This CL separates the two directions of MediaChannel into two separate objects that do not couple with each other.
The notable API change is that receiver local SSRC now has to be set explicitly - before, it was done implicitly when the send-side MediaChannel had a stream added to it.
Bug: webrtc:13931
Change-Id: I83c2e3c8e79f89872d5adda1bc2899f7049748b3
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/288400
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#39340}
This is in preparation for splitting MediaChannel into sender and
receiver channels, with independent objects.
Bug: webrtc:13931
Change-Id: I8e34b0c80b4d76132394efcda658a8face3ab873
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/288750
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#38998}
This detaches the implementation (which is still merged)
from the objects used to interface to it.
Bug: webrtc:13931
Change-Id: I872ee10e4ed9fa432bfa231f723af1d3989d79d4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/288080
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#38906}
The implementation here has a number of changes that force the callers
that called the "channel" functions into specific interfaces rather than
just letting C++ take care of it; this should go away once there stops
being a common implementation class for those interfaces.
Bug: webrtc:13931
Change-Id: Ic4e279528a341bc0a0e88d2e1e76c90bc43a1035
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/287640
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#38888}
This allows callers to differentiate on whether they need the
channel for sending or receiving purposes.
Note: This CL is incomplete, in that many places cast the pointers
to the concrete subclasses "VideoMediaChannel" and "AudioMediaChannel", which are not split into sending and receiving APIs.
The long term goal is to make two MediaChannel-like class APIs, with distinct implementations, and let the RtpSender and RtpReceiver manage those objects, rather than keeping them in the RtpTransceiver.
Bug: webrtc:13931
Change-Id: I8d56defe2287bd6552b71571cc6a5ec842927fa4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/287040
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#38844}
This also hides the existence of the classes VideoChannel and
VoiceChannel from anything that does not include "channel.h".
Bug: webrtc:13931
Change-Id: I080a692b6acfd5d2d0401ec20d59c3a684eddb05
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/260944
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36746}
This makes the channel manager object into a factory, not a manager.
Bug: webrtc:13931
Change-Id: I59f7d818a739797a7c0a7a32e6583450834df122
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/260467
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#36718}
These class comments can help people to keep the relationship between
various classes / levels of abstraction straight.
Bug: None
Change-Id: I9fc899ac9c1080a406b06e88a38883cc3f00093d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/248169
Reviewed-by: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Tomas Gunnarsson <tommi@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35769}
The interface is implemented by the ChannelManager and contains methods
to create and destroy media channel objects as used by a transceiver.
This will subsequently allow us to delete the channel objects from
the transceiver class where ownership really lies rather than from
the outside - which is currently required by some tests that keep
channel objects on the stack. We'll furthermore be able to do the
destruction asynchronously without additional Invoke()s as we do now
which will remove an Invoke when making sdp changes.
With introducing the interface, the following simplifications were made:
* ChannelManager constructed on the signaling thread.
Before, there was an Invoke in the context class, which existed
for the purposes of calling MediaEngine::Init() (which in turn is
only needed for the VoiceEngine). This Invoke has now been moved
into the CM (more tbd).
* The CM now has a pointer to the signaling thread (since that's the
construction thread). That allows us to remove the signaling thread
parameter from the CreateFooChannel methods.
* The ssrc_generator (UniqueRandomIdGenerator) instance for SSRCs moved
from SdpOfferAnswerHandler to the CM, as it's always used in
combination with the CM. This simplifies the CreateFooChannel methods
as well as a couple of other classes that have a CM dependency.
* Removed DestroyFooChannel related code from SdpOfferAnswerHandler since
the channel type detail can be taken care of by the CM.
Bug: webrtc:11992, webrtc:13540
Change-Id: I04938a803734de8489ba31e6212d9eaecc244126
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/247904
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35766}
This is a slight refactoring while doing some other changes, so not
strictly necessary, but the error param is always supplied in practice
so it made sense to update the tests to reflect that, test that error
values are reported in (at least) some cases and remove the additional
code that checks for whether or not error information is requested.
Bug: none
Change-Id: Ia5739a18ea2beb6970eabf9d809c24dfa43466b1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/244097
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35632}
Because of this (seemingly simple) change, I had to change the return
type of transport_name from `const std::string&` to `absl::string_view`
to handle the case when there's no transport assigned.
That in turn caused an avalanche of required updates.
Bug: webrtc:12230, webrtc:11993
Change-Id: I16ec6c6a5fc2f5f7c7de572355a3c6ca924bb9d4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/244084
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Tomas Gunnarsson <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35617}
The channel classes have stored the negotiated headers but a more
natural place to store them is in the RtpTransceiver class where
RtpHeaderExtension state is managed as well as the implementation of
the only method that depends on the stored state,
HeaderExtensionsNegotiated().
Also adding a TODO for further improvements where we're unnecessarily
storing state in the channel classes for the purposes of the transports.
Bug: webrtc:12726
Change-Id: If36668e3e49782ddeada23ebed126ee2c4935b8c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/216691
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33917}
This changes the notification to a single std::function pointer
instead of being a sigslot::signal1<> collection.
Summary:
* Remove SignalFirstPacketReceived_, the last sigslot member variable.
(still inherits from sigslot::has_slots<>)
* BaseChannel doesn't post to the signaling thread anymore. The only
reason that remains for the signaling_thread_ variable, is for
thread checking.
