About system: URL: https://linux-hardware.org/?probe=ac5d29c839

While there is a lot of info and logs included in the link in comment 1, I don't
see anything showing plasma crashing.
 
The output of "journalctl -b --no-hostname", may include better info. Also,
please describe what you're seeing on the screen. Any error messages, or does
it just immediately return the the login screen? Which login manager (sddm, gdm,
xdm, etc.)?

CC: (none) => davidwhodgins

(In reply to Dave Hodgins from comment #2)
> While there is a lot of info and logs included in the link in comment 1, I
> don't
> see anything showing plasma crashing.
>  
> The output of "journalctl -b --no-hostname", may include better info. Also,
> please describe what you're seeing on the screen. Any error messages, or does
> it just immediately return the the login screen? Which login manager (sddm,
> gdm,
> xdm, etc.)?

Login manager - sddm. But I see on boot the dark displey with simple lines - terminal-like instead sddm

What you're describing is a text login screen. If it's sddm that's failing,
try logging in using your normal user id and password, and then entering
the command "startx". If that fails, then it's likely an Xorg problem.

Either way, once logged in, "journalctl -b --no-hostname" should give better
info about what's failing, and why.

(In reply to Dave Hodgins from comment #4)
> What you're describing is a text login screen. If it's sddm that's failing,
> try logging in using your normal user id and password, and then entering
> the command "startx". If that fails, then it's likely an Xorg problem.
> 
> Either way, once logged in, "journalctl -b --no-hostname" should give better
> info about what's failing, and why.

As I wrote above, after startx, the simplest graphical interface appears without Plasma elements, which is not the norm. I assumed that the reason is in the driver, which is being recompiled when updating the kernel. Changed from Nvidia to Nouveau - everything worked. Then I returned Nvidia back - so far it works fine.

It's likely starting icewm. Try "startx startkde".

Also see https://wiki.mageia.org/en/Mageia_8_Errata#Desktops for sddm issues

I am puzzled where this 5.16.18 kernel came from. On Mageia 8, I see on my up-to date system only 5.15.* (currently 5.15.32-desktop-1.mga8). And there is nothing lined up in updates_testing. Is this a Cauldron affair?

> Changed from Nvidia to Nouveau - everything worked.
> Then I returned Nvidia back - so far it works fine
So for the moment, no problem to play with. Please however post the output of: 
 $ inxi -MSGxx
to describe your system.

CC: (none) => lewyssmith

@Lewis: It's a kernel from core-backports.

Thanks. I was tempted to try this, but do not have Nvidia graphics, which seems to be a key factor.

The interesting part from my view is that IceWM starts. So this is in my opinion no Nvidia driver or kernel problem. More a damaged Plasma installation. But thats only a guess...

We have to be able to reproduce it to investigate it further.

If you read comment 5 again:

"Changed from Nvidia to Nouveau - everything worked. Then I returned Nvidia back - so far it works fine."


So seems nvidia for some reason failed the first time, but switching to nouveau and back to nvidia has restored the system to working order...

seems I didn't read comment 7 :/

(In reply to Lewis Smith from comment #7)
> I am puzzled where this 5.16.18 kernel came from. On Mageia 8, I see on my
> up-to date system only 5.15.* (currently 5.15.32-desktop-1.mga8). And there
> is nothing lined up in updates_testing. Is this a Cauldron affair?
> 
> > Changed from Nvidia to Nouveau - everything worked.
> > Then I returned Nvidia back - so far it works fine
> So for the moment, no problem to play with. Please however post the output
> of: 
>  $ inxi -MSGxx
> to describe your system.

Good day. Full info here https://linux-hardware.org/?probe=ac5d29c839

(In reply to Levchenko Roman from comment #1 & comment 14)
> About system: URL: https://linux-hardware.org/?probe=ac5d29c839
> Full info here https://linux-hardware.org/?probe=ac5d29c839
Oh dear! I overlooked this previously. The inxi output would have been more concise... The graphic essentials are roughly:

Device:
PCI 	10de:128b:1462:8c93 »
/ 03-00 	NVIDIA Corporation 	GK208B [GeForce GT 710] 	graphics card 	nvidia_drm 	detected

lsmod:
nvidia              35373056  636 nvidia_uvm,nvidia_modeset
nvidia_drm             69632  5
nvidia_modeset       1204224  13 nvidia_drm
nvidia_uvm           1187840  0

modprobe:
display-driver.conf
-------------------
install nvidia /sbin/modprobe nvidia-current $CMDLINE_OPTS

dmesg:
13.619650] nvidia: loading out-of-tree module taints kernel.
[   13.619663] nvidia: module license 'NVIDIA' taints kernel.
[   13.619664] Disabling lock debugging due to kernel taint
[   13.649649] nvidia-nvlink: Nvlink Core is being initialized, major device number 242

[   13.650843] nvidia 0000:01:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem
[   13.854656] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  470.94  Mon Dec  6 22:42:02 UTC 2021
[   14.948233] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  470.94  Mon Dec  6 22:31:27 UTC 2021
[   15.868667] [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
[   15.868670] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 0
[   16.744384] nvidia_uvm: module uses symbols from proprietary module nvidia, inheriting taint.
[   16.752051] nvidia-uvm: Loaded the UVM driver, major device number 239.
Nothing related to the graphics environment.

The best evidence would be the system journal taken at the point of error, as suggested in comment 4. I do not think we can deduce anything from what you gave.

(In reply to Thomas Backlund from comment #12)
> If you read comment 5 again:
> "Changed from Nvidia to Nouveau - everything worked. Then I returned Nvidia
> back - so far it works fine."
> So seems nvidia for some reason failed the first time, but switching to
> nouveau and back to nvidia has restored the system to working order...
This is how it looks. It does not change that the problem remains invisible.

@ Levchenko Roman,

Did the problem occur again after the latest Mageia 8 kernel update?

If not, then please close this bug report.

Keywords: (none) => NEEDINFO
CC: (none) => marja11

Closing due to no response.

Status: NEW => RESOLVED
Resolution: (none) => OLD