| /* |
| * Copyright (C) 2009 The Android Open Source Project |
| * |
| * Licensed under the Apache License, Version 2.0 (the "License"); |
| * you may not use this file except in compliance with the License. |
| * You may obtain a copy of the License at |
| * |
| * http://www.apache.org/licenses/LICENSE-2.0 |
| * |
| * Unless required by applicable law or agreed to in writing, software |
| * distributed under the License is distributed on an "AS IS" BASIS, |
| * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| * See the License for the specific language governing permissions and |
| * limitations under the License. |
| */ |
| |
| //#define LOG_NDEBUG 0 |
| #define LOG_TAG "MPEG4Extractor" |
| |
| #include <ctype.h> |
| #include <inttypes.h> |
| #include <memory> |
| #include <stdint.h> |
| #include <stdlib.h> |
| #include <string.h> |
| |
| #include <log/log.h> |
| #include <utils/Log.h> |
| |
| #include "AC4Parser.h" |
| #include "MPEG4Extractor.h" |
| #include "SampleTable.h" |
| #include "ItemTable.h" |
| #include "include/ESDS.h" |
| |
| #include <media/DataSourceBase.h> |
| #include <media/ExtractorUtils.h> |
| #include <media/stagefright/foundation/ABitReader.h> |
| #include <media/stagefright/foundation/ABuffer.h> |
| #include <media/stagefright/foundation/ADebug.h> |
| #include <media/stagefright/foundation/AMessage.h> |
| #include <media/stagefright/foundation/AudioPresentationInfo.h> |
| #include <media/stagefright/foundation/AUtils.h> |
| #include <media/stagefright/foundation/ByteUtils.h> |
| #include <media/stagefright/foundation/ColorUtils.h> |
| #include <media/stagefright/foundation/avc_utils.h> |
| #include <media/stagefright/foundation/hexdump.h> |
| #include <media/stagefright/foundation/OpusHeader.h> |
| #include <media/stagefright/MediaBufferGroup.h> |
| #include <media/stagefright/MediaDefs.h> |
| #include <media/stagefright/MetaDataBase.h> |
| #include <utils/String8.h> |
| |
| #include <byteswap.h> |
| #include "include/ID3.h" |
| |
| #ifndef UINT32_MAX |
| #define UINT32_MAX (4294967295U) |
| #endif |
| |
| #define ALAC_SPECIFIC_INFO_SIZE (36) |
| |
| namespace android { |
| |
| enum { |
| // max track header chunk to return |
| kMaxTrackHeaderSize = 32, |
| |
| // maximum size of an atom. Some atoms can be bigger according to the spec, |
| // but we only allow up to this size. |
| kMaxAtomSize = 64 * 1024 * 1024, |
| }; |
| |
| class MPEG4Source : public MediaTrackHelper { |
| static const size_t kMaxPcmFrameSize = 8192; |
| public: |
| // Caller retains ownership of both "dataSource" and "sampleTable". |
| MPEG4Source(AMediaFormat *format, |
| DataSourceHelper *dataSource, |
| int32_t timeScale, |
| const sp<SampleTable> &sampleTable, |
| Vector<SidxEntry> &sidx, |
| const Trex *trex, |
| off64_t firstMoofOffset, |
| const sp<ItemTable> &itemTable, |
| uint64_t elstShiftStartTicks); |
| virtual status_t init(); |
| |
| virtual media_status_t start(); |
| virtual media_status_t stop(); |
| |
| virtual media_status_t getFormat(AMediaFormat *); |
| |
| virtual media_status_t read(MediaBufferHelper **buffer, const ReadOptions *options = NULL); |
| bool supportsNonBlockingRead() override { return true; } |
| virtual media_status_t fragmentedRead( |
| MediaBufferHelper **buffer, const ReadOptions *options = NULL); |
| |
| virtual ~MPEG4Source(); |
| |
| private: |
| Mutex mLock; |
| |
| AMediaFormat *mFormat; |
| DataSourceHelper *mDataSource; |
| int32_t mTimescale; |
| sp<SampleTable> mSampleTable; |
| uint32_t mCurrentSampleIndex; |
| uint32_t mCurrentFragmentIndex; |
| Vector<SidxEntry> &mSegments; |
| const Trex *mTrex; |
| off64_t mFirstMoofOffset; |
| off64_t mCurrentMoofOffset; |
| off64_t mNextMoofOffset; |
| uint32_t mCurrentTime; // in media timescale ticks |
| int32_t mLastParsedTrackId; |
| int32_t mTrackId; |
| |
| int32_t mCryptoMode; // passed in from extractor |
| int32_t mDefaultIVSize; // passed in from extractor |
| uint8_t mCryptoKey[16]; // passed in from extractor |
| int32_t mDefaultEncryptedByteBlock; |
| int32_t mDefaultSkipByteBlock; |
| uint32_t mCurrentAuxInfoType; |
| uint32_t mCurrentAuxInfoTypeParameter; |
| int32_t mCurrentDefaultSampleInfoSize; |
| uint32_t mCurrentSampleInfoCount; |
| uint32_t mCurrentSampleInfoAllocSize; |
| uint8_t* mCurrentSampleInfoSizes; |
| uint32_t mCurrentSampleInfoOffsetCount; |
| uint32_t mCurrentSampleInfoOffsetsAllocSize; |
| uint64_t* mCurrentSampleInfoOffsets; |
| |
| bool mIsAVC; |
| bool mIsHEVC; |
| bool mIsAC4; |
| bool mIsPcm; |
| size_t mNALLengthSize; |
| |
| bool mStarted; |
| |
| MediaBufferHelper *mBuffer; |
| |
| uint8_t *mSrcBuffer; |
| |
| bool mIsHeif; |
| bool mIsAudio; |
| sp<ItemTable> mItemTable; |
| |
| // Start offset from composition time to presentation time. |
| // Support shift only for video tracks through mElstShiftStartTicks for now. |
| uint64_t mElstShiftStartTicks; |
| |
| size_t parseNALSize(const uint8_t *data) const; |
| status_t parseChunk(off64_t *offset); |
| status_t parseTrackFragmentHeader(off64_t offset, off64_t size); |
| status_t parseTrackFragmentRun(off64_t offset, off64_t size); |
| status_t parseSampleAuxiliaryInformationSizes(off64_t offset, off64_t size); |
| status_t parseSampleAuxiliaryInformationOffsets(off64_t offset, off64_t size); |
| status_t parseClearEncryptedSizes(off64_t offset, bool isSubsampleEncryption, uint32_t flags); |
| status_t parseSampleEncryption(off64_t offset); |
| // returns -1 for invalid layer ID |
| int32_t parseHEVCLayerId(const uint8_t *data, size_t size); |
| |
| struct TrackFragmentHeaderInfo { |
| enum Flags { |
| kBaseDataOffsetPresent = 0x01, |
| kSampleDescriptionIndexPresent = 0x02, |
| kDefaultSampleDurationPresent = 0x08, |
| kDefaultSampleSizePresent = 0x10, |
| kDefaultSampleFlagsPresent = 0x20, |
| kDurationIsEmpty = 0x10000, |
| }; |
| |
| uint32_t mTrackID; |
| uint32_t mFlags; |
| uint64_t mBaseDataOffset; |
| uint32_t mSampleDescriptionIndex; |
| uint32_t mDefaultSampleDuration; |
| uint32_t mDefaultSampleSize; |
| uint32_t mDefaultSampleFlags; |
| |
| uint64_t mDataOffset; |
| }; |
| TrackFragmentHeaderInfo mTrackFragmentHeaderInfo; |
| |
| struct Sample { |
| off64_t offset; |
| size_t size; |
| uint32_t duration; |
| int32_t compositionOffset; |
| uint8_t iv[16]; |
| Vector<size_t> clearsizes; |
| Vector<size_t> encryptedsizes; |
| }; |
| Vector<Sample> mCurrentSamples; |
| |
| MPEG4Source(const MPEG4Source &); |
| MPEG4Source &operator=(const MPEG4Source &); |
| }; |
| |
| // This custom data source wraps an existing one and satisfies requests |
| // falling entirely within a cached range from the cache while forwarding |
| // all remaining requests to the wrapped datasource. |
| // This is used to cache the full sampletable metadata for a single track, |
| // possibly wrapping multiple times to cover all tracks, i.e. |
| // Each CachedRangedDataSource caches the sampletable metadata for a single track. |
| |
| class CachedRangedDataSource : public DataSourceHelper { |
| public: |
| explicit CachedRangedDataSource(DataSourceHelper *source); |
| virtual ~CachedRangedDataSource(); |
| |
| ssize_t readAt(off64_t offset, void *data, size_t size) override; |
| status_t getSize(off64_t *size) override; |
| uint32_t flags() override; |
| |
| status_t setCachedRange(off64_t offset, size_t size, bool assumeSourceOwnershipOnSuccess); |
| |
| |
| private: |
| Mutex mLock; |
| |
| DataSourceHelper *mSource; |
| bool mOwnsDataSource; |
| off64_t mCachedOffset; |
| size_t mCachedSize; |
| uint8_t *mCache; |
| |
| void clearCache(); |
| |
| CachedRangedDataSource(const CachedRangedDataSource &); |
| CachedRangedDataSource &operator=(const CachedRangedDataSource &); |
| }; |
| |
| CachedRangedDataSource::CachedRangedDataSource(DataSourceHelper *source) |
| : DataSourceHelper(source), |
| mSource(source), |
| mOwnsDataSource(false), |
| mCachedOffset(0), |
| mCachedSize(0), |
| mCache(NULL) { |
| } |
| |
| CachedRangedDataSource::~CachedRangedDataSource() { |
| clearCache(); |
| if (mOwnsDataSource) { |
| delete mSource; |
| } |
| } |
| |
| void CachedRangedDataSource::clearCache() { |
| if (mCache) { |
| free(mCache); |
| mCache = NULL; |
| } |
| |
| mCachedOffset = 0; |
| mCachedSize = 0; |
| } |
| |
| ssize_t CachedRangedDataSource::readAt(off64_t offset, void *data, size_t size) { |
| Mutex::Autolock autoLock(mLock); |
| |
| if (isInRange(mCachedOffset, mCachedSize, offset, size)) { |
| memcpy(data, &mCache[offset - mCachedOffset], size); |
| return size; |
| } |
| |
| return mSource->readAt(offset, data, size); |
| } |
| |
| status_t CachedRangedDataSource::getSize(off64_t *size) { |
| return mSource->getSize(size); |
| } |
| |
| uint32_t CachedRangedDataSource::flags() { |
| return mSource->flags(); |
| } |
| |
| status_t CachedRangedDataSource::setCachedRange(off64_t offset, |
| size_t size, |
| bool assumeSourceOwnershipOnSuccess) { |
| Mutex::Autolock autoLock(mLock); |
| |
| clearCache(); |
| |
| mCache = (uint8_t *)malloc(size); |
| |
| if (mCache == NULL) { |
| return -ENOMEM; |
| } |
| |
| mCachedOffset = offset; |
| mCachedSize = size; |
| |
| ssize_t err = mSource->readAt(mCachedOffset, mCache, mCachedSize); |
| |
| if (err < (ssize_t)size) { |
| clearCache(); |
| |
| return ERROR_IO; |
| } |
| mOwnsDataSource = assumeSourceOwnershipOnSuccess; |
| return OK; |
| } |
| |
| //////////////////////////////////////////////////////////////////////////////// |
| |
| static const bool kUseHexDump = false; |
| |
| static const char *FourCC2MIME(uint32_t fourcc) { |
| switch (fourcc) { |
| case FOURCC("mp4a"): |
| return MEDIA_MIMETYPE_AUDIO_AAC; |
| |
| case FOURCC("samr"): |
| return MEDIA_MIMETYPE_AUDIO_AMR_NB; |
| |
| case FOURCC("sawb"): |
| return MEDIA_MIMETYPE_AUDIO_AMR_WB; |
| |
| case FOURCC("ec-3"): |
| return MEDIA_MIMETYPE_AUDIO_EAC3; |
| |
| case FOURCC("mp4v"): |
| return MEDIA_MIMETYPE_VIDEO_MPEG4; |
| |
| case FOURCC("s263"): |
| case FOURCC("h263"): |
| case FOURCC("H263"): |
| return MEDIA_MIMETYPE_VIDEO_H263; |
| |
| case FOURCC("avc1"): |
| return MEDIA_MIMETYPE_VIDEO_AVC; |
| |
| case FOURCC("hvc1"): |
| case FOURCC("hev1"): |
| return MEDIA_MIMETYPE_VIDEO_HEVC; |
| case FOURCC("ac-4"): |
