[llvm] [VPlan] Remove loop region in optimizeForVFAndUF. (PR #108378)

[via llvm-commits]

================
@@ -676,6 +678,8 @@ class InnerLoopVectorizer {
   BasicBlock *AdditionalBypassBlock = nullptr;
 
   VPlan &Plan;
+
+  VPBasicBlock *VectorPHVPBB;
----------------
ayalz wrote:

Follow-up: when skeleton creation moves to VPlan creation, ILV should not depend on VPlan or its VPBB's, as was originally the case.

https://github.com/llvm/llvm-project/pull/108378