core1 WILL be paused

[mcl-uk]

Hello

At: https://arduino-pico.readthedocs.io/en/latest/multicore.html#core-1-operation

It says: "By default, core1 (the second core) has no non-user written code running on it.
No interrupts, exceptions, or other background processing is done
(but the core is still subject to hardware stalls due to on-die memory resource conflicts).
When flash erase or write operations (i.e. LittleFS or EEPROM) are called from core0, core1 will be paused."

My question is: Is this the case even if I'm running core1's main loop() from ram?

Thanks for your help.

[earlephilhower]

Yes, it will be paused.

[mcl-uk]

Thanks for taking the time to answer so quickly.
Is this a fundamental issue or something that could be patched or addressed in a future release?

[earlephilhower]

In general, there's no way for this core to know it's safe to not freeze the other core. Even as a user, if GCC puts in an intrinsic (i.e. some of the 64b ops like * or /) those will generally be in flash but not "visible" in the actual source. Same for STL or templated operations.

It gets worse with the RP2350 with PSRAM which shares that QSPI bus. Memory accesses to it also need to be frozen while doing flash ops since the bus is taken over. So your code could all be in RAM but if you get passed in a pmalloc'd block to work on, boom.

It's not something that will be changed here that I can imagine, sorry. If you really need that then I think you'll need to jump to the bare SDK.

[mcl-uk]

Thanks again, good to know. Looks like the solution has to be an external eeprom chip :-(