How can I create a custom Linux image for my custom board to use the Zynq workflow or Intel SoC workflow with MATLAB/Simulink?

MathWorks offers an integrated workflow for targeting Zynq and Intel SoC platforms using HDL Coder and Embedded Coder. The supported boards are:

• For Zynq workflow: Zedboard, ZC702, ZC706, and ZCU102 evaluation kits.
• For Intel SoC workflow: Intel Cyclone V SoC Development Kit and Arrow SoCkit

However, I am using a Zynq board or Intel SoC that is not supported by MathWorks (e.g. MicroZed, PicoZed, Arty, Terasic DE10-Nano), or a completely custom-made board. Because of this, I was trying to follow the following demos:

• Define Custom Board and Reference Design for Zynq Workflow: https://www.mathworks.com/help/hdlcoder/ug/define-and-register-custom-board-and-reference-design-for-zynq-workflow.html
• Define Custom Board and Reference Design for Intel SoC Workflow: https://www.mathworks.com/help/hdlcoder/ug/define-and-register-custom-board-and-reference-design-for-intel-soc-workflow.html

In these demos, MathWorks provided the Linux image for the Zybo and Terasic DE10-SoC boards, but I am looking for a similar thing for my board. Can MathWorks provide this image for my board, or do you have some documentation teaching users how to create their own image?

Answer 1

[MathWorks Support Team]

The above-mentioned demos are aimed to define a custom board for HDL Code generation, not necessarily for creating a custom image, which is why the demo includes a pre-built image for the Zybo and Terasic DE10-SoC boards. These images were not tested on other boards.

For unsupported or custom boards, you would have 2 options:

Option 1: Use the Mathworks Buildroot environment from our GitHub repository and create a new image for your board: https://github.com/mathworks/buildroot

This would involve creating a new FSBL (First Stage Bootloader), device tree, and possibly reconfiguring the kernel to include additional drivers required by your board.

Refer to the following readme for basic instructions on how to use the Buildroot: https://github.com/mathworks/buildroot/blob/master/board/mathworks/doc/readme.md 

In MATLAB R2022a, we added a documentation example "Author a Xilinx Zynq Linux Image for a Custom Zynq Board by Using MathWorks Buildroot" to help you understand the basic steps involved: https://www.mathworks.com/help/hdlcoder/ug/xilinx-zynq-linux-image-for-custom-boards.html

Option 2: Use some existing Linux image or other build frameworks (e.g. Petalinux, Yocto), configure and build only the required MathWorks drivers against your kernel source, and add them to your image.

 

NOTE: Both of the above options are non-trivial and require considerable knowledge and experience with embedded Linux. Since the MathWorks Buildroot is not part of our product, we cannot guarantee it will work for all custom boards and we can only provide limited support.

If you are unfamiliar with how to build a custom Linux kernel and custom device tree, MathWorks Consulting Service may be able to help with this: https://www.mathworks.com/services/consulting.html

You can also contact MathWorks Training Services regarding our "Embedded Linux and System Integration for Zynq" training: ​​https://www.mathworks.com/learn/training/embedded-linux-and-system-integration-for-zynq.html