%PDF-
%PDF-
Mini Shell
Mini Shell
// Copyright 2012 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/heap/heap.h"
#include <atomic>
#include <cinttypes>
#include <iomanip>
#include <memory>
#include <unordered_map>
#include <unordered_set>
#include "include/v8-locker.h"
#include "src/api/api-inl.h"
#include "src/base/bits.h"
#include "src/base/flags.h"
#include "src/base/logging.h"
#include "src/base/macros.h"
#include "src/base/once.h"
#include "src/base/optional.h"
#include "src/base/platform/memory.h"
#include "src/base/platform/mutex.h"
#include "src/base/utils/random-number-generator.h"
#include "src/builtins/accessors.h"
#include "src/codegen/assembler-inl.h"
#include "src/codegen/compilation-cache.h"
#include "src/common/assert-scope.h"
#include "src/common/globals.h"
#include "src/compiler-dispatcher/optimizing-compile-dispatcher.h"
#include "src/debug/debug.h"
#include "src/deoptimizer/deoptimizer.h"
#include "src/execution/embedder-state.h"
#include "src/execution/isolate-utils-inl.h"
#include "src/execution/microtask-queue.h"
#include "src/execution/v8threads.h"
#include "src/execution/vm-state-inl.h"
#include "src/flags/flags.h"
#include "src/handles/global-handles-inl.h"
#include "src/handles/traced-handles.h"
#include "src/heap/allocation-observer.h"
#include "src/heap/array-buffer-sweeper.h"
#include "src/heap/base/stack.h"
#include "src/heap/base/worklist.h"
#include "src/heap/basic-memory-chunk.h"
#include "src/heap/code-range.h"
#include "src/heap/code-stats.h"
#include "src/heap/collection-barrier.h"
#include "src/heap/combined-heap.h"
#include "src/heap/concurrent-allocator.h"
#include "src/heap/concurrent-marking.h"
#include "src/heap/cppgc-js/cpp-heap.h"
#include "src/heap/ephemeron-remembered-set.h"
#include "src/heap/evacuation-verifier-inl.h"
#include "src/heap/finalization-registry-cleanup-task.h"
#include "src/heap/gc-callbacks.h"
#include "src/heap/gc-idle-time-handler.h"
#include "src/heap/gc-tracer-inl.h"
#include "src/heap/gc-tracer.h"
#include "src/heap/heap-allocator.h"
#include "src/heap/heap-controller.h"
#include "src/heap/heap-layout-tracer.h"
#include "src/heap/heap-write-barrier-inl.h"
#include "src/heap/incremental-marking-inl.h"
#include "src/heap/incremental-marking.h"
#include "src/heap/large-spaces.h"
#include "src/heap/local-heap.h"
#include "src/heap/mark-compact-inl.h"
#include "src/heap/mark-compact.h"
#include "src/heap/marking-barrier-inl.h"
#include "src/heap/marking-barrier.h"
#include "src/heap/marking-state-inl.h"
#include "src/heap/marking-state.h"
#include "src/heap/memory-balancer.h"
#include "src/heap/memory-chunk-inl.h"
#include "src/heap/memory-chunk-layout.h"
#include "src/heap/memory-measurement.h"
#include "src/heap/memory-reducer.h"
#include "src/heap/minor-gc-job.h"
#include "src/heap/minor-mark-sweep.h"
#include "src/heap/new-spaces.h"
#include "src/heap/object-lock.h"
#include "src/heap/object-stats.h"
#include "src/heap/objects-visiting-inl.h"
#include "src/heap/objects-visiting.h"
#include "src/heap/paged-spaces-inl.h"
#include "src/heap/parked-scope.h"
#include "src/heap/pretenuring-handler.h"
#include "src/heap/read-only-heap.h"
#include "src/heap/remembered-set.h"
#include "src/heap/safepoint.h"
#include "src/heap/scavenger-inl.h"
#include "src/heap/stress-scavenge-observer.h"
#include "src/heap/sweeper.h"
#include "src/heap/zapping.h"
#include "src/init/bootstrapper.h"
#include "src/init/v8.h"
#include "src/interpreter/interpreter.h"
#include "src/logging/log.h"
#include "src/logging/runtime-call-stats-scope.h"
#include "src/numbers/conversions.h"
#include "src/objects/data-handler.h"
#include "src/objects/feedback-vector.h"
#include "src/objects/free-space-inl.h"
#include "src/objects/hash-table-inl.h"
#include "src/objects/hash-table.h"
#include "src/objects/instance-type.h"
#include "src/objects/maybe-object.h"
#include "src/objects/objects.h"
#include "src/objects/shared-function-info.h"
#include "src/objects/slots-atomic-inl.h"
#include "src/objects/slots-inl.h"
#include "src/objects/visitors.h"
#include "src/profiler/heap-profiler.h"
#include "src/regexp/regexp.h"
#include "src/snapshot/embedded/embedded-data.h"
#include "src/snapshot/serializer-deserializer.h"
#include "src/snapshot/snapshot.h"
#include "src/strings/string-stream.h"
#include "src/strings/unicode-decoder.h"
#include "src/strings/unicode-inl.h"
#include "src/tracing/trace-event.h"
#include "src/utils/utils-inl.h"
#include "src/utils/utils.h"
#ifdef V8_ENABLE_CONSERVATIVE_STACK_SCANNING
#include "src/heap/conservative-stack-visitor.h"
#endif // V8_ENABLE_CONSERVATIVE_STACK_SCANNING
// Has to be the last include (doesn't have include guards):
#include "src/objects/object-macros.h"
namespace v8 {
namespace internal {
#ifdef V8_ENABLE_THIRD_PARTY_HEAP
Isolate* Heap::GetIsolateFromWritableObject(Tagged<HeapObject> object) {
return reinterpret_cast<Isolate*>(
third_party_heap::Heap::GetIsolate(object.address()));
}
#endif
// These are outside the Heap class so they can be forward-declared
// in heap-write-barrier-inl.h.
bool Heap_PageFlagsAreConsistent(Tagged<HeapObject> object) {
return Heap::PageFlagsAreConsistent(object);
}
void Heap_CombinedGenerationalAndSharedBarrierSlow(Tagged<HeapObject> object,
Address slot,
Tagged<HeapObject> value) {
Heap::CombinedGenerationalAndSharedBarrierSlow(object, slot, value);
}
void Heap_CombinedGenerationalAndSharedEphemeronBarrierSlow(
Tagged<EphemeronHashTable> table, Address slot, Tagged<HeapObject> value) {
Heap::CombinedGenerationalAndSharedEphemeronBarrierSlow(table, slot, value);
}
void Heap_GenerationalBarrierForCodeSlow(Tagged<InstructionStream> host,
RelocInfo* rinfo,
Tagged<HeapObject> object) {
Heap::GenerationalBarrierForCodeSlow(host, rinfo, object);
}
void Heap::SetConstructStubCreateDeoptPCOffset(int pc_offset) {
DCHECK_EQ(Smi::zero(), construct_stub_create_deopt_pc_offset());
set_construct_stub_create_deopt_pc_offset(Smi::FromInt(pc_offset));
}
void Heap::SetConstructStubInvokeDeoptPCOffset(int pc_offset) {
DCHECK_EQ(Smi::zero(), construct_stub_invoke_deopt_pc_offset());
set_construct_stub_invoke_deopt_pc_offset(Smi::FromInt(pc_offset));
}
void Heap::SetInterpreterEntryReturnPCOffset(int pc_offset) {
DCHECK_EQ(Smi::zero(), interpreter_entry_return_pc_offset());
set_interpreter_entry_return_pc_offset(Smi::FromInt(pc_offset));
}
void Heap::SetSerializedObjects(Tagged<FixedArray> objects) {
DCHECK(isolate()->serializer_enabled());
set_serialized_objects(objects);
}
void Heap::SetSerializedGlobalProxySizes(Tagged<FixedArray> sizes) {
DCHECK(isolate()->serializer_enabled());
set_serialized_global_proxy_sizes(sizes);
}
void Heap::SetBasicBlockProfilingData(Handle<ArrayList> list) {
set_basic_block_profiling_data(*list);
}
class ScheduleMinorGCTaskObserver final : public AllocationObserver {
public:
explicit ScheduleMinorGCTaskObserver(Heap* heap)
: AllocationObserver(kNotUsingFixedStepSize), heap_(heap) {
// Register GC callback for all atomic pause types.
heap_->main_thread_local_heap()->AddGCEpilogueCallback(
&GCEpilogueCallback, this, GCCallbacksInSafepoint::GCType::kLocal);
AddToNewSpace();
}
~ScheduleMinorGCTaskObserver() final {
RemoveFromNewSpace();
heap_->main_thread_local_heap()->RemoveGCEpilogueCallback(
&GCEpilogueCallback, this);
}
intptr_t GetNextStepSize() final {
size_t new_space_threshold =
MinorGCJob::YoungGenerationTaskTriggerSize(heap_);
size_t new_space_size = heap_->new_space()->Size();
if (new_space_size < new_space_threshold) {
return new_space_threshold - new_space_size;
}
// Force a step on next allocation.
return 1;
}
void Step(int, Address, size_t) final {
heap_->ScheduleMinorGCTaskIfNeeded();
// Remove this observer. It will be re-added after a GC.
DCHECK(was_added_to_space_);
heap_->allocator()->new_space_allocator()->RemoveAllocationObserver(this);
was_added_to_space_ = false;
}
protected:
static void GCEpilogueCallback(void* observer) {
reinterpret_cast<ScheduleMinorGCTaskObserver*>(observer)
->RemoveFromNewSpace();
reinterpret_cast<ScheduleMinorGCTaskObserver*>(observer)->AddToNewSpace();
}
void AddToNewSpace() {
DCHECK(!was_added_to_space_);
DCHECK_IMPLIES(v8_flags.minor_ms,
!heap_->allocator()->new_space_allocator()->IsLabValid());
heap_->allocator()->new_space_allocator()->AddAllocationObserver(this);
was_added_to_space_ = true;
}
void RemoveFromNewSpace() {
if (!was_added_to_space_) return;
heap_->allocator()->new_space_allocator()->RemoveAllocationObserver(this);
was_added_to_space_ = false;
}
Heap* heap_;
bool was_added_to_space_ = false;
};
Heap::Heap()
: isolate_(isolate()),
heap_allocator_(this),
memory_pressure_level_(MemoryPressureLevel::kNone),
safepoint_(std::make_unique<IsolateSafepoint>(this)),
external_string_table_(this),
allocation_type_for_in_place_internalizable_strings_(
isolate()->OwnsStringTables() ? AllocationType::kOld
: AllocationType::kSharedOld),
marking_state_(isolate_),
non_atomic_marking_state_(isolate_),
pretenuring_handler_(this) {
// Ensure old_generation_size_ is a multiple of kPageSize.
DCHECK_EQ(0, max_old_generation_size() & (Page::kPageSize - 1));
max_regular_code_object_size_ = MemoryChunkLayout::MaxRegularCodeObjectSize();
set_native_contexts_list(Smi::zero());
// Put a dummy entry in the remembered pages so we can find the list the
// minidump even if there are no real unmapped pages.
RememberUnmappedPage(kNullAddress, false);
}
Heap::~Heap() = default;
size_t Heap::MaxReserved() const {
const size_t kMaxNewLargeObjectSpaceSize = max_semi_space_size_;
return static_cast<size_t>(
(v8_flags.minor_ms ? 1 : 2) * max_semi_space_size_ +
kMaxNewLargeObjectSpaceSize + max_old_generation_size());
}
size_t Heap::YoungGenerationSizeFromOldGenerationSize(size_t old_generation) {
// Compute the semi space size and cap it.
bool is_low_memory = old_generation <= kOldGenerationLowMemory;
size_t semi_space;
if (v8_flags.minor_ms && !is_low_memory) {
semi_space = DefaultMaxSemiSpaceSize();
} else {
size_t ratio = is_low_memory ? OldGenerationToSemiSpaceRatioLowMemory()
: OldGenerationToSemiSpaceRatio();
semi_space = old_generation / ratio;
semi_space = std::min({semi_space, DefaultMaxSemiSpaceSize()});
semi_space = std::max({semi_space, DefaultMinSemiSpaceSize()});
semi_space = RoundUp(semi_space, Page::kPageSize);
}
return YoungGenerationSizeFromSemiSpaceSize(semi_space);
}
size_t Heap::HeapSizeFromPhysicalMemory(uint64_t physical_memory) {
// Compute the old generation size and cap it.
uint64_t old_generation = physical_memory /
kPhysicalMemoryToOldGenerationRatio *
kHeapLimitMultiplier;
old_generation =
std::min(old_generation,
static_cast<uint64_t>(MaxOldGenerationSize(physical_memory)));
old_generation =
std::max({old_generation, static_cast<uint64_t>(V8HeapTrait::kMinSize)});
old_generation = RoundUp(old_generation, Page::kPageSize);
size_t young_generation = YoungGenerationSizeFromOldGenerationSize(
static_cast<size_t>(old_generation));
return static_cast<size_t>(old_generation) + young_generation;
}
void Heap::GenerationSizesFromHeapSize(size_t heap_size,
size_t* young_generation_size,
size_t* old_generation_size) {
// Initialize values for the case when the given heap size is too small.
*young_generation_size = 0;
*old_generation_size = 0;
// Binary search for the largest old generation size that fits to the given
// heap limit considering the correspondingly sized young generation.
size_t lower = 0, upper = heap_size;
while (lower + 1 < upper) {
size_t old_generation = lower + (upper - lower) / 2;
size_t young_generation =
YoungGenerationSizeFromOldGenerationSize(old_generation);
if (old_generation + young_generation <= heap_size) {
// This size configuration fits into the given heap limit.
