[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>

tests/ut/attention/test_mla_v1.py

@@ -1112,6 +1112,7 @@ class TestAscendMLAImpl(TestBase):
             MagicMock(), MagicMock()
         ]
         self.impl.num_kv_heads = self.impl.num_heads
+        self.impl.is_kv_producer = False
 
         decode_res, prefill_res = self.impl._mla_preprocess(
             "mock_layer",