Date: Thu, 13 Apr 2017 19:59:07 +0200
From: Peter Zijlstra <peterz@infradead.org>
To: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: linux-kernel@vger.kernel.org, mingo@kernel.org, jiangshanlai@gmail.com,
        dipankar@in.ibm.com, akpm@linux-foundation.org,
        mathieu.desnoyers@efficios.com, josh@joshtriplett.org,
        tglx@linutronix.de, rostedt@goodmis.org, dhowells@redhat.com,
        edumazet@google.com, fweisbec@gmail.com, oleg@redhat.com,
        bobby.prani@gmail.com, dvyukov@google.com, will.deacon@arm.com
Subject: Re: [PATCH tip/core/rcu 07/13] rcu: Add smp_mb__after_atomic() to
 sync_exp_work_done()
Message-ID: <20170413175907.hcrlh2z55zivs4v4@hirez.programming.kicks-ass.net>
References: <20170412165441.GA17149@linux.vnet.ibm.com>
 <1492016149-18834-7-git-send-email-paulmck@linux.vnet.ibm.com>
 <20170413091832.phnfppqjjy6sislo@hirez.programming.kicks-ass.net>
 <20170413161042.GA3956@linux.vnet.ibm.com>
 <20170413162409.q5gsqfytjyirgfep@hirez.programming.kicks-ass.net>
 <20170413165755.GJ3956@linux.vnet.ibm.com>
 <20170413171027.snjqn4u54t2kdzgx@hirez.programming.kicks-ass.net>
 <20170413173951.GM3956@linux.vnet.ibm.com>
 <20170413175136.5qnzvqrmzyuvlqsj@hirez.programming.kicks-ass.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170413175136.5qnzvqrmzyuvlqsj@hirez.programming.kicks-ass.net>
User-Agent: NeoMutt/20170113 (1.7.2)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1680
Lines: 41

On Thu, Apr 13, 2017 at 07:51:36PM +0200, Peter Zijlstra wrote:

> > I suppose that one alternative is the new variant of kerneldoc, though
> > very few of these functions have comment headers, let alone kerneldoc
> > headers.  Which reminds me, the question of spin_unlock_wait() and
> > spin_is_locked() semantics came up a bit ago.  Here is what I believe
> > to be the case.  Does this match others' expectations?
> > 
> > o	spin_unlock_wait() semantics:
> > 
> > 	1.	Any access in any critical section prior to the
> > 		spin_unlock_wait() is visible to all code following
> > 		(in program order) the spin_unlock_wait().
> > 
> > 	2.	Any access prior (in program order) to the
> > 		spin_unlock_wait() is visible to any critical
> > 		section following the spin_unlock_wait().
> > 
> > o	spin_is_locked() semantics: Half of spin_unlock_wait(),
> > 	but only if it returns false:
> > 
> > 	1.	Any access in any critical section prior to the
> > 		spin_unlock_wait() is visible to all code following
> > 		(in program order) the spin_unlock_wait().
> 
> Urgh.. yes those are pain. The best advise is to not use them.
> 
>   055ce0fd1b86 ("locking/qspinlock: Add comments")

The big problem with spin_unlock_wait(), aside from the icky barrier
semantics, is that it tends to end up prone to starvation. So where
spin_lock()+spin_unlock() have guaranteed fwd progress if the lock is
fair (ticket,queued,etc..) spin_unlock_wait() must often lack that
guarantee.

Equally, spin_unlock_wait() was intended to be 'cheap' and be a
read-only loop, but in order to satisfy the barrier requirements, it
ends up doing stores anyway (see for example the arm64 and ppc
implementations).