*young_generation_size = young_generation;
*old_generation_size = old_generation;
lower = old_generation;
} else {
upper = old_generation;
}
}
}
size_t Heap::MinYoungGenerationSize() {
return YoungGenerationSizeFromSemiSpaceSize(DefaultMinSemiSpaceSize());
}
size_t Heap::MinOldGenerationSize() {
size_t paged_space_count =
LAST_GROWABLE_PAGED_SPACE - FIRST_GROWABLE_PAGED_SPACE + 1;
return paged_space_count * Page::kPageSize;
}
size_t Heap::AllocatorLimitOnMaxOldGenerationSize() {
#ifdef V8_COMPRESS_POINTERS
// Isolate and the young generation are also allocated on the heap.
return kPtrComprCageReservationSize -
YoungGenerationSizeFromSemiSpaceSize(DefaultMaxSemiSpaceSize()) -
RoundUp(sizeof(Isolate), size_t{1} << kPageSizeBits);
#else
return std::numeric_limits<size_t>::max();
#endif
}
size_t Heap::MaxOldGenerationSize(uint64_t physical_memory) {
size_t max_size = V8HeapTrait::kMaxSize;
// Increase the heap size from 2GB to 4GB for 64-bit systems with physical
// memory at least 16GB. The theshold is set to 15GB to accomodate for some
// memory being reserved by the hardware.
constexpr bool x64_bit = Heap::kHeapLimitMultiplier >= 2;
if (v8_flags.huge_max_old_generation_size && x64_bit &&
(physical_memory / GB) >= 15) {
DCHECK_EQ(max_size / GB, 2u);
max_size *= 2;
}
return std::min(max_size, AllocatorLimitOnMaxOldGenerationSize());
}
namespace {
int NumberOfSemiSpaces() { return v8_flags.minor_ms ? 1 : 2; }
} // namespace
size_t Heap::YoungGenerationSizeFromSemiSpaceSize(size_t semi_space_size) {
return semi_space_size *
(NumberOfSemiSpaces() + kNewLargeObjectSpaceToSemiSpaceRatio);
}
size_t Heap::SemiSpaceSizeFromYoungGenerationSize(
size_t young_generation_size) {
return young_generation_size /
(NumberOfSemiSpaces() + kNewLargeObjectSpaceToSemiSpaceRatio);
}
size_t Heap::Capacity() {
if (!HasBeenSetUp()) return 0;
if (v8_flags.enable_third_party_heap) return tp_heap_->Capacity();
return NewSpaceCapacity() + OldGenerationCapacity();
}
size_t Heap::OldGenerationCapacity() const {
if (!HasBeenSetUp()) return 0;
PagedSpaceIterator spaces(this);
size_t total = 0;
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
total += space->Capacity();
}
if (shared_lo_space_) {
total += shared_lo_space_->SizeOfObjects();
}
return total + lo_space_->SizeOfObjects() + code_lo_space_->SizeOfObjects();
}
size_t Heap::CommittedOldGenerationMemory() {
if (!HasBeenSetUp()) return 0;
PagedSpaceIterator spaces(this);
size_t total = 0;
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
total += space->CommittedMemory();
}
if (shared_lo_space_) {
total += shared_lo_space_->Size();
}
return total + lo_space_->Size() + code_lo_space_->Size() +
trusted_lo_space_->Size();
}
size_t Heap::CommittedMemoryOfUnmapper() {
if (!HasBeenSetUp()) return 0;
return memory_allocator()->unmapper()->CommittedBufferedMemory();
}
size_t Heap::CommittedMemory() {
if (!HasBeenSetUp()) return 0;
size_t new_space_committed = new_space_ ? new_space_->CommittedMemory() : 0;
size_t new_lo_space_committed = new_lo_space_ ? new_lo_space_->Size() : 0;
return new_space_committed + new_lo_space_committed +
CommittedOldGenerationMemory();
}
size_t Heap::CommittedPhysicalMemory() {
if (!HasBeenSetUp()) return 0;
size_t total = 0;
for (SpaceIterator it(this); it.HasNext();) {
total += it.Next()->CommittedPhysicalMemory();
}
return total;
}
size_t Heap::CommittedMemoryExecutable() {
if (!HasBeenSetUp()) return 0;
return static_cast<size_t>(memory_allocator()->SizeExecutable());
}
void Heap::UpdateMaximumCommitted() {
if (!HasBeenSetUp()) return;
const size_t current_committed_memory = CommittedMemory();
if (current_committed_memory > maximum_committed_) {
maximum_committed_ = current_committed_memory;
}
}
size_t Heap::Available() {
if (!HasBeenSetUp()) return 0;
size_t total = 0;
for (SpaceIterator it(this); it.HasNext();) {
total += it.Next()->Available();
}
total += memory_allocator()->Available();
return total;
}
bool Heap::CanExpandOldGeneration(size_t size) const {
if (force_oom_ || force_gc_on_next_allocation_) return false;
if (OldGenerationCapacity() + size > max_old_generation_size()) return false;
// Stay below `MaxReserved()` such that it is more likely that committing the
// second semi space at the beginning of a GC succeeds.
return memory_allocator()->Size() + size <= MaxReserved();
}
bool Heap::IsOldGenerationExpansionAllowed(
size_t size, const base::MutexGuard& expansion_mutex_guard) const {
return OldGenerationCapacity() + size <= max_old_generation_size();
}
namespace {
bool IsIsolateDeserializationActive(LocalHeap* local_heap) {
return local_heap && !local_heap->heap()->deserialization_complete();
}
} // anonymous namespace
bool Heap::CanPromoteYoungAndExpandOldGeneration(size_t size) const {
size_t new_space_capacity = NewSpaceTargetCapacity();
size_t new_lo_space_capacity = new_lo_space_ ? new_lo_space_->Size() : 0;
// Over-estimate the new space size using capacity to allow some slack.
return CanExpandOldGeneration(size + new_space_capacity +
new_lo_space_capacity);
}
bool Heap::HasBeenSetUp() const {
// We will always have an old space when the heap is set up.
return old_space_ != nullptr;
}
bool Heap::ShouldUseBackgroundThreads() const {
return !v8_flags.single_threaded_gc_in_background ||
!isolate()->IsIsolateInBackground();
}
GarbageCollector Heap::SelectGarbageCollector(AllocationSpace space,
GarbageCollectionReason gc_reason,
const char** reason) const {
if (gc_reason == GarbageCollectionReason::kFinalizeConcurrentMinorMS) {
DCHECK(new_space());
*reason = "Concurrent MinorMS needs finalization";
return GarbageCollector::MINOR_MARK_SWEEPER;
}
// Is global GC requested?
if (space != NEW_SPACE && space != NEW_LO_SPACE) {
isolate_->counters()->gc_compactor_caused_by_request()->Increment();
*reason = "GC in old space requested";
return GarbageCollector::MARK_COMPACTOR;
}
if (v8_flags.gc_global || ShouldStressCompaction() || !new_space()) {
*reason = "GC in old space forced by flags";
return GarbageCollector::MARK_COMPACTOR;
}
if (incremental_marking()->IsMajorMarking() &&
incremental_marking()->IsMajorMarkingComplete() &&
AllocationLimitOvershotByLargeMargin()) {
*reason = "Incremental marking needs finalization";
return GarbageCollector::MARK_COMPACTOR;
}
if (v8_flags.separate_gc_phases && incremental_marking()->IsMajorMarking()) {
// TODO(v8:12503): Remove previous condition when flag gets removed.
*reason = "Incremental marking forced finalization";
return GarbageCollector::MARK_COMPACTOR;
}
if (!CanPromoteYoungAndExpandOldGeneration(0)) {
isolate_->counters()
->gc_compactor_caused_by_oldspace_exhaustion()
->Increment();
*reason = "scavenge might not succeed";
return GarbageCollector::MARK_COMPACTOR;
}
DCHECK(!v8_flags.single_generation);
DCHECK(!v8_flags.gc_global);
// Default
*reason = nullptr;
return YoungGenerationCollector();
}
void Heap::SetGCState(HeapState state) {
gc_state_.store(state, std::memory_order_relaxed);
}
bool Heap::IsGCWithStack() const {
return embedder_stack_state_ ==
cppgc::EmbedderStackState::kMayContainHeapPointers;
}
bool Heap::CanShortcutStringsDuringGC(GarbageCollector collector) const {
if (!v8_flags.shortcut_strings_with_stack && IsGCWithStack()) return false;
switch (collector) {
case GarbageCollector::MINOR_MARK_SWEEPER:
if (!v8_flags.minor_ms_shortcut_strings) return false;
DCHECK(!incremental_marking()->IsMajorMarking());
// Minor MS cannot short cut strings during concurrent marking.
if (incremental_marking()->IsMinorMarking()) return false;
// Minor MS uses static roots to check for strings to shortcut.
if (!V8_STATIC_ROOTS_BOOL) return false;
break;
case GarbageCollector::SCAVENGER:
// Scavenger cannot short cut strings during incremental marking.
if (incremental_marking()->IsMajorMarking()) return false;
if (isolate()->has_shared_space() &&
!isolate()->is_shared_space_isolate() &&
isolate()
->shared_space_isolate()
->heap()
->incremental_marking()
->IsMarking()) {
DCHECK(isolate()
->shared_space_isolate()
->heap()
->incremental_marking()
->IsMajorMarking());
return false;
}
break;
default:
UNREACHABLE();
}
return true;
}
void Heap::PrintShortHeapStatistics() {
if (!v8_flags.trace_gc_verbose) return;
PrintIsolate(isolate_,
"Memory allocator, used: %6zu KB,"
" available: %6zu KB\n",
memory_allocator()->Size() / KB,
memory_allocator()->Available() / KB);
PrintIsolate(isolate_,
"Read-only space, used: %6zu KB"
", available: %6zu KB"
", committed: %6zu KB\n",
read_only_space_->Size() / KB, size_t{0},
read_only_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"New space, used: %6zu KB"
", available: %6zu KB%s"
", committed: %6zu KB\n",
NewSpaceSize() / KB, new_space_->Available() / KB,
(v8_flags.minor_ms && minor_sweeping_in_progress()) ? "*" : "",
new_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"New large object space, used: %6zu KB"
", available: %6zu KB"
", committed: %6zu KB\n",
new_lo_space_->SizeOfObjects() / KB,
new_lo_space_->Available() / KB,
new_lo_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"Old space, used: %6zu KB"
", available: %6zu KB%s"
", committed: %6zu KB\n",
old_space_->SizeOfObjects() / KB, old_space_->Available() / KB,
sweeping_in_progress() ? "*" : "",
old_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"Code space, used: %6zu KB"
", available: %6zu KB%s"
", committed: %6zu KB\n",
code_space_->SizeOfObjects() / KB, code_space_->Available() / KB,
major_sweeping_in_progress() ? "*" : "",
code_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"Large object space, used: %6zu KB"
", available: %6zu KB"
", committed: %6zu KB\n",
lo_space_->SizeOfObjects() / KB, lo_space_->Available() / KB,
lo_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"Code large object space, used: %6zu KB"
", available: %6zu KB"
", committed: %6zu KB\n",
code_lo_space_->SizeOfObjects() / KB,
code_lo_space_->Available() / KB,
code_lo_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"Trusted space, used: %6zu KB"
", available: %6zu KB%s"
", committed: %6zu KB\n",
trusted_space_->SizeOfObjects() / KB,
trusted_space_->Available() / KB,
sweeping_in_progress() ? "*" : "",
trusted_space_->CommittedMemory() / KB);
PrintIsolate(isolate_,
"Trusted large object space, used: %6zu KB"
", available: %6zu KB"
", committed: %6zu KB\n",
trusted_lo_space_->SizeOfObjects() / KB,
trusted_lo_space_->Available() / KB,
trusted_lo_space_->CommittedMemory() / KB);
ReadOnlySpace* const ro_space = read_only_space_;
PrintIsolate(isolate_,
"All spaces, used: %6zu KB"
", available: %6zu KB%s"
", committed: %6zu KB\n",
(this->SizeOfObjects() + ro_space->Size()) / KB,
(this->Available()) / KB, sweeping_in_progress() ? "*" : "",
(this->CommittedMemory() + ro_space->CommittedMemory()) / KB);
PrintIsolate(isolate_,
"Unmapper buffering %zu chunks of committed: %6zu KB\n",
memory_allocator()->unmapper()->NumberOfCommittedChunks(),
CommittedMemoryOfUnmapper() / KB);
PrintIsolate(isolate_, "External memory reported: %6" PRId64 " KB\n",
external_memory_.total() / KB);
PrintIsolate(isolate_, "Backing store memory: %6" PRIu64 " KB\n",
backing_store_bytes() / KB);
PrintIsolate(isolate_, "External memory global %zu KB\n",
external_memory_callback_() / KB);
PrintIsolate(isolate_, "Total time spent in GC : %.1f ms\n",
total_gc_time_ms_.InMillisecondsF());
if (sweeping_in_progress()) {
PrintIsolate(isolate_,
"(*) Sweeping is still in progress, making available sizes "
"inaccurate.\n");
}
}
void Heap::PrintFreeListsStats() {
DCHECK(v8_flags.trace_gc_freelists);
if (v8_flags.trace_gc_freelists_verbose) {
PrintIsolate(isolate_,
"Freelists statistics per Page: "
"[category: length || total free bytes]\n");
}
std::vector<int> categories_lengths(
old_space()->free_list()->number_of_categories(), 0);
std::vector<size_t> categories_sums(
old_space()->free_list()->number_of_categories(), 0);
unsigned int pageCnt = 0;
// This loops computes freelists lengths and sum.
// If v8_flags.trace_gc_freelists_verbose is enabled, it also prints
// the stats of each FreeListCategory of each Page.
for (Page* page : *old_space()) {
std::ostringstream out_str;
if (v8_flags.trace_gc_freelists_verbose) {
out_str << "Page " << std::setw(4) << pageCnt;
}
for (int cat = kFirstCategory;
cat <= old_space()->free_list()->last_category(); cat++) {
FreeListCategory* free_list =
page->free_list_category(static_cast<FreeListCategoryType>(cat));
int length = free_list->FreeListLength();
size_t sum = free_list->SumFreeList();
if (v8_flags.trace_gc_freelists_verbose) {
out_str << "[" << cat << ": " << std::setw(4) << length << " || "
<< std::setw(6) << sum << " ]"
<< (cat == old_space()->free_list()->last_category() ? "\n"
: ", ");
}
categories_lengths[cat] += length;
categories_sums[cat] += sum;
}
if (v8_flags.trace_gc_freelists_verbose) {
PrintIsolate(isolate_, "%s", out_str.str().c_str());
}
pageCnt++;
}
// Print statistics about old_space (pages, free/wasted/used memory...).
PrintIsolate(
isolate_,
"%d pages. Free space: %.1f MB (waste: %.2f). "
"Usage: %.1f/%.1f (MB) -> %.2f%%.\n",
pageCnt, static_cast<double>(old_space_->Available()) / MB,
static_cast<double>(old_space_->Waste()) / MB,
static_cast<double>(old_space_->Size()) / MB,
static_cast<double>(old_space_->Capacity()) / MB,
static_cast<double>(old_space_->Size()) / old_space_->Capacity() * 100);
// Print global statistics of each FreeListCategory (length & sum).
