This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: [PATCH] x86: VZEROALL/VZEROUPPER are WIG
- From: "H.J. Lu" <hjl dot tools at gmail dot com>
- To: Jan Beulich <JBeulich at suse dot com>
- Cc: Binutils <binutils at sourceware dot org>
- Date: Fri, 14 Sep 2018 13:04:57 -0700
- Subject: Re: [PATCH] x86: VZEROALL/VZEROUPPER are WIG
- References: <5B8F9D8402000078001E5687@prv1-mh.provo.novell.com> <CAMe9rOoMSaZOfMKnSK7DNsrBYk4xH+ktvU=WKW3aNbxOuZSBSg@mail.gmail.com>
On Wed, Sep 12, 2018 at 2:03 PM, H.J. Lu <hjl.tools@gmail.com> wrote:
> On Wed, Sep 5, 2018 at 2:10 AM, Jan Beulich <JBeulich@suse.com> wrote:
>> opcodes/
>> 2018-09-05 Jan Beulich <jbeulich@suse.com>
>>
>> * i386-dis.c (VEX_W_0F77_P_0): Delete.
>> * (prefix_table): Use VZERO here.
>> * (vex_w_table): Delete VEX_W_0F77_P_0 entry.
>> ---
>> Note regarding the lack of a testcase addition: I would have added
>> something if there was a pre-existing testcase checking VEX-encoded WIG
>> insns, but something like this exists for EVEX encoded ones only
>> (including a command line option to control the choice).
>>
>>
>
> I am adding the -mvexwig= option. Let's wait until it is in.
>
Please check out users/hjl/wig branch:
https://sourceware.org/git/?p=binutils-gdb.git;a=shortlog;h=refs/heads/users/hjl/wig
Many -mvexwig=1 tests fail due to disassembler bug:
FAIL: i386 AVX WIG insns with -mvexwig=1
FAIL: i386 AVX2 WIG insns with -mvexwig=1
FAIL: x86-64 AVX WIG insns with -mvexwig=1
FAIL: x86-64 AVX2 WIG insns with -mvexwig=1
--
H.J.