onStateChanged can be used to just clean some structure, for example
with TlsTurnEndpoint. Also, if called after move, the data can be
scrapped and incorrect.
Also, fix a cast in TlsSocketEndpoint when using ICE.

Change-Id: I8104bc8a0fd8e9cd3dae92e06eee45c22feced45

Adrien Béraud authored 4 years ago and Kateryna Kostiuk committed 4 years ago

Change-Id: I0bbcdd7e5d496ce79f31ed2c172c9b530c73f160

Ice destruction can take too long and delay media initialization.
This causes first packets sent by the peer to be missed.

This patch fix multi-device call bug with Android (and MediaCodec)

Change-Id: I4ceb1838e668d5be2ed36d1e0d2f00113833f63c

Change-Id: Iafd7d28c0cf808d73e5cfc3dc0a7916d4b3006a9

Mohamed Amine Younes Bouacida authored 5 years ago and Adrien Béraud committed 4 years ago

Change-Id: I2ff7d879de55e18bedcbddce74f743ffe8755ca6

Change-Id: I8495f79ead4e15df756238624198f46ec72032f5

Change-Id: I54c876133ec39454d4d71b8122b744da82dd030b

Sébastien Blin authored 4 years ago and Yang Wang committed 4 years ago

isPasswordValid must return a boolean

Change-Id: I6465bfc73b08f203b2fa0b53deaf71b3406cabd6

A shared_ptr<Call> could be deleted during call to hangup,
causing call to detachCall() and deadlock.

Copy id set to avoid potential deadlock.

Change-Id: Ic43280351f19408fc308362f3d287090b90c736d

Change-Id: Ice7459ba421c28f72c05bfda5283b733acef4695

Change-Id: Ib6a956ac1b8b4667e017f6bda7dd26488076bb47

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

Change-Id: Ib3d36408841d1b339a00bf9135b308afaa93b218

Vladimir Stoiakin authored 6 years ago and Adrien Béraud committed 4 years ago

Change-Id: I331522a09257d4ae787893ff72c270886c1aa210
Reviewed-by: Sébastien Blin <sebastien.blin@savoirfairelinux.com>

Change-Id: If1ff8a0213b9687b0899b40d17feb56e491b35ca

Change-Id: I969de308275f02556bd9f82cf5d5c0973652dcab

Kateryna Kostiuk authored 4 years ago and Sébastien Blin committed 4 years ago

Change-Id: I2cbc8347859136e54235a38afdbf52b5475a355b

Change-Id: Ia12da6f9e58ec0cd0100eaba206e1d9e711ca0e1

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

Better compromise between quality and size of the file

Change-Id: I0fa59683f4f49e108fcbe41aefa428f291309095

Change-Id: I017cf5a0fd76659efb8ff8e7a65567b5c357963d

Change-Id: Icddfa964f7980af5915b110745cb922adc10dba4

Change-Id: I59414f3d0c55f13aa8a91e498ee8c694fa73ffe3

git.xiph.org doesn't exists anymore

Change-Id: I93a478888319bc1de5aa11713358d20a19e8bd55

Change-Id: I44c1c8c1ae653e08ba2824272d22a90b47becd57

Change-Id: Ic4a90183ae7589d4ddf7dd196e54cd6e8cba02f3

This aims to fix the some negotiation failure when:
UserA (TURN ipv4) calls UserB (ipv4 only)

In fact, the first user will take some time to know the userB
addresses to allow on the TURN. So, during this time, the TURN will
give some EPIPE errors. EPIPE is code 32 defined by 120032 in pjsip
(PJ_STATUS_FROM_OS(EPIPE)).

So previous code was a mistake

Change-Id: If38c5d36dc38bfe89e3b2ccfbf1b64c934639adb

In the case whare two packets like this were received:

P1
+----------------+-------------+---------------+
| data (X bytes) | header (2b) | data (1 byte) |
+----------------+-------------+---------------+

P2
+----------------+
| data (X bytes) |
+----------------+

The RTP loop used the last byte of p1 and the first
byte of P2 as the header. Leading to malformed packets

Change-Id: Ie714fcf0bab1e372f7433342ed60ed4e6d20aff3

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

Change-Id: Icd33b5a7d054e83a141f99f2a4fb1b539b125cb0

Change-Id: I66b07df09098833a1ad59df008f04c521edf1449

This patch move checkAudio() out of call mutex. Because it could cause
deadlock if subcallStateChanged() called at the same time.

Change-Id: Iee06874fb79c8e0953e061f7a669b3f61f13d8b9

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

Change-Id: I5c991b37011b4c3d97114c9f2b8f05b742e029a3

Ming Rui Zhang authored 4 years ago and Adrien Béraud committed 4 years ago

it is to prevent pj_shutdown from doing harmful actions to sipLink destruction

Change-Id: Icb472b2f3350f39f31370a5dba313abe3c8a7bb1

Before this patch when the the sender is restarted, we get the last sequence number from media_encoder that is refreshed on each frame only. But the sequence number continue to be incremented on each RTP packet sent, this result in rollover counter (ROC) desynchronisation in srtp context because the largest sequence number sent is now superior to sequence number that is supposed to occur only when sequence number overflow 65535.

With the patch when the sender is restarted, we keep the last sequence number from socket_pair (RTP level)

Change-Id: I531e3d0a073c251c78bbf9f0ffc702aafaf6ccc8

Restart encoder when Codec or hardware API don't support dynamic bitrate.

Change-Id: Ic5fb95d405dc7694125ccb3c80020af39441b2b7

With these changes, Jami only got on_data_sent for data that Jami
send. In previous state, some on_data_sent comes from TURN refresh
because it's managed by the ice_session in pjsip.

Change-Id: I76109f9662aeca01ffccd884dac2a5aab7ab2503

Change-Id: Ie9a7fdca48dc086af7e21990a490c93041f12428

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

Change-Id: I2dcb7e71738f6f71ebb9db770902d64b24920022

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

The current design describes SIPVoIPLink as a Singleton. This cause
some behaviors where the link is destroyed then immediately created
or vice-versa and its creation/destruction can't really be planned.
This design was made to allow multiple Managers to exist. However,
we never do that, so let's keep it simple, there is only one
Manager and all accounts needs that SIPVoIPLink

This patch change this behavior and the voip link is now owned
by the manager.

Change-Id: I248e41742d342cf452a5503f532fe5ab862166e6

- Decrease drops threshold for bitrate diminution (10% to 5%)
- Space number of decrease from 2 to 1 every 500ms for delays detection
- Always use bitrate from MediaStream during encoder init

Change-Id: If3d696b8fc1bc7f6fabe98f70652921d4113b7b7

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

Use RateMode parameter instead of autoQuality, it makes more sense for the encoder

Change-Id: Ia5b6acdfb08fb146dec11b75f1d26127a0ed520b

Change-Id: I2a1a29a8fb99cf8eea40d35c511dadebd0fb46a2