Re: [PATCH 4/4] locking: Introduce smp_cond_acquire()

From: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
To: Boqun Feng <boqun.feng@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>,
	mingo@kernel.org, linux-kernel@vger.kernel.org, corbet@lwn.net,
	mhocko@kernel.org, dhowells@redhat.com,
	torvalds@linux-foundation.org, will.deacon@arm.com,
	Michael Ellerman <mpe@ellerman.id.au>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>,
	Paul Mackerras <paulus@samba.org>
Subject: Re: [PATCH 4/4] locking: Introduce smp_cond_acquire()
Date: Thu, 12 Nov 2015 07:02:51 -0800	[thread overview]
Message-ID: <20151112150251.GZ3972@linux.vnet.ibm.com> (raw)
In-Reply-To: <20151112144902.GA4549@fixme-laptop.cn.ibm.com>

On Thu, Nov 12, 2015 at 10:49:02PM +0800, Boqun Feng wrote:
> On Thu, Nov 12, 2015 at 06:40:04AM -0800, Paul E. McKenney wrote:
> [snip]
> > 
> > I cannot resist suggesting that any lock that interacts with
> > spin_unlock_wait() must have all relevant acquisitions followed by
> > smp_mb__after_unlock_lock().
> > 
> 
> But
> 
> 1.	This would expand the purpose of smp_mb__after_unlock_lock(),
> 	right? smp_mb__after_unlock_lock() is for making UNLOCK-LOCK
> 	pair global transitive rather than guaranteeing no operations
> 	can be reorder before the STORE part of LOCK/ACQUIRE.

Indeed it would.  Which might be OK.

> 2.	If ARM64 has the same problem as PPC now,
> 	smp_mb__after_unlock_lock() can't help, as it's a no-op on
> 	ARM64.

Agreed, and that is why we need Will to weigh in.

							Thanx, Paul