From the test result, using AVX2/FMA is 1.60x faster than SSE on atlas.
Bug: webrtc:11663
Test: common_audio_unittests on atlas and octopus.
Change-Id: Ibd45ea46aa97d5790a24e5116f741592b95f6416
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/176382
Reviewed-by: Per Åhgren <peah@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31810}
This CL changes the style of logging for an API which is essential when
WebRTC is used in Chrome. By changing the format, we can more easily
tie in (search for tags etc.) logs from WebRTC with logs in Chrome.
See e.g.
https://chromium-review.googlesource.com/c/chromium/src/+/2093443
for more details.
I decided to use a new private method to avoid using rtc::StringBuilder.
The idea was to make the log statements less complex and more condensed.
Tbr: mbonadei
Bug: webrtc:11493
Change-Id: I46b4a933ad62ac1db376743b4a41b62c5f8c6ac6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/172841
Commit-Queue: Henrik Andreassson <henrika@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31808}
Port allocator can be created on one thread and then initialized and
used on another. So we can avoid sync invoke to network thread to create
port allocator.
Bug: webrtc:11799
Change-Id: I5020093a41acbf7e372f2e4970e016ce14a7f406
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/180122
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31805}
This is a reland of d4089cae47334a4228b69d6bb23f2e49ebb7496e
with the following fix:
Invoke MaybeStartGathering as the last step of DoSetLocalDescription.
This ensures that candidates and onicegatheringstatechange does not
happen before SLD is resolved. This is important for passing
external/wpt/webrtc/RTCPeerConnection-iceGatheringState.html.
Original change's description:
> [Perfect Negotiation] Implement non-racy version of SetLocalDescription.
>
> BACKGROUND
>
> When SLD is invoked with SetSessionDescriptionObserver, the observer is
> called by posting a message back to the execution thread, delaying the
> call. This delay is "artificial" - it's not necessary; the operation is
> already complete. It's a post from the signaling thread to the signaling
> thread. The rationale for the post was to avoid the observer making
> recursive calls back into the PeerConnection. The problem with this is
> that by the time the observer is called, the PeerConnection could
> already have executed other operations and modified its states.
>
> This causes the referenced bug: one can have a race where SLD is
> resolved "too late" (after a pending SRD is executed) and the signaling
> state observed when SLD resolves doesn't make sense.
>
> When implementing Unified Plan, we fixed similar issues for SRD by
> adding a version that takes SetRemoteDescriptionObserverInterface as
> argument instead of SetSessionDescriptionObserver. The new version did
> not have the delay. The old version had to be kept around not to break
> downstream projects that had dependencies both on he delay and on
> allowing the PC to be destroyed midst-operation without informing its
> observers.
>
> THIS CL
>
> This does the old SRD fix for SLD as well: A new observer interface is
> added, SetLocalDescriptionObserverInterface, and
> PeerConnection::SetLocalDescription() is overloaded. If you call it with
> the old observer, you get the delay, but if you call it with the new
> observer, you don't get a delay.
>
> - SetLocalDescriptionObserverInterface is added.
> - SetLocalDescription is overloaded.
> - The adapter for SetSessionDescriptionObserver that causes the delay
> previously only used for SRD is updated to handle both SLD and SRD.
> - FakeSetLocalDescriptionObserver is added and
> MockSetRemoteDescriptionObserver is renamed "Fake...".
>
> Bug: chromium:1071733
> Change-Id: I920368e648bede481058ac22f5b8794752a220b3
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/179100
> Commit-Queue: Henrik Boström <hbos@webrtc.org>
> Reviewed-by: Harald Alvestrand <hta@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#31798}
TBR=hta@webrtc.org
Bug: chromium:1071733
Change-Id: Ic6e8d96afa1c19604762f373716c08dbfa9d178c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/180481
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31804}
This log line is causing test failures due to excessive logging (see
referenced bug); reducing log level.
Bug: chromium:984879
Change-Id: Ic94ba0a39b91b4253a58ad54de0cba1ca49882e2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/175913
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Marina Ciocea <marinaciocea@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31803}
This reverts commit d4089cae47334a4228b69d6bb23f2e49ebb7496e.
Reason for revert: Breaks chromium WPT that is timing sensitive to onicegatheringstatechanges.
This CL accidentally moved the MaybeStartGatheringIceCandidates to after completing the SLD call. The fix is to move it back. I'll do that in a re-land.
Original change's description:
> [Perfect Negotiation] Implement non-racy version of SetLocalDescription.
>
> BACKGROUND
>
> When SLD is invoked with SetSessionDescriptionObserver, the observer is
> called by posting a message back to the execution thread, delaying the
> call. This delay is "artificial" - it's not necessary; the operation is
> already complete. It's a post from the signaling thread to the signaling
> thread. The rationale for the post was to avoid the observer making
> recursive calls back into the PeerConnection. The problem with this is
> that by the time the observer is called, the PeerConnection could
> already have executed other operations and modified its states.
>
> This causes the referenced bug: one can have a race where SLD is
> resolved "too late" (after a pending SRD is executed) and the signaling
> state observed when SLD resolves doesn't make sense.
>
> When implementing Unified Plan, we fixed similar issues for SRD by
> adding a version that takes SetRemoteDescriptionObserverInterface as
> argument instead of SetSessionDescriptionObserver. The new version did
> not have the delay. The old version had to be kept around not to break
> downstream projects that had dependencies both on he delay and on
> allowing the PC to be destroyed midst-operation without informing its
> observers.
