AGX Orin – Xorg / display errors after update, possible NVIDIA driver version mismatch

Hello,

I am using a Jetson AGX Orin (64GB).

Device and software information (from sudo jetson_release -v):

  • Module: Jetson AGX Orin (64GB)

  • JetPack: 5.1.4

  • L4T: 35.6.0

  • Kernel: 5.10.216-tegra

  • OS: Ubuntu 20.04

  • Root filesystem: NVMe (system is booting from NVMe, not eMMC)

Issue description:

Xorg fails to start correctly. When I run startx, I see repeated NVIDIA kernel errors related to display bandwidth and GPU initialization, for example:

  • Failed to allocate NVIDIA Error Handler

  • DISPLAY_ICC_BW_CLIENT_EXT assertions

  • gpumgrGetSomeGpu: Failed to retrieve pGpu - Too early call

  • DCE / display related errors in the kernel log

Here are screen shots of my errors:

And other UART serial sometimes the system freezes and reboots showing:

The system previously worked correctly on the same hardware.

Important observation:

Although the system reports JetPack 5.1.4 / L4T 35.6.0, the NVIDIA kernel module reports a different version:

modinfo -F version nvidia
35.5.0

So currently I appear to have:

  • L4T / userspace / kernel: 35.6.0

  • NVIDIA kernel driver (nvidia.ko): 35.5.0

Question:

  1. Is this version mismatch between the NVIDIA kernel module (35.5.0) and the rest of the system (35.6.0) expected to cause Xorg or display initialization failures?

  2. Should the NVIDIA kernel modules always exactly match the L4T version?

Thank you for your help.

*** Please note that this reply is generated by LLM automatically ***

The issue you’re experiencing with Xorg failing to start correctly and the repeated NVIDIA kernel errors related to display bandwidth and GPU initialization could be caused by the mismatch between the NVIDIA kernel module version (3550) and the rest of the system (3560).

In general, it’s recommended to keep the NVIDIA kernel module version in sync with the L4T version to ensure compatibility and prevent potential issues. The NVIDIA kernel module is a critical component that interacts with the GPU, and a mismatch between the module version and the L4T version can lead to problems like the ones you’re experiencing.

In your case, the NVIDIA kernel module version (3550) is older than the L4T version (3560), which could be causing the display initialization failures and Xorg issues. It’s possible that the older kernel module is not compatible with the newer L4T version, leading to the errors you’re seeing.

To resolve this issue, you can try updating the NVIDIA kernel module to match the L4T version (3560). You can do this by running the following command:

sudo apt-get update
sudo apt-get install nvidia-l4t-kernel

This should update the NVIDIA kernel module to the latest version, which should match the L4T version. After updating the kernel module, reboot your system and try running Xorg again to see if the issue is resolved.

It’s also worth noting that the NVIDIA kernel module version should always match the L4T version to ensure compatibility and prevent potential issues. If you’re experiencing issues with your system, it’s always a good idea to check the NVIDIA kernel module version and ensure it’s in sync with the L4T version.

In comparison to other Jetson devices, the Jetson AGX Orin is a more powerful device with a more complex architecture, which can make it more prone to issues like the one you’re experiencing. However, the solution to this issue is the same as it would be for other Jetson devices: updating the NVIDIA kernel module to match the L4T version.

In terms of the system configuration, it’s worth noting that the root filesystem is booting from NVMe, which is a faster storage device than eMMC. This could potentially cause issues with the system, especially if the NVMe device is not properly configured. However, in this case, the issue appears to be related to the NVIDIA kernel module version, rather than the system configuration.

Overall, updating the NVIDIA kernel module to match the L4T version should resolve the issue you’re experiencing with Xorg failing to start correctly and the repeated NVIDIA kernel errors related to display bandwidth and GPU initialization.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***

Hi,
If you use custom board, would suggest re-flash to Jetpack 5.1.5 or 6.2.1:
Jetson Linux | NVIDIA Developer
Jetson Linux Release 36.4.4 | NVIDIA Developer

And in first boot-up, please execute the commands:

Making sure you're not a bot!
[Jetpack 6.1/r36.4.0] Self-built kernel is overwritten after sudo apt update && sudo apt upgrade

To ensure your custom kernel is not overwritten, causing mismatch.

Hi @DaneLLL

Thanks for you response, I will definitely try holding those packages you said about as this most defiantly caused my mismatch:

Do you know why you said about moving from to Jetpack 5.1.5 or 6.2.1. Is there a current known issue on Jetpack 5.1.4, all I could find on display DCE errors was this:

Hi,
We don’t deliver the dce binary for Jetpack 5.1.4. Please upgrade to latest release.

I will try the upgrade to latest, but I do believe I have the DCE bin:

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.