Re: [RFC][PATCH 2/3] perf: Add a bit of paranoia

From: Peter Zijlstra <peterz@infradead.org>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Vince Weaver <vincent.weaver@maine.edu>,
	mingo@kernel.org, linux-kernel@vger.kernel.org,
	eranian@gmail.com, mark.rutland@arm.com,
	torvalds@linux-foundation.org, tglx@linutronix.de
Subject: Re: [RFC][PATCH 2/3] perf: Add a bit of paranoia
Date: Wed, 4 Feb 2015 17:43:27 +0100	[thread overview]
Message-ID: <20150204164327.GL24151@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <20150204163336.GC5029@twins.programming.kicks-ass.net>

On Wed, Feb 04, 2015 at 05:33:36PM +0100, Peter Zijlstra wrote:
> On Wed, Feb 04, 2015 at 03:51:36PM +0100, Jiri Olsa wrote:
> > On Mon, Feb 02, 2015 at 06:32:32PM +0100, Peter Zijlstra wrote:
> > > > That looks like tail recursive fun! An irq work that raises and irq work
> > > > ad infinitum. Lemme see if I can squash that.. didn't we have something
> > > > like this before... /me goes look.
> > > 
> > > 
> > > Does this make it go away?
> > > 
> > > --- a/kernel/events/core.c
> > > +++ b/kernel/events/core.c
> > > @@ -4413,6 +4413,8 @@ static void perf_pending_event(struct ir
> > >  	struct perf_event *event = container_of(entry,
> > >  			struct perf_event, pending);
> > >  
> > > +	int rctx = perf_swevent_get_recursion_context();
> > > +
> > 
> > hum, you should check the rctx
> > 
> > 	if (rctx == -1)
> > 		return;
> 
> D'uh, yes.

Hmm, that's not actually so, we need to process the irq_work, but we
want no further nested swevents.

We cannot not do the irq_work; so I think what we want is a conditional
put, seeing how if we fail the get, the recursion counter is already
raised and nested events won't happen.

Now in practise its very unlikely to ever happen; you need nested IRQs
for this and most of those have been killed, some legacy drivers might
maybe still use them but I forgot.