Re: [PATCH tip/core/rcu 9/9] rcu: Provide counterpart to rcu_dereference() for non-RCU situations

[Oleg Nesterov <oleg@redhat.com>]

On 10/29, Paul E. McKenney wrote:
>
> On Wed, Oct 29, 2014 at 11:57:04AM +0100, Peter Zijlstra wrote:
> > On Tue, Oct 28, 2014 at 03:10:02PM -0700, Paul E. McKenney wrote:
> > > From: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
> > >
> > > Although rcu_dereference() and friends can be used in situations where
> > > object lifetimes are being managed by something other than RCU, the
> > > resulting sparse and lockdep-RCU noise can be annoying.  This commit
> > > therefore supplies a lockless_dereference(), which provides the
> > > protection for dereferences without the RCU-related debugging noise.
> > >
> > > Reported-by: Al Viro <viro@ZenIV.linux.org.uk>
> > > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > > ---
> >
> > > +#define lockless_dereference(p) \
> > > +({ \
> > > +	typeof(p) _________p1 = ACCESS_ONCE(p); \
> > > +	smp_read_barrier_depends(); /* Dependency order vs. p above. */ \
> > > +	(_________p1); \
> > > +})
> >
> > Should we not have at least a single user along with this?
>
> And we do.  In fact, Al Viro has pulled this into his vfs.git tree and
> so I will be dropping this patch in favor of his.

And it seems that most of smp_read_barrier_depends() users can be changed
to use this helper.

Oleg.