[llvm] [VPlan] Don't use the legacy cost model for loop conditions (PR #156864)

[Florian Hahn via llvm-commits]

https://github.com/fhahn commented:

Likely related @ElvisWang123 's https://github.com/llvm/llvm-project/pull/125640. IIRC the latest status was that there still were some regressions needing investigation

https://github.com/llvm/llvm-project/pull/156864