[llvm] clastb representation in existing IR, and AArch64 codegen (PR #112738)

[Graham Hunter via llvm-commits]

huntergr-arm wrote:

> Bad TargetLowering API with no use context. SelectionDAGBuilder should directly translate into dag nodes and not be a point of configuration

I was following a pattern of existing TLI functions (starting with `shouldExpandGetActiveLaneMask` 3 years ago). If that's the wrong way to do it, then I guess we have some cleanup to do later in addition to removing the function from this patch.

I'll investigate leaving it as an intrinsic and moving the current code in SelectionDAGBuilder to an expand legalization function instead.

https://github.com/llvm/llvm-project/pull/112738