Re: perf: fuzzer gets CPU stuck in perf_callchain()

From: Peter Zijlstra <peterz@infradead.org>
To: Vince Weaver <vincent.weaver@maine.edu>
Cc: linux-kernel@vger.kernel.org, Paul Mackerras <paulus@samba.org>,
	Ingo Molnar <mingo@redhat.com>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	Jiri Olsa <jolsa@redhat.com>,
	Steven Rostedt <rostedt@goodmis.org>
Subject: Re: perf: fuzzer gets CPU stuck in perf_callchain()
Date: Thu, 19 Feb 2015 18:03:11 +0100	[thread overview]
Message-ID: <20150219170311.GH21418@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <alpine.DEB.2.11.1502191149140.28751@vincent-weaver-1.umelst.maine.edu>

On Thu, Feb 19, 2015 at 11:54:40AM -0500, Vince Weaver wrote:

> [ 7938.802139]  [<ffffffff8115e114>] perf_tp_event+0xc4/0x210
> [ 7938.861174]  [<ffffffff810b4f8a>] perf_trace_lock+0x12a/0x160
> [ 7938.882197]  [<ffffffff810b7f10>] lock_release+0x130/0x260
> [ 7938.888754]  [<ffffffff816c7474>] _raw_spin_unlock_irqrestore+0x24/0x40
> [ 7938.896510]  [<ffffffff8107bb4d>] do_send_sig_info+0x5d/0x80
> [ 7938.903240]  [<ffffffff811f69df>] send_sigio_to_task+0x12f/0x1a0
> [ 7938.923894]  [<ffffffff811f71ce>] send_sigio+0xae/0x100
> [ 7938.930132]  [<ffffffff811f72b7>] kill_fasync+0x97/0xf0
> [ 7938.942707]  [<ffffffff8115d0b4>] perf_event_wakeup+0xd4/0xf0
> [ 7938.956367]  [<ffffffff8115d103>] perf_pending_event+0x33/0x60
> [ 7938.963199]  [<ffffffff8114e3fc>] irq_work_run_list+0x4c/0x80
> [ 7938.969932]  [<ffffffff8114e448>] irq_work_run+0x18/0x40
> [ 7938.976212]  [<ffffffff810196af>] smp_trace_irq_work_interrupt+0x3f/0xc0
> [ 7938.983957]  [<ffffffff816c99bd>] trace_irq_work_interrupt+0x6d/0x80

Ah, it looks like the never ending irq_work..

Could you try this?

---
 kernel/events/core.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index f04daabfd1cf..088e81ae43cb 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4574,6 +4574,13 @@ static void perf_pending_event(struct irq_work *entry)
 {
 	struct perf_event *event = container_of(entry,
 			struct perf_event, pending);
+	int rctx;
+
+	rctx = perf_swevent_get_recursion_context();
+	/*
+	 * If we 'fail' here, that's OK, it means recursion is already disabled
+	 * and we won't recurse 'further'.
+	 */
 
 	if (event->pending_disable) {
 		event->pending_disable = 0;
@@ -4584,6 +4591,9 @@ static void perf_pending_event(struct irq_work *entry)
 		event->pending_wakeup = 0;
 		perf_event_wakeup(event);
 	}
+
+	if (rctx >= 0)
+		perf_swevent_put_recursion_context(rctx);
 }
 
 /*
