[llvm] [AArch64] Optimize vector multiplications by certain constants for v2i64 (PR #183827)

[Aly ElAshram via llvm-commits]

AlyElashram wrote:

@davemgreen I moved the optimization to be placed in lowerMul to avoid unnecessary checks , I think it's a bit cleaner and it still applies to any vector that could benefit from the optimization if there are no legal MUL instructions. This also avoids the need for the sext/zext check in the performMulCombine , and the tests as well. Let me know what you think 👀 

https://github.com/llvm/llvm-project/pull/183827