Biggest change is to Remove MediaType as argument to RtcEventLog::LogRtpHeader and RtcEventLog::LogRtcpHeader.
Since the type is used by tools, these tools are rewritten to figure out the media type from the configurations instead.
BUG=webrtc:7538
TBR=solenberg@webrtc.org // For call.cc and voiceengine.cc
Review-Url: https://codereview.webrtc.org/2855143002
Cr-Commit-Position: refs/heads/master@{#18324}
Before this CL, the RTP state would be re-randomized after a
recreation of VideoSendStream. That might lead to us sending
a non-compliant RTP stream, which is avoided after the
changes in this CL.
BUG=webrtc:5654
TBR=pbos@webrtc.org # Trivial change to fuzzer.
Review-Url: https://codereview.webrtc.org/2912713002
Cr-Commit-Position: refs/heads/master@{#18322}
This change:
Reduces complexity for audio playout by removing a redundant memcopy in the output audio path.
Adds support for iOS simulator for playout since we now allow the audio layer to ask for different sizes
of audio buffers at each callback. Real iOS devices always asks for the same size, simulators does not.
This change comes without any new cost for real devices.
BUG=b/37580746
Review-Url: https://codereview.webrtc.org/2894873002
Cr-Commit-Position: refs/heads/master@{#18321}
Reason for revert:
Take three of relanding this after all internal issues have been resolved.
Original issue's description:
> Revert of Adding backward compatibility header (patchset #1 id:1 of https://codereview.webrtc.org/2909923002/ )
>
> Reason for revert:
> Breaks downstream project.
>
> Original issue's description:
> > Adding backward compatibility header
> >
> > This header will be removed ad soon as downstream projects will be
> > updated.
> >
> > BUG=webrtc:4867
> > NOTRY=True
> > TBR=kjellander@webrtc.org
> >
> > Review-Url: https://codereview.webrtc.org/2909923002
> > Cr-Commit-Position: refs/heads/master@{#18306}
> > Committed: 4fa8a97cc3
>
> TBR=kjellander@webrtc.org,magjed@webrtc.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:4867
>
> Review-Url: https://codereview.webrtc.org/2908143003
> Cr-Commit-Position: refs/heads/master@{#18308}
> Committed: 6684fdce93TBR=kjellander@webrtc.org,magjed@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:4867
Review-Url: https://codereview.webrtc.org/2912773003
Cr-Commit-Position: refs/heads/master@{#18320}
After this CL, reconfiguring the FlexFEC payload type at the
WebRtcVideoChannel2 level will no longer lead to the recreation of
the VideoReceiveStream. This means that the jitter buffer will
not be destroyed and a smoother video playback is achieved during
SDP renegotiation.
BUG=webrtc:5654
Review-Url: https://codereview.webrtc.org/2911913002
Cr-Commit-Position: refs/heads/master@{#18318}
Previously, the base class PlanarYuvBuffer was used directly. Having
separate base classes will allow us to improve type safety in some
places.
BUG=webrtc:7632
TBR=stefan@webrtc.org
Review-Url: https://codereview.webrtc.org/2914463002
Cr-Commit-Position: refs/heads/master@{#18317}
This CL removes |default_recv_ssrc_| from DefaultUnsignalledSsrcHandler
and replaces it with calls to a new member function
WebRtcVideoChannel2::GetDefaultReceiveStreamSsrc. The latter checks
the |default_stream_| member on the
WebRtcVideoChannel2::WebRtcVideoReceiveStreams to know which stream
is the current default stream.
This change removes duplicate state and fixes an issue where
incoming unsignaled SSRCs would compete for being the default
receive stream.
BUG=webrtc:7725
Review-Url: https://codereview.webrtc.org/2906893002
Cr-Commit-Position: refs/heads/master@{#18314}
Reason for revert:
Breaks downstream project.
Original issue's description:
> Adding backward compatibility header
>
> This header will be removed ad soon as downstream projects will be
> updated.
>
> BUG=webrtc:4867
> NOTRY=True
> TBR=kjellander@webrtc.org
>
> Review-Url: https://codereview.webrtc.org/2909923002
> Cr-Commit-Position: refs/heads/master@{#18306}
> Committed: 4fa8a97cc3TBR=kjellander@webrtc.org,magjed@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:4867
Review-Url: https://codereview.webrtc.org/2908143003
Cr-Commit-Position: refs/heads/master@{#18308}
Reason for revert:
Revert of revert of revert of revert of 'Activating..'. Or "reland of reland of 'Activate..'".
*Now* the internal projects are fixed and the fix is verified.
Original issue's description:
> Revert of Activate 'offload debug dump recordings from audio thread to TaskQueue'. (patchset #1 id:1 of https://codereview.webrtc.org/2903153005/ )
>
> Reason for revert:
> Reverting again: internal project issues were apparently not completely fixed.
>
> Original issue's description:
> > Reland of Activate 'offload debug dump recordings from audio thread to TaskQueue'. (patchset #1 id:1 of https://codereview.webrtc.org/2904893002/ )
> >
> > Reason for revert:
> > Revert the revert now that internal projects are updated.
> >
> > Original issue's description:
> > > Revert of Activate 'offload debug dump recordings from audio thread to TaskQueue'. (patchset #4 id:160001 of https://codereview.webrtc.org/2896813002/ )
> > >
> > > Reason for revert:
> > > Breaks internal project.
