[EN] After yesterdays upgrade sleep function ist broken

[Sir_Aim]

I did an upgrade yesterday (no full-upgrade as some packages would have been removed).

Since then my system is not able to wake up after i put it to sleep.
When i press the power button while it's sleeping the system powers on (USB devices get power) but the screen stays black. I can not switch to a new terminal - it does not respond to any input. So reset is the only option.

Does anyone have an idea what this issue is cause by?

[hendrikL]

Please provide us the output of 'inxi -SG'.

Take a look at journalctl and/or dmesg, grep for errors, the graphics ...!

[Traciuarez]

Please provide us the output of 'inxi -SG'.

Take a look at journalctl and/or dmesg, space bar clicker, grep for errors, the graphics ...!

I'm also in the same situation. And where can I find 'inxi -SG?

[hendrikL]

Open a terminal and and type the command.

[Sir_Aim]

Code: [Select]

inxi -SG
System:
  Host: i9 Kernel: 6.9.7-1-siduction-amd64 arch: x86_64 bits: 64
  Desktop: KDE Plasma v: 5.27.11 Distro: siduction 2023.1.1 giants - kde -
    (202403291554)
Graphics:
  Device-1: NVIDIA TU116 [GeForce GTX 1660 Ti] driver: nvidia v: 550.67
  Display: unspecified server: X.Org v: 21.1.11 with: Xwayland v: 24.1.0
    driver: X: loaded: nvidia unloaded: fbdev,modesetting,vesa gpu: nvidia
    resolution: 1: 1920x1200~60Hz 2: 1920x1200~60Hz 3: 1920x1200~60Hz
  API: EGL v: 1.5 drivers: kms_swrast,nvidia,swrast
    platforms: gbm,x11,surfaceless,device
  API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: nvidia mesa v: 550.67
    renderer: NVIDIA GeForce GTX 1660 Ti/PCIe/SSE2
I found those in journalctl :

Code: [Select]

Jul 02 12:53:01 i9 kernel: rcu: INFO: rcu_preempt detected expedited stalls on CPUs/tasks: { 7-...D } 21 jiffies s: 4193 root: 0x80/.
Jul 02 12:53:01 i9 kernel: rcu: blocking rcu_node structures (internal RCU debug):
That was about the time I tried to wake the system up today. And those are also the last entries for the 2 boots before.

And those in dmesg:

Code: [Select]

[Di Jul  2 12:54:17 2024] nvidia-gpu 0000:01:00.3: i2c timeout error e0000000
[Di Jul  2 12:54:17 2024] ucsi_ccg 0-0008: probe with driver ucsi_ccg failed with error -110

[hendrikL]

What happens when the USB device is not connected?

[Sir_Aim]

I tried without USB devices (exept mouse and keyboard) - does not make a difference.

But I could login via ssh from my Laptop - so the system wakes up, but ignores any local input and fails to resume X.
I took a look with htop if a process got stuck but i couldn't see anything unusual.

Errors in logs are the same as before.

[hendrikL]

Did you try an older kernel?
Maybe Nvidia ...
Did you check the apt logs what has changed?
How long did you an upgrade and not a full-upgrade?
Did you try a full-upgrade with the --solver 3.0 option?
What Nvidia driver you are using?

[Sir_Aim]

So, I tried with the previous kernel - did not work.

The upgrade was a larger one as I did the last full-uprade about 5-6 weeks ago. Since then it always wanted to remove something important and I kept waiting for the issues beeing resolved. So this was the first time i did upgrade and not full-upgrade.

I checked the apt log and found this packages that might have to do with it:

Code: [Select]

udev
kwayland-integration
kwin-wayland
xserver-xorg-core
kscreen
qdbus-qt5
kdegraphics
plasma-workspace-wayland
systemd
xserver-xorg-legacy
kwin-x11
libkscreenlocker5
xwayland
sddm
nvidia-modprobe
I was able to run a full-upgrade today (using --solver 3.0) and a normal one after that.
But still not working.

Nvidia driver was version 550.67.
But I just purged it and switched to nouveau - now the system wakes up correctly.

So it seems there's something broken in combination with nvidia drivers...

[hendrikL]

You might be right, but I can't help you with Nvidia because I don't have the hardware.
One question, do you use Xorg or Wayland?

[Sir_Aim]

I did a fresh install at the end of march with a fresh siduction build - so I'm using whatever the default was back then then.

Distro: siduction 2023.1.1 giants - kde - (202403291554)
Graphics: server: X.Org v: 21.1.11 with: Xwayland v: 24.1.0

So I guess I'm using Xorg with Wayland... ?

I'm already very thankful for your help at this point!
Nouveau as a workaround will do for some time, but I'm planing to go back to some blender projects where nvidia drivers with cuda support perform worlds better than nouveau...

[thomv]

You are using either Xorg or wayland.
You can check which with

Code: [Select]

$ echo $XDG_SESSION_TYPE

[Sir_Aim]

Ok, now I got it. I thought Wayland is a new rendering engine in X - but it's a complete replacement.

This wakeup issue with nvidia-drivers was with Xorg.

Since I did the fresh install I was not able to login with sddm as it would freeze trying.
I always switched to a new terminal, logged in there and used startx to start KDE.

I just chose X11 for sddm and voila KDE launched successfully.
So Wayland is not working at all, neither with nvidia-drivers nor nouveau.

[hendrikL]

So I did a little research, and definitely you are not alone!

As example to solve some issues I found that, maybe you to -> https://wiki.archlinux.org/title/NVIDIA/Troubleshooting

But the net is full of reports about the problems you have with NVIDIA and Wayland/Xorg.
It's funny, but most people choose to start with startx/startplasma-wayland from Getty (terminal) to have the ability to suspend the system and wake it up again.
So you have to decide between a working login-manager or a working suspend.