PrintIsolate(isolate_,
"FreeLists global statistics: "
"[category: length || total free KB]\n");
std::ostringstream out_str;
for (int cat = kFirstCategory;
cat <= old_space()->free_list()->last_category(); cat++) {
out_str << "[" << cat << ": " << categories_lengths[cat] << " || "
<< std::fixed << std::setprecision(2)
<< static_cast<double>(categories_sums[cat]) / KB << " KB]"
<< (cat == old_space()->free_list()->last_category() ? "\n" : ", ");
}
PrintIsolate(isolate_, "%s", out_str.str().c_str());
}
void Heap::DumpJSONHeapStatistics(std::stringstream& stream) {
HeapStatistics stats;
reinterpret_cast<v8::Isolate*>(isolate())->GetHeapStatistics(&stats);
// clang-format off
#define DICT(s) "{" << s << "}"
#define LIST(s) "[" << s << "]"
#define QUOTE(s) "\"" << s << "\""
#define MEMBER(s) QUOTE(s) << ":"
auto SpaceStatistics = [this](int space_index) {
HeapSpaceStatistics space_stats;
reinterpret_cast<v8::Isolate*>(isolate())->GetHeapSpaceStatistics(
&space_stats, space_index);
std::stringstream stream;
stream << DICT(
MEMBER("name")
<< QUOTE(ToString(
static_cast<AllocationSpace>(space_index)))
<< ","
MEMBER("size") << space_stats.space_size() << ","
MEMBER("used_size") << space_stats.space_used_size() << ","
MEMBER("available_size") << space_stats.space_available_size() << ","
MEMBER("physical_size") << space_stats.physical_space_size());
return stream.str();
};
stream << DICT(
MEMBER("isolate") << QUOTE(reinterpret_cast<void*>(isolate())) << ","
MEMBER("id") << gc_count() << ","
MEMBER("time_ms") << isolate()->time_millis_since_init() << ","
MEMBER("total_heap_size") << stats.total_heap_size() << ","
MEMBER("total_heap_size_executable")
<< stats.total_heap_size_executable() << ","
MEMBER("total_physical_size") << stats.total_physical_size() << ","
MEMBER("total_available_size") << stats.total_available_size() << ","
MEMBER("used_heap_size") << stats.used_heap_size() << ","
MEMBER("heap_size_limit") << stats.heap_size_limit() << ","
MEMBER("malloced_memory") << stats.malloced_memory() << ","
MEMBER("external_memory") << stats.external_memory() << ","
MEMBER("peak_malloced_memory") << stats.peak_malloced_memory() << ","
MEMBER("spaces") << LIST(
SpaceStatistics(RO_SPACE) << "," <<
SpaceStatistics(NEW_SPACE) << "," <<
SpaceStatistics(OLD_SPACE) << "," <<
SpaceStatistics(CODE_SPACE) << "," <<
SpaceStatistics(LO_SPACE) << "," <<
SpaceStatistics(CODE_LO_SPACE) << "," <<
SpaceStatistics(NEW_LO_SPACE) << "," <<
SpaceStatistics(TRUSTED_SPACE) << "," <<
SpaceStatistics(TRUSTED_LO_SPACE)));
#undef DICT
#undef LIST
#undef QUOTE
#undef MEMBER
// clang-format on
}
void Heap::ReportStatisticsAfterGC() {
for (int i = 0; i < static_cast<int>(v8::Isolate::kUseCounterFeatureCount);
++i) {
isolate()->CountUsage(static_cast<v8::Isolate::UseCounterFeature>(i),
deferred_counters_[i]);
deferred_counters_[i] = 0;
}
}
class Heap::AllocationTrackerForDebugging final
: public HeapObjectAllocationTracker {
public:
static bool IsNeeded() {
return v8_flags.verify_predictable || v8_flags.fuzzer_gc_analysis ||
(v8_flags.trace_allocation_stack_interval > 0);
}
explicit AllocationTrackerForDebugging(Heap* heap) : heap_(heap) {
CHECK(IsNeeded());
heap_->AddHeapObjectAllocationTracker(this);
}
~AllocationTrackerForDebugging() final {
heap_->RemoveHeapObjectAllocationTracker(this);
if (v8_flags.verify_predictable || v8_flags.fuzzer_gc_analysis) {
PrintAllocationsHash();
}
}
void AllocationEvent(Address addr, int size) final {
if (v8_flags.verify_predictable) {
allocations_count_.fetch_add(1, std::memory_order_relaxed);
// Advance synthetic time by making a time request.
heap_->MonotonicallyIncreasingTimeInMs();
UpdateAllocationsHash(HeapObject::FromAddress(addr));
UpdateAllocationsHash(size);
if (allocations_count_ % v8_flags.dump_allocations_digest_at_alloc == 0) {
PrintAllocationsHash();
}
} else if (v8_flags.fuzzer_gc_analysis) {
allocations_count_.fetch_add(1, std::memory_order_relaxed);
} else if (v8_flags.trace_allocation_stack_interval > 0) {
allocations_count_.fetch_add(1, std::memory_order_relaxed);
if (allocations_count_ % v8_flags.trace_allocation_stack_interval == 0) {
heap_->isolate()->PrintStack(stdout, Isolate::kPrintStackConcise);
}
}
}
void MoveEvent(Address source, Address target, int size) final {
if (v8_flags.verify_predictable) {
allocations_count_.fetch_add(1, std::memory_order_relaxed);
// Advance synthetic time by making a time request.
heap_->MonotonicallyIncreasingTimeInMs();
UpdateAllocationsHash(HeapObject::FromAddress(source));
UpdateAllocationsHash(HeapObject::FromAddress(target));
UpdateAllocationsHash(size);
if (allocations_count_ % v8_flags.dump_allocations_digest_at_alloc == 0) {
PrintAllocationsHash();
}
} else if (v8_flags.fuzzer_gc_analysis) {
allocations_count_.fetch_add(1, std::memory_order_relaxed);
}
}
void UpdateObjectSizeEvent(Address, int) final {}
private:
void UpdateAllocationsHash(Tagged<HeapObject> object) {
Address object_address = object.address();
MemoryChunk* memory_chunk = MemoryChunk::FromAddress(object_address);
AllocationSpace allocation_space = memory_chunk->owner_identity();
static_assert(kSpaceTagSize + kPageSizeBits <= 32);
uint32_t value =
static_cast<uint32_t>(object_address - memory_chunk->address()) |
(static_cast<uint32_t>(allocation_space) << kPageSizeBits);
UpdateAllocationsHash(value);
}
void UpdateAllocationsHash(uint32_t value) {
const uint16_t c1 = static_cast<uint16_t>(value);
const uint16_t c2 = static_cast<uint16_t>(value >> 16);
raw_allocations_hash_ =
StringHasher::AddCharacterCore(raw_allocations_hash_, c1);
raw_allocations_hash_ =
StringHasher::AddCharacterCore(raw_allocations_hash_, c2);
}
void PrintAllocationsHash() {
uint32_t hash = StringHasher::GetHashCore(raw_allocations_hash_);
PrintF("\n### Allocations = %zu, hash = 0x%08x\n",
allocations_count_.load(std::memory_order_relaxed), hash);
}
Heap* const heap_;
// Count of all allocations performed through C++ bottlenecks. This needs to
// be atomic as objects are moved in parallel in the GC which counts as
// allocations.
std::atomic<size_t> allocations_count_{0};
// Running hash over allocations performed.
uint32_t raw_allocations_hash_ = 0;
};
void Heap::AddHeapObjectAllocationTracker(
HeapObjectAllocationTracker* tracker) {
if (allocation_trackers_.empty() && v8_flags.inline_new) {
DisableInlineAllocation();
}
allocation_trackers_.push_back(tracker);
if (allocation_trackers_.size() == 1) {
isolate_->UpdateLogObjectRelocation();
}
}
void Heap::RemoveHeapObjectAllocationTracker(
HeapObjectAllocationTracker* tracker) {
allocation_trackers_.erase(std::remove(allocation_trackers_.begin(),
allocation_trackers_.end(), tracker),
allocation_trackers_.end());
if (allocation_trackers_.empty()) {
isolate_->UpdateLogObjectRelocation();
}
if (allocation_trackers_.empty() && v8_flags.inline_new) {
EnableInlineAllocation();
}
}
void Heap::AddRetainingPathTarget(Handle<HeapObject> object,
RetainingPathOption option) {
if (!v8_flags.track_retaining_path) {
PrintF("Retaining path tracking requires --track-retaining-path\n");
} else {
Handle<WeakArrayList> array(retaining_path_targets(), isolate());
int index = array->length();
array = WeakArrayList::AddToEnd(isolate(), array,
MaybeObjectHandle::Weak(object));
set_retaining_path_targets(*array);
DCHECK_EQ(array->length(), index + 1);
retaining_path_target_option_[index] = option;
}
}
bool Heap::IsRetainingPathTarget(Tagged<HeapObject> object,
RetainingPathOption* option) {
Tagged<WeakArrayList> targets = retaining_path_targets();
int length = targets->length();
MaybeObject object_to_check = HeapObjectReference::Weak(object);
for (int i = 0; i < length; i++) {
MaybeObject target = targets->Get(i);
DCHECK(target->IsWeakOrCleared());
if (target == object_to_check) {
DCHECK(retaining_path_target_option_.count(i));
*option = retaining_path_target_option_[i];
return true;
}
}
return false;
}
void Heap::PrintRetainingPath(Tagged<HeapObject> target,
RetainingPathOption option) {
PrintF("\n\n\n");
PrintF("#################################################\n");
PrintF("Retaining path for %p:\n", reinterpret_cast<void*>(target.ptr()));
Tagged<HeapObject> object = target;
std::vector<std::pair<Tagged<HeapObject>, bool>> retaining_path;
base::Optional<Root> root;
bool ephemeron = false;
while (true) {
retaining_path.push_back(std::make_pair(object, ephemeron));
if (option == RetainingPathOption::kTrackEphemeronPath &&
ephemeron_retainer_.count(object)) {
object = ephemeron_retainer_[object];
ephemeron = true;
} else if (retainer_.count(object)) {
object = retainer_[object];
ephemeron = false;
} else {
if (retaining_root_.count(object)) {
root.emplace(retaining_root_[object]);
}
break;
}
}
int distance = static_cast<int>(retaining_path.size());
for (auto node : retaining_path) {
Tagged<HeapObject> node_object = node.first;
bool node_ephemeron = node.second;
PrintF("\n");
PrintF("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n");
PrintF("Distance from root %d%s: ", distance,
node_ephemeron ? " (ephemeron)" : "");
ShortPrint(*node_object);
PrintF("\n");
#ifdef OBJECT_PRINT
i::Print(*node_object);
PrintF("\n");
#endif
--distance;
}
PrintF("\n");
PrintF("^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n");
PrintF("Root: %s\n",
root.has_value() ? RootVisitor::RootName(root.value()) : "Unknown");
PrintF("-------------------------------------------------\n");
}
void UpdateRetainersMapAfterScavenge(
UnorderedHeapObjectMap<Tagged<HeapObject>>* map) {
// This is only used for Scavenger.
DCHECK(!v8_flags.minor_ms);
UnorderedHeapObjectMap<Tagged<HeapObject>> updated_map;
for (auto pair : *map) {
Tagged<HeapObject> object = pair.first;
Tagged<HeapObject> retainer = pair.second;
if (Heap::InFromPage(object)) {
MapWord map_word = object->map_word(kRelaxedLoad);
if (!map_word.IsForwardingAddress()) continue;
object = map_word.ToForwardingAddress(object);
}
if (Heap::InFromPage(retainer)) {
MapWord map_word = retainer->map_word(kRelaxedLoad);
if (!map_word.IsForwardingAddress()) continue;
retainer = map_word.ToForwardingAddress(retainer);
}
updated_map[object] = retainer;
}
*map = std::move(updated_map);
}
void Heap::UpdateRetainersAfterScavenge() {
if (!incremental_marking()->IsMarking()) return;
DCHECK(incremental_marking()->IsMajorMarking());
// This is only used for Scavenger.
DCHECK(!v8_flags.minor_ms);
UpdateRetainersMapAfterScavenge(&retainer_);
UpdateRetainersMapAfterScavenge(&ephemeron_retainer_);
UnorderedHeapObjectMap<Root> updated_retaining_root;
for (auto pair : retaining_root_) {
Tagged<HeapObject> object = pair.first;
if (Heap::InFromPage(object)) {
MapWord map_word = object->map_word(kRelaxedLoad);
if (!map_word.IsForwardingAddress()) continue;
object = map_word.ToForwardingAddress(object);
}
updated_retaining_root[object] = pair.second;
}
retaining_root_ = std::move(updated_retaining_root);
}
void Heap::AddRetainer(Tagged<HeapObject> retainer, Tagged<HeapObject> object) {
if (retainer_.count(object)) return;
retainer_[object] = retainer;
RetainingPathOption option = RetainingPathOption::kDefault;
if (IsRetainingPathTarget(object, &option)) {
// Check if the retaining path was already printed in
// AddEphemeronRetainer().
if (ephemeron_retainer_.count(object) == 0 ||
option == RetainingPathOption::kDefault) {
PrintRetainingPath(object, option);
}
}
}
void Heap::AddEphemeronRetainer(Tagged<HeapObject> retainer,
Tagged<HeapObject> object) {
if (ephemeron_retainer_.count(object)) return;
ephemeron_retainer_[object] = retainer;
RetainingPathOption option = RetainingPathOption::kDefault;
if (IsRetainingPathTarget(object, &option) &&
option == RetainingPathOption::kTrackEphemeronPath) {
// Check if the retaining path was already printed in AddRetainer().
if (retainer_.count(object) == 0) {
PrintRetainingPath(object, option);
}
}
}
void Heap::AddRetainingRoot(Root root, Tagged<HeapObject> object) {
if (retaining_root_.count(object)) return;
retaining_root_[object] = root;
RetainingPathOption option = RetainingPathOption::kDefault;
if (IsRetainingPathTarget(object, &option)) {
PrintRetainingPath(object, option);
}
}
void Heap::IncrementDeferredCount(v8::Isolate::UseCounterFeature feature) {
deferred_counters_[feature]++;
}
void Heap::GarbageCollectionPrologue(
GarbageCollectionReason gc_reason,
const v8::GCCallbackFlags gc_callback_flags) {
TRACE_GC(tracer(), GCTracer::Scope::HEAP_PROLOGUE);
is_current_gc_forced_ = gc_callback_flags & v8::kGCCallbackFlagForced ||
current_gc_flags_ & GCFlag::kForced ||
force_gc_on_next_allocation_;
is_current_gc_for_heap_profiler_ =
gc_reason == GarbageCollectionReason::kHeapProfiler;
if (force_gc_on_next_allocation_) force_gc_on_next_allocation_ = false;
#ifdef V8_ENABLE_ALLOCATION_TIMEOUT
heap_allocator_.UpdateAllocationTimeout();
#endif // V8_ENABLE_ALLOCATION_TIMEOUT
// There may be an allocation memento behind objects in new space. Upon
// evacuation of a non-full new space (or if we are on the last page) there
// may be uninitialized memory behind top. We fill the remainder of the page
// with a filler.
