Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Joel Fernandes <joel@joelfernandes.org>
To:     linux-kernel@vger.kernel.org
Cc:     kernel-team@android.com,
        "Joel Fernandes (Google)" <joel@joelfernandes.org>,
        Boqun Feng <boqun.feng@gmail.com>,
        Byungchul Park <byungchul.park@lge.com>,
        Erick Reyes <erickreyes@google.com>,
        Ingo Molnar <mingo@redhat.com>,
        Julia Cartwright <julia@ni.com>,
        Masami Hiramatsu <mhiramat@kernel.org>,
        Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        Namhyung Kim <namhyung@kernel.org>,
        Paul McKenney <paulmck@linux.vnet.ibm.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Steven Rostedt <rostedt@goodmis.org>,
        Thomas Glexiner <tglx@linutronix.de>,
        Todd Kjos <tkjos@google.com>,
        Tom Zanussi <tom.zanussi@linux.intel.com>,
        Will Deacon <will.deacon@arm.com>
Subject: [PATCH v10 2/3] tracepoint: Make rcuidle tracepoint callers use SRCU
Date:   Fri, 13 Jul 2018 14:55:46 -0700
Message-Id: <20180713215547.255620-3-joel@joelfernandes.org>
In-Reply-To: <20180713215547.255620-1-joel@joelfernandes.org>
References: <20180713215547.255620-1-joel@joelfernandes.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

From: "Joel Fernandes (Google)" <joel@joelfernandes.org>

In recent tests with IRQ on/off tracepoints, a large performance
overhead ~10% is noticed when running hackbench. This is root caused to
calls to rcu_irq_enter_irqson and rcu_irq_exit_irqson from the
tracepoint code. Following a long discussion on the list [1] about this,
we concluded that srcu is a better alternative for use during rcu idle.
Although it does involve extra barriers, its lighter than the sched-rcu
version which has to do additional RCU calls to notify RCU idle about
entry into RCU sections.

In this patch, we change the underlying implementation of the
trace_*_rcuidle API to use SRCU. This has shown to improve performance
alot for the high frequency irq enable/disable tracepoints.

Test: Tested idle and preempt/irq tracepoints.

Here are some performance numbers:

With a run of the following 30 times on a single core x86 Qemu instance
with 1GB memory:
hackbench -g 4 -f 2 -l 3000

Completion times in seconds. CONFIG_PROVE_LOCKING=y.

No patches (without this series)
Mean: 3.048
Median: 3.025
Std Dev: 0.064

With Lockdep using irq tracepoints with RCU implementation:
Mean: 3.451   (-11.66 %)
Median: 3.447 (-12.22%)
Std Dev: 0.049

With Lockdep using irq tracepoints with SRCU implementation (this series):
Mean: 3.020   (I would consider the improvement against the "without
	       this series" case as just noise).
Median: 3.013
Std Dev: 0.033

[1] https://patchwork.kernel.org/patch/10344297/

Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cleaned-up-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
---
 include/linux/tracepoint.h | 45 +++++++++++++++++++++++++++++++-------
 kernel/tracepoint.c        | 16 +++++++++++++-
 2 files changed, 52 insertions(+), 9 deletions(-)

diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h
index 19a690b559ca..97e1d365a817 100644
--- a/include/linux/tracepoint.h
+++ b/include/linux/tracepoint.h
@@ -15,6 +15,7 @@
  */
 
 #include <linux/smp.h>
+#include <linux/srcu.h>
 #include <linux/errno.h>
 #include <linux/types.h>
 #include <linux/cpumask.h>
@@ -33,6 +34,8 @@ struct trace_eval_map {
 
 #define TRACEPOINT_DEFAULT_PRIO	10
 
+extern struct srcu_struct tracepoint_srcu;
+
 extern int
 tracepoint_probe_register(struct tracepoint *tp, void *probe, void *data);
 extern int
@@ -75,10 +78,16 @@ int unregister_tracepoint_module_notifier(struct notifier_block *nb)
  * probe unregistration and the end of module exit to make sure there is no
  * caller executing a probe when it is freed.
  */
+#ifdef CONFIG_TRACEPOINTS
 static inline void tracepoint_synchronize_unregister(void)
 {
+	synchronize_srcu(&tracepoint_srcu);
 	synchronize_sched();
 }
+#else
+static inline void tracepoint_synchronize_unregister(void)
+{ }
+#endif
 
