shaka-packager/packager/media/crypto/subsample_generator_unittes...

508 lines
19 KiB
C++

// Copyright 2018 Google LLC. All rights reserved.
//
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file or at
// https://developers.google.com/open-source/licenses/bsd
#include "packager/media/crypto/subsample_generator.h"
#include <gmock/gmock.h>
#include <gtest/gtest.h>
#include "packager/media/base/audio_stream_info.h"
#include "packager/media/base/video_stream_info.h"
#include "packager/media/codecs/av1_parser.h"
#include "packager/media/codecs/video_slice_header_parser.h"
#include "packager/media/codecs/vpx_parser.h"
#include "packager/status_test_util.h"
namespace shaka {
namespace media {
namespace {
using ::testing::_;
using ::testing::AtLeast;
using ::testing::DoAll;
using ::testing::ElementsAre;
using ::testing::ElementsAreArray;
using ::testing::Return;
using ::testing::SetArgPointee;
using ::testing::Test;
using ::testing::Values;
using ::testing::WithParamInterface;
const bool kVP9SubsampleEncryption = true;
const uint8_t kH264CodecConfig[] = {
// clang-format off
// Header
0x01, 0x64, 0x00, 0x1e, 0xff,
// SPS count (ignore top three bits)
0xe1,
// SPS
0x00, 0x19, // Size
0x67, 0x64, 0x00, 0x1e, 0xac, 0xd9, 0x40, 0xa0, 0x2f, 0xf9, 0x70, 0x11,
0x00, 0x00, 0x03, 0x03, 0xe9, 0x00, 0x00, 0xea, 0x60, 0x0f, 0x16, 0x2d,
0x96,
// PPS count
0x01,
// PPS
0x00, 0x06, // Size
0x68, 0xeb, 0xe3, 0xcb, 0x22, 0xc0,
// clang-format on
};
const uint8_t kAV1CodecConfig[] = {0x00, 0x01, 0x02, 0x03};
const int kTrackId = 1;
const uint32_t kTimeScale = 1000;
const uint64_t kDuration = 10000;
const char kCodecString[] = "codec string";
const char kLanguage[] = "eng";
const bool kEncrypted = true;
VideoStreamInfo GetVideoStreamInfo(Codec codec) {
const uint16_t kWidth = 10u;
const uint16_t kHeight = 20u;
const uint32_t kPixelWidth = 2u;
const uint32_t kPixelHeight = 3u;
const int16_t kTrickPlayFactor = 0;
const uint8_t kNaluLengthSize = 1u;
const uint8_t* codec_config = nullptr;
size_t codec_config_size = 0;
switch (codec) {
case kCodecH264:
codec_config = kH264CodecConfig;
codec_config_size = sizeof(kH264CodecConfig);
break;
case kCodecAV1:
codec_config = kAV1CodecConfig;
codec_config_size = sizeof(kAV1CodecConfig);
break;
default:
// We do not care about the codec configs for other codecs in this file.
break;
}
return VideoStreamInfo(kTrackId, kTimeScale, kDuration, codec,
H26xStreamFormat::kUnSpecified, kCodecString,
codec_config, codec_config_size, kWidth, kHeight,
kPixelWidth, kPixelHeight, kTrickPlayFactor,
kNaluLengthSize, kLanguage, !kEncrypted);
}
AudioStreamInfo GetAudioStreamInfo(Codec codec) {
const uint8_t kSampleBits = 1;
const uint8_t kNumChannels = 2;
const uint32_t kSamplingFrequency = 48000;
const uint64_t kSeekPrerollNs = 12345;
const uint64_t kCodecDelayNs = 56789;
const uint32_t kMaxBitrate = 13579;
const uint32_t kAvgBitrate = 13000;
const uint8_t kCodecConfig[] = {0x00};
return AudioStreamInfo(kTrackId, kTimeScale, kDuration, codec, kCodecString,
kCodecConfig, sizeof(kCodecConfig), kSampleBits,
kNumChannels, kSamplingFrequency, kSeekPrerollNs,
kCodecDelayNs, kMaxBitrate, kAvgBitrate, kLanguage,
!kEncrypted);
}
} // namespace
inline bool operator==(const SubsampleEntry& lhs, const SubsampleEntry& rhs) {
return lhs.clear_bytes == rhs.clear_bytes &&
lhs.cipher_bytes == rhs.cipher_bytes;
}
class MockVPxParser : public VPxParser {
public:
MOCK_METHOD3(Parse,
bool(const uint8_t* data,
size_t data_size,
std::vector<VPxFrameInfo>* vpx_frames));
};
class MockVideoSliceHeaderParser : public VideoSliceHeaderParser {
public:
MOCK_METHOD1(Initialize,
bool(const std::vector<uint8_t>& decoder_configuration));
MOCK_METHOD1(ProcessNalu, bool(const Nalu& nalu));
MOCK_METHOD1(GetHeaderSize, int64_t(const Nalu& nalu));
};
class MockAV1Parser : public AV1Parser {
public:
