[2/N][Refactor] Refactor V1 attention for better extensibility (#1995)

### What this PR does / why we need it?

Refactor V1 Attention for better extensibility (prepared for torchair
attention refactor).

**Main changes:**
- Move different kinds of foward into their method respectively, e.g.,
`_forward_prefill_no_cache()`, `_forward_prefill_cache_hit()`,
`_forward_decode_only()`, `_forward_v1_style()`.

### Does this PR introduce _any_ user-facing change?

No.

- vLLM version: v0.10.0
- vLLM main:
14a5d903ab

Signed-off-by: shen-shanshan <467638484@qq.com>

vllm_ascend/attention/attention_v1.py

@@ -120,7 +120,7 @@ class AscendAttentionState(Enum):
 @dataclass
 class AscendMetadata:
 
-    # **************************** Basic Properties ****************************
+    # **************************** Basic Properties ************************** #
     attn_mask: Optional[torch.Tensor] = None
     # Current state of this attention run.
     attn_state: AscendAttentionState = AscendAttentionState.ChunkedPrefill
@@ -138,7 +138,7 @@ class AscendMetadata:
     # Maximum query length in the batch (None for decoding).
     max_query_len: Optional[int] = None
 
-    # ********************** KV Cache Related Properties ***********************
+    # ********************** KV Cache Related Properties ********************* #
     # Block addresses per sequence (Seq id -> list of physical block).
     # (batch_size, max_blocks_per_seq)
     block_tables: torch.Tensor = None
@@ -150,6 +150,7 @@ class AscendMetadata:
     # (num_tokens,)
     slot_mapping: torch.Tensor = None
 
+    # *************************** Other Properties *************************** #
     enable_dbo_across_dp: bool = False
     is_only_prefill: bool = False
 
@@ -245,6 +246,144 @@ class AscendAttentionBackendImpl(AttentionImpl):
         self.key_cache = None
         self.value_cache = None
 
