; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
; RUN: opt -passes='loop-vectorize' -force-vector-width=2 -S < %s | FileCheck %s
;
; Forcing VF=2 to trigger vector code gen
;
; This is a test case that let's vectorizer's code gen to generate
; more than one BasicBlocks in the loop body (emulated masked scatter)
; for those targets that do not support masked scatter. Broadcast
; code generation was previously dependent on loop body being
; a single basic block and this test case exposed incorrect code gen
; resulting in an assert in IL verification.
;
@a = external global [2 x i16], align 1

define void @f1(ptr noalias %b, i1 %c, i32 %start) {
; CHECK-LABEL: define void @f1
; CHECK-SAME: (ptr noalias [[B:%.*]], i1 [[C:%.*]], i32 [[START:%.*]]) {
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[START]], 1
; CHECK-NEXT:    [[SMIN1:%.*]] = call i32 @llvm.smin.i32(i32 [[START]], i32 1)
; CHECK-NEXT:    [[TMP1:%.*]] = sub i32 [[TMP0]], [[SMIN1]]
; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[TMP1]], 2
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_SCEVCHECK:%.*]]
; CHECK:       vector.scevcheck:
; CHECK-NEXT:    [[SMIN:%.*]] = call i32 @llvm.smin.i32(i32 [[START]], i32 1)
; CHECK-NEXT:    [[TMP2:%.*]] = sub i32 [[START]], [[SMIN]]
; CHECK-NEXT:    [[TMP3:%.*]] = trunc i32 [[START]] to i16
; CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP2]] to i16
; CHECK-NEXT:    [[MUL:%.*]] = call { i16, i1 } @llvm.umul.with.overflow.i16(i16 1, i16 [[TMP4]])
; CHECK-NEXT:    [[MUL_RESULT:%.*]] = extractvalue { i16, i1 } [[MUL]], 0
; CHECK-NEXT:    [[MUL_OVERFLOW:%.*]] = extractvalue { i16, i1 } [[MUL]], 1
; CHECK-NEXT:    [[TMP5:%.*]] = sub i16 [[TMP3]], [[MUL_RESULT]]
; CHECK-NEXT:    [[TMP6:%.*]] = icmp sgt i16 [[TMP5]], [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = or i1 [[TMP6]], [[MUL_OVERFLOW]]
; CHECK-NEXT:    [[TMP8:%.*]] = icmp ugt i32 [[TMP2]], 65535
; CHECK-NEXT:    [[TMP9:%.*]] = or i1 [[TMP7]], [[TMP8]]
; CHECK-NEXT:    br i1 [[TMP9]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
; CHECK:       vector.ph:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i32 [[TMP1]], 2
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i32 [[TMP1]], [[N_MOD_VF]]
; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <2 x i1> poison, i1 [[C]], i64 0
; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <2 x i1> [[BROADCAST_SPLATINSERT]], <2 x i1> poison, <2 x i32> zeroinitializer
; CHECK-NEXT:    [[TMP10:%.*]] = sub i32 [[START]], [[N_VEC]]
; CHECK-NEXT:    [[TMP11:%.*]] = xor <2 x i1> [[BROADCAST_SPLAT]], splat (i1 true)
; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[PRED_STORE_CONTINUE3:%.*]] ]
; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = sub i32 [[START]], [[INDEX]]
; CHECK-NEXT:    [[TMP12:%.*]] = trunc i32 [[OFFSET_IDX]] to i16
; CHECK-NEXT:    [[TMP14:%.*]] = extractelement <2 x i1> [[TMP11]], i32 0
; CHECK-NEXT:    br i1 [[TMP14]], label [[PRED_STORE_IF:%.*]], label [[PRED_STORE_CONTINUE:%.*]]
; CHECK:       pred.store.if:
; CHECK-NEXT:    store i32 10, ptr [[B]], align 1
; CHECK-NEXT:    br label [[PRED_STORE_CONTINUE]]
; CHECK:       pred.store.continue:
; CHECK-NEXT:    [[TMP15:%.*]] = extractelement <2 x i1> [[TMP11]], i32 1
; CHECK-NEXT:    br i1 [[TMP15]], label [[PRED_STORE_IF2:%.*]], label [[PRED_STORE_CONTINUE3]]
; CHECK:       pred.store.if2:
; CHECK-NEXT:    store i32 10, ptr [[B]], align 1
; CHECK-NEXT:    br label [[PRED_STORE_CONTINUE3]]
; CHECK:       pred.store.continue3:
; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 [[TMP12]]
; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i16, ptr [[TMP16]], i32 0
; CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i16, ptr [[TMP17]], i32 -1
; CHECK-NEXT:    store <2 x i16> zeroinitializer, ptr [[TMP18]], align 1
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 2
; CHECK-NEXT:    [[TMP19:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP19]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       middle.block:
; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i32 [[TMP1]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; CHECK:       scalar.ph:
; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[TMP10]], [[MIDDLE_BLOCK]] ], [ [[START]], [[ENTRY:%.*]] ], [ [[START]], [[VECTOR_SCEVCHECK]] ]
; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
; CHECK:       for.body:
; CHECK-NEXT:    [[TMP20:%.*]] = phi i32 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[DEC:%.*]], [[LAND_END:%.*]] ]
; CHECK-NEXT:    br i1 [[C]], label [[LAND_END]], label [[LAND_RHS:%.*]]
; CHECK:       land.rhs:
; CHECK-NEXT:    store i32 10, ptr [[B]], align 1
; CHECK-NEXT:    br label [[LAND_END]]
; CHECK:       land.end:
; CHECK-NEXT:    [[T:%.*]] = trunc i32 [[TMP20]] to i16
; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 [[T]]
; CHECK-NEXT:    store i16 0, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT:    [[DEC]] = add nsw i32 [[TMP20]], -1
; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[TMP20]], 1
; CHECK-NEXT:    br i1 [[CMP]], label [[FOR_BODY]], label [[EXIT]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK:       exit:
; CHECK-NEXT:    ret void
;
entry:
  br label %for.body

