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

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Tue, 2015-07-14 at 00:15 +0200, Peter Zijlstra wrote:

> > This is instead the sequence that is of concern:
> > 
> > 	store a
> > 	unlock M
> > 	lock N
> > 	load b
> 
> So its late and that table didn't parse, but that should be ordered too.
> The load of b should not be able to escape the lock N.
> 
> If only because LWSYNC is a valid RMB and any LOCK implementation must
> load the lock state to observe it unlocked.

What happens is that the load passes the store conditional, though it
doesn't pass the load with reserve. However, both store A and unlock M
being just stores with an lwsync, can pass a load, so they can pass the
load with reserve. And thus inside the LL/SC loop, our store A has
passed our load B.

> > > Additionally, the assertion in Documentation/memory_barriers.txt that
> > > the sequence above can be reordered as
> > > 
> > >   LOCK N
> > >   store b
> > >   store a
> > >   UNLOCK M
> > > 
> > > is not true on any existing arch in Linux.
> > 
> > It was at one time and might be again.
> 
> What would be required to make this true? I'm having a hard time seeing
> how things can get reordered like that.