>>>>> "Bruce" == Bruce Albrecht <bruce@zuhause.mn.org> writes:

Bruce> I vote for keeping the SBIC driver as is, although I think that
Bruce> it should either be modified to disable sync mode on an unit
Bruce> basis, OR make disable sync mode a kernel configuration
Bruce> parameter.

Back in the early days, it *was* settable on a unit basis.  When the
change came to allow multiple sbic-based cards someone was lazy and
made it a single static.  I changed it on my system to be per-unit
settable again, but was also lazy and never commited the patches.  Now
I own a Zeus so I don't actually use it myself.  However if people
like to use the patches they're free to visit
http://www.cd.chalmers.se/~nh/netbsd.html and fetch the NH-amiga diffs
and carefully pick out the pieces relating to sbic.  NOTE! Don't apply
all of this diff!  I make no guarantees whatsoever that it will work
for you.  Esp. it contains the ISA kit and a new interruptsystem not
ready for non-hackerish work!

I hope what I've said is true (ie. that the sync-per-unit patches are
in there), it ought to be, anyway...  If it isn't, maybe MTK has it
still.  MTK?

On the other hand I think that all binpatchable variables should have
a kernel compiletime configurable default...

Niklas