[P/D] Improve the performance of Layerwise Connector (#5303)

### What this PR does / why we need it?
Improve the performance of Layerwise Connector, mainly includes the
following points:
1. Use event synchronize to replace stream synchronize.
2. Access metaserver when scheduling.
3. Transfer kvcache each Chunk prefill segmentation.

### Does this PR introduce _any_ user-facing change?
No.

### How was this patch tested?
By CI.
- vLLM version: release/v0.13.0
- vLLM main:
5fbfa8d9ef

---------

Signed-off-by: nwpu-zxr <zhouxuerong2@huawei.com>
Signed-off-by: liziyu <liziyu16@huawei.com>
Signed-off-by: wangxiaoteng <wangxiaoteng@huawei.com>
Co-authored-by: liziyu <liziyu16@huawei.com>
Co-authored-by: wangxiaoteng <wangxiaoteng@huawei.com>

vllm_ascend/attention/attention_v1.py

@@ -176,6 +176,8 @@ class AscendMetadata:
     causal: bool = True
     # runner_type in model_config.
     model_runner_type: str = ""
+    # prefill reshape_and_cache event
+    reshape_cache_event: torch.npu.Event = None
 
     # sliding window attention mask
     swa_mask: Optional[torch.Tensor] = None
@@ -333,6 +335,7 @@ class AscendAttentionBackendImpl(AttentionImpl):
         self.num_queries_per_kv = self.num_heads // self.num_kv_heads
         self.key_cache = None
         self.value_cache = None
+        self.is_kv_producer = self.vllm_config.kv_transfer_config is not None and self.vllm_config.kv_transfer_config.is_kv_producer
 
     def full_graph_fia(self, query: torch.Tensor, key: torch.Tensor,
                        value: torch.Tensor, attn_metadata: AscendMetadata,
@@ -654,6 +657,8 @@ class AscendAttentionBackendImpl(AttentionImpl):
     ):
 
         if len(kv_cache) > 1:
+            if self.is_kv_producer:
+                attn_metadata.reshape_cache_event = torch.npu.Event()
             if self.key_cache is None:
                 self.key_cache, self.value_cache = kv_cache[0], kv_cache[1]
             slots = attn_metadata.slot_mapping
@@ -674,6 +679,8 @@ class AscendAttentionBackendImpl(AttentionImpl):
                     key_cache=self.key_cache,
                     value_cache=self.value_cache,
                     slot_indices=slots[:attn_metadata.num_actual_tokens])
+            if self.is_kv_producer:
+                attn_metadata.reshape_cache_event.record()
         return key, value
 
     def forward_impl(