[PATCH 4.14 40/74] tracing: Fix lock inversion in trace_event_enable_tgid_record()

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

From: Prateek Sood <prsood@codeaurora.org>

commit 3a53acf1d9bea11b57c1f6205e3fe73f9d8a3688 upstream.

       Task T2                             Task T3
trace_options_core_write()            subsystem_open()

 mutex_lock(trace_types_lock)           mutex_lock(event_mutex)

 set_tracer_flag()

   trace_event_enable_tgid_record()       mutex_lock(trace_types_lock)

    mutex_lock(event_mutex)

This gives a circular dependency deadlock between trace_types_lock and
event_mutex. To fix this invert the usage of trace_types_lock and
event_mutex in trace_options_core_write(). This keeps the sequence of
lock usage consistent.

Link: http://lkml.kernel.org/r/0101016eef175e38-8ca71caf-a4eb-480d-a1e6-6f0bbc015495-000000@us-west-2.amazonses.com

Cc: stable@vger.kernel.org
Fixes: d914ba37d7145 ("tracing: Add support for recording tgid of tasks")
Signed-off-by: Prateek Sood <prsood@codeaurora.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 kernel/trace/trace.c        |    8 ++++++++
 kernel/trace/trace_events.c |    8 ++++----
 2 files changed, 12 insertions(+), 4 deletions(-)

--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -4368,6 +4368,10 @@ int trace_keep_overwrite(struct tracer *
 
 int set_tracer_flag(struct trace_array *tr, unsigned int mask, int enabled)
 {
+	if ((mask == TRACE_ITER_RECORD_TGID) ||
+	    (mask == TRACE_ITER_RECORD_CMD))
+		lockdep_assert_held(&event_mutex);
+
 	/* do nothing if flag is already set */
 	if (!!(tr->trace_flags & mask) == !!enabled)
 		return 0;
@@ -4433,6 +4437,7 @@ static int trace_set_options(struct trac
 		cmp += 2;
 	}
 
+	mutex_lock(&event_mutex);
 	mutex_lock(&trace_types_lock);
 
 	for (i = 0; trace_options[i]; i++) {
@@ -4447,6 +4452,7 @@ static int trace_set_options(struct trac
 		ret = set_tracer_option(tr, cmp, neg);
 
 	mutex_unlock(&trace_types_lock);
+	mutex_unlock(&event_mutex);
 
 	/*
 	 * If the first trailing whitespace is replaced with '\0' by strstrip,
@@ -7373,9 +7379,11 @@ trace_options_core_write(struct file *fi
 	if (val != 0 && val != 1)
 		return -EINVAL;
 
+	mutex_lock(&event_mutex);
 	mutex_lock(&trace_types_lock);
 	ret = set_tracer_flag(tr, 1 << index, val);
 	mutex_unlock(&trace_types_lock);
+	mutex_unlock(&event_mutex);
 
 	if (ret < 0)
 		return ret;
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -326,7 +326,8 @@ void trace_event_enable_cmd_record(bool
 	struct trace_event_file *file;
 	struct trace_array *tr;
 
-	mutex_lock(&event_mutex);
+	lockdep_assert_held(&event_mutex);
+
 	do_for_each_event_file(tr, file) {
 
 		if (!(file->flags & EVENT_FILE_FL_ENABLED))
@@ -340,7 +341,6 @@ void trace_event_enable_cmd_record(bool
 			clear_bit(EVENT_FILE_FL_RECORDED_CMD_BIT, &file->flags);
 		}
 	} while_for_each_event_file();
-	mutex_unlock(&event_mutex);
 }
 
 void trace_event_enable_tgid_record(bool enable)
@@ -348,7 +348,8 @@ void trace_event_enable_tgid_record(bool
 	struct trace_event_file *file;
 	struct trace_array *tr;
 
-	mutex_lock(&event_mutex);
+	lockdep_assert_held(&event_mutex);
+
 	do_for_each_event_file(tr, file) {
 		if (!(file->flags & EVENT_FILE_FL_ENABLED))
 			continue;
@@ -362,7 +363,6 @@ void trace_event_enable_tgid_record(bool
 				  &file->flags);
 		}
 	} while_for_each_event_file();
-	mutex_unlock(&event_mutex);
 }
 
 static int __ftrace_event_enable_disable(struct trace_event_file *file,