Except for a use of std::multimap (for which there currently is no
drop-in replacement), use webrtc::flat_map and flat_set for increased
performance.
The number of ssrcs/mids/payload types/etc is likely to be small and is
generally updated very rarely, but looked up a lot.
RtpDemuxer::ResolveSink's use of std::map represents about 0.6% CPU in
heavily loaded media servers. It does quite a few map lookups, and by
taking advantage of the increased cache locality of the flat_map and
flat_set containers, performance should be increased.
A previous attempt to use std::unordered_map failed in change 216243,
which was reverted. This is the second attempt.
Bug: webrtc:12689
Change-Id: Ifdbec63b2fd8f2f52e45ebaf12834b11dd7a41c5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/224662
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Victor Boivie <boivie@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34429}
This is a step in the direction of being able to make configuration
changes without having to tear down and reconstruct the object
during renegotiation.
Bug: none
Change-Id: If594fd41f3a561060f64212c479a25d19adf8598
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/223740
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34402}
This change achieves an Idle Wakeup savings of 200 Hz.
ModuleRtcRtcpImpl2 had Process() logic only active if TMMBR() is
enabled in RtcpSender, which it never is. Hence the Module
inheritance could be removed. The change removes all known
dependencies of the module inheritance, and any related mentions
of ProcessThread.
Fixed: webrtc:11581
Change-Id: I440942f07187fdb9ac18186dab088633969b340e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/222604
Reviewed-by: Tommi <tommi@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Markus Handell <handellm@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34358}
* Remove unnecessary decoder factory pointer.
* Set video decoder factory in the ctor of the config class.
* Prepare SetRecvParameters for not needing RecreateWebRtcVideoStream.
Bug: none
Change-Id: I48fbf2920c9fe50f3995ceab5667eb2f70618f25
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/223067
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34351}
This is a reland of e9ae4729e03f60dbe3b1828dd9009b401097cd3f
TBR=philipel@webrtc.org,terelius@webrtc.org
Original change's description:
> Correctly handle retransmissions/padding in early loss detection.
>
> This CL makes sure we don't cull packets from the history based on
> incorrect ack mapping, just like it's predecessor:
> https://webrtc-review.googlesource.com/c/src/+/218000
>
> It also changes the logic to make sure retransmits counts towards
> history pruning - and properly ignores padding/fec.
>
> Bug: webrtc:12713
> Change-Id: I7835d10e483687e960a9cce41d4e2f1a6c3189b4
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221863
> Reviewed-by: Björn Terelius <terelius@webrtc.org>
> Reviewed-by: Philip Eliasson <philipel@webrtc.org>
> Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
> Commit-Queue: Erik Språng <sprang@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#34293}
Bug: webrtc:12713
Change-Id: Iec123d71edafea98fe289acde007b57e212681f4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/222640
Commit-Queue: Erik Språng <sprang@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34297}
This reverts commit e9ae4729e03f60dbe3b1828dd9009b401097cd3f.
Reason for revert: Internal test failure
Original change's description:
> Correctly handle retransmissions/padding in early loss detection.
>
> This CL makes sure we don't cull packets from the history based on
> incorrect ack mapping, just like it's predecessor:
> https://webrtc-review.googlesource.com/c/src/+/218000
>
> It also changes the logic to make sure retransmits counts towards
> history pruning - and properly ignores padding/fec.
>
> Bug: webrtc:12713
> Change-Id: I7835d10e483687e960a9cce41d4e2f1a6c3189b4
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221863
> Reviewed-by: Björn Terelius <terelius@webrtc.org>
> Reviewed-by: Philip Eliasson <philipel@webrtc.org>
> Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
> Commit-Queue: Erik Språng <sprang@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#34293}
TBR=danilchap@webrtc.org,terelius@webrtc.org,sprang@webrtc.org,philipel@webrtc.org,webrtc-scoped@luci-project-accounts.iam.gserviceaccount.com
Change-Id: Iaca6dc7739d953e97add5f5d516139b4819e43ee
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:12713
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/222601
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34294}
This CL makes sure we don't cull packets from the history based on
incorrect ack mapping, just like it's predecessor:
https://webrtc-review.googlesource.com/c/src/+/218000
It also changes the logic to make sure retransmits counts towards
history pruning - and properly ignores padding/fec.
