diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
index 0be866c91f62..4a68273b29d5 100644
--- a/include/trace/events/sched.h
+++ b/include/trace/events/sched.h
@@ -226,7 +226,7 @@ DECLARE_EVENT_CLASS(sched_process_template,
 DEFINE_EVENT(sched_process_template, sched_process_free,
 	     TP_PROTO(struct task_struct *p),
 	     TP_ARGS(p));
-	     
+
 
 /*
  * Tracepoint for a task exiting:
@@ -380,6 +380,30 @@ DEFINE_EVENT(sched_stat_template, sched_stat_blocked,
 	     TP_PROTO(struct task_struct *tsk, u64 delay),
 	     TP_ARGS(tsk, delay));
 
+/*
+ * Tracepoint for recording the cause of uninterruptible sleep.
+ */
+TRACE_EVENT(sched_blocked_reason,
+
+	TP_PROTO(struct task_struct *tsk),
+
+	TP_ARGS(tsk),
+
+	TP_STRUCT__entry(
+		__field( pid_t,	pid	)
+		__field( void*, caller	)
+		__field( bool, io_wait	)
+	),
+
+	TP_fast_assign(
+		__entry->pid	= tsk->pid;
+		__entry->caller = (void*)get_wchan(tsk);
+		__entry->io_wait = tsk->in_iowait;
+	),
+
+	TP_printk("pid=%d iowait=%d caller=%pS", __entry->pid, __entry->io_wait, __entry->caller)
+);
+
 /*
  * Tracepoint for accounting runtime (time the task is executing
  * on a CPU).
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index b39fb596f6c1..cc6edbea9cf0 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -941,6 +941,7 @@ update_stats_enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se)
 			}
 
 			trace_sched_stat_blocked(tsk, delta);
+			trace_sched_blocked_reason(tsk);
 
 			/*
 			 * Blocking time is in units of nanosecs, so shift by