Add queued_time_ms trace for events in message loop

base/message_loop/message_loop.cc

Index: base/message_loop/message_loop.cc
diff --git a/base/message_loop/message_loop.cc b/base/message_loop/message_loop.cc
index 9e33a426eb9f1241be625c25aee1a1585898c6fc..8e68a4b99d39a057315fb1d4117753b44c55fe29 100644
--- a/base/message_loop/message_loop.cc
+++ b/base/message_loop/message_loop.cc
@@ -459,8 +459,10 @@ void MessageLoop::RunTask(const PendingTask& pending_task) {
   TRACE_EVENT_FLOW_END0("task", "MessageLoop::PostTask",
       TRACE_ID_MANGLE(GetTaskTraceID(pending_task, this)));
   TRACE_EVENT2("task", "MessageLoop::RunTask",
-               "src_file", pending_task.posted_from.file_name(),
-               "src_func", pending_task.posted_from.function_name());
+               "src", pending_task.posted_from.ToTraceFormat(),
+               "queued_time_ms",
+               pending_task.DurationSinceExpectedRunTime().InMilliseconds());

[jar (doing other things), 2013/07/08 17:08:33]

Possibly not a big deal...  I'm always wary of the cost of conversion to
milliseconds.  It uses 64 bit arithmetic (on Windows) to do a divide-by-1000,
which commonly requires calling a library call (sadly), and hence is also pretty
expensive.

[Xianzhu, 2013/07/08 18:51:39]

Changed DurationSinceExpectedRunTime() to return tracked_objects::Duration.

+
   DCHECK(nestable_tasks_allowed_);
   // Execute the task and assume the worst: It is probably not reentrant.
   nestable_tasks_allowed_ = false;