| return MEDIA_MIMETYPE_AUDIO_AC4; |
| case FOURCC("Opus"): |
| return MEDIA_MIMETYPE_AUDIO_OPUS; |
| |
| case FOURCC("twos"): |
| case FOURCC("sowt"): |
| return MEDIA_MIMETYPE_AUDIO_RAW; |
| case FOURCC("alac"): |
| return MEDIA_MIMETYPE_AUDIO_ALAC; |
| case FOURCC("fLaC"): |
| return MEDIA_MIMETYPE_AUDIO_FLAC; |
| case FOURCC("av01"): |
| return MEDIA_MIMETYPE_VIDEO_AV1; |
| case FOURCC(".mp3"): |
| case 0x6D730055: // "ms U" mp3 audio |
| return MEDIA_MIMETYPE_AUDIO_MPEG; |
| default: |
| ALOGW("Unknown fourcc: %c%c%c%c", |
| (fourcc >> 24) & 0xff, |
| (fourcc >> 16) & 0xff, |
| (fourcc >> 8) & 0xff, |
| fourcc & 0xff |
| ); |
| return "application/octet-stream"; |
| } |
| } |
| |
| static bool AdjustChannelsAndRate(uint32_t fourcc, uint32_t *channels, uint32_t *rate) { |
| if (!strcasecmp(MEDIA_MIMETYPE_AUDIO_AMR_NB, FourCC2MIME(fourcc))) { |
| // AMR NB audio is always mono, 8kHz |
| *channels = 1; |
| *rate = 8000; |
| return true; |
| } else if (!strcasecmp(MEDIA_MIMETYPE_AUDIO_AMR_WB, FourCC2MIME(fourcc))) { |
| // AMR WB audio is always mono, 16kHz |
| *channels = 1; |
| *rate = 16000; |
| return true; |
| } |
| return false; |
| } |
| |
| MPEG4Extractor::MPEG4Extractor(DataSourceHelper *source, const char *mime) |
| : mMoofOffset(0), |
| mMoofFound(false), |
| mMdatFound(false), |
| mDataSource(source), |
| mInitCheck(NO_INIT), |
| mHeaderTimescale(0), |
| mIsQT(false), |
| mIsHeif(false), |
| mHasMoovBox(false), |
| mPreferHeif(mime != NULL && !strcasecmp(mime, MEDIA_MIMETYPE_CONTAINER_HEIF)), |
| mFirstTrack(NULL), |
| mLastTrack(NULL) { |
| ALOGV("mime=%s, mPreferHeif=%d", mime, mPreferHeif); |
| mFileMetaData = AMediaFormat_new(); |
| } |
| |
| MPEG4Extractor::~MPEG4Extractor() { |
| Track *track = mFirstTrack; |
| while (track) { |
| Track *next = track->next; |
| |
| delete track; |
| track = next; |
| } |
| mFirstTrack = mLastTrack = NULL; |
| |
| for (size_t i = 0; i < mPssh.size(); i++) { |
| delete [] mPssh[i].data; |
| } |
| mPssh.clear(); |
| |
| delete mDataSource; |
| AMediaFormat_delete(mFileMetaData); |
| } |
| |
| uint32_t MPEG4Extractor::flags() const { |
| return CAN_PAUSE | |
| ((mMoofOffset == 0 || mSidxEntries.size() != 0) ? |
| (CAN_SEEK_BACKWARD | CAN_SEEK_FORWARD | CAN_SEEK) : 0); |
| } |
| |
| media_status_t MPEG4Extractor::getMetaData(AMediaFormat *meta) { |
| status_t err; |
| if ((err = readMetaData()) != OK) { |
| return AMEDIA_ERROR_UNKNOWN; |
| } |
| AMediaFormat_copy(meta, mFileMetaData); |
| return AMEDIA_OK; |
| } |
| |
| size_t MPEG4Extractor::countTracks() { |
| status_t err; |
| if ((err = readMetaData()) != OK) { |
| ALOGV("MPEG4Extractor::countTracks: no tracks"); |
| return 0; |
| } |
| |
| size_t n = 0; |
| Track *track = mFirstTrack; |
| while (track) { |
| ++n; |
| track = track->next; |
| } |
| |
| ALOGV("MPEG4Extractor::countTracks: %zu tracks", n); |
| return n; |
| } |
| |
| media_status_t MPEG4Extractor::getTrackMetaData( |
| AMediaFormat *meta, |
| size_t index, uint32_t flags) { |
| status_t err; |
| if ((err = readMetaData()) != OK) { |
| return AMEDIA_ERROR_UNKNOWN; |
| } |
| |
| Track *track = mFirstTrack; |
| while (index > 0) { |
| if (track == NULL) { |
| return AMEDIA_ERROR_UNKNOWN; |
| } |
| |
| track = track->next; |
| --index; |
| } |
| |
| if (track == NULL) { |
| return AMEDIA_ERROR_UNKNOWN; |
| } |
| |
| [=] { |
| int64_t duration; |
| int32_t samplerate; |
| // Only for audio track. |
| if (track->has_elst && mHeaderTimescale != 0 && |
| AMediaFormat_getInt64(track->meta, AMEDIAFORMAT_KEY_DURATION, &duration) && |
| AMediaFormat_getInt32(track->meta, AMEDIAFORMAT_KEY_SAMPLE_RATE, &samplerate)) { |
| |
| // Elst has to be processed only the first time this function is called. |
| track->has_elst = false; |
| |
| if (track->elst_segment_duration > INT64_MAX) { |
| return; |
| } |
| int64_t segment_duration = track->elst_segment_duration; |
| int64_t media_time = track->elst_media_time; |
| int64_t halfscale = track->timescale / 2; |
| |
| ALOGV("segment_duration = %" PRId64 ", media_time = %" PRId64 |
| ", halfscale = %" PRId64 ", mdhd_timescale = %d, track_timescale = %u", |
| segment_duration, media_time, |
| halfscale, mHeaderTimescale, track->timescale); |
| |
| if ((uint32_t)samplerate != track->timescale){ |
| ALOGV("samplerate:%" PRId32 ", track->timescale and samplerate are different!", |
| samplerate); |
| } |
| // Both delay and paddingsamples have to be set inorder for either to be |
| // effective in the lower layers. |
| int64_t delay = 0; |
| if (media_time > 0) { // Gapless playback |
| // delay = ((media_time * samplerate) + halfscale) / track->timescale; |
| if (__builtin_mul_overflow(media_time, samplerate, &delay) || |
| __builtin_add_overflow(delay, halfscale, &delay) || |
| (delay /= track->timescale, false) || |
| delay > INT32_MAX || |
| delay < INT32_MIN) { |
| ALOGW("ignoring edit list with bogus values"); |
| return; |
| } |
| } |
| ALOGV("delay = %" PRId64, delay); |
| AMediaFormat_setInt32(track->meta, AMEDIAFORMAT_KEY_ENCODER_DELAY, delay); |
| |
| int64_t paddingsamples = 0; |
| if (segment_duration > 0) { |
| int64_t scaled_duration; |
| // scaled_duration = duration * mHeaderTimescale; |
| if (__builtin_mul_overflow(duration, mHeaderTimescale, &scaled_duration)) { |
| return; |
| } |
| ALOGV("scaled_duration = %" PRId64, scaled_duration); |
| |
| int64_t segment_end; |
| int64_t padding; |
| int64_t segment_duration_e6; |
| int64_t media_time_scaled_e6; |
| int64_t media_time_scaled; |
| // padding = scaled_duration - ((segment_duration * 1000000) + |
| // ((media_time * mHeaderTimescale * 1000000)/track->timescale) ) |
| // segment_duration is based on timescale in movie header box(mdhd) |
| // media_time is based on timescale track header/media timescale |
| if (__builtin_mul_overflow(segment_duration, 1000000, &segment_duration_e6) || |
| __builtin_mul_overflow(media_time, mHeaderTimescale, &media_time_scaled) || |
| __builtin_mul_overflow(media_time_scaled, 1000000, &media_time_scaled_e6)) { |
| return; |
| } |
| media_time_scaled_e6 /= track->timescale; |
| if (__builtin_add_overflow(segment_duration_e6, media_time_scaled_e6, &segment_end) |
| || __builtin_sub_overflow(scaled_duration, segment_end, &padding)) { |
| return; |
| } |
| ALOGV("segment_end = %" PRId64 ", padding = %" PRId64, segment_end, padding); |
| // track duration from media header (which is what AMEDIAFORMAT_KEY_DURATION is) |
| // might be slightly shorter than the segment duration, which would make the |
| // padding negative. Clamp to zero. |
| if (padding > 0) { |
| int64_t halfscale_mht = mHeaderTimescale / 2; |
| int64_t halfscale_e6; |
| int64_t timescale_e6; |
| // paddingsamples = ((padding * samplerate) + (halfscale_mht * 1000000)) |
| // / (mHeaderTimescale * 1000000); |
| if (__builtin_mul_overflow(padding, samplerate, &paddingsamples) || |
| __builtin_mul_overflow(halfscale_mht, 1000000, &halfscale_e6) || |
| __builtin_mul_overflow(mHeaderTimescale, 1000000, ×cale_e6) || |
| __builtin_add_overflow(paddingsamples, halfscale_e6, &paddingsamples) || |
| (paddingsamples /= timescale_e6, false) || |
| paddingsamples > INT32_MAX) { |
| return; |
| } |
| } |
| } |
| ALOGV("paddingsamples = %" PRId64, paddingsamples); |
| AMediaFormat_setInt32(track->meta, AMEDIAFORMAT_KEY_ENCODER_PADDING, paddingsamples); |
| } |
| }(); |
| |
| if ((flags & kIncludeExtensiveMetaData) |
| && !track->includes_expensive_metadata) { |
| track->includes_expensive_metadata = true; |
| |
| const char *mime; |
| CHECK(AMediaFormat_getString(track->meta, AMEDIAFORMAT_KEY_MIME, &mime)); |
| if (!strncasecmp("video/", mime, 6)) { |
| // MPEG2 tracks do not provide CSD, so read the stream header |
| if (!strcmp(mime, MEDIA_MIMETYPE_VIDEO_MPEG2)) { |
| off64_t offset; |
| size_t size; |
| if (track->sampleTable->getMetaDataForSample( |
| 0 /* sampleIndex */, &offset, &size, NULL /* sampleTime */) == OK) { |
| if (size > kMaxTrackHeaderSize) { |
| size = kMaxTrackHeaderSize; |
| } |
| uint8_t header[kMaxTrackHeaderSize]; |
| if (mDataSource->readAt(offset, &header, size) == (ssize_t)size) { |
| AMediaFormat_setBuffer(track->meta, |
| AMEDIAFORMAT_KEY_MPEG2_STREAM_HEADER, header, size); |
| } |
| } |
| } |
| |
| if (mMoofOffset > 0) { |
| int64_t duration; |
| if (AMediaFormat_getInt64(track->meta, |
| AMEDIAFORMAT_KEY_DURATION, &duration)) { |
| // nothing fancy, just pick a frame near 1/4th of the duration |
| AMediaFormat_setInt64(track->meta, |
| AMEDIAFORMAT_KEY_THUMBNAIL_TIME, duration / 4); |
| } |
| } else { |
| uint32_t sampleIndex; |
| uint64_t sampleTime; |
| if (track->timescale != 0 && |
| track->sampleTable->findThumbnailSample(&sampleIndex) == OK |
| && track->sampleTable->getMetaDataForSample( |
| sampleIndex, NULL /* offset */, NULL /* size */, |
| &sampleTime) == OK) { |
| AMediaFormat_setInt64(track->meta, |
| AMEDIAFORMAT_KEY_THUMBNAIL_TIME, |
| ((int64_t)sampleTime * 1000000) / track->timescale); |
| } |
| } |
| } |
| } |
| |
| AMediaFormat_copy(meta, track->meta); |
| return AMEDIA_OK; |
| } |
| |
| status_t MPEG4Extractor::readMetaData() { |
| if (mInitCheck != NO_INIT) { |
| return mInitCheck; |
| } |
| |
| off64_t offset = 0; |
| status_t err; |
| bool sawMoovOrSidx = false; |
| |
| while (!((mHasMoovBox && sawMoovOrSidx && (mMdatFound || mMoofFound)) || |
| (mIsHeif && (mPreferHeif || !mHasMoovBox) && |
| (mItemTable != NULL) && mItemTable->isValid()))) { |
| off64_t orig_offset = offset; |
| err = parseChunk(&offset, 0); |
| |
| if (err != OK && err != UNKNOWN_ERROR) { |
| break; |
| } else if (offset <= orig_offset) { |
| // only continue parsing if the offset was advanced, |
| // otherwise we might end up in an infinite loop |
| ALOGE("did not advance: %lld->%lld", (long long)orig_offset, (long long)offset); |
| err = ERROR_MALFORMED; |
| break; |
| } else if (err == UNKNOWN_ERROR) { |
| sawMoovOrSidx = true; |
| } |
| } |
| |
| if (mIsHeif && (mItemTable != NULL) && (mItemTable->countImages() > 0)) { |
| off64_t exifOffset; |
| size_t exifSize; |
| if (mItemTable->getExifOffsetAndSize(&exifOffset, &exifSize) == OK) { |
| AMediaFormat_setInt64(mFileMetaData, |
| AMEDIAFORMAT_KEY_EXIF_OFFSET, (int64_t)exifOffset); |
