add attention docs

Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

Author: sudhakarsingh27

docs/envvars.rst

@@ -143,8 +143,8 @@ Attention Backend Selection
 .. envvar:: NVTE_FUSED_ATTN_BACKEND
 
    :Type: ``int`` (1 or 2)
-   :Default: Auto-selected
-   :Description: Force a specific FusedAttention backend. ``1`` = F16_arbitrary_seqlen (cuDNN, any seq len), ``2`` = FP8 backend. If not set, the backend is automatically selected based on the input configuration.
+   :Default: Ignored
+   :Description: Historical option for forcing a FusedAttention sub-backend. The current PyTorch DotProductAttention selector does not read this variable; the FusedAttention sub-backend is selected automatically from the input configuration and cuDNN support.
 
 .. envvar:: NVTE_FUSED_ATTN_FORCE_WORKSPACE_OPT