rtc: fail fast instead of hanging when FFI is used across fork()

The native runtime (tokio worker/IO threads in liblivekit_ffi) does not
survive fork() and cannot be rebuilt by re-initializing, so a child that
inherited an initialized FFI singleton would hang on every request. Guard
FfiClient.instance with the owning pid and raise a clear error instead.

Author: theomonnom

livekit-rtc/livekit/rtc/_ffi_client.py

@@ -213,11 +213,23 @@ class FfiClient:
 
     @classproperty
     def instance(cls) -> "FfiClient":
-        if cls._instance is None:
-            cls._instance = FfiClient()
-        return cls._instance
+        inst = cls._instance
+        if inst is not None and inst._pid != os.getpid():
+            # The native runtime (tokio worker/IO threads inside liblivekit_ffi)
+            # does not survive fork(), and re-initializing does not rebuild it,
+            # so a child that inherited an initialized FFI hangs on every request.
+            # Fail loudly instead of hanging.
+            raise RuntimeError(
+                "livekit.rtc was used in a parent process before fork(); the native "
+                "runtime cannot be used across fork(). Do not create or connect a Room "
+                "(or use any livekit.rtc object) before forking child processes."
+            )
+        if inst is None:
+            cls._instance = inst = FfiClient()
+        return inst
 
     def __init__(self) -> None:
+        self._pid = os.getpid()
         self._lock = threading.RLock()
         self._queue = FfiQueue[proto_ffi.FfiEvent]()