gecko-dev/dom/media/MediaDecoderStateMachine.cpp

3602 строки
124 KiB
C++

/* vim:set ts=2 sw=2 sts=2 et cindent: */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#ifdef XP_WIN
// Include Windows headers required for enabling high precision timers.
#include "windows.h"
#include "mmsystem.h"
#endif
#include "mozilla/DebugOnly.h"
#include <stdint.h>
#include "MediaDecoderStateMachine.h"
#include "MediaDecoderStateMachineScheduler.h"
#include "AudioSink.h"
#include "nsTArray.h"
#include "MediaDecoder.h"
#include "MediaDecoderReader.h"
#include "mozilla/MathAlgorithms.h"
#include "mozilla/mozalloc.h"
#include "VideoUtils.h"
#include "mozilla/dom/TimeRanges.h"
#include "nsDeque.h"
#include "AudioSegment.h"
#include "VideoSegment.h"
#include "ImageContainer.h"
#include "nsComponentManagerUtils.h"
#include "nsITimer.h"
#include "nsContentUtils.h"
#include "MediaShutdownManager.h"
#include "SharedThreadPool.h"
#include "MediaTaskQueue.h"
#include "nsIEventTarget.h"
#include "prenv.h"
#include "mozilla/Preferences.h"
#include "gfx2DGlue.h"
#include "nsPrintfCString.h"
#include "DOMMediaStream.h"
#include <algorithm>
namespace mozilla {
using namespace mozilla::layers;
using namespace mozilla::dom;
using namespace mozilla::gfx;
// avoid redefined macro in unified build
#undef DECODER_LOG
#undef VERBOSE_LOG
#ifdef PR_LOGGING
extern PRLogModuleInfo* gMediaDecoderLog;
#define DECODER_LOG(x, ...) \
PR_LOG(gMediaDecoderLog, PR_LOG_DEBUG, ("Decoder=%p " x, mDecoder.get(), ##__VA_ARGS__))
#define VERBOSE_LOG(x, ...) \
PR_BEGIN_MACRO \
if (!PR_GetEnv("MOZ_QUIET")) { \
DECODER_LOG(x, ##__VA_ARGS__); \
} \
PR_END_MACRO
#define SAMPLE_LOG(x, ...) \
PR_BEGIN_MACRO \
if (PR_GetEnv("MEDIA_LOG_SAMPLES")) { \
DECODER_LOG(x, ##__VA_ARGS__); \
} \
PR_END_MACRO
#else
#define DECODER_LOG(x, ...)
#define VERBOSE_LOG(x, ...)
#define SAMPLE_LOG(x, ...)
#endif
// Somehow MSVC doesn't correctly delete the comma before ##__VA_ARGS__
// when __VA_ARGS__ expands to nothing. This is a workaround for it.
#define DECODER_WARN_HELPER(a, b) NS_WARNING b
#define DECODER_WARN(x, ...) \
DECODER_WARN_HELPER(0, (nsPrintfCString("Decoder=%p " x, mDecoder.get(), ##__VA_ARGS__).get()))
// GetCurrentTime is defined in winbase.h as zero argument macro forwarding to
// GetTickCount() and conflicts with MediaDecoderStateMachine::GetCurrentTime
// implementation. With unified builds, putting this in headers is not enough.
#ifdef GetCurrentTime
#undef GetCurrentTime
#endif
// Certain constants get stored as member variables and then adjusted by various
// scale factors on a per-decoder basis. We want to make sure to avoid using these
// constants directly, so we put them in a namespace.
namespace detail {
// If audio queue has less than this many usecs of decoded audio, we won't risk
// trying to decode the video, we'll skip decoding video up to the next
// keyframe. We may increase this value for an individual decoder if we
// encounter video frames which take a long time to decode.
static const uint32_t LOW_AUDIO_USECS = 300000;
// If more than this many usecs of decoded audio is queued, we'll hold off
// decoding more audio. If we increase the low audio threshold (see
// LOW_AUDIO_USECS above) we'll also increase this value to ensure it's not
// less than the low audio threshold.
const int64_t AMPLE_AUDIO_USECS = 1000000;
} // namespace detail
// When we're only playing audio and we don't have a video stream, we divide
// AMPLE_AUDIO_USECS and LOW_AUDIO_USECS by the following value. This reduces
// the amount of decoded audio we buffer, reducing our memory usage. We only
// need to decode far ahead when we're decoding video using software decoding,
// as otherwise a long video decode could cause an audio underrun.
const int64_t NO_VIDEO_AMPLE_AUDIO_DIVISOR = 8;
// If we have fewer than LOW_VIDEO_FRAMES decoded frames, and
// we're not "prerolling video", we'll skip the video up to the next keyframe
// which is at or after the current playback position.
static const uint32_t LOW_VIDEO_FRAMES = 1;
// Threshold in usecs that used to check if we are low on decoded video.
// If the last video frame's end time |mDecodedVideoEndTime| doesn't exceed
// |clock time + LOW_VIDEO_THRESHOLD_USECS*mPlaybackRate| calculation in
// Advanceframe(), we are low on decoded video frames and trying to skip to next
// keyframe.
static const int32_t LOW_VIDEO_THRESHOLD_USECS = 16000;
// Arbitrary "frame duration" when playing only audio.
static const int AUDIO_DURATION_USECS = 40000;
// If we increase our "low audio threshold" (see LOW_AUDIO_USECS above), we
// use this as a factor in all our calculations. Increasing this will cause
// us to be more likely to increase our low audio threshold, and to
// increase it by more.
static const int THRESHOLD_FACTOR = 2;
namespace detail {
// If we have less than this much undecoded data available, we'll consider
// ourselves to be running low on undecoded data. We determine how much
// undecoded data we have remaining using the reader's GetBuffered()
// implementation.
static const int64_t LOW_DATA_THRESHOLD_USECS = 5000000;
// LOW_DATA_THRESHOLD_USECS needs to be greater than AMPLE_AUDIO_USECS, otherwise
// the skip-to-keyframe logic can activate when we're running low on data.
static_assert(LOW_DATA_THRESHOLD_USECS > AMPLE_AUDIO_USECS,
"LOW_DATA_THRESHOLD_USECS is too small");
} // namespace detail
// Amount of excess usecs of data to add in to the "should we buffer" calculation.
static const uint32_t EXHAUSTED_DATA_MARGIN_USECS = 60000;
// If we enter buffering within QUICK_BUFFER_THRESHOLD_USECS seconds of starting
// decoding, we'll enter "quick buffering" mode, which exits a lot sooner than
// normal buffering mode. This exists so that if the decode-ahead exhausts the
// downloaded data while decode/playback is just starting up (for example
// after a seek while the media is still playing, or when playing a media
// as soon as it's load started), we won't necessarily stop for 30s and wait
// for buffering. We may actually be able to playback in this case, so exit
// buffering early and try to play. If it turns out we can't play, we'll fall
// back to buffering normally.
static const uint32_t QUICK_BUFFER_THRESHOLD_USECS = 2000000;
namespace detail {
// If we're quick buffering, we'll remain in buffering mode while we have less than
// QUICK_BUFFERING_LOW_DATA_USECS of decoded data available.
static const uint32_t QUICK_BUFFERING_LOW_DATA_USECS = 1000000;
// If QUICK_BUFFERING_LOW_DATA_USECS is > AMPLE_AUDIO_USECS, we won't exit
// quick buffering in a timely fashion, as the decode pauses when it
// reaches AMPLE_AUDIO_USECS decoded data, and thus we'll never reach
// QUICK_BUFFERING_LOW_DATA_USECS.
static_assert(QUICK_BUFFERING_LOW_DATA_USECS <= AMPLE_AUDIO_USECS,
"QUICK_BUFFERING_LOW_DATA_USECS is too large");
} // namespace detail
// The amount of instability we tollerate in calls to
// MediaDecoderStateMachine::UpdateEstimatedDuration(); changes of duration
// less than this are ignored, as they're assumed to be the result of
// instability in the duration estimation.
static const uint64_t ESTIMATED_DURATION_FUZZ_FACTOR_USECS = USECS_PER_S / 2;
static TimeDuration UsecsToDuration(int64_t aUsecs) {
return TimeDuration::FromMicroseconds(aUsecs);
}
static int64_t DurationToUsecs(TimeDuration aDuration) {
return static_cast<int64_t>(aDuration.ToSeconds() * USECS_PER_S);
}
MediaDecoderStateMachine::MediaDecoderStateMachine(MediaDecoder* aDecoder,
MediaDecoderReader* aReader,
bool aRealTime) :
mDecoder(aDecoder),
mScheduler(new MediaDecoderStateMachineScheduler(
aDecoder->GetReentrantMonitor(),
&MediaDecoderStateMachine::TimeoutExpired, this, aRealTime)),
mState(DECODER_STATE_DECODING_NONE),
mSyncPointInMediaStream(-1),
mSyncPointInDecodedStream(-1),
mPlayDuration(0),
mStartTime(-1),
mEndTime(-1),
mDurationSet(false),
mFragmentEndTime(-1),
mReader(aReader),
mCurrentFrameTime(0),
mAudioStartTime(-1),
mAudioEndTime(-1),
mDecodedAudioEndTime(-1),
mVideoFrameEndTime(-1),
mDecodedVideoEndTime(-1),
mVolume(1.0),
mPlaybackRate(1.0),
mPreservesPitch(true),
mAmpleVideoFrames(2),
mLowAudioThresholdUsecs(detail::LOW_AUDIO_USECS),
mAmpleAudioThresholdUsecs(detail::AMPLE_AUDIO_USECS),
mQuickBufferingLowDataThresholdUsecs(detail::QUICK_BUFFERING_LOW_DATA_USECS),
mIsAudioPrerolling(false),
mIsVideoPrerolling(false),
mAudioRequestStatus(RequestStatus::Idle),
mVideoRequestStatus(RequestStatus::Idle),
mAudioCaptured(false),
mPositionChangeQueued(false),
mAudioCompleted(false),
mGotDurationFromMetaData(false),
mDispatchedEventToDecode(false),
mStopAudioThread(true),
mQuickBuffering(false),
mMinimizePreroll(false),
mDecodeThreadWaiting(false),
mDropAudioUntilNextDiscontinuity(false),
mDropVideoUntilNextDiscontinuity(false),
mDecodeToSeekTarget(false),
mWaitingForDecoderSeek(false),
mCancelingSeek(false),
mCurrentTimeBeforeSeek(0),
mLastFrameStatus(MediaDecoderOwner::NEXT_FRAME_UNINITIALIZED),
mDecodingFrozenAtStateDecoding(false),
mSentLoadedMetadataEvent(false),
mSentFirstFrameLoadedEvent(false)
{
MOZ_COUNT_CTOR(MediaDecoderStateMachine);
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
mAmpleVideoFrames =
std::max<uint32_t>(Preferences::GetUint("media.video-queue.default-size", 10), 3);
mBufferingWait = IsRealTime() ? 0 : 30;
mLowDataThresholdUsecs = IsRealTime() ? 0 : detail::LOW_DATA_THRESHOLD_USECS;
#ifdef XP_WIN
// Ensure high precision timers are enabled on Windows, otherwise the state
// machine thread isn't woken up at reliable intervals to set the next frame,
// and we drop frames while painting. Note that multiple calls to this
// function per-process is OK, provided each call is matched by a corresponding
// timeEndPeriod() call.
timeBeginPeriod(1);
#endif
}
MediaDecoderStateMachine::~MediaDecoderStateMachine()
{
MOZ_ASSERT(NS_IsMainThread(), "Should be on main thread.");
MOZ_COUNT_DTOR(MediaDecoderStateMachine);
NS_ASSERTION(!mPendingWakeDecoder.get(),
"WakeDecoder should have been revoked already");
mReader = nullptr;
#ifdef XP_WIN
timeEndPeriod(1);
#endif
}
bool MediaDecoderStateMachine::HasFutureAudio() {
AssertCurrentThreadInMonitor();
NS_ASSERTION(HasAudio(), "Should only call HasFutureAudio() when we have audio");
// We've got audio ready to play if:
// 1. We've not completed playback of audio, and
// 2. we either have more than the threshold of decoded audio available, or
// we've completely decoded all audio (but not finished playing it yet
// as per 1).
return !mAudioCompleted &&
(AudioDecodedUsecs() >
mLowAudioThresholdUsecs * mPlaybackRate ||
AudioQueue().IsFinished());
}
bool MediaDecoderStateMachine::HaveNextFrameData() {
AssertCurrentThreadInMonitor();
return (!HasAudio() || HasFutureAudio()) &&
(!HasVideo() || VideoQueue().GetSize() > 0);
}
int64_t MediaDecoderStateMachine::GetDecodedAudioDuration() {
NS_ASSERTION(OnDecodeThread() || OnStateMachineThread(),
"Should be on decode thread or state machine thread");
AssertCurrentThreadInMonitor();
int64_t audioDecoded = AudioQueue().Duration();
if (mAudioEndTime != -1) {
audioDecoded += mAudioEndTime - GetMediaTime();
}
return audioDecoded;
}
void MediaDecoderStateMachine::SendStreamAudio(AudioData* aAudio,
DecodedStreamData* aStream,
AudioSegment* aOutput)
{
NS_ASSERTION(OnDecodeThread() ||
OnStateMachineThread(), "Should be on decode thread or state machine thread");
AssertCurrentThreadInMonitor();
if (aAudio->mTime <= aStream->mLastAudioPacketTime) {
// ignore packet that we've already processed
return;
}
aStream->mLastAudioPacketTime = aAudio->mTime;
aStream->mLastAudioPacketEndTime = aAudio->GetEndTime();
// This logic has to mimic AudioSink closely to make sure we write
// the exact same silences
CheckedInt64 audioWrittenOffset = UsecsToFrames(mInfo.mAudio.mRate,
aStream->mInitialTime + mStartTime) + aStream->mAudioFramesWritten;
CheckedInt64 frameOffset = UsecsToFrames(mInfo.mAudio.mRate, aAudio->mTime);
if (!audioWrittenOffset.isValid() || !frameOffset.isValid())
return;
if (audioWrittenOffset.value() < frameOffset.value()) {
// Write silence to catch up
VERBOSE_LOG("writing %d frames of silence to MediaStream",
int32_t(frameOffset.value() - audioWrittenOffset.value()));
AudioSegment silence;
silence.InsertNullDataAtStart(frameOffset.value() - audioWrittenOffset.value());
aStream->mAudioFramesWritten += silence.GetDuration();
aOutput->AppendFrom(&silence);
}
int64_t offset;
if (aStream->mAudioFramesWritten == 0) {
NS_ASSERTION(frameOffset.value() <= audioWrittenOffset.value(),
"Otherwise we'd have taken the write-silence path");
// We're starting in the middle of a packet. Split the packet.
offset = audioWrittenOffset.value() - frameOffset.value();
} else {
// Write the entire packet.
offset = 0;
}
if (offset >= aAudio->mFrames)
return;
size_t framesToWrite = aAudio->mFrames - offset;
aAudio->EnsureAudioBuffer();
nsRefPtr<SharedBuffer> buffer = aAudio->mAudioBuffer;
AudioDataValue* bufferData = static_cast<AudioDataValue*>(buffer->Data());
nsAutoTArray<const AudioDataValue*,2> channels;
for (uint32_t i = 0; i < aAudio->mChannels; ++i) {
channels.AppendElement(bufferData + i*aAudio->mFrames + offset);
}
aOutput->AppendFrames(buffer.forget(), channels, framesToWrite);
VERBOSE_LOG("writing %u frames of data to MediaStream for AudioData at %lld",
static_cast<unsigned>(framesToWrite),
aAudio->mTime);
aStream->mAudioFramesWritten += framesToWrite;
aOutput->ApplyVolume(mVolume);
}
static void WriteVideoToMediaStream(MediaStream* aStream,
layers::Image* aImage,
int64_t aEndMicroseconds,
int64_t aStartMicroseconds,
const IntSize& aIntrinsicSize,
VideoSegment* aOutput)
{
nsRefPtr<layers::Image> image = aImage;
StreamTime duration =
aStream->MicrosecondsToStreamTimeRoundDown(aEndMicroseconds) -
aStream->MicrosecondsToStreamTimeRoundDown(aStartMicroseconds);
aOutput->AppendFrame(image.forget(), duration, aIntrinsicSize);
}
void MediaDecoderStateMachine::SendStreamData()
{
NS_ASSERTION(OnDecodeThread() || OnStateMachineThread(),
"Should be on decode thread or state machine thread");
AssertCurrentThreadInMonitor();
MOZ_ASSERT(mState != DECODER_STATE_DECODING_NONE);
DecodedStreamData* stream = mDecoder->GetDecodedStream();
if (!stream) {
return;
}
if (mState == DECODER_STATE_DECODING_METADATA ||
mState == DECODER_STATE_DECODING_FIRSTFRAME) {
return;
}
// If there's still an audio sink alive, then we can't send any stream
// data yet since both SendStreamData and the audio sink want to be in
// charge of popping the audio queue. We're waiting for the audio sink
if (mAudioSink) {
return;
}
int64_t minLastAudioPacketTime = INT64_MAX;
bool finished =
(!mInfo.HasAudio() || AudioQueue().IsFinished()) &&
(!mInfo.HasVideo() || VideoQueue().IsFinished());
if (mDecoder->IsSameOriginMedia()) {
SourceMediaStream* mediaStream = stream->mStream;
StreamTime endPosition = 0;
if (!stream->mStreamInitialized) {
if (mInfo.HasAudio()) {
AudioSegment* audio = new AudioSegment();
mediaStream->AddAudioTrack(kAudioTrack, mInfo.mAudio.mRate, 0, audio);
stream->mStream->DispatchWhenNotEnoughBuffered(kAudioTrack,
GetStateMachineThread(), GetWakeDecoderRunnable());
}
if (mInfo.HasVideo()) {
VideoSegment* video = new VideoSegment();
mediaStream->AddTrack(kVideoTrack, 0, video);
stream->mStream->DispatchWhenNotEnoughBuffered(kVideoTrack,
GetStateMachineThread(), GetWakeDecoderRunnable());
}
stream->mStreamInitialized = true;
}
if (mInfo.HasAudio()) {
nsAutoTArray<nsRefPtr<AudioData>,10> audio;
// It's OK to hold references to the AudioData because while audio
// is captured, only the decoder thread pops from the queue (see below).
