memory-lowering.cc

Go to the documentation of this file.

// Copyright 2019 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
 
#include "src/compiler/memory-lowering.h"
 
#include "src/codegen/interface-descriptors-inl.h"
#include "src/common/globals.h"
#include "src/compiler/js-graph.h"
#include "src/compiler/linkage.h"
#include "src/compiler/node-matchers.h"
#include "src/compiler/node-properties.h"
#include "src/compiler/node.h"
#include "src/compiler/simplified-operator.h"
#include "src/roots/roots-inl.h"
#include "src/sandbox/external-pointer-inl.h"
 
#if V8_ENABLE_WEBASSEMBLY
#include "src/wasm/wasm-linkage.h"
#include "src/wasm/wasm-objects.h"
#endif
namespace v8 {
namespace internal {
namespace compiler {
 
// An allocation group represents a set of allocations that have been folded
// together.
class MemoryLowering::AllocationGroup final : public ZoneObject {
 public:
  AllocationGroup(Node* node, AllocationType allocation, Zone* zone);
  AllocationGroup(Node* node, AllocationType allocation, Node* size,
                  Zone* zone);
  ~AllocationGroup() = default;
 
  void Add(Node* object);
  bool Contains(Node* object) const;
  bool IsYoungGenerationAllocation() const {
    return allocation() == AllocationType::kYoung;
  }
 
  AllocationType allocation() const { return allocation_; }
  Node* size() const { return size_; }
 
 private:
  ZoneSet<NodeId> node_ids_;
  AllocationType const allocation_;
  Node* const size_;
 
  static inline AllocationType CheckAllocationType(AllocationType allocation) {
    // For non-generational heap, all young allocations are redirected to old
    // space.
    if (v8_flags.single_generation && allocation == AllocationType::kYoung) {
      return AllocationType::kOld;
    }
    return allocation;
  }
 
  DISALLOW_IMPLICIT_CONSTRUCTORS(AllocationGroup);
};
 
MemoryLowering::MemoryLowering(JSGraph* jsgraph, Zone* zone,
                               JSGraphAssembler* graph_assembler, bool is_wasm,
                               AllocationFolding allocation_folding,
                               WriteBarrierAssertFailedCallback callback,
                               const char* function_debug_name)
    : isolate_(jsgraph->isolate()),
      zone_(zone),
      graph_(jsgraph->graph()),
      common_(jsgraph->common()),
      machine_(jsgraph->machine()),
      graph_assembler_(graph_assembler),
      is_wasm_(is_wasm),
      allocation_folding_(allocation_folding),
      write_barrier_assert_failed_(callback),
      function_debug_name_(function_debug_name) {}
 
Zone* MemoryLowering::graph_zone() const { return graph()->zone(); }
 
Reduction MemoryLowering::Reduce(Node* node) {
  switch (node->opcode()) {
    case IrOpcode::kAllocate:
      // Allocate nodes were purged from the graph in effect-control
      // linearization.
      UNREACHABLE();
    case IrOpcode::kAllocateRaw:
      return ReduceAllocateRaw(node);
    case IrOpcode::kLoadFromObject:
    case IrOpcode::kLoadImmutableFromObject:
      return ReduceLoadFromObject(node);
    case IrOpcode::kLoadElement:
      return ReduceLoadElement(node);
    case IrOpcode::kLoadField:
      return ReduceLoadField(node);
    case IrOpcode::kStoreToObject:
    case IrOpcode::kInitializeImmutableInObject:
      return ReduceStoreToObject(node);
    case IrOpcode::kStoreElement:
      return ReduceStoreElement(node);
    case IrOpcode::kStoreField:
      return ReduceStoreField(node);
    case IrOpcode::kStore:
      return ReduceStore(node);
    default:
      return NoChange();
  }
}
 
void MemoryLowering::EnsureAllocateOperator() {
  if (allocate_operator_.is_set()) return;
 
  auto descriptor = AllocateDescriptor{};
  StubCallMode mode = isolate_ != nullptr ? StubCallMode::kCallCodeObject
                                          : StubCallMode::kCallBuiltinPointer;
  auto call_descriptor = Linkage::GetStubCallDescriptor(
      graph_zone(), descriptor, descriptor.GetStackParameterCount(),
      CallDescriptor::kCanUseRoots, Operator::kNoThrow, mode);
  allocate_operator_.set(common()->Call(call_descriptor));
}
 
