Re: [PATCH v2] tools/memory-model: Add extra ordering for locks and remove it for ordinary release/acquire

[Peter Zijlstra]

On Thu, Jul 12, 2018 at 10:28:38AM -0700, Paul E. McKenney wrote:
> Look for the uses of raw_spin_lock_irq_rcu_node() and friends in
> kernel/rcu and include/linux/*rcu*, along with the explanation in
> Documentation/RCU/Design/Memory-Ordering/Tree-RCU-Memory-Ordering.html
> 
> I must confess that I am not following exactly what Peter is calling
> out as the failure.  My best guess is that he is leading up to his
> call for RCsc locks, but I might have missed a turn.

Yes, it is part of my call for RCsc locks.

The locking pattern is fairly simple and shows where RCpc comes apart
from expectation real nice.

Yes I helped clean up the barrier usage, but as always it'd be even
cleaner if we could just do away with the special barrier entirely.