AudioQueue().GetElementsAfter(stream->mLastAudioPacketTime, &audio);
AudioSegment output;
for (uint32_t i = 0; i < audio.Length(); ++i) {
SendStreamAudio(audio[i], stream, &output);
}
if (output.GetDuration() > 0) {
mediaStream->AppendToTrack(kAudioTrack, &output);
}
if (AudioQueue().IsFinished() && !stream->mHaveSentFinishAudio) {
mediaStream->EndTrack(kAudioTrack);
stream->mHaveSentFinishAudio = true;
}
minLastAudioPacketTime = std::min(minLastAudioPacketTime, stream->mLastAudioPacketTime);
endPosition = std::max(endPosition,
mediaStream->TicksToTimeRoundDown(mInfo.mAudio.mRate,
stream->mAudioFramesWritten));
}
if (mInfo.HasVideo()) {
nsAutoTArray<nsRefPtr<VideoData>,10> video;
// It's OK to hold references to the VideoData only the decoder thread
// pops from the queue.
VideoQueue().GetElementsAfter(stream->mNextVideoTime, &video);
VideoSegment output;
for (uint32_t i = 0; i < video.Length(); ++i) {
VideoData* v = video[i];
if (stream->mNextVideoTime < v->mTime) {
VERBOSE_LOG("writing last video to MediaStream %p for %lldus",
mediaStream, v->mTime - stream->mNextVideoTime);
// Write last video frame to catch up. mLastVideoImage can be null here
// which is fine, it just means there's no video.
WriteVideoToMediaStream(mediaStream, stream->mLastVideoImage,
v->mTime, stream->mNextVideoTime, stream->mLastVideoImageDisplaySize,
&output);
stream->mNextVideoTime = v->mTime;
}
if (stream->mNextVideoTime < v->GetEndTime()) {
VERBOSE_LOG("writing video frame %lldus to MediaStream %p for %lldus",
v->mTime, mediaStream, v->GetEndTime() - stream->mNextVideoTime);
WriteVideoToMediaStream(mediaStream, v->mImage,
v->GetEndTime(), stream->mNextVideoTime, v->mDisplay,
&output);
stream->mNextVideoTime = v->GetEndTime();
stream->mLastVideoImage = v->mImage;
stream->mLastVideoImageDisplaySize = v->mDisplay;
} else {
VERBOSE_LOG("skipping writing video frame %lldus (end %lldus) to MediaStream",
v->mTime, v->GetEndTime());
}
}
if (output.GetDuration() > 0) {
mediaStream->AppendToTrack(kVideoTrack, &output);
}
if (VideoQueue().IsFinished() && !stream->mHaveSentFinishVideo) {
mediaStream->EndTrack(kVideoTrack);
stream->mHaveSentFinishVideo = true;
}
endPosition = std::max(endPosition,
mediaStream->MicrosecondsToStreamTimeRoundDown(
stream->mNextVideoTime - stream->mInitialTime));
}
if (!stream->mHaveSentFinish) {
stream->mStream->AdvanceKnownTracksTime(endPosition);
}
if (finished && !stream->mHaveSentFinish) {
stream->mHaveSentFinish = true;
stream->mStream->Finish();
}
}
if (mAudioCaptured) {
// Discard audio packets that are no longer needed.
while (true) {
const AudioData* a = AudioQueue().PeekFront();
// Packet times are not 100% reliable so this may discard packets that
// actually contain data for mCurrentFrameTime. This means if someone might
// create a new output stream and we actually don't have the audio for the
// very start. That's OK, we'll play silence instead for a brief moment.
// That's OK. Seeking to this time would have a similar issue for such
// badly muxed resources.
if (!a || a->GetEndTime() >= minLastAudioPacketTime)
break;
OnAudioEndTimeUpdate(std::max(mAudioEndTime, a->GetEndTime()));
nsRefPtr<AudioData> releaseMe = AudioQueue().PopFront();
}
if (finished) {
mAudioCompleted = true;
UpdateReadyState();
}
}
}
MediaDecoderStateMachine::WakeDecoderRunnable*
MediaDecoderStateMachine::GetWakeDecoderRunnable()
{
AssertCurrentThreadInMonitor();
if (!mPendingWakeDecoder.get()) {
mPendingWakeDecoder = new WakeDecoderRunnable(this);
}
return mPendingWakeDecoder.get();
}
bool MediaDecoderStateMachine::HaveEnoughDecodedAudio(int64_t aAmpleAudioUSecs)
{
AssertCurrentThreadInMonitor();
if (AudioQueue().GetSize() == 0 ||
GetDecodedAudioDuration() < aAmpleAudioUSecs) {
return false;
}
if (!mAudioCaptured) {
return true;
}
DecodedStreamData* stream = mDecoder->GetDecodedStream();
// Since stream is initialized in SendStreamData() which is called when
// audio/video samples are received, we need to keep decoding to ensure
// the stream is initialized.
if (stream && !stream->mStreamInitialized) {
return false;
}
if (stream && stream->mStreamInitialized && !stream->mHaveSentFinishAudio) {
if (!stream->mStream->HaveEnoughBuffered(kAudioTrack)) {
return false;
}
stream->mStream->DispatchWhenNotEnoughBuffered(kAudioTrack,
GetStateMachineThread(), GetWakeDecoderRunnable());
}
return true;
}
bool MediaDecoderStateMachine::HaveEnoughDecodedVideo()
{
AssertCurrentThreadInMonitor();
if (static_cast<uint32_t>(VideoQueue().GetSize()) < mAmpleVideoFrames * mPlaybackRate) {
return false;
}
DecodedStreamData* stream = mDecoder->GetDecodedStream();
if (stream && !stream->mStreamInitialized) {
return false;
}
if (stream && stream->mStreamInitialized && !stream->mHaveSentFinishVideo) {
if (!stream->mStream->HaveEnoughBuffered(kVideoTrack)) {
return false;
}
stream->mStream->DispatchWhenNotEnoughBuffered(kVideoTrack,
GetStateMachineThread(), GetWakeDecoderRunnable());
}
return true;
}
bool
MediaDecoderStateMachine::NeedToDecodeVideo()
{
AssertCurrentThreadInMonitor();
return IsVideoDecoding() &&
((mState == DECODER_STATE_SEEKING && mDecodeToSeekTarget) ||
(mState == DECODER_STATE_DECODING_FIRSTFRAME &&
IsVideoDecoding() && VideoQueue().GetSize() == 0) ||
(!mMinimizePreroll && !HaveEnoughDecodedVideo()));
}
bool
MediaDecoderStateMachine::NeedToSkipToNextKeyframe()
{
AssertCurrentThreadInMonitor();
if (mState == DECODER_STATE_DECODING_FIRSTFRAME) {
return false;
}
MOZ_ASSERT(mState == DECODER_STATE_DECODING ||
mState == DECODER_STATE_BUFFERING ||
mState == DECODER_STATE_SEEKING);
// We are in seeking or buffering states, don't skip frame.
if (!IsVideoDecoding() || mState == DECODER_STATE_BUFFERING ||
mState == DECODER_STATE_SEEKING) {
return false;
}
// Don't skip frame for video-only decoded stream because the clock time of
// the stream relies on the video frame.
if (mDecoder->GetDecodedStream() && !HasAudio()) {
DECODER_LOG("Video-only decoded stream, set skipToNextKeyFrame to false");
return false;
}
// We'll skip the video decode to the nearest keyframe if we're low on
// audio, or if we're low on video, provided we're not running low on
// data to decode. If we're running low on downloaded data to decode,
// we won't start keyframe skipping, as we'll be pausing playback to buffer
// soon anyway and we'll want to be able to display frames immediately
// after buffering finishes. We ignore the low audio calculations for
// readers that are async, as since their audio decode runs on a different
// task queue it should never run low and skipping won't help their decode.
bool isLowOnDecodedAudio = !mReader->IsAsync() &&
!mIsAudioPrerolling && IsAudioDecoding() &&
(GetDecodedAudioDuration() <
mLowAudioThresholdUsecs * mPlaybackRate);
bool isLowOnDecodedVideo = !mIsVideoPrerolling &&
(mDecodedVideoEndTime - GetClock() <
LOW_VIDEO_THRESHOLD_USECS * mPlaybackRate);
bool lowUndecoded = HasLowUndecodedData();
if ((isLowOnDecodedAudio || isLowOnDecodedVideo) && !lowUndecoded) {
DECODER_LOG("Skipping video decode to the next keyframe lowAudio=%d lowVideo=%d lowUndecoded=%d async=%d",
isLowOnDecodedAudio, isLowOnDecodedVideo, lowUndecoded, mReader->IsAsync());
return true;
}
return false;
}
void
MediaDecoderStateMachine::DecodeVideo()
{
int64_t currentTime = 0;
bool skipToNextKeyFrame = false;
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
if (mState != DECODER_STATE_DECODING &&
mState != DECODER_STATE_DECODING_FIRSTFRAME &&
mState != DECODER_STATE_BUFFERING &&
mState != DECODER_STATE_SEEKING) {
mVideoRequestStatus = RequestStatus::Idle;
DispatchDecodeTasksIfNeeded();
return;
}
skipToNextKeyFrame = NeedToSkipToNextKeyframe();
currentTime = mState == DECODER_STATE_SEEKING ? 0 : GetMediaTime();
// Time the video decode, so that if it's slow, we can increase our low
// audio threshold to reduce the chance of an audio underrun while we're
// waiting for a video decode to complete.
mVideoDecodeStartTime = TimeStamp::Now();
}
SAMPLE_LOG("DecodeVideo() queued=%i, decoder-queued=%o, skip=%i, time=%lld",
VideoQueue().GetSize(), mReader->SizeOfVideoQueueInFrames(), skipToNextKeyFrame, currentTime);
mReader->RequestVideoData(skipToNextKeyFrame, currentTime)
->Then(DecodeTaskQueue(), __func__, this,
&MediaDecoderStateMachine::OnVideoDecoded,
&MediaDecoderStateMachine::OnVideoNotDecoded);
}
bool
MediaDecoderStateMachine::NeedToDecodeAudio()
{
AssertCurrentThreadInMonitor();
SAMPLE_LOG("NeedToDecodeAudio() isDec=%d decToTar=%d minPrl=%d seek=%d enufAud=%d",
IsAudioDecoding(), mDecodeToSeekTarget, mMinimizePreroll,
mState == DECODER_STATE_SEEKING,
HaveEnoughDecodedAudio(mAmpleAudioThresholdUsecs * mPlaybackRate));
return IsAudioDecoding() &&
((mState == DECODER_STATE_SEEKING && mDecodeToSeekTarget) ||
(mState == DECODER_STATE_DECODING_FIRSTFRAME &&
IsAudioDecoding() && AudioQueue().GetSize() == 0) ||
(!mMinimizePreroll &&
!HaveEnoughDecodedAudio(mAmpleAudioThresholdUsecs * mPlaybackRate) &&
(mState != DECODER_STATE_SEEKING || mDecodeToSeekTarget)));
}
void
MediaDecoderStateMachine::DecodeAudio()
{
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
if (mState != DECODER_STATE_DECODING &&
mState != DECODER_STATE_DECODING_FIRSTFRAME &&
mState != DECODER_STATE_BUFFERING &&
mState != DECODER_STATE_SEEKING) {
mAudioRequestStatus = RequestStatus::Idle;
DispatchDecodeTasksIfNeeded();
mon.NotifyAll();
return;
}
}
SAMPLE_LOG("DecodeAudio() queued=%i, decoder-queued=%o",
AudioQueue().GetSize(), mReader->SizeOfAudioQueueInFrames());
mReader->RequestAudioData()->Then(DecodeTaskQueue(), __func__, this,
&MediaDecoderStateMachine::OnAudioDecoded,
&MediaDecoderStateMachine::OnAudioNotDecoded);
}
bool
MediaDecoderStateMachine::IsAudioSeekComplete()
{
AssertCurrentThreadInMonitor();
SAMPLE_LOG("IsAudioSeekComplete() curTarVal=%d mAudDis=%d aqFin=%d aqSz=%d",
mCurrentSeekTarget.IsValid(), mDropAudioUntilNextDiscontinuity, AudioQueue().IsFinished(), AudioQueue().GetSize());
return
!HasAudio() ||
(mCurrentSeekTarget.IsValid() &&
!mDropAudioUntilNextDiscontinuity &&
(AudioQueue().IsFinished() || AudioQueue().GetSize() > 0));
}
bool
MediaDecoderStateMachine::IsVideoSeekComplete()
{
AssertCurrentThreadInMonitor();
SAMPLE_LOG("IsVideoSeekComplete() curTarVal=%d mVidDis=%d vqFin=%d vqSz=%d",
mCurrentSeekTarget.IsValid(), mDropVideoUntilNextDiscontinuity, VideoQueue().IsFinished(), VideoQueue().GetSize());
return
!HasVideo() ||
(mCurrentSeekTarget.IsValid() &&
!mDropVideoUntilNextDiscontinuity &&
(VideoQueue().IsFinished() || VideoQueue().GetSize() > 0));
}
void
MediaDecoderStateMachine::OnAudioDecoded(AudioData* aAudioSample)
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
nsRefPtr<AudioData> audio(aAudioSample);
MOZ_ASSERT(audio);
mAudioRequestStatus = RequestStatus::Idle;
mDecodedAudioEndTime = audio->GetEndTime();
SAMPLE_LOG("OnAudioDecoded [%lld,%lld] disc=%d",
(audio ? audio->mTime : -1),
(audio ? audio->GetEndTime() : -1),
(audio ? audio->mDiscontinuity : 0));
switch (mState) {
case DECODER_STATE_DECODING_FIRSTFRAME: {
Push(audio);
MaybeFinishDecodeFirstFrame();
return;
}
case DECODER_STATE_BUFFERING: {
// If we're buffering, this may be the sample we need to stop buffering.
// Save it and schedule the state machine.
