Commit graph

121 commits

Author SHA1 Message Date
Jan Schmidt
09d870a39c webrtc: Fixes for matching pads to unassociated transceivers
Fix an inverted condition when checking if sink pad caps match
the codec-preference of an unassociated transceiver, and
fix a condition check for transceiver media kind to
avoid matching sinkpad requests where caps aren't provided
against unassociated transceivers where the caps might
not match later.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7237>
2024-08-05 13:15:38 +00:00
Jan Schmidt
87a7a7567f webrtcbin: tracked maximum pad serial better
If a sink pad with a specific index is requested, also
increase the maximum pad serial number if necessary, so
that mixing fixed sink_X requests with unspecific sink_%u
requests works.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7237>
2024-08-05 13:15:38 +00:00
Carlos Bentzen
efa0a3ec6a webrtcbin: connect output stream on recv transceivers
With MR 7156, transceivers and transports are created earlier,
but for sendrecv media we could get `not-linked` errors due to
transportreceivebin not being connected to rtpbin yet when incoming
data arrives.

This condition wasn't being tested in elements_webrtcbin, but could be
reproduced in the webrtcbidirectional example. This commit now also
adds a test for this, so that this doesn't regress anymore.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7294>
2024-08-05 08:25:04 +00:00
Carlos Bentzen
cad3e63546 webrtcbin: reverse direction from remote media
This had been overlooked from the spec. We need to reverse
the remote media direction when setting the transceiver direction.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7294>
2024-08-05 08:25:04 +00:00
Carlos Bentzen
48ae40f477 webrtcbin: create and associate transceivers earlier in negotation
According to https://w3c.github.io/webrtc-pc/#set-the-session-description
(steps in 4.6.10.), we should be creating and associating transceivers when
setting session descriptions.

Before this commit, webrtcbin deviated from the spec:
1. Transceivers from sink pads where created when the sink pad was
   requested, but not associated after setting local description, only
   when signaling is STABLE.
2. Transceivers from remote offers were not created after applying the
   the remote description, only when the answer is created, and were then
   only associated once signaling is STABLE.

This commit makes webrtcbin follow the spec more closely with regards to
timing of transceivers creation and association.

A unit test is added, checking that the transceivers are created and
associated after every session description is set.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7156>
2024-08-01 07:38:46 +00:00
Philippe Normand
299a000917 webrtcbin: Allow session level setup attribute in SDP
An SDP answer can declare its setup attribute at the session level or at the
media level. Until this patch we were validating only the latter case and an
assert was raised in the former case.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6930>
2024-05-27 14:21:15 +00:00
Matthew Waters
c9af1d486a webrtc: request-aux-sender, only sink floating refs
Don't add an extra ref if non-floating as that ref will never be
unreffed.

gst_bin_add() is transfer floating (alias to transfer none).

Fixes a leak when a non-floating ref was provided as a return value in
the request-aux-sender signal.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6807>
2024-05-06 14:21:38 +00:00
François Laignel
7d5bb1ea7a webrtc: add all SSRC attributes getting CAPS for a PT
The transport stream only returned the CAPS for the first matching PT entry
from the `ptmap`. Other SSRC with the same PT where not included. For a stream
which bundled multiple audio streams for instance, only the first SSRC was
knowed to the SSRC demux and downstream elements.

This commit adds all the `ssrc-` attributes from the matching PT entries.

The RTP jitter buffer can now find the CNAME corresponding its SSRC even if it
was not the first to be registered for a particular PT.

The RTP PT demux removes `ssrc-*` attributes cooresponding to other SSRCs
before pushing SSRC specific CAPS to downstream elements.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6119>
2024-03-08 10:28:15 +00:00
François Laignel
aeaef7a7f0 webrtcbin: RFC5576 - early CNAME support
See RFC5576: have CNAME available to the rtpjitterbuffer before the the first
RTCP SR is received, for rapid synchronization. Similar to what was done for
RTSP (last 2 commits) of [MR 2132].

[RFC5576]: https://www.rfc-editor.org/rfc/rfc5576
[MR 2132]: https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2132

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6119>
2024-03-08 10:28:15 +00:00
Mathieu Duponchelle
91317aacaf webrtcbin, rtpbin: check before setting properties on jitterbuffer
In rtpbin we already systematically check for all property names
except latency, correct that.

In webrtcbin we need to check before trying to use the do-retransmission
property.

