This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: fold_const.c/tree_swap_operands_p change from Oct 11 breaks m68k
- From: Roger Sayle <roger at eyesopen dot com>
- To: Richard Henderson <rth at redhat dot com>
- Cc: Gunther Nikl <gni at gecko dot de>, Bernado Innocenti <bernie at develer dot com>, <gcc at gcc dot gnu dot org>
- Date: Sat, 25 Oct 2003 15:06:19 -0600 (MDT)
- Subject: Re: fold_const.c/tree_swap_operands_p change from Oct 11 breaks m68k
On Sat, 25 Oct 2003, Richard Henderson wrote:
> On Sat, Oct 25, 2003 at 01:35:54PM -0600, Roger Sayle wrote:
> > I'm not really a back-end expert. Here is the relevant define_insn
> > from line 196 of m68k.md. Can anyone see an obvious mistake?
>
> Let me introduce you to NOTICE_UPDATE_CC.
Many thanks.
> Incidentally, it's not your bug.
Indeed, but my responsibility to atleast analyse the true source of
the latent bug that I unwittingly exposed.
> Indeed, the bug appears to be that
>
> case ASHIFT: case ASHIFTRT: case LSHIFTRT:
> case ROTATE: case ROTATERT:
>
> are not special-cased at all.
These have been "#if 0"'d inside m68k.c's notice_update_cc since the
start of current CVS. Is there a way of finding out why the special
casing of these RTX codes was originally disabled?
Many thanks for your help. I guess the m68k maintainers can take it
from here, as I'd be unable to test the impact of tweaking m68k.c's
notice_update_cc.
Roger
--