webrtc/modules/video_coding/packet_buffer.cc
Johannes Kron a370556270 Refactor to free up PacketBuffer as soon as possible
The packets belonging to a frame were kept in PacketBuffer
until the frame was decoded. This CL clears the dependencies
of an existing RtpFrameObject to PacketBuffer so that we can
free up PacketBuffer as soon as the RtpFrameObject is created.

Bug: none
Change-Id: Ic939be91815519ae1d1c67ada82006417b2d26a3
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/149818
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Johannes Kron <kron@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28977}
2019-08-28 08:07:32 +00:00

548 lines
19 KiB
C++

/*
* Copyright (c) 2016 The WebRTC project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include "modules/video_coding/packet_buffer.h"
#include <string.h>
#include <algorithm>
#include <cstdint>
#include <utility>
#include "absl/types/variant.h"
#include "api/video/encoded_frame.h"
#include "common_video/h264/h264_common.h"
#include "modules/rtp_rtcp/source/rtp_video_header.h"
#include "modules/video_coding/codecs/h264/include/h264_globals.h"
#include "modules/video_coding/frame_object.h"
#include "rtc_base/atomic_ops.h"
#include "rtc_base/checks.h"
#include "rtc_base/logging.h"
#include "rtc_base/numerics/mod_ops.h"
#include "system_wrappers/include/clock.h"
#include "system_wrappers/include/field_trial.h"
namespace webrtc {
namespace video_coding {
rtc::scoped_refptr<PacketBuffer> PacketBuffer::Create(
Clock* clock,
size_t start_buffer_size,
size_t max_buffer_size,
OnAssembledFrameCallback* assembled_frame_callback) {
return rtc::scoped_refptr<PacketBuffer>(new PacketBuffer(
clock, start_buffer_size, max_buffer_size, assembled_frame_callback));
}
PacketBuffer::PacketBuffer(Clock* clock,
size_t start_buffer_size,
size_t max_buffer_size,
OnAssembledFrameCallback* assembled_frame_callback)
: clock_(clock),
size_(start_buffer_size),
max_size_(max_buffer_size),
first_seq_num_(0),
first_packet_received_(false),
is_cleared_to_first_seq_num_(false),
data_buffer_(start_buffer_size),
sequence_buffer_(start_buffer_size),
assembled_frame_callback_(assembled_frame_callback),
unique_frames_seen_(0),
sps_pps_idr_is_h264_keyframe_(
field_trial::IsEnabled("WebRTC-SpsPpsIdrIsH264Keyframe")) {
RTC_DCHECK_LE(start_buffer_size, max_buffer_size);
// Buffer size must always be a power of 2.
RTC_DCHECK((start_buffer_size & (start_buffer_size - 1)) == 0);
RTC_DCHECK((max_buffer_size & (max_buffer_size - 1)) == 0);
}
PacketBuffer::~PacketBuffer() {
Clear();
}
bool PacketBuffer::InsertPacket(VCMPacket* packet) {
std::vector<std::unique_ptr<RtpFrameObject>> found_frames;
{
rtc::CritScope lock(&crit_);
OnTimestampReceived(packet->timestamp);
uint16_t seq_num = packet->seqNum;
size_t index = seq_num % size_;
if (!first_packet_received_) {
first_seq_num_ = seq_num;
first_packet_received_ = true;
} else if (AheadOf(first_seq_num_, seq_num)) {
// If we have explicitly cleared past this packet then it's old,
// don't insert it, just silently ignore it.
if (is_cleared_to_first_seq_num_) {
delete[] packet->dataPtr;
packet->dataPtr = nullptr;
return true;
}
first_seq_num_ = seq_num;
}
if (sequence_buffer_[index].used) {
// Duplicate packet, just delete the payload.
if (data_buffer_[index].seqNum == packet->seqNum) {
delete[] packet->dataPtr;
packet->dataPtr = nullptr;
return true;
}
// The packet buffer is full, try to expand the buffer.
while (ExpandBufferSize() && sequence_buffer_[seq_num % size_].used) {
}
index = seq_num % size_;
// Packet buffer is still full since we were unable to expand the buffer.
if (sequence_buffer_[index].used) {
// Clear the buffer, delete payload, and return false to signal that a
// new keyframe is needed.
