This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: f77 on sco5 fails on GOTOFF references
- To: Richard Henderson <rth at cygnus dot com>
- Subject: Re: f77 on sco5 fails on GOTOFF references
- From: Robert Lipe <robertl at dgii dot com>
- Date: Fri, 9 Jan 1998 13:34:42 -0600
- Cc: egcs at cygnus dot com, jcardoso at inescn dot pt, scox at cygnus dot com
- References: <19971230095243.14797@dgii.com> <199712310143.UAA10553@contents-vnder-pressvre.MIT.EDU> <19971231010534.48570@dgii.com> <19971231150445.51268@dot.cygnus.com> <19980109005815.21508@dot.cygnus.com>
Richard Henderson wrote:
> On Wed, Dec 31, 1997 at 03:04:45PM -0800, Richard Henderson wrote:
> > I believe the problem to be between LEGITIMATE_PIC_OPERAND_P and the
> > pushl "i" constraint, and the fact that the i386 backend lies to gcc
> > about how pic really works. If the later were fixed (big overhaul),
> > I'm sure many things would work much better.
>
> I still believe that hiding the use of %ebx from the compiler for
> constant pool addresses is wrong, but I did not have the patience
> to walk through all of the lossage when turning that on.
That reminds me of something I'd meant to mention in this group.
Since the construct that GCC is emitting isn't even a legal Intel
addressing mode, but GAS was accepting it, I dropped a note to the
gas2 list. Ian agreed that this was bad and said he'd make the
assembler reject this syntax.
So, just as a "heads-up" to anyone using GAS x86 snapshots before
this is all worked out in egcs or gcc, this stuff may soon start
failing in a more obvious way.
RJL