webrtc_m130/modules/rtp_rtcp/source/rtp_sender_egress.cc

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

657 lines
24 KiB
C++
Raw Normal View History

/*
* Copyright (c) 2019 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/rtp_rtcp/source/rtp_sender_egress.h"
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
#include <algorithm>
#include <limits>
#include <memory>
#include <utility>
#include "absl/strings/match.h"
#include "api/transport/field_trial_based_config.h"
#include "logging/rtc_event_log/events/rtc_event_rtp_packet_outgoing.h"
#include "rtc_base/logging.h"
namespace webrtc {
namespace {
constexpr uint32_t kTimestampTicksPerMs = 90;
constexpr int kSendSideDelayWindowMs = 1000;
constexpr int kBitrateStatisticsWindowMs = 1000;
constexpr size_t kRtpSequenceNumberMapMaxEntries = 1 << 13;
constexpr TimeDelta kUpdateInterval =
TimeDelta::Millis(kBitrateStatisticsWindowMs);
bool IsTrialSetTo(const FieldTrialsView* field_trials,
absl::string_view name,
absl::string_view value) {
FieldTrialBasedConfig default_trials;
auto& trials = field_trials ? *field_trials : default_trials;
return absl::StartsWith(trials.Lookup(name), value);
}
} // namespace
RtpSenderEgress::NonPacedPacketSender::NonPacedPacketSender(
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
RtpSenderEgress* sender,
PacketSequencer* sequencer)
: transport_sequence_number_(0), sender_(sender), sequencer_(sequencer) {
RTC_DCHECK(sequencer);
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
}
RtpSenderEgress::NonPacedPacketSender::~NonPacedPacketSender() = default;
void RtpSenderEgress::NonPacedPacketSender::EnqueuePackets(
std::vector<std::unique_ptr<RtpPacketToSend>> packets) {
for (auto& packet : packets) {
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
PrepareForSend(packet.get());
Revert "Reland "Allows FEC generation after pacer step."" This reverts commit 19df870d924662e3b6efb86078d31a8e086b38b5. Reason for revert: Downstream project failure Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} TBR=sprang@webrtc.org,srte@webrtc.org Change-Id: I3b2b25898ce88b64c2322f68ef83f9f86ac2edb0 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: webrtc:11340 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178563 Reviewed-by: Erik Språng <sprang@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31614}
2020-07-02 12:02:36 +00:00
sender_->SendPacket(packet.get(), PacedPacketInfo());
Reland "Allows FEC generation after pacer step." This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 Patchset 2 contains a fix. Old code can in factor call RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec is not supported there - we shouldn't crash. Original change's description: > Allows FEC generation after pacer step. > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > This CL enables FEC packets to be generated as media packets are sent, > rather than generated, i.e. media packets are inserted into the fec > generator after the pacing stage rather than at packetization time. > > This may have some small impact of performance. FEC packets are > typically only generated when a new packet with a marker bit is added, > which means FEC packets protecting a frame will now be sent after all > of the media packets, rather than (potentially) interleaved with them. > Therefore this feature is currently behind a flag so we can examine the > impact. Once we are comfortable with the behavior we'll make it default > and remove the old code. > > Note that this change does not include the "protect all header > extensions" part of the original CL - that will be a follow-up. > > Bug: webrtc:11340 > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31558} Bug: webrtc:11340 Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 Commit-Queue: Erik Språng <sprang@webrtc.org> Reviewed-by: Sebastian Jansson <srte@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31613}
2020-07-01 17:45:23 +02:00
}
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
auto fec_packets = sender_->FetchFecPackets();
if (!fec_packets.empty()) {
EnqueuePackets(std::move(fec_packets));
}
}
void RtpSenderEgress::NonPacedPacketSender::PrepareForSend(
RtpPacketToSend* packet) {
// Assign sequence numbers, but not for flexfec which is already running on
// an internally maintained sequence number series.
