Fix NONE_HASH import: use module ref instead of from-import (value binding bug)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

third_party/vllm/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py

@@ -142,6 +142,7 @@ class MooncakeConnectorMetadata(KVConnectorMetadata):
         # Hash table sync: scheduler → worker (for direct RDMA read)
         self.hash_table_updates: dict[str, int] = {}  # hex hash → block_id
         self.hash_table_removals: set[str] = set()
+        self.token_hash_updates: dict[str, int] = {}  # str(hash(tokens)) → block_id
 
     def add_new_req(
         self,

third_party/vllm/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_utils.py

@@ -69,10 +69,11 @@ class MooncakeBootstrapServer:
         self.server: uvicorn.Server | None = None
 
         # Direct RDMA read support
-        self._hash_table: dict[str, int] = {}  # hex BlockHash → block_id (legacy)
-        self._kv_info: dict | None = None  # set by worker at register_kv_caches
-        self._pinned: dict[str, list[int]] = {}  # pin_token → block_ids
-        self._block_pool = None  # set by scheduler for token-based lookup
+        self._hash_table: dict[str, int] = {}  # hex BlockHash → block_id
+        self._token_hash_table: dict[str, int] = {}  # str(hash(token_tuple)) → block_id
+        self._kv_info: dict | None = None
+        self._pinned: dict[str, list[int]] = {}
+        self._block_pool = None
 
     def __del__(self):
         self.shutdown()
@@ -227,14 +228,10 @@ class MooncakeBootstrapServer:
     ) -> tuple[list[int | None], list[int]]:
         """Look up cached blocks by computing hashes using the synced hash table.
 
-        Computes block hashes from token_ids using the same hash function
-        as the scheduler, then looks up in the synced _hash_table.
+        Uses module-level NONE_HASH (accessed via module ref to get latest value
+        after init_none_hash is called).
         """
-        from vllm.v1.core.kv_cache_utils import (
-            BlockHash,
-            hash_block_tokens,
-            NONE_HASH,
-        )
+        import vllm.v1.core.kv_cache_utils as kv_utils
         from vllm.utils.hashing import sha256_cbor
 
         block_size = self._kv_info.get("block_size", 512) if self._kv_info else 512
@@ -244,19 +241,20 @@ class MooncakeBootstrapServer:
 
         block_ids: list[int | None] = []
         pinned_ids: list[int] = []
-        prev_hash = NONE_HASH
+        prev_hash = kv_utils.NONE_HASH  # module-level ref, always current
 
         for i in range(num_blocks):
             block_tokens = tuple(token_ids[i * block_size:(i + 1) * block_size])
-            block_hash = hash_block_tokens(sha256_cbor, prev_hash, block_tokens, None)
+            block_hash = kv_utils.hash_block_tokens(
+                sha256_cbor, prev_hash, block_tokens, None)
             prev_hash = block_hash
 
             bid = self._hash_table.get(block_hash.hex())
             if i == 0:
                 table_sample = next(iter(self._hash_table)) if self._hash_table else "empty"
                 logger.info(
-                    "_lookup_by_tokens: block0 hash=%s, NONE_HASH=%s, table_sample=%s",
-                    block_hash.hex()[:16], NONE_HASH.hex()[:16], table_sample[:16])
+                    "_lookup: hash=%s NONE=%s tbl=%s",
+                    block_hash.hex()[:12], kv_utils.NONE_HASH.hex()[:12], table_sample[:12])
             if bid is not None:
                 block_ids.append(bid)
                 pinned_ids.append(bid)