| Date | Fri, 31 Aug 2012 11:40:56 -0700 | From | Josh Triplett <> | Subject | Re: [PATCH tip/core/rcu 12/15] rcu: Remove redundant memory barrier from __call_rcu() |
| |
On Fri, Aug 31, 2012 at 11:30:35AM -0700, Josh Triplett wrote: > On Thu, Aug 30, 2012 at 11:56:25AM -0700, Paul E. McKenney wrote: > > From: "Paul E. McKenney" <paul.mckenney@linaro.org> > > > > The first memory barrier in __call_rcu() is supposed to order any > > updates done beforehand by the caller against the actual queuing > > of the callback. However, the second memory barrier (which is intended > > to order incrementing the queue lengths before queuing the callback) > > is also between the caller's updates and the queuing of the callback. > > The second memory barrier can therefore serve both purposes. > > > > This commit therefore removes the first memory barrier. > > I don't see any such second memory barrier in __call_rcu(), at least not > in current master. Right after this smp_mb(), __call_rcu() enqueues the > callback and increments the queue length. > > Did you add a second memory barrier in some other patch that hasn't made > it upstream yet? If so, could you note that patch dependency explicitly > in the commit message?
Argh, nevermind. Looked at the wrong branch, not master. Looking at master, I do indeed see the second smp_mb().
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
|