Push(audio);
ScheduleStateMachine();
return;
}
case DECODER_STATE_DECODING: {
Push(audio);
if (mIsAudioPrerolling && DonePrerollingAudio()) {
StopPrerollingAudio();
}
return;
}
case DECODER_STATE_SEEKING: {
if (!mCurrentSeekTarget.IsValid()) {
// We've received a sample from a previous decode. Discard it.
return;
}
if (audio->mDiscontinuity) {
mDropAudioUntilNextDiscontinuity = false;
}
if (!mDropAudioUntilNextDiscontinuity) {
// We must be after the discontinuity; we're receiving samples
// at or after the seek target.
if (mCurrentSeekTarget.mType == SeekTarget::PrevSyncPoint &&
mCurrentSeekTarget.mTime > mCurrentTimeBeforeSeek &&
audio->mTime < mCurrentTimeBeforeSeek) {
// We are doing a fastSeek, but we ended up *before* the previous
// playback position. This is surprising UX, so switch to an accurate
// seek and decode to the seek target. This is not conformant to the
// spec, fastSeek should always be fast, but until we get the time to
// change all Readers to seek to the keyframe after the currentTime
// in this case, we'll just decode forward. Bug 1026330.
mCurrentSeekTarget.mType = SeekTarget::Accurate;
}
if (mCurrentSeekTarget.mType == SeekTarget::PrevSyncPoint) {
// Non-precise seek; we can stop the seek at the first sample.
AudioQueue().Push(audio);
} else {
// We're doing an accurate seek. We must discard
// MediaData up to the one containing exact seek target.
if (NS_FAILED(DropAudioUpToSeekTarget(audio))) {
DecodeError();
return;
}
}
}
CheckIfSeekComplete();
return;
}
default: {
// Ignore other cases.
return;
}
}
}
void
MediaDecoderStateMachine::Push(AudioData* aSample)
{
MOZ_ASSERT(aSample);
// TODO: Send aSample to MSG and recalculate readystate before pushing,
// otherwise AdvanceFrame may pop the sample before we have a chance
// to reach playing.
AudioQueue().Push(aSample);
if (mState > DECODER_STATE_DECODING_FIRSTFRAME) {
SendStreamData();
// The ready state can change when we've decoded data, so update the
// ready state, so that DOM events can fire.
UpdateReadyState();
DispatchDecodeTasksIfNeeded();
mDecoder->GetReentrantMonitor().NotifyAll();
}
}
void
MediaDecoderStateMachine::Push(VideoData* aSample)
{
MOZ_ASSERT(aSample);
// TODO: Send aSample to MSG and recalculate readystate before pushing,
// otherwise AdvanceFrame may pop the sample before we have a chance
// to reach playing.
VideoQueue().Push(aSample);
if (mState > DECODER_STATE_DECODING_FIRSTFRAME) {
SendStreamData();
// The ready state can change when we've decoded data, so update the
// ready state, so that DOM events can fire.
UpdateReadyState();
DispatchDecodeTasksIfNeeded();
mDecoder->GetReentrantMonitor().NotifyAll();
}
}
void
MediaDecoderStateMachine::OnNotDecoded(MediaData::Type aType,
MediaDecoderReader::NotDecodedReason aReason)
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
SAMPLE_LOG("OnNotDecoded (aType=%u, aReason=%u)", aType, aReason);
bool isAudio = aType == MediaData::AUDIO_DATA;
MOZ_ASSERT_IF(!isAudio, aType == MediaData::VIDEO_DATA);
// This callback means that the pending request is dead.
RequestStatusRef(aType) = RequestStatus::Idle;
// If this is a decode error, delegate to the generic error path.
if (aReason == MediaDecoderReader::DECODE_ERROR) {
DecodeError();
return;
}
// If the decoder is waiting for data, we tell it to call us back when the
// data arrives.
if (aReason == MediaDecoderReader::WAITING_FOR_DATA) {
MOZ_ASSERT(mReader->IsWaitForDataSupported(),
"Readers that send WAITING_FOR_DATA need to implement WaitForData");
RequestStatusRef(aType) = RequestStatus::Waiting;
mReader->WaitForData(aType)->Then(DecodeTaskQueue(), __func__, this,
&MediaDecoderStateMachine::OnWaitForDataResolved,
&MediaDecoderStateMachine::OnWaitForDataRejected);
return;
}
if (aReason == MediaDecoderReader::CANCELED) {
DispatchDecodeTasksIfNeeded();
return;
}
// This is an EOS. Finish off the queue, and then handle things based on our
// state.
MOZ_ASSERT(aReason == MediaDecoderReader::END_OF_STREAM);
if (!isAudio && mState == DECODER_STATE_SEEKING &&
mCurrentSeekTarget.IsValid() && mFirstVideoFrameAfterSeek) {
// Null sample. Hit end of stream. If we have decoded a frame,
// insert it into the queue so that we have something to display.
// We make sure to do this before invoking VideoQueue().Finish()
// below.
VideoQueue().Push(mFirstVideoFrameAfterSeek);
mFirstVideoFrameAfterSeek = nullptr;
}
if (isAudio) {
AudioQueue().Finish();
StopPrerollingAudio();
} else {
VideoQueue().Finish();
StopPrerollingVideo();
}
switch (mState) {
case DECODER_STATE_DECODING_FIRSTFRAME: {
MaybeFinishDecodeFirstFrame();
return;
}
case DECODER_STATE_BUFFERING:
case DECODER_STATE_DECODING: {
CheckIfDecodeComplete();
SendStreamData();
// The ready state can change when we've decoded data, so update the
// ready state, so that DOM events can fire.
UpdateReadyState();
mDecoder->GetReentrantMonitor().NotifyAll();
return;
}
case DECODER_STATE_SEEKING: {
if (!mCurrentSeekTarget.IsValid()) {
// We've received a sample from a previous decode. Discard it.
return;
}
if (isAudio) {
mDropAudioUntilNextDiscontinuity = false;
} else {
mDropVideoUntilNextDiscontinuity = false;
}
CheckIfSeekComplete();
return;
}
default: {
return;
}
}
}
void
MediaDecoderStateMachine::AcquireMonitorAndInvokeDecodeError()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
DecodeError();
}
void
MediaDecoderStateMachine::MaybeFinishDecodeFirstFrame()
{
AssertCurrentThreadInMonitor();
if ((IsAudioDecoding() && AudioQueue().GetSize() == 0) ||
(IsVideoDecoding() && VideoQueue().GetSize() == 0)) {
return;
}
if (NS_FAILED(FinishDecodeFirstFrame())) {
DecodeError();
}
}
void
MediaDecoderStateMachine::OnVideoDecoded(VideoData* aVideoSample)
{
MOZ_ASSERT(OnDecodeThread());
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
nsRefPtr<VideoData> video(aVideoSample);
mVideoRequestStatus = RequestStatus::Idle;
mDecodedVideoEndTime = video ? video->GetEndTime() : mDecodedVideoEndTime;
SAMPLE_LOG("OnVideoDecoded [%lld,%lld] disc=%d",
(video ? video->mTime : -1),
(video ? video->GetEndTime() : -1),
(video ? video->mDiscontinuity : 0));
switch (mState) {
case DECODER_STATE_DECODING_FIRSTFRAME: {
Push(video);
MaybeFinishDecodeFirstFrame();
return;
}
case DECODER_STATE_BUFFERING: {
// If we're buffering, this may be the sample we need to stop buffering.
// Save it and schedule the state machine.
Push(video);
ScheduleStateMachine();
return;
}
case DECODER_STATE_DECODING: {
Push(video);
if (mIsVideoPrerolling && DonePrerollingVideo()) {
StopPrerollingVideo();
}
// For non async readers, if the requested video sample was slow to
// arrive, increase the amount of audio we buffer to ensure that we
// don't run out of audio. This is unnecessary for async readers,
// since they decode audio and video on different threads so they
// are unlikely to run out of decoded audio.
if (mReader->IsAsync()) {
return;
}
TimeDuration decodeTime = TimeStamp::Now() - mVideoDecodeStartTime;
if (THRESHOLD_FACTOR * DurationToUsecs(decodeTime) > mLowAudioThresholdUsecs &&
!HasLowUndecodedData())
{
mLowAudioThresholdUsecs =
std::min(THRESHOLD_FACTOR * DurationToUsecs(decodeTime), mAmpleAudioThresholdUsecs);
mAmpleAudioThresholdUsecs = std::max(THRESHOLD_FACTOR * mLowAudioThresholdUsecs,
mAmpleAudioThresholdUsecs);
DECODER_LOG("Slow video decode, set mLowAudioThresholdUsecs=%lld mAmpleAudioThresholdUsecs=%lld",
mLowAudioThresholdUsecs, mAmpleAudioThresholdUsecs);
}
return;
}
case DECODER_STATE_SEEKING: {
if (!mCurrentSeekTarget.IsValid()) {
// We've received a sample from a previous decode. Discard it.
return;
}
if (mDropVideoUntilNextDiscontinuity) {
if (video->mDiscontinuity) {
mDropVideoUntilNextDiscontinuity = false;
}
}
if (!mDropVideoUntilNextDiscontinuity) {
// We must be after the discontinuity; we're receiving samples
// at or after the seek target.
if (mCurrentSeekTarget.mType == SeekTarget::PrevSyncPoint &&
mCurrentSeekTarget.mTime > mCurrentTimeBeforeSeek &&
video->mTime < mCurrentTimeBeforeSeek) {
// We are doing a fastSeek, but we ended up *before* the previous
// playback position. This is surprising UX, so switch to an accurate
// seek and decode to the seek target. This is not conformant to the
// spec, fastSeek should always be fast, but until we get the time to
// change all Readers to seek to the keyframe after the currentTime
// in this case, we'll just decode forward. Bug 1026330.
mCurrentSeekTarget.mType = SeekTarget::Accurate;
}
if (mCurrentSeekTarget.mType == SeekTarget::PrevSyncPoint) {
// Non-precise seek; we can stop the seek at the first sample.
VideoQueue().Push(video);
} else {
// We're doing an accurate seek. We still need to discard
// MediaData up to the one containing exact seek target.
if (NS_FAILED(DropVideoUpToSeekTarget(video))) {
DecodeError();
return;
}
}
}
CheckIfSeekComplete();
return;
}
default: {
// Ignore other cases.
return;
}
}
}
void
MediaDecoderStateMachine::CheckIfSeekComplete()
{
AssertCurrentThreadInMonitor();
const bool videoSeekComplete = IsVideoSeekComplete();
if (HasVideo() && !videoSeekComplete) {
// We haven't reached the target. Ensure we have requested another sample.
if (NS_FAILED(EnsureVideoDecodeTaskQueued())) {
DECODER_WARN("Failed to request video during seek");
DecodeError();
}
}
const bool audioSeekComplete = IsAudioSeekComplete();
if (HasAudio() && !audioSeekComplete) {
// We haven't reached the target. Ensure we have requested another sample.
if (NS_FAILED(EnsureAudioDecodeTaskQueued())) {
DECODER_WARN("Failed to request audio during seek");
DecodeError();
}
}
SAMPLE_LOG("CheckIfSeekComplete() audioSeekComplete=%d videoSeekComplete=%d",
audioSeekComplete, videoSeekComplete);
if (audioSeekComplete && videoSeekComplete) {
mDecodeToSeekTarget = false;
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::SeekCompleted));
nsresult rv = DecodeTaskQueue()->Dispatch(task);
if (NS_FAILED(rv)) {
DecodeError();
}
}
}
bool
MediaDecoderStateMachine::IsAudioDecoding()
{
AssertCurrentThreadInMonitor();
return HasAudio() && !AudioQueue().IsFinished();
}
bool
MediaDecoderStateMachine::IsVideoDecoding()
{
AssertCurrentThreadInMonitor();
return HasVideo() && !VideoQueue().IsFinished();
}
void
MediaDecoderStateMachine::CheckIfDecodeComplete()
{
AssertCurrentThreadInMonitor();
if (mState == DECODER_STATE_SHUTDOWN ||
mState == DECODER_STATE_SEEKING ||
mState == DECODER_STATE_COMPLETED) {
// Don't change our state if we've already been shutdown, or we're seeking,
// since we don't want to abort the shutdown or seek processes.
return;
}
if (!IsVideoDecoding() && !IsAudioDecoding()) {
// We've finished decoding all active streams,
// so move to COMPLETED state.
SetState(DECODER_STATE_COMPLETED);
DispatchDecodeTasksIfNeeded();
ScheduleStateMachine();
}
DECODER_LOG("CheckIfDecodeComplete %scompleted",
((mState == DECODER_STATE_COMPLETED) ? "" : "NOT "));
}
bool MediaDecoderStateMachine::IsPlaying() const
{
AssertCurrentThreadInMonitor();
return !mPlayStartTime.IsNull();
}
nsresult MediaDecoderStateMachine::Init(MediaDecoderStateMachine* aCloneDonor)
{
MOZ_ASSERT(NS_IsMainThread());
if (NS_WARN_IF(!mReader->EnsureTaskQueue())) {
return NS_ERROR_FAILURE;
}
MediaDecoderReader* cloneReader = nullptr;
if (aCloneDonor) {
cloneReader = aCloneDonor->mReader;
}
nsresult rv = mScheduler->Init();
NS_ENSURE_SUCCESS(rv, rv);
rv = mReader->Init(cloneReader);
NS_ENSURE_SUCCESS(rv, rv);
return NS_OK;
}
void MediaDecoderStateMachine::StopPlayback()
{
DECODER_LOG("StopPlayback()");
AssertCurrentThreadInMonitor();
mDecoder->NotifyPlaybackStopped();
if (IsPlaying()) {
mPlayDuration = GetClock() - mStartTime;
SetPlayStartTime(TimeStamp());
}
// Notify the audio sink, so that it notices that we've stopped playing,
// so it can pause audio playback.
mDecoder->GetReentrantMonitor().NotifyAll();
NS_ASSERTION(!IsPlaying(), "Should report not playing at end of StopPlayback()");
mDecoder->UpdateStreamBlockingForStateMachinePlaying();
DispatchDecodeTasksIfNeeded();
}
void MediaDecoderStateMachine::SetSyncPointForMediaStream()
{
AssertCurrentThreadInMonitor();
DecodedStreamData* stream = mDecoder->GetDecodedStream();
if (!stream) {
return;
}
mSyncPointInMediaStream = stream->GetLastOutputTime();
TimeDuration timeSincePlayStart = mPlayStartTime.IsNull() ? TimeDuration(0) :
TimeStamp::Now() - mPlayStartTime;
mSyncPointInDecodedStream = mStartTime + mPlayDuration +
timeSincePlayStart.ToMicroseconds();
DECODER_LOG("SetSyncPointForMediaStream MediaStream=%lldus, DecodedStream=%lldus",
mSyncPointInMediaStream, mSyncPointInDecodedStream);
}
void MediaDecoderStateMachine::ResyncMediaStreamClock()
{
AssertCurrentThreadInMonitor();
MOZ_ASSERT(mDecoder->GetDecodedStream());
if (IsPlaying()) {
SetPlayStartTime(TimeStamp::Now());
mPlayDuration = GetCurrentTimeViaMediaStreamSync() - mStartTime;
}
}
int64_t MediaDecoderStateMachine::GetCurrentTimeViaMediaStreamSync() const
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(mSyncPointInDecodedStream >= 0, "Should have set up sync point");
DecodedStreamData* stream = mDecoder->GetDecodedStream();
int64_t streamDelta = stream->GetLastOutputTime() - mSyncPointInMediaStream;
return mSyncPointInDecodedStream + streamDelta;
}
void MediaDecoderStateMachine::MaybeStartPlayback()
{
AssertCurrentThreadInMonitor();
if (IsPlaying()) {
// Logging this case is really spammy - don't do it.
return;
}
bool playStatePermits = mDecoder->GetState() == MediaDecoder::PLAY_STATE_PLAYING;
bool decodeStatePermits = mState == DECODER_STATE_DECODING || mState == DECODER_STATE_COMPLETED;
if (!playStatePermits || !decodeStatePermits || mIsAudioPrerolling || mIsVideoPrerolling) {
DECODER_LOG("Not starting playback [playStatePermits: %d, decodeStatePermits: %d, "
"mIsAudioPrerolling: %d, mIsVideoPrerolling: %d]", (int) playStatePermits,
(int) decodeStatePermits, (int) mIsAudioPrerolling, (int) mIsVideoPrerolling);
return;
}
if (mDecoder->CheckDecoderCanOffloadAudio()) {
DECODER_LOG("Offloading playback");
return;
}
DECODER_LOG("MaybeStartPlayback() starting playback");
mDecoder->NotifyPlaybackStarted();
SetPlayStartTime(TimeStamp::Now());
MOZ_ASSERT(IsPlaying());
nsresult rv = StartAudioThread();
NS_ENSURE_SUCCESS_VOID(rv);
mDecoder->GetReentrantMonitor().NotifyAll();
mDecoder->UpdateStreamBlockingForStateMachinePlaying();
DispatchDecodeTasksIfNeeded();
}
void MediaDecoderStateMachine::UpdatePlaybackPositionInternal(int64_t aTime)
{
SAMPLE_LOG("UpdatePlaybackPositionInternal(%lld) (mStartTime=%lld)", aTime, mStartTime);
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine thread.");
AssertCurrentThreadInMonitor();
NS_ASSERTION(mStartTime >= 0, "Should have positive mStartTime");
mCurrentFrameTime = aTime - mStartTime;
NS_ASSERTION(mCurrentFrameTime >= 0, "CurrentTime should be positive!");
if (aTime > mEndTime) {
NS_ASSERTION(mCurrentFrameTime > GetDuration(),
"CurrentTime must be after duration if aTime > endTime!");
mEndTime = aTime;
nsCOMPtr<nsIRunnable> event =
NS_NewRunnableMethod(mDecoder, &MediaDecoder::DurationChanged);
NS_DispatchToMainThread(event);
}
}
void MediaDecoderStateMachine::UpdatePlaybackPosition(int64_t aTime)
{
UpdatePlaybackPositionInternal(aTime);
bool fragmentEnded = mFragmentEndTime >= 0 && GetMediaTime() >= mFragmentEndTime;
if (!mPositionChangeQueued || fragmentEnded) {
mPositionChangeQueued = true;
nsCOMPtr<nsIRunnable> event =
NS_NewRunnableMethod(mDecoder, &MediaDecoder::PlaybackPositionChanged);
NS_DispatchToMainThread(event);
}
mMetadataManager.DispatchMetadataIfNeeded(mDecoder, aTime);
if (fragmentEnded) {
StopPlayback();
}
}
void MediaDecoderStateMachine::ClearPositionChangeFlag()
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
AssertCurrentThreadInMonitor();
mPositionChangeQueued = false;
}
MediaDecoderOwner::NextFrameStatus MediaDecoderStateMachine::GetNextFrameStatus()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (IsBuffering()) {
return MediaDecoderOwner::NEXT_FRAME_UNAVAILABLE_BUFFERING;
} else if (IsSeeking()) {
return MediaDecoderOwner::NEXT_FRAME_UNAVAILABLE_SEEKING;
} else if (HaveNextFrameData()) {
return MediaDecoderOwner::NEXT_FRAME_AVAILABLE;
}
return MediaDecoderOwner::NEXT_FRAME_UNAVAILABLE;
}
static const char* const gMachineStateStr[] = {
"NONE",
"DECODING_METADATA",
"WAIT_FOR_RESOURCES",
"DECODING_FIRSTFRAME",
"DORMANT",
"DECODING",
"SEEKING",
"BUFFERING",
"COMPLETED",
"SHUTDOWN"
};
void MediaDecoderStateMachine::SetState(State aState)
{
AssertCurrentThreadInMonitor();
if (mState == aState) {
return;
}
DECODER_LOG("Change machine state from %s to %s",
gMachineStateStr[mState], gMachineStateStr[aState]);
mState = aState;
}
void MediaDecoderStateMachine::SetVolume(double volume)
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
mVolume = volume;
if (mAudioSink) {
mAudioSink->SetVolume(mVolume);
}
}
void MediaDecoderStateMachine::SetAudioCaptured(bool aCaptured)
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (!mAudioCaptured && aCaptured && !mStopAudioThread) {
// Make sure the state machine runs as soon as possible. That will
// stop the audio sink.
// If mStopAudioThread is true then we're already stopping the audio sink
// and since we set mAudioCaptured to true, nothing can start it again.
ScheduleStateMachine();
if (HasAudio()) {
// The audio clock is active so force a resync now in case the audio
// clock is ahead of us (observed on Android), since after mAudioCaptured
// gets set can't call GetAudioClock().
ResyncAudioClock();
}
}
mAudioCaptured = aCaptured;
}
double MediaDecoderStateMachine::GetCurrentTime() const
{
NS_ASSERTION(NS_IsMainThread() ||
OnStateMachineThread() ||
OnDecodeThread(),
"Should be on main, decode, or state machine thread.");
return static_cast<double>(mCurrentFrameTime) / static_cast<double>(USECS_PER_S);
}
bool MediaDecoderStateMachine::IsRealTime() const {
return mScheduler->IsRealTime();
}
int64_t MediaDecoderStateMachine::GetDuration()
{
AssertCurrentThreadInMonitor();
if (mEndTime == -1 || mStartTime == -1)
return -1;
return mEndTime - mStartTime;
}
int64_t MediaDecoderStateMachine::GetEndTime()
{
if (mEndTime == -1 && mDurationSet) {
return INT64_MAX;
}
return mEndTime;
}
void MediaDecoderStateMachine::SetDuration(int64_t aDuration)
{
NS_ASSERTION(NS_IsMainThread() || OnDecodeThread(),
"Should be on main or decode thread.");
AssertCurrentThreadInMonitor();
if (aDuration == -1) {
mDurationSet = false;
return;
}
mDurationSet = true;
if (mStartTime == -1) {
SetStartTime(0);
}
if (aDuration == INT64_MAX) {
mEndTime = -1;
return;
}
mEndTime = mStartTime + aDuration;
}
void MediaDecoderStateMachine::UpdateEstimatedDuration(int64_t aDuration)
{
AssertCurrentThreadInMonitor();
int64_t duration = GetDuration();
if (aDuration != duration &&
mozilla::Abs(aDuration - duration) > ESTIMATED_DURATION_FUZZ_FACTOR_USECS) {
SetDuration(aDuration);
nsCOMPtr<nsIRunnable> event =
NS_NewRunnableMethod(mDecoder, &MediaDecoder::DurationChanged);
NS_DispatchToMainThread(event);
}
}
void MediaDecoderStateMachine::SetMediaEndTime(int64_t aEndTime)
{
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread");
AssertCurrentThreadInMonitor();
mEndTime = aEndTime;
}
void MediaDecoderStateMachine::SetFragmentEndTime(int64_t aEndTime)
{
AssertCurrentThreadInMonitor();
mFragmentEndTime = aEndTime < 0 ? aEndTime : aEndTime + mStartTime;
}
bool MediaDecoderStateMachine::IsDormantNeeded()
{
return mReader->IsDormantNeeded();
}
void MediaDecoderStateMachine::SetDormant(bool aDormant)
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
AssertCurrentThreadInMonitor();
if (!mReader) {
return;
}
DECODER_LOG("SetDormant=%d", aDormant);
if (aDormant) {
if (mState == DECODER_STATE_SEEKING && !mQueuedSeekTarget.IsValid()) {
if (mSeekTarget.IsValid()) {
mQueuedSeekTarget = mSeekTarget;
} else if (mCurrentSeekTarget.IsValid()) {
mQueuedSeekTarget = mCurrentSeekTarget;
}
}
mSeekTarget.Reset();
mCurrentSeekTarget.Reset();
ScheduleStateMachine();
SetState(DECODER_STATE_DORMANT);
StopPlayback();
mDecoder->GetReentrantMonitor().NotifyAll();
} else if ((aDormant != true) && (mState == DECODER_STATE_DORMANT)) {
mDecodingFrozenAtStateDecoding = true;
ScheduleStateMachine();
mCurrentFrameTime = 0;
SetState(DECODER_STATE_DECODING_NONE);
mDecoder->GetReentrantMonitor().NotifyAll();
}
}
void MediaDecoderStateMachine::Shutdown()
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
// Once we've entered the shutdown state here there's no going back.
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
// Change state before issuing shutdown request to threads so those
// threads can start exiting cleanly during the Shutdown call.
DECODER_LOG("Changed state to SHUTDOWN");
SetState(DECODER_STATE_SHUTDOWN);
mScheduler->ScheduleAndShutdown();
if (mAudioSink) {
mAudioSink->PrepareToShutdown();
}
mDecoder->GetReentrantMonitor().NotifyAll();
}
void MediaDecoderStateMachine::StartDecoding()
{
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (mState == DECODER_STATE_DECODING) {
return;
}
SetState(DECODER_STATE_DECODING);
mDecodeStartTime = TimeStamp::Now();
CheckIfDecodeComplete();
if (mState == DECODER_STATE_COMPLETED) {
return;
}
// Reset other state to pristine values before starting decode.
mIsAudioPrerolling = !DonePrerollingAudio();
mIsVideoPrerolling = !DonePrerollingVideo();
// Ensure that we've got tasks enqueued to decode data if we need to.
DispatchDecodeTasksIfNeeded();
ScheduleStateMachine();
}
void MediaDecoderStateMachine::StartWaitForResources()
{
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
AssertCurrentThreadInMonitor();
SetState(DECODER_STATE_WAIT_FOR_RESOURCES);
DECODER_LOG("StartWaitForResources");
}
void MediaDecoderStateMachine::NotifyWaitingForResourcesStatusChanged()
{
AssertCurrentThreadInMonitor();
DECODER_LOG("NotifyWaitingForResourcesStatusChanged");
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this,
&MediaDecoderStateMachine::DoNotifyWaitingForResourcesStatusChanged));
DecodeTaskQueue()->Dispatch(task);
}
void MediaDecoderStateMachine::DoNotifyWaitingForResourcesStatusChanged()
{
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (mState != DECODER_STATE_WAIT_FOR_RESOURCES) {
return;
}
DECODER_LOG("DoNotifyWaitingForResourcesStatusChanged");
// The reader is no longer waiting for resources (say a hardware decoder),
// we can now proceed to decode metadata.
SetState(DECODER_STATE_DECODING_NONE);
ScheduleStateMachine();
}
void MediaDecoderStateMachine::PlayInternal()
{
NS_ASSERTION(OnStateMachineThread(), "Should be on state machine thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
// Once we start playing, we don't want to minimize our prerolling, as we
// assume the user is likely to want to keep playing in future. This needs to
// happen before we invoke StartDecoding().
if (mMinimizePreroll) {
mMinimizePreroll = false;
DispatchDecodeTasksIfNeeded();
}
// Some state transitions still happen synchronously on the main thread. So
// if the main thread invokes Play() and then Seek(), the seek will initiate
// synchronously on the main thread, and the asynchronous PlayInternal task
// will arrive when it's no longer valid. The proper thing to do is to move
// all state transitions to the state machine thread, but for now we just
// make sure that none of the possible main-thread state transitions (Seek(),
// SetDormant(), and Shutdown()) have not occurred.
if (mState != DECODER_STATE_DECODING && mState != DECODER_STATE_BUFFERING &&
mState != DECODER_STATE_COMPLETED)
{
DECODER_LOG("Unexpected state - Bailing out of PlayInternal()");
return;
}
// When asked to play, switch to decoding state only if
// we are currently buffering. In other cases, we'll start playing anyway
// when the state machine notices the decoder's state change to PLAYING.
if (mState == DECODER_STATE_BUFFERING) {
StartDecoding();
}
if (mDecodingFrozenAtStateDecoding) {
mDecodingFrozenAtStateDecoding = false;
DispatchDecodeTasksIfNeeded();
}
ScheduleStateMachine();
}
void MediaDecoderStateMachine::ResetPlayback()
{
AssertCurrentThreadInMonitor();
MOZ_ASSERT(mState == DECODER_STATE_SEEKING ||
mState == DECODER_STATE_SHUTDOWN ||
mState == DECODER_STATE_DORMANT);
// Audio thread should've been stopped at the moment. Otherwise, AudioSink
// might be accessing AudioQueue outside of the decoder monitor while we
// are clearing the queue and causes crash for no samples to be popped.
MOZ_ASSERT(!mAudioSink);
mVideoFrameEndTime = -1;
mDecodedVideoEndTime = -1;
mAudioStartTime = -1;
mAudioEndTime = -1;
mDecodedAudioEndTime = -1;
mAudioCompleted = false;
AudioQueue().Reset();
VideoQueue().Reset();
mFirstVideoFrameAfterSeek = nullptr;
mDropAudioUntilNextDiscontinuity = true;
mDropVideoUntilNextDiscontinuity = true;
}
void MediaDecoderStateMachine::NotifyDataArrived(const char* aBuffer,
uint32_t aLength,
int64_t aOffset)
{
NS_ASSERTION(NS_IsMainThread(), "Only call on main thread");
mReader->NotifyDataArrived(aBuffer, aLength, aOffset);
// While playing an unseekable stream of unknown duration, mEndTime is
// updated (in AdvanceFrame()) as we play. But if data is being downloaded
// faster than played, mEndTime won't reflect the end of playable data
// since we haven't played the frame at the end of buffered data. So update
// mEndTime here as new data is downloaded to prevent such a lag.
//
// Make sure to only do this if we have a start time, otherwise the reader
// doesn't know how to compute GetBuffered.
nsRefPtr<dom::TimeRanges> buffered = new dom::TimeRanges();
if (mDecoder->IsInfinite() && (mStartTime != -1) &&
NS_SUCCEEDED(mDecoder->GetBuffered(buffered)))
{
uint32_t length = 0;
buffered->GetLength(&length);
if (length) {
double end = 0;
buffered->End(length - 1, &end);
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
mEndTime = std::max<int64_t>(mEndTime, end * USECS_PER_S);
}
}
}
void MediaDecoderStateMachine::Seek(const SeekTarget& aTarget)
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
mDecodingFrozenAtStateDecoding = false;
if (mState == DECODER_STATE_SHUTDOWN) {
return;
}
// We need to be able to seek both at a transport level and at a media level
// to seek.
if (!mDecoder->IsMediaSeekable()) {
DECODER_WARN("Seek() function should not be called on a non-seekable state machine");
return;
}
NS_ASSERTION(mState > DECODER_STATE_DECODING_METADATA,
"We should have got duration already");
if (mState < DECODER_STATE_DECODING) {
DECODER_LOG("Seek() Not Enough Data to continue at this stage, queuing seek");
mQueuedSeekTarget = aTarget;
return;
}
mQueuedSeekTarget.Reset();
StartSeek(aTarget);
}
void
MediaDecoderStateMachine::EnqueueStartQueuedSeekTask()
{
nsCOMPtr<nsIRunnable> event =
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::StartQueuedSeek);
NS_DispatchToMainThread(event, NS_DISPATCH_NORMAL);
}
void
MediaDecoderStateMachine::StartQueuedSeek()
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (!mQueuedSeekTarget.IsValid()) {
return;
}
StartSeek(mQueuedSeekTarget);
mQueuedSeekTarget.Reset();
}
void
MediaDecoderStateMachine::StartSeek(const SeekTarget& aTarget)
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
AssertCurrentThreadInMonitor();
MOZ_ASSERT(mState >= DECODER_STATE_DECODING);
if (mState == DECODER_STATE_SHUTDOWN) {
return;
}
// Bound the seek time to be inside the media range.
int64_t end = GetEndTime();
NS_ASSERTION(mStartTime != -1, "Should know start time by now");
NS_ASSERTION(end != -1, "Should know end time by now");
int64_t seekTime = aTarget.mTime + mStartTime;
seekTime = std::min(seekTime, end);
seekTime = std::max(mStartTime, seekTime);
NS_ASSERTION(seekTime >= mStartTime && seekTime <= end,
"Can only seek in range [0,duration]");
mSeekTarget = SeekTarget(seekTime, aTarget.mType);
DECODER_LOG("Changed state to SEEKING (to %lld)", mSeekTarget.mTime);
SetState(DECODER_STATE_SEEKING);
mDecoder->RecreateDecodedStreamIfNecessary(seekTime - mStartTime);
ScheduleStateMachine();
}
void MediaDecoderStateMachine::StopAudioThread()
{
NS_ASSERTION(OnDecodeThread() || OnStateMachineThread(),
"Should be on decode thread or state machine thread");
AssertCurrentThreadInMonitor();
if (mStopAudioThread) {
// Audio sink is being stopped in another thread. Wait until finished.
while (mAudioSink) {
mDecoder->GetReentrantMonitor().Wait();
}
return;
}
mStopAudioThread = true;
// Wake up audio sink so that it can reach the finish line.
mDecoder->GetReentrantMonitor().NotifyAll();
if (mAudioSink) {
DECODER_LOG("Shutdown audio thread");
mAudioSink->PrepareToShutdown();
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
mAudioSink->Shutdown();
}
mAudioSink = nullptr;
// Now that the audio sink is dead, try sending data to our MediaStream(s).
