lang: two protocols with same name, defined over differents structs

We have some protocols that do the following:
```
#[idle]
prot idle() {
// ....
}

#[idle]
prot idle() {
// ...
}
```

This seems like a totally legitimate thing to do. However, it kind of violates some rules around duplicate names that one would expect from a programming language. Namely that you cannot have two functions in the same scope that are both named `idle`.

Maybe it is time to switch to more of a `impl` style protocol definition. Like:

```
impl ReadSubordinate {
#[idle]
prot idle(DUT: Self) {
// ....
}
}

impl WriteSubordinate{
#[idle]
prot idle(DUT: Self) {
// ....
}
}
```

I don't necessarily like the `DUT: Self`. Another option would be a magic `self` or even a magic `dut`.

No implementation file, test, or entry point is named. Start by locating the parser and name-resolution code for protocol definitions and the existing duplicate-name checks. Done means agreeing on the protocol syntax and defining how same-named protocols over different structs are represented and validated.

Written by the indexing model from the issue text.