[MLIR][GPU] Disallow llvm tanh intrinsics when lowering to NVVM/ROCm.
The lowering to NVVM and ROCm handles tanh operations differently by
mapping them to NVVM/ROCm specific intrinsics. This conflicts with
the lowering to LLVM, which uses the default llvm intrinsic. This change
declares the LLVM intrinsics to be illegal, hence disallowing the
Differential Revision: https://reviews.llvm.org/D74389