if (packet->Ssrc() != sender_->FlexFecSsrc()) {
sequencer_->Sequence(*packet);
}
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
if (!packet->SetExtension<TransportSequenceNumber>(
++transport_sequence_number_)) {
--transport_sequence_number_;
}
packet->ReserveExtension<TransmissionOffset>();
packet->ReserveExtension<AbsoluteSendTime>();
}
RtpSenderEgress::RtpSenderEgress(const RtpRtcpInterface::Configuration& config,
RtpPacketHistory* packet_history)
: worker_queue_(TaskQueueBase::Current()),
ssrc_(config.local_media_ssrc),
rtx_ssrc_(config.rtx_send_ssrc),
Reland "Reland "Refactors UlpFec and FlexFec to use a common interface."" This is a reland of 49734dc0faa69616a58a1a95c7fc61a4610793cf Patchset 2 contains a fix for the fuzzer set up. Since we now parse an RtpPacket out of the fuzzer data, the header needs to be correct, otherwise we fail before even reaching the FEC code that we actually want to test. Bug: webrtc:11340, chromium:1052323, chromium:1055974 TBR=stefan@webrtc.org Original change's description: > Reland "Refactors UlpFec and FlexFec to use a common interface." > > This is a reland of 11af1d7444fd7438766b7bc52cbd64752d72e32e > > Original change's description: > > Refactors UlpFec and FlexFec to use a common interface. > > > > The new VideoFecGenerator is now injected into RtpSenderVideo, > > and generalizes the usage. > > This also prepares for being able to genera FEC in the RTP egress > > module. > > > > Bug: webrtc:11340 > > Change-Id: I8aa873129b2fb4131eb3399ee88f6ea2747155a3 > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/168347 > > Reviewed-by: Stefan Holmer <stefan@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Reviewed-by: Rasmus Brandt <brandtr@webrtc.org> > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#30515} > > Bug: webrtc:11340, chromium:1052323 > Change-Id: Id646047365f1c46cca9e6f3e8eefa5151207b4a0 > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/168608 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Stefan Holmer <stefan@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#30593} Bug: webrtc:11340, chromium:1052323 Change-Id: Ib8925f44e2edfcfeadc95c845c3bfc23822604ed Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/169222 Commit-Queue: Erik Språng <sprang@webrtc.org> Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org> Cr-Commit-Position: refs/heads/master@{#30724}
2020-03-05 10:14:04 +01:00
flexfec_ssrc_(config.fec_generator ? config.fec_generator->FecSsrc()
: absl::nullopt),
populate_network2_timestamp_(config.populate_network2_timestamp),
send_side_bwe_with_overhead_(
!IsTrialSetTo(config.field_trials,
"WebRTC-SendSideBwe-WithOverhead",
"Disabled")),
clock_(config.clock),
packet_history_(packet_history),
transport_(config.outgoing_transport),
event_log_(config.event_log),
#if BWE_TEST_LOGGING_COMPILE_TIME_ENABLE
is_audio_(config.audio),
#endif
need_rtp_packet_infos_(config.need_rtp_packet_infos),
fec_generator_(config.fec_generator),
transport_feedback_observer_(config.transport_feedback_callback),
send_side_delay_observer_(config.send_side_delay_observer),
send_packet_observer_(config.send_packet_observer),
rtp_stats_callback_(config.rtp_stats_callback),
bitrate_callback_(config.send_bitrate_observer),
media_has_been_sent_(false),
force_part_of_allocation_(false),
timestamp_offset_(0),
max_delay_it_(send_delays_.end()),
sum_delays_ms_(0),
total_packet_send_delay_ms_(0),
send_rates_(kNumMediaTypes,
{kBitrateStatisticsWindowMs, RateStatistics::kBpsScale}),
rtp_sequence_number_map_(need_rtp_packet_infos_
? std::make_unique<RtpSequenceNumberMap>(
kRtpSequenceNumberMapMaxEntries)
: nullptr) {
RTC_DCHECK(worker_queue_);
pacer_checker_.Detach();
if (bitrate_callback_) {
update_task_ = RepeatingTaskHandle::DelayedStart(worker_queue_,
kUpdateInterval, [this]() {
PeriodicUpdate();
