How to compile tegra X1 source code

Hi,

I am learning how to rebuild the code for tegra X1 and I have found several posts from people asking about how to do it, I am adding my notes to the following wiki page in case someone wants to give it a try as well:

http://developer.ridgerun.com/wiki/index.php?title=Compiling_Tegra_X1_source_code

Hopefully they will save you some time.

-David

nice work. thank you.

I feel this topic should be pinned to the forum, many users will find David Soto’s wiki useful.

Additionally I have compiled some more up to date notes on how to setup a cross compile environment:
https://gist.github.com/chutsu/9bb6abe6f61924c88521adec859c7006

I feel that Nvidia should have made slightly more effort in making this as painless as possible, the nvidia docs on this matter is not clear enough. I feel as though they should really have a Wiki of some sort to communicate this vital information instead of having users troll through numerous forums to find answers they seek.

The Jetson TX1 wiki is located here: http://elinux.org/Jetson_TX1

There is a link to the Ridgerun article on the wiki under “Guides and Tutorials -> Compiling the Kernel”. It appears to have been placed there August 10, 2016.

I am not associated with NVIDIA. You can contribute to the wiki if you feel community knowledge is lacking in any particular area.

Any knows if this will work on a docker environment?

Hi spottybadrabbit, What do you mean?

I have boot2docker running on a server. I want to compile the code for the TX1 on there so I can simulate a virtual environment. Reason is there are a lot of packages I want to test and build for production before deploying to the TX1. Docker makes it easier to revert and deploy different version of the image. In the event one the TX1 kernel got corrupted during compilation. Would the Ridge tutorial run work in that environment or does the OS have to be Ubuntu with the TX1 kernel compiled for it/

Thanks for your response. We haven’t tried it.

It does but not enough space on dev board. Booting from SD or SSD card allows you to install nvidia-docker and then virtualize anything.