if (new_space()) {
new_space()->main_allocator()->MakeLinearAllocationAreaIterable();
DCHECK_NOT_NULL(minor_gc_job());
minor_gc_job()->CancelTaskIfScheduled();
}
// Reset GC statistics.
promoted_objects_size_ = 0;
previous_new_space_surviving_object_size_ = new_space_surviving_object_size_;
new_space_surviving_object_size_ = 0;
nodes_died_in_new_space_ = 0;
nodes_copied_in_new_space_ = 0;
nodes_promoted_ = 0;
UpdateMaximumCommitted();
#ifdef DEBUG
DCHECK(!AllowGarbageCollection::IsAllowed());
DCHECK_EQ(gc_state(), NOT_IN_GC);
if (v8_flags.gc_verbose) Print();
#endif // DEBUG
memory_allocator()->unmapper()->PrepareForGC();
}
void Heap::GarbageCollectionPrologueInSafepoint() {
TRACE_GC(tracer(), GCTracer::Scope::HEAP_PROLOGUE_SAFEPOINT);
gc_count_++;
DCHECK_EQ(ResizeNewSpaceMode::kNone, resize_new_space_mode_);
if (new_space_) {
UpdateNewSpaceAllocationCounter();
if (!v8_flags.minor_ms) {
resize_new_space_mode_ = ShouldResizeNewSpace();
// Pretenuring heuristics require that new space grows before pretenuring
// feedback is processed.
if (resize_new_space_mode_ == ResizeNewSpaceMode::kGrow) {
ExpandNewSpaceSize();
}
SemiSpaceNewSpace::From(new_space_)->ResetParkedAllocationBuffers();
}
}
}
void Heap::UpdateNewSpaceAllocationCounter() {
new_space_allocation_counter_ = NewSpaceAllocationCounter();
}
size_t Heap::NewSpaceAllocationCounter() {
return new_space_allocation_counter_ +
(new_space_ ? new_space()->AllocatedSinceLastGC() : 0);
}
size_t Heap::SizeOfObjects() {
size_t total = 0;
for (SpaceIterator it(this); it.HasNext();) {
total += it.Next()->SizeOfObjects();
}
return total;
}
size_t Heap::TotalGlobalHandlesSize() {
return isolate_->global_handles()->TotalSize() +
isolate_->traced_handles()->total_size_bytes();
}
size_t Heap::UsedGlobalHandlesSize() {
return isolate_->global_handles()->UsedSize() +
isolate_->traced_handles()->used_size_bytes();
}
void Heap::AddAllocationObserversToAllSpaces(
AllocationObserver* observer, AllocationObserver* new_space_observer) {
DCHECK(observer && new_space_observer);
allocator()->AddAllocationObserver(observer, new_space_observer);
}
void Heap::RemoveAllocationObserversFromAllSpaces(
AllocationObserver* observer, AllocationObserver* new_space_observer) {
DCHECK(observer && new_space_observer);
allocator()->RemoveAllocationObserver(observer, new_space_observer);
}
void Heap::PublishPendingAllocations() {
if (v8_flags.enable_third_party_heap) return;
if (new_space_) new_space_->main_allocator()->MarkLabStartInitialized();
PagedSpaceIterator spaces(this);
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
space->main_allocator()->MoveOriginalTopForward();
}
lo_space_->ResetPendingObject();
if (new_lo_space_) new_lo_space_->ResetPendingObject();
code_lo_space_->ResetPendingObject();
trusted_lo_space_->ResetPendingObject();
}
void Heap::DeoptMarkedAllocationSites() {
// TODO(hpayer): If iterating over the allocation sites list becomes a
// performance issue, use a cache data structure in heap instead.
ForeachAllocationSite(
allocation_sites_list(), [this](Tagged<AllocationSite> site) {
if (site->deopt_dependent_code()) {
DependentCode::MarkCodeForDeoptimization(
isolate_, site,
DependentCode::kAllocationSiteTenuringChangedGroup);
site->set_deopt_dependent_code(false);
}
});
Deoptimizer::DeoptimizeMarkedCode(isolate_);
}
static GCType GetGCTypeFromGarbageCollector(GarbageCollector collector) {
switch (collector) {
case GarbageCollector::MARK_COMPACTOR:
return kGCTypeMarkSweepCompact;
case GarbageCollector::SCAVENGER:
return kGCTypeScavenge;
case GarbageCollector::MINOR_MARK_SWEEPER:
return kGCTypeMinorMarkSweep;
default:
UNREACHABLE();
}
}
void Heap::GarbageCollectionEpilogueInSafepoint(GarbageCollector collector) {
if (collector == GarbageCollector::MARK_COMPACTOR) {
memory_pressure_level_.store(MemoryPressureLevel::kNone,
std::memory_order_relaxed);
if (v8_flags.stress_marking > 0) {
stress_marking_percentage_ = NextStressMarkingLimit();
}
}
TRACE_GC(tracer(), GCTracer::Scope::HEAP_EPILOGUE_SAFEPOINT);
{
// Allows handle derefs for all threads/isolates from this thread.
AllowHandleDereferenceAllThreads allow_all_handle_derefs;
safepoint()->IterateLocalHeaps([](LocalHeap* local_heap) {
local_heap->InvokeGCEpilogueCallbacksInSafepoint(
GCCallbacksInSafepoint::GCType::kLocal);
});
if (collector == GarbageCollector::MARK_COMPACTOR &&
isolate()->is_shared_space_isolate()) {
isolate()->global_safepoint()->IterateClientIsolates([](Isolate* client) {
client->heap()->safepoint()->IterateLocalHeaps(
[](LocalHeap* local_heap) {
local_heap->InvokeGCEpilogueCallbacksInSafepoint(
GCCallbacksInSafepoint::GCType::kShared);
});
});
}
}
#define UPDATE_COUNTERS_FOR_SPACE(space) \
isolate_->counters()->space##_bytes_available()->Set( \
static_cast<int>(space()->Available())); \
isolate_->counters()->space##_bytes_committed()->Set( \
static_cast<int>(space()->CommittedMemory())); \
isolate_->counters()->space##_bytes_used()->Set( \
static_cast<int>(space()->SizeOfObjects()));
#define UPDATE_FRAGMENTATION_FOR_SPACE(space) \
if (space()->CommittedMemory() > 0) { \
isolate_->counters()->external_fragmentation_##space()->AddSample( \
static_cast<int>(100 - (space()->SizeOfObjects() * 100.0) / \
space()->CommittedMemory())); \
}
#define UPDATE_COUNTERS_AND_FRAGMENTATION_FOR_SPACE(space) \
UPDATE_COUNTERS_FOR_SPACE(space) \
UPDATE_FRAGMENTATION_FOR_SPACE(space)
if (new_space()) {
UPDATE_COUNTERS_FOR_SPACE(new_space)
}
UPDATE_COUNTERS_AND_FRAGMENTATION_FOR_SPACE(old_space)
UPDATE_COUNTERS_AND_FRAGMENTATION_FOR_SPACE(code_space)
UPDATE_COUNTERS_AND_FRAGMENTATION_FOR_SPACE(lo_space)
#undef UPDATE_COUNTERS_FOR_SPACE
#undef UPDATE_FRAGMENTATION_FOR_SPACE
#undef UPDATE_COUNTERS_AND_FRAGMENTATION_FOR_SPACE
#ifdef DEBUG
if (v8_flags.print_global_handles) isolate_->global_handles()->Print();
if (v8_flags.print_handles) PrintHandles();
if (v8_flags.check_handle_count) CheckHandleCount();
#endif
if (new_space() && !v8_flags.minor_ms) {
SemiSpaceNewSpace* semi_space_new_space =
SemiSpaceNewSpace::From(new_space());
if (heap::ShouldZapGarbage() || v8_flags.clear_free_memory) {
semi_space_new_space->ZapUnusedMemory();
}
{
TRACE_GC(tracer(), GCTracer::Scope::HEAP_EPILOGUE_REDUCE_NEW_SPACE);
if (resize_new_space_mode_ == ResizeNewSpaceMode::kShrink) {
ReduceNewSpaceSize();
}
}
resize_new_space_mode_ = ResizeNewSpaceMode::kNone;
semi_space_new_space->MakeAllPagesInFromSpaceIterable();
}
// Ensure that unmapper task isn't running during full GC. We need access to
// those pages for accessing page flags when processing old-to-new slots.
DCHECK_IMPLIES(collector == GarbageCollector::MARK_COMPACTOR,
!memory_allocator()->unmapper()->IsRunning());
// Start concurrent unmapper tasks to free pages queued during GC.
memory_allocator()->unmapper()->FreeQueuedChunks();
// Remove CollectionRequested flag from main thread state, as the collection
// was just performed.
safepoint()->AssertActive();
LocalHeap::ThreadState old_state =
main_thread_local_heap()->state_.ClearCollectionRequested();
CHECK(old_state.IsRunning());
// Resume all threads waiting for the GC.
collection_barrier_->ResumeThreadsAwaitingCollection();
}
void Heap::GarbageCollectionEpilogue(GarbageCollector collector) {
TRACE_GC(tracer(), GCTracer::Scope::HEAP_EPILOGUE);
AllowGarbageCollection for_the_rest_of_the_epilogue;
UpdateMaximumCommitted();
if (v8_flags.track_retaining_path &&
collector == GarbageCollector::MARK_COMPACTOR) {
retainer_.clear();
ephemeron_retainer_.clear();
retaining_root_.clear();
}
isolate_->counters()->alive_after_last_gc()->Set(
static_cast<int>(SizeOfObjects()));
if (CommittedMemory() > 0) {
isolate_->counters()->external_fragmentation_total()->AddSample(
static_cast<int>(100 - (SizeOfObjects() * 100.0) / CommittedMemory()));
isolate_->counters()->heap_sample_total_committed()->AddSample(
static_cast<int>(CommittedMemory() / KB));
isolate_->counters()->heap_sample_total_used()->AddSample(
static_cast<int>(SizeOfObjects() / KB));
isolate_->counters()->heap_sample_code_space_committed()->AddSample(
static_cast<int>(code_space()->CommittedMemory() / KB));
isolate_->counters()->heap_sample_maximum_committed()->AddSample(
static_cast<int>(MaximumCommittedMemory() / KB));
}
#ifdef DEBUG
ReportStatisticsAfterGC();
if (v8_flags.code_stats) ReportCodeStatistics("After GC");
#endif // DEBUG
last_gc_time_ = MonotonicallyIncreasingTimeInMs();
}
GCCallbacksScope::GCCallbacksScope(Heap* heap) : heap_(heap) {
heap_->gc_callbacks_depth_++;
}
GCCallbacksScope::~GCCallbacksScope() { heap_->gc_callbacks_depth_--; }
bool GCCallbacksScope::CheckReenter() const {
return heap_->gc_callbacks_depth_ == 1;
}
void Heap::HandleGCRequest() {
if (IsStressingScavenge() && stress_scavenge_observer_->HasRequestedGC()) {
CollectGarbage(NEW_SPACE, GarbageCollectionReason::kTesting);
stress_scavenge_observer_->RequestedGCDone();
} else if (HighMemoryPressure()) {
CheckMemoryPressure();
} else if (CollectionRequested()) {
CheckCollectionRequested();
} else if (incremental_marking()->MajorCollectionRequested()) {
CollectAllGarbage(current_gc_flags_,
GarbageCollectionReason::kFinalizeMarkingViaStackGuard,
current_gc_callback_flags_);
} else if (minor_mark_sweep_collector()->gc_finalization_requsted()) {
CollectGarbage(NEW_SPACE,
GarbageCollectionReason::kFinalizeConcurrentMinorMS);
}
}
void Heap::ScheduleMinorGCTaskIfNeeded() {
DCHECK_NOT_NULL(minor_gc_job_);
minor_gc_job_->ScheduleTask();
}
namespace {
size_t MinorMSConcurrentMarkingTrigger(Heap* heap) {
return heap->new_space()->TotalCapacity() *
v8_flags.minor_ms_concurrent_marking_trigger / 100;
}
} // namespace
void Heap::StartMinorMSIncrementalMarkingIfNeeded() {
DCHECK(!incremental_marking()->IsMarking());
if (v8_flags.concurrent_minor_ms_marking && !IsTearingDown() &&
!ShouldOptimizeForLoadTime() && incremental_marking()->CanBeStarted() &&
V8_LIKELY(!v8_flags.gc_global) &&
(paged_new_space()->paged_space()->UsableCapacity() >=
v8_flags.minor_ms_min_new_space_capacity_for_concurrent_marking_mb *
MB) &&
new_space()->Size() >= MinorMSConcurrentMarkingTrigger(this)) {
StartIncrementalMarking(GCFlag::kNoFlags, GarbageCollectionReason::kTask,
kNoGCCallbackFlags,
GarbageCollector::MINOR_MARK_SWEEPER);
// Schedule a task for finalizing the GC if needed.
