[tip:perfcounters/core] perf_counter: optimize perf_counter_task_tick()

[tip-bot for Peter Zijlstra <a.p.zijlstra@chello.nl>]

Commit-ID:  7fc23a5380797012e92a9633169440f2f4a21253
Gitweb:     http://git.kernel.org/tip/7fc23a5380797012e92a9633169440f2f4a21253
Author:     Peter Zijlstra <a.p.zijlstra@chello.nl>
AuthorDate: Fri, 8 May 2009 18:52:21 +0200
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Fri, 8 May 2009 20:36:57 +0200

perf_counter: optimize perf_counter_task_tick()

perf_counter_task_tick() does way too much work to find out
there's nothing to do. Provide an easy short-circuit for the
normal case where there are no counters on the system.

[ Impact: micro-optimization ]

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090508170028.750619201@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 kernel/perf_counter.c |   13 +++++++++++--
 1 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/kernel/perf_counter.c b/kernel/perf_counter.c
index 60e55f0..fdb0d24 100644
--- a/kernel/perf_counter.c
+++ b/kernel/perf_counter.c
@@ -39,6 +39,7 @@ int perf_max_counters __read_mostly = 1;
 static int perf_reserved_percpu __read_mostly;
 static int perf_overcommit __read_mostly = 1;
 
+static atomic_t nr_counters __read_mostly;
 static atomic_t nr_mmap_tracking __read_mostly;
 static atomic_t nr_munmap_tracking __read_mostly;
 static atomic_t nr_comm_tracking __read_mostly;
@@ -1076,8 +1077,14 @@ static void rotate_ctx(struct perf_counter_context *ctx)
 
 void perf_counter_task_tick(struct task_struct *curr, int cpu)
 {
-	struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu);
-	struct perf_counter_context *ctx = &curr->perf_counter_ctx;
+	struct perf_cpu_context *cpuctx;
+	struct perf_counter_context *ctx;
+
+	if (!atomic_read(&nr_counters))
+		return;
+
+	cpuctx = &per_cpu(perf_cpu_context, cpu);
+	ctx = &curr->perf_counter_ctx;
 
 	perf_counter_cpu_sched_out(cpuctx);
 	perf_counter_task_sched_out(curr, cpu);
@@ -1197,6 +1204,7 @@ static void free_counter(struct perf_counter *counter)
 {
 	perf_pending_sync(counter);
 
+	atomic_dec(&nr_counters);
 	if (counter->hw_event.mmap)
 		atomic_dec(&nr_mmap_tracking);
 	if (counter->hw_event.munmap)
@@ -2861,6 +2869,7 @@ done:
 
 	counter->pmu = pmu;
 
+	atomic_inc(&nr_counters);
 	if (counter->hw_event.mmap)
 		atomic_inc(&nr_mmap_tracking);
 	if (counter->hw_event.munmap)