Reason for revert:
Reland
Original issue's description:
> Revert of Make the acceptable queue in the cwnd experiment configurable. (patchset #1 id:1 of https://codereview.webrtc.org/2998753002/ )
>
> Reason for revert:
> Speculative revert to see if this caused regressions in android perf tests.
>
> Original issue's description:
> > Make the acceptable queue in the cwnd experiment configurable.
> >
> > BUG=webrtc:7926
> >
> > Review-Url: https://codereview.webrtc.org/2998753002
> > Cr-Commit-Position: refs/heads/master@{#19320}
> > Committed: 7c83c56b6d
>
> TBR=philipel@webrtc.org
> # Not skipping CQ checks because original CL landed more than 1 days ago.
> BUG=webrtc:7926
>
> Review-Url: https://codereview.webrtc.org/2999893002
> Cr-Commit-Position: refs/heads/master@{#19337}
> Committed: c5d9e63c2bTBR=philipel@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/2999083002
Cr-Commit-Position: refs/heads/master@{#19377}
Reason for revert:
Reland
Original issue's description:
> Revert of Add functionality which limits the number of bytes on the network. (patchset #26 id:500001 of https://codereview.webrtc.org/2918323002/ )
>
> Reason for revert:
> Speculative revert to see if this caused regressions in android perf tests.
>
> Original issue's description:
> > Add functionality which limits the number of bytes on the network.
> >
> > The limit is based on the bandwidth delay product, but also adds some additional slack to compensate for the sawtooth-like BWE pattern and the slowness of the encoder rate control. The delay is estimated based on the time from sending a packet until an ack is received. Since acks are received in bursts (feedback is only sent periodically), a min filter is used to estimate the rtt.
> >
> > Whenever the in flight bytes reaches the congestion window, the pacer is paused, which in turn will result in send-side queues growing. Eventually the encoders will be paused as the pacer queue grows large (currently 2 seconds).
> >
> > BUG=webrtc:7926
> >
> > Review-Url: https://codereview.webrtc.org/2918323002
> > Cr-Commit-Position: refs/heads/master@{#19289}
> > Committed: 8497fdde43
>
> TBR=terelius@webrtc.org,philipel@webrtc.org,tschumim@webrtc.org,gnish@webrtc.org
> # Not skipping CQ checks because original CL landed more than 1 days ago.
> BUG=webrtc:7926
>
> Review-Url: https://codereview.webrtc.org/3001653002
> Cr-Commit-Position: refs/heads/master@{#19339}
> Committed: 64136af364TBR=terelius@webrtc.org,philipel@webrtc.org,tschumim@webrtc.org,gnish@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/2994343002
Cr-Commit-Position: refs/heads/master@{#19373}
Reason for revert:
Speculative revert to see if this caused regressions in android perf tests.
Original issue's description:
> Add functionality which limits the number of bytes on the network.
>
> The limit is based on the bandwidth delay product, but also adds some additional slack to compensate for the sawtooth-like BWE pattern and the slowness of the encoder rate control. The delay is estimated based on the time from sending a packet until an ack is received. Since acks are received in bursts (feedback is only sent periodically), a min filter is used to estimate the rtt.
>
> Whenever the in flight bytes reaches the congestion window, the pacer is paused, which in turn will result in send-side queues growing. Eventually the encoders will be paused as the pacer queue grows large (currently 2 seconds).
>
> BUG=webrtc:7926
>
> Review-Url: https://codereview.webrtc.org/2918323002
> Cr-Commit-Position: refs/heads/master@{#19289}
> Committed: 8497fdde43TBR=terelius@webrtc.org,philipel@webrtc.org,tschumim@webrtc.org,gnish@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/3001653002
Cr-Commit-Position: refs/heads/master@{#19339}
Reason for revert:
Speculative revert to see if this caused regressions in android perf tests.
Original issue's description:
> Make the acceptable queue in the cwnd experiment configurable.
