Pass decoded picture size from VDA to client

content/common/gpu/media/v4l2_video_decode_accelerator.cc

Index: content/common/gpu/media/v4l2_video_decode_accelerator.cc
diff --git a/content/common/gpu/media/v4l2_video_decode_accelerator.cc b/content/common/gpu/media/v4l2_video_decode_accelerator.cc
index 1b47516c0e7f03828859425c3435d0cd429e07db..7f372ece564f67fb756106c40d3b47bb36c8993b 100644
--- a/content/common/gpu/media/v4l2_video_decode_accelerator.cc
+++ b/content/common/gpu/media/v4l2_video_decode_accelerator.cc
@@ -1092,7 +1092,9 @@ void V4L2VideoDecodeAccelerator::Dequeue() {
       DVLOG(3) << "Dequeue(): returning input_id=" << dqbuf.timestamp.tv_sec
                << " as picture_id=" << output_record.picture_id;
       const media::Picture& picture =
-          media::Picture(output_record.picture_id, dqbuf.timestamp.tv_sec);
+          media::Picture(output_record.picture_id,
+                         dqbuf.timestamp.tv_sec,
+                         frame_buffer_size_);
       pending_picture_ready_.push(
           PictureRecord(output_record.cleared, picture));
       SendPictureReady();