return kUpdateInterval;
});
}
}
RtpSenderEgress::~RtpSenderEgress() {
RTC_DCHECK_RUN_ON(worker_queue_);
update_task_.Stop();
}
void RtpSenderEgress::SendPacket(RtpPacketToSend* packet,
const PacedPacketInfo& pacing_info) {
RTC_DCHECK_RUN_ON(&pacer_checker_);
RTC_DCHECK(packet);
if (packet->Ssrc() == ssrc_ &&
packet->packet_type() != RtpPacketMediaType::kRetransmission) {
if (last_sent_seq_.has_value()) {
RTC_DCHECK_EQ(static_cast<uint16_t>(*last_sent_seq_ + 1),
packet->SequenceNumber());
}
last_sent_seq_ = packet->SequenceNumber();
} else if (packet->Ssrc() == rtx_ssrc_) {
if (last_sent_rtx_seq_.has_value()) {
RTC_DCHECK_EQ(static_cast<uint16_t>(*last_sent_rtx_seq_ + 1),
packet->SequenceNumber());
}
last_sent_rtx_seq_ = packet->SequenceNumber();
}
RTC_DCHECK(packet->packet_type().has_value());
RTC_DCHECK(HasCorrectSsrc(*packet));
if (packet->packet_type() == RtpPacketMediaType::kRetransmission) {
RTC_DCHECK(packet->retransmitted_sequence_number().has_value());
}
const uint32_t packet_ssrc = packet->Ssrc();
const Timestamp now = clock_->CurrentTime();
#if BWE_TEST_LOGGING_COMPILE_TIME_ENABLE
worker_queue_->PostTask(
SafeTask(task_safety_.flag(), [this, now, packet_ssrc]() {
BweTestLoggingPlot(now.ms(), packet_ssrc);
}));
#endif
if (need_rtp_packet_infos_ &&
packet->packet_type() == RtpPacketToSend::Type::kVideo) {
worker_queue_->PostTask(SafeTask(
task_safety_.flag(),
[this, packet_timestamp = packet->Timestamp(),
is_first_packet_of_frame = packet->is_first_packet_of_frame(),
is_last_packet_of_frame = packet->Marker(),
sequence_number = packet->SequenceNumber()]() {
RTC_DCHECK_RUN_ON(worker_queue_);
// Last packet of a frame, add it to sequence number info map.
const uint32_t timestamp = packet_timestamp - timestamp_offset_;
rtp_sequence_number_map_->InsertPacket(
sequence_number,
RtpSequenceNumberMap::Info(timestamp, is_first_packet_of_frame,
is_last_packet_of_frame));
}));
}
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
if (fec_generator_ && packet->fec_protect_packet()) {
// This packet should be protected by FEC, add it to packet generator.
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
RTC_DCHECK(fec_generator_);
RTC_DCHECK(packet->packet_type() == RtpPacketMediaType::kVideo);
absl::optional<std::pair<FecProtectionParams, FecProtectionParams>>
new_fec_params;
{
MutexLock lock(&lock_);
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
new_fec_params.swap(pending_fec_params_);
}
if (new_fec_params) {
fec_generator_->SetProtectionParameters(new_fec_params->first,
new_fec_params->second);
}
if (packet->is_red()) {
RtpPacketToSend unpacked_packet(*packet);
const rtc::CopyOnWriteBuffer buffer = packet->Buffer();
// Grab media payload type from RED header.
const size_t headers_size = packet->headers_size();
unpacked_packet.SetPayloadType(buffer[headers_size]);
// Copy the media payload into the unpacked buffer.
uint8_t* payload_buffer =
unpacked_packet.SetPayloadSize(packet->payload_size() - 1);
std::copy(&packet->payload()[0] + 1,
&packet->payload()[0] + packet->payload_size(), payload_buffer);
fec_generator_->AddPacketAndGenerateFec(unpacked_packet);
} else {
// If not RED encapsulated - we can just insert packet directly.
fec_generator_->AddPacketAndGenerateFec(*packet);
}
}
// Bug webrtc:7859. While FEC is invoked from rtp_sender_video, and not after
// the pacer, these modifications of the header below are happening after the
// FEC protection packets are calculated. This will corrupt recovered packets
// at the same place. It's not an issue for extensions, which are present in
// all the packets (their content just may be incorrect on recovered packets).
// In case of VideoTimingExtension, since it's present not in every packet,
// data after rtp header may be corrupted if these packets are protected by
// the FEC.
