AVR-Chat

--- In AVR-Chat@yahoogroups.com, Dave VanHorn <dvanhorn@d...> wrote:
> I was there. Using 6 MHz Z-80 systems and designing 4 MHz Z-80 
systems.
> But, the undocumented opcodes were then, and still are, a gamble.
> You have no idea if the next rev will do anything similar, or even 
> anything at all.

Back 'in the day' when Apple II's, Commodore 64's, etc. were the PCs 
most mere mortals could afford, undocumented CPU op-codes were not 
uncommonly used by copy protection routines in software, the idea 
being that it would take that much more effort by someone trying to 
bypass the protection if they didn't even know what the instructions 
were!  Of course, eventually the undocumented op-codes operation 
became pretty well known, and there were even assemblers and 
disassemblers that would accept them directly.

I suppose that 'undocumented' is something of a misnomer -- it was 
really just op-codes that weren't specifically 'designed' into 
place, so what they ended up doing was considered a 'don't care,' 
although -- as others have pointed out -- sometimes they did 
something vaguely useful.  The vast majority of the time, though, 
they just did something 'weird' that wasn't particularly handy.

Apple II's and Commodore 64's eventually came out with accelerators 
and newer models that used fancier, faster processors such as the 
65C02 and the 65816 (in 6502 compatibility mode) -- this did, 
indeed, break any code that relied on these op-codes.  (In 
particular, the 65C02 took the chance to add a bunch of new op-codes 
that were oft-requested by users.)

It's interesting to me that many of these old processors never die, 
they just tend to now end up as part of some embedded system.  The 
newer low-end HP calcultors (e.g., HP 33s) use a 6502-compatible CPU 
(ostensibly because it's dirt cheap in quantity), the Z-80 lives on 
in those Rabbit Semiconductor CPUs, the 68k is now quite popular in 
PDAs, etc.

---Joel Kolstad