[llvm] [LV] Split checking if tail-folding is possible, collecting masked ops. (PR #77612)

[Florian Hahn via llvm-commits]

================
@@ -1601,8 +1601,24 @@ bool LoopVectorizationLegality::prepareToFoldTailByMasking() {
 
   LLVM_DEBUG(dbgs() << "LV: can fold tail by masking.\n");
 
-  MaskedOp.insert(TmpMaskedOp.begin(), TmpMaskedOp.end());
   return true;
 }
 
+void LoopVectorizationLegality::prepareToFoldTailByMasking() {
+  // The list of pointers that we can safely read and write to remains empty.
+  SmallPtrSet<Value *, 8> SafePointers;
+
+  // Collect masked ops in temporary set first to avoid partially populating
+  // MaskedOp if a block cannot be predicated.
+  SmallPtrSet<const Instruction *, 8> TmpMaskedOp;
+
+  // Check and mark all blocks for predication, including those that ordinarily
+  // do not need predication such as the header block.
+  for (BasicBlock *BB : TheLoop->blocks()) {
+    bool R = blockCanBePredicated(BB, SafePointers, MaskedOp);
----------------
fhahn wrote:

Left as is for now for this PR to keep the number of changes small, but will check as follow-up. More generally, we should probably collect and consider safe pointers with tail folding as well

https://github.com/llvm/llvm-project/pull/77612