As of https://webrtc-review.googlesource.com/c/src/+/158899, FEC may be
used on packets with VideoTimingExtension. This may result in creation
of FEC packets that exceed the maximum configured RTP packet size.
This problem occurs most frequently with datagram transports that define a
smaller maximum packet size.
Bug: webrtc:9719
Change-Id: I842216a6696a695f0a3f01a221e538605fc5b9bd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161557
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Bjorn Mellem <mellem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30045}
The RWLockWin::Create() function returns NULL on some Windows platforms because it cannot load kernel32.dll. This causes a crash.
RWLockWin tries to load kernel32.dll to check if the Slim Reader/Writer Lock APIs are present in kernel32.dll but on newer Windows platforms, kernel32.dll does not exist and the APIs are exported by kernelbase.dll instead.
The fix is quite simple: There is no need to try to load any DLL to check if the Slim Reader/Writer Lock APIs are present, because these APIs
are always present in all Windows versions since Windows Vista.
I am removing the code that attempts to load kernel32.dll. This prevents the crash on platforms that use kernelbase.dll.
If the WINUWP preprocessor symbol is defined, RWLockWin was already doing the right thing. But this issue is not limited to WINUWP and in
some scenarios, building for WINUWP is not the right solution because it causes other problems. So, my fix is essentially to use the WINUWP
code path for all Windows builds.
The only version of Windows which does not have the Slim Reader/Writer Lock APIs is Windows XP (and older ones, of course.)
However, since the current code does not fall back to an alternative implementation when the Slim Reader/Writer Lock APIs are missing,
WebRTC is already broken on such old versions of Windows.
Bug: webrtc:11186
Change-Id: I34aad066e18b924792d47c244ecee00669e86c4d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161472
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30044}
This reverts commit 18314bd8d2cb27fa58e4d304bbc428e3ed1736ba.
Reason for revert: Breaks downstream test.
Original change's description:
> Distinguish between send and receive video codecs
>
> Even though send and receive codecs are the same,
> they might have different support in HW.
> Distinguish between send and receive codecs to be able to keep
> track of which codecs have HW support.
>
> Bug: chromium:1029737
> Change-Id: I16a80da44c5061ca42f2aabda76e6bf0b879bf7b
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161306
> Reviewed-by: Anders Carlsson <andersc@webrtc.org>
> Reviewed-by: Steve Anton <steveanton@webrtc.org>
> Commit-Queue: Johannes Kron <kron@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#30041}
TBR=steveanton@webrtc.org,andersc@webrtc.org,kron@webrtc.org
Change-Id: I7e5807460006db613e9b3b369ec6036b88f164fd
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:1029737
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161662
Reviewed-by: Johannes Kron <kron@webrtc.org>
Commit-Queue: Johannes Kron <kron@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30042}
Even though send and receive codecs are the same,
they might have different support in HW.
Distinguish between send and receive codecs to be able to keep
track of which codecs have HW support.
Bug: chromium:1029737
Change-Id: I16a80da44c5061ca42f2aabda76e6bf0b879bf7b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161306
Reviewed-by: Anders Carlsson <andersc@webrtc.org>
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Commit-Queue: Johannes Kron <kron@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30041}
This CL removes the deprecated legacy AEC code.
Note that this CL should not be landed before the M80 release has been cut.
Bug: webrtc:11165
Change-Id: I59ee94526e62f702bb9fa9fa2d38c4e48f44753c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161238
Commit-Queue: Per Åhgren <peah@webrtc.org>
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30036}
If the task to call OnEncodedImage is posted to the encoder task queue
just after VideoStreamEncoder::Stop post the task to release the
encoder, the destruction sequence of java HardwareVideoEncoder
deadlocks in outputBuffersBusyCount.waitForZero();
Encoders are generally allowed to call OnEncodedImage on any internal
encoder thread, so posting to the encoder task queue seems unnecessary.
Bug: webrtc:9378
Change-Id: Iee14f151d9efdc5ab348f9c86069fdb762e6a0dc
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161447
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30035}
This is the start of generating compliant errors, including diagnostics,
when datachannels close because of errors.
