On Mon Sep 28, 2026 at 7:39 PM CEST, Gary Guo wrote:
> I was quite skeptical whether the complication of adding the dma setup token 
> is
> worth the effort, but the capacity struct sounds like a very nice design as 
> it's
> infinitely extensible. I can foresee that we might want to eventually get rid 
> of
> the special `Core` typestate and just replace everything with setup
> capabilities.

I had a similar thought (which is one of the reasons I proposed it), but I don't
think we can replace the Core type state entirely.

There's more than just one bus callback that holds the Core type state, so I
think the bus setup token is rather the (much) better version of a Probe type
state.

Reply via email to