| AMediaFormat_setInt64(mFileMetaData, |
| AMEDIAFORMAT_KEY_EXIF_SIZE, (int64_t)exifSize); |
| } |
| for (uint32_t imageIndex = 0; |
| imageIndex < mItemTable->countImages(); imageIndex++) { |
| AMediaFormat *meta = mItemTable->getImageMeta(imageIndex); |
| if (meta == NULL) { |
| ALOGE("heif image %u has no meta!", imageIndex); |
| continue; |
| } |
| // Some heif files advertise image sequence brands (eg. 'hevc') in |
| // ftyp box, but don't have any valid tracks in them. Instead of |
| // reporting the entire file as malformed, we override the error |
| // to allow still images to be extracted. |
| if (err != OK) { |
| ALOGW("Extracting still images only"); |
| err = OK; |
| } |
| mInitCheck = OK; |
| |
| ALOGV("adding HEIF image track %u", imageIndex); |
| Track *track = new Track; |
| if (mLastTrack != NULL) { |
| mLastTrack->next = track; |
| } else { |
| mFirstTrack = track; |
| } |
| mLastTrack = track; |
| |
| track->meta = meta; |
| AMediaFormat_setInt32(track->meta, AMEDIAFORMAT_KEY_TRACK_ID, imageIndex); |
| track->timescale = 1000000; |
| } |
| } |
| |
| if (mInitCheck == OK) { |
| if (findTrackByMimePrefix("video/") != NULL) { |
| AMediaFormat_setString(mFileMetaData, |
| AMEDIAFORMAT_KEY_MIME, MEDIA_MIMETYPE_CONTAINER_MPEG4); |
| } else if (findTrackByMimePrefix("audio/") != NULL) { |
| AMediaFormat_setString(mFileMetaData, |
| AMEDIAFORMAT_KEY_MIME, "audio/mp4"); |
| } else if (findTrackByMimePrefix( |
| MEDIA_MIMETYPE_IMAGE_ANDROID_HEIC) != NULL) { |
| AMediaFormat_setString(mFileMetaData, |
| AMEDIAFORMAT_KEY_MIME, MEDIA_MIMETYPE_CONTAINER_HEIF); |
| } else { |
| AMediaFormat_setString(mFileMetaData, |
| AMEDIAFORMAT_KEY_MIME, "application/octet-stream"); |
| } |
| } else { |
| mInitCheck = err; |
| } |
| |
| CHECK_NE(err, (status_t)NO_INIT); |
| |
| // copy pssh data into file metadata |
| uint64_t psshsize = 0; |
| for (size_t i = 0; i < mPssh.size(); i++) { |
| psshsize += 20 + mPssh[i].datalen; |
| } |
| if (psshsize > 0 && psshsize <= UINT32_MAX) { |
| char *buf = (char*)malloc(psshsize); |
| if (!buf) { |
| ALOGE("b/28471206"); |
| return NO_MEMORY; |
| } |
| char *ptr = buf; |
| for (size_t i = 0; i < mPssh.size(); i++) { |
| memcpy(ptr, mPssh[i].uuid, 20); // uuid + length |
| memcpy(ptr + 20, mPssh[i].data, mPssh[i].datalen); |
| ptr += (20 + mPssh[i].datalen); |
| } |
| AMediaFormat_setBuffer(mFileMetaData, AMEDIAFORMAT_KEY_PSSH, buf, psshsize); |
| free(buf); |
| } |
| |
| return mInitCheck; |
| } |
| |
| struct PathAdder { |
| PathAdder(Vector<uint32_t> *path, uint32_t chunkType) |
| : mPath(path) { |
| mPath->push(chunkType); |
| } |
| |
| ~PathAdder() { |
| mPath->pop(); |
| } |
| |
| private: |
| Vector<uint32_t> *mPath; |
| |
| PathAdder(const PathAdder &); |
| PathAdder &operator=(const PathAdder &); |
| }; |
| |
| static bool underMetaDataPath(const Vector<uint32_t> &path) { |
| return path.size() >= 5 |
| && path[0] == FOURCC("moov") |
| && path[1] == FOURCC("udta") |
| && path[2] == FOURCC("meta") |
| && path[3] == FOURCC("ilst"); |
| } |
| |
| static bool underQTMetaPath(const Vector<uint32_t> &path, int32_t depth) { |
| return path.size() >= 2 |
| && path[0] == FOURCC("moov") |
| && path[1] == FOURCC("meta") |
| && (depth == 2 |
| || (depth == 3 |
| && (path[2] == FOURCC("hdlr") |
| || path[2] == FOURCC("ilst") |
| || path[2] == FOURCC("keys")))); |
| } |
| |
| // Given a time in seconds since Jan 1 1904, produce a human-readable string. |
| static bool convertTimeToDate(int64_t time_1904, String8 *s) { |
| // delta between mpeg4 time and unix epoch time |
| static const int64_t delta = (((66 * 365 + 17) * 24) * 3600); |
| if (time_1904 < INT64_MIN + delta) { |
| return false; |
| } |
| time_t time_1970 = time_1904 - delta; |
| |
| char tmp[32]; |
| struct tm* tm = gmtime(&time_1970); |
| if (tm != NULL && |
| strftime(tmp, sizeof(tmp), "%Y%m%dT%H%M%S.000Z", tm) > 0) { |
| s->setTo(tmp); |
| return true; |
| } |
| return false; |
| } |
| |
| status_t MPEG4Extractor::parseChunk(off64_t *offset, int depth) { |
| ALOGV("entering parseChunk %lld/%d", (long long)*offset, depth); |
| |
| if (*offset < 0) { |
| ALOGE("b/23540914"); |
| return ERROR_MALFORMED; |
| } |
| if (depth > 100) { |
| ALOGE("b/27456299"); |
| return ERROR_MALFORMED; |
| } |
| uint32_t hdr[2]; |
| if (mDataSource->readAt(*offset, hdr, 8) < 8) { |
| return ERROR_IO; |
| } |
| uint64_t chunk_size = ntohl(hdr[0]); |
| int32_t chunk_type = ntohl(hdr[1]); |
| off64_t data_offset = *offset + 8; |
| |
| if (chunk_size == 1) { |
| if (mDataSource->readAt(*offset + 8, &chunk_size, 8) < 8) { |
| return ERROR_IO; |
| } |
| chunk_size = ntoh64(chunk_size); |
| data_offset += 8; |
| |
| if (chunk_size < 16) { |
| // The smallest valid chunk is 16 bytes long in this case. |
| return ERROR_MALFORMED; |
| } |
| } else if (chunk_size == 0) { |
| if (depth == 0) { |
| // atom extends to end of file |
| off64_t sourceSize; |
| if (mDataSource->getSize(&sourceSize) == OK) { |
| chunk_size = (sourceSize - *offset); |
| } else { |
| // XXX could we just pick a "sufficiently large" value here? |
| ALOGE("atom size is 0, and data source has no size"); |
| return ERROR_MALFORMED; |
| } |
| } else { |
| // not allowed for non-toplevel atoms, skip it |
| *offset += 4; |
| return OK; |
| } |
| } else if (chunk_size < 8) { |
| // The smallest valid chunk is 8 bytes long. |
| ALOGE("invalid chunk size: %" PRIu64, chunk_size); |
| return ERROR_MALFORMED; |
| } |
| |
| char chunk[5]; |
| MakeFourCCString(chunk_type, chunk); |
| ALOGV("chunk: %s @ %lld, %d", chunk, (long long)*offset, depth); |
| |
| if (kUseHexDump) { |
| static const char kWhitespace[] = " "; |
| const char *indent = &kWhitespace[sizeof(kWhitespace) - 1 - 2 * depth]; |
| printf("%sfound chunk '%s' of size %" PRIu64 "\n", indent, chunk, chunk_size); |
| |
| char buffer[256]; |
| size_t n = chunk_size; |
| if (n > sizeof(buffer)) { |
| n = sizeof(buffer); |
| } |
| if (mDataSource->readAt(*offset, buffer, n) |
| < (ssize_t)n) { |
| return ERROR_IO; |
| } |
| |
| hexdump(buffer, n); |
| } |
| |
| PathAdder autoAdder(&mPath, chunk_type); |
| |
| // (data_offset - *offset) is either 8 or 16 |
| off64_t chunk_data_size = chunk_size - (data_offset - *offset); |
| if (chunk_data_size < 0) { |
| ALOGE("b/23540914"); |
| return ERROR_MALFORMED; |
| } |
| if (chunk_type != FOURCC("mdat") && chunk_data_size > kMaxAtomSize) { |
| char errMsg[100]; |
| sprintf(errMsg, "%s atom has size %" PRId64, chunk, chunk_data_size); |
| ALOGE("%s (b/28615448)", errMsg); |
| android_errorWriteWithInfoLog(0x534e4554, "28615448", -1, errMsg, strlen(errMsg)); |
| return ERROR_MALFORMED; |
| } |
| |
| if (chunk_type != FOURCC("cprt") |
| && chunk_type != FOURCC("covr") |
| && mPath.size() == 5 && underMetaDataPath(mPath)) { |
| off64_t stop_offset = *offset + chunk_size; |
| *offset = data_offset; |
| while (*offset < stop_offset) { |
| status_t err = parseChunk(offset, depth + 1); |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| if (*offset != stop_offset) { |
| return ERROR_MALFORMED; |
| } |
| |
| return OK; |
| } |
| |
| switch(chunk_type) { |
| case FOURCC("moov"): |
| case FOURCC("trak"): |
| case FOURCC("mdia"): |
| case FOURCC("minf"): |
| case FOURCC("dinf"): |
| case FOURCC("stbl"): |
| case FOURCC("mvex"): |
| case FOURCC("moof"): |
| case FOURCC("traf"): |
| case FOURCC("mfra"): |
| case FOURCC("udta"): |
| case FOURCC("ilst"): |
| case FOURCC("sinf"): |
| case FOURCC("schi"): |
| case FOURCC("edts"): |
| case FOURCC("wave"): |
| { |
| if (chunk_type == FOURCC("moov") && depth != 0) { |
| ALOGE("moov: depth %d", depth); |
| return ERROR_MALFORMED; |
| } |
| |
| if (chunk_type == FOURCC("moov") && mInitCheck == OK) { |
| ALOGE("duplicate moov"); |
| return ERROR_MALFORMED; |
| } |
| |
| if (chunk_type == FOURCC("moof") && !mMoofFound) { |
| // store the offset of the first segment |
| mMoofFound = true; |
| mMoofOffset = *offset; |
| } |
| |
| if (chunk_type == FOURCC("stbl")) { |
| ALOGV("sampleTable chunk is %" PRIu64 " bytes long.", chunk_size); |
| |
| if (mDataSource->flags() |
| & (DataSourceBase::kWantsPrefetching |
| | DataSourceBase::kIsCachingDataSource)) { |
| CachedRangedDataSource *cachedSource = |
| new CachedRangedDataSource(mDataSource); |
| |
| if (cachedSource->setCachedRange( |
| *offset, chunk_size, |
| true /* assume ownership on success */) == OK) { |
| mDataSource = cachedSource; |
| } else { |
| delete cachedSource; |
| } |
| } |
| |
| if (mLastTrack == NULL) { |
| return ERROR_MALFORMED; |
| } |
| |
| mLastTrack->sampleTable = new SampleTable(mDataSource); |
| } |
| |
| bool isTrack = false; |
| if (chunk_type == FOURCC("trak")) { |
| if (depth != 1) { |
| ALOGE("trak: depth %d", depth); |
| return ERROR_MALFORMED; |
| } |
| isTrack = true; |
| |
| ALOGV("adding new track"); |
| Track *track = new Track; |
| if (mLastTrack) { |
| mLastTrack->next = track; |
| } else { |
| mFirstTrack = track; |
| } |
| mLastTrack = track; |
| |
| track->meta = AMediaFormat_new(); |
| AMediaFormat_setString(track->meta, |
| AMEDIAFORMAT_KEY_MIME, "application/octet-stream"); |
| } |
| |
| off64_t stop_offset = *offset + chunk_size; |
| *offset = data_offset; |
| while (*offset < stop_offset) { |
| |
| // pass udata terminate |
| if (mIsQT && stop_offset - *offset == 4 && chunk_type == FOURCC("udta")) { |
| // handle the case that udta terminates with terminate code x00000000 |
| // note that 0 terminator is optional and we just handle this case. |
| uint32_t terminate_code = 1; |
| mDataSource->readAt(*offset, &terminate_code, 4); |
| if (0 == terminate_code) { |
| *offset += 4; |
| ALOGD("Terminal code for udta"); |
| continue; |
| } else { |
| ALOGW("invalid udta Terminal code"); |
| } |
| } |
| |
| status_t err = parseChunk(offset, depth + 1); |
| if (err != OK) { |
| if (isTrack) { |
| mLastTrack->skipTrack = true; |
| break; |
| } |
| return err; |
| } |
| } |
| |
| if (*offset != stop_offset) { |
| return ERROR_MALFORMED; |
| } |
| |
| if (isTrack) { |
| int32_t trackId; |
| // There must be exactly one track header per track. |
| |
| if (!AMediaFormat_getInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_TRACK_ID, &trackId)) { |
