’ve installed from the AUR. To get a list of them, you can run:
pacman -Qoq /usr/lib/python3.9
And to rebuild them all at once with an AUR helper such as yay, you can do:
yay -S $(pacman -Qoq /usr/lib/python3.9) --answerclean All
But if any of the packages don’t work with Python 3.10 yet, this might fail halfway through and you’ll have to do rebuild the remaining ones one or a few at a time.
Re-enable any per-user enabled pipewire-media-session.service manually:
You should re-enable any per-user pipewire-media-session.service manually to establish a new service alias for pipewire-session-manager.service added with 0.3.38.
We recommend to switch over to Nouveau drivers if your card is mentioned here and the feature you need is not in red TO DOhere for your particular card.
If you however still need proprietary drivers, you may want to execute the following to ensure all parts of the nVidia drivers get installed properly:
sudo mhwd -a pci nonfree 0300
If proprietary drivers don't fix everything, please click the ► at the beginning of this line to view more info
For AMD GPU users having a black screen with kernel 5.10
Due to a bug in the AMD drivers, please try the following first:
For GRUB:
Open a terminal or a TTY
Open /etc/default/grub in your favourite CLI editor (nano vi, emacs`)
Find the line: GRUB_CMDLINE_LINUX_DEFAULT="
Add amdgpu.dc=0
Save
Execute sudo update-grub and reboot
For systemd-boot:
Open a terminal or a TTY
Open /boot/loader/entries/manjarolinux5.10.conf in your favourite CLI editor (nano vi, emacs`)
Add amdgpu.dc=0 to the end of the line options
Save & reboot
For rEFInd:
Open a terminal or a TTY
Open /boot/refind_linux.conf in your favourite CLI editor (nano vi, emacs`)
Find the line: "Boot using default options" "root=
Add amdgpu.dc=0
Save & reboot
Possible 2 Min delay on shutdown with Gnome
If you got the problem with shutdown delay (about 2min) on Gnome, here is a workaround:
Edit /usr/lib/systemd/user/gnome-session-restart-dbus.service in your favourite editor as root and add Slice=-.slice as a line all by itself like this:
$PKGNAME warning: directory permissions differ on $Directoryname
followed by:
Filesystem: NNN package MMM
The solution:
sudo chmod MMM DirectoryName
Where obviously MMM is the second number you see (the correct one)
The explanation:
Your package expects the security permissions to be MMM but your system is set to NNN. This is just a warning (today) but to ensure you remain up-to-date with the latest and greatest security rules it’s advised to execute the above command.
Systemd error message: .slice: Failed to migrate controller cgroups
If you have a similar error message like this:
systemd[1004]: -.slice: Failed to migrate controller cgroups from /user.slice/user-1000.slice/user@1000.service, ignoring: Permission denied
… you may need to add this Grub parameter to your /etc/default/grub file in line GRUB_CMDLINE_LINUX_DEFAULT: systemd.unified_cgroup_hierarchy=true. To update your grub menu, call: sudo update-grub. For more information on that issue, see also Arch Forum.
Python 3.9 Requires Rebuilds
Instructions for quickly and easily rebuilding affected AUR packages:
# Install yay
sudo pacman -S --needed --noconfirm yay
# Rebuild AUR python packages
yay -S --noconfirm $(pacman -Qqo "/usr/lib/python3.8/site-packages")
The libtraceevent package prior to version 5.9-1 was missing a soname link. This has been fixed in 5.9-1, so the upgrade will need to overwrite the untracked files created by ldconfig. If you get any of these errors
libtraceevent: /usr/lib/libtraceevent.so.1 exists in filesystem
PAM and PAMBASE got updated, which might prevent you from login
Due to updates from pambase and pam you might take care about any .pacnew files in /etc/pam.d as for example pam_tally, pam_tally2 and pam_cracklib got deprecated. Read in the ArchWiki about managing those files.
Typical issue:
For recovery, it is enough to boot with kernel option “systemd.unit=rescue.target”, then proceed into /etc/pam.d and merge “system-auth” and “system-auth.pacnew”
System takes a long time to boot
If you’ve got errors like Failed to start Network Manager Wait Online. you can try removing systemd from passwd and group in /etc/nsswitch.conf as described here Update: If you can see the line dbus-daemon[1453]: [system] Connection has not authenticated soon enough, closing it (auth_timeout=30000ms, elapsed: 45146ms)
in in your dbus log ( journalctl -b -u dbus ), the new/better workaround is to switch cups from service to socket as described here: systemctl disable cups.service systemctl enable cups.socket
I think I've some font problems
With the update of fontconfig some major rules how fonts get applied changed. Make sure that your system has ttf-dejavu installed. A more in depth explanation can be found here.
Please RTFT (Read This Fine Thread) first before reporting the same issues over and over again!