MOCK_METHOD3(Parse,
bool(const uint8_t* data,
size_t data_size,
std::vector<Tile>* tiles));
};
class SubsampleGeneratorTest : public Test, public WithParamInterface<FourCC> {
public:
SubsampleGeneratorTest() : protection_scheme_(GetParam()) {}
protected:
FourCC protection_scheme_;
};
TEST_P(SubsampleGeneratorTest, VP9FullSampleEncryption) {
SubsampleGenerator generator(!kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecVP9)));
constexpr size_t kFrameSize = 50;
constexpr uint8_t kFrame[kFrameSize] = {};
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
EXPECT_THAT(subsamples, ElementsAre());
}
TEST_P(SubsampleGeneratorTest, VP9ParseFailed) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecVP9)));
constexpr size_t kFrameSize = 50;
constexpr uint8_t kFrame[kFrameSize] = {};
std::unique_ptr<MockVPxParser> mock_vpx_parser(new MockVPxParser);
EXPECT_CALL(*mock_vpx_parser, Parse(kFrame, kFrameSize, _))
.WillOnce(Return(false));
generator.InjectVpxParserForTesting(std::move(mock_vpx_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_NOT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
}
TEST_P(SubsampleGeneratorTest, VP9SubsampleEncryption) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecVP9)));
constexpr size_t kFrameSize = 50;
constexpr uint8_t kFrame[kFrameSize] = {};
constexpr size_t kUncompressedHeaderSize = 20;
// VP9 block align protected data for all protection schemes.
const SubsampleEntry kExpectedSubsamples[] = {
// {20,30} block aligned.
{34, 16},
};
std::vector<VPxFrameInfo> vpx_frame_info(1);
vpx_frame_info[0].frame_size = kFrameSize;
vpx_frame_info[0].uncompressed_header_size = kUncompressedHeaderSize;
std::unique_ptr<MockVPxParser> mock_vpx_parser(new MockVPxParser);
EXPECT_CALL(*mock_vpx_parser, Parse(kFrame, kFrameSize, _))
.WillOnce(DoAll(SetArgPointee<2>(vpx_frame_info), Return(true)));
generator.InjectVpxParserForTesting(std::move(mock_vpx_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
EXPECT_THAT(subsamples, ElementsAreArray(kExpectedSubsamples));
}
TEST_P(SubsampleGeneratorTest, VP9SubsampleEncryptionWithSuperFrame) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecVP9)));
constexpr size_t kFrameSize = 50;
constexpr uint8_t kFrame[kFrameSize] = {};
// Super frame with two subframes.
constexpr size_t kSubFrameSizes[] = {10, 34};
constexpr size_t kUncompressedHeaderSizes[] = {4, 1};
// VP9 block align protected data for all protection schemes.
const SubsampleEntry kExpectedSubsamples[] = {
// {4,6},{1,33} block aligned => {10,0},{2,32}
// Then merge consecutive clear-only subsamples.
{12, 32},
// Superframe index (50 - 10 - 34).
{6, 0},
};
std::vector<VPxFrameInfo> vpx_frame_info(2);
for (int i = 0; i < 2; i++) {
vpx_frame_info[i].frame_size = kSubFrameSizes[i];
vpx_frame_info[i].uncompressed_header_size = kUncompressedHeaderSizes[i];
}
std::unique_ptr<MockVPxParser> mock_vpx_parser(new MockVPxParser);
EXPECT_CALL(*mock_vpx_parser, Parse(kFrame, kFrameSize, _))
.WillOnce(DoAll(SetArgPointee<2>(vpx_frame_info), Return(true)));
generator.InjectVpxParserForTesting(std::move(mock_vpx_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
EXPECT_THAT(subsamples, ElementsAreArray(kExpectedSubsamples));
}
TEST_P(SubsampleGeneratorTest, VP9SubsampleEncryptionWithLargeSuperFrame) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecVP9)));
constexpr size_t kFrameSize = 0x23456;
constexpr uint8_t kFrame[kFrameSize] = {};
// Super frame with two subframes.
constexpr size_t kSubFrameSizes[] = {0x10, 0x23000, 0x440};
constexpr size_t kUncompressedHeaderSizes[] = {4, 0x21000, 2};
// VP9 block align protected data for all protection schemes.
const SubsampleEntry kExpectedSubsamples[] = {
// {4,12},{1,0x23000-1} block aligned => {16,0},{0x21000,0x2000}
// Then split big clear_bytes, merge consecutive clear-only subsamples.
