diff --git a/vllm/v1/core/kv_cache_manager.py b/vllm/v1/core/kv_cache_manager.py
index 41d9f1b65c23..33cc33beb2ae 100644
--- a/vllm/v1/core/kv_cache_manager.py
+++ b/vllm/v1/core/kv_cache_manager.py
@@ -291,6 +291,7 @@ def allocate_slots(
             # cache a block after is has finished recving.
             self.num_cached_block[request.request_id] = len(
                 new_computed_blocks)
+            return new_blocks
 
         self.cache_blocks(
             request=request,