This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: ppc completely soft patches
> > libgcc is not only slower, but does not handle exceptions or rounding
> > modes.
>
> Why can't it be enhanced to do this?
Well, mainly laziness. The proper solution would be to use the
soft-fp/ infrastructure in glibc in gcc. We'd gain speed, and
functionality. And we'd get not 1 not 2, but 3 copies of soft-fp
(linux kernel, libgcc, and glibc), sigh.
> > And even if it did handle, say rounding modes-- how would you set it?
> > Glibc's <fenv.h> routines set that.
>
> You'd define an interface that libgcc provides---perhaps a global
> variable, perhaps some nice thread-safe functions---and that fenv.h uses.
That's not a bad idea, actually. I don't have the cycles right now to
do so, but I'm thinking it would definitely be more maintainable, and
any port could use it, not just PPC like it is right now.
> If you've decided that these routines should live in glibc, then the
> proper way to deal with this is to not generate them in libgcc.
> Having one copy in libgcc and another copy in libc with different
> semantics is just asking for trouble.
Point taken.
Aldy