| mLastTrack->skipTrack = true; |
| } |
| |
| status_t err = verifyTrack(mLastTrack); |
| if (err != OK) { |
| mLastTrack->skipTrack = true; |
| } |
| |
| |
| if (mLastTrack->skipTrack) { |
| ALOGV("skipping this track..."); |
| Track *cur = mFirstTrack; |
| |
| if (cur == mLastTrack) { |
| delete cur; |
| mFirstTrack = mLastTrack = NULL; |
| } else { |
| while (cur && cur->next != mLastTrack) { |
| cur = cur->next; |
| } |
| if (cur) { |
| cur->next = NULL; |
| } |
| delete mLastTrack; |
| mLastTrack = cur; |
| } |
| |
| return OK; |
| } |
| |
| // place things we built elsewhere into their final locations |
| |
| // put aggregated tx3g data into the metadata |
| if (mLastTrack->mTx3gFilled > 0) { |
| ALOGV("Putting %zu bytes of tx3g data into meta data", |
| mLastTrack->mTx3gFilled); |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_TEXT_FORMAT_DATA, |
| mLastTrack->mTx3gBuffer, mLastTrack->mTx3gFilled); |
| // drop it now to reduce our footprint |
| free(mLastTrack->mTx3gBuffer); |
| mLastTrack->mTx3gBuffer = NULL; |
| } |
| |
| } else if (chunk_type == FOURCC("moov")) { |
| mInitCheck = OK; |
| |
| return UNKNOWN_ERROR; // Return a dummy error. |
| } |
| break; |
| } |
| |
| case FOURCC("schm"): |
| { |
| |
| *offset += chunk_size; |
| if (!mLastTrack) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint32_t scheme_type; |
| if (mDataSource->readAt(data_offset + 4, &scheme_type, 4) < 4) { |
| return ERROR_IO; |
| } |
| scheme_type = ntohl(scheme_type); |
| int32_t mode = kCryptoModeUnencrypted; |
| switch(scheme_type) { |
| case FOURCC("cbc1"): |
| { |
| mode = kCryptoModeAesCbc; |
| break; |
| } |
| case FOURCC("cbcs"): |
| { |
| mode = kCryptoModeAesCbc; |
| mLastTrack->subsample_encryption = true; |
| break; |
| } |
| case FOURCC("cenc"): |
| { |
| mode = kCryptoModeAesCtr; |
| break; |
| } |
| case FOURCC("cens"): |
| { |
| mode = kCryptoModeAesCtr; |
| mLastTrack->subsample_encryption = true; |
| break; |
| } |
| } |
| if (mode != kCryptoModeUnencrypted) { |
| AMediaFormat_setInt32(mLastTrack->meta, AMEDIAFORMAT_KEY_CRYPTO_MODE, mode); |
| } |
| break; |
| } |
| |
| |
| case FOURCC("elst"): |
| { |
| *offset += chunk_size; |
| |
| if (!mLastTrack) { |
| return ERROR_MALFORMED; |
| } |
| |
| // See 14496-12 8.6.6 |
| uint8_t version; |
| if (mDataSource->readAt(data_offset, &version, 1) < 1) { |
| return ERROR_IO; |
| } |
| |
| uint32_t entry_count; |
| if (!mDataSource->getUInt32(data_offset + 4, &entry_count)) { |
| return ERROR_IO; |
| } |
| |
| if (entry_count != 1) { |
| // we only support a single entry at the moment, for gapless playback |
| // or start offset |
| ALOGW("ignoring edit list with %d entries", entry_count); |
| } else { |
| off64_t entriesoffset = data_offset + 8; |
| uint64_t segment_duration; |
| int64_t media_time; |
| |
| if (version == 1) { |
| if (!mDataSource->getUInt64(entriesoffset, &segment_duration) || |
| !mDataSource->getUInt64(entriesoffset + 8, (uint64_t*)&media_time)) { |
| return ERROR_IO; |
| } |
| } else if (version == 0) { |
| uint32_t sd; |
| int32_t mt; |
| if (!mDataSource->getUInt32(entriesoffset, &sd) || |
| !mDataSource->getUInt32(entriesoffset + 4, (uint32_t*)&mt)) { |
| return ERROR_IO; |
| } |
| segment_duration = sd; |
| media_time = mt; |
| } else { |
| return ERROR_IO; |
| } |
| |
| // save these for later, because the elst atom might precede |
| // the atoms that actually gives us the duration and sample rate |
| // needed to calculate the padding and delay values |
| mLastTrack->has_elst = true; |
| mLastTrack->elst_media_time = media_time; |
| mLastTrack->elst_segment_duration = segment_duration; |
| } |
| break; |
| } |
| |
| case FOURCC("frma"): |
| { |
| *offset += chunk_size; |
| |
| uint32_t original_fourcc; |
| if (mDataSource->readAt(data_offset, &original_fourcc, 4) < 4) { |
| return ERROR_IO; |
| } |
| original_fourcc = ntohl(original_fourcc); |
| ALOGV("read original format: %d", original_fourcc); |
| |
| if (mLastTrack == NULL) { |
| return ERROR_MALFORMED; |
| } |
| |
| AMediaFormat_setString(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MIME, FourCC2MIME(original_fourcc)); |
| uint32_t num_channels = 0; |
| uint32_t sample_rate = 0; |
| if (AdjustChannelsAndRate(original_fourcc, &num_channels, &sample_rate)) { |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CHANNEL_COUNT, num_channels); |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_SAMPLE_RATE, sample_rate); |
| } |
| |
| if (!mIsQT && original_fourcc == FOURCC("alac")) { |
| off64_t tmpOffset = *offset; |
| status_t err = parseALACSampleEntry(&tmpOffset); |
| if (err != OK) { |
| ALOGE("parseALACSampleEntry err:%d Line:%d", err, __LINE__); |
| return err; |
| } |
| *offset = tmpOffset + 8; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("tenc"): |
| { |
| *offset += chunk_size; |
| |
| if (chunk_size < 32) { |
| return ERROR_MALFORMED; |
| } |
| |
| // tenc box contains 1 byte version, 3 byte flags, 3 byte default algorithm id, one byte |
| // default IV size, 16 bytes default KeyID |
| // (ISO 23001-7) |
| |
| uint8_t version; |
| if (mDataSource->readAt(data_offset, &version, sizeof(version)) |
| < (ssize_t)sizeof(version)) { |
| return ERROR_IO; |
| } |
| |
| uint8_t buf[4]; |
| memset(buf, 0, 4); |
| if (mDataSource->readAt(data_offset + 4, buf + 1, 3) < 3) { |
| return ERROR_IO; |
| } |
| |
| if (mLastTrack == NULL) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint8_t defaultEncryptedByteBlock = 0; |
| uint8_t defaultSkipByteBlock = 0; |
| uint32_t defaultAlgorithmId = ntohl(*((int32_t*)buf)); |
| if (version == 1) { |
| uint32_t pattern = buf[2]; |
| defaultEncryptedByteBlock = pattern >> 4; |
| defaultSkipByteBlock = pattern & 0xf; |
| if (defaultEncryptedByteBlock == 0 && defaultSkipByteBlock == 0) { |
| // use (1,0) to mean "encrypt everything" |
| defaultEncryptedByteBlock = 1; |
| } |
| } else if (mLastTrack->subsample_encryption) { |
| ALOGW("subsample_encryption should be version 1"); |
| } else if (defaultAlgorithmId > 1) { |
| // only 0 (clear) and 1 (AES-128) are valid |
| ALOGW("defaultAlgorithmId: %u is a reserved value", defaultAlgorithmId); |
| defaultAlgorithmId = 1; |
| } |
| |
| memset(buf, 0, 4); |
| if (mDataSource->readAt(data_offset + 7, buf + 3, 1) < 1) { |
| return ERROR_IO; |
| } |
| uint32_t defaultIVSize = ntohl(*((int32_t*)buf)); |
| |
| if (defaultAlgorithmId == 0 && defaultIVSize != 0) { |
| // only unencrypted data must have 0 IV size |
| return ERROR_MALFORMED; |
| } else if (defaultIVSize != 0 && |
| defaultIVSize != 8 && |
| defaultIVSize != 16) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint8_t defaultKeyId[16]; |
| |
| if (mDataSource->readAt(data_offset + 8, &defaultKeyId, 16) < 16) { |
| return ERROR_IO; |
| } |
| |
| sp<ABuffer> defaultConstantIv; |
| if (defaultAlgorithmId != 0 && defaultIVSize == 0) { |
| |
| uint8_t ivlength; |
| if (mDataSource->readAt(data_offset + 24, &ivlength, sizeof(ivlength)) |
| < (ssize_t)sizeof(ivlength)) { |
| return ERROR_IO; |
| } |
| |
| if (ivlength != 8 && ivlength != 16) { |
| ALOGW("unsupported IV length: %u", ivlength); |
| return ERROR_MALFORMED; |
| } |
| |
| defaultConstantIv = new ABuffer(ivlength); |
| if (mDataSource->readAt(data_offset + 25, defaultConstantIv->data(), ivlength) |
| < (ssize_t)ivlength) { |
| return ERROR_IO; |
| } |
| |
| defaultConstantIv->setRange(0, ivlength); |
| } |
| |
| int32_t tmpAlgorithmId; |
| if (!AMediaFormat_getInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CRYPTO_MODE, &tmpAlgorithmId)) { |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CRYPTO_MODE, defaultAlgorithmId); |
| } |
| |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CRYPTO_DEFAULT_IV_SIZE, defaultIVSize); |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CRYPTO_KEY, defaultKeyId, 16); |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CRYPTO_ENCRYPTED_BYTE_BLOCK, defaultEncryptedByteBlock); |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CRYPTO_SKIP_BYTE_BLOCK, defaultSkipByteBlock); |
| if (defaultConstantIv != NULL) { |
| AMediaFormat_setBuffer(mLastTrack->meta, AMEDIAFORMAT_KEY_CRYPTO_IV, |
| defaultConstantIv->data(), defaultConstantIv->size()); |
| } |
| break; |
| } |
| |
| case FOURCC("tkhd"): |
| { |
| *offset += chunk_size; |
| |
| status_t err; |
| if ((err = parseTrackHeader(data_offset, chunk_data_size)) != OK) { |
| return err; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("tref"): |
| { |
| off64_t stop_offset = *offset + chunk_size; |
| *offset = data_offset; |
| while (*offset < stop_offset) { |
| status_t err = parseChunk(offset, depth + 1); |
| if (err != OK) { |
| return err; |
| } |
| } |
| if (*offset != stop_offset) { |
| return ERROR_MALFORMED; |
| } |
| break; |
| } |
| |
| case FOURCC("thmb"): |
| { |
| *offset += chunk_size; |
| |
| if (mLastTrack != NULL) { |
| // Skip thumbnail track for now since we don't have an |
| // API to retrieve it yet. |
| // The thumbnail track can't be accessed by negative index or time, |
| // because each timed sample has its own corresponding thumbnail |
| // in the thumbnail track. We'll need a dedicated API to retrieve |
| // thumbnail at time instead. |
| mLastTrack->skipTrack = true; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("pssh"): |
| { |
| *offset += chunk_size; |
| |
| PsshInfo pssh; |
| |
| if (mDataSource->readAt(data_offset + 4, &pssh.uuid, 16) < 16) { |
| return ERROR_IO; |
| } |
| |
| uint32_t psshdatalen = 0; |
| if (mDataSource->readAt(data_offset + 20, &psshdatalen, 4) < 4) { |
| return ERROR_IO; |
| } |
| pssh.datalen = ntohl(psshdatalen); |
| ALOGV("pssh data size: %d", pssh.datalen); |
| if (chunk_size < 20 || pssh.datalen > chunk_size - 20) { |
| // pssh data length exceeds size of containing box |
| return ERROR_MALFORMED; |
| } |
| |
| pssh.data = new (std::nothrow) uint8_t[pssh.datalen]; |
| if (pssh.data == NULL) { |
| return ERROR_MALFORMED; |
| } |
| ALOGV("allocated pssh @ %p", pssh.data); |
| ssize_t requested = (ssize_t) pssh.datalen; |
| if (mDataSource->readAt(data_offset + 24, pssh.data, requested) < requested) { |
| delete[] pssh.data; |
| return ERROR_IO; |
| } |
| mPssh.push_back(pssh); |
