Re: x86 WBINVD prefix allocations

On Tue, Sep 1, 2026 at 3:16 AM Jan Beulich <[email protected]> wrote:
> >> Along the lines of the above, F2 and F3 used together and gaining
> >> meaning other than "last one takes effect" also goes against any
> >> existing use of those prefixes (afaict).
> >
> > HLE permits XACQUIRE with XRELEASE.  :-)
>
> Of course you can specify both, but not with both taking effect. Quote
> from SDM Vol 2: "If an encoded byte sequence that meets XACQUIRE/XRELEASE
> requirements includes both prefixes, then the HLE semantic is determined
> by the prefix byte that is placed closest to the instruction opcode."

yeah... not saying XA+XR or XR+XA makes sense

nor that HLE (deprecated) offers good precedence

> I.e. the same rule as for any other use of F2 vs F3.
>
> > And P5 vs P6/P4 disagree over first/last REPs.
>
> With their REP meaning, you mean?

yes – on the P5, the first REP wins, not the last one

> There wasn't any other meaning to REP prefixes yet in the P5, was there?

there was one additional quirk: REPs with LOOPs –
I added a small note for it on sandpile's 1-byte page

> > WBINVD isn't getting promoted to VEX/EVEX.
>
> I read the latter part of your initial mail as meaning to be more generic.
> Was that a mistake of mine?

for clarity: just good old WBINVD; up to 10 prefixes

to make WBINVD less of a... giant useless hammer

because giant CLFLUSHOPT loops are useless too

hth,

--
C.