Missing RT kernel header in /rootfs/usr/src

Hi,

I have followed the following instructions to build RT kernel for r35.5.
https://docs.nvidia.com/jetson/archives/r35.5.0/DeveloperGuide/SD/Kernel/KernelCustomization.html

I can boot up the orin in RT(verify by uname -r) but currently my /usr/src does not contain the linux-headers-5.10.192-rt96-tegra directory.
WIthout the kernel headers, i am unable to compile kernel drivers.

Is there any instructions to cross build this?

Thanks!

Hi,
If the device cannot be flashed/booted, please refer to the page to get uart log from the device:
Jetson/General debug - eLinux.org
And get logs of host PC and Jetson device for reference. If you are using custom board, you can compare uart log of developer kit and custom board to get more information.
Also please check FAQs:
Jetson AGX Orin FAQ
If possible, we would suggest follow quick start in developer guide to re-flash the system:
Quick Start — NVIDIA Jetson Linux Developer Guide 1 documentation
And see if the issue still persists on a clean-flashed system.
Thanks!

Hi,
I have flashed and boot up the device successfully and i am able to run cyclictest from rt_test properly.
However, I need to compile kernel drivers but the directory linux-headers-5.10.192-rt96-tegra-ubuntu20.04_aarch64 is missing.

I have checked the /Linux_for_Tegra/rootfs, I only have linux-headers-5.10.192-tegra-ubuntu20.04_aarch64 directory.

But in the lib/modules directory, I have both 5.10.192-tegra and 5.10.192-rt96-tegra directories.

Hi,
Please follow developer guide to build it on host PC:

https://docs.nvidia.com/jetson/archives/r36.4/DeveloperGuide/SD/Kernel/KernelCustomization.html

You can customize generic_rt_build.sh to enable additional configs for your use-case.

I have quite some codes that are dependent on cuda 11.4 and so I might not be able to use r36.4

I suppose the problem that I faced is similar to the following thread:

But, I don’t see any solution for r35.x to generate linux kernel rt header?

Hi,
Please share a simple code and steps so that we can replicate the issue. We think CUDA sample should be built with userspace header files. Not quite familiar with the use-case that it depends on kernel header files.

Hi DaneLLL,

Thanks for your reply, what I mean is that I have codes in userspace that are dependent on cuda 11.4 and it might require efforts to update it to cuda 12.6.

However, I have kernel driver source codes that needs linux kernel header to compile to use the frame grabber cards.

I have tried to build the drivers but it needs the linux headers.

Although I can run the jetson r35.5.0 RT kernel, i could not compile the driver codes since the linux headers in /usr/src directory is not available.

Hi,
We will evaluate to add this in future release. On Jetpack 5.1.3, please copy the driver code to host PC and build it.

Hi DaneLLL,

Thanks for the tip.

I prefer not to meddle too much with the driver’s build script. Therefore, instead of cross building, I copied the kernel source to Orin and built the full kernel in Orin.

Then, I used it to build the drivers.
I was successful and the built drivers were successfully loaded and detected by the userspace application.

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