// That may have been waiting for the audio thread to stop.
SendStreamData();
}
// Wake up those waiting for audio sink to finish.
mDecoder->GetReentrantMonitor().NotifyAll();
}
nsresult
MediaDecoderStateMachine::EnqueueDecodeMetadataTask()
{
AssertCurrentThreadInMonitor();
MOZ_ASSERT(mState == DECODER_STATE_DECODING_METADATA);
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::CallDecodeMetadata));
nsresult rv = DecodeTaskQueue()->Dispatch(task);
NS_ENSURE_SUCCESS(rv, rv);
return NS_OK;
}
nsresult
MediaDecoderStateMachine::EnqueueDecodeFirstFrameTask()
{
AssertCurrentThreadInMonitor();
MOZ_ASSERT(mState == DECODER_STATE_DECODING_FIRSTFRAME);
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::CallDecodeFirstFrame));
nsresult rv = DecodeTaskQueue()->Dispatch(task);
NS_ENSURE_SUCCESS(rv, rv);
return NS_OK;
}
void
MediaDecoderStateMachine::SetReaderIdle()
{
#ifdef PR_LOGGING
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
DECODER_LOG("SetReaderIdle() audioQueue=%lld videoQueue=%lld",
GetDecodedAudioDuration(),
VideoQueue().Duration());
}
#endif
MOZ_ASSERT(OnDecodeThread());
mReader->SetIdle();
}
void
MediaDecoderStateMachine::DispatchDecodeTasksIfNeeded()
{
AssertCurrentThreadInMonitor();
if (mState != DECODER_STATE_DECODING &&
mState != DECODER_STATE_DECODING_FIRSTFRAME &&
mState != DECODER_STATE_BUFFERING &&
mState != DECODER_STATE_SEEKING) {
return;
}
if (mState == DECODER_STATE_DECODING && mDecodingFrozenAtStateDecoding) {
DECODER_LOG("DispatchDecodeTasksIfNeeded return due to "
"mFreezeDecodingAtStateDecoding");
return;
}
// NeedToDecodeAudio() can go from false to true while we hold the
// monitor, but it can't go from true to false. This can happen because
// NeedToDecodeAudio() takes into account the amount of decoded audio
// that's been written to the AudioStream but not played yet. So if we
// were calling NeedToDecodeAudio() twice and we thread-context switch
// between the calls, audio can play, which can affect the return value
// of NeedToDecodeAudio() giving inconsistent results. So we cache the
// value returned by NeedToDecodeAudio(), and make decisions
// based on the cached value. If NeedToDecodeAudio() has
// returned false, and then subsequently returns true and we're not
// playing, it will probably be OK since we don't need to consume data
// anyway.
const bool needToDecodeAudio = NeedToDecodeAudio();
const bool needToDecodeVideo = NeedToDecodeVideo();
// If we're in completed state, we should not need to decode anything else.
MOZ_ASSERT(mState != DECODER_STATE_COMPLETED ||
(!needToDecodeAudio && !needToDecodeVideo));
bool needIdle = !mDecoder->IsLogicallyPlaying() &&
mState != DECODER_STATE_SEEKING &&
!needToDecodeAudio &&
!needToDecodeVideo &&
!IsPlaying();
SAMPLE_LOG("DispatchDecodeTasksIfNeeded needAudio=%d audioStatus=%d needVideo=%d videoStatus=%d needIdle=%d",
needToDecodeAudio, mAudioRequestStatus,
needToDecodeVideo, mVideoRequestStatus,
needIdle);
if (needToDecodeAudio) {
EnsureAudioDecodeTaskQueued();
}
if (needToDecodeVideo) {
EnsureVideoDecodeTaskQueued();
}
if (needIdle) {
RefPtr<nsIRunnable> event = NS_NewRunnableMethod(
this, &MediaDecoderStateMachine::SetReaderIdle);
nsresult rv = DecodeTaskQueue()->Dispatch(event.forget());
if (NS_FAILED(rv) && mState != DECODER_STATE_SHUTDOWN) {
DECODER_WARN("Failed to dispatch event to set decoder idle state");
}
}
}
nsresult
MediaDecoderStateMachine::EnqueueDecodeSeekTask()
{
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
AssertCurrentThreadInMonitor();
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::DecodeSeek));
nsresult rv = DecodeTaskQueue()->Dispatch(task);
if (NS_FAILED(rv)) {
DECODER_WARN("Dispatch DecodeSeek task failed.");
DecodeError();
}
return rv;
}
nsresult
MediaDecoderStateMachine::DispatchAudioDecodeTaskIfNeeded()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
if (NeedToDecodeAudio()) {
return EnsureAudioDecodeTaskQueued();
}
return NS_OK;
}
nsresult
MediaDecoderStateMachine::EnsureAudioDecodeTaskQueued()
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
SAMPLE_LOG("EnsureAudioDecodeTaskQueued isDecoding=%d status=%d",
IsAudioDecoding(), mAudioRequestStatus);
if (mState >= DECODER_STATE_COMPLETED) {
return NS_OK;
}
MOZ_ASSERT(mState >= DECODER_STATE_DECODING_FIRSTFRAME);
if (IsAudioDecoding() && mAudioRequestStatus == RequestStatus::Idle && !mWaitingForDecoderSeek) {
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::DecodeAudio));
nsresult rv = DecodeTaskQueue()->Dispatch(task);
if (NS_SUCCEEDED(rv)) {
mAudioRequestStatus = RequestStatus::Pending;
} else {
DECODER_WARN("Failed to dispatch task to decode audio");
}
}
return NS_OK;
}
nsresult
MediaDecoderStateMachine::DispatchVideoDecodeTaskIfNeeded()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
if (NeedToDecodeVideo()) {
return EnsureVideoDecodeTaskQueued();
}
return NS_OK;
}
nsresult
MediaDecoderStateMachine::EnsureVideoDecodeTaskQueued()
{
AssertCurrentThreadInMonitor();
SAMPLE_LOG("EnsureVideoDecodeTaskQueued isDecoding=%d status=%d",
IsVideoDecoding(), mVideoRequestStatus);
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
if (mState >= DECODER_STATE_COMPLETED) {
return NS_OK;
}
MOZ_ASSERT(mState >= DECODER_STATE_DECODING_FIRSTFRAME);
if (IsVideoDecoding() && mVideoRequestStatus == RequestStatus::Idle && !mWaitingForDecoderSeek) {
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::DecodeVideo));
nsresult rv = DecodeTaskQueue()->Dispatch(task);
if (NS_SUCCEEDED(rv)) {
mVideoRequestStatus = RequestStatus::Pending;
} else {
DECODER_WARN("Failed to dispatch task to decode video");
}
}
return NS_OK;
}
nsresult
MediaDecoderStateMachine::StartAudioThread()
{
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
AssertCurrentThreadInMonitor();
if (mAudioCaptured) {
NS_ASSERTION(mStopAudioThread, "mStopAudioThread must always be true if audio is captured");
return NS_OK;
}
mStopAudioThread = false;
if (HasAudio() && !mAudioSink) {
// The audio end time should always be at least the audio start time.
mAudioEndTime = mAudioStartTime;
MOZ_ASSERT(mAudioStartTime == GetMediaTime());
mAudioCompleted = false;
mAudioSink = new AudioSink(this, mAudioStartTime,
mInfo.mAudio, mDecoder->GetAudioChannel());
// OnAudioSinkError() will be called before Init() returns if an error
// occurs during initialization.
nsresult rv = mAudioSink->Init();
NS_ENSURE_SUCCESS(rv, rv);
mAudioSink->SetVolume(mVolume);
mAudioSink->SetPlaybackRate(mPlaybackRate);
mAudioSink->SetPreservesPitch(mPreservesPitch);
}
return NS_OK;
}
int64_t MediaDecoderStateMachine::AudioDecodedUsecs()
{
NS_ASSERTION(HasAudio(),
"Should only call AudioDecodedUsecs() when we have audio");
// The amount of audio we have decoded is the amount of audio data we've
// already decoded and pushed to the hardware, plus the amount of audio
// data waiting to be pushed to the hardware.
int64_t pushed = (mAudioEndTime != -1) ? (mAudioEndTime - GetMediaTime()) : 0;
// Currently for real time streams, AudioQueue().Duration() produce
// wrong values (Bug 1114434), so we use frame counts to calculate duration.
if (IsRealTime()) {
return pushed + FramesToUsecs(AudioQueue().FrameCount(), mInfo.mAudio.mRate).value();
}
return pushed + AudioQueue().Duration();
}
bool MediaDecoderStateMachine::HasLowDecodedData(int64_t aAudioUsecs)
{
AssertCurrentThreadInMonitor();
MOZ_ASSERT(mReader->UseBufferingHeuristics());
// We consider ourselves low on decoded data if we're low on audio,
// provided we've not decoded to the end of the audio stream, or
// if we're low on video frames, provided
// we've not decoded to the end of the video stream.
return ((IsAudioDecoding() && AudioDecodedUsecs() < aAudioUsecs) ||
(IsVideoDecoding() &&
static_cast<uint32_t>(VideoQueue().GetSize()) < LOW_VIDEO_FRAMES));
}
bool MediaDecoderStateMachine::HasLowUndecodedData()
{
return HasLowUndecodedData(mLowDataThresholdUsecs);
}
bool MediaDecoderStateMachine::HasLowUndecodedData(int64_t aUsecs)
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(mState > DECODER_STATE_DECODING_FIRSTFRAME,
"Must have loaded first frame for GetBuffered() to work");
nsRefPtr<dom::TimeRanges> buffered = new dom::TimeRanges();
nsresult rv = mReader->GetBuffered(buffered.get());
NS_ENSURE_SUCCESS(rv, false);
int64_t endOfDecodedVideoData = INT64_MAX;
if (HasVideo() && !VideoQueue().AtEndOfStream()) {
endOfDecodedVideoData = VideoQueue().Peek() ? VideoQueue().Peek()->GetEndTime() : mVideoFrameEndTime;
}
int64_t endOfDecodedAudioData = INT64_MAX;
if (HasAudio() && !AudioQueue().AtEndOfStream()) {
// mDecodedAudioEndTime could be -1 when no audio samples are decoded.
// But that is fine since we consider ourself as low in decoded data when
// we don't have any decoded audio samples at all.
endOfDecodedAudioData = mDecodedAudioEndTime;
}
int64_t endOfDecodedData = std::min(endOfDecodedVideoData, endOfDecodedAudioData);
return endOfDecodedData != INT64_MAX &&
!buffered->Contains(static_cast<double>(endOfDecodedData) / USECS_PER_S,
static_cast<double>(std::min(endOfDecodedData + aUsecs, GetDuration())) / USECS_PER_S);
}
void
MediaDecoderStateMachine::DecodeError()
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
if (mState == DECODER_STATE_SHUTDOWN) {
// Already shutdown.
return;
}
// Change state to shutdown before sending error report to MediaDecoder
// and the HTMLMediaElement, so that our pipeline can start exiting
// cleanly during the sync dispatch below.
DECODER_WARN("Decode error, changed state to SHUTDOWN due to error");
SetState(DECODER_STATE_SHUTDOWN);
mScheduler->ScheduleAndShutdown();
mDecoder->GetReentrantMonitor().NotifyAll();
// Dispatch the event to call DecodeError synchronously. This ensures
// we're in shutdown state by the time we exit the decode thread.
// If we just moved to shutdown state here on the decode thread, we may
// cause the state machine to shutdown/free memory without closing its
// media stream properly, and we'll get callbacks from the media stream
// causing a crash.
{
nsCOMPtr<nsIRunnable> event =
NS_NewRunnableMethod(mDecoder, &MediaDecoder::DecodeError);
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
NS_DispatchToMainThread(event, NS_DISPATCH_SYNC);
}
}
void
MediaDecoderStateMachine::CallDecodeMetadata()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (mState != DECODER_STATE_DECODING_METADATA) {
return;
}
if (NS_FAILED(DecodeMetadata())) {
DECODER_WARN("Decode metadata failed, shutting down decoder");
DecodeError();
}
}
nsresult MediaDecoderStateMachine::DecodeMetadata()
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
MOZ_ASSERT(mState == DECODER_STATE_DECODING_METADATA);
DECODER_LOG("Decoding Media Headers");
mReader->PreReadMetadata();
if (mReader->IsWaitingMediaResources()) {
StartWaitForResources();
return NS_OK;
}
nsresult res;
MediaInfo info;
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
res = mReader->ReadMetadata(&info, getter_Transfers(mMetadataTags));
}
if (NS_SUCCEEDED(res)) {
if (mState == DECODER_STATE_DECODING_METADATA &&
mReader->IsWaitingMediaResources()) {
// change state to DECODER_STATE_WAIT_FOR_RESOURCES
StartWaitForResources();
// affect values only if ReadMetadata succeeds
return NS_OK;
}
}
if (NS_SUCCEEDED(res)) {
mDecoder->SetMediaSeekable(mReader->IsMediaSeekable());
}
mInfo = info;
if (NS_FAILED(res) || (!info.HasValidMedia())) {
DECODER_WARN("ReadMetadata failed, res=%x HasValidMedia=%d", res, info.HasValidMedia());
return NS_ERROR_FAILURE;
}
mDecoder->StartProgressUpdates();
mGotDurationFromMetaData = (GetDuration() != -1) || mDurationSet;
if (mGotDurationFromMetaData) {
// We have all the information required: duration and size
// Inform the element that we've loaded the metadata.
EnqueueLoadedMetadataEvent();
}
if (mState == DECODER_STATE_DECODING_METADATA) {
SetState(DECODER_STATE_DECODING_FIRSTFRAME);
res = EnqueueDecodeFirstFrameTask();
if (NS_FAILED(res)) {
return NS_ERROR_FAILURE;
}
}
ScheduleStateMachine();
return NS_OK;
}
void
MediaDecoderStateMachine::EnqueueLoadedMetadataEvent()
{
nsAutoPtr<MediaInfo> info(new MediaInfo());
*info = mInfo;
nsCOMPtr<nsIRunnable> metadataLoadedEvent =
new MetadataEventRunner(mDecoder, info, mMetadataTags, mSentLoadedMetadataEvent);
NS_DispatchToMainThread(metadataLoadedEvent, NS_DISPATCH_NORMAL);
mSentLoadedMetadataEvent = true;
}
void
MediaDecoderStateMachine::EnqueueFirstFrameLoadedEvent()
{
nsAutoPtr<MediaInfo> info(new MediaInfo());
*info = mInfo;
nsCOMPtr<nsIRunnable> event =
new FirstFrameLoadedEventRunner(mDecoder, info, mSentFirstFrameLoadedEvent);
NS_DispatchToMainThread(event, NS_DISPATCH_NORMAL);
mSentFirstFrameLoadedEvent = true;
}
void
MediaDecoderStateMachine::CallDecodeFirstFrame()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (mState != DECODER_STATE_DECODING_FIRSTFRAME) {
return;
}
if (NS_FAILED(DecodeFirstFrame())) {
DECODER_WARN("Decode failed to start, shutting down decoder");
DecodeError();
}
}
nsresult
MediaDecoderStateMachine::DecodeFirstFrame()
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
MOZ_ASSERT(mState == DECODER_STATE_DECODING_FIRSTFRAME);
DECODER_LOG("DecodeFirstFrame started");
if (HasAudio()) {
RefPtr<nsIRunnable> decodeTask(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::DispatchAudioDecodeTaskIfNeeded));
AudioQueue().AddPopListener(decodeTask, DecodeTaskQueue());
}
if (HasVideo()) {
RefPtr<nsIRunnable> decodeTask(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::DispatchVideoDecodeTaskIfNeeded));
VideoQueue().AddPopListener(decodeTask, DecodeTaskQueue());
}
if (IsRealTime()) {
SetStartTime(0);
nsresult res = FinishDecodeFirstFrame();
NS_ENSURE_SUCCESS(res, res);
} else if (mSentFirstFrameLoadedEvent) {
// We're resuming from dormant state, so we don't need to request
// the first samples in order to determine the media start time,
// we have the start time from last time we loaded.
