OK, in this code it mostly appears that the wraparound is correct, but
in some cases it is very hard to say without specific knowledge of the code:
neigh->confirmed = jiffies - (n->parms->base_reachable_time<<1);
.
.
.
if ((state&NUD_VALID) &&
now - neigh->confirmed < neigh->parms->reachable_time) {
How are we to know that the above calculation won't be wrong because of
jiffies wrap?
> Please show me a bad case where casting to signed is necessary.
I don't know enough about the net code to say for sure, but for example
in drivers/sound/sb_common.c:
limit = jiffies + HZ / 10; /* Timeout */
for (i = 0; i < 500000 && (limit-jiffies)>0; i++)
Since limit and jiffies are both unsigned, the value will always be > 0,
unless they are equal.
> I actually ran through the tree the other night myself starting to
> convert these things, then I noticed that I couldn't even convince
> myself that the code was incorrect.
I don't disagree that it is possible to make correct code without the
macros, but it is easier to guarantee that it IS correct with the macros.
Rather than evaluate each jiffies usage on a case-by-case basis, it is
much easier to do a code audit and fix all comparisons. I don't see that
it harms anything to use the macros instead.
Cheers, Andreas
-- Andreas Dilger http://sourceforge.net/projects/ext2resize/ http://www-mddsp.enel.ucalgary.ca/People/adilger/- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/