[V1][BugFix] Free encoder cache for aborted requests (#12545)

Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>

vllm/v1/core/encoder_cache_manager.py

@@ -38,7 +38,8 @@ class EncoderCacheManager:
     def get_cached_input_ids(self, request: Request) -> Set[int]:
         return self.cached.get(request.request_id, set())
 
-    def free(self, request: Request, input_id: int) -> None:
+    def free_encoder_input(self, request: Request, input_id: int) -> None:
+        """Free a single encoder input id for the request."""
         req_id = request.request_id
         if req_id not in self.cached:
             return
@@ -49,6 +50,12 @@ class EncoderCacheManager:
         self.num_free_slots += request.get_num_encoder_tokens(input_id)
         self.freed.append((req_id, input_id))
 
+    def free(self, request: Request) -> None:
+        """Free all cached input ids for the request."""
+        input_ids = self.get_cached_input_ids(request)
+        for input_id in input_ids:
+            self.free_encoder_input(request, input_id)
+
     def get_freed_ids(self) -> List[Tuple[str, int]]:
         freed = self.freed
         self.freed = []

vllm/v1/core/scheduler.py

@@ -202,7 +202,7 @@ class Scheduler:
                 # which have output tokens.
                 num_new_tokens = request.num_tokens - num_computed_tokens
                 if num_new_tokens == 0:
-                    # The happens when prompt length is divisible by the block
+                    # This happens when prompt length is divisible by the block
                     # size and all blocks are cached. Now we force to recompute
                     # the last block. Note that we have to re-compute an entire
                     # block because allocate_slots() assumes num_computed_tokens
@@ -269,6 +269,7 @@ class Scheduler:
 
         # Get the longest common prefix among all requests in the running queue.
         # This can be potentially used for cascade attention.
+        num_common_prefix_blocks = 0
         if self.running:
             any_request = self.running[0]
             num_common_prefix_blocks = (
@@ -433,7 +434,8 @@ class Scheduler:
                     if start_pos + num_tokens <= request.num_computed_tokens:
                         # The encoder output is already processed and stored
                         # in the decoder's KV cache.
-                        self.encoder_cache_manager.free(request, input_id)
+                        self.encoder_cache_manager.free_encoder_input(
+                            request, input_id)
 
             if request.num_computed_tokens == request.num_tokens:
                 req_index = model_runner_output.req_id_to_index[req_id]
@@ -445,8 +447,10 @@ class Scheduler:
                 # TODO: Update the KV cache manager for prefix caching.
 
                 # Check for stop and update request state.
-                # This must be called before me make the EngineCoreOutput.
+                # This must be called before we make the EngineCoreOutput.
                 stopped = self._check_stop(request)
+                if stopped:
+                    self._free_request(request)
 
                 # Add EngineCoreOutput for this Request.
                 output = EngineCoreOutput(
@@ -472,7 +476,6 @@ class Scheduler:
         if (request.num_tokens >= self.max_model_len
                 or request.num_output_tokens >= request.max_tokens):
             request.status = RequestStatus.FINISHED_LENGTH_CAPPED
-            self._free_request(request)
             return True
 
         sampling_params = request.sampling_params
@@ -480,13 +483,11 @@ class Scheduler:
         if (not sampling_params.ignore_eos
                 and last_token_id == request.eos_token_id):
             request.status = RequestStatus.FINISHED_STOPPED
-            self._free_request(request)
             return True
 
         if last_token_id in (sampling_params.stop_token_ids or ()):
             request.status = RequestStatus.FINISHED_STOPPED
             request.stop_reason = last_token_id
-            self._free_request(request)
             return True
         return False
 
@@ -525,6 +526,7 @@ class Scheduler:
     def _free_request(self, request: Request) -> None:
         assert request.is_finished()
         self.kv_cache_manager.free(request)
+        self.encoder_cache_manager.free(request)
         self.running_reqs_data.pop(request.request_id, None)
         del self.requests[request.request_id]
         self.finished_req_ids.add(request.request_id)