[LV] Fix scalar cost for tail predicated loops

When it comes to the scalar cost of any predicated block, the loop
vectorizer by default regards this predication as a sign that it is
looking at an if-conversion and divides the scalar cost of the block by
2, assuming it would only be executed half the time. This however makes
no sense if the predication has been introduced to tail predicate the
loop.

Original patch by Anna Welker

Differential Revision: https://reviews.llvm.org/D86452

Author: davemgreen

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

@@ -6483,9 +6483,10 @@ LoopVectorizationCostModel::expectedCost(ElementCount VF) {
     // if-converted. This means that the block's instructions (aside from
     // stores and instructions that may divide by zero) will now be
     // unconditionally executed. For the scalar case, we may not always execute
-    // the predicated block. Thus, scale the block's cost by the probability of
-    // executing it.
-    if (VF.isScalar() && blockNeedsPredication(BB))
+    // the predicated block, if it is an if-else block. Thus, scale the block's
+    // cost by the probability of executing it. blockNeedsPredication from
+    // Legal is used so as to not include all blocks in tail folded loops.
+    if (VF.isScalar() && Legal->blockNeedsPredication(BB))
       BlockCost.first /= getReciprocalPredBlockProb();
 
     Cost.first += BlockCost.first;

llvm/test/Transforms/LoopVectorize/ARM/scalar-block-cost.ll

@@ -15,7 +15,7 @@ define void @pred_loop(i32* %off, i32* %data, i32* %dst, i32 %n) #0 {
 ; CHECK-COST-NEXT: LV: Found an estimated cost of 1 for VF 1 For instruction:   store i32 %add1, i32* %arrayidx2, align 4
 ; CHECK-COST-NEXT: LV: Found an estimated cost of 1 for VF 1 For instruction:   %exitcond.not = icmp eq i32 %add, %n
 ; CHECK-COST-NEXT: LV: Found an estimated cost of 0 for VF 1 For instruction:   br i1 %exitcond.not, label %exit.loopexit, label %for.body
-; CHECK-COST-NEXT: LV: Scalar loop costs: 2.
+; CHECK-COST-NEXT: LV: Scalar loop costs: 5.
 
 entry:
   %cmp8 = icmp sgt i32 %n, 0