[PATCH v3 10/10] x86-64: Intel64 adjustments for insns dealing with far pointers

[H.J. Lu]

On Wed, Jan 15, 2020 at 9:03 AM Jan Beulich <jbeulich@suse.com> wrote:
>
> On 15.01.2020 14:28, H.J. Lu wrote:
> > On Tue, Jan 14, 2020 at 11:49 PM Jan Beulich <jbeulich@suse.com> wrote:
> >>
> >> On 14.01.2020 18:05, H.J. Lu wrote:
> >>> On Tue, Jan 14, 2020 at 7:00 AM Jan Beulich <jbeulich@suse.com> wrote:
> >>>>
> >>>> On 13.01.2020 18:42, H.J. Lu wrote:
> >>>>> On Fri, Dec 27, 2019 at 1:27 AM Jan Beulich <JBeulich@suse.com> wrote:
> >>>>>>
> >>>>>> AMD and Intel differ in their handling of far indirect branches as well
> >>>>>> as LFS/LGS/LSS: AMD CPUs ignore REX.W while Intel ones honors it. (Note
> >>>>>> how the latter three were hybrids so far, while far branches were fully
> >>>>>> AMD-like.)
> >>>>>>
> >>>>>> gas/
> >>>>>> 2020-01-XX  Jan Beulich  <jbeulich@suse.com>
> >>>>>>
> >>>>>>         PR gas/24546
> >>>>>>         * config/tc-i386.c (match_template): Apply AMD64/Intel64 check
> >>>>>>         to 64-bit code only.
> >>>>>>         * config/tc-i386.c (i386_intel_operand): Also handle CALL/JMP in
> >>>>>>         O_tbyte_ptr case.
> >>>>>>         * testsuite/gas/i386/x86-64-branch-3.s,
> >>>>>>         testsuite/gas/i386/x86-64-intel64.s: Add 64-bit far call cases.
> >>>>>>         * testsuite/gas/i386/x86-64-branch-3.d,
> >>>>>>         testsuite/gas/i386/x86-64-intel64.d: Adjust expectations.
> >>>>>>         * testsuite/gas/i386/x86-64-branch-5.l,
> >>>>>>         testsuite/gas/i386/x86-64-branch-5.s: New.
> >>>>>>         * testsuite/gas/i386/i386.exp: Run new test.
> >>>>>>
> >>>>>> opcodes/
> >>>>>> 2020-01-XX  Jan Beulich  <jbeulich@suse.com>
> >>>>>>
> >>>>>>         PR gas/24546
> >>>>>>         * i386-dis.c (putop): Handle REX.W in '^' case for Intel64 mode.
> >>>>>>         * i386-opc.tbl (lfs, lgs, lss, lcall, ljmp): Split into
> >>>>>>         AMD64 and Intel64 templates.
> >>>>>>         (call, jmp): Likewise for far indirect variants. Dro
> >>>>>>         Unspecified.
> >>>>>>         * i386-tbl.h: Re-generate.
> >>>>>
> >>>>> Please add some documentations to describe how they are used in
> >>>>>  AMD64 and Intel64 in AT&T/Intel syntax.
> >>>>
> >>>> There's nothing unusual or unexpected here. I wouldn't even know
> >>>> where this would belong - are there descriptions like what you
> >>>> ask for somewhere already for a fair set of other insns? I don't
> >>>> recall any similar additions for any half way recent ISA
> >>>> extensions, and there it might be better justified to supply
> >>>> such than it is here. I guess I'm confused by the request ...
> >>>
> >>> We should document how new far pointer branches are used
> >>> in Intel and AT&T syntax.
> >>
> >> I have no idea at all what you're after. Could you give me a
> >> pointer to existing text I could use to clone or at least refer
> >> to, which then also would give me a hint where in the docs I
> >> should insert such a piece of information which - as said - I
> >> don't think has a need to be there in the first place? Once
> >> again
> >> - usage is consistent with other insns (none of which have any
> >>   such piece of documentation afaict),
> >> - we shouldn't duplicate SDM / PM information.
> >>
> >
> > The current assembler supports far pointer branches.  Your change
> > added a new one.  You should document how it differs from the
> > existing one.
>
> This
>
> @cindex return instructions, i386
> @cindex i386 jump, call, return
> @cindex return instructions, x86-64
> @cindex x86-64 jump, call, return
> @item
> Immediate form long jumps and calls are
> @samp{lcall/ljmp $@var{section}, $@var{offset}} in AT&T syntax; the
> Intel syntax is
> @samp{call/jmp far @var{section}:@var{offset}}.  Also, the far return
> instruction
> is @samp{lret $@var{stack-adjust}} in AT&T syntax; Intel syntax is
> @samp{ret far @var{stack-adjust}}.
>
> and this
>
> @cindex jump instructions, i386
> @cindex call instructions, i386
> @cindex jump instructions, x86-64
> @cindex call instructions, x86-64
> Far call/jump instructions are @samp{lcall} and @samp{ljmp} in
> AT&T syntax, but are @samp{call far} and @samp{jump far} in Intel
> convention.
>
> is what I've been able to find. Neither covers - for existing
> forms - what you ask for, so I'm still entirely unclear what

They do cover existing far branches, lcall and ljmp in AT&T syntax.
Since you are adding new far call and jmp, how should they be used?

> form of documentation of what specific aspects you're after.
> Sadly you didn't address my prior request to point me at
> something that I could go from. I'm afraid if you want me to
> do this, you'll need to be more specific. An alternative might
> be for you to put together whatever you think needs documenting.


-- 
H.J.