> 
> Yes, but *prior* to that point, the internal SCSI would still be able
> to compete for access to main memory whereas an SBUS card would
> lose the competition to a memory bound CPU (?)

Yes - the scsi (ie sbus) only gets in while the cpu is otherwise
busy.  After all there might be enough bus idle time before the cpu
needs the SCSI data, you wouldn't want to hold the cpu off memory then
have the memory bus idle later on, now would you...
> 
> Ah, wait.  You (?) said the onboard stuff is really *also*
> part of the SBUS subsystem -- just wired at a "higher slot
> number".  In that case, the CPU *always* wins, right?
> (regardless of where the SCSI controller resides)

Higher slot numbers don't mean higher priority... The Mbus - Sbus bridge
had slot 0 (motherboard I/O) 'high' priority and allowed one of the other
slots to be set 'high priority' - nothing to do with slot number.
> 
> (it's still early in the morning... I'm a bit "slow"  :>)

'Tis mornig here now...

(FYI this is all conjecture, I never got to read the datasheet for the
relevant lump of sand)

    David