ScheduleMinorGCTaskIfNeeded();
}
}
void Heap::CollectAllGarbage(GCFlags gc_flags,
GarbageCollectionReason gc_reason,
const v8::GCCallbackFlags gc_callback_flags) {
current_gc_flags_ = gc_flags;
CollectGarbage(OLD_SPACE, gc_reason, gc_callback_flags);
current_gc_flags_ = GCFlag::kNoFlags;
}
namespace {
intptr_t CompareWords(int size, Tagged<HeapObject> a, Tagged<HeapObject> b) {
int slots = size / kTaggedSize;
DCHECK_EQ(a->Size(), size);
DCHECK_EQ(b->Size(), size);
Tagged_t* slot_a = reinterpret_cast<Tagged_t*>(a.address());
Tagged_t* slot_b = reinterpret_cast<Tagged_t*>(b.address());
for (int i = 0; i < slots; i++) {
if (*slot_a != *slot_b) {
return *slot_a - *slot_b;
}
slot_a++;
slot_b++;
}
return 0;
}
void ReportDuplicates(int size, std::vector<Tagged<HeapObject>>* objects) {
if (objects->size() == 0) return;
sort(objects->begin(), objects->end(),
[size](Tagged<HeapObject> a, Tagged<HeapObject> b) {
intptr_t c = CompareWords(size, a, b);
if (c != 0) return c < 0;
return a < b;
});
std::vector<std::pair<int, Tagged<HeapObject>>> duplicates;
Tagged<HeapObject> current = (*objects)[0];
int count = 1;
for (size_t i = 1; i < objects->size(); i++) {
if (CompareWords(size, current, (*objects)[i]) == 0) {
count++;
} else {
if (count > 1) {
duplicates.push_back(std::make_pair(count - 1, current));
}
count = 1;
current = (*objects)[i];
}
}
if (count > 1) {
duplicates.push_back(std::make_pair(count - 1, current));
}
int threshold = v8_flags.trace_duplicate_threshold_kb * KB;
sort(duplicates.begin(), duplicates.end());
for (auto it = duplicates.rbegin(); it != duplicates.rend(); ++it) {
int duplicate_bytes = it->first * size;
if (duplicate_bytes < threshold) break;
PrintF("%d duplicates of size %d each (%dKB)\n", it->first, size,
duplicate_bytes / KB);
PrintF("Sample object: ");
Print(it->second);
PrintF("============================\n");
}
}
} // anonymous namespace
void Heap::CollectAllAvailableGarbage(GarbageCollectionReason gc_reason) {
// Min and max number of attempts for GC. The method will continue with more
// GCs until the root set is stable.
static constexpr int kMaxNumberOfAttempts = 7;
static constexpr int kMinNumberOfAttempts = 2;
// Returns the number of roots. We assume stack layout is stable but global
// roots could change between GCs due to finalizers and weak callbacks.
const auto num_roots = [this]() {
size_t js_roots = 0;
js_roots += isolate()->global_handles()->handles_count();
js_roots += isolate()->eternal_handles()->handles_count();
size_t cpp_roots = 0;
if (auto* cpp_heap = CppHeap::From(cpp_heap_)) {
cpp_roots += cpp_heap->GetStrongPersistentRegion().NodesInUse();
cpp_roots +=
cpp_heap->GetStrongCrossThreadPersistentRegion().NodesInUse();
}
return js_roots + cpp_roots;
};
if (gc_reason == GarbageCollectionReason::kLastResort) {
InvokeNearHeapLimitCallback();
}
RCS_SCOPE(isolate(), RuntimeCallCounterId::kGC_Custom_AllAvailableGarbage);
// The optimizing compiler may be unnecessarily holding on to memory.
isolate()->AbortConcurrentOptimization(BlockingBehavior::kDontBlock);
isolate()->ClearSerializerData();
isolate()->compilation_cache()->Clear();
current_gc_flags_ =
GCFlag::kReduceMemoryFootprint |
(gc_reason == GarbageCollectionReason::kLowMemoryNotification
? GCFlag::kForced
: GCFlag::kNoFlags);
for (int attempt = 0; attempt < kMaxNumberOfAttempts; attempt++) {
const size_t roots_before = num_roots();
CollectGarbage(OLD_SPACE, gc_reason, kNoGCCallbackFlags);
if ((roots_before == num_roots()) &&
((attempt + 1) >= kMinNumberOfAttempts)) {
break;
}
}
current_gc_flags_ = GCFlag::kNoFlags;
EagerlyFreeExternalMemory();
if (v8_flags.trace_duplicate_threshold_kb) {
std::map<int, std::vector<Tagged<HeapObject>>> objects_by_size;
PagedSpaceIterator spaces(this);
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
PagedSpaceObjectIterator it(this, space);
for (Tagged<HeapObject> obj = it.Next(); !obj.is_null();
obj = it.Next()) {
objects_by_size[obj->Size()].push_back(obj);
}
}
{
LargeObjectSpaceObjectIterator it(lo_space());
for (Tagged<HeapObject> obj = it.Next(); !obj.is_null();
obj = it.Next()) {
objects_by_size[obj->Size()].push_back(obj);
}
}
for (auto it = objects_by_size.rbegin(); it != objects_by_size.rend();
++it) {
ReportDuplicates(it->first, &it->second);
}
}
if (gc_reason == GarbageCollectionReason::kLastResort &&
v8_flags.heap_snapshot_on_oom) {
isolate()->heap_profiler()->WriteSnapshotToDiskAfterGC();
}
}
void Heap::PreciseCollectAllGarbage(GCFlags gc_flags,
GarbageCollectionReason gc_reason,
const GCCallbackFlags gc_callback_flags) {
if (!incremental_marking()->IsStopped()) {
FinalizeIncrementalMarkingAtomically(gc_reason);
}
CollectAllGarbage(gc_flags, gc_reason, gc_callback_flags);
}
void Heap::ReportExternalMemoryPressure() {
const GCCallbackFlags kGCCallbackFlagsForExternalMemory =
static_cast<GCCallbackFlags>(
kGCCallbackFlagSynchronousPhantomCallbackProcessing |
kGCCallbackFlagCollectAllExternalMemory);
int64_t current = external_memory_.total();
int64_t baseline = external_memory_.low_since_mark_compact();
int64_t limit = external_memory_.limit();
TRACE_EVENT2(
"devtools.timeline,v8", "V8.ExternalMemoryPressure", "external_memory_mb",
static_cast<int>((current - baseline) / MB), "external_memory_limit_mb",
static_cast<int>((limit - baseline) / MB));
if (current > baseline + external_memory_hard_limit()) {
CollectAllGarbage(
GCFlag::kReduceMemoryFootprint,
GarbageCollectionReason::kExternalMemoryPressure,
static_cast<GCCallbackFlags>(kGCCallbackFlagCollectAllAvailableGarbage |
kGCCallbackFlagsForExternalMemory));
return;
}
if (incremental_marking()->IsStopped()) {
if (incremental_marking()->CanBeStarted()) {
StartIncrementalMarking(GCFlagsForIncrementalMarking(),
GarbageCollectionReason::kExternalMemoryPressure,
kGCCallbackFlagsForExternalMemory);
} else {
CollectAllGarbage(i::GCFlag::kNoFlags,
GarbageCollectionReason::kExternalMemoryPressure,
kGCCallbackFlagsForExternalMemory);
}
} else {
// Incremental marking is turned on and has already been started.
current_gc_callback_flags_ = static_cast<GCCallbackFlags>(
current_gc_callback_flags_ | kGCCallbackFlagsForExternalMemory);
incremental_marking()->AdvanceAndFinalizeIfNecessary();
}
}
int64_t Heap::external_memory_limit() { return external_memory_.limit(); }
Heap::DevToolsTraceEventScope::DevToolsTraceEventScope(Heap* heap,
const char* event_name,
const char* event_type)
: heap_(heap), event_name_(event_name) {
TRACE_EVENT_BEGIN2("devtools.timeline,v8", event_name_, "usedHeapSizeBefore",
heap_->SizeOfObjects(), "type", event_type);
}
Heap::DevToolsTraceEventScope::~DevToolsTraceEventScope() {
TRACE_EVENT_END1("devtools.timeline,v8", event_name_, "usedHeapSizeAfter",
heap_->SizeOfObjects());
}
namespace {
template <typename Callback>
void InvokeExternalCallbacks(Isolate* isolate, Callback callback) {
DCHECK(!AllowJavascriptExecution::IsAllowed(isolate));
AllowGarbageCollection allow_gc;
// Temporary override any embedder stack state as callbacks may create
// their own state on the stack and recursively trigger GC.
EmbedderStackStateScope embedder_scope(
isolate->heap(), EmbedderStackStateScope::kExplicitInvocation,
StackState::kMayContainHeapPointers);
VMState<EXTERNAL> callback_state(isolate);
callback();
}
size_t GlobalMemorySizeFromV8Size(size_t v8_size) {
const size_t kGlobalMemoryToV8Ratio = 2;
return std::min(static_cast<uint64_t>(std::numeric_limits<size_t>::max()),
static_cast<uint64_t>(v8_size) * kGlobalMemoryToV8Ratio);
}
} // anonymous namespace
void Heap::SetOldGenerationAndGlobalMaximumSize(
size_t max_old_generation_size) {
max_old_generation_size_.store(max_old_generation_size,
std::memory_order_relaxed);
max_global_memory_size_ = GlobalMemorySizeFromV8Size(max_old_generation_size);
}
void Heap::SetOldGenerationAndGlobalAllocationLimit(
size_t new_old_generation_allocation_limit,
size_t new_global_allocation_limit) {
CHECK_GE(new_global_allocation_limit, new_old_generation_allocation_limit);
old_generation_allocation_limit_.store(new_old_generation_allocation_limit,
std::memory_order_relaxed);
global_allocation_limit_ = new_global_allocation_limit;
old_generation_allocation_limit_configured_ = true;
}
void Heap::ResetOldGenerationAndGlobalAllocationLimit() {
SetOldGenerationAndGlobalAllocationLimit(
initial_old_generation_size_,
GlobalMemorySizeFromV8Size(initial_old_generation_size_));
old_generation_allocation_limit_configured_ = false;
}
void Heap::CollectGarbage(AllocationSpace space,
GarbageCollectionReason gc_reason,
const v8::GCCallbackFlags gc_callback_flags) {
if (V8_UNLIKELY(!deserialization_complete_)) {
// During isolate initialization heap always grows. GC is only requested
// if a new page allocation fails. In such a case we should crash with
// an out-of-memory instead of performing GC because the prologue/epilogue
// callbacks may see objects that are not yet deserialized.
CHECK(always_allocate());
FatalProcessOutOfMemory("GC during deserialization");
}
// CollectGarbage consists of three parts:
// 1. The prologue part which may execute callbacks. These callbacks may
// allocate and trigger another garbage collection.
// 2. The main garbage collection phase.
// 3. The epilogue part which may execute callbacks. These callbacks may
// allocate and trigger another garbage collection
// Part 1: Invoke all callbacks which should happen before the actual garbage
// collection is triggered. Note that these callbacks may trigger another
// garbage collection since they may allocate.
// JS execution is not allowed in any of the callbacks.
DisallowJavascriptExecution no_js(isolate());
DCHECK(AllowGarbageCollection::IsAllowed());
const char* collector_reason = nullptr;
const GarbageCollector collector =
SelectGarbageCollector(space, gc_reason, &collector_reason);
current_or_last_garbage_collector_ = collector;
if (collector == GarbageCollector::MARK_COMPACTOR &&
incremental_marking()->IsMinorMarking()) {
CollectGarbage(NEW_SPACE,
GarbageCollectionReason::kFinalizeConcurrentMinorMS);
}
const GCType gc_type = GetGCTypeFromGarbageCollector(collector);
// Prologue callbacks. These callbacks may trigger GC themselves and thus
// cannot be related exactly to garbage collection cycles.
//
// GCTracer scopes are managed by callees.
InvokeExternalCallbacks(isolate(), [this, gc_callback_flags, gc_type]() {
// Ensure that all pending phantom callbacks are invoked.
isolate()->global_handles()->InvokeSecondPassPhantomCallbacks();
// Prologue callbacks registered with Heap.
CallGCPrologueCallbacks(gc_type, gc_callback_flags,
GCTracer::Scope::HEAP_EXTERNAL_PROLOGUE);
});
// The main garbage collection phase.
//
// We need a stack marker at the top of all entry points to allow
// deterministic passes over the stack. E.g., a verifier that should only
// find a subset of references of the marker.
//
// TODO(chromium:1056170): Consider adding a component that keeps track
// of relevant GC stack regions where interesting pointers can be found.
stack().SetMarkerIfNeededAndCallback([this, collector, gc_reason,
collector_reason, gc_callback_flags]() {
DisallowGarbageCollection no_gc_during_gc;
size_t committed_memory_before =
collector == GarbageCollector::MARK_COMPACTOR
? CommittedOldGenerationMemory()
: 0;
tracer()->StartObservablePause(base::TimeTicks::Now());
VMState<GC> state(isolate());
DevToolsTraceEventScope devtools_trace_event_scope(
this, IsYoungGenerationCollector(collector) ? "MinorGC" : "MajorGC",
ToString(gc_reason));
GarbageCollectionPrologue(gc_reason, gc_callback_flags);
{
GCTracer::RecordGCPhasesInfo record_gc_phases_info(this, collector,
gc_reason);
base::Optional<TimedHistogramScope> histogram_timer_scope;
base::Optional<OptionalTimedHistogramScope>
histogram_timer_priority_scope;
TRACE_EVENT0("v8", record_gc_phases_info.trace_event_name());
if (record_gc_phases_info.type_timer()) {
histogram_timer_scope.emplace(record_gc_phases_info.type_timer(),
isolate_);
}
if (record_gc_phases_info.type_priority_timer()) {
histogram_timer_priority_scope.emplace(
record_gc_phases_info.type_priority_timer(), isolate_,
OptionalTimedHistogramScopeMode::TAKE_TIME);
}
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) {
tp_heap_->CollectGarbage();
} else {
PerformGarbageCollection(collector, gc_reason, collector_reason);
}
// Clear flags describing the current GC now that the current GC is
// complete. Do this before GarbageCollectionEpilogue() since that could
// trigger another unforced GC.
is_current_gc_forced_ = false;
is_current_gc_for_heap_profiler_ = false;
if (collector == GarbageCollector::MARK_COMPACTOR ||
collector == GarbageCollector::SCAVENGER) {
tracer()->RecordGCPhasesHistograms(record_gc_phases_info.mode());
}
}
GarbageCollectionEpilogue(collector);
if (collector == GarbageCollector::MARK_COMPACTOR &&
v8_flags.track_detached_contexts) {
isolate()->CheckDetachedContextsAfterGC();
}
if (collector == GarbageCollector::MARK_COMPACTOR) {
if (memory_reducer_ != nullptr) {
memory_reducer_->NotifyMarkCompact(committed_memory_before);
}
if (initial_max_old_generation_size_ < max_old_generation_size() &&
OldGenerationSizeOfObjects() <
initial_max_old_generation_size_threshold_) {
SetOldGenerationAndGlobalMaximumSize(initial_max_old_generation_size_);
}
}
tracer()->StopAtomicPause();
tracer()->StopObservablePause(collector, base::TimeTicks::Now());
// Young generation cycles finish atomically. It is important that
// StopObservablePause, and StopCycle are called in this
// order; the latter may replace the current event with that of an
// interrupted full cycle.
if (IsYoungGenerationCollector(collector)) {
tracer()->StopYoungCycleIfNeeded();
} else {
tracer()->StopFullCycleIfNeeded();
}
});
// Epilogue callbacks. These callbacks may trigger GC themselves and thus
// cannot be related exactly to garbage collection cycles.
//
// GCTracer scopes are managed by callees.
InvokeExternalCallbacks(isolate(), [this, gc_callback_flags, gc_type]() {
// Epilogue callbacks registered with Heap.
