diff --git a/tests/pytorch/fused_attn/test_fused_attn.py b/tests/pytorch/fused_attn/test_fused_attn.py
index 43f1d4488..16791266e 100644
--- a/tests/pytorch/fused_attn/test_fused_attn.py
+++ b/tests/pytorch/fused_attn/test_fused_attn.py
@@ -219,8 +219,8 @@ def test():
         )
         (
             use_flash_attention,
-            use_fused_attention,
             flash_attention_backend,
+            use_fused_attention,
             fused_attention_backend,
             use_unfused_attention,
             available_backends,
@@ -368,6 +368,7 @@ def test_dot_product_attention(
             and config.attn_mask_type in ["causal", "padding_causal"]
         )
         and (config.window_size[0] == -1 or FlashAttentionUtils.v2_3_plus)
+        and not is_mla
     ):
         flash_attn_supported = True