No longer publicly expose the PeripheralClockControl struct

[jessebraham]

I have been meaning to do this for some time now, finally got around to it.

Rather than exposing the PeripheralClockControl struck via the SystemParts struct, the drivers just handle this all internally. Since PeripheralClockControl was already a ZST, I've just removed the &mut self argument of the enable function and instead converted it to an associated function. This type's scope was also limited to the crate-level.

This is obviously a pretty breaking change, but it's more than worth it IMO.

These changes were pretty straight forward, but also fairly invasive, so hopefully I haven't broken anything! If anybody has any questions/concerns/suggestions please let me know!

Closes #776

[MabezDev]

one nit pick, but LGTM otherwise!

[bjoernQ]

Probably enable for targets other than C6/H2 should be done in a critical section since at least in theory it could get interrupted between reading a register and writing it. If in an interrupt another peripheral gets enabled we might write back a bad value

Not very likely t.b.h. but at least we don't prevent it anymore

[jessebraham]

Sorry I'm not quite sure I follow, what do you mean by "we don't prevent it anymore"? What were we doing before exactly that we're not doing now?

[bjoernQ]

Before the user had to pass a mutable reference to SystemClockControl - there can be always only one at a time so access had to be serialized before

Maybe there is already something to prevent it in this PR ... using a smartphone right now - not ideal to look at this amount of changes

[jessebraham]

Oh I see what you mean now, sorry about that.

[MabezDev]

Before the user had to pass a mutable reference to SystemClockControl - there can be always only one at a time so access had to be serialized before

Maybe there is already something to prevent it in this PR ... using a smartphone right now - not ideal to look at this amount of changes

The desktop page doesn't work much better either :D, so laggy.

[AnthonyGrondin]

Thank you.

This should fix #776

[bugadani]

I always appreciate changes where the update process is one step:

• delete code

fairly invasive, so hopefully I haven't broken anything!

I'd love to try this in my project, but unfortunately, I am unable to because of two separate build errors introduced recently into the HAL. (#818)

[bjoernQ]

LGTM - I still think there should be a critical-section for non C6/H2 however