[Bug] Batch invariant: Fix flash attn MLA RuntimeError: scheduler_metadata must have shape (metadata_size) (vllm-project#27884)

Signed-off-by: xuebwang-amd <[email protected]>

Author: yewentao256

vllm/model_executor/layers/batch_invariant.py

@@ -1,6 +1,7 @@
 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 import contextlib
+import functools
 import os
 from collections import namedtuple
 from collections.abc import Callable
@@ -846,6 +847,7 @@ def get_batch_invariant_attention_block_size() -> AttentionBlockSize:
     return AttentionBlockSize(block_m=16, block_n=16)
 
 
+@functools.cache
 def vllm_is_batch_invariant():
     env_key = "VLLM_BATCH_INVARIANT"
     is_overridden = False

vllm/v1/attention/backends/mla/flashattn_mla.py

@@ -163,6 +163,9 @@ def _build_decode(
             # we only set num_splits when using cuda graphs.
             max_num_splits = self.max_num_splits
 
+        if vllm_is_batch_invariant():
+            max_num_splits = 1
+
         scheduler_metadata = self._schedule_decode(
             num_reqs=seq_lens_cpu.numel(),
             cu_query_lens=query_start_loc_device,
@@ -188,9 +191,6 @@ def _build_decode(
             self.scheduler_metadata[n:] = 0
             scheduler_metadata = self.scheduler_metadata[:n]
 
-        if vllm_is_batch_invariant():
-            max_num_splits = 1
-
         metadata = FlashAttnMLADecodeMetadata(
             block_table=block_table_tensor,
             seq_lens=seq_lens_device,