{0xffff, 0},
{0xffff, 0},
{0x1012, 0x2000},
// {2,0x440-2} block aligned.
{0x10, 0x430},
// Superframe index.
{6, 0},
};
std::vector<VPxFrameInfo> vpx_frame_info(3);
for (int i = 0; i < 3; i++) {
vpx_frame_info[i].frame_size = kSubFrameSizes[i];
vpx_frame_info[i].uncompressed_header_size = kUncompressedHeaderSizes[i];
}
std::unique_ptr<MockVPxParser> mock_vpx_parser(new MockVPxParser);
EXPECT_CALL(*mock_vpx_parser, Parse(kFrame, kFrameSize, _))
.WillOnce(DoAll(SetArgPointee<2>(vpx_frame_info), Return(true)));
generator.InjectVpxParserForTesting(std::move(mock_vpx_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
EXPECT_THAT(subsamples, ElementsAreArray(kExpectedSubsamples));
}
TEST_P(SubsampleGeneratorTest, H264ParseFailed) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecH264)));
constexpr uint8_t kFrame[] = {
// First NALU (nalu_size = 9).
0x09, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09};
constexpr size_t kFrameSize = sizeof(kFrame);
std::unique_ptr<MockVideoSliceHeaderParser> mock_video_slice_header_parser(
new MockVideoSliceHeaderParser);
EXPECT_CALL(*mock_video_slice_header_parser, ProcessNalu(_))
.WillOnce(Return(true));
EXPECT_CALL(*mock_video_slice_header_parser, GetHeaderSize(_))
.WillOnce(Return(-1));
generator.InjectVideoSliceHeaderParserForTesting(
std::move(mock_video_slice_header_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_NOT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
}
TEST_P(SubsampleGeneratorTest, H264SubsampleEncryption) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecH264)));
constexpr uint8_t kFrame[] = {
// First NALU (nalu_size = 9).
0x09, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09,
// Second NALU (nalu_size = 0x25).
0x27, 0x25, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 0x21, 0x22, 0x23,
0x24, 0x25, 0x26, 0x27,
// Third non-video-slice NALU (nalu_size = 0x32).
0x32, 0x67, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 0x21, 0x22, 0x23,
0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f,
0x30, 0x31, 0x32};
constexpr size_t kFrameSize = sizeof(kFrame);
// There are two video slices.
const size_t kSliceHeaderSize[] = {4, 5};
const SubsampleEntry kExpectedUnalignedSubsamples[] = {
// clear_bytes = nalu_length_size (1) + type_size (1) + header_size (4).
// encrypted_bytes = nalu_size (9) - type_size (1) - header_size (4).
{6, 4},
// clear_bytes = nalu_length_size (1) + type_size (1) + header_size (5).
// encrypted_bytes = nalu_size (0x27) - type_size (1) - header_size (5).
{7, 0x21},
// Non-video slice, clear_bytes = nalu_length_size (1) + nalu_size (0x32).
// encrypted_bytes = 0.
{0x33, 0},
};
const SubsampleEntry kExpectedAlignedSubsamples[] = {
// {6,4},{7,0x21} block aligned => {10,0},{8,0x20}
// Then merge consecutive clear-only subsamples.
{18, 0x20},
{0x33, 0},
};
std::unique_ptr<MockVideoSliceHeaderParser> mock_video_slice_header_parser(
new MockVideoSliceHeaderParser);
EXPECT_CALL(*mock_video_slice_header_parser, ProcessNalu(_))
.Times(AtLeast(2))
.WillRepeatedly(Return(true));
EXPECT_CALL(*mock_video_slice_header_parser, GetHeaderSize(_))
.WillOnce(Return(kSliceHeaderSize[0]))
.WillOnce(Return(kSliceHeaderSize[1]));
generator.InjectVideoSliceHeaderParserForTesting(
std::move(mock_video_slice_header_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
// Align subsamples for all CENC protection schemes except for cbcs.
if (protection_scheme_ == FOURCC_cbcs)
EXPECT_THAT(subsamples, ElementsAreArray(kExpectedUnalignedSubsamples));
else
EXPECT_THAT(subsamples, ElementsAreArray(kExpectedAlignedSubsamples));
}
TEST_P(SubsampleGeneratorTest, AV1ParserFailed) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecAV1)));
constexpr size_t kFrameSize = 50;
constexpr uint8_t kFrame[kFrameSize] = {};
std::unique_ptr<MockAV1Parser> mock_av1_parser(new MockAV1Parser);
EXPECT_CALL(*mock_av1_parser, Parse(kFrame, kFrameSize, _))
.WillOnce(Return(false));
generator.InjectAV1ParserForTesting(std::move(mock_av1_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_NOT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
}
TEST_P(SubsampleGeneratorTest, AV1SubsampleEncryption) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetVideoStreamInfo(kCodecAV1)));
constexpr size_t kFrameSize = 50;
constexpr uint8_t kFrame[kFrameSize] = {};
constexpr size_t kTileOffsets[] = {4, 11};
constexpr size_t kTileSizes[] = {6, 33};
// AV1 block align protected data for all protection schemes.
