gas x86-32: allow non-absolute segment values for lcall/ljmp?

H.J. Lu hjl.tools@gmail.com
Sun Oct 4 16:48:53 GMT 2020


On Sun, Oct 4, 2020 at 9:20 AM TK Chia <u1049321969@caramail.com> wrote:
>
> Hello all,
>
> >>>> $ gas/as-new --32 test5.s
> >>>> test5.s: Assembler messages:
> >>>> test5.s:2: Error: can't handle non absolute segment in `lcall'
> >>>> ======================================================================
>
> Enclosed please find my updated proposed patch to remove this
> restriction.  I have also submitted the patch on Bugzilla
> (https://sourceware.org/bugzilla/show_bug.cgi?id=26694).
>
> gas changes:
>
> 2020-10-04  T.K. Chia  <u1049321969@caramail.com>
>
>         PR 26694
>         * config/tc-i386.c (output_interseg_jump): Allow non-absolute
>         segment operand for immediate lcall/ljmp.

The change looks straight forward.  But I don't know if it can be considered
as trivial.  Does this patch require copyright paper with the FSF?

>         * testsuite/gas/i386/jump.d,
>         * testsuite/gas/i386/jump.s,
>         * testsuite/gas/i386/jump16.d,
>         * testsuite/gas/i386/jump16.e,
>         * testsuite/gas/i386/jump16.s: Add tests for non-absolute
>         segment operand for immediate ljmp.
>
> ld changes:
>
> 2020-10-04  T.K. Chia  <u1049321969@caramail.com>
>
>         PR gas/26694
>         * testsuite/ld-i386/ljmp.s,
>         * testsuite/ld-i386/ljmp1.d,
>         * testsuite/ld-i386/ljmp1.s,
>         * testsuite/ld-i386/ljmp2.d,
>         * testsuite/ld-i386/ljmp2.s,
>         * testsuite/ld-x86-64/ljmp1.d,
>         * testsuite/ld-x86-64/ljmp2.d: New testcases.
>         * testsuite/ld-i386/i386.exp,
>         * testsuite/ld-x86-64/x86-64.exp: Run them.
>
> Thank you very much!
>
> --
> https://github.com/tkchia



-- 
H.J.


More information about the Binutils mailing list