>
> BUG=webrtc:7926
>
> Review-Url: https://codereview.webrtc.org/2998753002
> Cr-Commit-Position: refs/heads/master@{#19320}
> Committed: 7c83c56b6dTBR=philipel@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/2999893002
Cr-Commit-Position: refs/heads/master@{#19337}
The limit is based on the bandwidth delay product, but also adds some additional slack to compensate for the sawtooth-like BWE pattern and the slowness of the encoder rate control. The delay is estimated based on the time from sending a packet until an ack is received. Since acks are received in bursts (feedback is only sent periodically), a min filter is used to estimate the rtt.
Whenever the in flight bytes reaches the congestion window, the pacer is paused, which in turn will result in send-side queues growing. Eventually the encoders will be paused as the pacer queue grows large (currently 2 seconds).
BUG=webrtc:7926
Review-Url: https://codereview.webrtc.org/2918323002
Cr-Commit-Position: refs/heads/master@{#19289}
I used a command like this to update the paths:
perl -pi -e "s/webrtc\/base/webrtc\/rtc_base/g" `find webrtc/rtc_base -name "*.cc" -o -name "*.h"`
The only manual edit is to add an include of webrtc/rtc_base/checks.h in
webrtc/modules/audio_device/android/opensles_common.h, which likely
was needed due to changed include paths due to 'git cl format'.
BUG=webrtc:7634
NOTRY=True
NOPRESUBMIT=True
Review-Url: https://codereview.webrtc.org/2969653002
Cr-Commit-Position: refs/heads/master@{#18871}
There are some functions in packet_router.cc and modules/congestion_controller that could be used by different threads, but they're protected using rtc::ThreadChecker which doesn't allow them to be called by more than one thread even if the calls are synchronised. This CL replaces those with rtc::RaceChecker, which allows serialized access of the functions from multiple threads.
BUG=webrtc:7826
Review-Url: https://codereview.webrtc.org/2940133003
Cr-Commit-Position: refs/heads/master@{#18628}
FakeRtpTransportController moves to a common header and its constructor is changed to take a SendSideCongestionController to enable injecting the mock.
BUG=webrtc:7395
Review-Url: https://codereview.webrtc.org/2834663003
Cr-Commit-Position: refs/heads/master@{#18055}
Instead of using the time on the first callback to Call::OnSentPacket, use the time when the first packet is sent from the pacer (to make sure this packet corresponds to an audio/video RTP packet).
BUG=webrtc:6244
Review-Url: https://codereview.webrtc.org/2825333002
Cr-Commit-Position: refs/heads/master@{#17777}
The point of this change is to make it possible to create the congestion
controller as part of creating RtpTransportController, later pass it to the
constructor of Call, and then let Call register itself as an observer.
BUG=webrtc:7135
Review-Url: https://codereview.webrtc.org/2795643002
Cr-Commit-Position: refs/heads/master@{#17504}
Implementation owned by call, and passed to VideoSendStream and
AudioSendStream.
BUG=webrtc:6847, webrtc:7135
Review-Url: https://codereview.webrtc.org/2685673003
Cr-Commit-Position: refs/heads/master@{#17389}
This CL is one in a series. To finish the work, the following CLs will be added:
1. CL for connecting RPLR as well
2. CL for RPLR-based FecController
3. CL for allowing experiment-driven configuration of the above (through both field-trials and protobuf)
BUG=webrtc:7058
Review-Url: https://codereview.webrtc.org/2638083002
Cr-Commit-Position: refs/heads/master@{#17365}
Reason for revert:
Problem was the rename of the include file. Intend to keep the old name for now, and then reland.
Original issue's description:
> Revert of Delete class MockCongestionController. (patchset #4 id:60001 of https://codereview.webrtc.org/2762023004/ )
>
> Reason for revert:
> This is breaking downstream build.
>
> Original issue's description:
> > Delete class MockCongestionController.
> >
> > It became unused in cl https://codereview.webrtc.org/2516983004.