TimeDelta diff = now - packet->capture_time();
if (packet->HasExtension<TransmissionOffset>()) {
packet->SetExtension<TransmissionOffset>(kTimestampTicksPerMs * diff.ms());
}
if (packet->HasExtension<AbsoluteSendTime>()) {
packet->SetExtension<AbsoluteSendTime>(AbsoluteSendTime::To24Bits(now));
}
if (packet->HasExtension<VideoTimingExtension>()) {
if (populate_network2_timestamp_) {
packet->set_network2_time(now);
} else {
packet->set_pacer_exit_time(now);
}
}
const bool is_media = packet->packet_type() == RtpPacketMediaType::kAudio ||
packet->packet_type() == RtpPacketMediaType::kVideo;
PacketOptions options;
{
MutexLock lock(&lock_);
options.included_in_allocation = force_part_of_allocation_;
}
// Downstream code actually uses this flag to distinguish between media and
// everything else.
options.is_retransmit = !is_media;
if (auto packet_id = packet->GetExtension<TransportSequenceNumber>()) {
options.packet_id = *packet_id;
options.included_in_feedback = true;
options.included_in_allocation = true;
AddPacketToTransportFeedback(*packet_id, *packet, pacing_info);
}
options.additional_data = packet->additional_data();
if (packet->packet_type() != RtpPacketMediaType::kPadding &&
packet->packet_type() != RtpPacketMediaType::kRetransmission) {
UpdateDelayStatistics(packet->capture_time().ms(), now.ms(), packet_ssrc);
Reland "Represent RtpPacketToSend::capture_time with Timestamp" This reverts commit 56db8d09529d5ba92d24954a1d78a90c8ea2cd85. Reason for revert: downstream problem addressed Original change's description: > Revert "Represent RtpPacketToSend::capture_time with Timestamp" > > This reverts commit 385eb9714daa80306d2f92d36678c42892dab555. > > Reason for revert: Causes problems downstream: > > # > # Fatal error in: rtc_base/units/unit_base.h, line 122 > # last system error: 0 > # Check failed: value >= 0 (-234 vs. 0) > > Original change's description: > > Represent RtpPacketToSend::capture_time with Timestamp > > > > Bug: webrtc:13757 > > Change-Id: I0ede22cd34e3a59afe1477d8edd495dce64e3242 > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/252586 > > Reviewed-by: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Rasmus Brandt <brandtr@webrtc.org> > > Commit-Queue: Danil Chapovalov <danilchap@webrtc.org> > > Cr-Commit-Position: refs/heads/main@{#36083} > > Bug: webrtc:13757 > Change-Id: I8442abd438be8726cf671d0f372d50ecfac6847e > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/252720 > Auto-Submit: Tomas Gunnarsson <tommi@webrtc.org> > Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com> > Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org> > Cr-Commit-Position: refs/heads/main@{#36087} Bug: webrtc:13757 Change-Id: I1fa852757480116f35deb2b6c3c27800bdf5e197 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/252781 Reviewed-by: Rasmus Brandt <brandtr@webrtc.org> Reviewed-by: Erik Språng <sprang@webrtc.org> Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org> Commit-Queue: Danil Chapovalov <danilchap@webrtc.org> Cr-Commit-Position: refs/heads/main@{#36093}
2022-02-27 21:10:55 +00:00
UpdateOnSendPacket(options.packet_id, packet->capture_time().ms(),
packet_ssrc);
}
const bool send_success = SendPacketToNetwork(*packet, options, pacing_info);
// Put packet in retransmission history or update pending status even if
// actual sending fails.
if (is_media && packet->allow_retransmission()) {
packet_history_->PutRtpPacket(std::make_unique<RtpPacketToSend>(*packet),
now);
} else if (packet->retransmitted_sequence_number()) {
packet_history_->MarkPacketAsSent(*packet->retransmitted_sequence_number());
}
if (send_success) {
// `media_has_been_sent_` is used by RTPSender to figure out if it can send
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
// padding in the absence of transport-cc or abs-send-time.
// In those cases media must be sent first to set a reference timestamp.
media_has_been_sent_ = true;
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
// TODO(sprang): Add support for FEC protecting all header extensions, add
// media packet to generator here instead.
