I was politely asked to avoid that, and to use
ICM R0,B'1111', X_UNAL Sensible comment
I now wonder if there was/is a performance penalty for ICM.
Roops
---
"Mundus sine Caesaribus"
On Tue, 2 Sept 2025, 17:08 Tony Harminc, <t...@harminc.com> wrote:
> On Tue, 2 Sept 2025 at 07:22, Binyamin Dissen <
> 00001773bcccb823-dmarc-requ...@listserv.uga.edu> wrote:
>
> > You can teach an old dog new tricks.
> >
> > I never knew about this and used the PUSH approach below.
> >
>
> I probably knew of it from other contexts if I thought about it, but it
> wouldn't have been top of mind.
>
> On that note, I imagine few people looking at that explicit 0 index
> register would know why it's there (and might be tempted to "fix" it as
> part of a code review/cleanup), so I'd suggest a concise comment on each
> usage. Something like "Explicit 0 index to avoid alignment msg".
>
> Tony H.
>
> On Tue, 2 Sep 2025 20:10:29 +1000 Peter Morrison
> > <0000160d7dfdc207-dmarc-requ...@listserv.uga.edu> wrote:
> >
> > :>Hello,
> > :>
> > :> Has anyone else used (0) at the end of an RX (or RXY)
> > instruction to suppress the assembler's alignment check?
> > :>
> > :> For Example:
> > :>
> > :> DC X'00' force next field to be
> > unaligned
> > :> X_UNAL DC AL4(0) Declare an unaligned fullword
> > :>
> > :> L R0,X_UNAL
> > gets a warning
> > :> L R0,X_UNAL(0)
> > same generated code (ix reg is 0) but no warning
> > :>
> > :> If so, where is it documented? (this is easier than a
> > <PUSH ACONTROL/ACONTROL NOALIGN/instr/POP ACONTROL> sequence (and takes
> no
> > extra lines))
> > :>
> > :>Peter Morrison
> >
> > --
> > Binyamin Dissen <bdis...@dissensoftware.com>
> > http://www.dissensoftware.com
> >
> > Director, Dissen Software, Bar & Grill - Israel
> >
>
