>To keep gcc from generating a copy of the function body for every
>single module just in case some luser compiles without -O. It makes
Why is it necessarily an error to attempt to call a non-inlined version of
these functions? What's stopping us from putting a normal, extern, copy of
inb() and friends into arch/i386/lib/io.c, and into libc for the benefit of
those using them from outside the kernel? What you're saying only makes sense
if we accept your assertion that it is an error to compile without
optimization. If you believe that, why don't you submit a patch to {gcc,egcs}
to remove the -O0 option and make -O1 the default? That would take care of
all of us lusers who think we have the right to choose our CFLAGS as we see
fit...
Hoping your next message will be more enlightening,
-- Alan Curry- 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/faq.html