viewer.git/indra/llwebrtc, branch 26.4

Merge tag 'Second_Life_Release#a4aa523-26.4' into 26.4

2026-09-04T15:25:54Z

Merge remote-tracking branch 'secondlife/release/26.4' into 26.4

2026-09-04T13:58:53Z

#6228 fix setting the default device after a region change

2026-09-02T11:19:15Z

Patch contributed by TJ Hecklezz

#6109 StopRecording() crash

2026-08-25T20:05:20Z

#6167 fix WebRTC logging level

2026-08-24T10:13:02Z

#6123 Fix worker thread crash on termination

2026-08-21T11:30:37Z

p#682 Fix shutdown crash and cut WebRTC teardown time

2026-08-12T21:29:29Z

terminate() released the thread unique_ptrs on the timeout path while the detached shutdown thread was still running the lambda that reads them, so a slow shutdown became a null deref at the next BlockingCall. Leave the members alone and report the timeout instead, so the caller leaks this object rather than deleting it out from under the detached thread. Reduce how often that timeout is reached at all: - Post the audio device stop instead of blocking on it. StopRecording and StopPlayout each wait up to 2s on a WASAPI thread, spending much of the 10s budget before the connections were touched. The device stop has to stay on the worker thread (AudioDeviceBuffer is guarded by a sequence checker bound to it), but the queue is FIFO, so ForceTerminate still runs after it while the signaling close gets on with its network-thread work. - Drain the peer connections in LLWebRTCVoiceClient::terminate() before handing off to llwebrtc::terminate(), bounded to 3s. An estate session can hold ten live connections -- the current region plus up to eight neighbours -- and closing them inline and serially under the shutdown budget is what made it tight. Draining also quiets the stats poll, so PeerConnection::Close() no longer blocks in WaitForPendingRequest() on an in-flight GetStats. Connections reaped by the drain are destroyed while webrtc is still fully alive, so they must unregister and release their peer connections normally; key the destructor's early-out off a new sWebRTCTerminated instead of isShuttingDown(). Notify signaling observers even when there was no peer connection to close, or a connection shut down before it finished initializing parks in VOICE_STATE_WAIT_FOR_CLOSE, which has no timeout of its own. Co-Authored-By: Claude Opus 5 (1M context)

Merge tag 'Second_Life_Release#932e8ac3-26.3' into 26.3

2026-08-07T23:01:40Z

Merge tag 'Second_Life_Release#0686467f-26.3' into 26.3

2026-08-07T02:24:49Z

p#682 Fix shutdown crash flushing peer connection stats

2026-08-06T23:54:44Z

LLWebRTCImpl::terminate() only posted the peer connection closes to the signaling thread, so they ran behind the BlockingCall that releases the factory -- after the factory and device module were gone, or not at all if the thread was destroyed with the task still queued. That matters for stats: PeerConnection::Close() flushes any in-flight GetStats request and delivers the report inline, and our callback walks mSignalingObserverList to hand it to the viewer. By then the viewer's LLVoiceWebRTCConnection objects have been destroyed -- on shutdown they deliberately skip unsetSignalingObserver, on the assumption that llwebrtc::terminate() already finished the job -- so the stats callback reaches into freed memory. Split the close out of LLWebRTCPeerConnectionImpl::terminate() into closeOnSignalingThread() and run it from a BlockingCall, so connections are closed and destroyed on the signaling thread before the factory, the device module or the observers go away. Also take ownership of the connection list up front so the detached thread in the timeout path isn't racing the main thread over it, clear the observer lists as part of the close, and skip the observer notification for a stats report delivered while shutting down. Co-Authored-By: Claude Opus 5 (1M context)