| |
| break; |
| } |
| |
| case FOURCC("mdhd"): |
| { |
| *offset += chunk_size; |
| |
| if (chunk_data_size < 4 || mLastTrack == NULL) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint8_t version; |
| if (mDataSource->readAt( |
| data_offset, &version, sizeof(version)) |
| < (ssize_t)sizeof(version)) { |
| return ERROR_IO; |
| } |
| |
| off64_t timescale_offset; |
| |
| if (version == 1) { |
| timescale_offset = data_offset + 4 + 16; |
| } else if (version == 0) { |
| timescale_offset = data_offset + 4 + 8; |
| } else { |
| return ERROR_IO; |
| } |
| |
| uint32_t timescale; |
| if (mDataSource->readAt( |
| timescale_offset, ×cale, sizeof(timescale)) |
| < (ssize_t)sizeof(timescale)) { |
| return ERROR_IO; |
| } |
| |
| if (!timescale) { |
| ALOGE("timescale should not be ZERO."); |
| return ERROR_MALFORMED; |
| } |
| |
| mLastTrack->timescale = ntohl(timescale); |
| |
| // 14496-12 says all ones means indeterminate, but some files seem to use |
| // 0 instead. We treat both the same. |
| int64_t duration = 0; |
| if (version == 1) { |
| if (mDataSource->readAt( |
| timescale_offset + 4, &duration, sizeof(duration)) |
| < (ssize_t)sizeof(duration)) { |
| return ERROR_IO; |
| } |
| if (duration != -1) { |
| duration = ntoh64(duration); |
| } |
| } else { |
| uint32_t duration32; |
| if (mDataSource->readAt( |
| timescale_offset + 4, &duration32, sizeof(duration32)) |
| < (ssize_t)sizeof(duration32)) { |
| return ERROR_IO; |
| } |
| if (duration32 != 0xffffffff) { |
| duration = ntohl(duration32); |
| } |
| } |
| if (duration != 0 && mLastTrack->timescale != 0) { |
| long double durationUs = ((long double)duration * 1000000) / mLastTrack->timescale; |
| if (durationUs < 0 || durationUs > INT64_MAX) { |
| ALOGE("cannot represent %lld * 1000000 / %lld in 64 bits", |
| (long long) duration, (long long) mLastTrack->timescale); |
| return ERROR_MALFORMED; |
| } |
| AMediaFormat_setInt64(mLastTrack->meta, AMEDIAFORMAT_KEY_DURATION, durationUs); |
| } |
| |
| uint8_t lang[2]; |
| off64_t lang_offset; |
| if (version == 1) { |
| lang_offset = timescale_offset + 4 + 8; |
| } else if (version == 0) { |
| lang_offset = timescale_offset + 4 + 4; |
| } else { |
| return ERROR_IO; |
| } |
| |
| if (mDataSource->readAt(lang_offset, &lang, sizeof(lang)) |
| < (ssize_t)sizeof(lang)) { |
| return ERROR_IO; |
| } |
| |
| // To get the ISO-639-2/T three character language code |
| // 1 bit pad followed by 3 5-bits characters. Each character |
| // is packed as the difference between its ASCII value and 0x60. |
| char lang_code[4]; |
| lang_code[0] = ((lang[0] >> 2) & 0x1f) + 0x60; |
| lang_code[1] = ((lang[0] & 0x3) << 3 | (lang[1] >> 5)) + 0x60; |
| lang_code[2] = (lang[1] & 0x1f) + 0x60; |
| lang_code[3] = '\0'; |
| |
| AMediaFormat_setString(mLastTrack->meta, AMEDIAFORMAT_KEY_LANGUAGE, lang_code); |
| |
| break; |
| } |
| |
| case FOURCC("stsd"): |
| { |
| uint8_t buffer[8]; |
| if (chunk_data_size < (off64_t)sizeof(buffer)) { |
| return ERROR_MALFORMED; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer, 8) < 8) { |
| return ERROR_IO; |
| } |
| |
| if (U32_AT(buffer) != 0) { |
| // Should be version 0, flags 0. |
| return ERROR_MALFORMED; |
| } |
| |
| uint32_t entry_count = U32_AT(&buffer[4]); |
| |
| if (entry_count > 1) { |
| // For 3GPP timed text, there could be multiple tx3g boxes contain |
| // multiple text display formats. These formats will be used to |
| // display the timed text. |
| // For encrypted files, there may also be more than one entry. |
| const char *mime; |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| CHECK(AMediaFormat_getString(mLastTrack->meta, AMEDIAFORMAT_KEY_MIME, &mime)); |
| if (strcasecmp(mime, MEDIA_MIMETYPE_TEXT_3GPP) && |
| strcasecmp(mime, "application/octet-stream")) { |
| // For now we only support a single type of media per track. |
| mLastTrack->skipTrack = true; |
| *offset += chunk_size; |
| break; |
| } |
| } |
| off64_t stop_offset = *offset + chunk_size; |
| *offset = data_offset + 8; |
| for (uint32_t i = 0; i < entry_count; ++i) { |
| status_t err = parseChunk(offset, depth + 1); |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| if (*offset != stop_offset) { |
| return ERROR_MALFORMED; |
| } |
| break; |
| } |
| case FOURCC("mett"): |
| { |
| *offset += chunk_size; |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| auto buffer = heapbuffer<uint8_t>(chunk_data_size); |
| if (buffer.get() == NULL) { |
| return NO_MEMORY; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer.get(), chunk_data_size) < chunk_data_size) { |
| return ERROR_IO; |
| } |
| |
| // Prior to API 29, the metadata track was not compliant with ISO/IEC |
| // 14496-12-2015. This led to some ISO-compliant parsers failing to read the |
| // metatrack. As of API 29 and onwards, a change was made to metadata track to |
| // make it compliant with the standard. The workaround is to write the |
| // null-terminated mime_format string twice. This allows compliant parsers to |
| // read the missing reserved, data_reference_index, and content_encoding fields |
| // from the first mime_type string. The actual mime_format field would then be |
| // read correctly from the second string. The non-compliant Android frameworks |
| // from API 28 and earlier would still be able to read the mime_format correctly |
| // as it would only read the first null-terminated mime_format string. To enable |
| // reading metadata tracks generated from both the non-compliant and compliant |
| // formats, a check needs to be done to see which format is used. |
| int null_pos = 0; |
| const unsigned char *str = buffer.get(); |
| while (null_pos < chunk_data_size) { |
| if (*(str + null_pos) == '\0') { |
| break; |
| } |
| ++null_pos; |
| } |
| |
| if (null_pos == chunk_data_size - 1) { |
| // This is not a standard ompliant metadata track. |
| String8 mimeFormat((const char *)(buffer.get()), chunk_data_size); |
| AMediaFormat_setString(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MIME, mimeFormat.string()); |
| } else { |
| // This is a standard compliant metadata track. |
| String8 contentEncoding((const char *)(buffer.get() + 8)); |
| String8 mimeFormat((const char *)(buffer.get() + 8 + contentEncoding.size() + 1), |
| chunk_data_size - 8 - contentEncoding.size() - 1); |
| AMediaFormat_setString(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MIME, mimeFormat.string()); |
| } |
| break; |
| } |
| |
| case FOURCC("mp4a"): |
| case FOURCC("enca"): |
| case FOURCC("samr"): |
| case FOURCC("sawb"): |
| case FOURCC("Opus"): |
| case FOURCC("twos"): |
| case FOURCC("sowt"): |
| case FOURCC("alac"): |
| case FOURCC("fLaC"): |
| case FOURCC(".mp3"): |
| case 0x6D730055: // "ms U" mp3 audio |
| { |
| if (mIsQT && depth >= 1 && mPath[depth - 1] == FOURCC("wave")) { |
| |
| if (chunk_type == FOURCC("alac")) { |
| off64_t offsetTmp = *offset; |
| status_t err = parseALACSampleEntry(&offsetTmp); |
| if (err != OK) { |
| ALOGE("parseALACSampleEntry err:%d Line:%d", err, __LINE__); |
| return err; |
| } |
| } |
| |
| // Ignore all atoms embedded in QT wave atom |
| ALOGV("Ignore all atoms embedded in QT wave atom"); |
| *offset += chunk_size; |
| break; |
| } |
| |
| uint8_t buffer[8 + 20]; |
| if (chunk_data_size < (ssize_t)sizeof(buffer)) { |
| // Basic AudioSampleEntry size. |
| return ERROR_MALFORMED; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer, sizeof(buffer)) < (ssize_t)sizeof(buffer)) { |
| return ERROR_IO; |
| } |
| |
| uint16_t data_ref_index __unused = U16_AT(&buffer[6]); |
| uint16_t version = U16_AT(&buffer[8]); |
| uint32_t num_channels = U16_AT(&buffer[16]); |
| |
| uint16_t sample_size = U16_AT(&buffer[18]); |
| uint32_t sample_rate = U32_AT(&buffer[24]) >> 16; |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| off64_t stop_offset = *offset + chunk_size; |
| *offset = data_offset + sizeof(buffer); |
| |
| if (mIsQT) { |
| if (version == 1) { |
| if (mDataSource->readAt(*offset, buffer, 16) < 16) { |
| return ERROR_IO; |
| } |
| |
| #if 0 |
| U32_AT(buffer); // samples per packet |
| U32_AT(&buffer[4]); // bytes per packet |
| U32_AT(&buffer[8]); // bytes per frame |
| U32_AT(&buffer[12]); // bytes per sample |
| #endif |
| *offset += 16; |
| } else if (version == 2) { |
| uint8_t v2buffer[36]; |
| if (mDataSource->readAt(*offset, v2buffer, 36) < 36) { |
| return ERROR_IO; |
| } |
| |
| #if 0 |
| U32_AT(v2buffer); // size of struct only |
| sample_rate = (uint32_t)U64_AT(&v2buffer[4]); // audio sample rate |
| num_channels = U32_AT(&v2buffer[12]); // num audio channels |
| U32_AT(&v2buffer[16]); // always 0x7f000000 |
| sample_size = (uint16_t)U32_AT(&v2buffer[20]); // const bits per channel |
| U32_AT(&v2buffer[24]); // format specifc flags |
| U32_AT(&v2buffer[28]); // const bytes per audio packet |
| U32_AT(&v2buffer[32]); // const LPCM frames per audio packet |
| #endif |
| *offset += 36; |
| } |
| } |
| |
| if (chunk_type != FOURCC("enca")) { |
| // if the chunk type is enca, we'll get the type from the frma box later |
| AMediaFormat_setString(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MIME, FourCC2MIME(chunk_type)); |
| AdjustChannelsAndRate(chunk_type, &num_channels, &sample_rate); |
| |
| if (!strcasecmp(MEDIA_MIMETYPE_AUDIO_RAW, FourCC2MIME(chunk_type))) { |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_BITS_PER_SAMPLE, sample_size); |
| if (chunk_type == FOURCC("twos")) { |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_PCM_BIG_ENDIAN, 1); |
| } |
| } |
| } |
| ALOGV("*** coding='%s' %d channels, size %d, rate %d\n", |
| chunk, num_channels, sample_size, sample_rate); |
| AMediaFormat_setInt32(mLastTrack->meta, AMEDIAFORMAT_KEY_CHANNEL_COUNT, num_channels); |
| AMediaFormat_setInt32(mLastTrack->meta, AMEDIAFORMAT_KEY_SAMPLE_RATE, sample_rate); |
| |
| if (chunk_type == FOURCC("Opus")) { |
| uint8_t opusInfo[AOPUS_OPUSHEAD_MAXSIZE]; |
| data_offset += sizeof(buffer); |
| size_t opusInfoSize = chunk_data_size - sizeof(buffer); |
| |
| if (opusInfoSize < AOPUS_OPUSHEAD_MINSIZE || |
| opusInfoSize > AOPUS_OPUSHEAD_MAXSIZE) { |
| return ERROR_MALFORMED; |
| } |
| // Read Opus Header |
| if (mDataSource->readAt( |