+    def _forward_prefill_no_cache(
+        self,
+        query: torch.Tensor,
+        key: torch.Tensor,
+        value: torch.Tensor,
+        attn_metadata: AscendMetadata,
+        output: Optional[torch.Tensor] = None,
+        num_tokens=0,
+    ) -> torch.Tensor:
+        assert attn_metadata is not None
+        assert attn_metadata.attn_mask is not None
+
+        mask = attn_metadata.attn_mask
+
+        if is_310p():
+            # align q k v output tensors
+            query = aligned_16(query)
+            key = aligned_16(key)
+            value = aligned_16(value)
+            output = aligned_16(output)
+            # do reformat in case of broadcasted tensors
+            mask = mask.repeat(attn_metadata.seq_lens.size(0), 1, 1, 1)
+            mask = torch_npu.npu_format_cast(mask.contiguous(),
+                                             ACL_FORMAT_FRACTAL_NZ)
+
+        torch_npu._npu_flash_attention(query=query,
+                                       key=key,
+                                       value=value,
+                                       mask=mask,
+                                       seq_len=attn_metadata.seq_lens,
+                                       scale_value=self.scale,
+                                       num_heads=self.num_heads,
+                                       num_kv_heads=self.num_kv_heads,
+                                       out=output)
+        assert output is not None
+        return output[:num_tokens, :, :]
+
+    def _forward_prefill_cache_hit(
+        self,
+        query: torch.Tensor,
+        attn_metadata: AscendMetadata,
+        output: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        assert attn_metadata is not None
+        assert attn_metadata.attn_mask is not None
+
+        compress_mask = attn_metadata.attn_mask
+        batch_size = attn_metadata.query_lens.shape[0]
+        block_table = attn_metadata.block_tables[:batch_size, :]
+
+        torch_npu._npu_flash_attention_qlens(
+            query=query,
+            key_cache=self.key_cache,
+            value_cache=self.value_cache,
+            block_table=block_table,
+            mask=compress_mask,
+            seq_len=attn_metadata.query_lens,
+            context_lens=attn_metadata.seq_lens,
+            num_kv_heads=self.num_kv_heads,
+            num_heads=self.num_heads,
+            scale_value=self.scale,
+            out=output)
+        return output
+
+    def _forward_decode_only(
+        self,
+        query: torch.Tensor,
+        attn_metadata: AscendMetadata,
+        output: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        if is_310p():
+            # seq_lens_tensor needs to be transferred to the device for 310P.
+            attn_metadata.seq_lens = \
+                attn_metadata.seq_lens.to(device=query.device)
+
+        torch_npu._npu_paged_attention(query=query,
+                                       key_cache=self.key_cache,
+                                       value_cache=self.value_cache,
+                                       num_kv_heads=self.num_kv_heads,
+                                       num_heads=self.num_heads,
+                                       scale_value=self.scale,
+                                       block_table=attn_metadata.block_tables,
+                                       context_lens=attn_metadata.seq_lens,
+                                       out=output)
+        return output
+
+    def _forward_v1_style(
+        self,
+        query: torch.Tensor,
+        attn_metadata: AscendMetadata,
+        output: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        # Use chunked prefill for head size 192 scenario, like deepseek
+        # paged_attention_splitfuse maybe crash at such scenario.
+        # TODO: vanilla path will be removed after the kernel support
+        # head_size 192 scenario.
+        if self.head_size == 192:
+            cu_seqlen_q = [0] + attn_metadata.query_lens.tolist()
+            cu_seqlen_k = [0] + attn_metadata.seq_lens.tolist()
+            cu_seqlen_q = torch.tensor(cu_seqlen_q, device=query.device)
+            cu_seqlen_k = torch.tensor(cu_seqlen_k, device=query.device)
+            cu_seqlen_q = torch.cumsum(cu_seqlen_q, dim=0)
+            cu_seqlen_k = torch.cumsum(cu_seqlen_k, dim=0)
+            max_seqlen_q = torch.max(attn_metadata.query_lens)
+            max_seqlen_k = torch.max(attn_metadata.seq_lens)
+            vanilla_chunked_prefill(output, query, self.key_cache,
+                                    self.value_cache,
+                                    attn_metadata.block_tables, cu_seqlen_q,
+                                    cu_seqlen_k, max_seqlen_q, max_seqlen_k,
+                                    self.scale, None, True)
+            return output
+
+        # Use paged attention.
+        assert attn_metadata is not None
+        assert attn_metadata.attn_mask is not None
+
+        if is_310p():
+            # Do reformat in case of broadcasted tensors.
+            attn_metadata.attn_mask = \
+                torch_npu.npu_format_cast(attn_metadata.attn_mask.contiguous(),
+                                          ACL_FORMAT_FRACTAL_NZ)
+            attn_metadata.seq_lens = \
+                attn_metadata.seq_lens.to(device=query.device)
+
+        torch_npu._npu_paged_attention_splitfuse(
+            query=query,
+            key_cache=self.key_cache,
+            value_cache=self.value_cache,
+            mask=attn_metadata.attn_mask,
+            block_table=attn_metadata.block_tables,
+            seq_len=attn_metadata.query_lens,
+            context_lens=attn_metadata.seq_lens,
+            num_kv_heads=self.num_kv_heads,
+            num_heads=self.num_heads,
+            scale_value=self.scale,
+            out=output)
+        return output
+
     def forward(
         self,
         layer: AttentionLayer,
@@ -325,109 +464,18 @@ class AscendAttentionBackendImpl(AttentionImpl):
 
