tracing: Use free_percpu instead of kfree

In the event->profile_enable() failure path, we release the per cpu
buffers using kfree which is wrong because they are per cpu pointers.
Although free_percpu only wraps kfree for now, that may change in the
future so lets use the correct way.
Reported-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Li Zefan <lizf@cn.fujitsu.com>

kernel/trace/trace_event_profile.c

@@ -53,8 +53,8 @@ static int ftrace_profile_enable_event(struct ftrace_event_call *event)
 
 fail_buf_nmi:
 	if (!total_profile_count) {
-		kfree(trace_profile_buf_nmi);
-		kfree(trace_profile_buf);
+		free_percpu(trace_profile_buf_nmi);
+		free_percpu(trace_profile_buf);
 		trace_profile_buf_nmi = NULL;
 		trace_profile_buf = NULL;
 	}