[PATCH] tracing/lockdep: report the time waited for a lock

[Frederic Weisbecker <fweisbec@gmail.com>]

While trying to optimize the new lock on reiserfs to replace
the bkl, I find the lock tracing very useful though it lacks
something important for performance (and latency) instrumentation:
the time a task waits for a lock.

That's what this patch implement:

            bash-4816  [000]   202.652815: lock_contended: lock_contended: &sb->s_type->i_mutex_key
            bash-4816  [000]   202.652819: lock_acquired: &rq->lock (0.000 us)
           <...>-4787  [000]   202.652825: lock_acquired: &rq->lock (0.000 us)
           <...>-4787  [000]   202.652829: lock_acquired: &rq->lock (0.000 us)
            bash-4816  [000]   202.652833: lock_acquired: &sb->s_type->i_mutex_key (16.005 us)

As shown above, the "lock acquired" field is followed by the time it has been waiting
for the lock. Usually, a lock contended entry is followed by a near lock_acquired entry
with a non-zero time waited.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 include/trace/lockdep_event_types.h |   23 ++++++++++++++++++-----
 kernel/lockdep.c                    |    8 ++++----
 2 files changed, 22 insertions(+), 9 deletions(-)

diff --git a/include/trace/lockdep_event_types.h b/include/trace/lockdep_event_types.h
index adccfcd..863f1e4 100644
--- a/include/trace/lockdep_event_types.h
+++ b/include/trace/lockdep_event_types.h
@@ -32,11 +32,24 @@ TRACE_FORMAT(lock_contended,
 	TP_FMT("%s", lock->name)
 	);
 
-TRACE_FORMAT(lock_acquired,
-	TP_PROTO(struct lockdep_map *lock, unsigned long ip),
-	TP_ARGS(lock, ip),
-	TP_FMT("%s", lock->name)
-	);
+TRACE_EVENT(lock_acquired,
+	TP_PROTO(struct lockdep_map *lock, unsigned long ip, s64 waittime),
+
+	TP_ARGS(lock, ip, waittime),
+
+	TP_STRUCT__entry(
+		__field(const char *, name)
+		__field(unsigned long, wait_usec)
+		__field(unsigned long, wait_nsec_rem)
+	),
+	TP_fast_assign(
+		__entry->name = lock->name;
+		__entry->wait_nsec_rem = do_div(waittime, NSEC_PER_USEC);
+		__entry->wait_usec = (unsigned long) waittime;
+	),
+	TP_printk("%s (%lu.%03lu us)", __entry->name, __entry->wait_usec,
+				       __entry->wait_nsec_rem)
+);
 
 #endif
 #endif
diff --git a/kernel/lockdep.c b/kernel/lockdep.c
index b0f0118..c4582a6 100644
--- a/kernel/lockdep.c
+++ b/kernel/lockdep.c
@@ -3061,6 +3061,8 @@ found_it:
 	put_lock_stats(stats);
 }
 
+DEFINE_TRACE(lock_acquired);
+
 static void
 __lock_acquired(struct lockdep_map *lock, unsigned long ip)
 {
@@ -3099,6 +3101,8 @@ found_it:
 		hlock->holdtime_stamp = now;
 	}
 
+	trace_lock_acquired(lock, ip, waittime);
+
 	stats = get_lock_stats(hlock_class(hlock));
 	if (waittime) {
 		if (hlock->read)
@@ -3137,14 +3141,10 @@ void lock_contended(struct lockdep_map *lock, unsigned long ip)
 }
 EXPORT_SYMBOL_GPL(lock_contended);
 
-DEFINE_TRACE(lock_acquired);
-
 void lock_acquired(struct lockdep_map *lock, unsigned long ip)
 {
 	unsigned long flags;
 
-	trace_lock_acquired(lock, ip);
-
 	if (unlikely(!lock_stat))
 		return;
 
-- 
1.6.1