Change-Id: I42375710410f3ceca06bd546a66bc3258c60d5fb

Pierre Lespagnol authored 4 years ago and Adrien Béraud committed 4 years ago

The ice transport destruction can take a while, move
this in the io pool to avoid blocking the main thread.
This patch fix the incorrect orientation for android
devices after conferences.

Change-Id: I7ce467628aa265c2e8afc140d3bb4980beebd2ae

Mohamed Chibani authored 4 years ago and Sébastien Blin committed 4 years ago

The external address of the IGD might not be always routable if
the router is not directly connected to the internet.
So we need to always add the published address returned by the DHT even if UPNP
is enabled and actif. Failing to do so, and if the peer use only relay candidates,
the TURN will reject all incoming connections from our public address because
it were not in the candidates list and thus not authorized by TURN.

The automatic reallocation of failing mappings was also improved.

Change-Id: I5fb1424d5dc61b990595c229be647099b017eba1
Gitlab: #335

Sébastien Blin authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: I9d11246efa4c86534d2e8940510df3165e1119f6
GitLab: #426

First, on_connect_complete dropped the connection errors but it
must be handled for restarting failed check if necessary.
Then, a problematic case was detected when the controlling agent
got only a few active checks to do, but a passive will succeed.
In the previous version the controlling was failing as soon as
all active checks were checked. Now, if all active are failing
the controlling agent will let some time for passive agents and
then will decide to set the negotiation as failed.

Change-Id: Ica728f55083a34496b031c859ed73b576e1eef26
GitLab: #411

Mohamed Chibani authored 4 years ago and Sébastien Blin committed 4 years ago

Calls to DeletePortMapping method are made synchronous which may cause
lags. This might be particulary critical if the IGD does not respond
fast enough since the calls are made on the shared main thread.

Change-Id: I57366a70ffeb294e6277c4e8d946b339a8277309
Gitlab: #420

answering and joining a participant can take quite some time,
avoid to lock main thread for that.

Change-Id: I4a2982ba8c35cdc7861ff5db01cffd0678777bea

Change-Id: I5d6f9d8fe3b44a5bbe2045cc3d48f11a81a2512c

Pierre Lespagnol authored 4 years ago and Adrien Béraud committed 4 years ago

Reduce the bandwidth and CPU/GPU usage

Change-Id: Icfc19b7d0232e01bf4d78a99649daeee3295fdbd

In the previous implementation, a failed mapping was not removed if it's
beeing used. The responsability was left to the owner to release it when
done with it. This requires that all the owners implement the mapping
callback and properly release the mapping. This might be somehow cumbersome
and prune to error.
Now, a failed mapping is automatically removed, and no further action is
required from the owner. The mapping's owner will still receive a notification
callback (if registered) when the state of the mapping changes to failed (or
to any other state).

Change-Id: I014ba3906b836efce3173fb22fbe298a4023db7e
Gitlab: #335

Sébastien Blin authored 4 years ago and Adrien Béraud committed 4 years ago

https://review.jami.net/c/ring-daemon/+/16369 introduced
setUpTransmissionData() which is unnecessary for startCall() and
sendTextMessage as "to" is for contact, not current account.

Change-Id: I23373f5419f50b45a649d840c073186715fe053d
GitLab: #406

GitLab: jami-plugins#10
Change-Id: Ib18829baec30ec949512c2ce215470771ca0ed70

Aline Gondim Santos authored 4 years ago and Aline Gondim Santos committed 4 years ago

- add internal preference for always turn on plugin

GitLab: jami-plugins#4
GitLab: #394
Change-Id: Ic37fb486531cf436df4346246ec74f6480e4c0fd

Change-Id: I8e637ec1e5d00a62d2d19cc7bb5cc39c8da4be3a

Adrien Béraud authored 4 years ago and Pierre Lespagnol committed 4 years ago

Change-Id: I5233425464054d1690a1556336e16d5b6aa04162

Adrien Béraud authored 4 years ago and Sébastien Blin committed 4 years ago

Change-Id: Id382d9c31b62c233af3561f0f33132d1a346e0d8

Pierre Lespagnol authored 4 years ago and Adrien Béraud committed 4 years ago

Before the frame transfer to the CPU could fail during the rendering,
the fix move the transfer part in the update.
This way, the source keep the last rendered frame if the transfer fail

Change-Id: I71848629023720ed09712a8191e556b59c514e1c
GitLab: #388

Pierre Lespagnol authored 4 years ago and Adrien Béraud committed 4 years ago

This fix audio bug with SILK/LBRR frames

Change-Id: Iee93a4727eba096068a19804b33a03be0aab56c7

Change-Id: If093cd75fc19e630209c518a747b1a50d389e2c8

Sébastien Blin authored 4 years ago and Adrien Béraud committed 4 years ago

This patch embeds multiple litte fixes detected while working on swarms:
+ When stopping an account, shutdownConnections() should be done
before dht->join() to avoid any lock (because some io::run() are done).
+ Also, some channel were not awoken when clearing sockets.
+ channelDatas_ must be locked while accessed
+ setOnRecv() can be done after first packets on channel, so to avoid
to miss first packets, re-inject current stored datas.

Change-Id: I1d1335b5b6873b281c5357d18121a4bb525646f8

Send a request to delete all mappings created by previous
JAMI instances as soon as a new upnp IGD is discovered.
This is necessary with some implementations since they may
not respond well if the list of existing mappings is large.
Also make new requests for mappings with failing states.

