Re: [RFC PATCH v2] memory-barriers: remove smp_mb__after_unlock_lock()

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Mon, Jul 13, 2015 at 06:50:29PM +0100, Will Deacon wrote:
> On Mon, Jul 13, 2015 at 04:54:47PM +0100, Peter Zijlstra wrote:
> > However I think we should look at the insides of the critical sections;
> > for example (from Documentation/memory-barriers.txt):
> > 
> > "       *A = a;
> >         RELEASE M
> >         ACQUIRE N
> >         *B = b;
> > 
> > could occur as:
> > 
> >         ACQUIRE N, STORE *B, STORE *A, RELEASE M"
> > 
> > This could not in fact happen, even though we could flip M and N, A and
> > B will remain strongly ordered.
> > 
> > That said, I don't think this could even happen on PPC because we have
> > load_acquire and store_release, this means that:
> > 
> > 	*A = a
> > 	lwsync
> > 	store_release M
> > 	load_acquire N
> > 	lwsync
> > 	*B = b
> > 
> > And since the store to M is wrapped inside two lwsync there must be
> > strong store order, and because the load from N is equally wrapped in
> > two lwsyncs there must also be strong load order.
> > 
> > In fact, no store/load can cross from before the first lwsync to after
> > the latter and the other way around.
> > 
> > So in that respect it does provide full load-store ordering. What it
> > does not provide is order for M and N, nor does it provide transitivity,
> > but looking at our documentation I'm not at all sure we guarantee that
> > in any case.
> 
> So if I'm following along, smp_mb__after_unlock_lock *does* provide
> transitivity when used with UNLOCK + LOCK, which is stronger than your
> example here.

Yes, that is indeed the intent.

> I don't think we want to make the same guarantee for general RELEASE +
> ACQUIRE, because we'd end up forcing most architectures to implement the
> expensive macro for a case that currently has no users.

Agreed, smp_mb__after_unlock_lock() makes a limited guarantee.

> In which case, it boils down to the question of how expensive it would
> be to implement an SC UNLOCK operation on PowerPC and whether that justifies
> the existence of a complicated barrier macro that isn't used outside of
> RCU.

Given that it is either smp_mb() or nothing, I am not seeing the
"complicated" part...

							Thanx, Paul