#if V8_ENABLE_WEBASSEMBLY
Node* MemoryLowering::GetWasmInstanceNode() {
  if (wasm_instance_node_.is_set()) return wasm_instance_node_.get();
  for (Node* use : graph()->start()->uses()) {
    if (use->opcode() == IrOpcode::kParameter &&
        ParameterIndexOf(use->op()) == wasm::kWasmInstanceDataParameterIndex) {
      wasm_instance_node_.set(use);
      return use;
    }
  }
  UNREACHABLE();  // The instance node must have been created before.
}
#endif  // V8_ENABLE_WEBASSEMBLY
 
#define __ gasm()->
 
Node* MemoryLowering::AlignToAllocationAlignment(Node* value) {
  if (!V8_COMPRESS_POINTERS_8GB_BOOL) return value;
 
  auto already_aligned = __ MakeLabel(MachineRepresentation::kWord64);
  Node* alignment_check = __ WordEqual(
      __ WordAnd(value, __ UintPtrConstant(kObjectAlignment8GbHeapMask)),
      __ UintPtrConstant(0));
 
  __ GotoIf(alignment_check, &already_aligned, value);
  {
    Node* aligned_value;
    if (kObjectAlignment8GbHeap == 2 * kTaggedSize) {
      aligned_value = __ IntPtrAdd(value, __ IntPtrConstant(kTaggedSize));
    } else {
      aligned_value = __ WordAnd(
          __ IntPtrAdd(value, __ IntPtrConstant(kObjectAlignment8GbHeapMask)),
          __ UintPtrConstant(~kObjectAlignment8GbHeapMask));
    }
    __ Goto(&already_aligned, aligned_value);
  }
 
  __ Bind(&already_aligned);
 
  return already_aligned.PhiAt(0);
}
 
Reduction MemoryLowering::ReduceAllocateRaw(Node* node,
                                            AllocationType allocation_type,
                                            AllocationState const** state_ptr) {
  DCHECK_EQ(IrOpcode::kAllocateRaw, node->opcode());
  DCHECK_IMPLIES(allocation_folding_ == AllocationFolding::kDoAllocationFolding,
                 state_ptr != nullptr);
  if (v8_flags.single_generation && allocation_type == AllocationType::kYoung) {
    allocation_type = AllocationType::kOld;
  }
  // InstructionStream objects may have a maximum size smaller than
  // kMaxHeapObjectSize due to guard pages. If we need to support allocating
  // code here we would need to call
  // MemoryChunkLayout::MaxRegularCodeObjectSize() at runtime.
  DCHECK_NE(allocation_type, AllocationType::kCode);
  Node* value;
  Node* size = node->InputAt(0);
  Node* effect = node->InputAt(1);
  Node* control = node->InputAt(2);
 
  gasm()->InitializeEffectControl(effect, control);
 
  Node* allocate_builtin;
  if (!is_wasm_) {
    if (allocation_type == AllocationType::kYoung) {
      allocate_builtin = __ AllocateInYoungGenerationStubConstant();
    } else {
      allocate_builtin = __ AllocateInOldGenerationStubConstant();
    }
  } else {
#if V8_ENABLE_WEBASSEMBLY
    // This lowering is used by Wasm, where we compile isolate-independent
    // code. Builtin calls simply encode the target builtin ID, which will
    // be patched to the builtin's address later.
    if (isolate_ == nullptr) {
      Builtin builtin;
      if (allocation_type == AllocationType::kYoung) {
        builtin = Builtin::kWasmAllocateInYoungGeneration;
      } else {
        builtin = Builtin::kWasmAllocateInOldGeneration;
      }
      static_assert(std::is_same<Smi, BuiltinPtr>(), "BuiltinPtr must be Smi");
      allocate_builtin =
          graph()->NewNode(common()->NumberConstant(static_cast<int>(builtin)));
    } else {
      if (allocation_type == AllocationType::kYoung) {
        allocate_builtin = __ WasmAllocateInYoungGenerationStubConstant();
      } else {
        allocate_builtin = __ WasmAllocateInOldGenerationStubConstant();
      }
    }
#else
    UNREACHABLE();
#endif
  }
 