RTC_LOG(LS_WARNING) << "Clear PacketBuffer and request key frame.";
Clear();
delete[] packet->dataPtr;
packet->dataPtr = nullptr;
return false;
}
}
sequence_buffer_[index].frame_begin = packet->is_first_packet_in_frame();
sequence_buffer_[index].frame_end = packet->is_last_packet_in_frame();
sequence_buffer_[index].seq_num = packet->seqNum;
sequence_buffer_[index].continuous = false;
sequence_buffer_[index].frame_created = false;
sequence_buffer_[index].used = true;
data_buffer_[index] = *packet;
packet->dataPtr = nullptr;
UpdateMissingPackets(packet->seqNum);
int64_t now_ms = clock_->TimeInMilliseconds();
last_received_packet_ms_ = now_ms;
if (packet->video_header.frame_type == VideoFrameType::kVideoFrameKey)
last_received_keyframe_packet_ms_ = now_ms;
found_frames = FindFrames(seq_num);
}
for (std::unique_ptr<RtpFrameObject>& frame : found_frames)
assembled_frame_callback_->OnAssembledFrame(std::move(frame));
return true;
}
void PacketBuffer::ClearTo(uint16_t seq_num) {
rtc::CritScope lock(&crit_);
// We have already cleared past this sequence number, no need to do anything.
if (is_cleared_to_first_seq_num_ &&
AheadOf<uint16_t>(first_seq_num_, seq_num)) {
return;
}
// If the packet buffer was cleared between a frame was created and returned.
if (!first_packet_received_)
return;
// Avoid iterating over the buffer more than once by capping the number of
// iterations to the |size_| of the buffer.
++seq_num;
size_t diff = ForwardDiff<uint16_t>(first_seq_num_, seq_num);
size_t iterations = std::min(diff, size_);
for (size_t i = 0; i < iterations; ++i) {
size_t index = first_seq_num_ % size_;
RTC_DCHECK_EQ(data_buffer_[index].seqNum, sequence_buffer_[index].seq_num);
if (AheadOf<uint16_t>(seq_num, sequence_buffer_[index].seq_num)) {
delete[] data_buffer_[index].dataPtr;
data_buffer_[index].dataPtr = nullptr;
sequence_buffer_[index].used = false;
}
++first_seq_num_;
}
// If |diff| is larger than |iterations| it means that we don't increment
// |first_seq_num_| until we reach |seq_num|, so we set it here.
first_seq_num_ = seq_num;
is_cleared_to_first_seq_num_ = true;
auto clear_to_it = missing_packets_.upper_bound(seq_num);
if (clear_to_it != missing_packets_.begin()) {
--clear_to_it;
missing_packets_.erase(missing_packets_.begin(), clear_to_it);
}
}
void PacketBuffer::ClearInterval(uint16_t start_seq_num,
uint16_t stop_seq_num) {
size_t iterations = ForwardDiff<uint16_t>(start_seq_num, stop_seq_num + 1);
RTC_DCHECK_LE(iterations, size_);
uint16_t seq_num = start_seq_num;
for (size_t i = 0; i < iterations; ++i) {
size_t index = seq_num % size_;
RTC_DCHECK_EQ(sequence_buffer_[index].seq_num, seq_num);
RTC_DCHECK_EQ(sequence_buffer_[index].seq_num, data_buffer_[index].seqNum);
delete[] data_buffer_[index].dataPtr;
data_buffer_[index].dataPtr = nullptr;
sequence_buffer_[index].used = false;
++seq_num;
}
}
void PacketBuffer::Clear() {
rtc::CritScope lock(&crit_);
for (size_t i = 0; i < size_; ++i) {
delete[] data_buffer_[i].dataPtr;
data_buffer_[i].dataPtr = nullptr;
sequence_buffer_[i].used = false;
}
first_packet_received_ = false;
is_cleared_to_first_seq_num_ = false;
last_received_packet_ms_.reset();
last_received_keyframe_packet_ms_.reset();
newest_inserted_seq_num_.reset();
missing_packets_.clear();
}