> >
> > BUG=None
> >
> > Review-Url: https://codereview.webrtc.org/2762023004
> > Cr-Commit-Position: refs/heads/master@{#17325}
> > Committed: d19bcb7116
>
> TBR=stefan@webrtc.org,nisse@webrtc.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=None
>
> Review-Url: https://codereview.webrtc.org/2762133003
> Cr-Commit-Position: refs/heads/master@{#17330}
> Committed: e27f1e764eTBR=stefan@webrtc.org,skvlad@webrtc.org
BUG=None
Review-Url: https://codereview.webrtc.org/2766133002
Cr-Commit-Position: refs/heads/master@{#17338}
New class ReceiveSideCongestionController, extracted from CongestionController, and responsible for the
OnReceivedPacket processing.
Rest of the CongestionController moved to a new class
SendSideCongestionController.
To avoid breaking applications, CongestionController is redefined
as a union of these two classes, with no intended change in behavior.
With one exception: CongestionController::SetBweBitrates used to call
remote_bitrate_estimator_.SetMinBitrate, but after remote_bitrate_estimator_ was moved to ReceiveSideCongestionController,
it no longer does this.
BUG=webrtc:6847
Review-Url: https://codereview.webrtc.org/2752233002
Cr-Commit-Position: refs/heads/master@{#17321}
It was only assigned at construction, and this improves consistency
with remote_estimator_proxy_.
The declaration of the private WrappingBitrateEstimator had to be
moved to the header file, and it was also converted from
system_wrappers' CriticalSectionWrapper to rtc::CriticalSection.
BUG=webrtc:6847
Review-Url: https://codereview.webrtc.org/2642363003
Cr-Commit-Position: refs/heads/master@{#16236}
This lets the RTP code be unaware of lower layers, and the
SetTransportOverhead method is deleted from RTPSender and RtpRtcp.
Instead, that method is added to CongestionController and
TransportFeedbackAdapter, where it is more appropriate.
BUG=wertc:6847
Review-Url: https://codereview.webrtc.org/2589743002
Cr-Commit-Position: refs/heads/master@{#15995}
Problem fixed: RTP header extensions were not properly set in tests.
BUG=webrtc:5654
Review-Url: https://codereview.webrtc.org/2593963003
Cr-Commit-Position: refs/heads/master@{#15741}
Also deletes one call to CongestionController::pacer.
BUG=None
Review-Url: https://codereview.webrtc.org/2542113003
Cr-Commit-Position: refs/heads/master@{#15479}
And delete the method CongestionController::packet_router.
BUG=None
Review-Url: https://codereview.webrtc.org/2516983004
Cr-Commit-Position: refs/heads/master@{#15323}
Now ProbeController can send periodic bandwidth probes when in
application-limited region. This will allow to maintain correct
bottleneck bandwidth estimate, even not all bandwidth is being used.
The feature is not enabled by default, but can be enabled with a flag.
Interval between probes is currently set to 5 seconds.
BUG=webrtc:6332
Review-Url: https://codereview.webrtc.org/2504023002
Cr-Commit-Position: refs/heads/master@{#15279}
Reason for revert:
Unfortunately, this change breaks internal projects. Specifically the change to the CongestionController interface means anything implementing it will be forced to change in lock-step.
Original issue's description:
> Pass time constanct to bwe smoothing filter.
>
> BUG=webrtc:6443, webrtc:6303
>
> Committed: https://crrev.com/9abbf5ae4ec7d688a9b4aa03a405f3faadb74b90
> Cr-Commit-Position: refs/heads/master@{#15266}
TBR=minyue@webrtc.org,stefan@webrtc.org,solenberg@webrtc.org,michaelt@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:6443, webrtc:6303
Review-Url: https://codereview.webrtc.org/2532993002
Cr-Commit-Position: refs/heads/master@{#15272}
To achieve this some refactoring was done to make it possible to synchronize
access to the DelayBasedBwe in TransportFeedbackAdapter:
- The callback was removed from DelayBasedBwe, it now instead returns its
result.
- TransportFeedbackAdapter was moved to modules/congestion_controller to avoid
unnecessary dependencies.
Reenables previously disabled flaky test. Can no longer reproduce flakiness with gtest-parallel and asan/tsan builds.
BUG=webrtc:6427, webrtc:6422
R=terelius@webrtc.org
Review URL: https://codereview.webrtc.org/2378103005 .
Cr-Commit-Position: refs/heads/master@{#14452}