| data_offset, opusInfo, opusInfoSize) < opusInfoSize) { |
| return ERROR_IO; |
| } |
| |
| // OpusHeader must start with this magic sequence, overwrite first 8 bytes |
| // http://wiki.xiph.org/OggOpus#ID_Header |
| strncpy((char *)opusInfo, "OpusHead", 8); |
| |
| // Version shall be 0 as per mp4 Opus Specific Box |
| // (https://opus-codec.org/docs/opus_in_isobmff.html#4.3.2) |
| if (opusInfo[8]) { |
| return ERROR_MALFORMED; |
| } |
| // Force version to 1 as per OpusHead definition |
| // (http://wiki.xiph.org/OggOpus#ID_Header) |
| opusInfo[8] = 1; |
| |
| // Read Opus Specific Box values |
| size_t opusOffset = 10; |
| uint16_t pre_skip = U16_AT(&opusInfo[opusOffset]); |
| uint32_t sample_rate = U32_AT(&opusInfo[opusOffset + 2]); |
| uint16_t out_gain = U16_AT(&opusInfo[opusOffset + 6]); |
| |
| // Convert Opus Specific Box values. ParseOpusHeader expects |
| // the values in LE, however MP4 stores these values as BE |
| // https://opus-codec.org/docs/opus_in_isobmff.html#4.3.2 |
| memcpy(&opusInfo[opusOffset], &pre_skip, sizeof(pre_skip)); |
| memcpy(&opusInfo[opusOffset + 2], &sample_rate, sizeof(sample_rate)); |
| memcpy(&opusInfo[opusOffset + 6], &out_gain, sizeof(out_gain)); |
| |
| static const int64_t kSeekPreRollNs = 80000000; // Fixed 80 msec |
| static const int32_t kOpusSampleRate = 48000; |
| int64_t codecDelay = pre_skip * 1000000000ll / kOpusSampleRate; |
| |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CSD_0, opusInfo, opusInfoSize); |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CSD_1, &codecDelay, sizeof(codecDelay)); |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CSD_2, &kSeekPreRollNs, sizeof(kSeekPreRollNs)); |
| |
| data_offset += opusInfoSize; |
| *offset = data_offset; |
| CHECK_EQ(*offset, stop_offset); |
| } |
| |
| if (!mIsQT && chunk_type == FOURCC("alac")) { |
| data_offset += sizeof(buffer); |
| |
| status_t err = parseALACSampleEntry(&data_offset); |
| if (err != OK) { |
| ALOGE("parseALACSampleEntry err:%d Line:%d", err, __LINE__); |
| return err; |
| } |
| *offset = data_offset; |
| CHECK_EQ(*offset, stop_offset); |
| } |
| |
| if (chunk_type == FOURCC("fLaC")) { |
| |
| // From https://github.com/xiph/flac/blob/master/doc/isoflac.txt |
| // 4 for mime, 4 for blockType and BlockLen, 34 for metadata |
| uint8_t flacInfo[4 + 4 + 34]; |
| // skipping dFla, version |
| data_offset += sizeof(buffer) + 12; |
| size_t flacOffset = 4; |
| // Add flaC header mime type to CSD |
| strncpy((char *)flacInfo, "fLaC", 4); |
| if (mDataSource->readAt( |
| data_offset, flacInfo + flacOffset, sizeof(flacInfo) - flacOffset) < |
| (ssize_t)sizeof(flacInfo) - flacOffset) { |
| return ERROR_IO; |
| } |
| data_offset += sizeof(flacInfo) - flacOffset; |
| |
| AMediaFormat_setBuffer(mLastTrack->meta, AMEDIAFORMAT_KEY_CSD_0, flacInfo, |
| sizeof(flacInfo)); |
| *offset = data_offset; |
| CHECK_EQ(*offset, stop_offset); |
| } |
| |
| while (*offset < stop_offset) { |
| status_t err = parseChunk(offset, depth + 1); |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| if (*offset != stop_offset) { |
| return ERROR_MALFORMED; |
| } |
| break; |
| } |
| |
| case FOURCC("mp4v"): |
| case FOURCC("encv"): |
| case FOURCC("s263"): |
| case FOURCC("H263"): |
| case FOURCC("h263"): |
| case FOURCC("avc1"): |
| case FOURCC("hvc1"): |
| case FOURCC("hev1"): |
| case FOURCC("av01"): |
| { |
| uint8_t buffer[78]; |
| if (chunk_data_size < (ssize_t)sizeof(buffer)) { |
| // Basic VideoSampleEntry size. |
| return ERROR_MALFORMED; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer, sizeof(buffer)) < (ssize_t)sizeof(buffer)) { |
| return ERROR_IO; |
| } |
| |
| uint16_t data_ref_index __unused = U16_AT(&buffer[6]); |
| uint16_t width = U16_AT(&buffer[6 + 18]); |
| uint16_t height = U16_AT(&buffer[6 + 20]); |
| |
| // The video sample is not standard-compliant if it has invalid dimension. |
| // Use some default width and height value, and |
| // let the decoder figure out the actual width and height (and thus |
| // be prepared for INFO_FOMRAT_CHANGED event). |
| if (width == 0) width = 352; |
| if (height == 0) height = 288; |
| |
| // printf("*** coding='%s' width=%d height=%d\n", |
| // chunk, width, height); |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| if (chunk_type != FOURCC("encv")) { |
| // if the chunk type is encv, we'll get the type from the frma box later |
| AMediaFormat_setString(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MIME, FourCC2MIME(chunk_type)); |
| } |
| AMediaFormat_setInt32(mLastTrack->meta, AMEDIAFORMAT_KEY_WIDTH, width); |
| AMediaFormat_setInt32(mLastTrack->meta, AMEDIAFORMAT_KEY_HEIGHT, height); |
| |
| off64_t stop_offset = *offset + chunk_size; |
| *offset = data_offset + sizeof(buffer); |
| while (*offset < stop_offset) { |
| status_t err = parseChunk(offset, depth + 1); |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| if (*offset != stop_offset) { |
| return ERROR_MALFORMED; |
| } |
| break; |
| } |
| |
| case FOURCC("stco"): |
| case FOURCC("co64"): |
| { |
| if ((mLastTrack == NULL) || (mLastTrack->sampleTable == NULL)) { |
| return ERROR_MALFORMED; |
| } |
| |
| status_t err = |
| mLastTrack->sampleTable->setChunkOffsetParams( |
| chunk_type, data_offset, chunk_data_size); |
| |
| *offset += chunk_size; |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("stsc"): |
| { |
| if ((mLastTrack == NULL) || (mLastTrack->sampleTable == NULL)) |
| return ERROR_MALFORMED; |
| |
| status_t err = |
| mLastTrack->sampleTable->setSampleToChunkParams( |
| data_offset, chunk_data_size); |
| |
| *offset += chunk_size; |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("stsz"): |
| case FOURCC("stz2"): |
| { |
| if ((mLastTrack == NULL) || (mLastTrack->sampleTable == NULL)) { |
| return ERROR_MALFORMED; |
| } |
| |
| status_t err = |
| mLastTrack->sampleTable->setSampleSizeParams( |
| chunk_type, data_offset, chunk_data_size); |
| |
| *offset += chunk_size; |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| adjustRawDefaultFrameSize(); |
| |
| size_t max_size; |
| err = mLastTrack->sampleTable->getMaxSampleSize(&max_size); |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| if (max_size != 0) { |
| // Assume that a given buffer only contains at most 10 chunks, |
| // each chunk originally prefixed with a 2 byte length will |
| // have a 4 byte header (0x00 0x00 0x00 0x01) after conversion, |
| // and thus will grow by 2 bytes per chunk. |
| if (max_size > SIZE_MAX - 10 * 2) { |
| ALOGE("max sample size too big: %zu", max_size); |
| return ERROR_MALFORMED; |
| } |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MAX_INPUT_SIZE, max_size + 10 * 2); |
| } else { |
| // No size was specified. Pick a conservatively large size. |
| uint32_t width, height; |
| if (!AMediaFormat_getInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_WIDTH, (int32_t*)&width) || |
| !AMediaFormat_getInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_HEIGHT,(int32_t*) &height)) { |
| ALOGE("No width or height, assuming worst case 1080p"); |
| width = 1920; |
| height = 1080; |
| } else { |
| // A resolution was specified, check that it's not too big. The values below |
| // were chosen so that the calculations below don't cause overflows, they're |
| // not indicating that resolutions up to 32kx32k are actually supported. |
| if (width > 32768 || height > 32768) { |
| ALOGE("can't support %u x %u video", width, height); |
| return ERROR_MALFORMED; |
| } |
| } |
| |
| const char *mime; |
| CHECK(AMediaFormat_getString(mLastTrack->meta, AMEDIAFORMAT_KEY_MIME, &mime)); |
| if (!strncmp(mime, "audio/", 6)) { |
| // for audio, use 128KB |
| max_size = 1024 * 128; |
| } else if (!strcmp(mime, MEDIA_MIMETYPE_VIDEO_AVC) |
| || !strcmp(mime, MEDIA_MIMETYPE_VIDEO_HEVC)) { |
| // AVC & HEVC requires compression ratio of at least 2, and uses |
| // macroblocks |
| max_size = ((width + 15) / 16) * ((height + 15) / 16) * 192; |
| } else { |
| // For all other formats there is no minimum compression |
| // ratio. Use compression ratio of 1. |
| max_size = width * height * 3 / 2; |
| } |
| // HACK: allow 10% overhead |
| // TODO: read sample size from traf atom for fragmented MPEG4. |
| max_size += max_size / 10; |
| AMediaFormat_setInt32(mLastTrack->meta, AMEDIAFORMAT_KEY_MAX_INPUT_SIZE, max_size); |
| } |
| |
| // NOTE: setting another piece of metadata invalidates any pointers (such as the |
| // mimetype) previously obtained, so don't cache them. |
| const char *mime; |
| CHECK(AMediaFormat_getString(mLastTrack->meta, AMEDIAFORMAT_KEY_MIME, &mime)); |
| // Calculate average frame rate. |
| if (!strncasecmp("video/", mime, 6)) { |
| size_t nSamples = mLastTrack->sampleTable->countSamples(); |
| if (nSamples == 0) { |
| int32_t trackId; |
| if (AMediaFormat_getInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_TRACK_ID, &trackId)) { |
| for (size_t i = 0; i < mTrex.size(); i++) { |
| Trex *t = &mTrex.editItemAt(i); |
| if (t->track_ID == (uint32_t) trackId) { |
| if (t->default_sample_duration > 0) { |
| int32_t frameRate = |
| mLastTrack->timescale / t->default_sample_duration; |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_FRAME_RATE, frameRate); |
| } |
| break; |
| } |
| } |
| } |
| } else { |
| int64_t durationUs; |
| if (AMediaFormat_getInt64(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_DURATION, &durationUs)) { |
| if (durationUs > 0) { |
| int32_t frameRate = (nSamples * 1000000LL + |
| (durationUs >> 1)) / durationUs; |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_FRAME_RATE, frameRate); |
| } |
| } |
| ALOGV("setting frame count %zu", nSamples); |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_FRAME_COUNT, nSamples); |
| } |
| } |
| |
| break; |
| } |
| |
| case FOURCC("stts"): |
| { |
| if ((mLastTrack == NULL) || (mLastTrack->sampleTable == NULL)) |
| return ERROR_MALFORMED; |
| |
| *offset += chunk_size; |
| |
| if (depth >= 1 && mPath[depth - 1] != FOURCC("stbl")) { |
| char chunk[5]; |
| MakeFourCCString(mPath[depth - 1], chunk); |
| ALOGW("stts's parent box (%s) is not stbl, skip it.", chunk); |
| break; |
| } |
| |
| status_t err = |
| mLastTrack->sampleTable->setTimeToSampleParams( |
