; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6
; RUN: opt -p loop-vectorize -force-vector-width=4 < %s -S | FileCheck %s

define void @reverse_copy(ptr noalias %src, ptr noalias %dst, i32 %n) {
; CHECK-LABEL: define void @reverse_copy(
; CHECK-SAME: ptr noalias [[SRC:%.*]], ptr noalias [[DST:%.*]], i32 [[N:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    [[TMP0:%.*]] = add nsw i32 [[N]], -1
; CHECK-NEXT:    [[SMIN:%.*]] = call i32 @llvm.smin.i32(i32 [[TMP0]], i32 0)
; CHECK-NEXT:    [[TMP1:%.*]] = sub i32 [[N]], [[SMIN]]
; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[TMP1]], 4
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = and i32 [[TMP1]], 3
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i32 [[TMP1]], [[N_MOD_VF]]
; CHECK-NEXT:    [[TMP2:%.*]] = sub i32 [[N]], [[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:    [[TMP3:%.*]] = sub i32 [[N]], [[INDEX]]
; CHECK-NEXT:    [[TMP4:%.*]] = add nsw i32 [[TMP3]], -1
; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr i32, ptr [[SRC]], i32 [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr i32, ptr [[TMP5]], i64 -3
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP6]], align 4
; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr i32, ptr [[DST]], i32 [[TMP4]]
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr i32, ptr [[TMP7]], i64 -3
; CHECK-NEXT:    store <4 x i32> [[WIDE_LOAD]], ptr [[TMP8]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP9]], 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 [ [[TMP2]], %[[MIDDLE_BLOCK]] ], [ [[N]], %[[ENTRY]] ]
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i32 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[IV_NEXT]] = add nsw i32 [[IV]], -1
; CHECK-NEXT:    [[GEP_SRC:%.*]] = getelementptr i32, ptr [[SRC]], i32 [[IV_NEXT]]
; CHECK-NEXT:    [[X:%.*]] = load i32, ptr [[GEP_SRC]], align 4
; CHECK-NEXT:    [[GEP_DST:%.*]] = getelementptr i32, ptr [[DST]], i32 [[IV_NEXT]]
; CHECK-NEXT:    store i32 [[X]], ptr [[GEP_DST]], align 4
; CHECK-NEXT:    [[EC:%.*]] = icmp sgt i32 [[IV_NEXT]], 0
; CHECK-NEXT:    br i1 [[EC]], label %[[LOOP]], label %[[EXIT]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i32 [%n, %entry], [%iv.next, %loop]
  %iv.next = add nsw i32 %iv, -1
  %gep.src = getelementptr i32, ptr %src, i32 %iv.next
  %x = load i32, ptr %gep.src
  %gep.dst = getelementptr i32, ptr %dst, i32 %iv.next
  store i32 %x, ptr %gep.dst
  %ec = icmp sgt i32 %iv.next, 0
  br i1 %ec, label %loop, label %exit

exit:
  ret void
}

define void @reverse_binop(ptr noalias %src, ptr noalias %dst, i32 %n) {
; CHECK-LABEL: define void @reverse_binop(
; CHECK-SAME: ptr noalias [[SRC:%.*]], ptr noalias [[DST:%.*]], i32 [[N:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    [[TMP0:%.*]] = add nsw i32 [[N]], -1
; CHECK-NEXT:    [[SMIN:%.*]] = call i32 @llvm.smin.i32(i32 [[TMP0]], i32 0)
; CHECK-NEXT:    [[TMP1:%.*]] = sub i32 [[N]], [[SMIN]]
; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[TMP1]], 4
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = and i32 [[TMP1]], 3
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i32 [[TMP1]], [[N_MOD_VF]]
; CHECK-NEXT:    [[TMP2:%.*]] = sub i32 [[N]], [[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:    [[TMP3:%.*]] = sub i32 [[N]], [[INDEX]]
; CHECK-NEXT:    [[TMP4:%.*]] = add nsw i32 [[TMP3]], -1
; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr i32, ptr [[SRC]], i32 [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr i32, ptr [[TMP5]], i64 -3
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP6]], align 4
; CHECK-NEXT:    [[TMP7:%.*]] = add <4 x i32> [[WIDE_LOAD]], splat (i32 1)
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr i32, ptr [[DST]], i32 [[TMP4]]
; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr i32, ptr [[TMP8]], i64 -3
; CHECK-NEXT:    store <4 x i32> [[TMP7]], ptr [[TMP9]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
; CHECK-NEXT:    [[TMP10:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP10]], 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 [ [[TMP2]], %[[MIDDLE_BLOCK]] ], [ [[N]], %[[ENTRY]] ]
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i32 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[IV_NEXT]] = add nsw i32 [[IV]], -1
; CHECK-NEXT:    [[GEP_SRC:%.*]] = getelementptr i32, ptr [[SRC]], i32 [[IV_NEXT]]
; CHECK-NEXT:    [[X:%.*]] = load i32, ptr [[GEP_SRC]], align 4
; CHECK-NEXT:    [[Y:%.*]] = add i32 [[X]], 1
; CHECK-NEXT:    [[GEP_DST:%.*]] = getelementptr i32, ptr [[DST]], i32 [[IV_NEXT]]
; CHECK-NEXT:    store i32 [[Y]], ptr [[GEP_DST]], align 4
; CHECK-NEXT:    [[EC:%.*]] = icmp sgt i32 [[IV_NEXT]], 0
; CHECK-NEXT:    br i1 [[EC]], label %[[LOOP]], label %[[EXIT]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i32 [%n, %entry], [%iv.next, %loop]
  %iv.next = add nsw i32 %iv, -1
  %gep.src = getelementptr i32, ptr %src, i32 %iv.next
  %x = load i32, ptr %gep.src
  %y = add i32 %x, 1
  %gep.dst = getelementptr i32, ptr %dst, i32 %iv.next
  store i32 %y, ptr %gep.dst
  %ec = icmp sgt i32 %iv.next, 0
  br i1 %ec, label %loop, label %exit

