Avoid fetching unused account data in sync. (#14973)

The per-room account data is no longer unconditionally
fetched, even if all rooms will be filtered out.

Global account data will not be fetched if it will all be
filtered out.

4 files changed, 57 insertions, 48 deletions

diff --git a/synapse/handlers/account_data.py b/synapse/handlers/account_data.py
index 67e789eef7..797de46dbc 100644
--- a/synapse/handlers/account_data.py
+++ b/synapse/handlers/account_data.py
@@ -343,10 +343,12 @@ class AccountDataEventSource(EventSource[int, JsonDict]):
                 }
             )
 
-        (
-            account_data,
-            room_account_data,
-        ) = await self.store.get_updated_account_data_for_user(user_id, last_stream_id)
+        account_data = await self.store.get_updated_global_account_data_for_user(
+            user_id, last_stream_id
+        )
+        room_account_data = await self.store.get_updated_room_account_data_for_user(
+            user_id, last_stream_id
+        )
 
         for account_data_type, content in account_data.items():
             results.append({"type": account_data_type, "content": content})
diff --git a/synapse/handlers/initial_sync.py b/synapse/handlers/initial_sync.py
index 191529bd8e..1a29abde98 100644
--- a/synapse/handlers/initial_sync.py
+++ b/synapse/handlers/initial_sync.py
@@ -154,9 +154,8 @@ class InitialSyncHandler:
 
         tags_by_room = await self.store.get_tags_for_user(user_id)
 
-        account_data, account_data_by_room = await self.store.get_account_data_for_user(
-            user_id
-        )
+        account_data = await self.store.get_global_account_data_for_user(user_id)
+        account_data_by_room = await self.store.get_room_account_data_for_user(user_id)
 
         public_room_ids = await self.store.get_public_room_ids()
 
diff --git a/synapse/handlers/room_member.py b/synapse/handlers/room_member.py
index d236cc09b5..6e7141d2ef 100644
--- a/synapse/handlers/room_member.py
+++ b/synapse/handlers/room_member.py
@@ -484,7 +484,7 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
             user_id: The user's ID.
         """
         # Retrieve user account data for predecessor room
-        user_account_data, _ = await self.store.get_account_data_for_user(user_id)
+        user_account_data = await self.store.get_global_account_data_for_user(user_id)
 
         # Copy direct message state if applicable
         direct_rooms = user_account_data.get(AccountDataTypes.DIRECT, {})
diff --git a/synapse/handlers/sync.py b/synapse/handlers/sync.py
index 202b35eee6..399685e5b7 100644
--- a/synapse/handlers/sync.py
+++ b/synapse/handlers/sync.py
@@ -1444,9 +1444,9 @@ class SyncHandler:
 
         logger.debug("Fetching account data")
 
-        account_data_by_room = await self._generate_sync_entry_for_account_data(
-            sync_result_builder
-        )
+        # Global account data is included if it is not filtered out.
+        if not sync_config.filter_collection.blocks_all_global_account_data():
+            await self._generate_sync_entry_for_account_data(sync_result_builder)
 
         # Presence data is included if the server has it enabled and not filtered out.
         include_presence_data = bool(
@@ -1472,9 +1472,7 @@ class SyncHandler:
             (
                 newly_joined_rooms,
                 newly_left_rooms,
-            ) = await self._generate_sync_entry_for_rooms(
-                sync_result_builder, account_data_by_room
-            )
+            ) = await self._generate_sync_entry_for_rooms(sync_result_builder)
 
             # Work out which users have joined or left rooms we're in. We use this
             # to build the presence and device_list parts of the sync response in
@@ -1717,35 +1715,29 @@ class SyncHandler:
 
     async def _generate_sync_entry_for_account_data(
         self, sync_result_builder: "SyncResultBuilder"
-    ) -> Dict[str, Dict[str, JsonDict]]:
-        """Generates the account data portion of the sync response.
+    ) -> None:
+        """Generates the global account data portion of the sync response.
 
         Account data (called "Client Config" in the spec) can be set either globally
         or for a specific room. Account data consists of a list of events which
         accumulate state, much like a room.
 
-        This function retrieves global and per-room account data. The former is written
-        to the given `sync_result_builder`. The latter is returned directly, to be
-        later written to the `sync_result_builder` on a room-by-room basis.
+        This function retrieves global account data and writes it to the given
+        `sync_result_builder`. See `_generate_sync_entry_for_rooms` for handling
+         of per-room account data.
 
