Delete class SSRCDatabase, and its global ssrc registry.

webrtc/voice_engine/channel.cc

 /*
  *  Copyright (c) 2012 The WebRTC project authors. All Rights Reserved.
  *
  *  Use of this source code is governed by a BSD-style license
  *  that can be found in the LICENSE file in the root of the source
  *  tree. An additional intellectual property rights grant can be found
  *  in the file PATENTS.  All contributing project authors may
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
@@ skipping 557 matching lines @@
   header.payload_type_frequency =
       rtp_payload_registry_->GetPayloadTypeFrequency(header.payloadType);
   if (header.payload_type_frequency < 0)
     return false;
   return ReceivePacket(rtp_packet, rtp_packet_length, header, false);
 }
 
 MixerParticipant::AudioFrameInfo Channel::GetAudioFrameWithMuted(
     int32_t id,
     AudioFrame* audioFrame) {
-  unsigned int ssrc;
-  RTC_CHECK_EQ(GetLocalSSRC(ssrc), 0);
-  event_log_proxy_->LogAudioPlayout(ssrc);
+
+  event_log_proxy_->LogAudioPlayout(0);

[the sun, 2017/02/16 13:52:32]

Hang on - RtcEventLog will have the same SSRC for all AudioPlayout events.

Further, this doesn't look right - I believe it should be using the remote SSRC.
terelius@, do you remember if there was any particular reason grouping by local
SSRC for incoming streams was selected?

[terelius, 2017/02/16 13:58:05]

I've never used the SSRC for audio playouts. ivoc@, can you answer the question
above?

[nisse-webrtc, 2017/02/16 14:01:59]

Ooops. I misunderstood the GetLocalSSRC call completely.

Background for this change was that ssrcs wasn't set at this point. I could try
undoing this part of the change and maybe it will now work, thanks to the random
default ssrc set by the ChannelManager.

[terelius, 2017/02/16 14:04:42]

If the ssrc is no longer considered useful, please send me an email so I can
remove the field from any future versions of the proto.

[ivoc, 2017/02/16 14:27:06]

The SSRC was added in https://codereview.webrtc.org/1340283002/. If I recall
correctly we had the problem that playout events were being logged for multiple
channels, and it was impossible to tell them apart without a unique ID. We
decided to use the SSRC for this purpose, so I don't think it should be removed.

[nisse-webrtc, 2017/02/16 14:29:22]

But if it's a received stream, we should be using the "remote" ssrc, not the
"local" one? Since in this code, local really means sending, if I've understood
it correctly.

[ivoc, 2017/02/16 14:35:38]

We only used it to tell the different channels apart, so we don't care about the
actual value. Feel free to change it to something that makes more sense.

[nisse-webrtc, 2017/02/17 09:59:47]

Undid this change, and then changed GetLocalSsrc to GetRemoteSsrc.

   // Get 10ms raw PCM data from the ACM (mixer limits output frequency)
   bool muted;
   if (audio_coding_->PlayoutData10Ms(audioFrame->sample_rate_hz_, audioFrame,
                                      &muted) == -1) {
     WEBRTC_TRACE(kTraceError, kTraceVoice, VoEId(_instanceId, _channelId),
                  "Channel::GetAudioFrame() PlayoutData10Ms() failed!");
     // In all likelihood, the audio in this frame is garbage. We return an
     // error so that the audio mixer module doesn't add it to the mix. As
     // a result, it won't be played out and the actions skipped here are
     // irrelevant.
@@ skipping 2575 matching lines @@
   int64_t min_rtt = 0;
   if (_rtpRtcpModule->RTT(remoteSSRC, &rtt, &avg_rtt, &min_rtt, &max_rtt) !=
       0) {
     return 0;
   }
   return rtt;
 }
 
 }  // namespace voe
 }  // namespace webrtc