CallGCEpilogueCallbacks(gc_type, gc_callback_flags,
GCTracer::Scope::HEAP_EXTERNAL_EPILOGUE);
isolate()->global_handles()->PostGarbageCollectionProcessing(
gc_callback_flags);
});
if (collector == GarbageCollector::MARK_COMPACTOR &&
(gc_callback_flags & (kGCCallbackFlagForced |
kGCCallbackFlagCollectAllAvailableGarbage)) != 0) {
isolate()->CountUsage(v8::Isolate::kForcedGC);
}
// Start incremental marking for the next cycle. We do this only for scavenger
// to avoid a loop where mark-compact causes another mark-compact.
if (collector == GarbageCollector::SCAVENGER) {
DCHECK(!v8_flags.minor_ms);
StartIncrementalMarkingIfAllocationLimitIsReached(
GCFlagsForIncrementalMarking(),
kGCCallbackScheduleIdleGarbageCollection);
}
if (!CanExpandOldGeneration(0)) {
InvokeNearHeapLimitCallback();
if (!CanExpandOldGeneration(0)) {
if (v8_flags.heap_snapshot_on_oom) {
isolate()->heap_profiler()->WriteSnapshotToDiskAfterGC();
}
FatalProcessOutOfMemory("Reached heap limit");
}
}
}
int Heap::NotifyContextDisposed(bool has_dependent_context) {
if (!has_dependent_context) {
tracer()->ResetSurvivalEvents();
ResetOldGenerationAndGlobalAllocationLimit();
if (memory_reducer_) {
memory_reducer_->NotifyPossibleGarbage();
}
}
isolate()->AbortConcurrentOptimization(BlockingBehavior::kDontBlock);
if (!isolate()->context().is_null()) {
RemoveDirtyFinalizationRegistriesOnContext(isolate()->raw_native_context());
isolate()->raw_native_context()->set_retained_maps(
ReadOnlyRoots(this).empty_weak_array_list());
}
return ++contexts_disposed_;
}
void Heap::StartIncrementalMarking(GCFlags gc_flags,
GarbageCollectionReason gc_reason,
GCCallbackFlags gc_callback_flags,
GarbageCollector collector) {
DCHECK(incremental_marking()->IsStopped());
// Delay incremental marking start while concurrent sweeping still has work.
// This helps avoid large CompleteSweep blocks on the main thread when major
// incremental marking should be scheduled following a minor GC.
if (sweeper()->AreMinorSweeperTasksRunning()) return;
if (IsYoungGenerationCollector(collector)) {
CompleteSweepingYoung();
} else {
// Sweeping needs to be completed such that markbits are all cleared before
// starting marking again.
CompleteSweepingFull();
}
base::Optional<SafepointScope> safepoint_scope;
{
AllowGarbageCollection allow_shared_gc;
SafepointKind safepoint_kind = isolate()->is_shared_space_isolate()
? SafepointKind::kGlobal
: SafepointKind::kIsolate;
safepoint_scope.emplace(isolate(), safepoint_kind);
}
#ifdef DEBUG
VerifyCountersAfterSweeping();
#endif
std::vector<Isolate*> paused_clients;
if (isolate()->is_shared_space_isolate()) {
isolate()->global_safepoint()->IterateClientIsolates(
[collector, &paused_clients](Isolate* client) {
if (v8_flags.concurrent_marking &&
client->heap()->concurrent_marking()->Pause()) {
paused_clients.push_back(client);
}
if (collector == GarbageCollector::MARK_COMPACTOR) {
Sweeper* const client_sweeper = client->heap()->sweeper();
client_sweeper->ContributeAndWaitForPromotedPagesIteration();
}
});
}
// Now that sweeping is completed, we can start the next full GC cycle.
tracer()->StartCycle(collector, gc_reason, nullptr,
GCTracer::MarkingType::kIncremental);
current_gc_flags_ = gc_flags;
current_gc_callback_flags_ = gc_callback_flags;
incremental_marking()->Start(collector, gc_reason);
if (isolate()->is_shared_space_isolate()) {
for (Isolate* client : paused_clients) {
client->heap()->concurrent_marking()->Resume();
}
} else {
DCHECK(paused_clients.empty());
}
}
namespace {
void CompleteArrayBufferSweeping(Heap* heap) {
auto* array_buffer_sweeper = heap->array_buffer_sweeper();
if (array_buffer_sweeper->sweeping_in_progress()) {
auto* tracer = heap->tracer();
GCTracer::Scope::ScopeId scope_id;
switch (tracer->GetCurrentCollector()) {
case GarbageCollector::MINOR_MARK_SWEEPER:
scope_id = GCTracer::Scope::MINOR_MS_COMPLETE_SWEEP_ARRAY_BUFFERS;
break;
case GarbageCollector::SCAVENGER:
scope_id = GCTracer::Scope::SCAVENGER_COMPLETE_SWEEP_ARRAY_BUFFERS;
break;
case GarbageCollector::MARK_COMPACTOR:
scope_id = GCTracer::Scope::MC_COMPLETE_SWEEP_ARRAY_BUFFERS;
}
TRACE_GC_EPOCH_WITH_FLOW(
tracer, scope_id, ThreadKind::kMain,
array_buffer_sweeper->GetTraceIdForFlowEvent(scope_id),
TRACE_EVENT_FLAG_FLOW_IN);
array_buffer_sweeper->EnsureFinished();
}
}
} // namespace
void Heap::CompleteSweepingFull() {
EnsureSweepingCompleted(SweepingForcedFinalizationMode::kUnifiedHeap);
DCHECK(!sweeping_in_progress());
DCHECK_IMPLIES(cpp_heap(),
!CppHeap::From(cpp_heap())->sweeper().IsSweepingInProgress());
DCHECK(!tracer()->IsSweepingInProgress());
}
void Heap::StartIncrementalMarkingOnInterrupt() {
StartIncrementalMarkingIfAllocationLimitIsReached(
GCFlagsForIncrementalMarking(), kGCCallbackScheduleIdleGarbageCollection);
}
void Heap::StartIncrementalMarkingIfAllocationLimitIsReached(
GCFlags gc_flags, const GCCallbackFlags gc_callback_flags) {
if (v8_flags.separate_gc_phases && gc_callbacks_depth_ > 0) {
// Do not start incremental marking while invoking GC callbacks.
// Heap::CollectGarbage already decided which GC is going to be invoked. In
// case it chose a young-gen GC, starting an incremental full GC during
// callbacks would break the separate GC phases guarantee.
return;
}
if (incremental_marking()->IsStopped()) {
switch (IncrementalMarkingLimitReached()) {
case IncrementalMarkingLimit::kHardLimit:
StartIncrementalMarking(
gc_flags,
OldGenerationSpaceAvailable() <= NewSpaceTargetCapacity()
? GarbageCollectionReason::kAllocationLimit
: GarbageCollectionReason::kGlobalAllocationLimit,
gc_callback_flags);
break;
case IncrementalMarkingLimit::kSoftLimit:
if (auto* job = incremental_marking()->incremental_marking_job()) {
job->ScheduleTask();
}
break;
case IncrementalMarkingLimit::kFallbackForEmbedderLimit:
// This is a fallback case where no appropriate limits have been
// configured yet.
if (memory_reducer() != nullptr) {
memory_reducer()->NotifyPossibleGarbage();
}
break;
case IncrementalMarkingLimit::kNoLimit:
break;
}
}
}
void Heap::StartIncrementalMarkingIfAllocationLimitIsReachedBackground() {
// TODO(v8:13012): Consider finalizing minor incremental marking when we need
// to start a full GC.
if (incremental_marking()->IsMarking() ||
!incremental_marking()->CanBeStarted()) {
return;
}
const size_t old_generation_space_available = OldGenerationSpaceAvailable();
if (old_generation_space_available < NewSpaceTargetCapacity()) {
if (auto* job = incremental_marking()->incremental_marking_job()) {
job->ScheduleTask();
}
if (old_generation_space_available == 0) {
// Fulfil the role of IncrementalMarkingLimitReached() == kHardLimit and
// try to start incremental marking ASAP.
// TODO(dinfuehr): Unify this logic with the main thread
// IncrementalMarkingLimitReached() call.
isolate()->stack_guard()->RequestStartIncrementalMarking();
}
}
}
void Heap::MoveRange(Tagged<HeapObject> dst_object, const ObjectSlot dst_slot,
const ObjectSlot src_slot, int len,
WriteBarrierMode mode) {
DCHECK_NE(len, 0);
DCHECK_NE(dst_object->map(), ReadOnlyRoots(this).fixed_cow_array_map());
const ObjectSlot dst_end(dst_slot + len);
// Ensure no range overflow.
DCHECK(dst_slot < dst_end);
DCHECK(src_slot < src_slot + len);
if ((v8_flags.concurrent_marking && incremental_marking()->IsMarking()) ||
(v8_flags.minor_ms && sweeper()->IsIteratingPromotedPages())) {
if (dst_slot < src_slot) {
// Copy tagged values forward using relaxed load/stores that do not
// involve value decompression.
const AtomicSlot atomic_dst_end(dst_end);
AtomicSlot dst(dst_slot);
AtomicSlot src(src_slot);
while (dst < atomic_dst_end) {
*dst = *src;
++dst;
++src;
}
} else {
// Copy tagged values backwards using relaxed load/stores that do not
// involve value decompression.
const AtomicSlot atomic_dst_begin(dst_slot);
AtomicSlot dst(dst_slot + len - 1);
AtomicSlot src(src_slot + len - 1);
while (dst >= atomic_dst_begin) {
*dst = *src;
--dst;
--src;
}
}
} else {
MemMove(dst_slot.ToVoidPtr(), src_slot.ToVoidPtr(), len * kTaggedSize);
}
if (mode == SKIP_WRITE_BARRIER) return;
WriteBarrierForRange(dst_object, dst_slot, dst_end);
}
// Instantiate Heap::CopyRange() for ObjectSlot and MaybeObjectSlot.
template void Heap::CopyRange<ObjectSlot>(Tagged<HeapObject> dst_object,
ObjectSlot dst_slot,
ObjectSlot src_slot, int len,
WriteBarrierMode mode);
template void Heap::CopyRange<MaybeObjectSlot>(Tagged<HeapObject> dst_object,
MaybeObjectSlot dst_slot,
MaybeObjectSlot src_slot,
int len, WriteBarrierMode mode);
template <typename TSlot>
void Heap::CopyRange(Tagged<HeapObject> dst_object, const TSlot dst_slot,
const TSlot src_slot, int len, WriteBarrierMode mode) {
DCHECK_NE(len, 0);
DCHECK_NE(dst_object->map(), ReadOnlyRoots(this).fixed_cow_array_map());
const TSlot dst_end(dst_slot + len);
// Ensure ranges do not overlap.
DCHECK(dst_end <= src_slot || (src_slot + len) <= dst_slot);
if ((v8_flags.concurrent_marking && incremental_marking()->IsMarking()) ||
(v8_flags.minor_ms && sweeper()->IsIteratingPromotedPages())) {
// Copy tagged values using relaxed load/stores that do not involve value
// decompression.
const AtomicSlot atomic_dst_end(dst_end);
AtomicSlot dst(dst_slot);
AtomicSlot src(src_slot);
while (dst < atomic_dst_end) {
*dst = *src;
++dst;
++src;
}
} else {
MemCopy(dst_slot.ToVoidPtr(), src_slot.ToVoidPtr(), len * kTaggedSize);
}
if (mode == SKIP_WRITE_BARRIER) return;
WriteBarrierForRange(dst_object, dst_slot, dst_end);
}
bool Heap::CollectionRequested() {
return collection_barrier_->WasGCRequested();
}
void Heap::CollectGarbageForBackground(LocalHeap* local_heap) {
CHECK(local_heap->is_main_thread());
CollectAllGarbage(current_gc_flags_,
GarbageCollectionReason::kBackgroundAllocationFailure,
current_gc_callback_flags_);
}
void Heap::CheckCollectionRequested() {
if (!CollectionRequested()) return;
CollectAllGarbage(current_gc_flags_,
GarbageCollectionReason::kBackgroundAllocationFailure,
current_gc_callback_flags_);
}
#if V8_ENABLE_WEBASSEMBLY
void Heap::EnsureWasmCanonicalRttsSize(int length) {
HandleScope scope(isolate());
Handle<WeakArrayList> current_rtts = handle(wasm_canonical_rtts(), isolate_);
if (length <= current_rtts->length()) return;
Handle<WeakArrayList> new_rtts = WeakArrayList::EnsureSpace(
isolate(), current_rtts, length, AllocationType::kOld);
new_rtts->set_length(length);
set_wasm_canonical_rtts(*new_rtts);
// Wrappers are indexed by canonical rtt length, and an additional boolean
// storing whether the corresponding function is imported or not.
int required_wrapper_length = 2 * length;
Handle<WeakArrayList> current_wrappers =
handle(js_to_wasm_wrappers(), isolate_);
if (required_wrapper_length <= current_wrappers->length()) return;
Handle<WeakArrayList> new_wrappers =
WeakArrayList::EnsureSpace(isolate(), current_wrappers,
required_wrapper_length, AllocationType::kOld);
new_wrappers->set_length(required_wrapper_length);
set_js_to_wasm_wrappers(*new_wrappers);
}
#endif
void Heap::UpdateSurvivalStatistics(int start_new_space_size) {
if (start_new_space_size == 0) return;
promotion_ratio_ = (static_cast<double>(promoted_objects_size_) /
static_cast<double>(start_new_space_size) * 100);
if (previous_new_space_surviving_object_size_ > 0) {
promotion_rate_ =
(static_cast<double>(promoted_objects_size_) /
static_cast<double>(previous_new_space_surviving_object_size_) * 100);
} else {
promotion_rate_ = 0;
}
new_space_surviving_rate_ =
(static_cast<double>(new_space_surviving_object_size_) /
static_cast<double>(start_new_space_size) * 100);
double survival_rate = promotion_ratio_ + new_space_surviving_rate_;
tracer()->AddSurvivalRatio(survival_rate);
}
namespace {
GCTracer::Scope::ScopeId CollectorScopeId(GarbageCollector collector) {
switch (collector) {
case GarbageCollector::MARK_COMPACTOR:
return GCTracer::Scope::ScopeId::MARK_COMPACTOR;
case GarbageCollector::MINOR_MARK_SWEEPER:
return GCTracer::Scope::ScopeId::MINOR_MARK_SWEEPER;
case GarbageCollector::SCAVENGER:
return GCTracer::Scope::ScopeId::SCAVENGER;
}
UNREACHABLE();
}
void ClearStubCaches(Isolate* isolate) {
isolate->load_stub_cache()->Clear();
isolate->store_stub_cache()->Clear();
if (isolate->is_shared_space_isolate()) {
isolate->global_safepoint()->IterateClientIsolates([](Isolate* client) {
client->load_stub_cache()->Clear();
client->store_stub_cache()->Clear();
});
}
}
} // namespace
void Heap::PerformGarbageCollection(GarbageCollector collector,
GarbageCollectionReason gc_reason,
const char* collector_reason) {
if (IsYoungGenerationCollector(collector)) {
CompleteSweepingYoung();
if (v8_flags.verify_heap) {
// If heap verification is enabled, we want to ensure that sweeping is
// completed here, as it will be triggered from Heap::Verify anyway.