| data_offset, chunk_data_size); |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("ctts"): |
| { |
| if ((mLastTrack == NULL) || (mLastTrack->sampleTable == NULL)) |
| return ERROR_MALFORMED; |
| |
| *offset += chunk_size; |
| |
| status_t err = |
| mLastTrack->sampleTable->setCompositionTimeToSampleParams( |
| data_offset, chunk_data_size); |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("stss"): |
| { |
| if ((mLastTrack == NULL) || (mLastTrack->sampleTable == NULL)) |
| return ERROR_MALFORMED; |
| |
| *offset += chunk_size; |
| |
| status_t err = |
| mLastTrack->sampleTable->setSyncSampleParams( |
| data_offset, chunk_data_size); |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| break; |
| } |
| |
| // \xA9xyz |
| case FOURCC("\251xyz"): |
| { |
| *offset += chunk_size; |
| |
| // Best case the total data length inside "\xA9xyz" box would |
| // be 9, for instance "\xA9xyz" + "\x00\x05\x15\xc7" + "+0+0/", |
| // where "\x00\x05" is the text string length with value = 5, |
| // "\0x15\xc7" is the language code = en, and "+0+0/" is a |
| // location (string) value with longitude = 0 and latitude = 0. |
| // Since some devices encountered in the wild omit the trailing |
| // slash, we'll allow that. |
| if (chunk_data_size < 8) { // 8 instead of 9 to allow for missing / |
| return ERROR_MALFORMED; |
| } |
| |
| uint16_t len; |
| if (!mDataSource->getUInt16(data_offset, &len)) { |
| return ERROR_IO; |
| } |
| |
| // allow "+0+0" without trailing slash |
| if (len < 4 || len > chunk_data_size - 4) { |
| return ERROR_MALFORMED; |
| } |
| // The location string following the language code is formatted |
| // according to ISO 6709:2008 (https://en.wikipedia.org/wiki/ISO_6709). |
| // Allocate 2 extra bytes, in case we need to add a trailing slash, |
| // and to add a terminating 0. |
| std::unique_ptr<char[]> buffer(new (std::nothrow) char[len+2]()); |
| if (!buffer) { |
| return NO_MEMORY; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset + 4, &buffer[0], len) < len) { |
| return ERROR_IO; |
| } |
| |
| len = strlen(&buffer[0]); |
| if (len < 4) { |
| return ERROR_MALFORMED; |
| } |
| // Add a trailing slash if there wasn't one. |
| if (buffer[len - 1] != '/') { |
| buffer[len] = '/'; |
| } |
| AMediaFormat_setString(mFileMetaData, AMEDIAFORMAT_KEY_LOCATION, &buffer[0]); |
| break; |
| } |
| |
| case FOURCC("esds"): |
| { |
| *offset += chunk_size; |
| |
| if (chunk_data_size < 4) { |
| return ERROR_MALFORMED; |
| } |
| |
| auto tmp = heapbuffer<uint8_t>(chunk_data_size); |
| uint8_t *buffer = tmp.get(); |
| if (buffer == NULL) { |
| return -ENOMEM; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer, chunk_data_size) < chunk_data_size) { |
| return ERROR_IO; |
| } |
| |
| if (U32_AT(buffer) != 0) { |
| // Should be version 0, flags 0. |
| return ERROR_MALFORMED; |
| } |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_ESDS, &buffer[4], chunk_data_size - 4); |
| |
| if (mPath.size() >= 2 |
| && mPath[mPath.size() - 2] == FOURCC("mp4a")) { |
| // Information from the ESDS must be relied on for proper |
| // setup of sample rate and channel count for MPEG4 Audio. |
| // The generic header appears to only contain generic |
| // information... |
| |
| status_t err = updateAudioTrackInfoFromESDS_MPEG4Audio( |
| &buffer[4], chunk_data_size - 4); |
| |
| if (err != OK) { |
| return err; |
| } |
| } |
| if (mPath.size() >= 2 |
| && mPath[mPath.size() - 2] == FOURCC("mp4v")) { |
| // Check if the video is MPEG2 |
| ESDS esds(&buffer[4], chunk_data_size - 4); |
| |
| uint8_t objectTypeIndication; |
| if (esds.getObjectTypeIndication(&objectTypeIndication) == OK) { |
| if (objectTypeIndication >= 0x60 && objectTypeIndication <= 0x65) { |
| AMediaFormat_setString(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MIME, MEDIA_MIMETYPE_VIDEO_MPEG2); |
| } |
| } |
| } |
| break; |
| } |
| |
| case FOURCC("btrt"): |
| { |
| *offset += chunk_size; |
| if (mLastTrack == NULL) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint8_t buffer[12]; |
| if (chunk_data_size != sizeof(buffer)) { |
| return ERROR_MALFORMED; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer, chunk_data_size) < chunk_data_size) { |
| return ERROR_IO; |
| } |
| |
| uint32_t maxBitrate = U32_AT(&buffer[4]); |
| uint32_t avgBitrate = U32_AT(&buffer[8]); |
| if (maxBitrate > 0 && maxBitrate < INT32_MAX) { |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MAX_BIT_RATE, (int32_t)maxBitrate); |
| } |
| if (avgBitrate > 0 && avgBitrate < INT32_MAX) { |
| AMediaFormat_setInt32(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_BIT_RATE, (int32_t)avgBitrate); |
| } |
| break; |
| } |
| |
| case FOURCC("avcC"): |
| { |
| *offset += chunk_size; |
| |
| auto buffer = heapbuffer<uint8_t>(chunk_data_size); |
| |
| if (buffer.get() == NULL) { |
| ALOGE("b/28471206"); |
| return NO_MEMORY; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer.get(), chunk_data_size) < chunk_data_size) { |
| return ERROR_IO; |
| } |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CSD_AVC, buffer.get(), chunk_data_size); |
| |
| break; |
| } |
| case FOURCC("hvcC"): |
| { |
| auto buffer = heapbuffer<uint8_t>(chunk_data_size); |
| |
| if (buffer.get() == NULL) { |
| ALOGE("b/28471206"); |
| return NO_MEMORY; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer.get(), chunk_data_size) < chunk_data_size) { |
| return ERROR_IO; |
| } |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CSD_HEVC, buffer.get(), chunk_data_size); |
| |
| *offset += chunk_size; |
| break; |
| } |
| case FOURCC("av1C"): |
| { |
| auto buffer = heapbuffer<uint8_t>(chunk_data_size); |
| |
| if (buffer.get() == NULL) { |
| ALOGE("b/28471206"); |
| return NO_MEMORY; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer.get(), chunk_data_size) < chunk_data_size) { |
| return ERROR_IO; |
| } |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_CSD_0, buffer.get(), chunk_data_size); |
| |
| *offset += chunk_size; |
| break; |
| } |
| case FOURCC("d263"): |
| { |
| *offset += chunk_size; |
| /* |
| * d263 contains a fixed 7 bytes part: |
| * vendor - 4 bytes |
| * version - 1 byte |
| * level - 1 byte |
| * profile - 1 byte |
| * optionally, "d263" box itself may contain a 16-byte |
| * bit rate box (bitr) |
| * average bit rate - 4 bytes |
| * max bit rate - 4 bytes |
| */ |
| char buffer[23]; |
| if (chunk_data_size != 7 && |
| chunk_data_size != 23) { |
| ALOGE("Incorrect D263 box size %lld", (long long)chunk_data_size); |
| return ERROR_MALFORMED; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer, chunk_data_size) < chunk_data_size) { |
| return ERROR_IO; |
| } |
| |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| AMediaFormat_setBuffer(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_D263, buffer, chunk_data_size); |
| |
| break; |
| } |
| |
| case FOURCC("meta"): |
| { |
| off64_t stop_offset = *offset + chunk_size; |
| *offset = data_offset; |
| bool isParsingMetaKeys = underQTMetaPath(mPath, 2); |
| if (!isParsingMetaKeys) { |
| uint8_t buffer[4]; |
| if (chunk_data_size < (off64_t)sizeof(buffer)) { |
| *offset = stop_offset; |
| return ERROR_MALFORMED; |
| } |
| |
| if (mDataSource->readAt( |
| data_offset, buffer, 4) < 4) { |
| *offset = stop_offset; |
| return ERROR_IO; |
| } |
| |
| if (U32_AT(buffer) != 0) { |
| // Should be version 0, flags 0. |
| |
| // If it's not, let's assume this is one of those |
| // apparently malformed chunks that don't have flags |
| // and completely different semantics than what's |
| // in the MPEG4 specs and skip it. |
| *offset = stop_offset; |
| return OK; |
| } |
| *offset += sizeof(buffer); |
| } |
| |
| while (*offset < stop_offset) { |
| status_t err = parseChunk(offset, depth + 1); |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| if (*offset != stop_offset) { |
| return ERROR_MALFORMED; |
| } |
| break; |
| } |
| |
| case FOURCC("iloc"): |
| case FOURCC("iinf"): |
| case FOURCC("iprp"): |
| case FOURCC("pitm"): |
| case FOURCC("idat"): |
| case FOURCC("iref"): |
| case FOURCC("ipro"): |
| { |
| if (mIsHeif) { |
| if (mItemTable == NULL) { |
| mItemTable = new ItemTable(mDataSource); |
| } |
| status_t err = mItemTable->parse( |
| chunk_type, data_offset, chunk_data_size); |
| if (err != OK) { |
| return err; |
| } |
| } |
| *offset += chunk_size; |
| break; |
| } |
| |
| case FOURCC("mean"): |
| case FOURCC("name"): |
| case FOURCC("data"): |
| { |
| *offset += chunk_size; |
| |
| if (mPath.size() == 6 && underMetaDataPath(mPath)) { |
| status_t err = parseITunesMetaData(data_offset, chunk_data_size); |
| |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| break; |
| } |
| |
| case FOURCC("mvhd"): |
| { |
| *offset += chunk_size; |
| |
| if (depth != 1) { |
| ALOGE("mvhd: depth %d", depth); |
| return ERROR_MALFORMED; |
| } |
| if (chunk_data_size < 32) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint8_t header[32]; |
| if (mDataSource->readAt( |
| data_offset, header, sizeof(header)) |
| < (ssize_t)sizeof(header)) { |
| return ERROR_IO; |
| } |
| |
| uint64_t creationTime; |
| uint64_t duration = 0; |
| if (header[0] == 1) { |
| creationTime = U64_AT(&header[4]); |
| mHeaderTimescale = U32_AT(&header[20]); |
| duration = U64_AT(&header[24]); |
| if (duration == 0xffffffffffffffff) { |
| duration = 0; |
| } |
| } else if (header[0] != 0) { |
| return ERROR_MALFORMED; |
| } else { |
| creationTime = U32_AT(&header[4]); |
| mHeaderTimescale = U32_AT(&header[12]); |
| uint32_t d32 = U32_AT(&header[16]); |
| if (d32 == 0xffffffff) { |
| d32 = 0; |
| } |
| duration = d32; |
| } |
| if (duration != 0 && mHeaderTimescale != 0 && duration < UINT64_MAX / 1000000) { |
| AMediaFormat_setInt64(mFileMetaData, |
| AMEDIAFORMAT_KEY_DURATION, duration * 1000000 / mHeaderTimescale); |
| } |
| |
| String8 s; |
| if (convertTimeToDate(creationTime, &s)) { |
| AMediaFormat_setString(mFileMetaData, AMEDIAFORMAT_KEY_DATE, s.string()); |
| } |
| |
| break; |
| } |
| |
| case FOURCC("mehd"): |
| { |
| *offset += chunk_size; |
| |
| if (chunk_data_size < 8) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint8_t flags[4]; |
| if (mDataSource->readAt( |
| data_offset, flags, sizeof(flags)) |
