[Fix] Refactor dummy attention metadata creation (#3497)

### What this PR does / why we need it?
The `force_attention` parameter is designed for flash infer kernel
warmup, we don't actually need it on Ascend device (at least for
now).And it tends to make things more complicated. So we replace the
`force_attention` parameter with `aclgraph_runtime_mode` in the
attention metadata creation logic.

This change makes the control flow more explicit by directly using the
graph runtime mode to determine how to build attention metadata, rather
than relying on an intermediate boolean flag. This simplification
removes redundant logic and clarifies the conditions for building
attention metadata for full decode graph mode.

### Does this PR introduce _any_ user-facing change?
None.

### How was this patch tested?
DP + `FULL_DECODE_ONLY` + online serving.

- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

---------

Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>

tests/ut/worker/test_worker_v1.py

@@ -456,9 +456,7 @@ class TestNPUWorker(TestBase):
 
             # Verify call
             mock_model_runner._dummy_run.assert_called_once_with(
-                num_tokens=mock_decode_token_per_req,
-                uniform_decode=True,
-                force_attention=False)
+                num_tokens=mock_decode_token_per_req, uniform_decode=True)
 
     @patch("vllm_ascend.worker.worker_v1.envs_vllm")
     @patch("vllm_ascend.worker.worker_v1.logger")

vllm_ascend/torchair/torchair_model_runner.py

@@ -19,13 +19,13 @@
 
 import math
 import types
-from typing import Optional
+from typing import Any, Optional
 
 import torch
 import torch.distributed as dist
 import torch.nn as nn
 import torch_npu
-from vllm.config import VllmConfig
+from vllm.config import CUDAGraphMode, VllmConfig
 from vllm.distributed import get_tensor_model_parallel_world_size
 from vllm.distributed.parallel_state import get_dp_group
 from vllm.forward_context import get_forward_context
@@ -147,14 +147,21 @@ class NPUTorchairModelRunner(NPUModelRunner):
 
         return maybe_padded_num_tokens, num_tokens_across_dp, with_prefill, enable_dbo
 
-    def _build_attention_metadata(self, with_prefill, num_reqs, num_tokens,
-                                  max_query_len, force_attention):
+    def _build_dummy_attn_metadata(
+        self,
+        with_prefill: bool,
+        num_reqs: int,
+        num_tokens: int,
+        max_query_len: int,
+        aclgraph_runtime_mode: Optional[CUDAGraphMode] = None,
+        force_attention: bool = False,
+    ) -> Optional[dict[str, Any]]:
         # NOTE: If torchair graph mode and not with_prefill,
         # we can't skip_attn, it will cause graph recompile.
         if with_prefill or self.enable_shared_expert_dp:
-            attn_metadata = super()._build_attention_metadata(
+            attn_metadata = super()._build_dummy_attn_metadata(
                 with_prefill, num_reqs, num_tokens, max_query_len,
-                force_attention)
+                aclgraph_runtime_mode, force_attention)
         else:
             common_attn_metadata = TorchairCommonAttentionMetadata(
                 num_reqs=num_reqs,

vllm_ascend/worker/model_runner_v1.py

@@ -2250,18 +2250,24 @@ class NPUModelRunner(LoRAModelRunnerMixin):
                 scheduler_output.finished_req_ids)
         return None, None
 
-    def _build_attention_metadata(self, create_mixed_batch, num_reqs,
-                                  num_tokens, max_query_len, force_attention):
+    def _build_dummy_attn_metadata(
+        self,
+        with_prefill: bool,
+        num_reqs: int,
+        num_tokens: int,
+        max_query_len: int,
+        aclgraph_runtime_mode: Optional[CUDAGraphMode] = None,
+        force_attention: bool = False,
+    ) -> Optional[dict[str, Any]]:
         attn_metadata: Optional[dict[str, Any]] = None
 
-        if force_attention:
+        if force_attention or aclgraph_runtime_mode == CUDAGraphMode.FULL:
+            assert with_prefill is False, \
+                "Full decode graph only supports uniform batch now."
+
             attn_metadata = {}
 
-            if create_mixed_batch:
-                raise NotImplementedError(
-                    "force_attention=True is not supported for mixed batches.")
-            else:
-                seq_lens = self.model_config.max_model_len
+            seq_lens = self.model_config.max_model_len
             self.seq_lens_np[:num_reqs] = seq_lens
             self.seq_lens_np[num_reqs:] = 0
 
@@ -2321,7 +2327,7 @@ class NPUModelRunner(LoRAModelRunnerMixin):
         forward_context = get_forward_context()
         assert forward_context is not None
         if forward_context.cudagraph_runtime_mode == CUDAGraphMode.FULL and \
-            not forward_context.capturing:
+            not forward_context.capturing and forward_context.attn_metadata is not None:
             if self.vllm_config.model_config.use_mla:
                 # FIXME: Try using `auto_dispatch_capture=True`
                 update_mla_attn_params(self.update_stream, forward_context,
@@ -2409,12 +2415,14 @@ class NPUModelRunner(LoRAModelRunnerMixin):
         if self.is_kv_producer and not self.is_kv_consumer:
             with_prefill = True
 
-        # TODO(cmq): check if with_prefill is reasonable
-        attn_metadata = self._build_attention_metadata(
+        # TODO(Mengqing): Set create_mixed_batch to False since it's only used in FI warmup
+        # and not supported in ASCEND now. We could remove it in the future.
+        attn_metadata = self._build_dummy_attn_metadata(
             False,
             num_reqs=num_reqs,
             num_tokens=num_tokens,
             max_query_len=max_query_len,
+            aclgraph_runtime_mode=aclgraph_runtime_mode,
             force_attention=force_attention,
         )
 

vllm_ascend/worker/worker_v1.py

@@ -26,7 +26,7 @@ import torch_npu
 import vllm.envs as envs_vllm
 from torch_npu.op_plugin.atb._atb_ops import _register_atb_extensions
 from torch_npu.profiler import dynamic_profile as dp
-from vllm.config import CUDAGraphMode, VllmConfig
+from vllm.config import VllmConfig
 from vllm.distributed import (ensure_model_parallel_initialized,
                               init_distributed_environment)
 from vllm.distributed.kv_transfer import ensure_kv_transfer_initialized
@@ -360,11 +360,9 @@ class NPUWorker(WorkerBase):
         return self.model_runner.pin_lora(lora_id)
 
     def execute_dummy_batch(self) -> None:
-        force_attention = self.compilation_config.cudagraph_mode == CUDAGraphMode.FULL_DECODE_ONLY
         self.model_runner._dummy_run(
             num_tokens=self.model_runner.decode_token_per_req,
-            uniform_decode=True,
-            force_attention=force_attention)
+            uniform_decode=True)
 
     def _init_worker_distributed_environment(self) -> None:
         """Initialize the distributed environment."""