This is a reland of 0cbcba7ea0dced1a7f353c64d6cf91d46ccb29f9.
Original change's description:
> Moved congestion controller to task queue.
>
> The goal of this work is to make it easier to experiment with the
> bandwidth estimation implementation. For this reason network control
> functionality is moved from SendSideCongestionController(SSCC),
> PacedSender and BitrateController to the newly created
> GoogCcNetworkController which implements the newly created
> NetworkControllerInterface. This allows the implementation to be
> replaced at runtime in the future.
>
> This is the first part of a split of a larger CL, see:
> https://webrtc-review.googlesource.com/c/src/+/39788/8
> For further explanations.
>
> Bug: webrtc:8415
> Change-Id: I770189c04cc31b313bd4e57821acff55fbcb1ad3
> Reviewed-on: https://webrtc-review.googlesource.com/43840
> Commit-Queue: Sebastian Jansson <srte@webrtc.org>
> Reviewed-by: Björn Terelius <terelius@webrtc.org>
> Reviewed-by: Stefan Holmer <stefan@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21868}
Bug: webrtc:8415
Change-Id: I1d1756a30deed5b421b1c91c1918a13b6bb455da
Reviewed-on: https://webrtc-review.googlesource.com/48000
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21899}
This can be helpful when debugging or analyzing logcats.
Bug: webrtc:8841
Change-Id: Ic3a18ee68321edbffd92e57ccb84a7b2710e16bd
Reviewed-on: https://webrtc-review.googlesource.com/47881
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21897}
VideoEncoderWrapper may be released and reused (Release() followed by
InitEncode()). This often happens back to back when encoders are
reconfigured. Because encoded frames are posted asynchronously to the
encoder queue, they may be processed after the encoder associated with
them has already been released.
In the existing code, if a frame for the new encoder had already been
received, the processing of the frame for the old encoder would clear
out the record for the new encoder's frame. This is now fixed by only
clearing out records that are older than the encoded frame being
processed.
A particularly bad symptom is when the new encoder is used for the same
stream as the old one (but was reconfigured for e.g. a change in
resolution). In that case, the new encoder's initial keyframe gets
dropped, and all subsequent difference frames are based off the last
sent frame from the old encoder. This all renders as garbage until a new
keyframe is sent.
Bug: webrtc:8849
Change-Id: I25094f12b38e03e158dc10ac66e92aa9ebaa5541
Reviewed-on: https://webrtc-review.googlesource.com/47549
Commit-Queue: Jonathan Yu <yujo@chromium.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21896}
OpenSSL is making a lot of data structure opaque, so we can no longer directly access internal data structure. Fortunately, API methods are provided for this purpose.
BoringSSL is sharing the same API.
Bug: webrtc:8817
Change-Id: Ia5090200f0e7c352f82e8191720ac4c14fbb5a85
Reviewed-on: https://webrtc-review.googlesource.com/47321
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Justin Uberti <juberti@webrtc.org>
Reviewed-by: Emad Omara <emadomara@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21895}
Translate --isolate-script-test-perf-output to --chartjson_result_file
and pass it to the test.
That way we can use Chromium's recipe code to report results to the
Perf dashboard.
TBR=phoglund@webrtc.org
Bug: chromium:807737
Change-Id: I2d3479fe29431cc1a8faf9a73b054a5f4ec610a4
Reviewed-on: https://webrtc-review.googlesource.com/47121
Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21894}
WHAT: made a BUILD.gn with library and tests in the Audio Processing
Module Voice Activity Detector directory. Updated depending
code. Fixed a Clang warning.
WHY: to make it possible for a target to depend on just the VAD and
not the whole APM. There are other benefits:
* Sometimes faster compilation.
* The VAD takes up 28000 bytes of libjingle_peerconnection_so.so. Making
a peerconnection shared object file without the VAD has to be done in
steps. The first step is a custom target for the VAD. Hence this Cl.
Change-Id: Iea0207a0b5979db26baaf46b24beaefbb1c431af
BUG: webrtc:5716, webrtc:7494
Reviewed-on: https://webrtc-review.googlesource.com/47521
Commit-Queue: Alex Loiko <aleloi@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Reviewed-by: Oleh Prypin <oprypin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21893}
These files were copied to Native/src but were kept around for
downstream projects that included them from their old locations.
Downstream projects have been updated so these can now be removed.
Bug: webrtc:8832
Change-Id: Ic28dc13e4b5bfced4b97ee872068683785d04bb3
Reviewed-on: https://webrtc-review.googlesource.com/47860
Reviewed-by: Kári Helgason <kthelgason@webrtc.org>
Commit-Queue: Anders Carlsson <andersc@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21892}
This patch modifies behaviour when TurnPort gets a lifetime
back from server that is shorter than 2 minutes.
