This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: [PATCH] x86: {f,}xsave64 / {f,}xrstor64 / xsaveopt64 should not allow q suffix
On Fri, Nov 10, 2017 at 5:17 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>>> On 10.11.17 at 14:05, <hjl.tools@gmail.com> wrote:
>> On Fri, Nov 10, 2017 at 4:31 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>> Just like is the case for xsave{s,c}64 and xrstors64 already. I wonder
>>> though why xsave{s,c} and xrstors don't allow for the q suffix, other
>>> than the other insns without the "64" suffix do.
>>
>> commit eacc9c891d71cfef7f5c1f152291daeab785b3d4
>> Author: H.J. Lu <hjl.tools@gmail.com>
>> Date: Fri Dec 4 07:51:41 2009 +0000
>>
>> Support fxsave64 and fxrstor64.
>>
>> gas/testsuite/
>>
>> 2009-12-03 H.J. Lu <hongjiu.lu@intel.com>
>>
>> * gas/i386/i386.exp: Run x86-64-fxsave and x86-64-fxsave-intel.
>>
>> * gas/i386/rex.d: Updated for fxsave64.
>>
>> * gas/i386/x86-64-fxsave-intel.d: New.
>> * gas/i386/x86-64-fxsave.d: Likewise.
>> * gas/i386/x86-64-fxsave.s: Likewise.
>>
>> were added after fxsaveq was introduced. Otherwise, fxsaveq won't have been
>> there.
>
> I'm aware of this, but not in full agreement: For AT&T syntax, using
> the q-suffixed form is more natural. Intel syntax, not normally using
> suffixes, would indeed prefer the "64" one, even if not in line with
> the SDM / PRM.
True, if there are no *64 instructions. But if there are *64 instructions,
the *q variants should be dropped.
--
H.J.