cortex-m3-rtos VS stm32mp1-baremetal

I will add my own project here https://github.com/t4th/cortex-m3-rtos

I decided to have similar architecture in my own private project ( https://github.com/t4th/cortex-m3-rtos ), because the underlying concept is that, it is all just an array of structs and I used C++ to Abstract those arrays as different modules for cleaner architecture.

Here is project that I was experimenting with C++14 instead of doing it in plain C: https://github.com/t4th/cortex-m3-rtos/

Here is repo https://github.com/t4th/cortex-m3-rtos You have two branches with simple prove of concepts if you are curious about technical problems only.

2x A7 cores for running an OS (Linux, Android) or you can run bare metal code on them too. (https://github.com/4ms/stm32mp1-baremetal)

stm32mp1-baremetal and forum post and summary, linker script

As others have said, it’s not real-time Linux because it’s not Linux. I also want to add that if you require hard-real time and are ok with not having Linux, you can also use the A7 core for hard-real time applications. Some example projects here: stm32mp1-baremetal

I didn’t go through the 29 bytes you’re sending to the codec, but I do see you call the array “master”. The codec should be setup as slave. It could cause problems if the codec and ARM chip are both trying to be the master (generating the clocks). You’ve double/triple checked those bytes set up the codec properly? You may want to spell them out like I’ve done here: https://github.com/4ms/stm32mp1-baremetal/blob/master/examples/audio_processor/lib/mdrivlib/drivers/codec_CS42L51_registers.h

And if you want more processing power and memory, there are some audio examples in this project using a Cortex A7 (STM32MP157) bare-metal (512MB RAM, 800MHz) https://github.com/4ms/stm32mp1-baremetal