Bug: chromium:1030631
Change-Id: I39aa41728efb25bca6193a782db4cbdaad8e0dc1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161304
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30034}
This is a reland of a0adf3d4409036d095480e9bfa0fc06990362f84
Original change's description:
> Reland "Implemented screen enumeration and selection for desktop capture under X11 using the X Resize and Rotate extension version 1.5."
>
> This is a reland of e7153012682ccd3d1eacc18f802cab7820e3bad3
>
> Original change's description:
> > Implemented screen enumeration and selection for desktop capture under X11 using the X Resize and Rotate entension version 1.5.
> >
> > Bug: chromium:396091
> > Change-Id: Ia1b36c771632c536bb8d15322461b479fabc409e
> > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/148768
> > Commit-Queue: Sergey Ulanov <sergeyu@chromium.org>
> > Reviewed-by: Sergey Ulanov <sergeyu@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#29083}
>
> Bug: chromium:396091
> Change-Id: I0d9171ae5f340e0489e4b45ce5d97bc52b0a4904
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/156067
> Commit-Queue: Tommi <tommi@webrtc.org>
> Reviewed-by: Tommi <tommi@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#29655}
Bug: chromium:396091
Change-Id: I47525911095fabc6cee613d03b0d83134b95b084
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/158900
Reviewed-by: Tomas Gunnarsson <tommi@chromium.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30032}
while suboptimal, these implementions are complete and allow to
swap code from using RtpDepacketizer interface to VideoRtpDepacketizer
Bug: webrtc:11152
Change-Id: Ie7823feeb5b0563b74754255aaedfada9d446ac5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161380
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30031}
This removes some code in the AudioDeviceWindowsCore::CoreAudioIsSupported function that was checking that every audio input and output device was functional. There are legitimate cases where some, or all, audio devices may not be accessible, and that was causing CoreAudioIsSupported to return false.
If CoreAudioIsSupported returns false, a subsequent RTC_CHECK call fails, which causes the entire app to exit.
After this change, the CoreAudioIsSupported() function simply checks if the Core Audio APIs are supported and no longer tries to do extra stuff unrelated to checking if the APIs are supported.
Note that Core Audio is actually supported in all versions of Windows after Windows XP. There were log messages in the code saying that if CoreAudioIsSupported() returns false, WebRTC will use the Wave Audio APIs instead. But this is no longer the case. The Wave Audio APIs would only be needed for Windows XP, and this code appears to have already been removed from WebRTC.
It is tempting to simply make CoreAudioIsSupported() do a "return true;" but for now I only removed the part of the logging messages that mentioned the Wave Audio APIs.
I understand that there is a new Audio Device Module (ADM) called WindowsCoreAudio2, which is now recommended for use by apps. Apps are supposed to instantiate WindowsCoreAudio2 and pass it in to WebRTC. When the app supplies its own ADM, CoreAudioIsSupported() does not get invoked, which avoids the bug. To help make it clearer that using WindowsCoreAudio2 is an acceptable solution, I am removing a comment that says that kWindowsCoreAudio2 is "experimental".
Bug: webrtc:11081
Change-Id: I7ed1684a276799f4c83006b45629e48814f0b18b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161463
Commit-Queue: Henrik Andreassson <henrika@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30025}
This CL follows the "Rule of zero".
Those constructors made no sense compared to default generated ones,
since all members are POD.
They were introduced to quiet a memory sanitizer warning,
which apparently was misleading.
As a bonus, the struct is now movable.
Bug: webrtc:11180, webrtc:9855
Change-Id: Iff9fd950bec8040bc6e7e7ece33cc49c5f453f5d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161381
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Yves Gerey <yvesg@google.com>
Cr-Commit-Position: refs/heads/master@{#30023}
This change avoids inadvertent capture of certain system windows (e.g.
the Start menu, other taskbar menus, and notification toasts) when
capturing a specific window on Windows.
It stops using EnumWindows for detection of overlapping windows, because
this API excludes these system windows from its enumeration. Using
FindWindowEx instead enumerates these windows.
The enumeration logic is refactored somewhat because a callback is no
longer necessary.