// In this way, sweeping finalization is accounted to the corresponding
// full GC cycle.
CompleteSweepingFull();
}
} else {
DCHECK_EQ(GarbageCollector::MARK_COMPACTOR, collector);
CompleteSweepingFull();
memory_allocator()->unmapper()->EnsureUnmappingCompleted();
}
const base::TimeTicks atomic_pause_start_time = base::TimeTicks::Now();
base::Optional<SafepointScope> safepoint_scope;
{
AllowGarbageCollection allow_shared_gc;
SafepointKind safepoint_kind = isolate()->is_shared_space_isolate()
? SafepointKind::kGlobal
: SafepointKind::kIsolate;
safepoint_scope.emplace(isolate(), safepoint_kind);
}
if (!incremental_marking_->IsMarking() ||
(collector == GarbageCollector::SCAVENGER)) {
tracer()->StartCycle(collector, gc_reason, collector_reason,
GCTracer::MarkingType::kAtomic);
}
tracer()->StartAtomicPause();
if ((!Heap::IsYoungGenerationCollector(collector) || v8_flags.minor_ms) &&
incremental_marking_->IsMarking()) {
DCHECK_IMPLIES(Heap::IsYoungGenerationCollector(collector),
incremental_marking_->IsMinorMarking());
tracer()->UpdateCurrentEvent(gc_reason, collector_reason);
}
DCHECK(tracer()->IsConsistentWithCollector(collector));
TRACE_GC_EPOCH(tracer(), CollectorScopeId(collector), ThreadKind::kMain);
collection_barrier_->StopTimeToCollectionTimer();
std::vector<Isolate*> paused_clients =
PauseConcurrentThreadsInClients(collector);
tracer()->StartInSafepoint(atomic_pause_start_time);
GarbageCollectionPrologueInSafepoint();
PerformHeapVerification();
if (new_space()) new_space()->Prologue();
size_t start_young_generation_size =
NewSpaceSize() + (new_lo_space() ? new_lo_space()->SizeOfObjects() : 0);
// Make sure allocation observers are disabled until the new new space
// capacity is set in the epilogue.
PauseAllocationObserversScope pause_observers(this);
size_t new_space_capacity_before_gc = NewSpaceTargetCapacity();
if (collector == GarbageCollector::MARK_COMPACTOR) {
MarkCompact();
} else if (collector == GarbageCollector::MINOR_MARK_SWEEPER) {
MinorMarkSweep();
} else {
DCHECK_EQ(GarbageCollector::SCAVENGER, collector);
Scavenge();
}
pretenuring_handler_.ProcessPretenuringFeedback(new_space_capacity_before_gc);
UpdateSurvivalStatistics(static_cast<int>(start_young_generation_size));
ShrinkOldGenerationAllocationLimitIfNotConfigured();
if (collector == GarbageCollector::SCAVENGER) {
// Objects that died in the new space might have been accounted
// as bytes marked ahead of schedule by the incremental marker.
incremental_marking()->UpdateMarkedBytesAfterScavenge(
start_young_generation_size - SurvivedYoungObjectSize());
}
isolate_->counters()->objs_since_last_young()->Set(0);
isolate_->eternal_handles()->PostGarbageCollectionProcessing();
// Update relocatables.
Relocatable::PostGarbageCollectionProcessing(isolate_);
if (isolate_->is_shared_space_isolate()) {
// Allows handle derefs for all threads/isolates from this thread.
AllowHandleDereferenceAllThreads allow_all_handle_derefs;
isolate()->global_safepoint()->IterateClientIsolates([](Isolate* client) {
Relocatable::PostGarbageCollectionProcessing(client);
});
}
// First round weak callbacks are not supposed to allocate and trigger
// nested GCs.
isolate_->global_handles()->InvokeFirstPassWeakCallbacks();
if (cpp_heap() && (collector == GarbageCollector::MARK_COMPACTOR ||
collector == GarbageCollector::MINOR_MARK_SWEEPER)) {
// TraceEpilogue may trigger operations that invalidate global handles. It
// has to be called *after* all other operations that potentially touch
// and reset global handles. It is also still part of the main garbage
// collection pause and thus needs to be called *before* any operation
// that can potentially trigger recursive garbage collections.
TRACE_GC(tracer(), GCTracer::Scope::HEAP_EMBEDDER_TRACING_EPILOGUE);
CppHeap::From(cpp_heap())->TraceEpilogue();
}
if (collector == GarbageCollector::MARK_COMPACTOR) {
ClearStubCaches(isolate());
}
PerformHeapVerification();
GarbageCollectionEpilogueInSafepoint(collector);
const base::TimeTicks atomic_pause_end_time = base::TimeTicks::Now();
tracer()->StopInSafepoint(atomic_pause_end_time);
ResumeConcurrentThreadsInClients(std::move(paused_clients));
RecomputeLimits(collector, atomic_pause_end_time);
// After every full GC the old generation allocation limit should be
// configured.
DCHECK_IMPLIES(collector == GarbageCollector::MARK_COMPACTOR,
old_generation_allocation_limit_configured_);
}
void Heap::PerformHeapVerification() {
HeapVerifier::VerifyHeapIfEnabled(this);
if (isolate()->is_shared_space_isolate()) {
isolate()->global_safepoint()->IterateClientIsolates([](Isolate* client) {
HeapVerifier::VerifyHeapIfEnabled(client->heap());
});
}
}
std::vector<Isolate*> Heap::PauseConcurrentThreadsInClients(
GarbageCollector collector) {
std::vector<Isolate*> paused_clients;
if (isolate()->is_shared_space_isolate()) {
isolate()->global_safepoint()->IterateClientIsolates(
[collector, &paused_clients](Isolate* client) {
CHECK(client->heap()->deserialization_complete());
if (v8_flags.concurrent_marking &&
client->heap()->concurrent_marking()->Pause()) {
paused_clients.push_back(client);
}
if (collector == GarbageCollector::MARK_COMPACTOR) {
Sweeper* const client_sweeper = client->heap()->sweeper();
client_sweeper->ContributeAndWaitForPromotedPagesIteration();
}
});
}
return paused_clients;
}
void Heap::ResumeConcurrentThreadsInClients(
std::vector<Isolate*> paused_clients) {
if (isolate()->is_shared_space_isolate()) {
for (Isolate* client : paused_clients) {
client->heap()->concurrent_marking()->Resume();
}
} else {
DCHECK(paused_clients.empty());
}
}
bool Heap::CollectGarbageShared(LocalHeap* local_heap,
GarbageCollectionReason gc_reason) {
CHECK(deserialization_complete());
DCHECK(isolate()->has_shared_space());
Isolate* shared_space_isolate = isolate()->shared_space_isolate();
return shared_space_isolate->heap()->CollectGarbageFromAnyThread(local_heap,
gc_reason);
}
bool Heap::CollectGarbageFromAnyThread(LocalHeap* local_heap,
GarbageCollectionReason gc_reason) {
DCHECK(local_heap->IsRunning());
if (isolate() == local_heap->heap()->isolate() &&
local_heap->is_main_thread()) {
CollectAllGarbage(current_gc_flags_, gc_reason, current_gc_callback_flags_);
return true;
} else {
if (!collection_barrier_->TryRequestGC()) return false;
const LocalHeap::ThreadState old_state =
main_thread_local_heap()->state_.SetCollectionRequested();
if (old_state.IsRunning()) {
const bool performed_gc =
collection_barrier_->AwaitCollectionBackground(local_heap);
return performed_gc;
} else {
DCHECK(old_state.IsParked());
return false;
}
}
}
void Heap::CompleteSweepingYoung() {
CompleteArrayBufferSweeping(this);
// If sweeping is in progress and there are no sweeper tasks running, finish
// the sweeping here, to avoid having to pause and resume during the young
// generation GC.
FinishSweepingIfOutOfWork();
if (v8_flags.minor_ms) {
EnsureYoungSweepingCompleted();
}
#if defined(CPPGC_YOUNG_GENERATION)
// Always complete sweeping if young generation is enabled.
if (cpp_heap()) {
if (auto* iheap = CppHeap::From(cpp_heap());
iheap->generational_gc_supported())
iheap->FinishSweepingIfRunning();
}
#endif // defined(CPPGC_YOUNG_GENERATION)
}
void Heap::EnsureSweepingCompletedForObject(Tagged<HeapObject> object) {
if (!sweeping_in_progress()) return;
BasicMemoryChunk* basic_chunk = BasicMemoryChunk::FromHeapObject(object);
if (basic_chunk->InReadOnlySpace()) return;
MemoryChunk* chunk = MemoryChunk::cast(basic_chunk);
if (chunk->SweepingDone()) return;
// SweepingDone() is always true for large pages.
DCHECK(!chunk->IsLargePage());
Page* page = Page::cast(chunk);
sweeper()->EnsurePageIsSwept(page);
}
void Heap::RecomputeLimits(GarbageCollector collector, base::TimeTicks time) {
if (!((collector == GarbageCollector::MARK_COMPACTOR) ||
(HasLowYoungGenerationAllocationRate() &&
old_generation_allocation_limit_configured_))) {
return;
}
double v8_gc_speed =
tracer()->CombinedMarkCompactSpeedInBytesPerMillisecond();
double v8_mutator_speed =
tracer()->CurrentOldGenerationAllocationThroughputInBytesPerMillisecond();
double v8_growing_factor = MemoryController<V8HeapTrait>::GrowingFactor(
this, max_old_generation_size(), v8_gc_speed, v8_mutator_speed);
double global_growing_factor = 0;
double embedder_gc_speed = tracer()->EmbedderSpeedInBytesPerMillisecond();
double embedder_speed =
tracer()->CurrentEmbedderAllocationThroughputInBytesPerMillisecond();
double embedder_growing_factor =
(embedder_gc_speed > 0 && embedder_speed > 0)
? MemoryController<GlobalMemoryTrait>::GrowingFactor(
this, max_global_memory_size_, embedder_gc_speed,
embedder_speed)
: 0;
global_growing_factor = std::max(v8_growing_factor, embedder_growing_factor);
size_t old_gen_size = OldGenerationSizeOfObjects();
size_t new_space_capacity = NewSpaceTargetCapacity();
HeapGrowingMode mode = CurrentHeapGrowingMode();
if (collector == GarbageCollector::MARK_COMPACTOR) {
external_memory_.ResetAfterGC();
size_t new_old_generation_allocation_limit =
MemoryController<V8HeapTrait>::CalculateAllocationLimit(
this, old_gen_size, min_old_generation_size_,
max_old_generation_size(), new_space_capacity, v8_growing_factor,
mode);
DCHECK_GT(global_growing_factor, 0);
size_t new_global_allocation_limit =
MemoryController<GlobalMemoryTrait>::CalculateAllocationLimit(
this, GlobalSizeOfObjects(), min_global_memory_size_,
max_global_memory_size_, new_space_capacity, global_growing_factor,
mode);
if (v8_flags.memory_balancer) {
// Now recompute the new allocation limit.
mb_->RecomputeLimits(
new_global_allocation_limit - new_old_generation_allocation_limit,
time);
} else {
SetOldGenerationAndGlobalAllocationLimit(
new_old_generation_allocation_limit, new_global_allocation_limit);
}
CheckIneffectiveMarkCompact(
old_gen_size, tracer()->AverageMarkCompactMutatorUtilization());
} else if (HasLowYoungGenerationAllocationRate() &&
old_generation_allocation_limit_configured_) {
size_t new_old_generation_allocation_limit =
MemoryController<V8HeapTrait>::CalculateAllocationLimit(
this, old_gen_size, min_old_generation_size_,
max_old_generation_size(), new_space_capacity, v8_growing_factor,
mode);
new_old_generation_allocation_limit = std::min(
new_old_generation_allocation_limit, old_generation_allocation_limit());
DCHECK_GT(global_growing_factor, 0);
size_t new_global_allocation_limit =
MemoryController<GlobalMemoryTrait>::CalculateAllocationLimit(
this, GlobalSizeOfObjects(), min_global_memory_size_,
max_global_memory_size_, new_space_capacity, global_growing_factor,
mode);
new_global_allocation_limit =
std::min(new_global_allocation_limit, global_allocation_limit_);
SetOldGenerationAndGlobalAllocationLimit(
new_old_generation_allocation_limit, new_global_allocation_limit);
}
CHECK_EQ(max_global_memory_size_,
GlobalMemorySizeFromV8Size(max_old_generation_size_));
CHECK_GE(global_allocation_limit_, old_generation_allocation_limit_);
}
void Heap::CallGCPrologueCallbacks(GCType gc_type, GCCallbackFlags flags,
GCTracer::Scope::ScopeId scope_id) {
if (gc_prologue_callbacks_.IsEmpty()) return;
GCCallbacksScope scope(this);
if (scope.CheckReenter()) {
RCS_SCOPE(isolate(), RuntimeCallCounterId::kGCPrologueCallback);
TRACE_GC(tracer(), scope_id);
HandleScope handle_scope(isolate());
gc_prologue_callbacks_.Invoke(gc_type, flags);
}
}
void Heap::CallGCEpilogueCallbacks(GCType gc_type, GCCallbackFlags flags,
GCTracer::Scope::ScopeId scope_id) {
if (gc_epilogue_callbacks_.IsEmpty()) return;
GCCallbacksScope scope(this);
if (scope.CheckReenter()) {
RCS_SCOPE(isolate(), RuntimeCallCounterId::kGCEpilogueCallback);
TRACE_GC(tracer(), scope_id);
HandleScope handle_scope(isolate());
gc_epilogue_callbacks_.Invoke(gc_type, flags);
}
}
void Heap::MarkCompact() {
SetGCState(MARK_COMPACT);
PROFILE(isolate_, CodeMovingGCEvent());
CodePageHeaderModificationScope code_modification(
"MarkCompact needs to access the marking bitmap in the page header");
UpdateOldGenerationAllocationCounter();
uint64_t size_of_objects_before_gc = SizeOfObjects();
mark_compact_collector()->Prepare();
ms_count_++;
contexts_disposed_ = 0;
MarkCompactPrologue();
mark_compact_collector()->CollectGarbage();
MarkCompactEpilogue();
if (v8_flags.allocation_site_pretenuring) {
EvaluateOldSpaceLocalPretenuring(size_of_objects_before_gc);
}
// This should be updated before PostGarbageCollectionProcessing, which
// can cause another GC. Take into account the objects promoted during
// GC.
