There was a small bug in the initialization code that would cause some
multiport boards to not get interrupts.  I believe I've fixed this,
although I don't currently have any multiport boards I can test with.

Anyway, there's a new snapshot with just that change at:

ftp://ftp.netbsd.org/pub/NetBSD/arch/i386/970219-11.shar.gz

As usual, feedback is appreciated.