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

; This tests exercises the out-of-band fixing-up of scalar resume values.

target triple = "aarch64"

define void @epilogue_vectorization_fix_scalar_resume_values(ptr %dst, i64 %n) {
; CHECK-LABEL: define void @epilogue_vectorization_fix_scalar_resume_values(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[N:%.*]]) {
; CHECK-NEXT:  [[ITER_CHECK:.*]]:
; CHECK-NEXT:    [[REM:%.*]] = urem i64 [[N]], 3
; CHECK-NEXT:    br i1 true, label %[[VEC_EPILOG_SCALAR_PH:.*]], label %[[VECTOR_MAIN_LOOP_ITER_CHECK:.*]]
; CHECK:       [[VECTOR_MAIN_LOOP_ITER_CHECK]]:
; CHECK-NEXT:    br i1 true, label %[[VEC_EPILOG_PH:.*]], label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[REM]], 32
; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[REM]], [[N_MOD_VF]]
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i8, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr i8, ptr [[TMP0]], i64 16
; CHECK-NEXT:    store <16 x i8> zeroinitializer, ptr [[TMP0]], align 1
; CHECK-NEXT:    store <16 x i8> zeroinitializer, ptr [[TMP1]], align 1
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 32
; CHECK-NEXT:    [[TMP2:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[TMP2]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[REM]], [[N_VEC]]
; CHECK-NEXT:    br i1 [[CMP_N]], label %[[EXIT:.*]], label %[[VEC_EPILOG_ITER_CHECK:.*]]
; CHECK:       [[VEC_EPILOG_ITER_CHECK]]:
; CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_MOD_VF]], 8
; CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK]], label %[[VEC_EPILOG_SCALAR_PH]], label %[[VEC_EPILOG_PH]], !prof [[PROF3:![0-9]+]]
; CHECK:       [[VEC_EPILOG_PH]]:
; CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[VEC_EPILOG_ITER_CHECK]] ], [ 0, %[[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT:    br label %[[VEC_EPILOG_VECTOR_BODY:.*]]
; CHECK:       [[VEC_EPILOG_VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX1:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], %[[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT2:%.*]], %[[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr i8, ptr [[DST]], i64 [[INDEX1]]
; CHECK-NEXT:    store <8 x i8> zeroinitializer, ptr [[TMP3]], align 1
; CHECK-NEXT:    [[INDEX_NEXT2]] = add nuw i64 [[INDEX1]], 8
; CHECK-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT2]], 0
; CHECK-NEXT:    br i1 [[TMP4]], label %[[VEC_EPILOG_MIDDLE_BLOCK:.*]], label %[[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK:       [[VEC_EPILOG_MIDDLE_BLOCK]]:
; CHECK-NEXT:    [[CMP_N3:%.*]] = icmp eq i64 [[REM]], 0
; CHECK-NEXT:    br i1 [[CMP_N3]], label %[[EXIT]], label %[[VEC_EPILOG_SCALAR_PH]]
; CHECK:       [[VEC_EPILOG_SCALAR_PH]]:
; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ 0, %[[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], %[[VEC_EPILOG_ITER_CHECK]] ], [ 0, %[[ITER_CHECK]] ]
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], %[[VEC_EPILOG_SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[GEP_DST_IV:%.*]] = getelementptr i8, ptr [[DST]], i64 [[IV]]
; CHECK-NEXT:    store i8 0, ptr [[GEP_DST_IV]], align 1
; CHECK-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
; CHECK-NEXT:    [[EXIT_COND:%.*]] = icmp eq i64 [[IV_NEXT]], [[REM]]
; CHECK-NEXT:    br i1 [[EXIT_COND]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  %rem = urem i64 %n, 3
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep.dst.iv = getelementptr i8, ptr %dst, i64 %iv
  store i8 0, ptr %gep.dst.iv, align 1
  %iv.next = add i64 %iv, 1
  %exit.cond = icmp eq i64 %iv.next, %rem
  br i1 %exit.cond, label %exit, label %loop

exit:
  ret void
}