 #ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS
 extern int syscall_regfunc(void);
@@ -129,18 +138,34 @@ extern void syscall_unregfunc(void);
  * as "(void *, void)". The DECLARE_TRACE_NOARGS() will pass in just
  * "void *data", where as the DECLARE_TRACE() will pass in "void *data, proto".
  */
-#define __DO_TRACE(tp, proto, args, cond, rcucheck)			\
+#define __DO_TRACE(tp, proto, args, cond, rcuidle)			\
 	do {								\
 		struct tracepoint_func *it_func_ptr;			\
 		void *it_func;						\
 		void *__data;						\
+		int __maybe_unused idx = 0;				\
 									\
 		if (!(cond))						\
 			return;						\
-		if (rcucheck)						\
-			rcu_irq_enter_irqson();				\
-		rcu_read_lock_sched_notrace();				\
-		it_func_ptr = rcu_dereference_sched((tp)->funcs);	\
+									\
+		/* srcu can't be used from NMI */			\
+		if (rcuidle && in_nmi())				\
+			WARN_ON_ONCE(1);				\
+									\
+		/* keep srcu and sched-rcu usage consistent */		\
+		preempt_disable_notrace();				\
+									\
+		/*							\
+		 * For rcuidle callers, use srcu since sched-rcu	\
+		 * doesn't work from the idle path.			\
+		 */							\
+		if (rcuidle)						\
+			idx = srcu_read_lock_notrace(&tracepoint_srcu);	\
+		else							\
+			rcu_read_lock_sched_notrace();			\
+									\
+		it_func_ptr = rcu_dereference_raw((tp)->funcs);		\
+									\
 		if (it_func_ptr) {					\
 			do {						\
 				it_func = (it_func_ptr)->func;		\
@@ -148,9 +173,13 @@ extern void syscall_unregfunc(void);
 				((void(*)(proto))(it_func))(args);	\
 			} while ((++it_func_ptr)->func);		\
 		}							\
-		rcu_read_unlock_sched_notrace();			\
-		if (rcucheck)						\
-			rcu_irq_exit_irqson();				\
+									\
+		if (rcuidle)						\
+			srcu_read_unlock_notrace(&tracepoint_srcu, idx);\
+		else							\
+			rcu_read_unlock_sched_notrace();		\
+									\
+		preempt_enable_notrace();				\
 	} while (0)
 
 #ifndef MODULE
diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c
index 6dc6356c3327..955148d91b74 100644
--- a/kernel/tracepoint.c
+++ b/kernel/tracepoint.c
@@ -31,6 +31,9 @@
 extern struct tracepoint * const __start___tracepoints_ptrs[];
 extern struct tracepoint * const __stop___tracepoints_ptrs[];
 
+DEFINE_SRCU(tracepoint_srcu);
+EXPORT_SYMBOL_GPL(tracepoint_srcu);
+
 /* Set to 1 to enable tracepoint debug output */
 static const int tracepoint_debug;
 
@@ -67,16 +70,27 @@ static inline void *allocate_probes(int count)
 	return p == NULL ? NULL : p->probes;
 }
 
-static void rcu_free_old_probes(struct rcu_head *head)
+static void srcu_free_old_probes(struct rcu_head *head)
 {
 	kfree(container_of(head, struct tp_probes, rcu));
 }
 
+static void rcu_free_old_probes(struct rcu_head *head)
+{
+	call_srcu(&tracepoint_srcu, head, srcu_free_old_probes);
+}
+
 static inline void release_probes(struct tracepoint_func *old)
 {
 	if (old) {
 		struct tp_probes *tp_probes = container_of(old,
 			struct tp_probes, probes[0]);
+		/*
+		 * Tracepoint probes are protected by both sched RCU and SRCU,
+		 * by calling the SRCU callback in the sched RCU callback we
+		 * cover both cases. So let us chain the SRCU and sched RCU
+		 * callbacks to wait for both grace periods.
+		 */
 		call_rcu_sched(&tp_probes->rcu, rcu_free_old_probes);
 	}
 }
-- 
2.18.0.203.gfac676dfb9-goog