Re: 68881 stuff

[David_McDonald@transarc.com]

Rob,

You are right that IBM uses the address in a load or store
instruction to determine the floating point operation to
perform on the MC68881.  I've forgotten the exact details
about how this works and the file server with the sources
is down so I can't look.  You should look at the IBM RT PC
Technical Reference Manual to see how you need to format
the MC68881 instruction as an address.  I believe this is
described in volume 2.  If you don't have a copy, Bob
McDivett used to have lots of them.  I believe that the
encoding included all the information needed by the MC68881
chip.

- Dave.