This is useful for the case where an element like identity gets passed
to rtpbin's request-jitterbuffer property, when the application wants
to use webrtcbin in an SFU situation, with no reordering and no added
latency

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6112>
2024-02-14 08:52:50 +00:00
sergey radionov
39f2afbd45 webrtcbin: it's better to have thread default main context
on thread bound to that main context.
fixes #3271

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6053>
2024-02-05 00:56:47 +00:00
Jan Schmidt
ef71c1319a webrtcbin: Improve SDP intersection for Opus
Remove optional sprop-stereo and sprop-maxcapture fields from Opus
remote offer caps before intersecting with local codec preferences.

According to https://datatracker.ietf.org/doc/html/rfc7587#section-7.1
those fields are sender-only informative, and don't affect
interoperability.

Fixes cases where the webrtc media will end up receive-only if the
local side wants to send stereo but the remote is sending mono, or
vice versa.

There may be other fields in other codecs, so the implementation
anticipates needing to add further fields and codecs in the future.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5993>
2024-01-25 13:37:21 +00:00
Olivier Crête
7104c867c0 webrtcsdp: Don't require fingerprint in inactive media
Inactive m-lines don't need a fingerprint as they may not
have a connection.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1118>
2023-11-13 20:51:31 +00:00
Olivier Crête
9eddf844f0 webrtcsdp: Remove comparison between media and session fingerprint
The code seems to validate that the media-level fingerprint matches
the fingerprint of the previous media or of the whole session. There
is no such requirement in any RFC I found. The session-session one
is just meant to act as a fallback when there is no media-level
fingerprint.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1118>
2023-11-13 20:51:31 +00:00
Johan Sternerup
5b64cfaca3 webrtcice: Add webrtc ALPN header for HTTP proxy
Section 3.4 in RFC8835 states that if a WebRTC endpoint uses an HTTP
proxy to access the Internet it MUST include the "ALPN" header. This
commit adds this header.

By default the ALPN used when connecting to the TURN/TCP server via a
proxy is set to "webrtc". It can be changed by adding an alpn url
option for the http-proxy. For example:

http://user:pass@my.http.proxy.com:8080?alpn=c-webrtc

This will add the header "ALPN: c-webrtc" to the HTTP proxy CONNECT
request.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4212>
2023-08-17 00:45:05 +00:00
Ryan Pavlik
e31407f9d2 webrtc: Fix docs for create-data-channel action signal
Initial line of the doc comment was incorrect, so the nicely written
docs were not being extracted.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5131>
2023-08-01 21:17:06 +00:00
Nirbheek Chauhan
d7d5d1ba93 webrtcbin: Fix support for glib older than 2.74
G_CONNECT_DEFAULT was added in 2.74, and passing `0` in older versions
gets the same behaviour.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5045>
2023-07-19 13:01:49 +00:00
Matthew Waters
6af8b3dd80 webrtcbin: don't hold the webrtc lock over on-new-transceiver emission
Could potentially produce a deadlock if the direction is changed in the
callback.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5045>
2023-07-19 13:01:49 +00:00
Matthew Waters
77e01571c8 webrtc: don't disallow transceiver direction changes
Initial testing seems to suggest that we support them reasonably well
(at least for BUNDLEd streams).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5045>
2023-07-19 13:01:49 +00:00
Matthew Waters
13f4066580 webrtc: add check for negotiation on transceiver direction changes
As required by the webrtc specification.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5045>
2023-07-19 13:01:49 +00:00
Philippe Normand
424a78c9b9 webrtcbin: Prevent critical warning when creating an additional data channel
The max_channels value wasn't clamped to 65534 in all situations.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5001>
2023-07-10 14:08:09 +00:00
Philippe Normand
d317379287 webrtcstats: Properly report IceCandidate type
strcmp returns a positive value if s1 is greater than s2, while we actually
needed to check equality here.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4952>
2023-07-03 03:51:53 +00:00
Sangchul Lee
2661bf6d9a webrtc: Add data-channels-opened/closed to get-stats signal documentation
With contributions from: Matthew Waters <matthew@centricular.com>

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2127>
2023-05-18 12:08:55 +00:00
Martin Nordholts
85e3f31740 webrtc: Track stats for data channels opened and closed
Track data channel stats for `dataChannelsOpened` and
`dataChannelsClosed` in `RTCPeerConnectionStats` as specified by
https://www.w3.org/TR/webrtc-stats/#dictionary-rtcpeerconnectionstats-members

