Re: [PATCH 3/6] Support Intel CMPccXADD

["H.J. Lu" <hjl.tools@gmail.com>]

On Wed, Nov 2, 2022 at 3:52 AM Jan Beulich <jbeulich@suse.com> wrote:
>
> On 31.10.2022 04:05, Haochen Jiang wrote:
> > --- a/gas/config/tc-i386.c
> > +++ b/gas/config/tc-i386.c
> > @@ -1098,6 +1098,7 @@ static const arch_entry cpu_arch[] =
> >    SUBARCH (prefetchi, PREFETCHI, ANY_PREFETCHI, false),
> >    SUBARCH (avx_ifma, AVX_IFMA, ANY_AVX_IFMA, false),
> >    SUBARCH (avx_vnni_int8, AVX_VNNI_INT8, ANY_AVX_VNNI_INT8, false),
> > +  SUBARCH (cmpccxadd, CMPCCXADD, ANY_CMPCCXADD, false),
>
> ANY_CMPCCXADD shouldn't have been re-introduced here. It now being
> there only means more code churn in my patch to sanitize all of this
> dependencies handling. (Unless, as I'll mention there as an option,
> we want to uniformly use ANY_... in [almost] all table entries).

We discussed it.  ANY_ doesn't hurt and it can be useful when there is
a dependency in the future.

> > --- a/opcodes/i386-opc.h
> > +++ b/opcodes/i386-opc.h
> > @@ -215,6 +215,8 @@ enum
> >    CpuAVX_IFMA,
> >    /* Intel AVX VNNI-INT8 Instructions support required.  */
> >    CpuAVX_VNNI_INT8,
> > +  /* Intel CMPccXADD instructions support required.  */
> > +  CpuCMPCCXADD,
>
> Just as a remark: Personally I would have considered it more readable
> overall if this had the "CC" infix in lower case (like in the comment).
>
> Jan



-- 
H.J.