[BugFix][MLA] Fix attn_mask bug for ring mla (#2704)

This PR fix a bug related to attention mask used in ring mla. Current
ring mla has supported compressed mask, so we can directly use a 512 *
512 attention mask.

- vLLM version: v0.10.1.1
- vLLM main:
b5ee1e3261

---------

Signed-off-by: whx-sjtu <2952154980@qq.com>

tests/ut/attention/test_mla_v1.py

@@ -238,13 +238,19 @@ class TestAscendMLAImpl(TestBase):
            new_callable=lambda: MagicMock(spec=GroupCoordinator))
     @patch("vllm.distributed.get_tensor_model_parallel_world_size",
            return_value=2)
-    @patch("vllm.config.get_current_vllm_config")
+    @patch("vllm_ascend.attention.mla_v1.get_current_vllm_config")
     @patch("vllm_ascend.attention.mla_v1.get_ascend_config")
-    def setUp(self, ascend_config, vllm_config, mock_get_tp_size, mock_tp):
+    def setUp(self, ascend_config, get_current_vllm_config, mock_get_tp_size,
+              mock_tp):
         mock_tp.world_size = 2
+        vllm_config = MagicMock()
         speculative_config = MagicMock()
+        model_config = MagicMock()
         speculative_config.num_speculative_tokens = 4
         vllm_config.speculative_config = speculative_config
+        model_config.dtype = torch.float16
+        vllm_config.model_config = model_config
+        get_current_vllm_config.return_value = vllm_config
 
         num_heads = 256
         head_size = 1024
@@ -622,4 +628,4 @@ class TestAscendMLAImpl(TestBase):
 
         self.assertEqual(result.shape[0], B)
         self.assertEqual(result.shape[1], N)
-        self.assertEqual(result.shape[2], HD)
+        self.assertEqual(result.shape[2], HD)

vllm_ascend/attention/mla_v1.py

@@ -483,10 +483,12 @@ class AscendMLAImpl(MLAAttentionImpl):
         self.enable_kv_nz = ascend_config.torchair_graph_config.enable_kv_nz
         self.chunked_prefill_for_mla = ascend_config.chunked_prefill_for_mla
 
+        vllm_config = get_current_vllm_config()
+        self.ring_mla_mask_size = 512
         self.prefill_mask = None
 
         # Adapt torch air graph mode with spec decoding.
-        speculative_config = get_current_vllm_config().speculative_config
+        speculative_config = vllm_config.speculative_config
         if speculative_config is not None:
             self.spec_token_num = speculative_config.num_speculative_tokens
             assert self.spec_token_num > 0
@@ -681,14 +683,10 @@ class AscendMLAImpl(MLAAttentionImpl):
                                    device=q_nope.device)
             if self.prefill_mask is None:
                 self.prefill_mask = torch.triu(
-                    torch.ones(512,
-                               512,
+                    torch.ones(self.ring_mla_mask_size,
+                               self.ring_mla_mask_size,
                                device=q_nope.device,
-                               dtype=q_nope.dtype),
-                    1)  # 512: mask only support 512
-            if attn_metadata.num_prefills > 1:
-                self.prefill_mask = self.prefill_mask.unsqueeze(0).repeat(
-                    attn_metadata.num_prefills, 1, 1)
+                               dtype=q_nope.dtype), 1)
             torch_npu.atb.npu_ring_mla(
                 q_nope=q_nope,
                 q_rope=q_pe,