LGTM, make this much clearer.

@9names @jannic I added a bypass for the workaround if it is running on rev B2 and above.
@dlkj would you mind testing it with your system?

@9names @jannic I added a bypass for the workaround if it is running on rev B2 and above.

I'm not 100% sure if I like this. This can lead to situations were a firmware works perfectly for the developer, who happens to use a B2 stepping, and breaks for users with older chips.

In that sense, I'd prefer something like #608: Make it a guaranteed panic if you have the e5 workaround enabled and forget to initialize bank0 before USB.
(That code is completely untested - please just treat it as an idea.)

And I forgot to say: The update of the doc comment is great!

I'm not 100% sure if I like this. This can lead to situations were a firmware works perfectly for the developer, who happens to use a B2 stepping, and breaks for users with older chips.

In that sense, I'd prefer something like #608: Make it a guaranteed panic if you have the e5 workaround enabled and forget to initialize bank0 before USB.
(That code is completely untested - please just treat it as an idea.)

Oh I didn't think about this indeed, you're totally right.
Do you mind if I incorporate your suggestions to this PR?

I think bypassing the workaround is a good thing on B2 revs to avoid messing needlessly with the hardware but checking correct implementation with the requirements of the feature make sense too.

Do you mind if I incorporate your suggestions to this PR?

Of course not, go ahead!

@dlkj have you been able to test this last iteration?
If it works as expected,m for you I'll press merge :)

I've not had the time to check this yet. I'll find some time this week probably.

I can't reproduce the panic now using the same pico and a very similar code base. LGTM.