On Tue, 3 Mar 1998 15:15:04 -0500 (EST),
mrz5149@cs.rit.edu (Michael R Zucca) writes:
> 
> However, the problem with IDE support isn't so much finding the base
> address. The problem from what I've heard is that Apple's implementation
> of IDE is kind of wacky and the chips they use aren't your every-day
> IDE chips. At least with SCSI we're working with something known...sort of. :)

Are you sure about this?  The vanilla IDE interface is just a 16-bit
wide port with a couple of lower address bits and control signals
attached.  EIDE (or ATA-2/3) has much more sophisticated capabilities
hence much more complicated (motherboard chipset must support it in
order to exploit all the capabilities), but it *can* be used as the
plain old IDE (much the same way on the AV Macs SCSI DMA is there but
ignored and SCSI I/O is driven by interrupt and/or polling).  The host
adapter doesn't have to be that sophisticated at all.  All you need
besides address decode logic is a couple of standard logic ICs,
something like two 245s and a 244 (or a 1645 and a 644).  Of course,
on your PC motherboard they are usually embedded in the chipset, so
you don't actually get to see these numbers.

This leads me to suspect IDE support on mac68k is a matter of finding
the port address and the interrupt vector, although I claim no
knowledge on Apple's IDE implementation.

Ken