Prevent OOB reads for zero-length H264 payloads.

webrtc/modules/rtp_rtcp/source/rtp_format_h264_unittest.cc

 /*
  *  Copyright (c) 2014 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 519 matching lines @@
 
   payload = RtpDepacketizer::ParsedPayload();
   ASSERT_TRUE(depacketizer_->Parse(&payload, packet3, sizeof(packet3)));
   ExpectPacket(&payload, kExpected3, sizeof(kExpected3));
   EXPECT_EQ(kVideoFrameKey, payload.frame_type);
   EXPECT_EQ(kRtpVideoH264, payload.type.Video.codec);
   EXPECT_FALSE(payload.type.Video.isFirstPacket);
   EXPECT_EQ(kH264FuA, payload.type.Video.codecHeader.H264.packetization_type);
   EXPECT_EQ(kIdr, payload.type.Video.codecHeader.H264.nalu_type);
 }
+
+TEST_F(RtpDepacketizerH264Test, TestEmptyPayload) {

[stefan-webrtc, 2015/06/30 10:20:03]

Seems like we may want to have a few generic tests for the depacketizer and
share some test code, wdyt?

[pbos-webrtc, 2015/06/30 11:22:05]

No, since they are different formats the only thing they would share as input is
zero-length, and that seems like a bunch of boilerplate. :)

+  // Using a wild pointer to crash on accesses from inside the depacketizer.
+  uint8_t* garbage_ptr = reinterpret_cast<uint8_t*>(0x4711);
+  RtpDepacketizer::ParsedPayload payload;
+  EXPECT_FALSE(depacketizer_->Parse(&payload, garbage_ptr, 0));
+}
+
 }  // namespace webrtc