z80asm: difference in 'SET b,(IY+d)' binary result

[electrickery]

Hi,
In a disassembly I want to convert to an asm-file for Z80pack/z80asm containing (IX+d) type instructions, at first I was confused about the signedness of the 'd' value, which isn't very well documented.

Applying formats like "(IY+-d)" works well for the LD instructions, but for the SET/RES instructions the result isn't equal to the original binary.

Original disassembly:

4E0E FD CB E9 D6  SET    2,(IY+E9)         ;...
4E12 FD CB ED B6  RES    6,(IY+ED)         ;...

Z80asm code:

        SET    2,(IY+-019h)         ;SET 2,(4369)
        RES    6,(IY+-01Dh)         ;RES 6 , (436D) => SYSTEM BC=N

Result:

4e0e  fd cb e7 d6   3914   3914         SET    2,(IY+-019h)         ;...
4e12  fd cb e3 b6   3915   3915         RES    6,(IY+-01Dh)         ;...

Here the third byte differs. How can I fix this?

Greetings & thank you for z8pack and z80asm,

Fred Jan Kraan

[udo-munk]

.fr-emoticon.fr-emoticon-img { background-repeat: no-repeat !important; font-size: inherit; height: 1em; width: 1em; min-height: 20px; min-width: 20px; display: inline-block; margin: -0.1em 0.1em 0.1em; line-height: 1; vertical-align: middle; } .fr-emoticon { font-weight: normal; font-family: "Apple Color Emoji","Segoe UI Emoji","NotoColorEmoji","Segoe UI Symbol","Android Emoji","EmojiSymbols"; display: inline; line-height: 0; } Hello Fred, the signedness of the IX/IY+d instructions is well documented, Zilog and Mostek Z80 datasheets mention this. I world suggest to use z80pack 1.38-dev available from Github: https://github.com/udo-munk/z80pack (https://github.com/udo-munk/z80pack" id="isPasted" rel="noopener noreferrer" target="_blank)and checkout the dev branch. There assembler and disassember and a lot of other things are much improvedand should fix your problem. Best regards,Udo

…

-----urspr&uuml;ngliche Nachricht----- Von: ***@***.*** ***@***.***" target="_blank) Gesendet: 02.03.2023 09:15 Uhr An: ***@***.*** ***@***.***" target="_blank) Kopie: ***@***.*** ***@***.***" target="_blank) Betreff: [udo-munk/z80pack] z80asm: difference in &#39;SET b,(IY+d)&#39; binary result (Discussion #266) Hi, In a disassembly I want to convert to an asm-file for Z80pack/z80asm containing (IX+d) type instructions, at first I was confused about the signedness of the &#39;d&#39; value, which isn&#39;t very well documented.Applying formats like &quot;(IY+-d)&quot; works well for the LD instructions, but for the SET/RES instructions the result isn&#39;t equal to the original binary.Original disassembly:4E0E FD CB E9 D6 SET 2,(IY+E9) ;...4E12 FD CB ED B6 RES 6,(IY+ED) ;...Z80asm code: SET 2,(IY+-019h) ;SET 2,(4369) RES 6,(IY+-01Dh) ;RES 6 , (436D) => SYSTEM BC=NResult:4e0e fd cb e7 d6 3914 3914 SET 2,(IY+-019h) ;...4e12 fd cb e3 b6 3915 3915 RES 6,(IY+-01Dh) ;...Here the third byte differs. How can I fix this?Greetings &amp; thank you for z8pack and z80asm,Fred Jan Kraan&mdash; Reply to this email directly, view it on GitHub (#266" target="_blank), or unsubscribe (https://github.com/notifications/unsubscribe-auth/AIYAKFHGIDUGSAHAJTCSRU3W2BJJDANCNFSM6AAAAAAVNBNFNE" target="_blank). You are receiving this because you are subscribed to this thread.<img src="https://github.com/notifications/beacon/AIYAKFGUDV4LAFW6CW64HSTW2BJJDA5CNFSM6AAAAAAVNBNFNGWGG33NNVSW45C7OR4XAZNKIRUXGY3VONZWS33OVJRW63LNMVXHIX3JMTHAASYIJQ.gif" alt="" class="fr-fic fr-dii" width="1" height="1">Message ID: ***@***.***> -----urspr&uuml;ngliche Nachricht Ende-----

[sneakywumpus]

Hello Fred,

The displacement values you use in your assembler source don't correspond to the values from the disassembly.

The displacement d for indexed addressing is a two's complement number in the range -128 <= d <= 127, i.e. 080H..0FFH for -128..-1 and 000H..07FH for 0..127. To get the positive value from a byte x that represents a negative value (080H..0FFH) calculate 0100H - x (or, alternatively NOT x + 1), i.e.

0100H - 0E9H (or, alternatively NOT 0E9H + 1 = 016H + 1) = 017H and
0100H - 0EDH (or, alternatively NOT 0EDH + 1 = 012H + 1) = 013H.

If you use these values, you get the expected third bytes.

Regards,
Thomas Eberhardt

[electrickery]

Hello Udo, Thomas,

Thank you for your fast response.

@Udo:
What I use is Linux and an up-to-date clone of the git repository (Master branch). The assembler mentions: "Z80 - Assembler Release 1.9, Copyright (C) 1987-2021 by Udo Munk".
As documentation I used an old Zaks book, but the Zilog CPU Programmer's Reference Guide, October 1982 I also have, is not much better (IMHO). But it at least taught me now that the offset (d) is in the third byte and the bit (b) is in the fourth.

@thomas:
The displacement might be wrong, but it did assemble, where the original disassembly code didn't. The solution with IY+(100h-019h) looks better than the IY+-... solution. But SET 2,(IY+100h-0E9h) still gets the third byte wron. (4e0e fd cb 17 d6). I tried several other values, but everything that would result in a third byte being higher than 7Fh fails.

4e0e  fd cb 00 d6   3914   3914         SET    2,(IY+100h-080h)         ;...
=> value out of range

Greetings,
Fred Jan

[sneakywumpus]

The assembler checks that the displacement value is in the allowed range [-128..127].
You should use the values -017h and -013h, i.e.

	SET	2,(IY+-017h) 		;SET 2,(4369)
        RES	6,(IY+-013h)		;RES 6 , (436D) => SYSTEM BC=N

If you use the "dev" branch, the assembler allows you to omit the "+", like M80, i.e. SET 2,(IY-017h)

[electrickery]

Hi Thomas,

The dev branch (Assembler Release 1.11-dev) does indeed work.;

4e0e  fd cb e9 d6   3914   3914         SET    2,(IY-(100h-0E9h))         ;...
4e12  fd cb ed b6   3916   3916         RES    6,(IY-(100H-0EDh))         ;...

Somewhat close to the original (incorrect) disassembly, but working.

Thank you & greetings,

Fred Jan