; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --filter-out-after "scalar.ph:" --version 5
; 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:    [[TMP5:%.*]] = sub i16 [[TMP3]], [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = icmp sgt i16 [[TMP5]], [[TMP3]]
; CHECK-NEXT:    [[TMP8:%.*]] = icmp ugt i32 [[TMP2]], 65535
; CHECK-NEXT:    [[TMP9:%.*]] = or i1 [[TMP6]], [[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:    [[TMP13:%.*]] = extractelement <2 x i1> [[TMP11]], i32 0
; CHECK-NEXT:    br i1 [[TMP13]], 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:    [[TMP14:%.*]] = extractelement <2 x i1> [[TMP11]], i32 1
; CHECK-NEXT:    br i1 [[TMP14]], 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:    [[TMP15:%.*]] = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 [[TMP12]]
; CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i16, ptr [[TMP15]], i64 0
; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i16, ptr [[TMP16]], i64 -1
; CHECK-NEXT:    store <2 x i16> zeroinitializer, ptr [[TMP17]], align 1
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 2
; CHECK-NEXT:    [[TMP18:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP18]], 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]], [[EXIT:label %.*]], label %[[SCALAR_PH]]
; CHECK:       [[SCALAR_PH]]:
;
entry:
  br label %loop.header

loop.header:
  %iv = phi i32 [ %start, %entry ], [ %iv.next, %loop.latch ]
  br i1 %c, label %loop.latch, label %else

else:
  store i32 10, ptr %b, align 1
  br label %loop.latch

loop.latch:
  %t = trunc i32 %iv to i16
  %arrayidx = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 %t
  store i16 0, ptr %arrayidx, align 1
  %iv.next = add nsw i32 %iv, -1
  %cmp = icmp sgt i32 %iv, 1
  br i1 %cmp, label %loop.header, 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:    [[TMP5:%.*]] = sub i16 [[TMP3]], [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = icmp sgt i16 [[TMP5]], [[TMP3]]
; CHECK-NEXT:    [[TMP8:%.*]] = icmp ugt i32 [[TMP2]], 65535
; CHECK-NEXT:    [[TMP9:%.*]] = or i1 [[TMP6]], [[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:    [[TMP12:%.*]] = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 [[TMP11]]
; CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i16, ptr [[TMP12]], i64 0
; CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i16, ptr [[TMP13]], i64 -1
; CHECK-NEXT:    store <2 x i16> zeroinitializer, ptr [[TMP14]], align 1
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 2
; CHECK-NEXT:    [[TMP15:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP15]], 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]], [[EXIT:label %.*]], label %[[SCALAR_PH]]
; CHECK:       [[SCALAR_PH]]:
;
entry:
  br label %loop.header

loop.header:
  %iv = phi i32 [ %start, %entry ], [ %iv.next, %loop.latch ]
  br i1 %c, label %loop.latch, label %else

else:
  %1 = load i32, ptr %b, align 1
  br label %loop.latch

loop.latch:
  %2 = trunc i32 %iv to i16
  %arrayidx = getelementptr inbounds [2 x i16], ptr @a, i16 0, i16 %2
  store i16 0, ptr %arrayidx, align 1
  %iv.next = add nsw i32 %iv, -1
  %cmp = icmp sgt i32 %iv, 1
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}