RTC_DCHECK(packet->packet_type().has_value());
RtpPacketMediaType packet_type = *packet->packet_type();
RtpPacketCounter counter(*packet);
size_t size = packet->size();
worker_queue_->PostTask(
SafeTask(task_safety_.flag(), [this, now, packet_ssrc, packet_type,
counter = std::move(counter), size]() {
RTC_DCHECK_RUN_ON(worker_queue_);
UpdateRtpStats(now.ms(), packet_ssrc, packet_type, std::move(counter),
size);
}));
}
}
RtpSendRates RtpSenderEgress::GetSendRates() const {
MutexLock lock(&lock_);
const int64_t now_ms = clock_->TimeInMilliseconds();
return GetSendRatesLocked(now_ms);
}
RtpSendRates RtpSenderEgress::GetSendRatesLocked(int64_t now_ms) const {
RtpSendRates current_rates;
for (size_t i = 0; i < kNumMediaTypes; ++i) {
RtpPacketMediaType type = static_cast<RtpPacketMediaType>(i);
current_rates[type] =
DataRate::BitsPerSec(send_rates_[i].Rate(now_ms).value_or(0));
}
return current_rates;
}
void RtpSenderEgress::GetDataCounters(StreamDataCounters* rtp_stats,
StreamDataCounters* rtx_stats) const {
// TODO(bugs.webrtc.org/11581): make sure rtx_rtp_stats_ and rtp_stats_ are
// only touched on the worker thread.
MutexLock lock(&lock_);
*rtp_stats = rtp_stats_;
*rtx_stats = rtx_rtp_stats_;
}
void RtpSenderEgress::ForceIncludeSendPacketsInAllocation(
bool part_of_allocation) {
MutexLock lock(&lock_);
force_part_of_allocation_ = part_of_allocation;
}
bool RtpSenderEgress::MediaHasBeenSent() const {
RTC_DCHECK_RUN_ON(&pacer_checker_);
return media_has_been_sent_;
}
void RtpSenderEgress::SetMediaHasBeenSent(bool media_sent) {
RTC_DCHECK_RUN_ON(&pacer_checker_);
media_has_been_sent_ = media_sent;
}
void RtpSenderEgress::SetTimestampOffset(uint32_t timestamp) {
RTC_DCHECK_RUN_ON(worker_queue_);
timestamp_offset_ = timestamp;
}
std::vector<RtpSequenceNumberMap::Info> RtpSenderEgress::GetSentRtpPacketInfos(
rtc::ArrayView<const uint16_t> sequence_numbers) const {
RTC_DCHECK_RUN_ON(worker_queue_);
RTC_DCHECK(!sequence_numbers.empty());
if (!need_rtp_packet_infos_) {
return std::vector<RtpSequenceNumberMap::Info>();
}
std::vector<RtpSequenceNumberMap::Info> results;
results.reserve(sequence_numbers.size());
for (uint16_t sequence_number : sequence_numbers) {
const auto& info = rtp_sequence_number_map_->Get(sequence_number);
if (!info) {
// The empty vector will be returned. We can delay the clearing
// of the vector until after we exit the critical section.
return std::vector<RtpSequenceNumberMap::Info>();
}
results.push_back(*info);
}
return results;
}
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
void RtpSenderEgress::SetFecProtectionParameters(
const FecProtectionParams& delta_params,
const FecProtectionParams& key_params) {
// TODO(sprang): Post task to pacer queue instead, one pacer is fully
// migrated to a task queue.
