I have developed an image based on the AWS IoT example which runs on a custom board. The custom overlay exposes some additional UARTs, LEDs, buttons, etc. in custom pin locations and removes things we don't need. Everything in our application image works fine if I either build with the stock MCUboot child image (without handing MCUboot a custom overlay or config or even trying to use MCUboot), or if I build without any child image at all. If I build with my custom child image, which I pass the same overlay I use for our application, when the child image boots into the application image the UARTs and ADC work but the GPIOs do not. The initialization code runs without errors, but any attempt to change which LEDs are on or off doesn't actually do anything.
- Windows 11 Pro
- VSCode 1.73.1
- nRF Connect for VSCode v2022.11.140
- nRF Device Tree v2022.11.153
- nRF Kconfig v2022.11.50
- nRF Connect SDK / Toolchain version 2.1.2
- All builds of the application image are with TFM
- The specified board for the child image is nrf9160dk_nrf9160
- The specified board for the application image is nrf9160dk_nrf9160_ns
- The CONFIG_NRF_GPIO0_PIN_MASK_SECURE is set to 0 in the generated .config for the application (same as when building with mcuboot)
- UARTE1 is the only peripheral configured as secure (For TFM) in the generated .config for the application
- The generated .config for the child image doesn't mention any *_SECURE peripheral or pin settings (neither does the generated mcuboot .config)
I do not know if my issues are from incorrect configuration, or if the child image is not booting into the application image correctly. I noticed that if I try to boot to where the application actually exists a reset occurs and execution starts over in the child image (over and over), but if I boot to where TFM starts execution does actually make it to my application image.
Please let me know if there is additional information I can provide to help figure this out, and thanks very much for any assistance.