void PacketBuffer::PaddingReceived(uint16_t seq_num) {
std::vector<std::unique_ptr<RtpFrameObject>> found_frames;
{
rtc::CritScope lock(&crit_);
UpdateMissingPackets(seq_num);
found_frames = FindFrames(static_cast<uint16_t>(seq_num + 1));
}
for (std::unique_ptr<RtpFrameObject>& frame : found_frames)
assembled_frame_callback_->OnAssembledFrame(std::move(frame));
}
absl::optional<int64_t> PacketBuffer::LastReceivedPacketMs() const {
rtc::CritScope lock(&crit_);
return last_received_packet_ms_;
}
absl::optional<int64_t> PacketBuffer::LastReceivedKeyframePacketMs() const {
rtc::CritScope lock(&crit_);
return last_received_keyframe_packet_ms_;
}
int PacketBuffer::GetUniqueFramesSeen() const {
rtc::CritScope lock(&crit_);
return unique_frames_seen_;
}
bool PacketBuffer::ExpandBufferSize() {
if (size_ == max_size_) {
RTC_LOG(LS_WARNING) << "PacketBuffer is already at max size (" << max_size_
<< "), failed to increase size.";
return false;
}
size_t new_size = std::min(max_size_, 2 * size_);
std::vector<VCMPacket> new_data_buffer(new_size);
std::vector<ContinuityInfo> new_sequence_buffer(new_size);
for (size_t i = 0; i < size_; ++i) {
if (sequence_buffer_[i].used) {
size_t index = sequence_buffer_[i].seq_num % new_size;
new_sequence_buffer[index] = sequence_buffer_[i];
new_data_buffer[index] = data_buffer_[i];
}
}
size_ = new_size;
sequence_buffer_ = std::move(new_sequence_buffer);
data_buffer_ = std::move(new_data_buffer);
RTC_LOG(LS_INFO) << "PacketBuffer size expanded to " << new_size;
return true;
}
bool PacketBuffer::PotentialNewFrame(uint16_t seq_num) const {
size_t index = seq_num % size_;
int prev_index = index > 0 ? index - 1 : size_ - 1;
if (!sequence_buffer_[index].used)
return false;
if (sequence_buffer_[index].seq_num != seq_num)
return false;
if (sequence_buffer_[index].frame_created)
return false;
if (sequence_buffer_[index].frame_begin)
return true;
if (!sequence_buffer_[prev_index].used)
return false;
if (sequence_buffer_[prev_index].frame_created)
return false;
if (sequence_buffer_[prev_index].seq_num !=
static_cast<uint16_t>(sequence_buffer_[index].seq_num - 1)) {
return false;
}
if (data_buffer_[prev_index].timestamp != data_buffer_[index].timestamp)
return false;
if (sequence_buffer_[prev_index].continuous)
return true;
return false;
}
std::vector<std::unique_ptr<RtpFrameObject>> PacketBuffer::FindFrames(
uint16_t seq_num) {
std::vector<std::unique_ptr<RtpFrameObject>> found_frames;
for (size_t i = 0; i < size_ && PotentialNewFrame(seq_num); ++i) {
size_t index = seq_num % size_;
sequence_buffer_[index].continuous = true;
// If all packets of the frame is continuous, find the first packet of the
// frame and create an RtpFrameObject.
if (sequence_buffer_[index].frame_end) {
size_t frame_size = 0;
int max_nack_count = -1;
uint16_t start_seq_num = seq_num;
int64_t min_recv_time = data_buffer_[index].packet_info.receive_time_ms();
int64_t max_recv_time = data_buffer_[index].packet_info.receive_time_ms();
RtpPacketInfos::vector_type packet_infos;
// Find the start index by searching backward until the packet with
// the |frame_begin| flag is set.
int start_index = index;
size_t tested_packets = 0;
int64_t frame_timestamp = data_buffer_[start_index].timestamp;
// Identify H.264 keyframes by means of SPS, PPS, and IDR.