MutexLock lock(&lock_);
Reland "Reland "Allows FEC generation after pacer step."" This is a reland of 19df870d924662e3b6efb86078d31a8e086b38b5 Patchset 1 is the original. Subsequent patchset changes threadchecker that crashed with downstream code. Original change's description: > Reland "Allows FEC generation after pacer step." > > This is a reland of 75fd127640bdf1729af6b4a25875e6d01f1570e0 > > Patchset 2 contains a fix. Old code can in factor call > RtpRtcpImpl::FetchFec(). It should only be a noop since deferred fec > is not supported there - we shouldn't crash. > > Original change's description: > > Allows FEC generation after pacer step. > > > > Split out from https://webrtc-review.googlesource.com/c/src/+/173708 > > This CL enables FEC packets to be generated as media packets are sent, > > rather than generated, i.e. media packets are inserted into the fec > > generator after the pacing stage rather than at packetization time. > > > > This may have some small impact of performance. FEC packets are > > typically only generated when a new packet with a marker bit is added, > > which means FEC packets protecting a frame will now be sent after all > > of the media packets, rather than (potentially) interleaved with them. > > Therefore this feature is currently behind a flag so we can examine the > > impact. Once we are comfortable with the behavior we'll make it default > > and remove the old code. > > > > Note that this change does not include the "protect all header > > extensions" part of the original CL - that will be a follow-up. > > > > Bug: webrtc:11340 > > Change-Id: I3fe139c5d53968579b75b91e2612075451ff0f5d > > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177760 > > Commit-Queue: Erik Språng <sprang@webrtc.org> > > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > > Cr-Commit-Position: refs/heads/master@{#31558} > > Bug: webrtc:11340 > Change-Id: I2ea49ee87ee9ff409044e34a777a7dd0ae0a077f > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/177984 > Commit-Queue: Erik Språng <sprang@webrtc.org> > Reviewed-by: Sebastian Jansson <srte@webrtc.org> > Cr-Commit-Position: refs/heads/master@{#31613} Bug: webrtc:11340 Change-Id: Ib741c8c284f523c959f8aca454088d9eee7b17f8 Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178600 Reviewed-by: Sebastian Jansson <srte@webrtc.org> Commit-Queue: Erik Språng <sprang@webrtc.org> Cr-Commit-Position: refs/heads/master@{#31619}
2020-07-02 17:41:32 +02:00
pending_fec_params_.emplace(delta_params, key_params);
}
std::vector<std::unique_ptr<RtpPacketToSend>>
RtpSenderEgress::FetchFecPackets() {
RTC_DCHECK_RUN_ON(&pacer_checker_);
if (fec_generator_) {
return fec_generator_->GetFecPackets();
}
return {};
}
void RtpSenderEgress::OnAbortedRetransmissions(
rtc::ArrayView<const uint16_t> sequence_numbers) {
RTC_DCHECK_RUN_ON(&pacer_checker_);
// Mark aborted retransmissions as sent, rather than leaving them in
// a 'pending' state - otherwise they can not be requested again and
// will not be cleared until the history has reached its max size.
for (uint16_t seq_no : sequence_numbers) {
packet_history_->MarkPacketAsSent(seq_no);
}
}
bool RtpSenderEgress::HasCorrectSsrc(const RtpPacketToSend& packet) const {
switch (*packet.packet_type()) {
case RtpPacketMediaType::kAudio:
case RtpPacketMediaType::kVideo:
return packet.Ssrc() == ssrc_;
case RtpPacketMediaType::kRetransmission:
case RtpPacketMediaType::kPadding:
// Both padding and retransmission must be on either the media or the
// RTX stream.
return packet.Ssrc() == rtx_ssrc_ || packet.Ssrc() == ssrc_;
case RtpPacketMediaType::kForwardErrorCorrection:
// FlexFEC is on separate SSRC, ULPFEC uses media SSRC.
return packet.Ssrc() == ssrc_ || packet.Ssrc() == flexfec_ssrc_;
}
return false;
}
void RtpSenderEgress::AddPacketToTransportFeedback(
uint16_t packet_id,
const RtpPacketToSend& packet,
const PacedPacketInfo& pacing_info) {
if (transport_feedback_observer_) {
size_t packet_size = packet.payload_size() + packet.padding_size();
if (send_side_bwe_with_overhead_) {
packet_size = packet.size();
}
RtpPacketSendInfo packet_info;
packet_info.transport_sequence_number = packet_id;
packet_info.rtp_timestamp = packet.Timestamp();
packet_info.length = packet_size;
packet_info.pacing_info = pacing_info;
packet_info.packet_type = packet.packet_type();
switch (*packet_info.packet_type) {
case RtpPacketMediaType::kAudio:
case RtpPacketMediaType::kVideo:
packet_info.media_ssrc = ssrc_;
packet_info.rtp_sequence_number = packet.SequenceNumber();
break;
case RtpPacketMediaType::kRetransmission:
// For retransmissions, we're want to remove the original media packet
// if the retransmit arrives - so populate that in the packet info.
