[llvm] [AArch64][Machine-Combiner] Split loads into lanes of neon vectors into multiple vectors when possible (PR #142941)

[Jonathan Cohen via llvm-commits]

jcohen-apple wrote:

Added an early exit for minsize, good catch.
Regarding supporting other ZIP types - technically you can, I couldn't find a case where it was actually profitable. I could address in a follow-up PR if you'd like.

https://github.com/llvm/llvm-project/pull/142941