[llvm] [X86] Truncate i64 add/sub/mul arithmetic to i32 with known zeros in higher bits (PR #143313)

[Simon Pilgrim via llvm-commits]

https://github.com/RKSimon requested changes to this pull request.

You're going to have much better test coverage for all of these - I suggest you split these into 3 PRs (add/sub/mul) - and we focus on the add first

https://github.com/llvm/llvm-project/pull/143313