Developing on NVIDIA® Jetson™ for AI on the Edge

Jetson TX2 Build Kernel for L4T 28.2.1 updated

A quick note, the repository for building the Linux kernel onboard the NVIDIA Jetson TX2 development kit has been updated. The repository is on the JetsonHacks account on Github, buildJetsonTX2Kernel.

I want to thank Shreeyak ( for pointing out a version identification issue, as well as building an alternative to pull the kernel sources from the NVIDIA Git repositories directly. While I ultimately chose not to use this approach in the JetsonHacks scripts, there is a pull request available which implements this. If you are developing on the Jetson professionally, you and your team should seriously consider this approach.


6 Responses

  1. I was wondering whether building the kernel on target also involves compiling the device tree from sources on target? We have our own carrier and need to change the device tree in order to get rid of device on the DevKit but not on our carrier and add GPIOs, SPI, etc.

    1. The way that the device tree works is different from earlier versions. The device tree is now signed, which means that it must be built on the host and then flashed to the device. Thanks for reading!

  2. Hi! Im novice using GPUs, i installed on my Jetson tx2 the 3.3 JetPack with L4T 28.2.1 kernel, as recommended on the NVIDIA website, but im having problems compiling the caffe examples (mnist), i installed the caffe version (jetsonhacks/video), is there a problem with this new version (3.3 JetPack with L4T 28.2.1 kernel) ? , and what version (JetPack & kernel) do you recommend?

Leave a Reply

Your email address will not be published. Required fields are marked *


Some links here are affiliate links. If you purchase through these links I will receive a small commission at no additional cost to you. As an Amazon Associate, I earn from qualifying purchases.

Books, Ideas & Other Curiosities