  // Determine the top/limit addresses.
  Node* top_address;
  Node* limit_address;
  if (isolate_ != nullptr) {
    top_address = __ ExternalConstant(
        allocation_type == AllocationType::kYoung
            ? ExternalReference::new_space_allocation_top_address(isolate())
            : ExternalReference::old_space_allocation_top_address(isolate()));
    limit_address = __ ExternalConstant(
        allocation_type == AllocationType::kYoung
            ? ExternalReference::new_space_allocation_limit_address(isolate())
            : ExternalReference::old_space_allocation_limit_address(isolate()));
  } else {
    // Wasm mode: producing isolate-independent code, loading the isolate
    // address at runtime.
#if V8_ENABLE_WEBASSEMBLY
    Node* instance_node = GetWasmInstanceNode();
    int top_address_offset =
        allocation_type == AllocationType::kYoung
            ? WasmTrustedInstanceData::kNewAllocationTopAddressOffset
            : WasmTrustedInstanceData::kOldAllocationTopAddressOffset;
    int limit_address_offset =
        allocation_type == AllocationType::kYoung
            ? WasmTrustedInstanceData::kNewAllocationLimitAddressOffset
            : WasmTrustedInstanceData::kOldAllocationLimitAddressOffset;
    top_address =
        __ Load(MachineType::Pointer(), instance_node,
                __ IntPtrConstant(top_address_offset - kHeapObjectTag));
    limit_address =
        __ Load(MachineType::Pointer(), instance_node,
                __ IntPtrConstant(limit_address_offset - kHeapObjectTag));
#else
    UNREACHABLE();
#endif  // V8_ENABLE_WEBASSEMBLY
  }
 
  // Check if we can fold this allocation into a previous allocation represented
  // by the incoming {state}.
  IntPtrMatcher m(size);
  if (m.IsInRange(0, kMaxRegularHeapObjectSize) && v8_flags.inline_new &&
      allocation_folding_ == AllocationFolding::kDoAllocationFolding) {
    intptr_t const object_size =
        ALIGN_TO_ALLOCATION_ALIGNMENT(m.ResolvedValue());
    AllocationState const* state = *state_ptr;
    if (state->size() <= kMaxRegularHeapObjectSize - object_size &&
        state->group()->allocation() == allocation_type) {
      // We can fold this Allocate {node} into the allocation {group}
      // represented by the given {state}. Compute the upper bound for
      // the new {state}.
      intptr_t const state_size = state->size() + object_size;
 
      // Update the reservation check to the actual maximum upper bound.
      AllocationGroup* const group = state->group();
      if (machine()->Is64()) {
        if (OpParameter<int64_t>(group->size()->op()) < state_size) {
          NodeProperties::ChangeOp(group->size(),
                                   common()->Int64Constant(state_size));
        }
      } else {
        if (OpParameter<int32_t>(group->size()->op()) < state_size) {
          NodeProperties::ChangeOp(
              group->size(),
              common()->Int32Constant(static_cast<int32_t>(state_size)));
        }
      }
 
      // Update the allocation top with the new object allocation.
      // TODO(bmeurer): Defer writing back top as much as possible.
      DCHECK_IMPLIES(V8_COMPRESS_POINTERS_8GB_BOOL,
                     IsAligned(object_size, kObjectAlignment8GbHeap));
      Node* top = __ IntAdd(state->top(), __ IntPtrConstant(object_size));
      __ Store(StoreRepresentation(MachineType::PointerRepresentation(),
                                   kNoWriteBarrier),
               top_address, __ IntPtrConstant(0), top);
 
      // Compute the effective inner allocated address.
      value = __ BitcastWordToTagged(
          __ IntAdd(state->top(), __ IntPtrConstant(kHeapObjectTag)));
      effect = gasm()->effect();
      control = gasm()->control();
 
