[PATCH 2/2] x86: adjust disassembly of insns operating on selector values

[H.J. Lu]

On Mon, Jul 17, 2023 at 12:08 AM Jan Beulich <jbeulich@suse.com> wrote:
>
> On 17.07.2023 04:04, Jiang, Haochen wrote:
> >> --- a/gas/testsuite/gas/i386/opcode-suffix.d
> >> +++ b/gas/testsuite/gas/i386/opcode-suffix.d
> >> @@ -102,7 +102,7 @@ Disassembly of section .text:
> >>   *[0-9a-f]+:        60[     ]+pushal
> >>   *[0-9a-f]+:        61[     ]+popal
> >>   *[0-9a-f]+:        62 90 90 90 90 90[      ]+boundl %edx,-0x6f6f6f70\(%eax\)
> >> - *[0-9a-f]+:        63 90 90 90 90 90[      ]+arpl[         ]+%dx,-0x6f6f6f70\(%eax\)
> >> + *[0-9a-f]+:        63 90 90 90 90 90[      ]+arpll[        ]+%edx,-
> >> 0x6f6f6f70\(%eax\)
> >>   *[0-9a-f]+:        68 90 90 90 90[         ]+pushl[        ]+\$0x90909090
> >>   *[0-9a-f]+:        69 90 90 90 90 90 90 90 90 90[  ]+imull[
> >>      ]+\$0x90909090,-0x6f6f6f70\(%eax\),%edx
> >>   *[0-9a-f]+:        6a 90[  ]+pushl[        ]+\$0xffffff90
> >> @@ -248,7 +248,7 @@ Disassembly of section .text:
> >>   *[0-9a-f]+:        fc[     ]+cld
> >>   *[0-9a-f]+:        fd[     ]+std
> >>   *[0-9a-f]+:        ff 90 90 90 90 90[      ]+calll[        ]+\*-0x6f6f6f70\(%eax\)
> >> - *[0-9a-f]+:        0f 00 90 90 90 90 90[   ]+lldt[         ]+-0x6f6f6f70\(%eax\)
> >> + *[0-9a-f]+:        0f 00 90 90 90 90 90[   ]+lldtw[        ]+-0x6f6f6f70\(%eax\)
> >
> > H.J. is on vacation till Wednesday but I suppose H.J. has this question in my patch
> > before and he might raise again if he is here.
> >
> > Is this suffix needed since m16 is the only allowed memory here?
>
> All suffixes should be appended in suffix-always mode. The question
> of whether one can be omitted arises only in the default mode, where
> optional suffixes are left off (for clarity / ease of reading). Plus
> please recall that the changes here are for consistency, and e.g. in
>
>  *[0-9a-f]+:    8c 90 90 90 90 90[      ]+movw[         ]+%ss,-0x6f6f6f70\(%eax\)
>
> the suffix is also present, no matter that only m16 is possible.
>
> In any event I'm not intending to commit this before the end of the
> week, so H.J. will have a chance to voice his opinion.
>
> Jan

Personally, I don't think this change is necessary.  But I am not
against it.

-- 
H.J.