[DispatchCreation] Don't pad on attention in producer dispatch

[jtuyls]

Towards: https://github.com/iree-org/iree/issues/20835, needed to get e2e Llama3 with padding to compile.

This PR avoids inserting padding encodings if the producer dispatch region contains an attention operation as that results in complicated multi-dimensional load/store access patterns which aren't well supported with padding in codegen.