Bug: webrtc:12713
Change-Id: I7835d10e483687e960a9cce41d4e2f1a6c3189b4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221863
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Reviewed-by: Philip Eliasson <philipel@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34293}
This is a consistent way to get to common config parameters for
all receive streams and avoids storing a copy of the extension
headers inside of Call. This is needed to get rid of the need of
keeping config and copies in sync, which currently is part of why
we repeatedly delete and recreate audio receive streams on config
changes.
Bug: webrtc:11993
Change-Id: Ia356b6cac1425c8c6766abd2e52fdeb73c4a4b4f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/222040
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34285}
Also including common Rtp config members.
Follow up changes will remove the ReceiveRtpConfig class in Call
and copy of extension headers, instead use the config directly
from the receive streams and not require stream recreation for changing
the headers.
Bug: webrtc:11993
Change-Id: I29ff3400d45d5bffddb3ad0a078403eb102afb65
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221983
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34283}
This is to be consistent with how things work on the video side but
also much less drastic than the current implementation. Aim is to
remove RecreateAudioReceiveStream(), which would improve efficiency
as well as allow for specific handling of the cases that currently
trigger recreation.
Bug: webrtc:11993
Change-Id: Ia81a5e66d44e41ea4eb2bff800e0b1583821c96a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221860
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Markus Handell <handellm@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34282}
This reverts commit 8a18e5b3c954a3f9cc006c90356a3d850bcc352f.
Reason for revert: Removing the problematic DCHECK.
Original change's description:
> Revert "Remove AudioReceiveStream::Reconfigure() method."
>
> This reverts commit e2561e17e29e62c02731f1d214d7ee5ffdaeb941.
>
> Reason for revert: Speculative revert: breaks an downstream project
>
> Original change's description:
> > Remove AudioReceiveStream::Reconfigure() method.
> >
> > Instead, adding specific setters that are needed at runtime:
> > * SetDepacketizerToDecoderFrameTransformer
> > * SetDecoderMap
> > * SetUseTransportCcAndNackHistory
> >
> > The whole config struct is big and much of the state it holds, needs to
> > be considered const. For that reason the Reconfigure() method is too
> > broad of an interface since it overwrites the whole config struct
> > and doesn't actually handle all the potential config changes that might
> > occur when the config changes.
> >
> > Bug: webrtc:11993
> > Change-Id: Ia5311978f56b2e136781467e44f0d18039f0bb2d
> > Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221363
> > Reviewed-by: Niels Moller <nisse@webrtc.org>
> > Commit-Queue: Tommi <tommi@webrtc.org>
> > Cr-Commit-Position: refs/heads/master@{#34252}
>
> TBR=saza@webrtc.org,nisse@webrtc.org,tommi@webrtc.org,webrtc-scoped@luci-project-accounts.iam.gserviceaccount.com
>
> Change-Id: I15ca2d8ee5fd612e13dc1f4b3bfb9c885c21dc66
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: webrtc:11993
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221746
> Commit-Queue: Artem Titov <titovartem@webrtc.org>
> Reviewed-by: Andrey Logvin <landrey@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#34253}
# Not skipping CQ checks because this is a reland.
Bug: webrtc:11993
Change-Id: I0d3bf9abdcdc8d3f9259d014e6074a5e6b6cc73c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221747
Reviewed-by: Tommi <tommi@webrtc.org>
Reviewed-by: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Andrey Logvin <landrey@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34255}
This reverts commit e2561e17e29e62c02731f1d214d7ee5ffdaeb941.
Reason for revert: Speculative revert: breaks an downstream project
Original change's description:
> Remove AudioReceiveStream::Reconfigure() method.
>
> Instead, adding specific setters that are needed at runtime:
> * SetDepacketizerToDecoderFrameTransformer
> * SetDecoderMap
> * SetUseTransportCcAndNackHistory
>
> The whole config struct is big and much of the state it holds, needs to
> be considered const. For that reason the Reconfigure() method is too
> broad of an interface since it overwrites the whole config struct
> and doesn't actually handle all the potential config changes that might
> occur when the config changes.
