timer stats: speedups

Make timer-stats have almost zero overhead when enabled in the config but
not used.  (this way distros can enable it more easily)

Also update the documentation about overhead of timer_stats - it was
written for the first version which had a global lock and a linear list
walk based lookup ;-)

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

3 files changed, 14 insertions, 5 deletions

diff --git a/Documentation/hrtimer/timer_stats.txt b/Documentation/hrtimer/timer_stats.txt
index 27f782e3593f..22b0814d0ad0 100644
--- a/Documentation/hrtimer/timer_stats.txt
+++ b/Documentation/hrtimer/timer_stats.txt
@@ -2,9 +2,10 @@ timer_stats - timer usage statistics
 ------------------------------------
 
 timer_stats is a debugging facility to make the timer (ab)usage in a Linux
-system visible to kernel and userspace developers. It is not intended for
-production usage as it adds significant overhead to the (hr)timer code and the
-(hr)timer data structures.
+system visible to kernel and userspace developers. If enabled in the config
+but not used it has almost zero runtime overhead, and a relatively small
+data structure overhead. Even if collection is enabled runtime all the
+locking is per-CPU and lookup is hashed.
 
 timer_stats should be used by kernel and userspace developers to verify that
 their code does not make unduly use of timers. This helps to avoid unnecessary
diff --git a/kernel/time/timer_stats.c b/kernel/time/timer_stats.c
index fa3d380ca8c0..321693724ad7 100644
--- a/kernel/time/timer_stats.c
+++ b/kernel/time/timer_stats.c
@@ -236,10 +236,15 @@ void timer_stats_update_stats(void *timer, pid_t pid, void *startf,
         /*
          * It doesnt matter which lock we take:
          */
-        spinlock_t *lock = &per_cpu(lookup_lock, raw_smp_processor_id());
+        spinlock_t *lock;
         struct entry *entry, input;
         unsigned long flags;
 
+        if (likely(!active))
+                return;
+
+        lock = &per_cpu(lookup_lock, raw_smp_processor_id());
+
         input.timer = timer;
         input.start_func = startf;
         input.expire_func = timerf;
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 1ba77ca7d165..da95e10cfd70 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -126,7 +126,10 @@ config TIMER_STATS
           reprogrammed. The statistics can be read from /proc/timer_stats.
           The statistics collection is started by writing 1 to /proc/timer_stats,
           writing 0 stops it. This feature is useful to collect information
-          about timer usage patterns in kernel and userspace.
+          about timer usage patterns in kernel and userspace. This feature
+          is lightweight if enabled in the kernel config but not activated
+          (it defaults to deactivated on bootup and will only be activated
+          if some application like powertop activates it explicitly).
 
 config DEBUG_SLAB
         bool "Debug slab memory allocations"