bool is_h264 = data_buffer_[start_index].codec() == kVideoCodecH264;
bool has_h264_sps = false;
bool has_h264_pps = false;
bool has_h264_idr = false;
bool is_h264_keyframe = false;
while (true) {
++tested_packets;
frame_size += data_buffer_[start_index].sizeBytes;
max_nack_count =
std::max(max_nack_count, data_buffer_[start_index].timesNacked);
sequence_buffer_[start_index].frame_created = true;
min_recv_time =
std::min(min_recv_time,
data_buffer_[start_index].packet_info.receive_time_ms());
max_recv_time =
std::max(max_recv_time,
data_buffer_[start_index].packet_info.receive_time_ms());
// Should use |push_front()| since the loop traverses backwards. But
// it's too inefficient to do so on a vector so we'll instead fix the
// order afterwards.
packet_infos.push_back(data_buffer_[start_index].packet_info);
if (!is_h264 && sequence_buffer_[start_index].frame_begin)
break;
if (is_h264 && !is_h264_keyframe) {
const auto* h264_header = absl::get_if<RTPVideoHeaderH264>(
&data_buffer_[start_index].video_header.video_type_header);
if (!h264_header || h264_header->nalus_length >= kMaxNalusPerPacket)
return found_frames;
for (size_t j = 0; j < h264_header->nalus_length; ++j) {
if (h264_header->nalus[j].type == H264::NaluType::kSps) {
has_h264_sps = true;
} else if (h264_header->nalus[j].type == H264::NaluType::kPps) {
has_h264_pps = true;
} else if (h264_header->nalus[j].type == H264::NaluType::kIdr) {
has_h264_idr = true;
}
}
if ((sps_pps_idr_is_h264_keyframe_ && has_h264_idr && has_h264_sps &&
has_h264_pps) ||
(!sps_pps_idr_is_h264_keyframe_ && has_h264_idr)) {
is_h264_keyframe = true;
}
}
if (tested_packets == size_)
break;
start_index = start_index > 0 ? start_index - 1 : size_ - 1;
// In the case of H264 we don't have a frame_begin bit (yes,
// |frame_begin| might be set to true but that is a lie). So instead
// we traverese backwards as long as we have a previous packet and
// the timestamp of that packet is the same as this one. This may cause
// the PacketBuffer to hand out incomplete frames.
// See: https://bugs.chromium.org/p/webrtc/issues/detail?id=7106
if (is_h264 &&
(!sequence_buffer_[start_index].used ||
data_buffer_[start_index].timestamp != frame_timestamp)) {
break;
}
--start_seq_num;
}
// Fix the order since the packet-finding loop traverses backwards.
std::reverse(packet_infos.begin(), packet_infos.end());
if (is_h264) {
// Warn if this is an unsafe frame.
if (has_h264_idr && (!has_h264_sps || !has_h264_pps)) {
RTC_LOG(LS_WARNING)
<< "Received H.264-IDR frame "
<< "(SPS: " << has_h264_sps << ", PPS: " << has_h264_pps
<< "). Treating as "
<< (sps_pps_idr_is_h264_keyframe_ ? "delta" : "key")
<< " frame since WebRTC-SpsPpsIdrIsH264Keyframe is "
<< (sps_pps_idr_is_h264_keyframe_ ? "enabled." : "disabled");
}
// Now that we have decided whether to treat this frame as a key frame
// or delta frame in the frame buffer, we update the field that
// determines if the RtpFrameObject is a key frame or delta frame.
const size_t first_packet_index = start_seq_num % size_;
RTC_CHECK_LT(first_packet_index, size_);
if (is_h264_keyframe) {
data_buffer_[first_packet_index].video_header.frame_type =
VideoFrameType::kVideoFrameKey;
} else {
data_buffer_[first_packet_index].video_header.frame_type =
VideoFrameType::kVideoFrameDelta;
}
// With IPPP, if this is not a keyframe, make sure there are no gaps
// in the packet sequence numbers up until this point.