Bug: webrtc:10835
Change-Id: I1cccd44d6ef07f13a68e8daf2d2573d422001201
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161153
Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
Commit-Queue: Jamie Walch <jamiewalch@chromium.org>
Cr-Commit-Position: refs/heads/master@{#30022}
The comment was stale and setting -Wextra actually turns on diagnostics
that are turned off by Chromium.
For example:
"-Wextra -Wno-deprecated-copy -Wextra" will turn on -Wdeprecated-copy
because starting from https://reviews.llvm.org/D70342
-Wdeprecated-copy is part of -Wextra.
Bug: webrtc:11180
Change-Id: Ia5d1e22bfe42d67cd892ae07620e7fd2daf9a7a4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161332
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30019}
Cause: VideoRtpReceiver::media_channel_ was used when it was null.
Fix: only use when provably not null.
Bug: chromium:1031013
Change-Id: I765e183186d895f39c122e26d50ac787216c44f7
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161328
Commit-Queue: Markus Handell <handellm@webrtc.org>
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30017}
This allows one to request the same sequence number again
in the case of resending an FIR to the a sender before the
sender has time to send a key-frame.
Bug: webrtc:11171
Change-Id: Idd8e8120ccbcc194cefb8d0cf3f7cc64e7f76aa5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161236
Commit-Queue: Evan Shrubsole <eshr@google.com>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30006}
This check just makes it more clear what the expectations are.
Pululating trials was made mandatory in an earlier CL, but if you don't
populate this field it will trigger a DCHECK at lower layer where we're
actually trying to parse an experiment. That is confusing and
misleading.
Bug: None
Change-Id: I1f520841a5a3b911048c8ee6d309eb7bb179e037
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161301
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30005}
This change finally wires up VideoRtpReceiver::OnGenerateKeyFrame and
OnEncodedSinkEnabled into internal::VideoReceiveStream so that encoded
frames can flow to sinks installed in VideoTrackSourceInterface.
Bug: chromium:1013590
Change-Id: I76f8226752294aee8fe137d1a78ee66548900cc2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161095
Commit-Queue: Markus Handell <handellm@webrtc.org>
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30003}
If WebRTC-AutomaticAnimationDetectionScreenshare experiment is enabled,
content type is screenshare and degradation preference is BALANCED,
then input resolution is restricted if update_rect of the incoming frames
is the same for considerable amount of time and is big enough.
This entails treating BALANCED degradation preference for screenshare as
MAINTAIN_RESOLUTION in adaptation logic.
Bug: webrtc:11058
Change-Id: I903dddf53fcbd7c8eac6c5b1447225b15fd8fe5f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161097
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30002}
This change ultimately enables wiring up VideoRtpReceiver::OnGenerateKeyFrame
and OnEncodedSinkEnabled into internal::VideoReceiveStream so that encoded
frames can flow to sinks installed in VideoTrackSourceInterface.
Bug: chromium:1013590
Change-Id: I136132c210e5811547f2522ddc371d0acac90664
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161093
Commit-Queue: Markus Handell <handellm@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30001}
This makes the WebRTC-KeepAbsSendTimeExtension field trial
always enabled. This means that we no longer avoid sending the
abs-send-time extension if we have negotiated sending of transport
wide sequence numbers.
The field trial WebRTC-FilterAbsSendTimeExtension is introduced to allow
reverting to the previous behavior.
Bug: webrtc:10234
Change-Id: Ifd9761d84dd1fe79af840f98ad0882a2e5adf0b0
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/159181
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Konrad Hofbauer <hofbauer@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29999}
This CL ensures that the high-pass filter is on whenever the echo
controller is on. This is important as the echo controller code assumes
that the external high-pass filter is active.
The CL also corrects the ToggleAec unit test (which started failing
after this code change).
Bug: webrtc:11159, chromium:1030179
Change-Id: Ief86eda8f7c67df1c25ac1a06d2cc0778e01196d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161228
Commit-Queue: Per Åhgren <peah@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29998}
I could find no reason for the extra complexity of doing messaging
in order to schedule a task to be done after the current cycle.
It also simplifies the peerconnection/datachannelcontroller coupling.
Bug: webrtc:11146
Change-Id: I68f45059b9f4a6869fb44b856e05a480f4652365
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161232
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#29997}