      // Extend the allocation {group}.
      group->Add(value);
      *state_ptr =
          AllocationState::Open(group, state_size, top, effect, zone());
    } else {
      auto call_runtime = __ MakeDeferredLabel();
      auto done = __ MakeLabel(MachineType::PointerRepresentation());
 
      // Setup a mutable reservation size node; will be patched as we fold
      // additional allocations into this new group.
      Node* reservation_size = __ UniqueIntPtrConstant(object_size);
 
      // Load allocation top and limit.
      Node* top =
          __ Load(MachineType::Pointer(), top_address, __ IntPtrConstant(0));
      Node* limit =
          __ Load(MachineType::Pointer(), limit_address, __ IntPtrConstant(0));
 
      // Check if we need to collect garbage before we can start bump pointer
      // allocation (always done for folded allocations).
      Node* check = __ UintLessThan(__ IntAdd(top, reservation_size), limit);
 
      __ GotoIfNot(check, &call_runtime);
      __ Goto(&done, top);
 
      __ Bind(&call_runtime);
      {
        EnsureAllocateOperator();
        Node* vfalse = __ BitcastTaggedToWord(__ Call(
            allocate_operator_.get(), allocate_builtin, reservation_size));
        vfalse = __ IntSub(vfalse, __ IntPtrConstant(kHeapObjectTag));
        __ Goto(&done, vfalse);
      }
 
      __ Bind(&done);
 
      // Compute the new top and write it back.
      top = __ IntAdd(done.PhiAt(0), __ IntPtrConstant(object_size));
      __ Store(StoreRepresentation(MachineType::PointerRepresentation(),
                                   kNoWriteBarrier),
               top_address, __ IntPtrConstant(0), top);
 
      // Compute the initial object address.
      value = __ BitcastWordToTagged(
          __ IntAdd(done.PhiAt(0), __ IntPtrConstant(kHeapObjectTag)));
      effect = gasm()->effect();
      control = gasm()->control();
 
      // Start a new allocation group.
      AllocationGroup* group = zone()->New<AllocationGroup>(
          value, allocation_type, reservation_size, zone());
      *state_ptr =
          AllocationState::Open(group, object_size, top, effect, zone());
    }
  } else {
    auto call_runtime = __ MakeDeferredLabel();
    auto done = __ MakeLabel(MachineRepresentation::kTaggedPointer);
 
    // Load allocation top and limit.
    Node* top =
        __ Load(MachineType::Pointer(), top_address, __ IntPtrConstant(0));
    Node* limit =
        __ Load(MachineType::Pointer(), limit_address, __ IntPtrConstant(0));
 
    // Compute the new top.
    Node* new_top = __ IntAdd(top, AlignToAllocationAlignment(size));
 
    // Check if we can do bump pointer allocation here.
    Node* check = __ UintLessThan(new_top, limit);
    __ GotoIfNot(check, &call_runtime);
    __ GotoIfNot(
        __ UintLessThan(size, __ IntPtrConstant(kMaxRegularHeapObjectSize)),
        &call_runtime);
    __ Store(StoreRepresentation(MachineType::PointerRepresentation(),
                                 kNoWriteBarrier),
             top_address, __ IntPtrConstant(0), new_top);
    __ Goto(&done, __ BitcastWordToTagged(
                       __ IntAdd(top, __ IntPtrConstant(kHeapObjectTag))));
 
    __ Bind(&call_runtime);
    EnsureAllocateOperator();
    __ Goto(&done, __ Call(allocate_operator_.get(), allocate_builtin, size));
 
    __ Bind(&done);
    value = done.PhiAt(0);
    effect = gasm()->effect();
    control = gasm()->control();
 
    if (state_ptr) {
      // Create an unfoldable allocation group.
      AllocationGroup* group =
          zone()->New<AllocationGroup>(value, allocation_type, zone());
      *state_ptr = AllocationState::Closed(group, effect, zone());
    }
  }
 
  return Replace(value);
}
 
Reduction MemoryLowering::ReduceLoadFromObject(Node* node) {
  DCHECK(node->opcode() == IrOpcode::kLoadFromObject ||
         node->opcode() == IrOpcode::kLoadImmutableFromObject);
  ObjectAccess const& access = ObjectAccessOf(node->op());
 