SetStartTime(mStartTime);
nsresult res = FinishDecodeFirstFrame();
NS_ENSURE_SUCCESS(res, res);
} else {
if (HasAudio()) {
mAudioRequestStatus = RequestStatus::Pending;
ReentrantMonitorAutoExit unlock(mDecoder->GetReentrantMonitor());
mReader->RequestAudioData()->Then(DecodeTaskQueue(), __func__, this,
&MediaDecoderStateMachine::OnAudioDecoded,
&MediaDecoderStateMachine::OnAudioNotDecoded);
}
if (HasVideo()) {
mVideoDecodeStartTime = TimeStamp::Now();
mVideoRequestStatus = RequestStatus::Pending;
ReentrantMonitorAutoExit unlock(mDecoder->GetReentrantMonitor());
mReader->RequestVideoData(false, 0)
->Then(DecodeTaskQueue(), __func__, this,
&MediaDecoderStateMachine::OnVideoDecoded,
&MediaDecoderStateMachine::OnVideoNotDecoded);
}
}
return NS_OK;
}
nsresult
MediaDecoderStateMachine::FinishDecodeFirstFrame()
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
DECODER_LOG("FinishDecodeFirstFrame");
if (mState == DECODER_STATE_SHUTDOWN) {
return NS_ERROR_FAILURE;
}
if (!IsRealTime() && !mSentFirstFrameLoadedEvent) {
const VideoData* v = VideoQueue().PeekFront();
const AudioData* a = AudioQueue().PeekFront();
SetStartTime(mReader->ComputeStartTime(v, a));
if (VideoQueue().GetSize()) {
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
RenderVideoFrame(VideoQueue().PeekFront(), TimeStamp::Now());
}
}
NS_ASSERTION(mStartTime != -1, "Must have start time");
MOZ_ASSERT(!(mDecoder->IsMediaSeekable() && mDecoder->IsTransportSeekable()) ||
(GetDuration() != -1) || mDurationSet,
"Seekable media should have duration");
DECODER_LOG("Media goes from %lld to %lld (duration %lld) "
"transportSeekable=%d, mediaSeekable=%d",
mStartTime, mEndTime, GetDuration(),
mDecoder->IsTransportSeekable(), mDecoder->IsMediaSeekable());
if (HasAudio() && !HasVideo()) {
// We're playing audio only. We don't need to worry about slow video
// decodes causing audio underruns, so don't buffer so much audio in
// order to reduce memory usage.
mAmpleAudioThresholdUsecs /= NO_VIDEO_AMPLE_AUDIO_DIVISOR;
mLowAudioThresholdUsecs /= NO_VIDEO_AMPLE_AUDIO_DIVISOR;
mQuickBufferingLowDataThresholdUsecs /= NO_VIDEO_AMPLE_AUDIO_DIVISOR;
}
// Get potentially updated metadata
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
mReader->ReadUpdatedMetadata(&mInfo);
}
nsAutoPtr<MediaInfo> info(new MediaInfo());
*info = mInfo;
if (!mGotDurationFromMetaData) {
// We now have a duration, we can fire the LoadedMetadata and
// FirstFrame event.
EnqueueLoadedMetadataEvent();
EnqueueFirstFrameLoadedEvent();
} else {
// Inform the element that we've loaded the first frame.
EnqueueFirstFrameLoadedEvent();
}
if (mState == DECODER_STATE_DECODING_FIRSTFRAME) {
StartDecoding();
}
// For very short media the first frame decode can decode the entire media.
// So we need to check if this has occurred, else our decode pipeline won't
// run (since it doesn't need to) and we won't detect end of stream.
CheckIfDecodeComplete();
MaybeStartPlayback();
if (mQueuedSeekTarget.IsValid()) {
EnqueueStartQueuedSeekTask();
}
return NS_OK;
}
void MediaDecoderStateMachine::DecodeSeek()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
if (mState != DECODER_STATE_SEEKING ||
!mSeekTarget.IsValid()) {
DECODER_LOG("Early returning from DecodeSeek");
return;
}
// If there's already an existing seek in progress, we need to handle that.
if (mCurrentSeekTarget.IsValid()) {
// There are 3 states we might be in, listed in the order that they occur:
// (1) Waiting for the seek to be resolved.
// (2) Waiting for the seek to be resolved, having already issued a cancel.
// (3) After seek resolution, waiting for SeekComplete to run.
//
// If we're in the first state, we move to the second. Otherwise, we just wait
// for things to sort themselves out.
if (mWaitingForDecoderSeek && !mCancelingSeek) {
mReader->CancelSeek();
mCancelingSeek = true;
}
return;
}
mCurrentSeekTarget = mSeekTarget;
mSeekTarget.Reset();
mDropAudioUntilNextDiscontinuity = HasAudio();
mDropVideoUntilNextDiscontinuity = HasVideo();
// During the seek, don't have a lock on the decoder state,
// otherwise long seek operations can block the main thread.
// The events dispatched to the main thread are SYNC calls.
// These calls are made outside of the decode monitor lock so
// it is safe for the main thread to makes calls that acquire
// the lock since it won't deadlock. We check the state when
// acquiring the lock again in case shutdown has occurred
// during the time when we didn't have the lock.
int64_t seekTime = mCurrentSeekTarget.mTime;
mDecoder->StopProgressUpdates();
bool currentTimeChanged = false;
mCurrentTimeBeforeSeek = GetMediaTime();
if (mCurrentTimeBeforeSeek != seekTime) {
currentTimeChanged = true;
// Stop playback now to ensure that while we're outside the monitor
// dispatching SeekingStarted, playback doesn't advance and mess with
// mCurrentFrameTime that we've setting to seekTime here.
StopPlayback();
UpdatePlaybackPositionInternal(seekTime);
}
// SeekingStarted will do a UpdateReadyStateForData which will
// inform the element and its users that we have no frames
// to display
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
nsCOMPtr<nsIRunnable> startEvent =
NS_NewRunnableMethod(mDecoder, &MediaDecoder::SeekingStarted);
NS_DispatchToMainThread(startEvent, NS_DISPATCH_SYNC);
}
if (mState != DECODER_STATE_SEEKING) {
// May have shutdown while we released the monitor.
return;
}
mDecodeToSeekTarget = false;
if (!currentTimeChanged) {
DECODER_LOG("Seek !currentTimeChanged...");
nsresult rv = DecodeTaskQueue()->Dispatch(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::SeekCompleted));
if (NS_FAILED(rv)) {
DecodeError();
}
} else {
// The seek target is different than the current playback position,
// we'll need to seek the playback position, so shutdown our decode
// thread and audio sink.
StopAudioThread();
ResetPlayback();
nsresult res;
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
// We must not hold the state machine monitor while we call into
// the reader, since it could do I/O or deadlock some other way.
res = mReader->ResetDecode();
if (NS_SUCCEEDED(res)) {
mReader->Seek(seekTime, GetEndTime())
->Then(DecodeTaskQueue(), __func__, this,
&MediaDecoderStateMachine::OnSeekCompleted,
&MediaDecoderStateMachine::OnSeekFailed);
}
}
if (NS_FAILED(res)) {
DecodeError();
return;
}
mWaitingForDecoderSeek = true;
}
}
void
MediaDecoderStateMachine::OnSeekCompleted(int64_t aTime)
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
mWaitingForDecoderSeek = false;
mCancelingSeek = false;
// We must decode the first samples of active streams, so we can determine
// the new stream time. So dispatch tasks to do that.
mDecodeToSeekTarget = true;
DispatchDecodeTasksIfNeeded();
}
void
MediaDecoderStateMachine::OnSeekFailed(nsresult aResult)
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
bool wasCanceled = mCancelingSeek;
mWaitingForDecoderSeek = false;
mCancelingSeek = false;
if (NS_FAILED(aResult)) {
DecodeError();
} else if (wasCanceled && mSeekTarget.IsValid() && mState == DECODER_STATE_SEEKING) {
// Try again.
mCurrentSeekTarget = mSeekTarget;
mSeekTarget.Reset();
mReader->Seek(mCurrentSeekTarget.mTime, mEndTime)
->Then(DecodeTaskQueue(), __func__, this,
&MediaDecoderStateMachine::OnSeekCompleted,
&MediaDecoderStateMachine::OnSeekFailed);
mWaitingForDecoderSeek = true;
}
}
void
MediaDecoderStateMachine::SeekCompleted()
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
// We must reset the seek target when exiting this function, but not
// before, as if we dropped the monitor in any function called here,
// we may begin a new seek on the state machine thread, and be in
// an inconsistent state.
AutoSetOnScopeExit<SeekTarget> reset(mCurrentSeekTarget, SeekTarget());
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
if (mState != DECODER_STATE_SEEKING) {
return;
}
int64_t seekTime = mCurrentSeekTarget.mTime;
int64_t newCurrentTime = mCurrentSeekTarget.mTime;
// Setup timestamp state.
VideoData* video = VideoQueue().PeekFront();
if (seekTime == mEndTime) {
newCurrentTime = mAudioStartTime = seekTime;
} else if (HasAudio()) {
AudioData* audio = AudioQueue().PeekFront();
newCurrentTime = mAudioStartTime = audio ? audio->mTime : seekTime;
} else {
newCurrentTime = video ? video->mTime : seekTime;
}
mPlayDuration = newCurrentTime - mStartTime;
if (HasVideo()) {
if (video) {
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
RenderVideoFrame(video, TimeStamp::Now());
}
nsCOMPtr<nsIRunnable> event =
NS_NewRunnableMethod(mDecoder, &MediaDecoder::Invalidate);
NS_DispatchToMainThread(event, NS_DISPATCH_NORMAL);
}
}
MOZ_ASSERT(mState != DECODER_STATE_DECODING_NONE);
mDecoder->StartProgressUpdates();
if (mState == DECODER_STATE_DECODING_METADATA ||
mState == DECODER_STATE_DECODING_FIRSTFRAME ||
mState == DECODER_STATE_DORMANT ||
mState == DECODER_STATE_SHUTDOWN) {
return;
}
// Change state to DECODING or COMPLETED now. SeekingStopped will
// call MediaDecoderStateMachine::Seek to reset our state to SEEKING
// if we need to seek again.
nsCOMPtr<nsIRunnable> stopEvent;
bool isLiveStream = mDecoder->GetResource()->GetLength() == -1;
if (mSeekTarget.IsValid()) {
// A new seek target came in while we were processing the old one. No rest
// for the seeking.
DECODER_LOG("A new seek came along while we were finishing the old one - staying in SEEKING");
SetState(DECODER_STATE_SEEKING);
} else if (GetMediaTime() == mEndTime && !isLiveStream) {
// Seeked to end of media, move to COMPLETED state. Note we don't do
// this if we're playing a live stream, since the end of media will advance
// once we download more data!
DECODER_LOG("Changed state from SEEKING (to %lld) to COMPLETED", seekTime);
stopEvent = NS_NewRunnableMethod(mDecoder, &MediaDecoder::SeekingStoppedAtEnd);
// Explicitly set our state so we don't decode further, and so
// we report playback ended to the media element.
SetState(DECODER_STATE_COMPLETED);
DispatchDecodeTasksIfNeeded();
} else {
DECODER_LOG("Changed state from SEEKING (to %lld) to DECODING", seekTime);
stopEvent = NS_NewRunnableMethod(mDecoder, &MediaDecoder::SeekingStopped);
StartDecoding();
}
// Ensure timestamps are up to date.
UpdatePlaybackPositionInternal(newCurrentTime);
if (mDecoder->GetDecodedStream()) {
SetSyncPointForMediaStream();
}
// Try to decode another frame to detect if we're at the end...
DECODER_LOG("Seek completed, mCurrentFrameTime=%lld", mCurrentFrameTime);
mCurrentSeekTarget = SeekTarget();
// Reset quick buffering status. This ensures that if we began the
// seek while quick-buffering, we won't bypass quick buffering mode
// if we need to buffer after the seek.
mQuickBuffering = false;
// Prevent changes in playback position before 'seeked' is fired for we
// expect currentTime equals seek target in 'seeked' callback.
mScheduler->FreezeScheduling();
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
NS_DispatchToMainThread(stopEvent, NS_DISPATCH_SYNC);
}
ScheduleStateMachine();
mScheduler->ThawScheduling();
}
// Runnable to dispose of the decoder and state machine on the main thread.
class nsDecoderDisposeEvent : public nsRunnable {
public:
nsDecoderDisposeEvent(already_AddRefed<MediaDecoder> aDecoder,
already_AddRefed<MediaDecoderStateMachine> aStateMachine)
: mDecoder(aDecoder), mStateMachine(aStateMachine) {}
NS_IMETHOD Run() {
NS_ASSERTION(NS_IsMainThread(), "Must be on main thread.");
MOZ_ASSERT(mStateMachine);
MOZ_ASSERT(mDecoder);
mStateMachine->BreakCycles();
mDecoder->BreakCycles();
mStateMachine = nullptr;
mDecoder = nullptr;
return NS_OK;
}
private:
nsRefPtr<MediaDecoder> mDecoder;
nsRefPtr<MediaDecoderStateMachine> mStateMachine;
};
// Runnable which dispatches an event to the main thread to dispose of the
// decoder and state machine. This runs on the state machine thread after
// the state machine has shutdown, and all events for that state machine have
// finished running.
class nsDispatchDisposeEvent : public nsRunnable {
public:
nsDispatchDisposeEvent(MediaDecoder* aDecoder,
MediaDecoderStateMachine* aStateMachine)
: mDecoder(aDecoder), mStateMachine(aStateMachine) {}
NS_IMETHOD Run() {
NS_DispatchToMainThread(new nsDecoderDisposeEvent(mDecoder.forget(),
mStateMachine.forget()));
return NS_OK;
}
private:
nsRefPtr<MediaDecoder> mDecoder;
nsRefPtr<MediaDecoderStateMachine> mStateMachine;
};
void
MediaDecoderStateMachine::ShutdownReader()
{
MOZ_ASSERT(OnDecodeThread());
mReader->Shutdown()->Then(GetStateMachineThread(), __func__, this,
&MediaDecoderStateMachine::FinishShutdown,
&MediaDecoderStateMachine::FinishShutdown);
}
void
MediaDecoderStateMachine::FinishShutdown()
{
MOZ_ASSERT(OnStateMachineThread());
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
// The reader's listeners hold references to the state machine,
// creating a cycle which keeps the state machine and its shared
// thread pools alive. So break it here.
AudioQueue().ClearListeners();
VideoQueue().ClearListeners();
// Now that those threads are stopped, there's no possibility of
// mPendingWakeDecoder being needed again. Revoke it.
mPendingWakeDecoder = nullptr;
MOZ_ASSERT(mState == DECODER_STATE_SHUTDOWN,
"How did we escape from the shutdown state?");
// We must daisy-chain these events to destroy the decoder. We must
// destroy the decoder on the main thread, but we can't destroy the
// decoder while this thread holds the decoder monitor. We can't
// dispatch an event to the main thread to destroy the decoder from
// here, as the event may run before the dispatch returns, and we
// hold the decoder monitor here. We also want to guarantee that the
// state machine is destroyed on the main thread, and so the
// event runner running this function (which holds a reference to the
// state machine) needs to finish and be released in order to allow
// that. So we dispatch an event to run after this event runner has
// finished and released its monitor/references. That event then will
// dispatch an event to the main thread to release the decoder and
// state machine.
GetStateMachineThread()->Dispatch(
new nsDispatchDisposeEvent(mDecoder, this), NS_DISPATCH_NORMAL);
DECODER_LOG("Dispose Event Dispatched");
}
nsresult MediaDecoderStateMachine::RunStateMachine()
{
AssertCurrentThreadInMonitor();
MediaResource* resource = mDecoder->GetResource();
NS_ENSURE_TRUE(resource, NS_ERROR_NULL_POINTER);
switch (mState) {
case DECODER_STATE_SHUTDOWN: {
if (IsPlaying()) {
StopPlayback();
}
StopAudioThread();
FlushDecoding();
// Put a task in the decode queue to shutdown the reader.
