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

[Florian Hahn via llvm-commits]

================
@@ -663,14 +670,16 @@ void VPBasicBlock::print(raw_ostream &O, const Twine &Indent,
 }
 #endif
 
-static std::pair<VPBlockBase *, VPBlockBase *> cloneFrom(VPBlockBase *Entry);
+static std::tuple<VPBlockBase *, VPBlockBase *, VPIRBasicBlock *>
+cloneFrom(VPBlockBase *Entry, VPIRBasicBlock *ScalarHeader = nullptr);
 
 // Clone the CFG for all nodes reachable from \p Entry, this includes cloning
 // the blocks and their recipes. Operands of cloned recipes will NOT be updated.
 // Remapping of operands must be done separately. Returns a pair with the new
 // entry and exiting blocks of the cloned region. If \p Entry isn't part of a
 // region, return nullptr for the exiting block.
-static std::pair<VPBlockBase *, VPBlockBase *> cloneFrom(VPBlockBase *Entry) {
+static std::tuple<VPBlockBase *, VPBlockBase *, VPIRBasicBlock *>
+cloneFrom(VPBlockBase *Entry, VPIRBasicBlock *ScalarHeader) {
----------------
fhahn wrote:

Updated to retrieve the cloned scalar header by traversing the cloned CFG to look for the correct VPIRBB. I think it would be difficult to identify the correct block directly in cloneFrom.

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