Add fuzzer test for various media parsers

media/filters/vp8_parser_fuzzertest.cc

+// Copyright 2016 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include <stddef.h>
+#include <stdint.h>
+
+#include "base/numerics/safe_conversions.h"
+#include "media/filters/ivf_parser.h"
+#include "media/filters/vp8_parser.h"
+
+// Entry point for LibFuzzer.
+extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) {
+  const uint8_t* ivf_payload = nullptr;
+  media::IvfParser ivf_parser;
+  media::IvfFileHeader ivf_file_header;
+  media::IvfFrameHeader ivf_frame_header;
+
+  if (!ivf_parser.Initialize(data, size, &ivf_file_header))
+    return 0;
+  if (ivf_file_header.fourcc != 0x30385056u)  // VP80

[DaleCurtis, 2016/03/05 01:30:06]

Does this trigger a dcheck otherwise? Seems you could combine the vp8/vp9 fuzzer
tests by conditionally going down one path or the other. Dunno if fuzzer team
would prefer them separate.

[jrummell, 2016/03/07 20:07:28]

Since we're testing with random files, I removed this check so that it still
attempts to parse it.

Doesn't matter to me if we combine the vp8 and vp9 tests. Just figured this way
the tests line up with the class they are testing. I would put both vp8 and vp9
files in the seed directories.

+    return 0;
+
+  // Parse until the end of stream/unsupported stream/error in stream is found.
+  while (ivf_parser.ParseNextFrame(&ivf_frame_header, &ivf_payload)) {
+    media::Vp8Parser vp8_parser;
+    media::Vp8FrameHeader vp8_frame_header;
+    vp8_parser.ParseFrame(ivf_payload, ivf_frame_header.frame_size,
+                          &vp8_frame_header);
+  }
+
+  return 0;
+}