stm32f1xx-hal
stm32f1xx-hal copied to clipboard
stm32-rs
Exporting the sdv2rust crate as device
For consistency, I propose to re export the sdv2rust crate as device. It is coherent with the name used by RTFM: https://japaric.github.io/cortex-m-rtfm/book/by-example/app.html
If there is approvals and no veto, I can propose such a PR.
Actually the agreed on terminology is pac. I would not go as far as insta-dropping stm32 but we can certainly mark it is deprecated and re-export the PAC as pac.
Funny enough even the proponents of that terminology are not using pac to re-export the PAC:
pub use nrf52832_pac as target;
Changing the name does sound like a good idea, I don't really like the current stm32 name, device makes a lot more sense. However, if pac is the agreed upon terminology, we should probably do that.
Out of curiosity, why was pac chosen, I don't really understand what it would mean in this context.
Of course we will do that with a new export. I don't really care of the name as long as it can be common between all hal implementation (not only stm32).
The ambiguity with pac is that there are 2 PAC crates: the core PAC (cortex-m) and the device PAC (stm32f1 in this case).
Is that so? I haven't thought of cortex-m as a PAC so far. Maybe we should raise the question as issue on the embedded-wg issue tracker to sort out potential conflicts before settling on an implementation.
If nothing else, the figure in the embedded book seems to indicate that the PAC is the only layer between the microcontroller and the HAL. A wg issue sounds like a good idea to get clarifications about this and to make sure it is standardised
Well, don't let the figure irritate you; I've created that. ;)
However https://rust-embedded.github.io/book/start/registers.html agrees with that view and that hadn't been done by me.
The thing is that makes me saying that cortex-m is a "PAC" is that there is a Peripherals object in cortex-m: https://docs.rs/cortex-m/0.5.8/cortex_m/peripheral/struct.Peripherals.html
In RTFM, there is a core peripheral object and a device peripheral object that is given to the init function.
#29 Closed this automatically but I don't think it's fixed since we didn't agree upon a name
@TheZoq2 Uhm, but with the latest changes we're exporting it as stm32, pac and device.
Yea, that's true, but I feel like we should eventually decide on one of them to avoid confusion. Allthough I suppose allowing all three also has some benefits
Well, if we have to chose one it should probably the documented one which is pac. 😏