On Fri, Dec 20, 2013 at 11:30 AM, Davidlohr Bueso <davidl...@hp.com> wrote:
>
> So we'd need the barrier right after the ticket increment (ie: the xadd
> TICKET_LOCK_INC in x86), and cannot rely on the barrier after the lock
> is taken as we could miss waiters that are just spinning trying to take
> it. Is this implicitly guaranteed across all archs?
Not necessarily. But I don't see why threads spinning on it would be
special? If you spin on things, you've already updated the head
counter, so even spinners *are* visible, even if they haven't actually
gotten the lock yet.
Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
