fix deepseek mla

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>

Author: wangxiyuan

vllm_ascend/models/layers/mla.py

@@ -132,8 +132,13 @@ def forward(
         output = torch.empty(output_shape,
                              dtype=hidden_states.dtype,
                              device=hidden_states.device)
+        if forward_context.attn_metadata:
+            attn_metadata = forward_context.attn_metadata[
+                self.mla_attn.layer_name]
+        else:
+            attn_metadata = forward_context.attn_metadata
         output = self.mla_attn.impl.forward(hidden_states, kv_cache,
-                                            forward_context.attn_metadata,
-                                            need_gather_q_kv, output)
+                                            attn_metadata, need_gather_q_kv,
+                                            output)
         output = output.view(-1, output_shape[-1])
         return output