[Bugfix]Add register_kv_cache in ucm_connector (#5657)

### What this PR does / why we need it?
To adapt different shapes of the KV cache, UCM optimized the
initialization of store by moving it into `register_kv_caches`.
Therefore, this update adds `register_kv_caches` interface to
UCMConnectorV1.

### How was this patch tested?

- vLLM version: v0.13.0
- vLLM main:
2f4e6548ef

Signed-off-by: UnifiedCacheManager <unifiedcachem@163.com>

vllm_ascend/distributed/ucm_connector.py

@@ -40,6 +40,15 @@ class UCMConnectorV1(KVConnectorBase_V1):
     # ==============================
     # Worker-side methods
     # ==============================
+    def register_kv_caches(self, kv_caches: dict[str, torch.Tensor]) -> None:
+        """
+        Initialize with the KV caches. Useful for pre-registering the
+        KV Caches in the KVConnector (e.g. for NIXL).
+        Args:
+            kv_caches: A dictionary mapping layer names to KV cache tensors.
+        """
+        self._ucm_engine.register_kv_caches(kv_caches)
+
     def start_load_kv(self, forward_context: "ForwardContext",
                       **kwargs: Any) -> None:
         """