Add a linearizer on (appservice, stream) when handling ephemeral events.

changelog.d/11207.bugfix

@@ -0,0 +1 @@
+Fix a long-standing bug which could result in serialization errors and potentially duplicate transaction data when sending ephemeral events to application services. Contributed by @Fizzadar at Beeper.

synapse/handlers/appservice.py

@@ -34,6 +34,7 @@
 )
 from synapse.storage.databases.main.directory import RoomAliasMapping
 from synapse.types import JsonDict, RoomAlias, RoomStreamToken, UserID
+from synapse.util.async_helpers import Linearizer
 from synapse.util.metrics import Measure
 
 if TYPE_CHECKING:
@@ -58,6 +59,10 @@ def __init__(self, hs: "HomeServer"):
         self.current_max = 0
         self.is_processing = False
 
+        self._ephemeral_events_linearizer = Linearizer(
+            name="appservice_ephemeral_events"
+        )
+
     def notify_interested_services(self, max_token: RoomStreamToken) -> None:
         """Notifies (pushes) all application services interested in this event.
 
@@ -248,26 +253,37 @@ async def _notify_interested_services_ephemeral(
                     events = await self._handle_typing(service, new_token)
                     if events:
                         self.scheduler.submit_ephemeral_events_for_as(service, events)
+                    continue
 
-                elif stream_key == "receipt_key":
-                    events = await self._handle_receipts(service)
-                    if events:
-                        self.scheduler.submit_ephemeral_events_for_as(service, events)
-
-                    # Persist the latest handled stream token for this appservice
-                    await self.store.set_type_stream_id_for_appservice(
-                        service, "read_receipt", new_token
+                # Since we read/update the stream position for this AS/stream
+                with (
+                    await self._ephemeral_events_linearizer.queue(
+                        (service.id, stream_key)
                     )
+                ):
+                    if stream_key == "receipt_key":
+                        events = await self._handle_receipts(service, new_token)
+                        if events:
+                            self.scheduler.submit_ephemeral_events_for_as(
+                                service, events
+                            )
+
+                        # Persist the latest handled stream token for this appservice
+                        await self.store.set_type_stream_id_for_appservice(
+                            service, "read_receipt", new_token
+                        )
 
-                elif stream_key == "presence_key":
-                    events = await self._handle_presence(service, users)
-                    if events:
-                        self.scheduler.submit_ephemeral_events_for_as(service, events)
+                    elif stream_key == "presence_key":
+                        events = await self._handle_presence(service, users, new_token)
+                        if events:
+                            self.scheduler.submit_ephemeral_events_for_as(
+                                service, events
+                            )
 
-                    # Persist the latest handled stream token for this appservice
-                    await self.store.set_type_stream_id_for_appservice(
-                        service, "presence", new_token
-                    )
+                        # Persist the latest handled stream token for this appservice
+                        await self.store.set_type_stream_id_for_appservice(
+                            service, "presence", new_token
+                        )
 
     async def _handle_typing(
         self, service: ApplicationService, new_token: int
@@ -304,7 +320,9 @@ async def _handle_typing(
         )
         return typing
 
-    async def _handle_receipts(self, service: ApplicationService) -> List[JsonDict]:
+    async def _handle_receipts(
+        self, service: ApplicationService, new_token: Optional[int]
+    ) -> List[JsonDict]:
         """
         Return the latest read receipts that the given application service should receive.
 
@@ -323,14 +341,20 @@ async def _handle_receipts(self, service: ApplicationService) -> List[JsonDict]:
         from_key = await self.store.get_type_stream_id_for_appservice(
             service, "read_receipt"
         )
+        if new_token is not None and new_token <= from_key:
+            raise Exception("Rejecting token lower than stored: %s" % (new_token,))
+
         receipts_source = self.event_sources.sources.receipt
         receipts, _ = await receipts_source.get_new_events_as(
             service=service, from_key=from_key
         )
         return receipts
 
     async def _handle_presence(
-        self, service: ApplicationService, users: Collection[Union[str, UserID]]
+        self,
+        service: ApplicationService,
+        users: Collection[Union[str, UserID]],
+        new_token: Optional[int],
     ) -> List[JsonDict]:
         """
         Return the latest presence updates that the given application service should receive.
@@ -353,6 +377,9 @@ async def _handle_presence(
         from_key = await self.store.get_type_stream_id_for_appservice(
             service, "presence"
         )
+        if new_token is not None and new_token <= from_key:
+            raise Exception("Rejecting token lower than stored: %s" % (new_token,))
+
         for user in users:
             if isinstance(user, str):
                 user = UserID.from_string(user)