> > >
> > > Original issue's description:
> > > > Activate 'offload debug dump recordings from audio thread to TaskQueue'.
> > > >
> > > > A low priority task queue is added to WebRTCVoiceEngine. The
> > > > start/stop debug calls make file logging happen on the task queue.
> > > >
> > > > In a dependent CL (https://codereview.webrtc.org/2888303003), the task queue is moved to PeerConnectionFactory,
> > > > so that it can be shared for low priority tasks between different
> > > > subcomponents. It will require some changes to MediaEngine,
> > > > CompositeMediaEngine, WebRTCVoiceEngine, and changes in internal
> > > > projects.
> > > >
> > > > A task queue must be created and destroyed from the same thread. With
> > > > this CL that will be the worker thread, which creates and destroys
> > > > WebRTCVoiceEngine. With the dependent CL, it will probably change to
> > > > the signaling thread.
> > > >
> > > > NOTRY=True # tests just passed
> > > >
> > > > BUG=webrtc:7404
> > > >
> > > > Review-Url: https://codereview.webrtc.org/2896813002
> > > > Cr-Commit-Position: refs/heads/master@{#18252}
> > > > Committed: c61bf947b4
> > >
> > > TBR=solenberg@webrtc.org,tommi@webrtc.org,perkj@webrtc.org,danilchap@webrtc.org,tommi@chromium.org
> > > # Skipping CQ checks because original CL landed less than 1 days ago.
> > > NOPRESUBMIT=true
> > > NOTREECHECKS=true
> > > NOTRY=true
> > > BUG=webrtc:7404
> > >
> > > Review-Url: https://codereview.webrtc.org/2904893002
> > > Cr-Commit-Position: refs/heads/master@{#18255}
> > > Committed: be68b72cfa
> >
> > TBR=solenberg@webrtc.org,tommi@webrtc.org,perkj@webrtc.org,danilchap@webrtc.org,tommi@chromium.org
> > # Not skipping CQ checks because original CL landed more than 1 days ago.
> > BUG=webrtc:7404
> >
> > Review-Url: https://codereview.webrtc.org/2903153005
> > Cr-Commit-Position: refs/heads/master@{#18270}
> > Committed: d2303a2338
>
> TBR=solenberg@webrtc.org,tommi@webrtc.org,perkj@webrtc.org,danilchap@webrtc.org,tommi@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:7404
>
> Review-Url: https://codereview.webrtc.org/2910633002
> Cr-Commit-Position: refs/heads/master@{#18272}
> Committed: fe9ecb07eaTBR=solenberg@webrtc.org,tommi@webrtc.org,perkj@webrtc.org,danilchap@webrtc.org,tommi@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:7404
Review-Url: https://codereview.webrtc.org/2904423002
Cr-Commit-Position: refs/heads/master@{#18300}
Reason for revert:
According to crbug.com/726706 this is now fixed so let's give it a try.
Original issue's description:
> Disable Goma for Linux32 (ARM) bots.
>
> This config is having issues to build reliably on Goma.
>
> BUG=chromium:726706
> TBR=mbonadei@webrtc.org
>
> Review-Url: https://codereview.webrtc.org/2911893002 .
> Cr-Commit-Position: refs/heads/master@{#18295}
> Committed: 3a343ad329TBR=mbonadei@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:726706
Review-Url: https://codereview.webrtc.org/2910963002
Cr-Commit-Position: refs/heads/master@{#18299}
The reason is that there might be implementations that do not want to implement all methods.
To allow easier modification of the RtcEventLog interface, allow these implementation to inherit the RtcEventLogNullImpl implementation.
BUG=none
Review-Url: https://codereview.webrtc.org/2903003002
Cr-Commit-Position: refs/heads/master@{#18298}
The previous code attempted to lock instance_count and instance with a
CriticalSection, but the CriticalSection was not static, so each
function invocation got its own instance. Locking this call-specific
instance doesn't actually stop any other threads from concurrently
accessing the same function-scope globals, so this function had a data
race, which broke tsan tests (and possibly other things).
Making the CriticalSection shared among function calls will actually
synchronize access to the globals and allow our tsan tests to pass.
BUG=webrtc:3062
Review-Url: https://codereview.webrtc.org/2890213002
Cr-Commit-Position: refs/heads/master@{#18296}
AsyncStunTCPSocket wasn't firing SignalSentPacket, which the bandwidth
estimator requires for every packet in order to look up send times when
feedback arrives. If the signal isn't fired, it always assumes feedback
is arriving extremely late, and decreases the bandwidth by a factor of
2 until it reaches the minimum of 10kbps.
BUG=webrtc:7717
TBR=pthatcher@webrtc.org
Review-Url: https://codereview.webrtc.org/2912523003
Cr-Commit-Position: refs/heads/master@{#18279}
This can occur (and by default, terminates the process) for apps that
don't use the "voip" UIBackgroundMode.
We're already doing a similar thing on Linux (using MSG_NOSIGNAL for every
packet sent).
BUG=webrtc:7686
Review-Url: https://codereview.webrtc.org/2903313002
Cr-Commit-Position: refs/heads/master@{#18277}
