linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: linux-kernel@vger.kernel.org
Cc: Ingo Molnar <mingo@kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
	"Joel Fernandes (Google)" <joel@joelfernandes.org>
Subject: [for-next][PATCH 15/34] tracepoint: Make rcuidle tracepoint callers use SRCU
Date: Sat, 11 Aug 2018 09:49:43 -0400	[thread overview]
Message-ID: <20180811135013.409952552@goodmis.org> (raw)
In-Reply-To: 20180811134928.034373761@goodmis.org

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/

[remove rcu_read_lock_sched_notrace as its the equivalent of
preempt_disable_notrace and is unnecessary to call in tracepoint code]
Link: http://lkml.kernel.org/r/20180730222423.196630-3-joel@joelfernandes.org

Cleaned-up-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
[ Simplified WARN_ON_ONCE() ]
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
---
 include/linux/tracepoint.h | 40 ++++++++++++++++++++++++++++++--------
 kernel/tracepoint.c        | 16 ++++++++++++++-
 2 files changed, 47 insertions(+), 9 deletions(-)

diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h
index 19a690b559ca..d9a084c72541 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 +138,31 @@ 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 */			\
+		WARN_ON_ONCE(rcuidle && in_nmi());			\
+									\
+		/* 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);	\
+									\
+		it_func_ptr = rcu_dereference_raw((tp)->funcs);		\
+									\
 		if (it_func_ptr) {					\
 			do {						\
 				it_func = (it_func_ptr)->func;		\
@@ -148,9 +170,11 @@ 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);\
+									\
+		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



  parent reply	other threads:[~2018-08-11 13:51 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-08-11 13:49 [for-next][PATCH 00/34] tracing: My official queue finally passed testing and ready for linux-next Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 01/34] srcu: Add notrace variants of srcu_read_{lock,unlock} Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 02/34] srcu: Add notrace variant of srcu_dereference Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 03/34] tracing/irqsoff: Split reset into separate functions Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 04/34] lib: Add module for testing preemptoff/irqsoff latency tracers Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 05/34] kselftests: Add tests for the preemptoff and irqsoff tracers Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 06/34] tracing: Make unregister_trigger() static Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 07/34] tracing: Remove orphaned function using_ftrace_ops_list_func() Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 08/34] tracing: Remove orphaned function ftrace_nr_registered_ops() Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 09/34] tracing/kprobes: Simplify the logic of enable_trace_kprobe() Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 10/34] tracing: preemptirq_delay_run() can be static Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 11/34] tracing: kprobes: Prohibit probing on notrace function Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 12/34] selftest/ftrace: Move kprobe selftest function to separate compile unit Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 13/34] selftests/ftrace: Fix kprobe string testcase to not probe notrace function Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 14/34] lockdep: Use this_cpu_ptr instead of get_cpu_var stats Steven Rostedt
2018-08-11 13:49 ` Steven Rostedt [this message]
2018-08-11 13:49 ` [for-next][PATCH 16/34] tracing: Centralize preemptirq tracepoints and unify their usage Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 17/34] tracefs: Annotate tracefs_ops with __ro_after_init Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 18/34] tracing: Do not call start/stop() functions when tracing_on does not change Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 19/34] ftrace: Add missing check for existing hwlat thread Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 20/34] tracing: Do a WARN_ON() if start_thread() in hwlat is called when thread exists Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 21/34] tracing: Make tracer_tracing_is_on() return bool Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 22/34] ring-buffer: Make ring_buffer_record_is_on() " Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 23/34] ring-buffer: Make ring_buffer_record_is_set_on() " Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 24/34] ftrace: Use true and false for boolean values in ops_references_rec() Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 25/34] tracing/kprobes: Fix within_notrace_func() to check only notrace functions Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 26/34] trace: Use rcu_dereference_raw for hooks from trace-event subsystem Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 27/34] tracing: irqsoff: Account for additional preempt_disable Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 28/34] tracing: Partial revert of "tracing: Centralize preemptirq tracepoints and unify their usage" Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 29/34] tracing/irqsoff: Handle preempt_count for different configs Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 30/34] tracing: More reverting of "tracing: Centralize preemptirq tracepoints and unify their usage" Steven Rostedt
2018-08-11 13:49 ` [for-next][PATCH 31/34] ftrace: Remove unused pointer ftrace_swapper_pid Steven Rostedt
2018-08-11 13:50 ` [for-next][PATCH 32/34] tracing: Fix synchronizing to event changes with tracepoint_synchronize_unregister() Steven Rostedt
2018-08-11 13:50 ` [for-next][PATCH 33/34] uprobes: Use synchronize_rcu() not synchronize_sched() Steven Rostedt
2018-08-11 13:50 ` [for-next][PATCH 34/34] tracepoints: Free early tracepoints after RCU is initialized Steven Rostedt

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180811135013.409952552@goodmis.org \
    --to=rostedt@goodmis.org \
    --cc=akpm@linux-foundation.org \
    --cc=joel@joelfernandes.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).