[PATCH] Support Intel AVX10.2 BF16 instructions

[Jiang, Haochen]

> From: Haochen Jiang <haochen.jiang@intel.com>
> Sent: Tuesday, October 15, 2024 3:42 PM
> 
>   - For vgetexppbf16, the encoding in the current SPEC will probably be
>     changed to the form that Jan pointed out in Intel forum. I have got
>     a patch to change them but not squashed into this patch yet since
>     the doc haven't been revised yet. Part of the patch would be:
> 
> diff --git a/opcodes/i386-dis-evex-prefix.h b/opcodes/i386-dis-evex-prefix.h
> index d98b761facb..11893818d73 100644
> --- a/opcodes/i386-dis-evex-prefix.h
> +++ b/opcodes/i386-dis-evex-prefix.h
> @@ -561,6 +561,12 @@
>      { Bad_Opcode },
>      { "vscalefp%XH",      { XM, Vex, EXxh, EXxEVexR }, 0 },
>    },
> +  /* PREFIX_EVEX_MAP6_42 */
> +  {
> +    { "vgetexpp%XB",   { XM, EXxh }, 0 },
> +    { Bad_Opcode },
> +    { "vgetexpp%XH",   { XM, EXxh, EXxEVexS }, 0 },
> +  },
>    /* PREFIX_EVEX_MAP6_4C */
>    {
>      { "vrcpp%XB",    { XM, EXxh }, 0 },
> 
> 
>     I suppose this part would show the change after the doc update.
> 
>     The doc might be updated within a month. Maybe some wait is needed
>     for that. If the wait is long, let's say after everything else solved
>     in this patch, I may send satcvt patch out and make this patch pending
>     for that change.
> 

Doc change happens just within 24 hours, I will send out a v2 patch reflecting
this change soon.

Thx,
Haochen