exit:
  ret void
}

define void @reverse_fneg(ptr noalias %src, ptr noalias %dst, i32 %n) {
; CHECK-LABEL: define void @reverse_fneg(
; CHECK-SAME: ptr noalias [[SRC:%.*]], ptr noalias [[DST:%.*]], i32 [[N:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    [[TMP0:%.*]] = add nsw i32 [[N]], -1
; CHECK-NEXT:    [[SMIN:%.*]] = call i32 @llvm.smin.i32(i32 [[TMP0]], i32 0)
; CHECK-NEXT:    [[TMP1:%.*]] = sub i32 [[N]], [[SMIN]]
; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[TMP1]], 4
; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = and i32 [[TMP1]], 3
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i32 [[TMP1]], [[N_MOD_VF]]
; CHECK-NEXT:    [[TMP2:%.*]] = sub i32 [[N]], [[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:    [[TMP3:%.*]] = sub i32 [[N]], [[INDEX]]
; CHECK-NEXT:    [[TMP4:%.*]] = add nsw i32 [[TMP3]], -1
; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr float, ptr [[SRC]], i32 [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr float, ptr [[TMP5]], i64 -3
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP6]], align 4
; CHECK-NEXT:    [[REVERSE1:%.*]] = fneg <4 x float> [[WIDE_LOAD]]
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr float, ptr [[DST]], i32 [[TMP4]]
; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr float, ptr [[TMP8]], i64 -3
; CHECK-NEXT:    store <4 x float> [[REVERSE1]], ptr [[TMP9]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
; CHECK-NEXT:    [[TMP10:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP10]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![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 [ [[TMP2]], %[[MIDDLE_BLOCK]] ], [ [[N]], %[[ENTRY]] ]
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i32 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[IV_NEXT]] = add nsw i32 [[IV]], -1
; CHECK-NEXT:    [[GEP_SRC:%.*]] = getelementptr float, ptr [[SRC]], i32 [[IV_NEXT]]
; CHECK-NEXT:    [[X:%.*]] = load float, ptr [[GEP_SRC]], align 4
; CHECK-NEXT:    [[Y:%.*]] = fneg float [[X]]
; CHECK-NEXT:    [[GEP_DST:%.*]] = getelementptr float, ptr [[DST]], i32 [[IV_NEXT]]
; CHECK-NEXT:    store float [[Y]], ptr [[GEP_DST]], align 4
; CHECK-NEXT:    [[EC:%.*]] = icmp sgt i32 [[IV_NEXT]], 0
; CHECK-NEXT:    br i1 [[EC]], label %[[LOOP]], label %[[EXIT]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i32 [%n, %entry], [%iv.next, %loop]
  %iv.next = add nsw i32 %iv, -1
  %gep.src = getelementptr float, ptr %src, i32 %iv.next
  %x = load float, ptr %gep.src
  %y = fneg float %x
  %gep.dst = getelementptr float, ptr %dst, i32 %iv.next
  store float %y, ptr %gep.dst
  %ec = icmp sgt i32 %iv.next, 0
  br i1 %ec, label %loop, label %exit

exit:
  ret void
}