[llvm] [RISCV][VLOPT] Add support for checkUsers when UserMI is a Single-Width Integer Reduction (PR #120345)

[Philip Reames via llvm-commits]

https://github.com/preames requested changes to this pull request.

This is (still) not correct.

I think you're getting yourself confused here, and need to take a step back and revisit.  This is (deceptively) simple, and you seem to be tripping up on that.  I'd suggest reorganizing this bit of code in terms of an accessor which gets the VL for a given use on an instruction, and then sharing all of the merge/validation logic between this codepath and the existing one.  Stop trying to make the current early continue idiom work.  It doesn't.  

https://github.com/llvm/llvm-project/pull/120345