2011/4/28 Jonathan Nieder <jrnieder@xxxxxxxxx>:
>
> Hm. This would be very sensitive to the compiler, since a too-smart
> optimizer could take this loop and rewrite it back to memcmp! So I
> wonder if it's possible to convey this to the compiler more precisely:
>
> return memcmp_probably_differs_early(sha1, sha2, 20);
>
> E.g., how would something like
>
> const unsigned int *start1 = (const unsigned int *) sha1;
> const unsigned int *start2 = (const unsigned int *) sha2;
>
> if (likely(*start1 != *start2)) {
> if (*start1 < *start2)
It can be a problem with unalligned access. So, IMHO, it is
better to use get_be32 here:
unsigned start1 = get_be32(sha1);
unsigned start2 = get_be32(sha2);
if (likely(start1 != start2)) {
if (start1 < start2)
...
Dmitry
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html