Fix host VP9 experiment for WebRTC connections

This is a simple change to allow VP9 to work reasonably well without
crashing, when connecting via the website with VP9 experiment enabled.

Future work will optimize VP9 encoding, and remove large-picture
detection in the frame scheduler (as it's not needed for VP9).

Change-Id: Id80297dd9a0abf420ccf0e22251fca23200c32e3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1489711Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
Commit-Queue: Lambros Lambrou <lambroslambrou@chromium.org>
Cr-Commit-Position: refs/heads/master@{#637936}

remoting/protocol/webrtc_dummy_video_encoder.cc

@@ -181,6 +181,12 @@ webrtc::EncodedImageCallback::Result WebrtcDummyVideoEncoder::SendEncodedFrame(
     vp9_info->num_spatial_layers = 1;
     vp9_info->gof_idx = webrtc::kNoGofIdx;
     vp9_info->temporal_idx = webrtc::kNoTemporalIdx;
+    vp9_info->flexible_mode = false;
+    vp9_info->temporal_up_switch = true;
+    vp9_info->inter_layer_predicted = false;
+    vp9_info->first_frame_in_picture = true;
+    vp9_info->end_of_picture = true;
+    vp9_info->spatial_layer_resolution_present = false;
   } else if (frame.codec == webrtc::kVideoCodecH264) {
 #if defined(USE_H264_ENCODER)
     webrtc::CodecSpecificInfoH264* h264_info =