Re: REP RET

> > Thanks a lot for the info. I guess if Microsoft uses it, then it
> > must be safe. Even if it wasn't safe before, now it will be - I
> > don't think Intel will dare break compatibility with Microsoft
> > software.
>
> Yeah, I have my doubts that REP RET will ever be used by Intel for
> anything now (except, perhaps, for the same thing AMD uses it for). It
> makes me wonder if VS 2003 didn't have this as well-- okay, checked,
> and VS 2003 didn't have REP RET in secchk.c. So this is definitely
> something new in VS 2005 (unless REP RET was used elsewhere in VS
> 2003).
>
> > I really battled to get anything out of either AMD or Intel. Intel
> > wanted me to sign an NDA, and the folks from AMD developer support
> > were clueless. :-(
>
> Yeah, the whole NDA thing was just silly, though AMD not knowing about
> it at all (since it was in their docs) was really weird.
>
> But obviously someone at Microsoft managed to get it sorted out. =)

Actually i think they did sign the NDA. In that case perhaps Intel did
indeed give them information about the future effect of REP RET.
However i still don't see why Intel would do such a thing, as this only
means advantage for their competitor.
.

Relevant Pages

  • Re: Intel cancels next-generation Xeon for even-more-next-generation Xeon
    ... >>Yeah, that was 3GIO, which became Arapahoe, which became PCI-Express. ... Of course PCI-E is no competition for HT, ... >>kept HT as simple as possible while Intel threw as much bling-bling ... >>rumours about AMD building a PCI-e link directly into its processors ...
    (comp.sys.ibm.pc.hardware.chips)
  • Re: REP RET
    ... In that case perhaps Intel did ... > indeed give them information about the future effect of REP RET. ... I guesstimate that the reason why AMD chose "rep" as the prefix is because ... the "ret" is good enough to avoid the penalty, so it is telling that AMD ...
    (borland.public.delphi.language.basm)
  • Re: [patch 13/26] Xen-paravirt_ops: Consistently wrap paravirt ops callsites to make them patcha
    ... Well most of Intel 64bit is P4 -- and Intel is still shipping ... Which generates an interrupt holdoff on the ret, ... Please read the FAQ at http://www.tux.org/lkml/ ...
    (Linux-Kernel)
  • Re: Fastcode Voting - Rep Ret Rule
    ... Intel contradicts themselves there, ... > who's to say that there's no problem with REP RET either? ... PAUSE is not an IA-32 instruction. ... It means that Intel doesn't guarantee ...
    (borland.public.delphi.language.basm)
  • Re: Fotos: Welche Anforderungen an die Grafikkarte?
    ... Der Grund warum Intel AMD nicht ganz so hart aus dem Markt drängt ist meines ... So Spielt der Markt. ... kurzfristig daran verdienen sondern auch neue Kunden gewinnen. ...
    (de.comp.hardware.graphik)