Before the patch such lifetime resulted in TurnPort not scheduling any
refresh, leading to timeout on the turn allocation.
After then patch lifetime shorter then 2 minutes leads to refresh
after half stipulated lifetime.
BUG=webrtc:8826
Change-Id: I80561100f2307bd9a6a91af0924bb2814102ddd3
Reviewed-on: https://webrtc-review.googlesource.com/46741
Commit-Queue: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21891}
This helps separate concerns, so that the VideoProcessorIntegrationTest
is almost oblivious to the fact that it needs to connect to the JVM
to get the Android HW codecs.
Bug: webrtc:8448
Change-Id: I4359b31f84be48eaf99d83525bcce6e593e874f8
Reviewed-on: https://webrtc-review.googlesource.com/47384
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Commit-Queue: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21890}
This change list passes the instance of RtcEventLog from Peerconnection
down to P2PTransportChannel, and binds the structured ICE logging with
ICE layer objects. Logs of ICE connectivity checks are injected for
candidate pairs.
TBR=terelius@webrtc.org
Bug: None
Change-Id: Ia979dbbac6d31dcf0f8988da1065bdfc3e461821
Reviewed-on: https://webrtc-review.googlesource.com/34660
Commit-Queue: Qingsi Wang <qingsi@google.com>
Reviewed-by: Peter Thatcher <pthatcher@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21884}
This removes use of the MockPeerConnection and replaces it with the
FakePeerConnectionForStats testing class.
Bug: webrtc:8764
Change-Id: I78553c5a4e4d68cb6666a83f443f72f7c25488dc
Reviewed-on: https://webrtc-review.googlesource.com/46940
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21882}
This is a reland of eed5aa8904d09179971d3f4e7e10c109d7c62bfc
Original change's description:
> Structured ICE logging via RtcEventLog.
>
> This change list contains the structured logging module for ICE using
> the RtcEventLog infrastructure, and also extension to the log parser
> and analyzer.
>
> Bug: None
> Change-Id: I6539cf282155c2cde4d3161c53500c0746671a02
> Reviewed-on: https://webrtc-review.googlesource.com/34622
> Commit-Queue: Qingsi Wang <qingsi@google.com>
> Reviewed-by: Björn Terelius <terelius@webrtc.org>
> Reviewed-by: Peter Thatcher <pthatcher@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21816}
TBR=pthatcher@webrtc.org,terelius@webrtc.org,deadbeef@webrtc.org
Bug: None
Change-Id: I3df585bf636315ceb0273967146111346a83be86
Reviewed-on: https://webrtc-review.googlesource.com/47545
Commit-Queue: Qingsi Wang <qingsi@google.com>
Reviewed-by: Qingsi Wang <qingsi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21881}
When a simulcast stream is enabled or disabled, we want this state
change to be reflected properly in the RtpRtcp modules. Each video send
stream can contain multiple rtp_rtcp_modules pertaining to different
simulcast streams. These modules are currently all turned on/off when
the send stream is started and stopped. This change allows for
individual modules to be turned on/off. This means if a module stops
sending it will send a bye message, so the receiving side will not
expect more frames to be sent when the stream is inactive and the
encoder is no longer encoding/sending images.
Bug: webrtc:8653
Change-Id: Ib6d00240f627b4ff1714646e847026f24c7c3aa4
Reviewed-on: https://webrtc-review.googlesource.com/42841
Commit-Queue: Seth Hampson <shampson@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21880}
This reverts commit 0cbcba7ea0dced1a7f353c64d6cf91d46ccb29f9.
Reason for revert: Major regressions on perf bots.
Original change's description:
> Moved congestion controller to task queue.
>
> The goal of this work is to make it easier to experiment with the
> bandwidth estimation implementation. For this reason network control
> functionality is moved from SendSideCongestionController(SSCC),
> PacedSender and BitrateController to the newly created
> GoogCcNetworkController which implements the newly created
> NetworkControllerInterface. This allows the implementation to be
> replaced at runtime in the future.
>
> This is the first part of a split of a larger CL, see:
> https://webrtc-review.googlesource.com/c/src/+/39788/8
> For further explanations.
