Re: [PATCH] X86: Disassemble primary opcode map's group 2 ModRM.reg == 6 aliases correctly

["H.J. Lu" <hjl dot tools at gmail dot com>]

On Tue, May 30, 2017 at 8:03 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>>> On 30.05.17 at 16:53, <hjl.tools@gmail.com> wrote:
>> On Tue, May 30, 2017 at 7:52 AM, H.J. Lu <hjl.tools@gmail.com> wrote:
>>> On Tue, May 30, 2017 at 7:48 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>>>>>> On 30.05.17 at 16:43, <hjl.tools@gmail.com> wrote:
>>>>> On Tue, May 30, 2017 at 7:41 AM, Jan Beulich <JBeulich@suse.com> wrote:
>>>>>>>>> On 30.05.17 at 16:35, <hjl.tools@gmail.com> wrote:
>>>>>>> On Tue, May 30, 2017 at 1:37 AM, Borislav Petkov <bp@suse.de> wrote:
>>>>>>>> On Tue, May 30, 2017 at 02:34:28AM -0600, Jan Beulich wrote:
>>>>>>>>> which I think should read
>>>>>>>>>
>>>>>>>>> "The CF flag contains the value of the last bit shifted out of the
>>>>>>>>>  destination operand; for instructions other than SAR it is undefined
>>>>>>>>>  when the count is greater than or equal to the size (in bits) of
>>>>>>>>>  the destination operand."
>>>>>>>>
>>>>>>>> Makes sense.
>>>>>>>
>>>>>>> I look another look.  Disassembler always display "shl".  Please
>>>>>>> don't use "sal" now.
>>>>>>
>>>>>> Which disassembler are you talking about, as you can't mean
>>>>>> objdump or anything else that's libopcode-based?
>>>>>
>>>>> [hjl@gnu-6 tmp]$ cat x.s
>>>>> shl %cl, %eax
>>>>> sal %cl, %eax
>>>>> shl $1, %eax
>>>>> sal $1, %eax
>>>>> [hjl@gnu-6 tmp]$ gcc -c x.s
>>>>> [hjl@gnu-6 tmp]$ objdump -dw x.o
>>>>>
>>>>> x.o:     file format elf64-x86-64
>>>>>
>>>>>
>>>>> Disassembly of section .text:
>>>>>
>>>>> 0000000000000000 <.text>:
>>>>>    0: d3 e0                 shl    %cl,%eax
>>>>>    2: d3 e0                 shl    %cl,%eax
>>>>>    4: d1 e0                 shl    %eax
>>>>>    6: d1 e0                 shl    %eax
>>>>> [hjl@gnu-6 tmp]$ objdump -dw -Mintel x.o
>>>>>
>>>>> x.o:     file format elf64-x86-64
>>>>>
>>>>>
>>>>> Disassembly of section .text:
>>>>>
>>>>> 0000000000000000 <.text>:
>>>>>    0: d3 e0                 shl    eax,cl
>>>>>    2: d3 e0                 shl    eax,cl
>>>>>    4: d1 e0                 shl    eax,1
>>>>>    6: d1 e0                 shl    eax,1
>>>>> [hjl@gnu-6 tmp]$
>>>>
>>>> Right, but here all instructions use the /4 encoding. We're talking
>>>> about how to display the /6 encoding, though, and I think it would
>>>> be helpful to the user if she could distinguish one from the other
>>>> without having to look at the opcode bytes.
>>>
>>> This is adjust an opcode alias.  These aren't the only instructions with
>>> identical opcodes.
>>>
>>
>> II meant "the same instruction with different opcodes".
>
> But here we have the obvious opportunity to make the difference
> visible. In other cases we simply have no suitable second mnemonic.
>

We are considering deprecating SAL from IA SDM.

-- 
H.J.