* Remove BaseChannel's reliance on MessageHandlerAutoCleanup
(still inherits from MessageHandler)
RtpTransceiver is the consumer of this event. That class is also the
class that sits between the PC classes and the channel object, holding
a pointer to the channel and managing calls that come in on the
signaling thread, such as SetChannel. The responsibility of delivering
the first packet received on the signaling thread is now with
RtpTransceiver:
* RtpTransceiver always requires a ChannelManager instance. Previously
this variable was sometimes set, but it's now required.
* Updated tests in rtp_transceiver_unittest.cc to include a
ChannelManager as well as fix them to include call expectations for
mock sender and receivers.
Bug: webrtc:11993, webrtc:11988
Change-Id: If49d6be157cd7599fa6fe3a42cd0a363464e3a74
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/215979
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33853}
* Changing return value from bool to void since the operation as async
effects anyway.
* Removing the `enabled()` accessor due to potential threading issues
and potential TOCTOU issues. It was only used in one place anyway.
* Applying thread restrictions to member variables.
Bug: none
Change-Id: I51949f5594339952d7b717cfd82f99b532e86b23
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/216182
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33830}
This reverts commit c1ad1ff178.
Reason for revert: This blocks the worker thread for a longer
contiguous period of time which can lead to delays in processing
packets. And due to other recent changes, the need to speed up
SetLocalDescription/SetRemoteDescription is reduced.
Still plan to reland some of the changes from the CL, just not the
part that groups the Invokes.
Original change's description:
> Do all BaseChannel operations within a single Thread::Invoke.
>
> Instead of doing a separate Invoke for each channel, this CL first
> gathers a list of operations to be performed on the signaling thread,
> then does a single Invoke on the worker thread (and nested Invoke
> on the network thread) to update all channels at once.
>
> This includes the methods:
> * Enable
> * SetLocalContent/SetRemoteContent
> * RegisterRtpDemuxerSink
> * UpdateRtpHeaderExtensionMap
>
> Also, removed the need for a network thread Invoke in
> IsReadyToSendMedia_w by moving ownership of was_ever_writable_ to the
> worker thread.
>
> Bug: webrtc:12266
> Change-Id: I31e61fe0758aeb053b09db84f234deb58dfb3d05
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/194181
> Commit-Queue: Taylor <deadbeef@webrtc.org>
> Reviewed-by: Harald Alvestrand <hta@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#32817}
TBR=deadbeef@webrtc.org,hta@webrtc.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: webrtc:12266
Change-Id: I40ec519a614dc740133219f775b5638a488529b1
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/203860
Reviewed-by: Taylor <deadbeef@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Taylor <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#33111}
Instead of doing a separate Invoke for each channel, this CL first
gathers a list of operations to be performed on the signaling thread,
then does a single Invoke on the worker thread (and nested Invoke
on the network thread) to update all channels at once.
This includes the methods:
* Enable
* SetLocalContent/SetRemoteContent
* RegisterRtpDemuxerSink
* UpdateRtpHeaderExtensionMap
Also, removed the need for a network thread Invoke in
IsReadyToSendMedia_w by moving ownership of was_ever_writable_ to the
worker thread.
Bug: webrtc:12266
Change-Id: I31e61fe0758aeb053b09db84f234deb58dfb3d05
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/194181
Commit-Queue: Taylor <deadbeef@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#32817}
This fixes regressions caused by:
https://webrtc-review.googlesource.com/c/src/+/183120
... which disabled payload type demuxing when multiple video tracks are
present, to avoid one channel creating a default track intended for
another channel.
However, this isn't an issue when not bundling, as each track will be
delivered on separate transport.
And it's also not an issue when each track uses a distinct set of
payload types (e.g., VP8 is mapped to PT 96 in one m= section, and PT 97
in another).
This CL addresses both of those cases; PT demuxing is only disabled
when two bundled m= sections have overlapping payload types.
Bug: chromium:1139052, webrtc:12029
Change-Id: Ied844bffac2a5fac29147c11b56a5f83a95ecb36
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/187560
Commit-Queue: Taylor <deadbeef@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#32419}
This fixes some edge cases where early media could cause default
stream that block the actual signaled media from beind delivered.
Bug: webrtc:11477
Change-Id: I8b26df63a690861bd19f083102d1395e882f8733
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/183120
Commit-Queue: Taylor <deadbeef@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#32030}
CreateOffer and CreateAnswer will now examine the layers on the
transceiver to determine if multiple layers are requested (Simulcast).
In this scenario RIDs will be used in the layers (instead of SSRCs).
When the offer is created, only RIDs are signalled in the offer.
When the offer is set locally SetLocalDescription() SSRCs will be
generated for each layer by the Channel and sent downstream to the
MediaChannel.
The MediaChannel receives configuration that looks identical to that of
legacy simulcast, and should be able to integrate the streams correctly
regardless of how they were signalled.
Setting multiple layers on the transciever is still not supported
through the API.
Bug: webrtc:10075
Change-Id: Id4ad3637b87b68ef6ca7eec69166fee2d9dfa36f
Reviewed-on: https://webrtc-review.googlesource.com/c/119780
Reviewed-by: Seth Hampson <shampson@webrtc.org>
Reviewed-by: Steve Anton <steveanton@webrtc.org>
Commit-Queue: Amit Hilbuch <amithi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26428}
