Remember I'm not implementing DIAG with LA. I don't know what the difference may be between my target opcodes and what constants I could use to generate the machine code. But I do know (or at least /think/ that I know...) that if I use instructions of the same format there won't be a difference.
RIL format, for instance, is what I'm looking at right now. i ------ Original Message ------ Received: 04:44 PM COT, 12/21/2013 From: Bernd Oppolzer <[email protected]> To: [email protected] Subject: Re: macros to implement opcodes > I don't get this. What is the difference of allowed operand formats > in the register operands in LA and the numbers in the AL.4 constant ? > > Or the displacement in LA and the displacement in the Y address constant? > > Kind regards > > Bernd > > > Am 21.12.2013 22:27, schrieb Ian S. Worthington: > > As the OP, the version leveraging a similar instruction is infinitely more > > useful to me as it allows the assembler to deal with the multitude of operand > > formats rather than making me have to. > > > > i > > > > ------ Original Message ------ > > Received: 04:16 PM COT, 12/21/2013 > > From: Mark Boonie <[email protected]> > > To: [email protected] > > Subject: Re: macros to implement opcodes > > > >> Stepping back from everything that's already been suggested, when I wrote > >> this from scratch, I used something simpler: > >> > >> MACRO , > >> &LABEL DIAG &RX,&RY,&DISP > >> &LABEL DC 0H,X'83',AL.4(&RX,&RY),Y(&DISP) > >> MEND , > >> > >> The overwriting of another instruction seems unneccesarily complex. I'm > >> guessing that the intent was to allow the assembler to produce register > >> cross-reference entries, but some Diags use only Rx and not Ry, some use an > >> implied Rx+1 and Ry+1, etc., which means that (the lack of) any > >> cross-reference entries can't be trusted anyway. If the overwriting was > >> done for another reason, hopefully someone will let me know. > >> > >> - mb > >>
