mark lewis wrote to Will Honea on 03-25-1998
ml> your testing shows much promise and things may not be as 
ml> bad as they seem(ed)... i gotta wonder, now, about the COM 
ml> ports ;-)
I spent several years qualifying hardware for a client to ship, so the
testing cycle got very short after a time - I fire up one machine and
let it do the tests for each board.  Some of what I wrote was a
compilation of lots of past tests as well.  The printer port has never
been much of a problem, but the comm ports, that's a whole new bucket
of worms.  So far, every mb that I've tested with integrated comm ports
has properly disabled them when asked.  Older add-in i/o boards would
frequently disable the port but leave it on the bus so that nothing
else could use the port address - at least the IRQ's were usually
jumpers!  0.10 inch air gaps are very effective disabling methods!
As for the integrated boards, I have had a couple with AWARD bios
chips that allowed IRQ 2/9, 3, 4, and 5 in various combinations.  I
gotten to where I either disable them completely of plan on assigning
my plugin cards to use other than 3 and 4.  I agree, it's a PITA but
here we are: if you plan on using the builtins on those IRQs and
everything else get other IRQs you should be OK.
One caveat about printer ports:  the accepted definions are that port
addresses 03BC and (I may be bass-ackwards on the next 2) 0378 use IRQ
7, while 0278 uses IRQ 5.  At least, the general consensus here seems
to support that with the possible exception that I reversed the last 2.
 Unfortunately, some mother boards seem to map ANY printer port to IRQ
7, so you may have an OS/2 problem if you guess wrong.
In any event, you can probably get by with using the printer IRQ for
something else if you defeat OS/2's use of it but the comm ports are
gonna bite you unless you just plain disable them or use them as
default com1/com2.
Will Honea 
--- Maximus/2 2.02
---------------