A fresh JP 6.0 installation and run nvstart-weston.sh in Terminal.
The Orin restart instantly, after restart, Settings → System → Windowing system still shows X11
weston-simple-egl gets a segmentation fault.
I’ve read Weston (Wayland) — NVIDIA Jetson Linux Developer Guide 1 documentation
Am I missing any step?
In my recent attempt, the AGX orin restarts instantly when I run nvstart-weston.sh
It stuck on restarting. I had to pull the power cable and re-plug it in. The monitor detected the signal but showed nothing, even during the UEFI boot stage.
Here’s the log: minicom.cap.txt (3.7 MB)
Hi @jasl9187
Is the issue still there after skipping oem-config?
I just think this issue shows a similar pattern to my flashing issue.
However, this one isn’t related to oem-config.
I run nvstart-weston.sh on a configured installation
I record a new output on the skip oem-config installation.
What I’ve done:
- Connect the AGX Orin to pure DP
- Manual flash the AGX Orin eMMC that skipping
oem-config (you taught me on my another post)
- Run
sudo apt update && sudo apt -y dist-upgrade && sudo apt -y autoremove and reboot
- Open the
Terminal app, run nvstart-weston.sh
- I saw it requires me to type my password
- After typing the password, it reboots instantly
- However, it stuck on rebooting (serial console flood CBB errors), so I had to plug out the power supply
- After reboot, Gnome Settings → About → Windowing system shows
X11
the output records the process: minicom.cap.txt (269.4 KB)
Even when I use pure DP, the CBB error is still there. It occurred on shutdown or rebooting.
I reflash to 5.1.3, nvstart-weston.sh still not working
- The monitor reports no signal
- Serial console shows it doesn’t reboot, and running normal except display
- I can get
/tmp/weston-startup.log in serial console
jasl@jasl-tegra:~$ cat /tmp/weston-startup.log
Date: 2024-05-06 CST
[02:22:39.778] weston 6.0.1
http://wayland.freedesktop.org
Bug reports to: https://bugs.freedesktop.org/enter_bug.cgi?product=Wayland&component=weston&version=3.0.0
Build: unknown (not built from git or tarball)
[02:22:39.779] Command line: /usr/bin/weston
[02:22:39.779] OS: Linux, 5.10.192-tegra, #1 SMP PREEMPT Mon Feb 19 20:19:53 PST 2024, aarch64
[02:22:39.780] Using config file '/etc/xdg/weston/weston.ini'
[02:22:39.781] Output repaint window is 8 ms maximum.
[02:22:39.781] Loading module '/usr/lib/aarch64-linux-gnu/weston//drm-backend.so'
[02:22:39.783] initializing drm backend
[02:22:39.783] no drm device found
[02:22:39.783] fatal: failed to create compositor backend
[02:22:39.783] Internal warning: debug scope 'drm-backend' has not been destroyed.
Could you follow these steps instead of script?
I guess " $ sudo modprobe nvidia_drm modeset=1" is missing.
I believe this line causes trouble,
The console shows:
[ 72.683332] [drm:nv_drm_exit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00020000] Failed to grab modeset ownership
[ 74.763332] [drm:nv_drm_exit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00020000] Failed to grab modeset ownership
This is on JP 5.1.3
Could you manually set weston up instead of running script from scratch?
Yes, manual set Weston up works.
It starts with a very basic desktop, not Gnome DE.
I got this error a couple weeks ago trying to install debian 12 and ubuntu 22.04 on a system with an amd r5 7600x and rtx 4070ti super. I never figured out what caused it or fixed it.
So, nvstart-weston.sh does not work on JP 5.1.3. The manual can start with a simple shell.
Can JP Wayland power Gnome DE?
Because of the JP 6.0 bug, I can’t try it. Could you help to test after the bug is fixed?
Wayland issue seems fixed on new 36.3. Please try.
Do you mean I need to apt upgrade?