[llvm] [WebAssembly] Use the same lowerings for f16x8 as other float vectors. (PR #127897)

[Brendan Dahl via llvm-commits]

brendandahl wrote:

Oh yeah, for the v8f16 we don't need to expand those. However, I'm missing the promote and truncate instructions, so these still ended up getting expanded out. I'll add those instructions in a different PR.

https://github.com/llvm/llvm-project/pull/127897