[llvm] [RISCV] Fold vector shift of sext/zext to widening multiply (PR #121563)

[Philip Reames via llvm-commits]

preames wrote:

> I've heard that this transform won't be profitable on other CPUs, so I added a commit that enables it on BPI's SpacemiT X60 only.

For my micro-architecture, this should be at most a very minor regression, and possible a slight improvement.  From our perspective, I'm 100% fine with this being default enabled without a tuning flag.  

https://github.com/llvm/llvm-project/pull/121563