  MachineType machine_type = access.machine_type;
 
  if (machine_type.IsMapWord()) {
    CHECK_EQ(machine_type.semantic(), MachineSemantic::kAny);
    return ReduceLoadMap(node);
  }
 
  MachineRepresentation rep = machine_type.representation();
  const Operator* load_op =
      ElementSizeInBytes(rep) > kTaggedSize &&
              !machine()->UnalignedLoadSupported(machine_type.representation())
          ? machine()->UnalignedLoad(machine_type)
          : machine()->Load(machine_type);
  NodeProperties::ChangeOp(node, load_op);
  return Changed(node);
}
 
Reduction MemoryLowering::ReduceLoadElement(Node* node) {
  DCHECK_EQ(IrOpcode::kLoadElement, node->opcode());
  ElementAccess const& access = ElementAccessOf(node->op());
  Node* index = node->InputAt(1);
  node->ReplaceInput(1, ComputeIndex(access, index));
  MachineType type = access.machine_type;
  DCHECK(!type.IsMapWord());
  NodeProperties::ChangeOp(node, machine()->Load(type));
  return Changed(node);
}
 
Reduction MemoryLowering::ReduceLoadExternalPointerField(Node* node) {
  DCHECK_EQ(node->opcode(), IrOpcode::kLoadField);
  FieldAccess const& access = FieldAccessOf(node->op());
 
#ifdef V8_ENABLE_SANDBOX
  ExternalPointerTag tag = access.external_pointer_tag;
  DCHECK_NE(tag, kExternalPointerNullTag);
  // Fields for sandboxed external pointer contain a 32-bit handle, not a
  // 64-bit raw pointer.
  NodeProperties::ChangeOp(node, machine()->Load(MachineType::Uint32()));
 
  Node* effect = NodeProperties::GetEffectInput(node);
  Node* control = NodeProperties::GetControlInput(node);
  __ InitializeEffectControl(effect, control);
 
  // Clone the load node and put it here.
  // TODO(turbofan): consider adding GraphAssembler::Clone() suitable for
  // cloning nodes from arbitrary locations in effect/control chains.
  static_assert(kExternalPointerIndexShift > kSystemPointerSizeLog2);
  Node* handle = __ AddNode(graph()->CloneNode(node));
  Node* shift_amount =
      __ Int32Constant(kExternalPointerIndexShift - kSystemPointerSizeLog2);
  Node* offset = __ Word32Shr(handle, shift_amount);
 
  // Uncomment this to generate a breakpoint for debugging purposes.
  // __ DebugBreak();
 
  // Decode loaded external pointer.
  //
  // Here we access the external pointer table through an ExternalReference.
  // Alternatively, we could also hardcode the address of the table since it
  // is never reallocated. However, in that case we must be able to guarantee
  // that the generated code is never executed under a different Isolate, as
  // that would allow access to external objects from different Isolates. It
  // also would break if the code is serialized/deserialized at some point.
  Node* table_address =
      IsSharedExternalPointerType(tag)
          ? __
            Load(MachineType::Pointer(),
                 __ ExternalConstant(
                     ExternalReference::
                         shared_external_pointer_table_address_address(
                             isolate())),
                 __ IntPtrConstant(0))
          : __ ExternalConstant(
                ExternalReference::external_pointer_table_address(isolate()));
  Node* table = __ Load(MachineType::Pointer(), table_address,
                        Internals::kExternalPointerTableBasePointerOffset);
  Node* pointer =
      __ Load(MachineType::Pointer(), table, __ ChangeUint32ToUint64(offset));
  Node* actual_tag =
      __ WordAnd(pointer, __ IntPtrConstant(kExternalPointerTagMask));
  actual_tag = __ TruncateInt64ToInt32(
      __ WordShr(actual_tag, __ IntPtrConstant(kExternalPointerTagShift)));
  Node* expected_tag = __ Int32Constant(tag);
  pointer =
      __ Word64And(pointer, __ IntPtrConstant(kExternalPointerPayloadMask));
  auto done = __ MakeLabel(MachineRepresentation::kWord64);
  __ GotoIf(__ WordEqual(actual_tag, expected_tag), &done, pointer);
  __ Goto(&done, __ IntPtrConstant(0));
  __ Bind(&done);
  return Replace(done.PhiAt(0));
#else
  NodeProperties::ChangeOp(node, machine()->Load(access.machine_type));
  return Changed(node);
#endif  // V8_ENABLE_SANDBOX
}
 