packet_info.media_ssrc = ssrc_;
packet_info.rtp_sequence_number =
*packet.retransmitted_sequence_number();
break;
case RtpPacketMediaType::kPadding:
case RtpPacketMediaType::kForwardErrorCorrection:
// We're not interested in feedback about these packets being received
// or lost.
break;
}
transport_feedback_observer_->OnAddPacket(packet_info);
}
}
void RtpSenderEgress::UpdateDelayStatistics(int64_t capture_time_ms,
int64_t now_ms,
uint32_t ssrc) {
if (!send_side_delay_observer_ || capture_time_ms <= 0)
return;
int avg_delay_ms = 0;
int max_delay_ms = 0;
uint64_t total_packet_send_delay_ms = 0;
{
MutexLock lock(&lock_);
// Compute the max and average of the recent capture-to-send delays.
// The time complexity of the current approach depends on the distribution
// of the delay values. This could be done more efficiently.
// Remove elements older than kSendSideDelayWindowMs.
auto lower_bound =
send_delays_.lower_bound(now_ms - kSendSideDelayWindowMs);
for (auto it = send_delays_.begin(); it != lower_bound; ++it) {
if (max_delay_it_ == it) {
max_delay_it_ = send_delays_.end();
}
sum_delays_ms_ -= it->second;
}
send_delays_.erase(send_delays_.begin(), lower_bound);
if (max_delay_it_ == send_delays_.end()) {
// Removed the previous max. Need to recompute.
RecomputeMaxSendDelay();
}
// Add the new element.
RTC_DCHECK_GE(now_ms, 0);
RTC_DCHECK_LE(now_ms, std::numeric_limits<int64_t>::max() / 2);
RTC_DCHECK_GE(capture_time_ms, 0);
RTC_DCHECK_LE(capture_time_ms, std::numeric_limits<int64_t>::max() / 2);
int64_t diff_ms = now_ms - capture_time_ms;
RTC_DCHECK_GE(diff_ms, static_cast<int64_t>(0));
RTC_DCHECK_LE(diff_ms, std::numeric_limits<int>::max());
int new_send_delay = rtc::dchecked_cast<int>(now_ms - capture_time_ms);
SendDelayMap::iterator it;
bool inserted;
std::tie(it, inserted) =
send_delays_.insert(std::make_pair(now_ms, new_send_delay));
if (!inserted) {
// TODO(terelius): If we have multiple delay measurements during the same
// millisecond then we keep the most recent one. It is not clear that this
// is the right decision, but it preserves an earlier behavior.
int previous_send_delay = it->second;
sum_delays_ms_ -= previous_send_delay;
it->second = new_send_delay;
if (max_delay_it_ == it && new_send_delay < previous_send_delay) {
RecomputeMaxSendDelay();
}
}
if (max_delay_it_ == send_delays_.end() ||
it->second >= max_delay_it_->second) {
max_delay_it_ = it;
}
sum_delays_ms_ += new_send_delay;
total_packet_send_delay_ms_ += new_send_delay;
total_packet_send_delay_ms = total_packet_send_delay_ms_;
size_t num_delays = send_delays_.size();
RTC_DCHECK(max_delay_it_ != send_delays_.end());
max_delay_ms = rtc::dchecked_cast<int>(max_delay_it_->second);
int64_t avg_ms = (sum_delays_ms_ + num_delays / 2) / num_delays;
RTC_DCHECK_GE(avg_ms, static_cast<int64_t>(0));
RTC_DCHECK_LE(avg_ms,
static_cast<int64_t>(std::numeric_limits<int>::max()));
avg_delay_ms =
rtc::dchecked_cast<int>((sum_delays_ms_ + num_delays / 2) / num_delays);
}
send_side_delay_observer_->SendSideDelayUpdated(
avg_delay_ms, max_delay_ms, total_packet_send_delay_ms, ssrc);
}
void RtpSenderEgress::RecomputeMaxSendDelay() {
max_delay_it_ = send_delays_.begin();
for (auto it = send_delays_.begin(); it != send_delays_.end(); ++it) {
if (it->second >= max_delay_it_->second) {
max_delay_it_ = it;
}
}
}
void RtpSenderEgress::UpdateOnSendPacket(int packet_id,
int64_t capture_time_ms,
uint32_t ssrc) {
if (!send_packet_observer_ || capture_time_ms <= 0 || packet_id == -1) {
return;
}
send_packet_observer_->OnSendPacket(packet_id, capture_time_ms, ssrc);
}
bool RtpSenderEgress::SendPacketToNetwork(const RtpPacketToSend& packet,
const PacketOptions& options,
const PacedPacketInfo& pacing_info) {
int bytes_sent = -1;
if (transport_) {
bytes_sent = transport_->SendRtp(packet.data(), packet.size(), options)
? static_cast<int>(packet.size())
: -1;
if (event_log_ && bytes_sent > 0) {
event_log_->Log(std::make_unique<RtcEventRtpPacketOutgoing>(
packet, pacing_info.probe_cluster_id));
}
}
if (bytes_sent <= 0) {
RTC_LOG(LS_WARNING) << "Transport failed to send packet.";
return false;
}
return true;
}
void RtpSenderEgress::UpdateRtpStats(int64_t now_ms,
uint32_t packet_ssrc,
RtpPacketMediaType packet_type,
RtpPacketCounter counter,
size_t packet_size) {
RTC_DCHECK_RUN_ON(worker_queue_);
// TODO(bugs.webrtc.org/11581): send_rates_ should be touched only on the
// worker thread.