for.body:                                         ; preds = %land.end, %entry
  %0 = phi i32 [ %start, %entry ], [ %dec, %land.end ]
  br i1 %c, label %land.end, label %land.rhs

land.rhs:                                         ; preds = %for.body
  store i32 10, ptr %b, align 1
  br label %land.end

land.end:                                         ; preds = %land.rhs, %for.body
  %t = trunc i32 %0 to i16
  %arrayidx = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 %t
  store i16 0, ptr %arrayidx, align 1
  %dec = add nsw i32 %0, -1
  %cmp = icmp sgt i32 %0, 1
  br i1 %cmp, label %for.body, label %exit

exit:
  ret void
}

define void @f2(ptr noalias %b, i1 %c, i32 %start) {
; CHECK-LABEL: define void @f2
; CHECK-SAME: (ptr noalias [[B:%.*]], i1 [[C:%.*]], i32 [[START:%.*]]) {
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[START]], 1
; CHECK-NEXT:    [[SMIN1:%.*]] = call i32 @llvm.smin.i32(i32 [[START]], i32 1)
; CHECK-NEXT:    [[TMP1:%.*]] = sub i32 [[TMP0]], [[SMIN1]]
; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[TMP1]], 2
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_SCEVCHECK:%.*]]
; CHECK:       vector.scevcheck:
; CHECK-NEXT:    [[SMIN:%.*]] = call i32 @llvm.smin.i32(i32 [[START]], i32 1)
; CHECK-NEXT:    [[TMP2:%.*]] = sub i32 [[START]], [[SMIN]]
; CHECK-NEXT:    [[TMP3:%.*]] = trunc i32 [[START]] to i16
; CHECK-NEXT:    [[TMP4:%.*]] = trunc i32 [[TMP2]] to i16
; CHECK-NEXT:    [[MUL:%.*]] = call { i16, i1 } @llvm.umul.with.overflow.i16(i16 1, i16 [[TMP4]])
; CHECK-NEXT:    [[MUL_RESULT:%.*]] = extractvalue { i16, i1 } [[MUL]], 0
; CHECK-NEXT:    [[MUL_OVERFLOW:%.*]] = extractvalue { i16, i1 } [[MUL]], 1
; CHECK-NEXT:    [[TMP5:%.*]] = sub i16 [[TMP3]], [[MUL_RESULT]]
; CHECK-NEXT:    [[TMP6:%.*]] = icmp sgt i16 [[TMP5]], [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = or i1 [[TMP6]], [[MUL_OVERFLOW]]
; CHECK-NEXT:    [[TMP8:%.*]] = icmp ugt i32 [[TMP2]], 65535
; CHECK-NEXT:    [[TMP9:%.*]] = or i1 [[TMP7]], [[TMP8]]
; CHECK-NEXT:    br i1 [[TMP9]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
; CHECK:       vector.ph:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i32 [[TMP1]], 2
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i32 [[TMP1]], [[N_MOD_VF]]
; CHECK-NEXT:    [[TMP10:%.*]] = sub i32 [[START]], [[N_VEC]]
; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = sub i32 [[START]], [[INDEX]]
; CHECK-NEXT:    [[TMP11:%.*]] = trunc i32 [[OFFSET_IDX]] to i16
; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 [[TMP11]]
; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i16, ptr [[TMP13]], i32 0
; CHECK-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i16, ptr [[TMP14]], i32 -1
; CHECK-NEXT:    store <2 x i16> zeroinitializer, ptr [[TMP15]], align 1
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 2
; CHECK-NEXT:    [[TMP16:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP16]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK:       middle.block:
; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i32 [[TMP1]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; CHECK:       scalar.ph:
; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i32 [ [[TMP10]], [[MIDDLE_BLOCK]] ], [ [[START]], [[ENTRY:%.*]] ], [ [[START]], [[VECTOR_SCEVCHECK]] ]
; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
; CHECK:       for.body:
; CHECK-NEXT:    [[TMP17:%.*]] = phi i32 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[DEC:%.*]], [[LAND_END:%.*]] ]
; CHECK-NEXT:    br i1 [[C]], label [[LAND_END]], label [[LAND_RHS:%.*]]
; CHECK:       land.rhs:
; CHECK-NEXT:    [[TMP18:%.*]] = load i32, ptr [[B]], align 1
; CHECK-NEXT:    br label [[LAND_END]]
; CHECK:       land.end:
; CHECK-NEXT:    [[TMP19:%.*]] = trunc i32 [[TMP17]] to i16
; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 [[TMP19]]
; CHECK-NEXT:    store i16 0, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT:    [[DEC]] = add nsw i32 [[TMP17]], -1
; CHECK-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[TMP17]], 1
; CHECK-NEXT:    br i1 [[CMP]], label [[FOR_BODY]], label [[EXIT]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK:       exit:
; CHECK-NEXT:    ret void
;
entry:
  br label %for.body

for.body:                                         ; preds = %land.end, %entry
  %0 = phi i32 [ %start, %entry ], [ %dec, %land.end ]
  br i1 %c, label %land.end, label %land.rhs

land.rhs:                                         ; preds = %for.body
  %1 = load i32, ptr %b, align 1
  br label %land.end

land.end:                                         ; preds = %land.rhs, %for.body
  %2 = trunc i32 %0 to i16
  %arrayidx = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 %2
  store i16 0, ptr %arrayidx, align 1
  %dec = add nsw i32 %0, -1
  %cmp = icmp sgt i32 %0, 1
  br i1 %cmp, label %for.body, label %exit

exit:
  ret void
}