[PATCH] Improve PR28796

Richard Guenther richard.guenther@gmail.com
Sat Nov 4 13:38:00 GMT 2006


On 10/21/06, Richard Guenther <richard.guenther@gmail.com> wrote:
> On 8/25/06, Roger Sayle <roger@eyesopen.com> wrote:
> >
> > Hi Richard,
> >
> > On Fri, 25 Aug 2006, Richard Guenther wrote:
> > > 2006-08-25  Richard Guenther  <rguenther@suse.de>
> > >
> > >       PR target/19116
> > >       * config/i386/i386.c (override_options): Do not set MASK_IEEE_FP
> > >       if flag_unsafe_math_optimizations is specified.  We have
> > >       flag_finite_math_only for that.
> > >       * config/i386/i386.md (sqrtxf2): Do not require TARGET_IEEE_FP
> > >       or flag_unsafe_math_optimizations.
> > >
> > > Ok for mainline if we open for general bugfixes?  Otherwise ok for 4.3?
> >
> > Thanks.  Yes, this is OK for 4.3, and for mainline should Mark decide to
> > reopen it for general bug fixes.  Alas, I've not been following the recent
> > release discusssions, but I agree that a temporary return to stage2
> > would be preferable to diluting efforts with a 4.3 basic improvements
> > branch, but even that shouldn't be necessary if the regression counts
> > keep dropping.
>
> Applied to the mainline.

Ok for the branches as well?  The patch is safe enough and I guess
nobody is relying on -funsafe-math-optimizations to turn on
-ffinite-math-only on i386/x86_64 (and only there).

Thanks,
Richard.



More information about the Gcc-patches mailing list