This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: Check rex_ignored


On Mon, Nov 16, 2009 at 7:38 AM, Jan Beulich <JBeulich@novell.com> wrote:
>>>> "H.J. Lu" <hjl.tools@gmail.com> 16.11.09 16:31 >>>
>>On Mon, Nov 16, 2009 at 1:07 AM, Jan Beulich <JBeulich@novell.com> wrote:
>>>>>> "H.J. Lu" <hongjiu.lu@intel.com> 14.11.09 00:14 >>>
>>>>For VEX insn, REX prefix will be ignored. I checked in this patch to
>>>>fix it.
>>>
>>> Intel® Advanced Vector Extensions Programming Reference (rev 006)
>>> section 4.1.3 says "Any VEX-encoded instruction with a REX prefix
>>> proceeding VEX will #UD", and I'd think the disassembler should express
>>> this.
>>
>>There are so ?many "#UD". I don't think the disassembler should do it.
>
> It should at least explicitly print instead of silently ignoring it.

It does:

[       ]*[0-9a-f]+:[   ]+40 c5 f9 28 00[       ]+rex vmovapd \(%rax\),%xmm0


>
>>> (Btw., would it be possible for you doc folks to include revision histories
>>> in documents like this? It's rather hard to see what changed from e.g.
>>> the previous rev of that doc.)
>>
>>If you take a look at "Order Number: 253666-031US" on the first page of
>>SDM, "031" is the revision.
>
> No, I didn't mean to add a revision number - it's been long ago I saw a
> document from Intel that didn't have one. I'm asking for having a revision
> history (i.e. an overview of what got changed).
>

I will see what I can do. Revision history can be very long over time.


-- 
H.J.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]