A few days ago I started to look for any hidden instructions, you know, like those found on Z80 !
Of course I began with the known (on Z80) ...
For now, I found only 1 (maybe 2) undocument eZ80 instruction which wouldn't reset my TI83PCE :
IN --,(c)
(see http://www.z80.info/z80undoc.htm)
and not in F,(c) as I badly wrote on Sax !

In order to check , I suggest this little program:


Code:


the result is %01000100 in f , that means 2 flags changed : Z and P/V

Bit 7 6 5 4 3 2 1 0
Flag S Z F5 H F3 P/V N C

Well, all that is fine but we need to confirm that no registers (amongst a,b,c,d,e,h,l, even ix and iy) have been changed after a "in --,(c)"

Let's run this:



Code:


the result is 255 , that means no registers changed !

I believe that in --,(c) does exist !.. And so does out (c),-- coded $ed,$71 (output a zero to the port)
And there might be more !

Are you sure it's not "in (hl),(bc)"?

The z80 uses three-bit codes in the lower three/six bits of a byte for registers. 0x60 is "ld h, b"; 70 is "ld (hl),b"; so it makes sense if ED60 is "in h, (bc)" that ED70 would be "in (hl),(bc).

Good hypothesis , Lirtosiast
I even had a doubt during a few seconds
I've checked, I confirm that it's actually not "in (hl),(bc)"

I've just found another one !
in0 --,(n) ( coded $ed,$30,n )

It means port ({UU,00h,n}) put to nowhere.
The flags changes the same way as in --,(c)

I can confirm that ED30 is in0 (n) and ED70 is in (bc) (or in0 f,(n) and in f,(bc) as I have seen them called, even though that is misleading). However, I tested ED71 and it is definitely a trap. Actually, now that I think about it, I can probably modify this program to find every trap....

Edit: Just checked all of the ED instructions on calc, and ED30/ED70 are the only ones that are undocumented.
Edit 2: Checked all CB insts (including indexed), sll and autocopy are all traps as expected.

thanks, Jacobly
I wish we could find some "non IO" hidden intructions ! Too bad!