Home Original page

[PATCH 1/3] x86: extend LEA's segment override warning

Jan Beulich jbeulich@suse.com
Thu Feb 13 16:38:00 GMT 2020 
On 13.02.2020 17:31, H.J. Lu wrote:
> On Thu, Feb 13, 2020 at 8:00 AM Jan Beulich <jbeulich@suse.com> wrote:
>>
>> On 13.02.2020 16:51, H.J. Lu wrote:
>>> On Thu, Feb 13, 2020 at 6:48 AM Jan Beulich <jbeulich@suse.com> wrote:
>>>>
>>>> On 13.02.2020 15:11, H.J. Lu wrote:
>>>>> On Thu, Feb 13, 2020 at 6:05 AM Jan Beulich <jbeulich@suse.com> wrote:
>>>>>>
>>>>>> For one both possible forms should be warned about. And then there are
>>>>>> a couple of MPX insns behaving LEA-like, which should be warned about in
>>>>>> the same way. Finally, to guard against future surprises, qualify the
>>>>>> original opcode check by excluding VEX/EVEX-like templates.
>>>>>>
>>>>>> gas/
>>>>>> 2020-02-XX  Jan Beulich  <jbeulich@suse.com>
>>>>>>
>>>>>>         * config/tc-i386.c (process_operands): Also check insn prefix
>>>>>>         for ineffectual segment override warning. Also cover BNDC* and
>>>>>>         BNDMK there. Don't cover possible VEX/EVEX encoded insns there.
>>>>>>         * testsuite/gas/i386/lea.s, testsuite/gas/i386/lea.d,
>>>>>>         testsuite/gas/i386/lea.e: New.
>>>>>>         * testsuite/gas/i386/i386.exp: Run new test.
>>>>>>
>>>>>
>>>>> Why should it be warning, not error?
>>>>
>>>> Because the code isn't wrong, just inefficient. I also don't think
>>>> converting from warning to error should be done in the same patch
>>>> as extending the coverage of what gets a diagnostic emitted.
>>>
>>> What do we gain to allow it?
>>
>> It has always been allowed, and in one of the two cases even silently.
>> We're liable to break people's working code if we changed this.
> 
> Given that MPX has been deprecated, MPX codes are very unlikely to
> change.  Assembler shouldn't bother with this.

What a strange position to take. Anyway, are you saying the change is
going to be okay if I remove the MPX aspects from it? (It is only
this 3rd reply of yours where you mention MPX, so I'm a little
puzzled.)

Jan
More information about the Binutils mailing list