Make OverheadObserver::OnOverheadChanged count RTP headers only

webrtc/modules/congestion_controller/transport_feedback_adapter.cc

Index: webrtc/modules/congestion_controller/transport_feedback_adapter.cc
diff --git a/webrtc/modules/congestion_controller/transport_feedback_adapter.cc b/webrtc/modules/congestion_controller/transport_feedback_adapter.cc
index 7756fee365da2dd872be36bf028e525480b5abe3..ba6e9ce9ca5b13f973496c5ba23e6ad971382712 100644
--- a/webrtc/modules/congestion_controller/transport_feedback_adapter.cc
+++ b/webrtc/modules/congestion_controller/transport_feedback_adapter.cc
@@ -19,6 +19,7 @@
 #include "webrtc/modules/congestion_controller/delay_based_bwe.h"
 #include "webrtc/modules/rtp_rtcp/source/rtcp_packet/transport_feedback.h"
 #include "webrtc/modules/utility/include/process_thread.h"
+#include "webrtc/system_wrappers/include/field_trial.h"
 
 namespace webrtc {
 
@@ -59,6 +60,10 @@ void TransportFeedbackAdapter::AddPacket(uint16_t sequence_number,
                                          size_t length,
                                          int probe_cluster_id) {
   rtc::CritScope cs(&lock_);
+  if (webrtc::field_trial::FindFullName("WebRTC-SendSideBwe-WithOverhead") ==
+      "Enabled") {
+    length += transport_overhead_bytes_per_packet_;
+  }
   send_time_history_.AddAndRemoveOld(sequence_number, length, probe_cluster_id);
 }
 
@@ -73,6 +78,13 @@ void TransportFeedbackAdapter::SetMinBitrate(int min_bitrate_bps) {
   delay_based_bwe_->SetMinBitrate(min_bitrate_bps);
 }
 
+void TransportFeedbackAdapter::SetTransportOverhead(
+    int transport_overhead_bytes_per_packet) {
+  // TODO(nisse): Which lock?

[the sun, 2016/12/21 10:33:02]

figure out before landing

[nisse-webrtc, 2017/01/09 16:02:24]

It's accessed with lock_ held, so that seems right.

Can anyone (Michael, Stefan?) say if this is safe? This lock mainly protects
send_time_history_, and is held while calling methods on that object. Is there
any danger that those code paths interact in some way with the code paths
calling SetTransportOverhead, with risk of dead lock?

[stefan-webrtc, 2017/01/10 08:42:58]

SendTimeHistory doesn't have any callbacks, so I don't think there's any risk.

+  rtc::CritScope cs(&lock_);
+  transport_overhead_bytes_per_packet_ = transport_overhead_bytes_per_packet;
+}
+
 int64_t TransportFeedbackAdapter::GetProbingIntervalMs() const {
   rtc::CritScope cs(&bwe_lock_);
   return delay_based_bwe_->GetProbingIntervalMs();