Pat Pannuto

@lschuermann Is this something you still want to merge?

Is the answer here that the uart parameters struct takes in some opaque `UartBaud` type (which on something like an STM chip might be just aliased to a `NonZeroU32`) and...

> At this point, how controversial is it to ask that this PR add an associate type to uart which each board defines to be the valid baud rates? From...

This was [discussed on the core team call](https://github.com/tock/tock/pull/4356) again today. @bradjc drew an analogy to chip-specific restrictions on SPI `chip select` lines which are captured in types. @lschuermann echoed the...

You are correct, that register is just a (simple integer) clock divider on that chip — I've linked details in the other PR. > Should I close this PR or...

From (1), I do think this is nicer: ```diff diff --git a/boards/hail/src/io.rs b/boards/hail/src/io.rs index 72d679e57..7955cce31 100644 --- a/boards/hail/src/io.rs +++ b/boards/hail/src/io.rs @@ -38,8 +38,7 @@ impl IoWrite for Writer { if...

re tutorials: They should be modified to still work at least. They don't necessarily all need to do full credential checking, e.g. using the 'accept any shortId in the TBF...

I wonder if the better analog (and maybe solution is simply name-bikeshedding) is that `init` is closer to `mount`? The `init` is somewhat heavyweight in that it will allocate space...

> One idea I previously considered was to do the initialization sequence on the first syscall an app makes to this capsule. So that would mean that the first syscall...

> Replace ubiquitous usize with specific sized integers and explicit pointer type I remember talking about this (long, long ago) on a core team call (or two) for a while,...