[llvm] [AArch64] Use vecshiftL64 instead of vecshiftR64 to match scalar SLI imm. (PR #139904)

[David Green via llvm-commits]

davemgreen wrote:

It can certainly be separate, this makes sense on its own. It is not necessary either way - it is a very minor missed optimization.

https://github.com/llvm/llvm-project/pull/139904