* WIP commit
* test padding only with a gap scenario
* Debug
* Fix video corruption, need buffer to include payload and translated header size
* Revert incorrect change
* Fix VP8 translation to return buffer with proper length
* Restore 7-bit mode
* WIP commit
* Clean up
* More clean up and tests compiling again
* Fix tests
* Allow participants to create new connection for publishing new tracks
new param 'publish=<name>' enables publishing new connection with publish
only grants. Also, identity is appended with '#<name>'
* Add integration test to duplicate publish only connection
Signed-off-by: shishir gowda <shishir@livekit.io>
* Reuse token in test and set subscribeGrant to false on publish param
Signed-off-by: shishir gowda <shishir@livekit.io>
* create auth token with grants
Signed-off-by: shishir gowda <shishir@livekit.io>
* Signal ice on negotiate in test client
Signed-off-by: shishir gowda <shishir@livekit.io>
* Update pkg/service/rtcservice.go
Co-authored-by: David Zhao <david@davidzhao.com>
Co-authored-by: David Zhao <david@davidzhao.com>
* Make VP8 packet translation thread-safe.
Was using one packet from pool for all VP8 translation which was not thread safe.
Grab packets from the pool when needed for VP8 translation and return to pool after done.
Do not grab packet from pool if the header size between incoming and translated matches.
That also saves copying the packet payload.
* Keep Get/Put in the same function.
* Use protocol friendly StreamedTracksUpdate
* WIP commit
* Stream allocator update
* subtract the requested bandwidth as delta from Allocate could be adding to bandwidth
* Calculate delta correctly
* correct comment
* Simplify eventCh per David's suggestion
* Simplifying (hopefully) sfu.DownTrack
* Remove unnecessary check as pdding only packets are dropped before that check
* Temporal filtering max layer
* - Split out forwarder bits into a separate structure
- Address comments from Jie and David
* Remove debug and unneeded stuff
* Fix test
* Remove unneeded default initialization
* WIP branch to hash out down stream allocator.
* Plug more bits of stream allocator
* update protocol
* remove SignalRequest_Simulcast (#154)
* Plug more bits of stream allocator
* Handle simulcast track available layers change
* WIP branch to hash out down stream allocator.
* Plug more bits of stream allocator
* Handle simulcast track available layers change
* Adopt signature of call to AdjustAllocation
* Move StreamAllocator to PCTransport and allocate only for subscriber.
* Move streamallocator to ion-sfu
* Start/Stop of streamallocator
* Use StreamAllocator for subscriber bandwidth management.
* Do not allocate in ADD_TRACK
* Set payload in constructor
* - Add some logging
- Protocol message to notify clients of paused/resumed streams
* named return
* oops correct sense of isPausing
* Update pkg/sfu/streamallocator.go
Committing David's suggestion.
Co-authored-by: David Zhao <david@davidzhao.com>
* - Log estimate changes/commits.
- Catch more than epsilon change oscillating for long time.
Co-authored-by: David Colburn <xero73@gmail.com>
Co-authored-by: David Zhao <david@davidzhao.com>
* configurable node track limit
* sample config
* todos
* end of file new line
* default max num tracks
* bandwidth limit
* client message for limit exceeded node
* 10 Gbps default network limit
Use max layer subscription of down tracks to
prioritize bandwidth allocation. The ordering is
- Highest spatial layer
- Highest temporal layer, if spatial layer matches
- Down track id if both layers match (this is to
prevent unnecessary re-ordering)
Testing:
--------
(Really need to make a DownTrack interface so that
we can mock it and use it in tests)
For now, just a sanity check.
Small step on the way to making StreamAllocator prioritization of tracks.
With the new callback into StreamAllocator, the idea is to use the
max layer information to do track prioritization.
Testing:
--------
Sanity check that sample app works
* track numParticipants in room
* only track participant if not a hidden participant
* adjust coding style to use ++
* fix typo
* fix missing nil check
* update roomstore with new numParticipants on participantChanged
* only update roomstore if participant is not hidden
* call StoreRoom directly after StoreParticipant when joining/leaving
* Handle multiple codecs in renegotiation
update pion to v3.1.9 for answer same order of codec as publisher.
register enable codecs in subscriber peerconnectin created.
add codec parameter to buffer.bind
buffer should use the codec of TrackRemote as it's codec mime.
sent h264blankframe when DownTrack closing
When we RLock during write cycles, the mutex spends the majority of its time
staying locked. As new participants join, they have to acquire the WLock
before downtracks could be add it.
In load test scenarios (25 participants joining together), it's common to see
goroutine dump showing MediaTrack.AddSubscriber -> DownTrack.storeDownTrack trying to acquire mutex, and never able to acquire it.
David Zhao