>
> Bug: webrtc:8415
> Change-Id: I770189c04cc31b313bd4e57821acff55fbcb1ad3
> Reviewed-on: https://webrtc-review.googlesource.com/43840
> Commit-Queue: Sebastian Jansson <srte@webrtc.org>
> Reviewed-by: Björn Terelius <terelius@webrtc.org>
> Reviewed-by: Stefan Holmer <stefan@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21868}
TBR=terelius@webrtc.org,stefan@webrtc.org,srte@webrtc.org
Change-Id: Ia8a273eb9e92b7d0d960c49658c228208170962d
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:8415
Reviewed-on: https://webrtc-review.googlesource.com/47560
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21877}
end_to_end_tests.cc was over 5000 lines and covered many different
areas in it's testing. In this change it is separated into multiple
smaller test sets separated by the functionality they are testing. The
reasoning behind this is that the fact that a test is working end to end
should be secondary to what functionality the test is actually testing.
A slight functional change is that for some of the tests the
parametrization over round robin pacing being controlled with a field
trial is removed since they are simple enough that they should not be
affected by the pacing method.
Bug: None
Change-Id: I4b7eba80fc142ecfc8fa642dab9b6f587d914048
Reviewed-on: https://webrtc-review.googlesource.com/46143
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21876}
WebRTC internal code should always use include paths that start
from the root of the project and that clearly identify the header file.
This allows 'gn check' to actually keep dependencies under control
because 'gn check' cannot enforce anything if the include path
is not fully qualified (starting from the root of the project).
Bug: webrtc:8815
Change-Id: I8a7ab64dfecdb3da4099fdec61e5fc27af4f8ccc
Reviewed-on: https://webrtc-review.googlesource.com/47380
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21874}
So we can report perf results using JSON and not parsing stdout.
I reordered the way the arguments are parsed, so that options go
at the end, and not at the middle, which is an awkward place to put them.
Regular usage specifying [-I], bottleneck_value, infile and outfile
shouldn't be affected.
Bug: chromium:807737
Change-Id: Ida863846400326c33e443d723f384971b891b6e5
Reviewed-on: https://webrtc-review.googlesource.com/47161
Reviewed-by: Henrik Lundin <henrik.lundin@webrtc.org>
Reviewed-by: Oleh Prypin <oprypin@webrtc.org>
Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21873}
In MediaCodecVideoEncoder, VideoFrame timestamp was used as a
presentation timestamp. With this change timestamp maintained in C++
code is used instead. This matches the behaviour with old frame
callbacks.
Bug: b/72832862
Change-Id: I1f0543ebe837ccac22c83a81a81f3ea128e2a866
Reviewed-on: https://webrtc-review.googlesource.com/47381
Reviewed-by: Anders Carlsson <andersc@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21872}
WebRTC internal code should always use include paths that start
from the root of the project and that clearly identify the header file.
This allows 'gn check' to actually keep dependencies under control
because 'gn check' cannot enforce anything if the include path
is not fully qualified (starting from the root of the project).
Bug: webrtc:8815
Change-Id: I36f01784fa5f5b77eefc02db479b1f7f6ee1a8c3
Reviewed-on: https://webrtc-review.googlesource.com/46263
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21871}
The goal of this work is to make it easier to experiment with the
bandwidth estimation implementation. For this reason network control
functionality is moved from SendSideCongestionController(SSCC),
PacedSender and BitrateController to the newly created
GoogCcNetworkController which implements the newly created
NetworkControllerInterface. This allows the implementation to be
replaced at runtime in the future.
This is the first part of a split of a larger CL, see:
https://webrtc-review.googlesource.com/c/src/+/39788/8
For further explanations.
Bug: webrtc:8415
Change-Id: I770189c04cc31b313bd4e57821acff55fbcb1ad3
Reviewed-on: https://webrtc-review.googlesource.com/43840
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21868}
The VideoCodecTest class is a fixture base class for the
libvpx-VP8, libvpx-VP9, and OpenH264 unit tests. It is unrelated
to the VideoProcessor tests, which we colloquially refer to as
the "codec test".
This rename is thus to reduce this confusion. It should have no
functional impact.
Bug: webrtc:8448
Change-Id: If73443bda5df0f29a71ce6ce069ac128795ff0ad
Reviewed-on: https://webrtc-review.googlesource.com/47160
Reviewed-by: Sergey Silkin <ssilkin@webrtc.org>
Commit-Queue: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21867}
Previously, the code which reported cipher stats to UMA for all
transports would classify the media type based on the transport name,
which is brittle and misleading with BUNDLE. This corrects the code to
track all media types (audio, video, data) which use the transport and
report once for each.
Bug: None
Change-Id: I8506f64f0011788b744b8386ac58518a21914b52
Reviewed-on: https://webrtc-review.googlesource.com/47247
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21863}
This is a reland of 001546da953275c7a39eb220592b440c9b47d756
Original change's description:
> Break up rtc_event_log_api to solve circular dependencies.
>
> The original rtc_event_log_api is refactored to a pure API target plus
> multiple targets coupled with WebRTC implementations.