             # V0-Style scheduler situation.
             if attn_metadata.attn_state == AscendAttentionState.PrefillNoCache:
-                assert attn_metadata is not None
+                output = self._forward_prefill_no_cache(
-                assert attn_metadata.attn_mask is not None
+                    query, key, value, attn_metadata, output, num_tokens)
-                mask = attn_metadata.attn_mask
+            elif attn_metadata.attn_state == \
-                if is_310p():
+                AscendAttentionState.PrefillCacheHit:
-                    # align q k v output tensors
+                output = self._forward_prefill_cache_hit(
-                    query = aligned_16(query)
+                    query, attn_metadata, output)
-                    key = aligned_16(key)
-                    value = aligned_16(value)
-                    output = aligned_16(output)
-
-                    # do reformat in case of broadcasted tensors
-                    mask = mask.repeat(attn_metadata.seq_lens.size(0), 1, 1, 1)
-                    mask = torch_npu.npu_format_cast(mask.contiguous(),
-                                                     ACL_FORMAT_FRACTAL_NZ)
-
-                torch_npu._npu_flash_attention(query=query,
-                                               key=key,
-                                               value=value,
-                                               mask=mask,
-                                               seq_len=attn_metadata.seq_lens,
-                                               scale_value=self.scale,
-                                               num_heads=self.num_heads,
-                                               num_kv_heads=self.num_kv_heads,
-                                               out=output)
-                output = output[:num_tokens, :, :]
-            elif attn_metadata.attn_state == AscendAttentionState.PrefillCacheHit:
-                assert attn_metadata is not None
-                assert attn_metadata.attn_mask is not None
-                compress_mask = attn_metadata.attn_mask
-                batch_size = attn_metadata.query_lens.shape[0]
-                block_table = attn_metadata.block_tables[:batch_size, :]
-                torch_npu._npu_flash_attention_qlens(
-                    query=query,
-                    key_cache=self.key_cache,
-                    value_cache=self.value_cache,
-                    block_table=block_table,
-                    mask=compress_mask,
-                    seq_len=attn_metadata.query_lens,
-                    context_lens=attn_metadata.seq_lens,
-                    num_kv_heads=self.num_kv_heads,
-                    num_heads=self.num_heads,
-                    scale_value=self.scale,
-                    out=output)
             elif attn_metadata.attn_state == AscendAttentionState.DecodeOnly:
-                if is_310p():
+                output = self._forward_decode_only(query, attn_metadata,
-                    # # seq_lens_tensor needs to be transferred to the device for 310P
+                                                   output)
-                    attn_metadata.seq_lens = \
-                        attn_metadata.seq_lens.to(device=query.device)
-                torch_npu._npu_paged_attention(
-                    query=query,
-                    key_cache=self.key_cache,
-                    value_cache=self.value_cache,
-                    num_kv_heads=self.num_kv_heads,
-                    num_heads=self.num_heads,
-                    scale_value=self.scale,
-                    block_table=attn_metadata.block_tables,
-                    context_lens=attn_metadata.seq_lens,
-                    out=output)
             # Normal V1 situation.
             else:
-                # use chunked prefill for head size 192 scenario, like deepseek
+                output = self._forward_v1_style(query, attn_metadata, output)
-                # paged_attention_splitfuse maybe crash at such scenario
-                # TODO: vanilla path will be removed after the kernel support
-                # head_size 192 scenario
-                if self.head_size == 192:
-                    cu_seqlen_q = [0] + attn_metadata.query_lens.tolist()
-                    cu_seqlen_k = [0] + attn_metadata.seq_lens.tolist()
-                    cu_seqlen_q = torch.tensor(cu_seqlen_q,
-                                               device=query.device)
-                    cu_seqlen_k = torch.tensor(cu_seqlen_k,
-                                               device=query.device)
-                    cu_seqlen_q = torch.cumsum(cu_seqlen_q, dim=0)
-                    cu_seqlen_k = torch.cumsum(cu_seqlen_k, dim=0)
-                    max_seqlen_q = torch.max(attn_metadata.query_lens)
-                    max_seqlen_k = torch.max(attn_metadata.seq_lens)
-                    vanilla_chunked_prefill(output, query, self.key_cache,
-                                            self.value_cache,
-                                            attn_metadata.block_tables,
-                                            cu_seqlen_q, cu_seqlen_k,
-                                            max_seqlen_q, max_seqlen_k,
-                                            self.scale, None, True)
-                else:
-                    # use paged attention
-                    assert attn_metadata is not None
-                    assert attn_metadata.attn_mask is not None
-                    if is_310p():
-                        # do reformat in case of broadcasted tensors
-                        attn_metadata.attn_mask = \
-                            torch_npu.npu_format_cast(attn_metadata.attn_mask.contiguous(), ACL_FORMAT_FRACTAL_NZ)
-                        attn_metadata.seq_lens = \
-                            attn_metadata.seq_lens.to(device=query.device)
-                    torch_npu._npu_paged_attention_splitfuse(
-                        query=query,
-                        key_cache=self.key_cache,
-                        value_cache=self.value_cache,
-                        mask=attn_metadata.attn_mask,
-                        block_table=attn_metadata.block_tables,
-                        seq_len=attn_metadata.query_lens,
-                        context_lens=attn_metadata.seq_lens,
-                        num_kv_heads=self.num_kv_heads,
-                        num_heads=self.num_heads,
-                        scale_value=self.scale,
-                        out=output)
 
         # to make in-place change to the output tensor
         if hasattr(layer, 'quant_method') and use_kv_cache_int8: