Re: 2.2.10-ac7 compile errors on Alpha

Jeff Garzik (jgarzik@pobox.com)
Fri, 02 Jul 1999 17:11:09 -0400

"Richard B. Johnson" wrote:
>
> On Fri, 2 Jul 1999, Matthew Wilcox wrote:
>
> > On Fri, Jul 02, 1999 at 07:53:40PM +0100, Alan Cox wrote:
> > > > 1) SLAB_DEBUG_SUPPORT (in mm/slab.c) does not work on Alpha (or any
> > > > other non-i386 architecture, I presume) for lack of a memchr()
> > >
> > > Care to submit a memchr 8)
> >
> > For the architectures which don't have it, let's have a na\"\ive one in C.
> > Since it's debug code, who cares, right?
> >
> > void *memchr(const void *s, int c, size_t n)
> > {
> > while (n-- != 0) {
> > if (c == (char *)s++) {
> > return s-1;
> ^^^^^^_______ here????
> > }
> > }
> > return NULL;
> > }
>
> 's' is declared as a (void *).
> What is the value of a (void *) -1 ????

See a gcc manual on void ptr arithmetic:
http://egcs.cygnus.com/onlinedocs/gcc_4.html#SEC78

> void *memchr(const void *s, int c, size_t n)
> {
> register char *p = (char *) s;
> while(n--)
> if(c == *p++)
> return (void *) (--p);
> return NULL;
> }

Isn't declaring something a 'register char' a really bad idea these
days? Something about taking away a register for the compiler to
optimize with. A smart compiler should also make 'p' completely
disappear...

Jeff

-- 
My, my, this here Anakin guy
Maybe Vader someday later, now he's just a small fry
He left his home and kissed his mommy goodbye
Saying "Soon I'm gonna be a Jedi"...


-
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/