>
> THIS CL
>
> This does the old SRD fix for SLD as well: A new observer interface is
> added, SetLocalDescriptionObserverInterface, and
> PeerConnection::SetLocalDescription() is overloaded. If you call it with
> the old observer, you get the delay, but if you call it with the new
> observer, you don't get a delay.
>
> - SetLocalDescriptionObserverInterface is added.
> - SetLocalDescription is overloaded.
> - The adapter for SetSessionDescriptionObserver that causes the delay
> previously only used for SRD is updated to handle both SLD and SRD.
> - FakeSetLocalDescriptionObserver is added and
> MockSetRemoteDescriptionObserver is renamed "Fake...".
>
> Bug: chromium:1071733
> Change-Id: I920368e648bede481058ac22f5b8794752a220b3
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/179100
> Commit-Queue: Henrik Boström <hbos@webrtc.org>
> Reviewed-by: Harald Alvestrand <hta@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#31798}
TBR=hbos@webrtc.org,hta@webrtc.org
Change-Id: Ie1e1ecc49f3b1d7a7e230db6d36decbc4cbe8c86
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:1071733
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/180480
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31802}
BACKGROUND
When SLD is invoked with SetSessionDescriptionObserver, the observer is
called by posting a message back to the execution thread, delaying the
call. This delay is "artificial" - it's not necessary; the operation is
already complete. It's a post from the signaling thread to the signaling
thread. The rationale for the post was to avoid the observer making
recursive calls back into the PeerConnection. The problem with this is
that by the time the observer is called, the PeerConnection could
already have executed other operations and modified its states.
This causes the referenced bug: one can have a race where SLD is
resolved "too late" (after a pending SRD is executed) and the signaling
state observed when SLD resolves doesn't make sense.
When implementing Unified Plan, we fixed similar issues for SRD by
adding a version that takes SetRemoteDescriptionObserverInterface as
argument instead of SetSessionDescriptionObserver. The new version did
not have the delay. The old version had to be kept around not to break
downstream projects that had dependencies both on he delay and on
allowing the PC to be destroyed midst-operation without informing its
observers.
THIS CL
This does the old SRD fix for SLD as well: A new observer interface is
added, SetLocalDescriptionObserverInterface, and
PeerConnection::SetLocalDescription() is overloaded. If you call it with
the old observer, you get the delay, but if you call it with the new
observer, you don't get a delay.
- SetLocalDescriptionObserverInterface is added.
- SetLocalDescription is overloaded.
- The adapter for SetSessionDescriptionObserver that causes the delay
previously only used for SRD is updated to handle both SLD and SRD.
- FakeSetLocalDescriptionObserver is added and
MockSetRemoteDescriptionObserver is renamed "Fake...".
Bug: chromium:1071733
Change-Id: I920368e648bede481058ac22f5b8794752a220b3
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/179100
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31798}
Internal counters in the RenderDelayBuffer can slip out of sync with external counters, leading to buffer misalignment.
This CL gives the RenderDelayBuffer an opportunity to update its counters.
Tested:
Passes: modules_unittests --gtest_filter=BlockProcessor.*
Fails as expected due to new unit test: modules_unittests --gtest_filter=BlockProcessor.* --force_fieldtrials="WebRTC-Aec3RenderBufferCallCounterUpdateKillSwitch/Enabled/"
audioproc_f with default AEC settings has been verified to be bit-exact on a large number of aecdumps.
Bug: webrtc:11803
Change-Id: I9363b834c8c8c934add0335013df60bf131da4bc
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/180126
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31795}
RTPFragmentationHeader is already ignored by H264 packetizer
and thus doesn't need to be provided and calculated.
Bug: webrtc:6471
Change-Id: I45bc22827f0dc811457e3ebe477a16293501c2fb
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/179843
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Reviewed-by: Marco Paniconi <marpan@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31791}
The time precision of delayed tasks is one millisecond, so the
TaskQueuePacedSender makes sure that is the minimum sleep time, and
then allows sending prior data as if it was on time.
Furthermore, if there already exists a pending task within 1ms of a
new desired process time - we don't schedule a new one with the same
motivation as above.
These two facts clashes somewhat with how BitrateProber works, and
especially if they coincide it can result in scheduled ProcessPackets()
that is 2ms late. The default timeout set in BitrateProber is 3ms, so
there is a higher risk of probes timing out.
This CL changes the TaskQueuePacedSender to allow scheduling a
ProcesPackets() call as soon as possible if we are probing - even if
that means executing up to 1ms earlier than expected (the BitrateProber
will compensate for that). The PacingController is updated in order to
allow early execution in this one case.
Bug: webrtc:10809
Change-Id: Ia5097ddc39aa80c05ebfe56369310c94ef0e0baf
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178901
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31778}
The app showcased the ability to send real-time voice data between two endpoints using the VoIP API.
Users can also configure session parameters such as the endpoint information and codec used.
Bug: webrtc:11723
Change-Id: I682f4aa743b707759536bce59e598789a77b7ec6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/178467
Reviewed-by: Kári Helgason <kthelgason@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Tim Na <natim@webrtc.org>
Commit-Queue: Tim Na <natim@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31775}
local_addresses is a list of IPs that were used to send data, which was
used during stats calculation.
Bug: webrtc:11756
Change-Id: Ie6307eaa69c73ebe9f69e44503752151be9e9ef6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/179841
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Reviewed-by: Andrey Logvin <landrey@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31774}
A packet's capture time may be -1 to indicate an unset value. We need to
check that this is the case before adjusting it when generating padding.
Otherwise, invalid values will result.
Bug: webrtc:11615
Change-Id: Ibbeb959f1d4d37dd4d65702494b97246642b57d6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/176281
Commit-Queue: Dan Minor <dminor@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#31766}
