> I recall that there was a technique using two 68000 processors that is
> different from the one already mentioned. This techique would use the
> first 68000 to run most code, and then when it hits a bus error it is
> simply stalled. The second processor fixes up the mess and then lets
> the first one continue.

The 68000 can't resume instruction execution after an exception in the
middle of the instruction.  The internal CPU state is not completely
saved on stack, so the CPU can't finish the instruction after the
exception is complete.  68010 _can_ resume execution, so 68010 made
virtual memory possible.

The trick of using two 68000's had to be used to get around the exception
limitation.  I can't recall the details, but the only way I can think of
is to run one CPU behind the other.  When the first one encounters a bus
error (or anything that requires swapping data from disk to RAM), the
condition is handled and then the second CPU runs the instruction that
caused the exception.  Or something like that. :-)

  -jm