[Stable Update] 2021-05-06 - Kernels, Mesa, VirtualBox, LxQt 0.17, Deepin, Gnome Apps 40.1

You did rebuild those after kernel update?

Budgie desktop 10.5.3 soon?

After installing them using this guide it does this automatically with each update.

Uninstalling nvidia drivers (back to Nouveau) solved the problem. Maybe the 340.108 no longer get along with the latest updates like that of Mesa.

1 Like

I see, thanks for letting me know! Will subscribe there to see if I can contribute at some point.

Remove whatever package requires those packages, then run the update again.

2 Likes

After the latest stable update, I’m having this problem with my Activities function in Gnome. It wasn’t happening to me before the update. In the thread they point at dash to dock as the possible cause, so I’ll try disabling it for a while and see if the problem persists.

Gnome, 5.4, Intel

Evolution, terminal, virtualbox, timeshift so far will not open. Brave opens, chrome opens, pamac opens.

Why you didn’t use the 5.10?

Hi,

The solution you provided for amd gpus (with grub) didn’t work for me, when I boot my machine just displays the asus boot screen and freezes on it, I had to chroot in order to make it bootable again.
Just in case, I did this change in /etc/default/grub :

    GRUB_CMDLINE_LINUX_DEFAULT="quiet apparmor=1 security=apparmor udev.log_priority=3"
    ### became : 
    GRUB_CMDLINE_LINUX_DEFAULT="quiet apparmor=1 security=apparmor udev.log_priority=3  amdgpu.dc=0"

Updated everything this morning (Manjaro Gnome), about 80-ish packages.

Everything went fine, asked me to reboot so I did.

On boot:

  1. Dell screen
  2. Password for disk encryption
  3. Another Dell screen
    – everything normal up until here –
  4. All black screen, nothing except for my keyboard lighting up when I click.

Tried alt + ctrl + F1 (and F2), nothing happened.

Did a force shutdown (power button for about 10sec.)

Boot again, and now it works. Might take a Timeshift backup next time because I nearly cr*pped my pants :sweat_smile:

Should I be worried or consider this a one time glitch in the maxtrix and forget about it if it doesn’t occur again?

Context:
I’m on 5.10 LTS, have the latest NVIDIA drivers (hybrid laptop) with optimus-manager.

After updating my multiseat-setup was broken, so that lightdm would refuse to start. I could start xfce with startx, but only one seat worked.

Today i solved the problem by disabling 2 config-files which were brought by recent updates to /usr/share/X11/xorg.conf.d These configfiles may break all multiseat-setups using loginctl see arch-multiseat

in short all you have to do is disabling these files:

sudo mv /usr/share/X11/xorg.conf.d/10-amdgpu-nvidia-drm-outputclass.conf /usr/share/X11/xorg.conf.d/10-amdgpu-nvidia-drm-outputclass.conx       
sudo mv /usr/share/X11/xorg.conf.d/10-intel-nvidia-drm-outputclass.conf /usr/share/X11/xorg.conf.d/10-intel-nvidia-drm-outputclass.conx

if you are interested in details see the full info at Multiseat-Setup broken by …

1 Like

The firmware for Ravenridge (Zen 1 + Vega APU (2*00U)) in 20210426.fa0efef is faulty again (also was with the last stable update in April/End of March):
It crashes amdgpu within 2 hours of runtime. The messages in the kernellog switching on every crash, so they’re not helpful. Last time I got

amdgpu 0000:04:00.0: amdgpu: [gfxhub0] retry page fault (src_id:0 ring:0 vmid:1 pasid:32769, for process Xorg pid 942 thread Xorg:cs0 pid 981)

, the crash before was throwing:

drm:amdgpu_gart_unbind [amdgpu]] ERROR Timeout waiting for VM flush ACK!

Tried Kernel 5.10.30 – 34 and 5.4.106 – 116 and some new parameters like amdgpu.noretry=0 but the only solution is to copy back the old firmwarefiles (/usr/lib/firmware/amdgpu/) and recreate initramsfs.

Therefore I’m pretty sure the source of this new instability is the new firmware (see also the releasedates of new firmware in AMD’s repository - great, I not allowed to link it -.-’ ).
According to another user from another forum this also concerns Navi 21 (but not Navi 10).

5.11.14 is better

more info here:

The only problem I’ve found is that suddenly my user gets locked out after multiple failed login attempts (sodding capslock!) which has never been a problem before. (And isn’t required behaviour on my home computer anyway).

The solution seems to be to edit /etc/security/faillock.conf, uncomment “deny = 3” and change it to “deny = 0”.

I tested deepin and Budgie then shared problem detailed system report in external topic for not occupying this topic.

https://forum.manjaro.org/t/system-starting-problem/65851

grub seems to have borked something, i switched to linux512 and removed linux58 and linux511 but grub still thinks linux58 is installed.

$ mhwd-kernel -li              
Currently running: 5.12.1-2-MANJARO (linux512)
The following kernels are installed in your system:
* linux510
* linux512
* linux54

yet

Found linux image: /boot/vmlinuz-5.8-x86_64
Found initrd image: /boot/amd-ucode.img /boot/initramfs-5.8-x86_64.img
Found initrd fallback image: /boot/initramfs-5.8-x86_64-fallback.img

No, grub sees the kernel and initramfs in /boot. Delete those manually.

well did not something break, as every time previously it has been removed automatically? Maybe incompatibility between mhwd-kernel or grub and linux512 ?

No.
The package removal of a kernel package should trigger a pacman hook (/usr/share/libalpm/hooks/60-mkinitcpio-remove.hook) which in turn calls the /usr/share/libalpm/scripts/mkinitcpio-remove script which should remove those remnants.
Don’t know what went wrong on your system.

2 Likes

Am I the only who has problems with “Print Screen” button after this update?

On KDE, “Print Screen” button on keyboard does nothing now (it used to capture screenshots with Spectacle in the past). I even re-configured keyboard shortcuts via settings - no luck. No matter what I do, “Print Screen” button alone doesn’t work anymore. If I configure shortcut to use SOME_OTHER_BUTTON + “Print Screen”, it works fine.