This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Avoid ICE in convert_modes (regression from 3.0.x)
- From: "Ulrich Weigand" <Ulrich dot Weigand at de dot ibm dot com>
- To: jakub at redhat dot com
- Cc: rth at redhat dot com, gcc-patches at gcc dot gnu dot org
- Date: Mon, 11 Mar 2002 14:22:08 +0100
- Subject: Re: [PATCH] Avoid ICE in convert_modes (regression from 3.0.x)
Hello Jakub,
>The following testcase causes ICE in convert_move.
>The problem is that convert_mode doesn't handle the case where one mode is
>floating and one is non-floating:
> if (to_real != from_real)
> abort ();
>but basic_induction_var happily passes such modes to convert_modes.
this looks like the same bug I've noticed in
http://gcc.gnu.org/ml/gcc/2002-03/msg00104.html
It was probably introduced by a recent change due to Richard Kenner, see
http://gcc.gnu.org/ml/gcc-patches/2002-03/msg00398.html
My attempted fix is a little bit different in that I avoid recursing
into the subreg into the first place. I'm not sure which fix it better.
Mit freundlichen Gruessen / Best Regards
Ulrich Weigand
--
Dr. Ulrich Weigand
Linux for S/390 Design & Development
IBM Deutschland Entwicklung GmbH, Schoenaicher Str. 220, 71032 Boeblingen
Phone: +49-7031/16-3727 --- Email: Ulrich.Weigand@de.ibm.com