Re: 65802 problem

[David.Empson@bbs.actrix.gen.nz]

Further to my previous posting on this topic: one possible
compatibility problem with a 65802 is interrupts.

The 6502 and 65C02 have three interrupt vectors: NMI ($FFFA), RESET
($FFFC) and BRK/IRQ ($FFFE).

The 65802 supports the same interrupt vectors when running in
'emulation' mode, but has one extra vector for the COP instruction,
which is at $FFF4.  If an $02 byte is fetched as an instruction, the
computer will fetch an interrupt vector from $FFF4, resulting in a
jump to a strange location, and probably a crash.

Even worse, if the 65802 is running in 'native' mode, there are
completely separate vectors: COP ($FFE4), BRK ($FFE6), NMI ($FFEA) and
IRQ ($FFEE).  These vectors will also point to invalid code (they are
part of the monitor's command lookup table).

This means that if you're writing code that runs in 65802 native mode
on a IIe (or IIc), you must have interrupts disabled, and make sure
you don't execute any BRK or COP instructions.  Interrupts should not
be enabled again until you return to emulation mode.

The only exception to this is if you're using the language card
yourself (maybe under DOS 3.3, but not ProDOS) - you can then set up
the vectors yourself.


The IIgs has legal values in all vectors, so it supports emulation mode
and native mode interrupts.  The IIgs always fetches interrupt vectors
from the ROM, so there is no problem with native mode interrupts under
ProDOS or Pascal.



Another possibile cause of the problem is if the RANA drive's
interface card is executing an illegal 6502 opcode.  Most of these
result in a 'no operation' on the 65C02, have peculiar effects on the
6502, but are valid instructions on the 65802.
-- 
David Empson

Internet: David.Empson@bbs.actrix.gen.nz    EMPSON_D@kosmos.wcc.govt.nz
Snail mail: P.O. Box 27-103, Wellington, New Zealand