Reduction MemoryLowering::ReduceLoadBoundedSize(Node* node) {
#ifdef V8_ENABLE_SANDBOX
  const Operator* load_op =
      !machine()->UnalignedLoadSupported(MachineRepresentation::kWord64)
          ? machine()->UnalignedLoad(MachineType::Uint64())
          : machine()->Load(MachineType::Uint64());
  NodeProperties::ChangeOp(node, load_op);
 
  Node* effect = NodeProperties::GetEffectInput(node);
  Node* control = NodeProperties::GetControlInput(node);
  __ InitializeEffectControl(effect, control);
 
  Node* raw_value = __ AddNode(graph()->CloneNode(node));
  Node* shift_amount = __ IntPtrConstant(kBoundedSizeShift);
  Node* decoded_size = __ Word64Shr(raw_value, shift_amount);
  return Replace(decoded_size);
#else
  UNREACHABLE();
#endif
}
 
Reduction MemoryLowering::ReduceLoadMap(Node* node) {
#ifdef V8_MAP_PACKING
  NodeProperties::ChangeOp(node, machine()->Load(MachineType::AnyTagged()));
 
  Node* effect = NodeProperties::GetEffectInput(node);
  Node* control = NodeProperties::GetControlInput(node);
  __ InitializeEffectControl(effect, control);
 
  node = __ AddNode(graph()->CloneNode(node));
  return Replace(__ UnpackMapWord(node));
#else
  NodeProperties::ChangeOp(node, machine()->Load(MachineType::TaggedPointer()));
  return Changed(node);
#endif
}
 
Reduction MemoryLowering::ReduceLoadField(Node* node) {
  DCHECK_EQ(IrOpcode::kLoadField, node->opcode());
  FieldAccess const& access = FieldAccessOf(node->op());
  Node* offset = __ IntPtrConstant(access.offset - access.tag());
  node->InsertInput(graph_zone(), 1, offset);
  MachineType type = access.machine_type;
 
  if (type.IsMapWord()) {
    DCHECK(!access.type.Is(Type::ExternalPointer()));
    return ReduceLoadMap(node);
  }
 
  if (access.type.Is(Type::ExternalPointer())) {
    return ReduceLoadExternalPointerField(node);
  }
 
  if (access.is_bounded_size_access) {
    return ReduceLoadBoundedSize(node);
  }
 
  NodeProperties::ChangeOp(node, machine()->Load(type));
 
  return Changed(node);
}
 
Reduction MemoryLowering::ReduceStoreToObject(Node* node,
                                              AllocationState const* state) {
  DCHECK(node->opcode() == IrOpcode::kStoreToObject ||
         node->opcode() == IrOpcode::kInitializeImmutableInObject);
  ObjectAccess const& access = ObjectAccessOf(node->op());
  Node* object = node->InputAt(0);
  Node* value = node->InputAt(2);
 
  WriteBarrierKind write_barrier_kind = ComputeWriteBarrierKind(
      node, object, value, state, access.write_barrier_kind);
  DCHECK(!access.machine_type.IsMapWord());
  MachineRepresentation rep = access.machine_type.representation();
  StoreRepresentation store_rep(rep, write_barrier_kind);
  const Operator* store_op = ElementSizeInBytes(rep) > kTaggedSize &&
                                     !machine()->UnalignedStoreSupported(rep)
                                 ? machine()->UnalignedStore(rep)
                                 : machine()->Store(store_rep);
  NodeProperties::ChangeOp(node, store_op);
  return Changed(node);
}
 
