Merge pull request #8572 from matrix-org/rav/cache_hacking/2

Push some deferred wrangling down into DeferredCache

changelog.d/8572.misc

@@ -0,0 +1 @@
+Modify `DeferredCache.get()` to return `Deferred`s instead of `ObservableDeferred`s.

synapse/util/caches/deferred_cache.py

@@ -57,7 +57,7 @@ class DeferredCache(Generic[KT, VT]):
     """Wraps an LruCache, adding support for Deferred results.
 
     It expects that each entry added with set() will be a Deferred; likewise get()
-    may return an ObservableDeferred.
+    will return a Deferred.
     """
 
     __slots__ = (
@@ -130,16 +130,22 @@ def get(
         key: KT,
         callback: Optional[Callable[[], None]] = None,
         update_metrics: bool = True,
-    ) -> Union[ObservableDeferred, VT]:
+    ) -> defer.Deferred:
         """Looks the key up in the caches.
 
+        For symmetry with set(), this method does *not* follow the synapse logcontext
+        rules: the logcontext will not be cleared on return, and the Deferred will run
+        its callbacks in the sentinel context. In other words: wrap the result with
+        make_deferred_yieldable() before `await`ing it.
+
         Args:
-            key(tuple)
-            callback(fn): Gets called when the entry in the cache is invalidated
+            key:
+            callback: Gets called when the entry in the cache is invalidated
             update_metrics (bool): whether to update the cache hit rate metrics
 
         Returns:
-            Either an ObservableDeferred or the result itself
+            A Deferred which completes with the result. Note that this may later fail
+            if there is an ongoing set() operation which later completes with a failure.
 
         Raises:
             KeyError if the key is not found in the cache
@@ -152,15 +158,15 @@ def get(
                 m = self.cache.metrics
                 assert m  # we always have a name, so should always have metrics
                 m.inc_hits()
-            return val.deferred
+            return val.deferred.observe()
 
         val2 = self.cache.get(
             key, _Sentinel.sentinel, callbacks=callbacks, update_metrics=update_metrics
         )
         if val2 is _Sentinel.sentinel:
             raise KeyError()
         else:
-            return val2
+            return defer.succeed(val2)
 
     def get_immediate(
         self, key: KT, default: T, update_metrics: bool = True
@@ -173,7 +179,36 @@ def set(
         key: KT,
         value: defer.Deferred,
         callback: Optional[Callable[[], None]] = None,
-    ) -> ObservableDeferred:
+    ) -> defer.Deferred:
+        """Adds a new entry to the cache (or updates an existing one).
+
+        The given `value` *must* be a Deferred.
+
+        First any existing entry for the same key is invalidated. Then a new entry
+        is added to the cache for the given key.
+
+        Until the `value` completes, calls to `get()` for the key will also result in an
+        incomplete Deferred, which will ultimately complete with the same result as
+        `value`.
+
+        If `value` completes successfully, subsequent calls to `get()` will then return
+        a completed deferred with the same result. If it *fails*, the cache is
+        invalidated and subequent calls to `get()` will raise a KeyError.
+
+        If another call to `set()` happens before `value` completes, then (a) any
+        invalidation callbacks registered in the interim will be called, (b) any
+        `get()`s in the interim will continue to complete with the result from the
+        *original* `value`, (c) any future calls to `get()` will complete with the
+        result from the *new* `value`.
+
+        It is expected that `value` does *not* follow the synapse logcontext rules - ie,
+        if it is incomplete, it runs its callbacks in the sentinel context.
+
+        Args:
+            key: Key to be set
+            value: a deferred which will complete with a result to add to the cache
+            callback: An optional callback to be called when the entry is invalidated
+        """
         if not isinstance(value, defer.Deferred):
             raise TypeError("not a Deferred")
 
@@ -187,6 +222,8 @@ def set(
         if existing_entry:
             existing_entry.invalidate()
 
+        # XXX: why don't we invalidate the entry in `self.cache` yet?
+
         self._pending_deferred_cache[key] = entry
 
         def compare_and_pop():
@@ -230,7 +267,9 @@ def eb(_fail):
         # _pending_deferred_cache to the real cache.
         #
         observer.addCallbacks(cb, eb)
-        return observable
+
+        # we return a new Deferred which will be called before any subsequent observers.
+        return observable.observe()
 
     def prefill(self, key: KT, value: VT, callback: Callable[[], None] = None):
         callbacks = [callback] if callback else []

synapse/util/caches/descriptors.py

@@ -23,7 +23,6 @@
 
 from synapse.logging.context import make_deferred_yieldable, preserve_fn
 from synapse.util import unwrapFirstError
-from synapse.util.async_helpers import ObservableDeferred
 from synapse.util.caches.deferred_cache import DeferredCache
 
 logger = logging.getLogger(__name__)
@@ -156,7 +155,7 @@ def __get__(self, obj, owner):
             keylen=self.num_args,
             tree=self.tree,
             iterable=self.iterable,
-        )  # type: DeferredCache[Tuple, Any]
+        )  # type: DeferredCache[CacheKey, Any]
 
         def get_cache_key_gen(args, kwargs):
             """Given some args/kwargs return a generator that resolves into
@@ -208,26 +207,12 @@ def _wrapped(*args, **kwargs):
                 kwargs["cache_context"] = _CacheContext.get_instance(cache, cache_key)
 
             try:
-                cached_result_d = cache.get(cache_key, callback=invalidate_callback)
-
-                if isinstance(cached_result_d, ObservableDeferred):
-                    observer = cached_result_d.observe()
-                else:
-                    observer = defer.succeed(cached_result_d)
-
+                ret = cache.get(cache_key, callback=invalidate_callback)
             except KeyError:
                 ret = defer.maybeDeferred(preserve_fn(self.orig), obj, *args, **kwargs)
+                ret = cache.set(cache_key, ret, callback=invalidate_callback)
 
-                def onErr(f):
-                    cache.invalidate(cache_key)
-                    return f
-
-                ret.addErrback(onErr)
-
-                result_d = cache.set(cache_key, ret, callback=invalidate_callback)
-                observer = result_d.observe()
-
-            return make_deferred_yieldable(observer)
+            return make_deferred_yieldable(ret)
 
         wrapped = cast(_CachedFunction, _wrapped)
 
@@ -286,7 +271,7 @@ def __init__(self, orig, cached_method_name, list_name, num_args=None):
 
     def __get__(self, obj, objtype=None):
         cached_method = getattr(obj, self.cached_method_name)
-        cache = cached_method.cache
+        cache = cached_method.cache  # type: DeferredCache[CacheKey, Any]
         num_args = cached_method.num_args
 
         @functools.wraps(self.orig)
@@ -326,14 +311,11 @@ def arg_to_cache_key(arg):
             for arg in list_args:
                 try:
                     res = cache.get(arg_to_cache_key(arg), callback=invalidate_callback)
-                    if not isinstance(res, ObservableDeferred):
-                        results[arg] = res
-                    elif not res.has_succeeded():
-                        res = res.observe()
+                    if not res.called:
                         res.addCallback(update_results_dict, arg)
                         cached_defers.append(res)
                     else:
-                        results[arg] = res.get_result()
+                        results[arg] = res.result
                 except KeyError:
                     missing.add(arg)