// the queue to spin down.
RefPtr<nsIRunnable> task;
task = NS_NewRunnableMethod(this, &MediaDecoderStateMachine::ShutdownReader);
DebugOnly<nsresult> rv = DecodeTaskQueue()->Dispatch(task);
MOZ_ASSERT(NS_SUCCEEDED(rv));
DECODER_LOG("Shutdown started");
return NS_OK;
}
case DECODER_STATE_DORMANT: {
if (IsPlaying()) {
StopPlayback();
}
StopAudioThread();
FlushDecoding();
// Now that those threads are stopped, there's no possibility of
// mPendingWakeDecoder being needed again. Revoke it.
mPendingWakeDecoder = nullptr;
{
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
// Wait for the thread decoding, if any, to exit.
DecodeTaskQueue()->AwaitIdle();
mReader->ReleaseMediaResources();
}
return NS_OK;
}
case DECODER_STATE_WAIT_FOR_RESOURCES: {
return NS_OK;
}
case DECODER_STATE_DECODING_NONE: {
SetState(DECODER_STATE_DECODING_METADATA);
// Ensure we have a decode thread to decode metadata.
return EnqueueDecodeMetadataTask();
}
case DECODER_STATE_DECODING_METADATA: {
return NS_OK;
}
case DECODER_STATE_DECODING_FIRSTFRAME: {
// DECODER_STATE_DECODING_FIRSTFRAME will be started by DecodeMetadata
return NS_OK;
}
case DECODER_STATE_DECODING: {
if (mDecoder->GetState() != MediaDecoder::PLAY_STATE_PLAYING &&
IsPlaying())
{
// We're playing, but the element/decoder is in paused state. Stop
// playing!
StopPlayback();
}
// Start playback if necessary so that the clock can be properly queried.
MaybeStartPlayback();
AdvanceFrame();
NS_ASSERTION(mDecoder->GetState() != MediaDecoder::PLAY_STATE_PLAYING ||
IsStateMachineScheduled() ||
mPlaybackRate == 0.0, "Must have timer scheduled");
return NS_OK;
}
case DECODER_STATE_BUFFERING: {
TimeStamp now = TimeStamp::Now();
NS_ASSERTION(!mBufferingStart.IsNull(), "Must know buffering start time.");
// With buffering heuristics we will remain in the buffering state if
// we've not decoded enough data to begin playback, or if we've not
// downloaded a reasonable amount of data inside our buffering time.
if (mReader->UseBufferingHeuristics()) {
TimeDuration elapsed = now - mBufferingStart;
bool isLiveStream = resource->GetLength() == -1;
if ((isLiveStream || !mDecoder->CanPlayThrough()) &&
elapsed < TimeDuration::FromSeconds(mBufferingWait * mPlaybackRate) &&
(mQuickBuffering ? HasLowDecodedData(mQuickBufferingLowDataThresholdUsecs)
: HasLowUndecodedData(mBufferingWait * USECS_PER_S)) &&
mDecoder->IsExpectingMoreData())
{
DECODER_LOG("Buffering: wait %ds, timeout in %.3lfs %s",
mBufferingWait, mBufferingWait - elapsed.ToSeconds(),
(mQuickBuffering ? "(quick exit)" : ""));
ScheduleStateMachine(USECS_PER_S);
return NS_OK;
}
} else if (OutOfDecodedAudio() || OutOfDecodedVideo()) {
MOZ_ASSERT(mReader->IsWaitForDataSupported(),
"Don't yet have a strategy for non-heuristic + non-WaitForData");
DispatchDecodeTasksIfNeeded();
MOZ_ASSERT_IF(!mMinimizePreroll && OutOfDecodedAudio(), mAudioRequestStatus != RequestStatus::Idle);
MOZ_ASSERT_IF(!mMinimizePreroll && OutOfDecodedVideo(), mVideoRequestStatus != RequestStatus::Idle);
DECODER_LOG("In buffering mode, waiting to be notified: outOfAudio: %d, "
"mAudioStatus: %d, outOfVideo: %d, mVideoStatus: %d",
OutOfDecodedAudio(), mAudioRequestStatus,
OutOfDecodedVideo(), mVideoRequestStatus);
return NS_OK;
}
DECODER_LOG("Changed state from BUFFERING to DECODING");
DECODER_LOG("Buffered for %.3lfs", (now - mBufferingStart).ToSeconds());
StartDecoding();
// Notify to allow blocked decoder thread to continue
mDecoder->GetReentrantMonitor().NotifyAll();
UpdateReadyState();
MaybeStartPlayback();
NS_ASSERTION(IsStateMachineScheduled(), "Must have timer scheduled");
return NS_OK;
}
case DECODER_STATE_SEEKING: {
return EnqueueDecodeSeekTask();
}
case DECODER_STATE_COMPLETED: {
// Play the remaining media. We want to run AdvanceFrame() at least
// once to ensure the current playback position is advanced to the
// end of the media, and so that we update the readyState.
if (VideoQueue().GetSize() > 0 ||
(HasAudio() && !mAudioCompleted) ||
(mDecoder->GetDecodedStream() && !mDecoder->GetDecodedStream()->IsFinished()))
{
AdvanceFrame();
NS_ASSERTION(mDecoder->GetState() != MediaDecoder::PLAY_STATE_PLAYING ||
mPlaybackRate == 0 ||
IsStateMachineScheduled(),
"Must have timer scheduled");
return NS_OK;
}
// StopPlayback in order to reset the IsPlaying() state so audio
// is restarted correctly.
StopPlayback();
if (mState != DECODER_STATE_COMPLETED) {
// While we're presenting a frame we can change state. Whatever changed
// our state should have scheduled another state machine run.
NS_ASSERTION(IsStateMachineScheduled(), "Must have timer scheduled");
return NS_OK;
}
StopAudioThread();
// When we're decoding to a stream, the stream's main-thread finish signal
// will take care of calling MediaDecoder::PlaybackEnded.
if (mDecoder->GetState() == MediaDecoder::PLAY_STATE_PLAYING &&
!mDecoder->GetDecodedStream()) {
int64_t clockTime = std::max(mAudioEndTime, mVideoFrameEndTime);
clockTime = std::max(int64_t(0), std::max(clockTime, mEndTime));
UpdatePlaybackPosition(clockTime);
{
// Wait for the state change is completed in the main thread,
// otherwise we might see |mDecoder->GetState() == MediaDecoder::PLAY_STATE_PLAYING|
// in next loop and send |MediaDecoder::PlaybackEnded| again to trigger 'ended'
// event twice in the media element.
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
nsCOMPtr<nsIRunnable> event =
NS_NewRunnableMethod(mDecoder, &MediaDecoder::PlaybackEnded);
NS_DispatchToMainThread(event, NS_DISPATCH_SYNC);
}
}
return NS_OK;
}
}
return NS_OK;
}
void
MediaDecoderStateMachine::FlushDecoding()
{
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
AssertCurrentThreadInMonitor();
{
// Put a task in the decode queue to abort any decoding operations.
// The reader is not supposed to put any tasks to deliver samples into
// the queue after this runs (unless we request another sample from it).
RefPtr<nsIRunnable> task;
task = NS_NewRunnableMethod(mReader, &MediaDecoderReader::ResetDecode);
// Wait for the ResetDecode to run and for the decoder to abort
// decoding operations and run any pending callbacks. This is
// important, as we don't want any pending tasks posted to the task
// queue by the reader to deliver any samples after we've posted the
// reader Shutdown() task below, as the sample-delivery tasks will
// keep video frames alive until after we've called Reader::Shutdown(),
// and shutdown on B2G will fail as there are outstanding video frames
// alive.
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
DecodeTaskQueue()->Dispatch(task);
DecodeTaskQueue()->AwaitIdle();
}
// We must reset playback so that all references to frames queued
// in the state machine are dropped, else subsequent calls to Shutdown()
// or ReleaseMediaResources() can fail on B2G.
ResetPlayback();
}
void MediaDecoderStateMachine::RenderVideoFrame(VideoData* aData,
TimeStamp aTarget)
{
NS_ASSERTION(OnStateMachineThread() || OnDecodeThread(),
"Should be on state machine or decode thread.");
mDecoder->GetReentrantMonitor().AssertNotCurrentThreadIn();
if (aData->mDuplicate) {
return;
}
VERBOSE_LOG("playing video frame %lld (queued=%i, state-machine=%i, decoder-queued=%i)",
aData->mTime, VideoQueue().GetSize() + mReader->SizeOfVideoQueueInFrames(),
VideoQueue().GetSize(), mReader->SizeOfVideoQueueInFrames());
VideoFrameContainer* container = mDecoder->GetVideoFrameContainer();
if (container) {
container->SetCurrentFrame(ThebesIntSize(aData->mDisplay), aData->mImage,
aTarget);
MOZ_ASSERT(container->GetFrameDelay() >= 0 || IsRealTime());
}
}
void MediaDecoderStateMachine::ResyncAudioClock()
{
AssertCurrentThreadInMonitor();
if (IsPlaying()) {
SetPlayStartTime(TimeStamp::Now());
mPlayDuration = GetAudioClock() - mStartTime;
}
}
int64_t
MediaDecoderStateMachine::GetAudioClock() const
{
// We must hold the decoder monitor while using the audio stream off the
// audio sink to ensure that it doesn't get destroyed on the audio sink
// while we're using it.
AssertCurrentThreadInMonitor();
MOZ_ASSERT(HasAudio() && !mAudioCaptured);
return mAudioStartTime +
(mAudioSink ? mAudioSink->GetPosition() : 0);
}
int64_t MediaDecoderStateMachine::GetVideoStreamPosition() const
{
AssertCurrentThreadInMonitor();
if (!IsPlaying()) {
return mPlayDuration + mStartTime;
}
// Time elapsed since we started playing.
int64_t delta = DurationToUsecs(TimeStamp::Now() - mPlayStartTime);
// Take playback rate into account.
delta *= mPlaybackRate;
return mStartTime + mPlayDuration + delta;
}
int64_t MediaDecoderStateMachine::GetClock() const
{
AssertCurrentThreadInMonitor();
// Determine the clock time. If we've got audio, and we've not reached
// the end of the audio, use the audio clock. However if we've finished
// audio, or don't have audio, use the system clock. If our output is being
// fed to a MediaStream, use that stream as the source of the clock.
int64_t clock_time = -1;
if (!IsPlaying()) {
clock_time = mPlayDuration + mStartTime;
} else {
if (mDecoder->GetDecodedStream()) {
clock_time = GetCurrentTimeViaMediaStreamSync();
} else if (HasAudio() && !mAudioCompleted && !mAudioCaptured) {
clock_time = GetAudioClock();
} else {
// Audio is disabled on this system. Sync to the system clock.
clock_time = GetVideoStreamPosition();
}
// Ensure the clock can never go backwards.
NS_ASSERTION(GetMediaTime() <= clock_time || mPlaybackRate <= 0,
"Clock should go forwards if the playback rate is > 0.");
}
return clock_time;
}
void MediaDecoderStateMachine::AdvanceFrame()
{
NS_ASSERTION(OnStateMachineThread(), "Should be on state machine thread.");
AssertCurrentThreadInMonitor();
NS_ASSERTION(!HasAudio() || mAudioStartTime != -1,
"Should know audio start time if we have audio.");
if (mDecoder->GetState() != MediaDecoder::PLAY_STATE_PLAYING) {
return;
}
// If playbackRate is 0.0, we should stop the progress, but not be in paused
// state, per spec.
if (mPlaybackRate == 0.0) {
return;
}
DecodedStreamData* stream = mDecoder->GetDecodedStream();
if (stream && !stream->mStreamInitialized) {
// Output streams exist but are not initialized yet.
// Send the data we already have to allow stream clock to progress and
// avoid stalling playback.
SendStreamData();
}
const int64_t clock_time = GetClock();
TimeStamp nowTime = TimeStamp::Now();
// Skip frames up to the frame at the playback position, and figure out
// the time remaining until it's time to display the next frame.
int64_t remainingTime = AUDIO_DURATION_USECS;
NS_ASSERTION(clock_time >= mStartTime, "Should have positive clock time.");
nsRefPtr<VideoData> currentFrame;
if (VideoQueue().GetSize() > 0) {
VideoData* frame = VideoQueue().PeekFront();
#ifdef PR_LOGGING
int32_t droppedFrames = 0;
#endif
while (IsRealTime() || clock_time >= frame->mTime) {
mVideoFrameEndTime = frame->GetEndTime();
#ifdef PR_LOGGING
if (currentFrame) {
VERBOSE_LOG("discarding video frame mTime=%lld clock_time=%lld (%d so far)",
currentFrame->mTime, clock_time, ++droppedFrames);
}
#endif
currentFrame = frame;
nsRefPtr<VideoData> releaseMe = VideoQueue().PopFront();
// Notify the decode thread that the video queue's buffers may have
// free'd up space for more frames.
mDecoder->GetReentrantMonitor().NotifyAll();
OnPlaybackOffsetUpdate(frame->mOffset);
if (VideoQueue().GetSize() == 0)
break;
frame = VideoQueue().PeekFront();
}
// Current frame has already been presented, wait until it's time to
// present the next frame.
if (frame && !currentFrame) {
int64_t now = IsPlaying() ? clock_time : mStartTime + mPlayDuration;
remainingTime = frame->mTime - now;
}
}
// Check to see if we don't have enough data to play up to the next frame.
// If we don't, switch to buffering mode.
if (mState == DECODER_STATE_DECODING &&
mDecoder->GetState() == MediaDecoder::PLAY_STATE_PLAYING &&
mDecoder->IsExpectingMoreData()) {
bool shouldBuffer;
if (mReader->UseBufferingHeuristics()) {
shouldBuffer = HasLowDecodedData(remainingTime + EXHAUSTED_DATA_MARGIN_USECS) &&
(JustExitedQuickBuffering() || HasLowUndecodedData());
} else {
MOZ_ASSERT(mReader->IsWaitForDataSupported());
shouldBuffer = (OutOfDecodedAudio() && mAudioRequestStatus == RequestStatus::Waiting) ||
(OutOfDecodedVideo() && mVideoRequestStatus == RequestStatus::Waiting);
}
if (shouldBuffer) {
if (currentFrame) {
VideoQueue().PushFront(currentFrame);
}
StartBuffering();
// Don't go straight back to the state machine loop since that might
// cause us to start decoding again and we could flip-flop between
// decoding and quick-buffering.
ScheduleStateMachine(USECS_PER_S);
return;
}
}
// We've got enough data to keep playing until at least the next frame.
// Start playing now if need be.
if ((mFragmentEndTime >= 0 && clock_time < mFragmentEndTime) || mFragmentEndTime < 0) {
MaybeStartPlayback();
}
if (currentFrame) {
// Decode one frame and display it.
int64_t delta = currentFrame->mTime - clock_time;
TimeStamp presTime = nowTime + TimeDuration::FromMicroseconds(delta / mPlaybackRate);
NS_ASSERTION(currentFrame->mTime >= mStartTime, "Should have positive frame time");
// Filter out invalid frames by checking the frame time. FrameTime could be
// zero if it's a initial frame.
int64_t frameTime = currentFrame->mTime - mStartTime;
if (frameTime > 0 || (frameTime == 0 && mPlayDuration == 0) ||
IsRealTime()) {
ReentrantMonitorAutoExit exitMon(mDecoder->GetReentrantMonitor());
// If we have video, we want to increment the clock in steps of the frame
// duration.
RenderVideoFrame(currentFrame, presTime);
}
// If we're no longer playing after dropping and reacquiring the lock,
// playback must've been stopped on the decode thread (by a seek, for
// example). In that case, the current frame is probably out of date.
if (!IsPlaying()) {
ScheduleStateMachine();
return;
}
MediaDecoder::FrameStatistics& frameStats = mDecoder->GetFrameStatistics();
frameStats.NotifyPresentedFrame();
remainingTime = currentFrame->GetEndTime() - clock_time;
currentFrame = nullptr;
}
// Cap the current time to the larger of the audio and video end time.
