[llvm] [VPlan] Introduce scalar loop header in plan, remove VPLiveOut. (PR #109975)

[via llvm-commits]

================
@@ -376,10 +376,10 @@ static bool mergeBlocksIntoPredecessors(VPlan &Plan) {
   SmallVector<VPBasicBlock *> WorkList;
   for (VPBasicBlock *VPBB : VPBlockUtils::blocksOnly<VPBasicBlock>(
            vp_depth_first_deep(Plan.getEntry()))) {
-    // Don't fold the exit block of the Plan into its single predecessor for
-    // now.
+    // Don't fold the blocks in the skeleton of the Plan into its single
----------------
ayalz wrote:

```suggestion
    // Don't fold the blocks in the skeleton of the Plan into their single
```

https://github.com/llvm/llvm-project/pull/109975