Up until this point, I’ve only written about embedded applications which get loaded into a device’s non-volatile memory. That’s how most “normal” programs work, but it’s not the only way to run code on a microcontroller. There are plenty of situations where you might want to run a one-off program on a device before resetting it to its previous state. Maybe your microcontroller has a large firmware image that takes a long time to erase and restore, or maybe you want to perform a one-time data transfer betwee
When you encounter those sorts of situations, you can write a program to run from your microcontroller’s RAM instead of its Flash memory. There are a few extra steps to be aware of, and the process of loading and running the program is a bit different, but you won’t need to change much in the application’s C code.
In this post, I’ll try to demonstrate how and why to run code from RAM with a couple of examples. First we’ll write a minimal RAM program to blink an LED. Then, once we know how to load code into RAM and run it, we’ll write an ephemeral program which can send a file from our computer to a QSPI Flash chip connected to the microcontroller.