>
> Bug: webrtc:11993
> Change-Id: Ia5311978f56b2e136781467e44f0d18039f0bb2d
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221363
> Reviewed-by: Niels Moller <nisse@webrtc.org>
> Commit-Queue: Tommi <tommi@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#34252}
TBR=saza@webrtc.org,nisse@webrtc.org,tommi@webrtc.org,webrtc-scoped@luci-project-accounts.iam.gserviceaccount.com
Change-Id: I15ca2d8ee5fd612e13dc1f4b3bfb9c885c21dc66
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:11993
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221746
Commit-Queue: Artem Titov <titovartem@webrtc.org>
Reviewed-by: Andrey Logvin <landrey@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34253}
Instead, adding specific setters that are needed at runtime:
* SetDepacketizerToDecoderFrameTransformer
* SetDecoderMap
* SetUseTransportCcAndNackHistory
The whole config struct is big and much of the state it holds, needs to
be considered const. For that reason the Reconfigure() method is too
broad of an interface since it overwrites the whole config struct
and doesn't actually handle all the potential config changes that might
occur when the config changes.
Bug: webrtc:11993
Change-Id: Ia5311978f56b2e136781467e44f0d18039f0bb2d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221363
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34252}
* Make VideoSendStream and VideoSendStreamImpl construction non-blocking.
* Move ownership of the rtp video sender to VideoSendStream.
* Most state is constructed in initializer lists.
* More state is now const (including VideoSendStreamImpl ptr)
* Adding thread checks to classes that appear to have had a race before
E.g. RtpTransportControllerSend. The change in threading now actually
fixes an issue we weren't aware of.
* Moved from using weak_ptr to safety flag and made some PostTask calls
cancellable that could potentially have been problematic. Initalizing
the flag without thread synchronization is also simpler.
This should speed up renegotiation significantly when there are
multiple channels. A follow-up change will improve SetSend as well
which is another costly step during renegotiation.
Bug: webrtc:12840
Change-Id: If4b28da5a085643ce132c7cfcf80a62cd1a625c5
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/221105
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Markus Handell <handellm@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34224}
This variable is not used, always set to false but complicates
things for `keyframe_generation_requested_` as setting keyframe_needed
requires keyframe_generation_requested_ to be read synchronously from
what soon will be a different thread than where SetAndGetRecordingState
is called on.
Bug: webrtc:11993
Change-Id: I25675d9b70c9ec96a2542e7cf5480c835ea984eb
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/220840
Reviewed-by: Stefan Holmer <stefan@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34188}
Call send statistic updates are initiated on the send
transport sequence which forced calls to PostTask to the
worker thread which keeps several related attributes
protected by it.
Change this by:
* Using std::atomics for three attributes where synchronization
doesn't really matter and which can be accessed on either
context.
* Introducing a thread-compatible internal class which keeps
the statistics protected by the send transport sequence, and
emits UMA statistics on destruction.
The change also achieves the following trivial changes:
* The call origin time is now tracked by a proper
webrtc::Timestamp.
* The explicit use of the |send_transport_queue_| was replaced by
a more relaxed sequence checker.
Bug: webrtc:11993
Change-Id: I428a4d98b5fd2fd31222f62e597a9d61a3d4899f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/220931
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34187}
This way we can have custom implementation of RtpTransportControllerSendInterface and pass it properly to Call.
Call relies on RtpTransportControllerSendInterface already so this is natural way to customize RTP related classes.
If there is custom factory present in dependencies it will be used, otherwise default factory will be used.
Intention behind this change is to have ability to have custom QoS with custom parameters.
Bug: webrtc:12778
Change-Id: I5b88957025621ef4bcd63eaa98c218ad213da9c8
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/217769
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Philipp Hancke <phancke@nvidia.com>
Cr-Commit-Position: refs/heads/master@{#34181}
This is in preparation for actually doing this initialization
differently in the Call class. This CL takes the registration
steps that are inherently network thread associated and makes
them separate from the ctor/dtor.
Inject Call* instead of worker_thread(), which will simplify upcoming
work that needs to access the network_thread() as well.
This is related to:
https://webrtc-review.googlesource.com/c/src/+/220608https://webrtc-review.googlesource.com/c/src/+/220609
Bug: webrtc:11993
Change-Id: I72769fd61de84967d9a645750c40d01660a2716b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/220764
Reviewed-by: Markus Handell <handellm@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#34172}
