On Wed, 23 Jun 1999, David A. Gatwood wrote:

> On Wed, 23 Jun 1999, Bill Studenmund wrote:
> 
> > I'm not sure how much of a problem that'll be. My main thought with "bart"
> > was that it'd act as an indicator of a nubus class machine. I don't think
> > at any point we'd be going out and twiddling bits in the bart chip itself.
> > It's more that we expect to find nubus-ish cards around.
> 
> You'll probably want to enable and disable the various slots, which is
> handled by BART, as well as switching on or off burst mode and setting the
> chip's speed.  There might also be a need to reset it for some reason
> (dunno), as well as determine the chip revision to know whether to check
> for a possible PDS device instead of slot E.  I'm not sure how to do that,
> though....  Really, you should probably be able to leave that part as
> MacOS leaves it.

Which reminds me, what's the best way to build up a table with information
about NuBus cards?  Is there a standard address in each space that you can
probe and find out whether there's a card there and if so, what it is?  Is
there a way of doing it without triggering a bus error (I haven't been
able to find a way to trap TEA->Machine Check on PPC...).  Maybe I should
ask this on port-mac68k....


Later,
David