Re: perf: aux area related crash and warnings

From: Peter Zijlstra <peterz@infradead.org>
To: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@kernel.org>,
	Vince Weaver <vincent.weaver@maine.edu>,
	linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	Stephane Eranian <eranian@gmail.com>,
	"Paul E. McKenney" <paulmck@us.ibm.com>
Subject: Re: perf: aux area related crash and warnings
Date: Thu, 18 Jun 2015 12:24:10 +0200	[thread overview]
Message-ID: <20150618102410.GJ19282@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <87bngd2wbg.fsf@ashishki-desk.ger.corp.intel.com>

On Thu, Jun 18, 2015 at 12:35:31PM +0300, Alexander Shishkin wrote:
> Ingo Molnar <mingo@kernel.org> writes:
> 
> > * Alexander Shishkin <alexander.shishkin@linux.intel.com> wrote:
> >
> >> +void rb_free_aux(struct ring_buffer *rb)
> >> +{
> >> +	/*
> >> +	 * hold rb::refcount to make sure rb doesn't disappear
> >> +	 * before aux pages are freed
> >> +	 */
> >> +	if (WARN_ON_ONCE(!atomic_inc_not_zero(&rb->refcount)))
> >> +		return;
> >> +
> >> +	if (atomic_dec_and_test(&rb->aux_refcount))
> >> +		call_rcu(&rb->rcu_head, rb_free_rcu);
> >> +	else
> >> +		ring_buffer_put(rb);	/* matches the increment above */
> >
> > Is call_rcu() NMI-safe? I don't think so ...
> 
> Actually, we couldn't establish that without Paul's help.

I was fairly sure it was not, Paul did confirm last night it is indeed
not.