CS> So what about the case where a program has two threads doing I/O
CS> on two  serial ports and the same code is used to do I/O on both 
CS> ports? It sounds like what you two are saying is that it is  not
CS> possible for both threads to be blocked on the same  DosRead() call
CS> in the code at the same time.

  I was referring to the idea of multiple processors executing the serial
port write calls at the same time, not just multiple threads.  In any
case,  there would be little difference between one or several
processors executing the code since the application would be mostly I/O
bound.

CS> It sounds like you have actually used OS/2 SMP.  I've been waiting
CS> (and  waiting) for IBM to release the Warp SMP

  Actually, I have only been drooling so far. :-)  In the meantime I have
been evaluating the suitability of our device drivers and application
code for SMP.

CS>    LOCK
CS>     
CS>     INC     [ax]
CS> 
CS> 
CS> Does the interrupt get serviced between the LOCK and INC, or after
CS> the INC?   It seems to me that for LOCK to be truly useful,the 
CS> interrupt should be serviced after the INC. 

  The LOCK - INC executes as a single instruction; think of it as an INC with 
a modifier.

   -Mike
--- Maximus/2 2.02