Hi everybody,
Is there a commonly used or already-built system for platform/chip abstraction between multiple STM32 parts? I have encountered various PAL's over time, some better than others, but I was curious if there is one common one that people use when doing their own designs, and not already using a PAL baked into some vendor's SDK?
I am about to set out work on a new module in a hardware system that has two possible paths in front of it, one blank-page and one backwards compatible.
Roughly speaking, I might be extending a module already designed with an STM32F2 or the module might throw off it's legacy needs and then we could build it with an STM32U5. Or there's even a third avenue wherein we'd update the legacy section of the system and target STM32U5 hardware.
I can generate the template code for the F2 and U5, and then include both in a larger project.
Conceptually, I can see how e.g. CAN functions would just need to be routed to each library's CAN implementations, and it's probably even the case that the STM32 HAL has identical semantics for both.
And conceptually I can see how generic resources like timebases and pulse generators could be written with a layer that would then just need to be shimmed to each chip's specifics, and a choice of actual silicon peripheral, e.g. LPTIMx vs TIMx, etc.
But ... will I need to invent all of that myself from scratch? Has someone already done it in a way that I can just borrow?
Any recommendations would be appreciated!
(yes, doing it all myself from scratch would be a good skill builder, but, well, my time in limited)
I have looked at the Zephy project. They seem to have a fairly complete system for this ... but all as part of their build and config system. In this case, I am constrained to using ThreadX (or ... ugh ... Green Hills), so trying to excise and re-use their PAL might be too big a task.