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4638>
2023-05-18 04:31:16 +00:00
Matthew Waters
b10ec569d7 webrtc: advertise end-of-candidate with an empty candidate string
Just like what is done in the browsers.  When this is sent to the peer,
they will be able to know that no more candidates are coming and can
complete ICE.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4598>
2023-05-12 04:52:22 +00:00
Philippe Normand
b75114983e webrtcdatachannel: Bind to parent webrtcbin using a weak reference
The previous approach of using a simple pointer could lead to a use-after-free
in case a data-channel was created and its parent webrtcbin was disposed soon
after.

Fixes #2103

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4160>
2023-05-08 19:20:22 +00:00
Philippe Normand
4db12345d1 webrtcbin: Fix potential deadlock when closing before any data was sent
A blocking pad probe is added on new sink pads, it's usually removed after the
caps have been negotiated or the signaling state switched to stable, but if that
never happens and the pad is released we kept the pad probe active, leaving the
pad blocked, preventing clean disposal.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4529>
2023-05-03 02:29:31 +00:00
Daniel Moberg
4c1de25e9d webrtc: do not tear down data channel before data is flushed
Current implementation can in some cases detect
that all data is sent but in reality it is not,
leading to a push to an unlinked pad.
This is a race between the probe used to track data sent and a
call to close.

This patch sends an EOS before starting the close procedure
and then waits for the EOS event to come through to the
src pad before commencing with tear down.
This ensures that any queued data before EOS is flushed.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4462>
2023-04-27 00:40:18 +00:00
Eva Pace
692d4a3a16 webrtcbin: Fix trace log 'from' value
`webrtc->signaling_state` (from) and `new_signaling_state` (to) had the
same value when printed in a trace log. This commit adds a
`old_signaling_state` variable to hold the previous value, so that the
print statement works as intented.

Spotted by: Mustafa Asım REYHAN
Fixes #1802

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4362>
2023-04-25 04:27:02 +00:00
Albert Sjölund
65bd020754 webrtc: patch leak caused by early return
In webrtc_data_channel_send functions, both data and string,
an early return on a non-open datachannel caused it to leak
the buffer used for pushing to appsrc, meaning any buffer
sent after leaving the open state was leaked in full.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4191>
2023-03-16 15:17:58 +00:00
Tim-Philipp Müller
99ee8af782 webrtc: fix g-i annotations for allow-none
'allow none' doesn't exist, and 'allow-none' is now deprecated.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4146>
2023-03-10 13:09:25 +00:00
Philippe Normand
906b90287c webrtcbin: Relay add-ice-candidate errors from Ice implementation to Application
The `add_candidate` vfunc of the GstWebRTCICE interface gained a GstPromise
argument, which is an ABI break. We're not aware of any external user of this
interface yet so we think it's OK.

This change is useful in cases where the application needs to bubble up errors
from the underlying ICE agent, for instance when the agent was given an invalid
ICE candidate.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3960>
2023-02-27 09:09:47 +00:00
Philippe Normand
cf96d96f6a webrtcbin: Add add-ice-candidate-full signal
The signal triggers an asynchronous task on the PC thread but in some cases it
can be useful for apps to be notified when the task completed. This method of
the PeerConnection spec also returns a Promise so the interface is now more
coherent with the spec.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3960>
2023-02-27 09:09:47 +00:00
Jan Schmidt
621604aa3e webrtc: Calculate the jitter for remote-inbound-rtp stats
Populate the clock-rate in the internal stats structure, so
it can be used by the _get_stats_from_remote_rtp_source_stats()
method to calculate remote receivers' jitter.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3900>
2023-02-07 04:58:04 +11:00
Jan Schmidt
615a019457 webrtcbin: Report full codec-stats for source pads
Use the current caps for webrtcbin srcpads, as received_caps
are only stored for sink pads based on incoming caps events.

Makes it so that webrtcbin stats reports contain fuller
codec information.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3900>
2023-02-07 04:49:34 +11:00
Philippe Normand
72884f141c webrtcbin: Support for setting kind attribute on RTCRtpStreamStats
The attribute maps the `kind` property of the associated transceiver.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3630>
2022-12-22 21:35:51 +00:00
Matthew Waters
993bc8fc01 webrtc: implement support for msid values
Local msid values are taken from sink pad property, or fallback to the
previously used cname.

