On Mon, 20 Sep 1999 14:01:51 -0400 
 Dan Winship <danw@MIT.EDU> wrote:

 > OK. It seems like the Right Thing would be to use "badaddr()" as part
 > of the check if it's done resetting, rather than just sleeping an
 > arbitrarily chosen amount of time. But that would require having
 > badaddr() on all ports. Should I just commit it with the DELAY or try
 > to do something cleverer?

Actually, probably the Right Thing to do is to read the Adaptec manual
and determine how many bus cycles the chip is unresponsive for, and
delay that long + some slop.

badaddr() isn't a MI function, and you can't even make that determination
on some systems.

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>