| < (ssize_t)sizeof(flags)) { |
| return ERROR_IO; |
| } |
| |
| uint64_t duration = 0; |
| if (flags[0] == 1) { |
| // 64 bit |
| if (chunk_data_size < 12) { |
| return ERROR_MALFORMED; |
| } |
| mDataSource->getUInt64(data_offset + 4, &duration); |
| if (duration == 0xffffffffffffffff) { |
| duration = 0; |
| } |
| } else if (flags[0] == 0) { |
| // 32 bit |
| uint32_t d32; |
| mDataSource->getUInt32(data_offset + 4, &d32); |
| if (d32 == 0xffffffff) { |
| d32 = 0; |
| } |
| duration = d32; |
| } else { |
| return ERROR_MALFORMED; |
| } |
| |
| if (duration != 0 && mHeaderTimescale != 0) { |
| AMediaFormat_setInt64(mFileMetaData, |
| AMEDIAFORMAT_KEY_DURATION, duration * 1000000 / mHeaderTimescale); |
| } |
| |
| break; |
| } |
| |
| case FOURCC("mdat"): |
| { |
| mMdatFound = true; |
| |
| *offset += chunk_size; |
| break; |
| } |
| |
| case FOURCC("hdlr"): |
| { |
| *offset += chunk_size; |
| |
| if (underQTMetaPath(mPath, 3)) { |
| break; |
| } |
| |
| uint32_t buffer; |
| if (mDataSource->readAt( |
| data_offset + 8, &buffer, 4) < 4) { |
| return ERROR_IO; |
| } |
| |
| uint32_t type = ntohl(buffer); |
| // For the 3GPP file format, the handler-type within the 'hdlr' box |
| // shall be 'text'. We also want to support 'sbtl' handler type |
| // for a practical reason as various MPEG4 containers use it. |
| if (type == FOURCC("text") || type == FOURCC("sbtl")) { |
| if (mLastTrack != NULL) { |
| AMediaFormat_setString(mLastTrack->meta, |
| AMEDIAFORMAT_KEY_MIME, MEDIA_MIMETYPE_TEXT_3GPP); |
| } |
| } |
| |
| break; |
| } |
| |
| case FOURCC("keys"): |
| { |
| *offset += chunk_size; |
| |
| if (underQTMetaPath(mPath, 3)) { |
| status_t err = parseQTMetaKey(data_offset, chunk_data_size); |
| if (err != OK) { |
| return err; |
| } |
| } |
| break; |
| } |
| |
| case FOURCC("trex"): |
| { |
| *offset += chunk_size; |
| |
| if (chunk_data_size < 24) { |
| return ERROR_IO; |
| } |
| Trex trex; |
| if (!mDataSource->getUInt32(data_offset + 4, &trex.track_ID) || |
| !mDataSource->getUInt32(data_offset + 8, &trex.default_sample_description_index) || |
| !mDataSource->getUInt32(data_offset + 12, &trex.default_sample_duration) || |
| !mDataSource->getUInt32(data_offset + 16, &trex.default_sample_size) || |
| !mDataSource->getUInt32(data_offset + 20, &trex.default_sample_flags)) { |
| return ERROR_IO; |
| } |
| mTrex.add(trex); |
| break; |
| } |
| |
| case FOURCC("tx3g"): |
| { |
| if (mLastTrack == NULL) |
| return ERROR_MALFORMED; |
| |
| // complain about ridiculous chunks |
| if (chunk_size > kMaxAtomSize) { |
| return ERROR_MALFORMED; |
| } |
| |
| // complain about empty atoms |
| if (chunk_data_size <= 0) { |
| ALOGE("b/124330204"); |
| android_errorWriteLog(0x534e4554, "124330204"); |
| return ERROR_MALFORMED; |
| } |
| |
| // should fill buffer based on "data_offset" and "chunk_data_size" |
| // instead of *offset and chunk_size; |
| // but we've been feeding the extra data to consumers for multiple releases and |
| // if those apps are compensating for it, we'd break them with such a change |
| // |
| |
| if (mLastTrack->mTx3gSize - mLastTrack->mTx3gFilled < chunk_size) { |
| size_t growth = kTx3gGrowth; |
| if (growth < chunk_size) { |
| growth = chunk_size; |
| } |
| // although this disallows 2 tx3g atoms of nearly kMaxAtomSize... |
| if ((uint64_t) mLastTrack->mTx3gSize + growth > kMaxAtomSize) { |
| ALOGE("b/124330204 - too much space"); |
| android_errorWriteLog(0x534e4554, "124330204"); |
| return ERROR_MALFORMED; |
| } |
| uint8_t *updated = (uint8_t *)realloc(mLastTrack->mTx3gBuffer, |
| mLastTrack->mTx3gSize + growth); |
| if (updated == NULL) { |
| return ERROR_MALFORMED; |
| } |
| mLastTrack->mTx3gBuffer = updated; |
| mLastTrack->mTx3gSize += growth; |
| } |
| |
| if ((size_t)(mDataSource->readAt(*offset, |
| mLastTrack->mTx3gBuffer + mLastTrack->mTx3gFilled, |
| chunk_size)) |
| < chunk_size) { |
| |
| // advance read pointer so we don't end up reading this again |
| *offset += chunk_size; |
| return ERROR_IO; |
| } |
| |
| mLastTrack->mTx3gFilled += chunk_size; |
| *offset += chunk_size; |
| break; |
| } |
| |
| case FOURCC("covr"): |
| { |
| *offset += chunk_size; |
| |
| ALOGV("chunk_data_size = %" PRId64 " and data_offset = %" PRId64, |
| chunk_data_size, data_offset); |
| |
| if (chunk_data_size < 0 || static_cast<uint64_t>(chunk_data_size) >= SIZE_MAX - 1) { |
| return ERROR_MALFORMED; |
| } |
| auto buffer = heapbuffer<uint8_t>(chunk_data_size); |
| if (buffer.get() == NULL) { |
| ALOGE("b/28471206"); |
| return NO_MEMORY; |
| } |
| if (mDataSource->readAt( |
| data_offset, buffer.get(), chunk_data_size) != (ssize_t)chunk_data_size) { |
| return ERROR_IO; |
| } |
| const int kSkipBytesOfDataBox = 16; |
| if (chunk_data_size <= kSkipBytesOfDataBox) { |
| return ERROR_MALFORMED; |
| } |
| |
| AMediaFormat_setBuffer(mFileMetaData, |
| AMEDIAFORMAT_KEY_ALBUMART, |
| buffer.get() + kSkipBytesOfDataBox, chunk_data_size - kSkipBytesOfDataBox); |
| |
| break; |
| } |
| |
| case FOURCC("colr"): |
| { |
| *offset += chunk_size; |
| // this must be in a VisualSampleEntry box under the Sample Description Box ('stsd') |
| // ignore otherwise |
| if (depth >= 2 && mPath[depth - 2] == FOURCC("stsd")) { |
| status_t err = parseColorInfo(data_offset, chunk_data_size); |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| break; |
| } |
| |
| case FOURCC("titl"): |
| case FOURCC("perf"): |
| case FOURCC("auth"): |
| case FOURCC("gnre"): |
| case FOURCC("albm"): |
| case FOURCC("yrrc"): |
| { |
| *offset += chunk_size; |
| |
| status_t err = parse3GPPMetaData(data_offset, chunk_data_size, depth); |
| |
| if (err != OK) { |
| return err; |
| } |
| |
| break; |
| } |
| |
| case FOURCC("ID32"): |
| { |
| *offset += chunk_size; |
| |
| if (chunk_data_size < 6) { |
| return ERROR_MALFORMED; |
| } |
| |
| parseID3v2MetaData(data_offset + 6); |
| |
| break; |
| } |
| |
| case FOURCC("----"): |
| { |
| mLastCommentMean.clear(); |
| mLastCommentName.clear(); |
| mLastCommentData.clear(); |
| *offset += chunk_size; |
| break; |
| } |
| |
| case FOURCC("sidx"): |
| { |
| status_t err = parseSegmentIndex(data_offset, chunk_data_size); |
| if (err != OK) { |
| return err; |
| } |
| *offset += chunk_size; |
| return UNKNOWN_ERROR; // stop parsing after sidx |
| } |
| |
| case FOURCC("ac-3"): |
| { |
| *offset += chunk_size; |
| // bypass ac-3 if parse fail |
| if (parseAC3SpecificBox(data_offset) != OK) { |
| if (mLastTrack != NULL) { |
| ALOGW("Fail to parse ac-3"); |
| mLastTrack->skipTrack = true; |
| } |
| } |
| return OK; |
| } |
| |
| case FOURCC("ec-3"): |
| { |
| *offset += chunk_size; |
| // bypass ec-3 if parse fail |
| if (parseEAC3SpecificBox(data_offset) != OK) { |
| if (mLastTrack != NULL) { |
| ALOGW("Fail to parse ec-3"); |
| mLastTrack->skipTrack = true; |
| } |
| } |
| return OK; |
| } |
| |
| case FOURCC("ac-4"): |
| { |
| *offset += chunk_size; |
| // bypass ac-4 if parse fail |
| if (parseAC4SpecificBox(data_offset) != OK) { |
| if (mLastTrack != NULL) { |
| ALOGW("Fail to parse ac-4"); |
| mLastTrack->skipTrack = true; |
| } |
| } |
| return OK; |
| } |
| |
| case FOURCC("ftyp"): |
| { |
| if (chunk_data_size < 8 || depth != 0) { |
| return ERROR_MALFORMED; |
| } |
| |
| off64_t stop_offset = *offset + chunk_size; |
| uint32_t numCompatibleBrands = (chunk_data_size - 8) / 4; |
| std::set<uint32_t> brandSet; |
| for (size_t i = 0; i < numCompatibleBrands + 2; ++i) { |
| if (i == 1) { |
| // Skip this index, it refers to the minorVersion, |
| // not a brand. |
| continue; |
| } |
| |
| uint32_t brand; |
| if (mDataSource->readAt(data_offset + 4 * i, &brand, 4) < 4) { |
| return ERROR_MALFORMED; |
| } |
| |
| brand = ntohl(brand); |
| brandSet.insert(brand); |
| } |
| |
| if (brandSet.count(FOURCC("qt ")) > 0) { |
| mIsQT = true; |
| } else { |
| if (brandSet.count(FOURCC("mif1")) > 0 |
| && brandSet.count(FOURCC("heic")) > 0) { |
| ALOGV("identified HEIF image"); |
| |
| mIsHeif = true; |
| brandSet.erase(FOURCC("mif1")); |
| brandSet.erase(FOURCC("heic")); |
| } |
| |
| if (!brandSet.empty()) { |
| // This means that the file should have moov box. |
| // It could be any iso files (mp4, heifs, etc.) |
| mHasMoovBox = true; |
| if (mIsHeif) { |
| ALOGV("identified HEIF image with other tracks"); |
| } |
| } |
| } |
| |
| *offset = stop_offset; |
| |
| break; |
| } |
| |
| default: |
| { |
| // check if we're parsing 'ilst' for meta keys |
| // if so, treat type as a number (key-id). |
| if (underQTMetaPath(mPath, 3)) { |
| status_t err = parseQTMetaVal(chunk_type, data_offset, chunk_data_size); |
| if (err != OK) { |
| return err; |
| } |
| } |
| |
| *offset += chunk_size; |
| break; |
| } |
| } |
| |
| return OK; |
| } |
| |
| status_t MPEG4Extractor::parseChannelCountSampleRate( |
| off64_t *offset, uint16_t *channelCount, uint16_t *sampleRate) { |
| // skip 16 bytes: |
| // + 6-byte reserved, |
| // + 2-byte data reference index, |
| // + 8-byte reserved |
| *offset += 16; |
| if (!mDataSource->getUInt16(*offset, channelCount)) { |
| ALOGE("MPEG4Extractor: error while reading sample entry box: cannot read channel count"); |
| return ERROR_MALFORMED; |
| } |
| // skip 8 bytes: |
| // + 2-byte channelCount, |
| // + 2-byte sample size, |
| // + 4-byte reserved |
| *offset += 8; |
| if (!mDataSource->getUInt16(*offset, sampleRate)) { |
| ALOGE("MPEG4Extractor: error while reading sample entry box: cannot read sample rate"); |
| return ERROR_MALFORMED; |
| } |
| // skip 4 bytes: |
| // + 2-byte sampleRate, |
| // + 2-byte reserved |
| *offset += 4; |
| return OK; |
| } |
| |
| status_t MPEG4Extractor::parseAC4SpecificBox(off64_t offset) { |
| if (mLastTrack == NULL) { |
| return ERROR_MALFORMED; |
| } |
| |
| uint16_t sampleRate, channelCount; |
| status_t status; |
| if ((status = parseChannelCountSampleRate(&offset, &channelCount, &sampleRate)) != OK) { |
| return status; |
| } |
| uint32_t size; |
| // + 4-byte size |
| // + 4-byte type |
| // + 3-byte payload |
| const uint32_t kAC4MinimumBoxSize = 4 + 4 + 3; |
| if (!mDataSource->getUInt32(offset, &size) || size < kAC4MinimumBoxSize) { |
| ALOGE("MPEG4Extractor: error while reading ac-4 block: cannot read specific box size"); |
|