[PATCH] x86: drop vex_encoding_vex2 enumerator

H.J. Lu hjl.tools@gmail.com
Fri Jan 17 14:49:00 GMT 2020


On Fri, Jan 17, 2020 at 1:03 AM Jan Beulich <jbeulich@suse.com> wrote:
>
> On 16.01.2020 23:32,  H.J. Lu  wrote:
> > How about this?
>
> For {vex2} "if possible" still leaves open what happens if this
> is not possible. Aiui there are two distinct cases:
> - 3-byte VEX could be used, in which case the prefix is ignored

Did you mean "must be used"?

> - only legacy or EVEX encoding is possible, in which case there'd
>   be an error (not really sure about EVEX here)
> I think it would be better to have {vex} and {vex3}.

Agreed.  I will make the change and keep {vex2} for backward compatibility.
How about this patch?

> Furthermore both say "for VEX instruction", leaving open what is
> to happen for non-VEX-encodable ones. It also leaves open what
> their interaction with XOP-encoded insns is.

Here is the updated document.

--
H.J.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-x86-Add-vex-pseudo-prefix.patch
Type: text/x-patch
Size: 6914 bytes
Desc: not available
URL: <https://sourceware.org/pipermail/binutils/attachments/20200117/80353545/attachment.bin>


More information about the Binutils mailing list