Hi
I have two periodic tasks, one on SC and one on MCU.
I use a small memory region to exchange data between these tasks (50B).
Tasks are independent, operate on different cores, and therefore they can read/write to this memory region concurrently.
I want to avoid a theoretically possible situation, where SC reads the region updated only partially by MCU.
How can I implement functionality to treat this memory region as being atomic?
When the SC reads this region, the MCU waits to write.
Reading the manual, I found AUX semaphore (AUX_SMPH), dedicated to managing peripheral ownership. There are three slots available to use.
Do I have to use this feature?
Is there anything more appropriate?
Regards
Adam