training-kas

Practical tutorial repository for learning a Yocto workflow through kas.

This project is intentionally small: it provides kas configuration files so you can build core-image-minimal for QEMU or Jetson AGX Orin without manually setting up a BitBake build directory.

Repository layout

• kas/base.yml: shared project configuration (poky, scarthgap, core-image-minimal)
• kas/qemu.yml: includes base.yml, sets qemux86-64, enables debug-tweaks, and sets parallel build values
• kas/tegra.yml: includes base.yml, adds meta-tegra, and sets jetson-agx-orin-devkit
• poky/: Yocto Project source tree used by the kas configs
• build/: generated build output directory

Prerequisites

Install kas (for example with pip):

python3 -m pip install --user kas

Depending on your Linux distribution, you may also need Yocto host dependencies (compiler toolchain, Python modules, etc.).

Recommended workflow

Inspect the resolved configuration:

kas dump kas/qemu.yml

Build the image:

kas build kas/qemu.yml

Build the Jetson AGX Orin image:

kas build kas/tegra.yml

Open a shell in the configured build environment:

kas shell kas/qemu.yml

Run the image in QEMU:

kas shell kas/qemu.yml -c "runqemu qemux86-64 nographic"

Notes

• The target is core-image-minimal (defined in kas/base.yml).
• The machine is jetson-agx-orin-devkit (defined in kas/tegra.yml).
• The poky and meta-tegra repos are pinned to the current scarthgap heads.
• Parallelism is currently pinned to 8 threads in kas/qemu.yml and kas/tegra.yml.
• Build artifacts are written under build/.