Multicore and EEPROM

[LazaroFilm]

I'm working on a project where I am trying to have an IMU sensor and Fusion algorithm running on core1 while core0 controls everything else (TFT screen, Touch, LVGL, Neopixels, and settings saves via EEPROM).

I'm using the multicore because I want the IMU and Fusion to run in a loop as fast as possible independently from what core0 does (the screen refreshing graphics can slow down the loop pretty badly). I am running into the situation where saving EEPROM freezes the whole program sometimes.

I tried using rp2040.idleOtherCore() but I'm not sure I'm using it properly. Can someone give me a simple exemple of a loop running on core1 while saving a value to EEPROM on core0?

Also, I'm saving the values from core1 to global variable defined at the top of the sketch and it seems to work. Is that okay, or should I use the FIFO (Which I don't quite understand either)

[earlephilhower]

...I tried using rp2040.idleOtherCore() but...

There's almost no reason in the world to use this outside of writing to flash memory outside of the EEPROM and LittleFS libraries (i.e. using the SDK calls directly). If you want atomic access to a multi-word state use an SDK mutex or semaphore. If it's just a word(32b) long element you need to protect, don't bother because 32b writes are atomic. (16b and 8b writes are not, however, AFAIK).

Writing to EEPROM has to freeze the other core because when you write to flash, you can't read from it. So the non-writing core needs to jump to an code-in-RAM busy loop while the writing is going on (hence, the embedded idleOtherCore call). It does this automatically, there is no need for you to do anything yourself other than use the EEPROM class.