Reduction MemoryLowering::ReduceStoreElement(Node* node,
                                             AllocationState const* state) {
  DCHECK_EQ(IrOpcode::kStoreElement, node->opcode());
  ElementAccess const& access = ElementAccessOf(node->op());
  Node* object = node->InputAt(0);
  Node* index = node->InputAt(1);
  Node* value = node->InputAt(2);
  node->ReplaceInput(1, ComputeIndex(access, index));
  WriteBarrierKind write_barrier_kind = ComputeWriteBarrierKind(
      node, object, value, state, access.write_barrier_kind);
  NodeProperties::ChangeOp(
      node, machine()->Store(StoreRepresentation(
                access.machine_type.representation(), write_barrier_kind)));
  return Changed(node);
}
 
Reduction MemoryLowering::ReduceStoreField(Node* node,
                                           AllocationState const* state) {
  DCHECK_EQ(IrOpcode::kStoreField, node->opcode());
  FieldAccess const& access = FieldAccessOf(node->op());
  // External pointer must never be stored by optimized code when sandbox is
  // turned on
  DCHECK(!access.type.Is(Type::ExternalPointer()) || !V8_ENABLE_SANDBOX_BOOL);
  // SandboxedPointers are not currently stored by optimized code.
  DCHECK(!access.type.Is(Type::SandboxedPointer()));
  // Bounded size fields are not currently stored by optimized code.
  DCHECK(!access.is_bounded_size_access);
  MachineType machine_type = access.machine_type;
  Node* object = node->InputAt(0);
  Node* value = node->InputAt(1);
 
  Node* effect = NodeProperties::GetEffectInput(node);
  Node* control = NodeProperties::GetControlInput(node);
  __ InitializeEffectControl(effect, control);
 
  WriteBarrierKind write_barrier_kind = ComputeWriteBarrierKind(
      node, object, value, state, access.write_barrier_kind);
  Node* offset = __ IntPtrConstant(access.offset - access.tag());
  node->InsertInput(graph_zone(), 1, offset);
 
  if (machine_type.IsMapWord()) {
    machine_type = MachineType::TaggedPointer();
#ifdef V8_MAP_PACKING
    Node* mapword = __ PackMapWord(TNode<Map>::UncheckedCast(value));
    node->ReplaceInput(2, mapword);
#endif
  }
  if (machine_type.representation() ==
      MachineRepresentation::kIndirectPointer) {
    // Indirect pointer stores require knowledge of the indirect pointer tag of
    // the field. This is technically only required for stores that need a
    // write barrier, but currently we track the tag for all such stores.
    DCHECK_NE(access.indirect_pointer_tag, kIndirectPointerNullTag);
    Node* tag = __ IntPtrConstant(access.indirect_pointer_tag);
    node->InsertInput(graph_zone(), 3, tag);
    NodeProperties::ChangeOp(
        node, machine()->StoreIndirectPointer(write_barrier_kind));
  } else {
    NodeProperties::ChangeOp(
        node, machine()->Store(StoreRepresentation(
                  machine_type.representation(), write_barrier_kind)));
  }
  return Changed(node);
}
 
Reduction MemoryLowering::ReduceStore(Node* node,
                                      AllocationState const* state) {
  DCHECK_EQ(IrOpcode::kStore, node->opcode());
  StoreRepresentation representation = StoreRepresentationOf(node->op());
  Node* object = node->InputAt(0);
  Node* value = node->InputAt(2);
  WriteBarrierKind write_barrier_kind = ComputeWriteBarrierKind(
      node, object, value, state, representation.write_barrier_kind());
  if (write_barrier_kind != representation.write_barrier_kind()) {
    NodeProperties::ChangeOp(
        node, machine()->Store(StoreRepresentation(
                  representation.representation(), write_barrier_kind)));
    return Changed(node);
  }
  return NoChange();
}
 
Node* MemoryLowering::ComputeIndex(ElementAccess const& access, Node* index) {
  int const element_size_shift =
      ElementSizeLog2Of(access.machine_type.representation());
  if (element_size_shift) {
    index = __ WordShl(index, __ IntPtrConstant(element_size_shift));
  }
  int const fixed_offset = access.header_size - access.tag();
  if (fixed_offset) {
    index = __ IntAdd(index, __ IntPtrConstant(fixed_offset));
  }
  return index;
}
 
