System boots into black screen with auto login for user dietpi

Required Information

  • DietPi version | 9.20
  • Distro version | trixie
  • Kernel version | 6.12.58
  • Architecture | arm64
  • SBC model | Pine A64
  • Power supply used | RPi 5V 1A
  • SD card used | Samsung Evo

Additional Information (if applicable)

  • Software title | X11 & Xfce

Steps to reproduce

  1. Install X11 and Xfce
  2. Set auto login to DE for user dietpi
  3. Reboot

Expected behaviour

  • System boots into Xfce

Actual behaviour

  • System only shows a black screen
  • When auto login is set to DE for user root, the screen shows Xfce

Since booting into Xfce using root works flawlessly, I suspect my dietpi user isn’t allowed to access the xserver. I tried to allow any user to use xinit (stackexchange) but there was no Xwrapper.config. I created that file and put in

allowed_users = anybody

but that didn’t work.

Using CTRL + ALT + F1 i see:

failed to open /dev/dri/renderD128: Permission denied
glx: failed to create dri3 screen
failed to load driver: sun41-drm
(II) AIGLX: Suspending AIGLX clients for VT Switch
(II) AIGLX: Suspending AIGLX clients for VT Switch
(II) AIGLX: Suspending AIGLX clients for VT Switch

Can somebody please help me? Is there any additional info I can provide to fix this?

Did you tried adding user dietpi to group video?

Thanks for the quick answer. There was only octoprint in the group video. I added user dietpi and rebooted. Sadly, the screen still turns black.

In virtual console 1, there are the same errors like before:

failed to open /dev/dri/renderD128: Permission denied
glx: failed to create dri3 screen
failed to load driver: sun41-drm
(II) AIGLX: Suspending AIGLX clients for VT Switch

(The last one is now displayed only once.)
Edit: (II) AIGLX: Suspending AIGLX clients for VT Switch is shown evertime i switch to tty1.

Virtual console 2 is black and in VT3 I get an error after logging in with user dietpi:

(EE) 
Fatal server error:
(EE) Server is alread active for display 0
        If this server is no longer running, remove /tmp/.X0-lock and start again

(EE)
(EE)
Please consult the The X.Org Foundation support at http://wiki.x.org for help.
(EE)
X connection to :0 broken (explicit kill or server shutdown).

I guess this is because another session uses X already.

Edit #2: I switched to Light DM login mask instead of auto login. This also results in a black screen. Auto login for user root still works.

can you check owner and group

ls -la /dev/dri/renderD128

Sure:

dietpi@OctoPine:~$ ls -la /dev/dri/renderD128
crw-rw---- 1 root render 226, 128 Jan 12 13:10 /dev/dri/renderD128

I tried sudo chmod 777 renderD128 and did ls -la renderD128 again (switched to the directory before that of course). Output:
crwxrwxrwx 1 root render 226, 128 Jan 12 13:12 /dev/dri/renderD128

But after reboot the permissions were reverted:

dietpi@OctoPine:~$ ls -la /dev/dri/renderD128
crw-rw---- 1 root render 226, 128 Jan 12 13:17 /dev/dri/renderD128

Since the owner’s group is render, I added user dietpi to that group. Then I switched desktop auto login to user dietpi and rebootet. Now it seems that everything works. I can’t see Xfce because Octodash directly shows up (it’s configured for auto boot) but that is the intended behavior so I guess I’m fine.

Thanks for your quick help! I’ll report back IF something is still off.

Edit: I marked this post as solution. Is that alright or is marking one’s own post as solution regarded arrogant?

All good. Yes adding user to render group was the solution. Sometimes video group is enough. Sometimes not :smile:

Which depends on where the actual issue is, right? So i’ts actually good when solving an issue to look at file permissions. Which makes sense in my mind because in linux, everything is a file.

Thanks again for your guidance!