RtpSendRates send_rates;
{
MutexLock lock(&lock_);
// TODO(bugs.webrtc.org/11581): make sure rtx_rtp_stats_ and rtp_stats_ are
// only touched on the worker thread.
StreamDataCounters* counters =
packet_ssrc == rtx_ssrc_ ? &rtx_rtp_stats_ : &rtp_stats_;
if (counters->first_packet_time_ms == -1) {
counters->first_packet_time_ms = now_ms;
}
if (packet_type == RtpPacketMediaType::kForwardErrorCorrection) {
counters->fec.Add(counter);
} else if (packet_type == RtpPacketMediaType::kRetransmission) {
counters->retransmitted.Add(counter);
}
counters->transmitted.Add(counter);
send_rates_[static_cast<size_t>(packet_type)].Update(packet_size, now_ms);
if (bitrate_callback_) {
send_rates = GetSendRatesLocked(now_ms);
}
if (rtp_stats_callback_) {
rtp_stats_callback_->DataCountersUpdated(*counters, packet_ssrc);
}
}
// The bitrate_callback_ and rtp_stats_callback_ pointers in practice point
// to the same object, so these callbacks could be consolidated into one.
if (bitrate_callback_) {
bitrate_callback_->Notify(
send_rates.Sum().bps(),
send_rates[RtpPacketMediaType::kRetransmission].bps(), ssrc_);
}
}
void RtpSenderEgress::PeriodicUpdate() {
RTC_DCHECK_RUN_ON(worker_queue_);
RTC_DCHECK(bitrate_callback_);
RtpSendRates send_rates = GetSendRates();
bitrate_callback_->Notify(
send_rates.Sum().bps(),
send_rates[RtpPacketMediaType::kRetransmission].bps(), ssrc_);
}
#if BWE_TEST_LOGGING_COMPILE_TIME_ENABLE
void RtpSenderEgress::BweTestLoggingPlot(int64_t now_ms, uint32_t packet_ssrc) {
RTC_DCHECK_RUN_ON(worker_queue_);
const auto rates = GetSendRates();
if (is_audio_) {
BWE_TEST_LOGGING_PLOT_WITH_SSRC(1, "AudioTotBitrate_kbps", now_ms,
rates.Sum().kbps(), packet_ssrc);
BWE_TEST_LOGGING_PLOT_WITH_SSRC(
1, "AudioNackBitrate_kbps", now_ms,
rates[RtpPacketMediaType::kRetransmission].kbps(), packet_ssrc);
} else {
BWE_TEST_LOGGING_PLOT_WITH_SSRC(1, "VideoTotBitrate_kbps", now_ms,
rates.Sum().kbps(), packet_ssrc);
BWE_TEST_LOGGING_PLOT_WITH_SSRC(
1, "VideoNackBitrate_kbps", now_ms,
rates[RtpPacketMediaType::kRetransmission].kbps(), packet_ssrc);
}
}
#endif // BWE_TEST_LOGGING_COMPILE_TIME_ENABLE
} // namespace webrtc