:: pipewire-jack and jack2 are in conflict (jack). Remove jack2? [y/N] y
error: failed to prepare transaction (could not satisfy dependencies)
:: removing jack2 breaks dependency 'jack2=1.9.20' required by lib32-jack2
Removing jack2 and lib32-jack2 seems to be impossible due to multiple dependencies:
error: failed to prepare transaction (could not satisfy dependencies)
:: removing jack2 breaks dependency 'jack' required by audacious-plugins
:: removing jack2 breaks dependency 'libjack.so=0-64' required by audacity
:: removing jack2 breaks dependency 'jack' required by ffmpeg
...
You can install linux-firmware-qlogic to get rid of those errors, but unless you actually have qlogic stuff, you shouldn’t be worried about those warnings.
I also can reproduce that. Interesting though: These warnings only occur with the fallback initcpios not the regular ones. Shouldn’t the fallbacks include less stuff than the regular initcpios?
Of course, this will only work if you are not using any SCSI hardware on your rig (which, probably, will be true for most of us here).
Small advantage: Leaving out all the SCSI drivers in the initcpio build process, will speed things up a little bit when running mkinitcpio -P.
For convenience, here’s a patch doing the changes discussed.
--- /usr/lib/initcpio/install/block 2021-12-03 16:06:50.000000000 +0100
+++ /usr/lib/initcpio/install/block 2022-01-24 12:55:17.792192681 +0100
@@ -5,8 +5,10 @@
map add_module sd_mod? sr_mod? usb_storage? mmc_block? firewire-sbp2? virtio_blk?
- # pata, sata, scsi, nvme
- for filter in 'scsi/.*ata' '/(block|scsi|fusion|nvme)/' 'ata/[ps]ata_' \
+ # pata, sata, nvme
+ # scsi removed to avoid unnecessary missing firmware warnings
+ # on systems not using scsi
+ for filter in '/(block|fusion|nvme)/' 'ata/[ps]ata_' \
'ata/(ahci|pdc_adma|ata_piix|ata_generic)'; do
add_checked_modules "$filter"
done
@@ -34,8 +36,11 @@
help() {
cat <<HELPEOF
This hook loads the necessary modules for most block devices using pata, sata,
-scsi, firewire, usb, or mmc. Detection will take place at runtime. To minimize
+firewire, usb, or mmc. Detection will take place at runtime. To minimize
the modules in the image, add the autodetect hook too.
+In this version of the hook, the scsi drivers have been omitted to avoid
+missing firmware warnings when running mkinitcpio -P.
+Hence, DO NOT use this hook if you are using SCSI in your system.
HELPEOF
}
in addition to the changes mentioned i also saw, “sd_mod? sr_mod?” being omitted in the SCSI-less configuration.
decided to keep things as is, building fallback-initramfs with less drivers will indeed take slightly less time and will make it smaller. however considering the frequency it takes place and meddling with driver configuration (applies globally) that i have no idea about, I’ll pass on this. great work finding this BTW
The kernel runs well here - only drawback is that the Nvidia driver 495 is not compatible with it:
In file included from /home/gerd/Downloads/nvidia-495xx/dkms/src/NVIDIA-Linux-x86_64-495.46-no-compat32/kernel/common/inc/nv-procfs.h:26,
from /home/gerd/Downloads/nvidia-495xx/dkms/src/NVIDIA-Linux-x86_64-495.46-no-compat32/kernel/nvidia/nv-procfs.c:31:
/home/gerd/Downloads/nvidia-495xx/dkms/src/NVIDIA-Linux-x86_64-495.46-no-compat32/kernel/nvidia/nv-procfs.c: In function 'nv_procfs_open_gpu_info':
/home/gerd/Downloads/nvidia-495xx/dkms/src/NVIDIA-Linux-x86_64-495.46-no-compat32/kernel/common/inc/nv-procfs-utils.h:106:28: error: implicit declaration of function 'PDE' [-Werror=implicit-function-declaration]
106 | #define NV_PDE_DATA(inode) PDE(inode)->data
| ^~~
/home/gerd/Downloads/nvidia-495xx/dkms/src/NVIDIA-Linux-x86_64-495.46-no-compat32/kernel/common/inc/nv-procfs-utils.h:127:27: note: in expansion of macro 'NV_PDE_DATA'
127 | NV_PDE_DATA(inode)); \
| ^~~~~~~~~~~
Looks like in this driver, some #includes are missing or set incorrectly.
The Nvidia beta driver 510.39.01 runs with kernel 5.17-rc1, without visible problems so far.
For driver 495.46, I have prepared a tentative patch and uploaded it here: The easiest way to host your text
Feel free to test it and kindly provide feedback.
You may need to rebuild any AUR packages that install files to site-packages …