The remote msid values are exposed on the relevant src pads.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3106>
2022-12-14 12:23:32 +11:00
Jan Schmidt
dfb5e3365e webrtcbin: Remove queue after rtpfunnel
The original BUNDLE support commit placed a queue after the
rtpfunnel that combines streams, but I don't see a good reason for
it. It has default settings, so if network output is slow might
accidentally store up to 1 second of pending data, increasing
latency.

Remove it in favour of doing any necessary buffering before
webrtcbin. If it turns out that there is a reason for it to
exist, the limits should probably be configurable and small.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3437>
2022-11-19 10:31:50 +00:00
Jan Schmidt
5fa4f0562c webrtcbin: Fix a typo in debug log
transceiever -> transceiver

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3434>
2022-11-19 13:12:58 +11:00
Johan Sternerup
e708543039 webrtcbin: Add settings for HTTP proxy
Pass this to libnice which has a simple HTTP 1.0 proxy with basic
authentication only.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2867>
2022-11-18 15:00:58 +00:00
Matthew Waters
5ca3988420 webrtc/datachannel: handle error messages from appsrc/sink
Fixes a possible race where closing a data channel may produce e.g.
not-linked errors.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3381>
2022-11-11 10:13:27 +00:00
Edward Hervey
a100f36b69 webrtcbin: Don't duplicate enum string values
Some were leaked when debugging was enabled. Instead just directly use the
static strings as-is.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3347>
2022-11-07 11:21:00 +00:00
Matthew Waters
0077d13304 webrtcbin: configure rtpulpfecdec passthrough property
This allows downstream (payloaders mostly) to be able to correctly
detect actual packet loss from rtp sequence numbers.

See
https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/issues/581
for background.

Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1407

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3212>
2022-10-23 23:44:07 +00:00
Matthew Waters
a633f5d287 webrtcbin: also add rtcp-fb ccm fir for video mlines by default
In addition to the 'nack pli' already added.  Both are supported by
rtpbin/rtpsession by default already.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3235>
2022-10-21 01:02:34 +00:00
Sangchul Lee
0f05be382b webrtcbin: Improve documentation of 'turn-server' property
Description about how to set time-limited credentials is added.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3229>
2022-10-20 15:30:07 +00:00
Sangchul Lee
93b896eb4e webrtcbin: Fix pointer dereference before null check
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3129>
2022-10-06 16:46:33 +00:00
Johan Sternerup
212c09a70e webrtc: return error when sending on non-open datachannel
According to W3C
specification (https://w3c.github.io/webrtc-pc/#datachannel-send) we
should return InvalidStateError exception when trying to send when the
channel is not open. In the world of C/glib/gstreamer we don't have
exceptions but have to rely on gboolean/GError instead. Introducing
these calls for a change in function signature of the action signals
used to send data on the datachannel. Changing the signature of the
existing "send-string" and "send-data" signals would mean an immediate
breaking change so instead we deprecate them. Furthermore, there is no
way to express GError** as an argument to an action signal in a way
that fits language bindings (pointer-to-pointer simply does not work)
and we have to use regular functions instead.

Therefore we introduce gst_webrtc_data_channel_send_data_full() and
gst_webrtc_data_channel_send_string_full() while deprecating the old
functions and corresponding signals.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1958>
2022-10-05 11:08:30 +00:00
Mathieu Duponchelle
b454ec972f webrtcbin: fix picking available payload types
When picking an available payload type, we need to pick one that is
available across all media.

The previous code, when multiple media were present, looked at the first one,
noticed it had pt 96 as the media pt, then simply looked at the next media,
noticed it didn't, and decided 96 was available.

Instead, check if the pt is used by any of the media, if it is, decide
it is not available and go to the next pt. I'm fairly sure that was the
original intent.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2984>
2022-09-07 03:22:34 +00:00
Olivier Crête
4b3b234f72 webrtcbin: Allow locked mlines with no caps, as the last ones
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2439>
2022-09-02 11:52:58 +02:00
Olivier Crête
0930c467d4 webrtcbin: Reject creating an offer if a locked mline has no caps
This avoids getting in a bunch of corner cases. We'd have to insert
a "rejected" line from the start as a place-holder to get around this,
but the rest of the code just becomes more complicated, so just
disallow it for now.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2439>
2022-09-02 11:52:58 +02:00