#undef __
 
namespace {
 
bool ValueNeedsWriteBarrier(Node* value, Isolate* isolate) {
  switch (value->opcode()) {
    case IrOpcode::kBitcastWordToTaggedSigned:
      return false;
    case IrOpcode::kHeapConstant: {
      RootIndex root_index;
      if (isolate->roots_table().IsRootHandle(HeapConstantOf(value->op()),
                                              &root_index) &&
          RootsTable::IsImmortalImmovable(root_index)) {
        return false;
      }
      break;
    }
    default:
      break;
  }
  return true;
}
 
}  // namespace
 
Reduction MemoryLowering::ReduceAllocateRaw(Node* node) {
  DCHECK_EQ(IrOpcode::kAllocateRaw, node->opcode());
  const AllocateParameters& allocation = AllocateParametersOf(node->op());
  return ReduceAllocateRaw(node, allocation.allocation_type(), nullptr);
}
 
WriteBarrierKind MemoryLowering::ComputeWriteBarrierKind(
    Node* node, Node* object, Node* value, AllocationState const* state,
    WriteBarrierKind write_barrier_kind) {
  if (state && state->IsYoungGenerationAllocation() &&
      state->group()->Contains(object)) {
    write_barrier_kind = kNoWriteBarrier;
  }
  if (!ValueNeedsWriteBarrier(value, isolate())) {
    write_barrier_kind = kNoWriteBarrier;
  }
  if (v8_flags.disable_write_barriers) {
    write_barrier_kind = kNoWriteBarrier;
  }
  if (write_barrier_kind == WriteBarrierKind::kAssertNoWriteBarrier) {
    write_barrier_assert_failed_(node, object, function_debug_name_, zone());
  }
  return write_barrier_kind;
}
 
MemoryLowering::AllocationGroup::AllocationGroup(Node* node,
                                                 AllocationType allocation,
                                                 Zone* zone)
    : node_ids_(zone),
      allocation_(CheckAllocationType(allocation)),
      size_(nullptr) {
  node_ids_.insert(node->id());
}
 
MemoryLowering::AllocationGroup::AllocationGroup(Node* node,
                                                 AllocationType allocation,
                                                 Node* size, Zone* zone)
    : node_ids_(zone),
      allocation_(CheckAllocationType(allocation)),
      size_(size) {
  node_ids_.insert(node->id());
}
 
void MemoryLowering::AllocationGroup::Add(Node* node) {
  node_ids_.insert(node->id());
}
 
bool MemoryLowering::AllocationGroup::Contains(Node* node) const {
  // Additions should stay within the same allocated object, so it's safe to
  // ignore them.
  while (node_ids_.find(node->id()) == node_ids_.end()) {
    switch (node->opcode()) {
      case IrOpcode::kBitcastTaggedToWord:
      case IrOpcode::kBitcastWordToTagged:
      case IrOpcode::kInt32Add:
      case IrOpcode::kInt64Add:
        node = NodeProperties::GetValueInput(node, 0);
        break;
      default:
        return false;
    }
  }
  return true;
}
 
MemoryLowering::AllocationState::AllocationState()
    : group_(nullptr),
      size_(std::numeric_limits<int>::max()),
      top_(nullptr),
      effect_(nullptr) {}
 
MemoryLowering::AllocationState::AllocationState(AllocationGroup* group,
                                                 Node* effect)
    : group_(group),
      size_(std::numeric_limits<int>::max()),
      top_(nullptr),
      effect_(effect) {}
 
MemoryLowering::AllocationState::AllocationState(AllocationGroup* group,
                                                 intptr_t size, Node* top,
                                                 Node* effect)
    : group_(group), size_(size), top_(top), effect_(effect) {}
 
bool MemoryLowering::AllocationState::IsYoungGenerationAllocation() const {
  return group() && group()->IsYoungGenerationAllocation();
}
 
}  // namespace compiler
}  // namespace internal
}  // namespace v8