Open alevy opened 6 years ago
I was probably a bit optimistic about sharing code with other STM32 families. I will cleanup stm32
in the coming days. Judging from the F4 port, most of the modules should get moved to the stm32f1
crate.
There may not be lot of sharable peripheral code between main families (F1, F4, L4, ...) but those families are huge with their own subfamilies/lines and there is probably going to be a lot sharable code between them.
I think the basic structure makes sense right now, with two exceptions:
It might be worth getting rid of the
tock
submodule, and instead depending on the main git repo, possibly tied to a specific commit. The biggest reason not to do this, I think, is that Signpost has been using a submodule pretty successfully and why rock the boat.The
cortex-m3
crate could probably make its way into the main Tock repo sooner than everything else. While we don't currently have another Cortex-M3 supported upstream, there is really no difference between the M3 and M4 w.r.t to the details in these crates. The core kernel, on the other hand, is fairly coupled with the arch crate still, and it might be easier to change those in lockstep.I don't think either need to change particularly quickly though.
As for the
stm32
crate, I believe it actually has some stm32f1 specific code in there, but it might be easiest to just let that be and fix it when another variant needs it to be organized differently.