// This ensures that if we're running off the system clock, we don't
// advance the clock to after the media end time.
if (mVideoFrameEndTime != -1 || mAudioEndTime != -1) {
// These will be non -1 if we've displayed a video frame, or played an audio frame.
int64_t t = std::min(clock_time, std::max(mVideoFrameEndTime, mAudioEndTime));
// FIXME: Bug 1091422 - chained ogg files hit this assertion.
//MOZ_ASSERT(t >= GetMediaTime());
if (t > GetMediaTime()) {
UpdatePlaybackPosition(t);
}
}
// If the number of audio/video frames queued has changed, either by
// this function popping and playing a video frame, or by the audio
// thread popping and playing an audio frame, we may need to update our
// ready state. Post an update to do so.
UpdateReadyState();
ScheduleStateMachine(remainingTime / mPlaybackRate);
}
nsresult
MediaDecoderStateMachine::DropVideoUpToSeekTarget(VideoData* aSample)
{
nsRefPtr<VideoData> video(aSample);
MOZ_ASSERT(video);
DECODER_LOG("DropVideoUpToSeekTarget() frame [%lld, %lld] dup=%d",
video->mTime, video->GetEndTime(), video->mDuplicate);
const int64_t target = mCurrentSeekTarget.mTime;
// Duplicate handling: if we're dropping frames up the seek target, we must
// be wary of Theora duplicate frames. They don't have an image, so if the
// target frame is in a run of duplicates, we won't have an image to draw
// after the seek. So store the last frame encountered while dropping, and
// copy its Image forward onto duplicate frames, so that every frame has
// an Image.
if (video->mDuplicate &&
mFirstVideoFrameAfterSeek &&
!mFirstVideoFrameAfterSeek->mDuplicate) {
nsRefPtr<VideoData> temp =
VideoData::ShallowCopyUpdateTimestampAndDuration(mFirstVideoFrameAfterSeek,
video->mTime,
video->mDuration);
video = temp;
}
// If the frame end time is less than the seek target, we won't want
// to display this frame after the seek, so discard it.
if (target >= video->GetEndTime()) {
DECODER_LOG("DropVideoUpToSeekTarget() pop video frame [%lld, %lld] target=%lld",
video->mTime, video->GetEndTime(), target);
mFirstVideoFrameAfterSeek = video;
} else {
if (target >= video->mTime && video->GetEndTime() >= target) {
// The seek target lies inside this frame's time slice. Adjust the frame's
// start time to match the seek target. We do this by replacing the
// first frame with a shallow copy which has the new timestamp.
nsRefPtr<VideoData> temp = VideoData::ShallowCopyUpdateTimestamp(video, target);
video = temp;
}
mFirstVideoFrameAfterSeek = nullptr;
DECODER_LOG("DropVideoUpToSeekTarget() found video frame [%lld, %lld] containing target=%lld",
video->mTime, video->GetEndTime(), target);
VideoQueue().PushFront(video);
}
return NS_OK;
}
nsresult
MediaDecoderStateMachine::DropAudioUpToSeekTarget(AudioData* aSample)
{
nsRefPtr<AudioData> audio(aSample);
MOZ_ASSERT(audio &&
mCurrentSeekTarget.IsValid() &&
mCurrentSeekTarget.mType == SeekTarget::Accurate);
CheckedInt64 startFrame = UsecsToFrames(audio->mTime,
mInfo.mAudio.mRate);
CheckedInt64 targetFrame = UsecsToFrames(mCurrentSeekTarget.mTime,
mInfo.mAudio.mRate);
if (!startFrame.isValid() || !targetFrame.isValid()) {
return NS_ERROR_FAILURE;
}
if (startFrame.value() + audio->mFrames <= targetFrame.value()) {
// Our seek target lies after the frames in this AudioData. Don't
// push it onto the audio queue, and keep decoding forwards.
return NS_OK;
}
if (startFrame.value() > targetFrame.value()) {
// The seek target doesn't lie in the audio block just after the last
// audio frames we've seen which were before the seek target. This
// could have been the first audio data we've seen after seek, i.e. the
// seek terminated after the seek target in the audio stream. Just
// abort the audio decode-to-target, the state machine will play
// silence to cover the gap. Typically this happens in poorly muxed
// files.
DECODER_WARN("Audio not synced after seek, maybe a poorly muxed file?");
AudioQueue().Push(audio);
return NS_OK;
}
// The seek target lies somewhere in this AudioData's frames, strip off
// any frames which lie before the seek target, so we'll begin playback
// exactly at the seek target.
NS_ASSERTION(targetFrame.value() >= startFrame.value(),
"Target must at or be after data start.");
NS_ASSERTION(targetFrame.value() < startFrame.value() + audio->mFrames,
"Data must end after target.");
int64_t framesToPrune = targetFrame.value() - startFrame.value();
if (framesToPrune > audio->mFrames) {
// We've messed up somehow. Don't try to trim frames, the |frames|
// variable below will overflow.
DECODER_WARN("Can't prune more frames that we have!");
return NS_ERROR_FAILURE;
}
uint32_t frames = audio->mFrames - static_cast<uint32_t>(framesToPrune);
uint32_t channels = audio->mChannels;
nsAutoArrayPtr<AudioDataValue> audioData(new AudioDataValue[frames * channels]);
memcpy(audioData.get(),
audio->mAudioData.get() + (framesToPrune * channels),
frames * channels * sizeof(AudioDataValue));
CheckedInt64 duration = FramesToUsecs(frames, mInfo.mAudio.mRate);
if (!duration.isValid()) {
return NS_ERROR_FAILURE;
}
nsRefPtr<AudioData> data(new AudioData(audio->mOffset,
mCurrentSeekTarget.mTime,
duration.value(),
frames,
audioData.forget(),
channels,
audio->mRate));
AudioQueue().PushFront(data);
return NS_OK;
}
void MediaDecoderStateMachine::SetStartTime(int64_t aStartTimeUsecs)
{
AssertCurrentThreadInMonitor();
DECODER_LOG("SetStartTime(%lld)", aStartTimeUsecs);
mStartTime = 0;
if (aStartTimeUsecs != 0) {
mStartTime = aStartTimeUsecs;
if (mGotDurationFromMetaData && GetEndTime() != INT64_MAX) {
NS_ASSERTION(mEndTime != -1,
"We should have mEndTime as supplied duration here");
// We were specified a duration from a Content-Duration HTTP header.
// Adjust mEndTime so that mEndTime-mStartTime matches the specified
// duration.
mEndTime = mStartTime + mEndTime;
}
}
// Pass along this immutable value to the reader so that it can make
// calculations independently of the state machine.
mReader->SetStartTime(mStartTime);
// Set the audio start time to be start of media. If this lies before the
// first actual audio frame we have, we'll inject silence during playback
// to ensure the audio starts at the correct time.
mAudioStartTime = mStartTime;
DECODER_LOG("Set media start time to %lld", mStartTime);
}
void MediaDecoderStateMachine::UpdateReadyState() {
AssertCurrentThreadInMonitor();
MediaDecoderOwner::NextFrameStatus nextFrameStatus = GetNextFrameStatus();
// FIXME: This optimization could result in inconsistent next frame status
// between the decoder and state machine when GetNextFrameStatus() is called
// by the decoder without updating mLastFrameStatus.
// Note not to regress bug 882027 when fixing this bug.
if (nextFrameStatus == mLastFrameStatus) {
return;
}
mLastFrameStatus = nextFrameStatus;
/* This is a bit tricky. MediaDecoder::UpdateReadyStateForData will run on
* the main thread and re-evaluate GetNextFrameStatus there, passing it to
* HTMLMediaElement::UpdateReadyStateForData. It doesn't use the value of
* GetNextFrameStatus we computed here, because what we're computing here
* could be stale by the time MediaDecoder::UpdateReadyStateForData runs.
* We only compute GetNextFrameStatus here to avoid posting runnables to the main
* thread unnecessarily.
*/
nsCOMPtr<nsIRunnable> event;
event = NS_NewRunnableMethod(mDecoder, &MediaDecoder::UpdateReadyStateForData);
NS_DispatchToMainThread(event);
}
bool MediaDecoderStateMachine::JustExitedQuickBuffering()
{
return !mDecodeStartTime.IsNull() &&
mQuickBuffering &&
(TimeStamp::Now() - mDecodeStartTime) < TimeDuration::FromMicroseconds(QUICK_BUFFER_THRESHOLD_USECS);
}
void MediaDecoderStateMachine::StartBuffering()
{
AssertCurrentThreadInMonitor();
if (mState != DECODER_STATE_DECODING) {
// We only move into BUFFERING state if we're actually decoding.
// If we're currently doing something else, we don't need to buffer,
// and more importantly, we shouldn't overwrite mState to interrupt
// the current operation, as that could leave us in an inconsistent
// state!
return;
}
if (IsPlaying()) {
StopPlayback();
}
TimeDuration decodeDuration = TimeStamp::Now() - mDecodeStartTime;
// Go into quick buffering mode provided we've not just left buffering using
// a "quick exit". This stops us flip-flopping between playing and buffering
// when the download speed is similar to the decode speed.
mQuickBuffering =
!JustExitedQuickBuffering() &&
decodeDuration < UsecsToDuration(QUICK_BUFFER_THRESHOLD_USECS);
mBufferingStart = TimeStamp::Now();
// We need to tell the element that buffering has started.
// We can't just directly send an asynchronous runnable that
// eventually fires the "waiting" event. The problem is that
// there might be pending main-thread events, such as "data
// received" notifications, that mean we're not actually still
// buffering by the time this runnable executes. So instead
// we just trigger UpdateReadyStateForData; when it runs, it
// will check the current state and decide whether to tell
// the element we're buffering or not.
SetState(DECODER_STATE_BUFFERING);
UpdateReadyState();
DECODER_LOG("Changed state from DECODING to BUFFERING, decoded for %.3lfs",
decodeDuration.ToSeconds());
#ifdef PR_LOGGING
MediaDecoder::Statistics stats = mDecoder->GetStatistics();
DECODER_LOG("Playback rate: %.1lfKB/s%s download rate: %.1lfKB/s%s",
stats.mPlaybackRate/1024, stats.mPlaybackRateReliable ? "" : " (unreliable)",
stats.mDownloadRate/1024, stats.mDownloadRateReliable ? "" : " (unreliable)");
#endif
}
void MediaDecoderStateMachine::SetPlayStartTime(const TimeStamp& aTimeStamp)
{
AssertCurrentThreadInMonitor();
mPlayStartTime = aTimeStamp;
if (!mAudioSink) {
return;
}
if (!mPlayStartTime.IsNull()) {
mAudioSink->StartPlayback();
} else {
mAudioSink->StopPlayback();
}
}
nsresult MediaDecoderStateMachine::CallRunStateMachine()
{
AssertCurrentThreadInMonitor();
NS_ASSERTION(OnStateMachineThread(), "Should be on state machine thread.");
// If audio is being captured, stop the audio sink if it's running
if (mAudioCaptured) {
StopAudioThread();
}
return RunStateMachine();
}
nsresult MediaDecoderStateMachine::TimeoutExpired(void* aClosure)
{
MediaDecoderStateMachine* p = static_cast<MediaDecoderStateMachine*>(aClosure);
return p->CallRunStateMachine();
}
void MediaDecoderStateMachine::ScheduleStateMachineWithLockAndWakeDecoder() {
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
DispatchAudioDecodeTaskIfNeeded();
DispatchVideoDecodeTaskIfNeeded();
}
nsresult MediaDecoderStateMachine::ScheduleStateMachine(int64_t aUsecs) {
return mScheduler->Schedule(aUsecs);
}
bool MediaDecoderStateMachine::OnDecodeThread() const
{
return !DecodeTaskQueue() || DecodeTaskQueue()->IsCurrentThreadIn();
}
bool MediaDecoderStateMachine::OnStateMachineThread() const
{
return mScheduler->OnStateMachineThread();
}
nsIEventTarget* MediaDecoderStateMachine::GetStateMachineThread() const
{
return mScheduler->GetStateMachineThread();
}
bool MediaDecoderStateMachine::IsStateMachineScheduled() const
{
return mScheduler->IsScheduled();
}
void MediaDecoderStateMachine::SetPlaybackRate(double aPlaybackRate)
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
NS_ASSERTION(aPlaybackRate != 0,
"PlaybackRate == 0 should be handled before this function.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
if (mPlaybackRate == aPlaybackRate) {
return;
}
// AudioStream will handle playback rate change when we have audio.
// Do nothing while we are not playing. Change in playback rate will
// take effect next time we start playing again.
if (!HasAudio() && IsPlaying()) {
// Remember how much time we've spent in playing the media
// for playback rate will change from now on.
mPlayDuration = GetVideoStreamPosition() - mStartTime;
SetPlayStartTime(TimeStamp::Now());
}
mPlaybackRate = aPlaybackRate;
if (mAudioSink) {
mAudioSink->SetPlaybackRate(mPlaybackRate);
}
}
void MediaDecoderStateMachine::SetPreservesPitch(bool aPreservesPitch)
{
NS_ASSERTION(NS_IsMainThread(), "Should be on main thread.");
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
mPreservesPitch = aPreservesPitch;
if (mAudioSink) {
mAudioSink->SetPreservesPitch(mPreservesPitch);
}
}
void
MediaDecoderStateMachine::SetMinimizePrerollUntilPlaybackStarts()
{
AssertCurrentThreadInMonitor();
DECODER_LOG("SetMinimizePrerollUntilPlaybackStarts()");
mMinimizePreroll = true;
}
bool MediaDecoderStateMachine::IsShutdown()
{
AssertCurrentThreadInMonitor();
return GetState() == DECODER_STATE_SHUTDOWN;
}
void MediaDecoderStateMachine::QueueMetadata(int64_t aPublishTime,
nsAutoPtr<MediaInfo> aInfo,
nsAutoPtr<MetadataTags> aTags)
{
NS_ASSERTION(OnDecodeThread(), "Should be on decode thread.");
AssertCurrentThreadInMonitor();
TimedMetadata* metadata = new TimedMetadata;
metadata->mPublishTime = aPublishTime;
metadata->mInfo = aInfo.forget();
metadata->mTags = aTags.forget();
mMetadataManager.QueueMetadata(metadata);
}
void MediaDecoderStateMachine::OnAudioEndTimeUpdate(int64_t aAudioEndTime)
{
ReentrantMonitorAutoEnter mon(mDecoder->GetReentrantMonitor());
MOZ_ASSERT(aAudioEndTime >= mAudioEndTime);
mAudioEndTime = aAudioEndTime;
}
void MediaDecoderStateMachine::OnPlaybackOffsetUpdate(int64_t aPlaybackOffset)
{
mDecoder->UpdatePlaybackOffset(aPlaybackOffset);
}
void MediaDecoderStateMachine::OnAudioSinkComplete()
{
AssertCurrentThreadInMonitor();
if (mAudioCaptured) {
return;
}
ResyncAudioClock();
mAudioCompleted = true;
UpdateReadyState();
// Kick the decode thread; it may be sleeping waiting for this to finish.
mDecoder->GetReentrantMonitor().NotifyAll();
}
void MediaDecoderStateMachine::OnAudioSinkError()
{
AssertCurrentThreadInMonitor();
// AudioSink not used with captured streams, so ignore errors in this case.
if (mAudioCaptured) {
return;
}
ResyncAudioClock();
mAudioCompleted = true;
// Make the best effort to continue playback when there is video.
if (HasVideo()) {
return;
}
// Otherwise notify media decoder/element about this error for it makes
// no sense to play an audio-only file without sound output.
RefPtr<nsIRunnable> task(
NS_NewRunnableMethod(this, &MediaDecoderStateMachine::AcquireMonitorAndInvokeDecodeError));
nsresult rv = DecodeTaskQueue()->Dispatch(task);
if (NS_FAILED(rv)) {
DECODER_WARN("Failed to dispatch AcquireMonitorAndInvokeDecodeError");
}
}
} // namespace mozilla
// avoid redefined macro in unified build
#undef DECODER_LOG
#undef VERBOSE_LOG
#undef DECODER_WARN
#undef DECODER_WARN_HELPER