Change-Id: I97308f1620128a63f111bcfc02122b472def5c85
Gitlab: #335

Sébastien Blin authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: Ib53fa87efdbbbc6b7561e3b37f27d7a8d51af163

Mohamed Chibani authored 4 years ago and Sébastien Blin committed 4 years ago

Implement a handler for automatic provisioning of port mappings through UPNP
and NAT-PMP protocols in NATed networks. A pool of UDP/TCP ports will be
allocated on start-up and periodically monitored such as when port allocation
is requested, it will be readily available. This will improve responsiveness
with some UPNP server implementations that may take several seconds to grant
a mapping request.

Gitlab: #335

Change-Id: I837945797d3b6bb337d3419bf77b4dc686558e5f

Mohamed Chibani authored 4 years ago and Sébastien Blin committed 4 years ago

Gitlab: #334

Change-Id: Iac00278c78ed2a4765d75143d692f1fd7539fd46

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

Add the patches that enable FEC in libopus codec in ffmpeg package.json
This ensures that the patches are properly applied, especially in the
Windows builds.

Issue: #5157
Change-Id: I438de63528894d3c7ded553f1982d664bda26e31

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

Add a rtcpCheckerThread to get packet loss estimation from the RTCP RR
packets, and send this information to the audio encoder.
This packet loss estimation is capped between 0 and 100.
When the new packet loss is higher than present estimation, the value
in the codec is set to the new value, in order to react to a quick
network quality drop.
When the new packet loss estimation is lower than present estimation,
the new value goes through a basic filter to avoid reducing the packet
loss estimation too fast.

Issue: #5157
Change-Id: Ibc1d3d6f6935e3b73c362bd212ef997cfd7b8d2f

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

Add setPacketLoss method which enable setting a new packet loss
estimation.
This method is based on the model of setBitrate in video_sender.

Issue: #5157
Change-Id: Ib757ce9c241f0b02b8a9a899b3d5a22e898203b3

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

Add a setPacketLoss method to enable setting a new packet loss
estimation in the codec.
This method follows the model of setBitrate and requires the
implementation of getCurrentAudioAVCtx and isDynPacketLossSupported,
which are also included in this patch.

Issue: #5157
Change-Id: I62f5246df1df0dc0867649299c6bbf6c12db43a2

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

The libopus codec only reads and impelemnts options from the
AVCodecContext at initialization. This means that packet loss
estimation can not be dynamically set.
This patch corrects this by reloading the packet loss estimation before
each encode. This ensures that each opus packet is encoded with FEC
data corresponding to the last packet loss estimation.
This patch does not check the previous packet loss estimation, as it
would increase software complexity, with one call of
opus_multistream_encoder_ctl() at best when packet loss is unchanged,
and two calls at most (one to get value, one to set value).

Issue: #5157
Change-Id: I3fce82d8b310a5751dc7fbc9a1b766d43c41c304

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

Adds a function to initialize the opus encoder with FEC support and a
default packet loss of 10%.
In future iterations, the default value should be set to 0 and the
packet loss adapted accordingly to RTCP control packets.

Issue: #5156
Change-Id: I79b9c5e38ed388466d7b71a394f961d2ff9b001f

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

Add option to enable FEC support in libopus encoder, and adds uses this
option to instanciate the opus encoder.
This option is disabled by default, in order to avoid changing default
ffmpeg behavior. If FEC support is enabled and packet loss is higher
than 0, the encoder adds ancilliary data to the packet so the decoder
can restore the previous packet, in case it is lost. As more data has
to be transmitted, this reduces the audio quality when using a fixed
bitrate.

Issue: #5156
Change-Id: Ic7d69ab122bd63c3ff880e46a0a0dcb316b00760

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

By default, sets the option to use FEC decoding with libopus when
preparing the decoder context.

Issue: #5155
Change-Id: Icc9dcbbd9df3f538eb2da794e2ff9b6a0380bdc2

Philip-Dylan Gleonec authored 4 years ago and Pierre Lespagnol committed 4 years ago

Patch ffmpeg to add an option (AVOption) to enable opus FEC decoding
using libopus and add FEC decoding when the option is enabled.

Issue: #5155
Change-Id: I78754a9d02a333c7d73da60b87fd40a6da544835

Pierre Lespagnol authored 4 years ago and Sébastien Blin committed 4 years ago

Fix unlinked hardware encoder/decoder cases (Conference, plugin, ...)

Change-Id: I045de3d24d89e93141309b6b353cd0bcabd8509f

Change-Id: Iab97236dce66a63a660251c495f64154c6db07b7

Change-Id: Ia5ba79e370642738df593493a6d1da52a9b1d9b8

Change-Id: I4c30fe6c8c99bdc518c0e81a89eaf5eb3ce2b726

Since https://review.jami.net/c/ring-daemon/+/16668 handleControlPacket
contains a shared_ptr to the mxsock pimpl. This means that
closeConnectionsWith don't call the destructor like before and the
shutdown is not triggered. So, force the shutdown to be called
to fix testShutdownCallbacks() and call correctly all shutdown
callbacks.

Change-Id: I8aa8bec05650066261ac61d5fd42e1f20f99830a

Change-Id: I5a31c01936ac4b8e3af6ebbb44665b03d0602a95

Change-Id: I2b64c929b2973b88d51786a328fcb0249560a687