Fix the condition error when checking fp8 attn in `get_attention_backend` (#1965)

Update utils.py

Fix the condition error of the FP8 attention in `get_attention_backend`
Signed-off-by: yuzhongw-nvidia <yuzhongw@nvidia.com>
Co-authored-by: Xiaowei Ren <103958965+xrennvidia@users.noreply.github.com>

transformer_engine/pytorch/attention/dot_product_attention/utils.py

@@ -609,7 +609,7 @@ def get_attention_backend(
                 " bias for THD format"
             )
             use_fused_attention = False
-        elif fp8 and head_dim_qk != head_dim_v:
+        elif fp8 and fp8_meta["recipe"].fp8_dpa and head_dim_qk != head_dim_v:
             logger.debug(
                 "Disabling FusedAttention as it does not support context parallelism with FP8"
                 " MLA attention"