>
> Bug: None
> Change-Id: Iab9eee3f7bf4228c52d94a5f26fc39bb99b5033f
> Reviewed-on: https://webrtc-review.googlesource.com/43247
> Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
> Reviewed-by: Björn Terelius <terelius@webrtc.org>
> Reviewed-by: Stefan Holmer <stefan@webrtc.org>
> Commit-Queue: Qingsi Wang <qingsi@google.com>
> Cr-Commit-Position: refs/heads/master@{#21811}
TBR=pthatcher@webrtc.org,deadbeef@webrtc.org,terelius@webrtc.org,stefan@webrtc.org
Bug: None
Change-Id: I3e7213733741cbfd5dd0076f32209e6bc42a0647
Reviewed-on: https://webrtc-review.googlesource.com/46900
Commit-Queue: Qingsi Wang <qingsi@google.com>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21862}
https://chromium-review.googlesource.com/c/external/webrtc/+/575910 pretty much made it a mandate to have OpenSSL 1.1.0 to compile webrtc.
So, let's be explicit about it and cleanup old code for older version support.
Also, generate a compiler error for older OpenSSL versions.
Bug: webrtc:8817
Change-Id: I28590348137b6a04503eabdcc6328297ecf5213e
Reviewed-on: https://webrtc-review.googlesource.com/46502
Reviewed-by: Justin Uberti <juberti@webrtc.org>
Commit-Queue: Jiawei Ou <ouj@fb.com>
Cr-Commit-Position: refs/heads/master@{#21861}
Original change's description:
> Parameterize PeerConnection signaling tests for Unified Plan
>
> This also changes the behavior of CreateAnswer to fail unless
> the signaling state is kHaveRemoteOffer or kHaveLocalPranswer,
> as per the WebRTC specification.
>
> Bug: webrtc:8765
> Change-Id: I60ac67cd92b17fcbff964afc14d049481e816a28
> Reviewed-on: https://webrtc-review.googlesource.com/41042
> Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
> Commit-Queue: Steve Anton <steveanton@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21779}
Bug: webrtc:8813
Change-Id: I9f608fcd0b7aca00b4c1092e271dbd9cd710c38a
Reviewed-on: https://webrtc-review.googlesource.com/46861
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21860}
This is intended to ensure compatibility between Plan B and
Unified Plan endpoints for the single audio - single video case.
If Unified Plan is the offerer, it will add a=msid and a=ssrc MSID
entries to its offer.
If Unified Plan is the answerer, it will use whatever MSID
signaling mechanism was used in the offer (either a=msid or
a=ssrc).
Bug: webrtc:7600
Change-Id: I6192dec19123fbb56f5d04540d2175c7fb30b9b6
Reviewed-on: https://webrtc-review.googlesource.com/44162
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21859}
Instead of chartjson-result-file, since that's the flag passed by the recipe.
TBR=phoglund@webrtc.org
Bug: chromium:807737
Change-Id: I3a679ab7e5c0a446e675d0f4647344cc4194b357
Reviewed-on: https://webrtc-review.googlesource.com/46541
Commit-Queue: Edward Lemur <ehmaldonado@webrtc.org>
Reviewed-by: Oleh Prypin <oprypin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21858}
This changes the behavior of CreateOffer/CreateAnswer when Unified
Plan is enabled to be in line with that specified in JSEP.
In particular, MSID information is now only included if the
RtpTransceiver is not stopped and either is sending or has ever
sent.
Bug: webrtc:7600
Change-Id: I6400f0583525c7776331eeb0e1bb53973bc02dfb
Reviewed-on: https://webrtc-review.googlesource.com/46400
Commit-Queue: Oleh Prypin <oprypin@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21857}
This is done to solve a problem where a string literal is implicitly cast
to a temporary std::string when calling webrtc::test::InitFieldTrialsFromString
which passes a pointer to the internal representation to
webrtc::field_trial::InitFieldTrialFromString(char*). This pointer is
stored for later use, but the temporary std::string is destroyed as soon
as the function returns.
Using webrtc::field_trial::InitFieldTrialFromString(char*) instead,
avoids the implicit casts (but the caller still needs to ensure that
the char* outlives the program). The validation previously done by
webrtc::test::InitFieldTrialsFromString can now be done by manually
calling webrtc::test::ValidateFieldTrialsStringOrDie(const std::string&).
Add system_wrappers:field_trial_default as a direct dependency to
various targets to allow including the field_trials_default.h header.
Bug: webrtc:8812
Change-Id: Ib5a641ea255b1c16a8f7f35e1fe67f6c38a61da6
Reviewed-on: https://webrtc-review.googlesource.com/46141
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Oleh Prypin <oprypin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21856}
