>>> On 01.03.17 at 20:56, wrote:
> On 01/03/17 14:19, Jan Beulich wrote:
> On 01.03.17 at 14:59, wrote:
>>> On 28/02/17 12:50, Jan Beulich wrote:
>> I also think avoiding two identical prefixes is (marginally) better
>> architecture-
>> wise.
>
> There is no specific advice in the AMD optim
On 01/03/17 14:19, Jan Beulich wrote:
On 01.03.17 at 14:59, wrote:
>> On 28/02/17 12:50, Jan Beulich wrote:
>>> Previously supported insns are being converted to the new model, and
>>> several new ones are being added.
>>>
>>> To keep the stub handling reasonably simple, integrate SET_SSE_PRE
>>> On 01.03.17 at 14:59, wrote:
> On 28/02/17 12:50, Jan Beulich wrote:
>> Previously supported insns are being converted to the new model, and
>> several new ones are being added.
>>
>> To keep the stub handling reasonably simple, integrate SET_SSE_PREFIX()
>> into copy_REX_VEX(), at once switch
On 28/02/17 12:50, Jan Beulich wrote:
> Previously supported insns are being converted to the new model, and
> several new ones are being added.
>
> To keep the stub handling reasonably simple, integrate SET_SSE_PREFIX()
> into copy_REX_VEX(), at once switching the stubs to use an empty REX
> prefi
