linux-2.3.3 doesn't boot when compiled with egcs-2.93.22

[Andi Kleen]

On Mon, May 24, 1999 at 07:45:34PM +0200, Martin v. Loewis wrote:
> > But will Linus like that?
> 
> It is very easy to argue that the code in question is undefined in ISO
> C, and that the "cast to word pointer" never was a gcc extension -
> earlier gcc versions just weren't so smart to exploit aliasing in
> optimizations.
> 
> Since Linus more or less accepted the existance of egcs (:-), he
> should accept that egcs requires special options to properly support
> unportable code.
> 
> If we have compiler-version dependent flags in kernel compilation,
> perhaps this is also a chance to get rid of -fno-strength-reduce - is
> this still needed?

Not needed, but it doesn't harm. From all I've seen strength-reduce 
generates worse code on i386 on all gcc versions because it increases 
register pressure.

More interesting would be to use the -m<CPU> options, to get more intelligent
alignment.





-Andi

-- 
This is like TV. I don't like TV.