         Args:
             sync_result_builder
-
-        Returns:
-            A dictionary whose keys (room ids) map to the per room account data for that
-            room.
         """
         sync_config = sync_result_builder.sync_config
         user_id = sync_result_builder.sync_config.user.to_string()
         since_token = sync_result_builder.since_token
 
         if since_token and not sync_result_builder.full_state:
-            # TODO Do not fetch room account data if it will be unused.
-            (
-                global_account_data,
-                account_data_by_room,
-            ) = await self.store.get_updated_account_data_for_user(
-                user_id, since_token.account_data_key
+            global_account_data = (
+                await self.store.get_updated_global_account_data_for_user(
+                    user_id, since_token.account_data_key
+                )
             )
 
             push_rules_changed = await self.store.have_push_rules_changed_for_user(
@@ -1758,28 +1750,26 @@ class SyncHandler:
                     sync_config.user
                 )
         else:
-            # TODO Do not fetch room account data if it will be unused.
-            (
-                global_account_data,
-                account_data_by_room,
-            ) = await self.store.get_account_data_for_user(sync_config.user.to_string())
+            all_global_account_data = await self.store.get_global_account_data_for_user(
+                user_id
+            )
 
-            global_account_data = dict(global_account_data)
+            global_account_data = dict(all_global_account_data)
             global_account_data["m.push_rules"] = await self.push_rules_for_user(
                 sync_config.user
             )
 
-        account_data_for_user = await sync_config.filter_collection.filter_account_data(
-            [
-                {"type": account_data_type, "content": content}
-                for account_data_type, content in global_account_data.items()
-            ]
+        account_data_for_user = (
+            await sync_config.filter_collection.filter_global_account_data(
+                [
+                    {"type": account_data_type, "content": content}
+                    for account_data_type, content in global_account_data.items()
+                ]
+            )
         )
 
         sync_result_builder.account_data = account_data_for_user
 
-        return account_data_by_room
-
     async def _generate_sync_entry_for_presence(
         self,
         sync_result_builder: "SyncResultBuilder",
@@ -1839,9 +1829,7 @@ class SyncHandler:
         sync_result_builder.presence = presence
 
     async def _generate_sync_entry_for_rooms(
-        self,
-        sync_result_builder: "SyncResultBuilder",
-        account_data_by_room: Dict[str, Dict[str, JsonDict]],
+        self, sync_result_builder: "SyncResultBuilder"
     ) -> Tuple[AbstractSet[str], AbstractSet[str]]:
         """Generates the rooms portion of the sync response. Populates the
         `sync_result_builder` with the result.
@@ -1852,7 +1840,6 @@ class SyncHandler:
 
         Args:
             sync_result_builder
-            account_data_by_room: Dictionary of per room account data
 
         Returns:
             Returns a 2-tuple describing rooms the user has joined or left.
@@ -1865,9 +1852,30 @@ class SyncHandler:
         since_token = sync_result_builder.since_token
         user_id = sync_result_builder.sync_config.user.to_string()
 
+        blocks_all_rooms = (
+            sync_result_builder.sync_config.filter_collection.blocks_all_rooms()
+        )
+
+        # 0. Start by fetching room account data (if required).
+        if (
+            blocks_all_rooms
+            or sync_result_builder.sync_config.filter_collection.blocks_all_room_account_data()
+        ):
+            account_data_by_room: Mapping[str, Mapping[str, JsonDict]] = {}
+        elif since_token and not sync_result_builder.full_state:
+            account_data_by_room = (
+                await self.store.get_updated_room_account_data_for_user(
+                    user_id, since_token.account_data_key
+                )
+            )
+        else:
+            account_data_by_room = await self.store.get_room_account_data_for_user(
+                user_id
+            )
+
         # 1. Start by fetching all ephemeral events in rooms we've joined (if required).
         block_all_room_ephemeral = (
-            sync_result_builder.sync_config.filter_collection.blocks_all_rooms()
+            blocks_all_rooms
             or sync_result_builder.sync_config.filter_collection.blocks_all_room_ephemeral()
         )
         if block_all_room_ephemeral:
@@ -2294,7 +2302,7 @@ class SyncHandler:
         room_builder: "RoomSyncResultBuilder",
         ephemeral: List[JsonDict],
         tags: Optional[Dict[str, Dict[str, Any]]],
-        account_data: Dict[str, JsonDict],
+        account_data: Mapping[str, JsonDict],
         always_include: bool = False,
     ) -> None:
         """Populates the `joined` and `archived` section of `sync_result_builder`