chrysn

> The main objective is to give a linker error if the firmware grows into the AUX section. That's also provided by FLASH_WRITABLE_INIT (just with less the hassle of components...

If MTD creation is what it's about, then I suggest to build on the existing mechanism -- we have flash-to-MTD and flash partitioning, so the building blocks are there. An...

I think that flash inside an application is generally used vastly differently from flash between versions -- but I'll ponder over commonalities. A user who places data in an area,...

(The IPv6 example may be more relevant than I realized initially, for if the struct were considered opaque and 1-2 functions were different, the implementation might easily flip from having...

Thanks for your input, I've pulled in changes and made adjustments. In particular, I found that the IPv6 example doesn't need more docs (because it is returned by value), and...

A few issues/PR have accumulated that should have been created as sub-tasks in here: * [ ] #14169 (Gcoap does not accept separate responses): ready to be reviewed at #14178...

If we need a way to tell which flash pages correspond to the text section, let's use that (introducing first_used / last_used, or just feed the relevant linker symbols through...

> I can change the functions to reflect this but it should be first_used / last_used based on the whole firmware not just the .text section right ? That's really...

But then still, deprecating the others sounds like a good thing to do at this point. (Currently that'd just be through documentation; https://github.com/RIOT-OS/RIOT/pull/18565 is still blocked on doxygen issues).

I see that point about sock being standalone (and ad "what other than RIOT": a portable application can take RIOT's sock API and implement it on the other platforms it...