[llvm] [X86] combineConcatVectorOps - recursively call combineConcatVectorOps instead of predicting when ops will freely concat (PR #130275)

[Simon Pilgrim via llvm-commits]

RKSimon wrote:

@phoebewang  I've restricted this to FADD/FSUB/FMUL instructions for the initial patch (ADD/SUB/MUL will still need test coverage in a later PR).

https://github.com/llvm/llvm-project/pull/130275