Example 1 – Secure versus non-secure hello world

In general, when you start a new project for an Arm Cortex-M device with TrustZone (such as the Cortex-M33 and Cortex-M55), the project will comprise two sub-projects: a secure and non-secure project. Secure and non-secure code have their own boot code and are compiled and linked independently in the sub-projects. Both secure and non-secure code run on the same processor but are loaded in isolated and independent areas of memory. All the code that handles security and configuration, such as boot code, firmware updates, and crypto libraries, is placed in the secure project. The rest of the application code is placed in the non-secure project. The objective is to minimize the amount of code in the secure project and run exhaustive checks on it for security vulnerabilities.

Important note

The implementation of secure software with TrustZone inherently follows the 10 security goals – specifically, Goal 7 of isolating...