const uint8_t h264tid =
data_buffer_[start_index].video_header.frame_marking.temporal_id;
if (h264tid == kNoTemporalIdx && !is_h264_keyframe &&
missing_packets_.upper_bound(start_seq_num) !=
missing_packets_.begin()) {
uint16_t stop_index = (index + 1) % size_;
while (start_index != stop_index) {
sequence_buffer_[start_index].frame_created = false;
start_index = (start_index + 1) % size_;
}
return found_frames;
}
}
missing_packets_.erase(missing_packets_.begin(),
missing_packets_.upper_bound(seq_num));
found_frames.emplace_back(
new RtpFrameObject(this, start_seq_num, seq_num, frame_size,
max_nack_count, min_recv_time, max_recv_time,
RtpPacketInfos(std::move(packet_infos))));
ClearInterval(start_seq_num, seq_num);
}
++seq_num;
}
return found_frames;
}
bool PacketBuffer::GetBitstream(const RtpFrameObject& frame,
uint8_t* destination) {
rtc::CritScope lock(&crit_);
size_t index = frame.first_seq_num() % size_;
size_t end = (frame.last_seq_num() + 1) % size_;
uint16_t seq_num = frame.first_seq_num();
uint32_t timestamp = frame.Timestamp();
uint8_t* destination_end = destination + frame.size();
do {
// Check both seq_num and timestamp to handle the case when seq_num wraps
// around too quickly for high packet rates.
if (!sequence_buffer_[index].used ||
sequence_buffer_[index].seq_num != seq_num ||
data_buffer_[index].timestamp != timestamp) {
return false;
}
RTC_DCHECK_EQ(data_buffer_[index].seqNum, sequence_buffer_[index].seq_num);
size_t length = data_buffer_[index].sizeBytes;
if (destination + length > destination_end) {
RTC_LOG(LS_WARNING) << "Frame (" << frame.id.picture_id << ":"
<< static_cast<int>(frame.id.spatial_layer) << ")"
<< " bitstream buffer is not large enough.";
return false;
}
const uint8_t* source = data_buffer_[index].dataPtr;
memcpy(destination, source, length);
destination += length;
index = (index + 1) % size_;
++seq_num;
} while (index != end);
return true;
}
VCMPacket* PacketBuffer::GetPacket(uint16_t seq_num) {
size_t index = seq_num % size_;
if (!sequence_buffer_[index].used ||
seq_num != sequence_buffer_[index].seq_num) {
return nullptr;
}
return &data_buffer_[index];
}
int PacketBuffer::AddRef() const {
return rtc::AtomicOps::Increment(&ref_count_);
}
int PacketBuffer::Release() const {
int count = rtc::AtomicOps::Decrement(&ref_count_);
if (!count) {
delete this;
}
return count;
}
void PacketBuffer::UpdateMissingPackets(uint16_t seq_num) {
if (!newest_inserted_seq_num_)
newest_inserted_seq_num_ = seq_num;
const int kMaxPaddingAge = 1000;
if (AheadOf(seq_num, *newest_inserted_seq_num_)) {
uint16_t old_seq_num = seq_num - kMaxPaddingAge;
auto erase_to = missing_packets_.lower_bound(old_seq_num);
missing_packets_.erase(missing_packets_.begin(), erase_to);
// Guard against inserting a large amount of missing packets if there is a
// jump in the sequence number.
if (AheadOf(old_seq_num, *newest_inserted_seq_num_))
*newest_inserted_seq_num_ = old_seq_num;
++*newest_inserted_seq_num_;
while (AheadOf(seq_num, *newest_inserted_seq_num_)) {
missing_packets_.insert(*newest_inserted_seq_num_);
++*newest_inserted_seq_num_;
}
} else {
missing_packets_.erase(seq_num);
}
}
void PacketBuffer::OnTimestampReceived(uint32_t rtp_timestamp) {
const size_t kMaxTimestampsHistory = 1000;
if (rtp_timestamps_history_set_.insert(rtp_timestamp).second) {
rtp_timestamps_history_queue_.push(rtp_timestamp);
++unique_frames_seen_;
if (rtp_timestamps_history_set_.size() > kMaxTimestampsHistory) {
uint32_t discarded_timestamp = rtp_timestamps_history_queue_.front();
rtp_timestamps_history_set_.erase(discarded_timestamp);
rtp_timestamps_history_queue_.pop();
}
}
}
} // namespace video_coding
} // namespace webrtc