> I think we've got two choices:
> 
> 	* Load the card's initialization routine and run it with an
> 	  emulated slot manager.
> 
> 	* Load the values from the slot manager in the booter and then
> 	  use those values at device probe time.
> 
> The former may get us into trouble if the init routines want a lot of
> the MacOS to be there.  We're already starting to emulate the basic
> functions, though, and it might not be unreasonable to work from this
> angle.  Loading the values in the booter is certainly easier, but every
> little bit that we add to the booter makes booting from a boot block
> that much harder.

It is probably also worth noting that the former option brings us a small =
step
closer to MacOS emulation.  MacOS emulation (while still unimportant =
compared to other project goals) is becoming more relevant as the '040s =
get closer and
closer to working.  Further, if Apple holds on its promise to release the
MkLinux source and we create NetBSD/macPPC, then MacOS emulation would =
become
quite desireable, IMHO.  BTW, what is the status of Apple releasing the =
MkLinux
source.  Does it look like they'll hold to their promise?

-- 
--------------------------------------------------------------------------=
------
    SMTP/MIME: jmoyer@cortland.com     |   Show your dogcow that you love =
it;
  http://daniel.cortland.com/jmoyer/   |     give it a can of Mountain =
Dew.
--------------------------------------------------------------------------=
------