On Thu, Jan 22, 2004 at 11:12:39AM -0500, Allen Briggs wrote:
> On Thu, Jan 22, 2004 at 10:21:45AM -0500, Daniel R. Killoran,Ph.D. wrote:
> > There is no obvious reason why one NuBus slot should be more 
> > important than another, so they are wired together.
> > The software can control their relative importance by the order in 
> > which they are interrogated.
> 
> Well, kind of.  There are other ways to get that flexibility.  They
> give each nubus slot some dedicated address space, it would make
> as much sense to give each slot a dedicated interrupt.  It would
> also be more robust because you could disable individual cards for
> whatever reason.
> 
> But unless you were there at the time, it's hard to say for sure what
> the reasoning is--we just have to deal with it now...  :-)

Can we keep the interrupt engine disabled and run the hardware strictly
through polling?  (Kinda like OS/2 tried to do when it could AFAIK)

Joshua Coombs