> According to the egcs people, the aliasing framework has
> been integrated into egcs since 1.1, but it had not been
> enabled by default. gcc-2.95 (AKA egcs-1.2) will switch
> it on by default. So, you have not yet seen the true
> size of the problem.
Yes, I have. i586, _one_ function miscompiled as far as I could find
out. Kernel wouldn't boot, BTW.
> Both, Dave Miller and Andy Kleen, tried to fix the TCP
> code and gave up on it. This seems to indicate much more
> work than just a few fixed .h files and explains why Linus
> pressed so strong (and quite futile) to change the
> egcs behavior.
OK, use -fno-strict-aliasing then for the time being, clean up as you
go. It will have to be done someday anyway...
-- Dr. Horst H. von Brand mailto:vonbrand@inf.utfsm.cl Departamento de Informatica Fono: +56 32 654431 Universidad Tecnica Federico Santa Maria +56 32 654239 Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/