[VPlan] Also allow extracts as users when converting to single scalars.

Extracts technically do not use scalars, but vectors, but if the operand
is a single scalar we do not need a vector and they should not block
forming single scalars.

Author: fhahn

llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp

@@ -1311,7 +1311,11 @@ static void narrowToSingleScalarRecipes(VPlan &Plan) {
       // scalar results used. In the latter case, we would introduce extra
       // broadcasts.
       if (!vputils::isSingleScalar(RepOrWidenR) ||
-          !vputils::onlyScalarValuesUsed(RepOrWidenR))
+          !all_of(RepOrWidenR->users(), [RepOrWidenR](const VPUser *U) {
+            return U->usesScalars(RepOrWidenR) ||
+                   match(cast<VPRecipeBase>(U),
+                         m_ExtractLastElement(m_VPValue()));
+          }))
         continue;
 
       auto *Clone = new VPReplicateRecipe(RepOrWidenR->getUnderlyingInstr(),

llvm/test/Transforms/LoopVectorize/pr66616.ll

@@ -13,14 +13,11 @@ define void @pr66616(ptr %ptr) {
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP0:%.*]] = load i32, ptr [[PTR]], align 4
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[TMP0]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP1:%.*]] = add <4 x i32> [[BROADCAST_SPLAT]], splat (i32 1)
+; CHECK-NEXT:    [[TMP3:%.*]] = add i32 [[TMP0]], 1
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP2:%.*]] = icmp eq i32 [[INDEX_NEXT]], 256
 ; CHECK-NEXT:    br i1 [[TMP2]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x i32> [[TMP1]], i32 3
 ; CHECK-NEXT:    br label [[PREHEADER:%.*]]
 ; CHECK:       scalar.ph:
 ; CHECK-NEXT:    br label [[LOOP_1:%.*]]