const SubsampleEntry kExpectedSubsamples[] = {
// {4,6},{11-4-6,33},{50-11-33,0} block aligned => {10,0},{2,32},{6,0}.
// Then merge consecutive clear-only subsamples.
{12, 32},
{6, 0},
};
std::vector<AV1Parser::Tile> tiles(2);
for (int i = 0; i < 2; i++) {
tiles[i].start_offset_in_bytes = kTileOffsets[i];
tiles[i].size_in_bytes = kTileSizes[i];
}
std::unique_ptr<MockAV1Parser> mock_av1_parser(new MockAV1Parser);
EXPECT_CALL(*mock_av1_parser, Parse(kFrame, kFrameSize, _))
.WillOnce(DoAll(SetArgPointee<2>(tiles), Return(true)));
generator.InjectAV1ParserForTesting(std::move(mock_av1_parser));
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
EXPECT_THAT(subsamples, ElementsAreArray(kExpectedSubsamples));
}
TEST_P(SubsampleGeneratorTest, AACIsFullSampleEncrypted) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(
generator.Initialize(protection_scheme_, GetAudioStreamInfo(kCodecAAC)));
constexpr size_t kFrameSize = 50;
constexpr uint8_t kFrame[kFrameSize] = {};
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
EXPECT_THAT(subsamples, ElementsAre());
}
INSTANTIATE_TEST_CASE_P(
CencProtectionSchemes,
SubsampleGeneratorTest,
Values(FOURCC_cenc, FOURCC_cens, FOURCC_cbc1, FOURCC_cbcs));
TEST(SampleAesSubsampleGeneratorTest, AAC) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(generator.Initialize(kAppleSampleAesProtectionScheme,
GetAudioStreamInfo(kCodecAAC)));
constexpr size_t kNumFrames = 4;
constexpr size_t kMaxFrameSize = 100;
constexpr size_t kFrameSizes[] = {6, 16, 17, 50};
constexpr uint8_t kFrames[kNumFrames][kMaxFrameSize] = {};
// 16 bytes clear lead.
const SubsampleEntry kExpectedSubsamples[] = {
{6, 0},
{16, 0},
{16, 1},
{16, 34},
};
for (int i = 0; i < 4; i++) {
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(
generator.GenerateSubsamples(kFrames[i], kFrameSizes[i], &subsamples));
EXPECT_THAT(subsamples, ElementsAre(kExpectedSubsamples[i]));
}
}
TEST(SampleAesSubsampleGeneratorTest, H264) {
SubsampleGenerator generator(kVP9SubsampleEncryption);
ASSERT_OK(generator.Initialize(kAppleSampleAesProtectionScheme,
GetVideoStreamInfo(kCodecH264)));
constexpr uint8_t kFrame[] = {
// First NALU (nalu_size = 9).
0x09, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09,
// Second NALU (nalu_size = 0x30).
0x30, 0x25, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 0x21, 0x22, 0x23,
0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f,
0x30,
// Third NALU (nalu_size = 0x31).
0x31, 0x25, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 0x21, 0x22, 0x23,
0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f,
0x30, 0x31,
// Fourth non-video-slice NALU (nalu_size = 6).
0x32, 0x67, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b,
0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 0x21, 0x22, 0x23,
0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f,
0x30, 0x31, 0x32};
constexpr size_t kFrameSize = sizeof(kFrame);
const SubsampleEntry kExpectedSubsamples[] = {
// NAL units with nalu_size <= 32+16 is not encrypted, so
// the first two NALUs are left in clear {1+9,0},{1+48,0}.
// The third NALUs has a fixed 32 bytes clear lead, +1 byte NALU length
// size, so it is {1+32, 17}.
// Then merge consecutive clear-only subsamples.
{1 + 9 + 1 + 48 + 1 + 32, 17},
// Non video slice is not encrypted.
{0x33, 0},
};
std::vector<SubsampleEntry> subsamples;
ASSERT_OK(generator.GenerateSubsamples(kFrame, kFrameSize, &subsamples));
EXPECT_THAT(subsamples, ElementsAreArray(kExpectedSubsamples));
}
} // namespace media
} // namespace shaka