old_generation_allocation_counter_at_last_gc_ +=
static_cast<size_t>(promoted_objects_size_);
old_generation_size_at_last_gc_ = OldGenerationSizeOfObjects();
global_memory_at_last_gc_ = GlobalSizeOfObjects();
}
void Heap::MinorMarkSweep() {
DCHECK(v8_flags.minor_ms);
CHECK_EQ(NOT_IN_GC, gc_state());
DCHECK(new_space());
DCHECK(!incremental_marking()->IsMajorMarking());
TRACE_GC(tracer(), GCTracer::Scope::MINOR_MS);
AlwaysAllocateScope always_allocate(this);
SetGCState(MINOR_MARK_SWEEP);
minor_mark_sweep_collector_->CollectGarbage();
SetGCState(NOT_IN_GC);
}
void Heap::MarkCompactEpilogue() {
TRACE_GC(tracer(), GCTracer::Scope::MC_EPILOGUE);
SetGCState(NOT_IN_GC);
isolate_->counters()->objs_since_last_full()->Set(0);
}
void Heap::MarkCompactPrologue() {
TRACE_GC(tracer(), GCTracer::Scope::MC_PROLOGUE);
isolate_->descriptor_lookup_cache()->Clear();
RegExpResultsCache::Clear(string_split_cache());
RegExpResultsCache::Clear(regexp_multiple_cache());
FlushNumberStringCache();
}
void Heap::Scavenge() {
DCHECK_NOT_NULL(new_space());
DCHECK_IMPLIES(v8_flags.separate_gc_phases,
!incremental_marking()->IsMarking());
if (v8_flags.trace_incremental_marking &&
!incremental_marking()->IsStopped()) {
isolate()->PrintWithTimestamp(
"[IncrementalMarking] Scavenge during marking.\n");
}
TRACE_GC(tracer(), GCTracer::Scope::SCAVENGER_SCAVENGE);
base::MutexGuard guard(relocation_mutex());
// Young generation garbage collection is orthogonal from full GC marking. It
// is possible that objects that are currently being processed for marking are
// reclaimed in the young generation GC that interleaves concurrent marking.
// Pause concurrent markers to allow processing them using
// `UpdateMarkingWorklistAfterYoungGenGC()`.
ConcurrentMarking::PauseScope pause_js_marking(concurrent_marking());
CppHeap::PauseConcurrentMarkingScope pause_cpp_marking(
CppHeap::From(cpp_heap_));
// There are soft limits in the allocation code, designed to trigger a mark
// sweep collection by failing allocations. There is no sense in trying to
// trigger one during scavenge: scavenges allocation should always succeed.
AlwaysAllocateScope scope(this);
// Bump-pointer allocations done during scavenge are not real allocations.
// Pause the inline allocation steps.
IncrementalMarking::PauseBlackAllocationScope pause_black_allocation(
incremental_marking());
SetGCState(SCAVENGE);
SemiSpaceNewSpace::From(new_space())->EvacuatePrologue();
// We also flip the young generation large object space. All large objects
// will be in the from space.
new_lo_space()->Flip();
new_lo_space()->ResetPendingObject();
// Implements Cheney's copying algorithm
scavenger_collector_->CollectGarbage();
SetGCState(NOT_IN_GC);
}
bool Heap::ExternalStringTable::Contains(Tagged<String> string) {
for (size_t i = 0; i < young_strings_.size(); ++i) {
if (young_strings_[i] == string) return true;
}
for (size_t i = 0; i < old_strings_.size(); ++i) {
if (old_strings_[i] == string) return true;
}
return false;
}
void Heap::UpdateExternalString(Tagged<String> string, size_t old_payload,
size_t new_payload) {
DCHECK(IsExternalString(string));
if (v8_flags.enable_third_party_heap) return;
Page* page = Page::FromHeapObject(string);
if (old_payload > new_payload) {
page->DecrementExternalBackingStoreBytes(
ExternalBackingStoreType::kExternalString, old_payload - new_payload);
} else {
page->IncrementExternalBackingStoreBytes(
ExternalBackingStoreType::kExternalString, new_payload - old_payload);
}
}
Tagged<String> Heap::UpdateYoungReferenceInExternalStringTableEntry(
Heap* heap, FullObjectSlot p) {
// This is only used for Scavenger.
DCHECK(!v8_flags.minor_ms);
PtrComprCageBase cage_base(heap->isolate());
Tagged<HeapObject> obj = HeapObject::cast(*p);
MapWord first_word = obj->map_word(cage_base, kRelaxedLoad);
Tagged<String> new_string;
if (InFromPage(obj)) {
if (!first_word.IsForwardingAddress()) {
// Unreachable external string can be finalized.
Tagged<String> string = Tagged<String>::cast(obj);
if (!IsExternalString(string, cage_base)) {
// Original external string has been internalized.
DCHECK(IsThinString(string, cage_base));
return Tagged<String>();
}
heap->FinalizeExternalString(string);
return Tagged<String>();
}
new_string = Tagged<String>::cast(first_word.ToForwardingAddress(obj));
} else {
new_string = Tagged<String>::cast(obj);
}
// String is still reachable.
if (IsThinString(new_string, cage_base)) {
// Filtering Thin strings out of the external string table.
return Tagged<String>();
} else if (IsExternalString(new_string, cage_base)) {
MemoryChunk::MoveExternalBackingStoreBytes(
ExternalBackingStoreType::kExternalString,
Page::FromAddress((*p).ptr()), Page::FromHeapObject(new_string),
ExternalString::cast(new_string)->ExternalPayloadSize());
return new_string;
}
// Internalization can replace external strings with non-external strings.
return IsExternalString(new_string, cage_base) ? new_string
: Tagged<String>();
}
void Heap::ExternalStringTable::VerifyYoung() {
#ifdef DEBUG
std::set<Tagged<String>> visited_map;
std::map<MemoryChunk*, size_t> size_map;
ExternalBackingStoreType type = ExternalBackingStoreType::kExternalString;
for (size_t i = 0; i < young_strings_.size(); ++i) {
Tagged<String> obj =
Tagged<String>::cast(Tagged<Object>(young_strings_[i]));
MemoryChunk* mc = MemoryChunk::FromHeapObject(obj);
DCHECK(mc->InYoungGeneration());
DCHECK(heap_->InYoungGeneration(obj));
DCHECK(!IsTheHole(obj, heap_->isolate()));
DCHECK(IsExternalString(obj));
// Note: we can have repeated elements in the table.
DCHECK_EQ(0, visited_map.count(obj));
visited_map.insert(obj);
size_map[mc] += Tagged<ExternalString>::cast(obj)->ExternalPayloadSize();
}
for (std::map<MemoryChunk*, size_t>::iterator it = size_map.begin();
it != size_map.end(); it++)
DCHECK_EQ(it->first->ExternalBackingStoreBytes(type), it->second);
#endif
}
void Heap::ExternalStringTable::Verify() {
#ifdef DEBUG
std::set<Tagged<String>> visited_map;
std::map<MemoryChunk*, size_t> size_map;
ExternalBackingStoreType type = ExternalBackingStoreType::kExternalString;
VerifyYoung();
for (size_t i = 0; i < old_strings_.size(); ++i) {
Tagged<String> obj = Tagged<String>::cast(Tagged<Object>(old_strings_[i]));
MemoryChunk* mc = MemoryChunk::FromHeapObject(obj);
DCHECK(!mc->InYoungGeneration());
DCHECK(!heap_->InYoungGeneration(obj));
DCHECK(!IsTheHole(obj, heap_->isolate()));
DCHECK(IsExternalString(obj));
// Note: we can have repeated elements in the table.
DCHECK_EQ(0, visited_map.count(obj));
visited_map.insert(obj);
size_map[mc] += Tagged<ExternalString>::cast(obj)->ExternalPayloadSize();
}
for (std::map<MemoryChunk*, size_t>::iterator it = size_map.begin();
it != size_map.end(); it++)
DCHECK_EQ(it->first->ExternalBackingStoreBytes(type), it->second);
#endif
}
void Heap::ExternalStringTable::UpdateYoungReferences(
Heap::ExternalStringTableUpdaterCallback updater_func) {
if (young_strings_.empty()) return;
FullObjectSlot start(young_strings_.data());
FullObjectSlot end(young_strings_.data() + young_strings_.size());
FullObjectSlot last = start;
for (FullObjectSlot p = start; p < end; ++p) {
Tagged<String> target = updater_func(heap_, p);
if (target.is_null()) continue;
DCHECK(IsExternalString(target));
if (InYoungGeneration(target)) {
// String is still in new space. Update the table entry.
last.store(target);
++last;
} else {
// String got promoted. Move it to the old string list.
old_strings_.push_back(target);
}
}
DCHECK(last <= end);
young_strings_.resize(last - start);
if (v8_flags.verify_heap) {
VerifyYoung();
}
}
void Heap::ExternalStringTable::PromoteYoung() {
old_strings_.reserve(old_strings_.size() + young_strings_.size());
std::move(std::begin(young_strings_), std::end(young_strings_),
std::back_inserter(old_strings_));
young_strings_.clear();
}
void Heap::ExternalStringTable::IterateYoung(RootVisitor* v) {
if (!young_strings_.empty()) {
v->VisitRootPointers(
Root::kExternalStringsTable, nullptr,
FullObjectSlot(young_strings_.data()),
FullObjectSlot(young_strings_.data() + young_strings_.size()));
}
}
void Heap::ExternalStringTable::IterateAll(RootVisitor* v) {
IterateYoung(v);
if (!old_strings_.empty()) {
v->VisitRootPointers(
Root::kExternalStringsTable, nullptr,
FullObjectSlot(old_strings_.data()),
FullObjectSlot(old_strings_.data() + old_strings_.size()));
}
}
void Heap::UpdateYoungReferencesInExternalStringTable(
ExternalStringTableUpdaterCallback updater_func) {
external_string_table_.UpdateYoungReferences(updater_func);
}
void Heap::ExternalStringTable::UpdateReferences(
Heap::ExternalStringTableUpdaterCallback updater_func) {
if (old_strings_.size() > 0) {
FullObjectSlot start(old_strings_.data());
FullObjectSlot end(old_strings_.data() + old_strings_.size());
for (FullObjectSlot p = start; p < end; ++p)
p.store(updater_func(heap_, p));
}
UpdateYoungReferences(updater_func);
}
void Heap::UpdateReferencesInExternalStringTable(
ExternalStringTableUpdaterCallback updater_func) {
external_string_table_.UpdateReferences(updater_func);
}
void Heap::ProcessAllWeakReferences(WeakObjectRetainer* retainer) {
ProcessNativeContexts(retainer);
ProcessAllocationSites(retainer);
ProcessDirtyJSFinalizationRegistries(retainer);
}
void Heap::ProcessNativeContexts(WeakObjectRetainer* retainer) {
Tagged<Object> head =
VisitWeakList<Context>(this, native_contexts_list(), retainer);
// Update the head of the list of contexts.
set_native_contexts_list(head);
}
void Heap::ProcessAllocationSites(WeakObjectRetainer* retainer) {
Tagged<Object> allocation_site_obj =
VisitWeakList<AllocationSite>(this, allocation_sites_list(), retainer);
set_allocation_sites_list(allocation_site_obj);
}
void Heap::ProcessDirtyJSFinalizationRegistries(WeakObjectRetainer* retainer) {
Tagged<Object> head = VisitWeakList<JSFinalizationRegistry>(
this, dirty_js_finalization_registries_list(), retainer);
set_dirty_js_finalization_registries_list(head);
// If the list is empty, set the tail to undefined. Otherwise the tail is set
// by WeakListVisitor<JSFinalizationRegistry>::VisitLiveObject.
if (IsUndefined(head, isolate())) {
set_dirty_js_finalization_registries_list_tail(head);
}
}
void Heap::ProcessWeakListRoots(WeakObjectRetainer* retainer) {
set_native_contexts_list(retainer->RetainAs(native_contexts_list()));
set_allocation_sites_list(retainer->RetainAs(allocation_sites_list()));
set_dirty_js_finalization_registries_list(
retainer->RetainAs(dirty_js_finalization_registries_list()));
set_dirty_js_finalization_registries_list_tail(
retainer->RetainAs(dirty_js_finalization_registries_list_tail()));
}
void Heap::ForeachAllocationSite(
Tagged<Object> list,
const std::function<void(Tagged<AllocationSite>)>& visitor) {
DisallowGarbageCollection no_gc;
Tagged<Object> current = list;
while (IsAllocationSite(current)) {
Tagged<AllocationSite> site = Tagged<AllocationSite>::cast(current);
visitor(site);
Tagged<Object> current_nested = site->nested_site();
while (IsAllocationSite(current_nested)) {
Tagged<AllocationSite> nested_site =
Tagged<AllocationSite>::cast(current_nested);
visitor(nested_site);
current_nested = nested_site->nested_site();
}
current = site->weak_next();
}
}
void Heap::ResetAllAllocationSitesDependentCode(AllocationType allocation) {
DisallowGarbageCollection no_gc_scope;
bool marked = false;
ForeachAllocationSite(
allocation_sites_list(),
[&marked, allocation, this](Tagged<AllocationSite> site) {
if (site->GetAllocationType() == allocation) {
site->ResetPretenureDecision();
site->set_deopt_dependent_code(true);
marked = true;
pretenuring_handler_.RemoveAllocationSitePretenuringFeedback(site);
return;
}
});
if (marked) isolate_->stack_guard()->RequestDeoptMarkedAllocationSites();
}
void Heap::EvaluateOldSpaceLocalPretenuring(
uint64_t size_of_objects_before_gc) {
uint64_t size_of_objects_after_gc = SizeOfObjects();
double old_generation_survival_rate =
(static_cast<double>(size_of_objects_after_gc) * 100) /
static_cast<double>(size_of_objects_before_gc);
if (old_generation_survival_rate < kOldSurvivalRateLowThreshold) {
// Too many objects died in the old generation, pretenuring of wrong
// allocation sites may be the cause for that. We have to deopt all
// dependent code registered in the allocation sites to re-evaluate
// our pretenuring decisions.
ResetAllAllocationSitesDependentCode(AllocationType::kOld);
if (v8_flags.trace_pretenuring) {
PrintF(
"Deopt all allocation sites dependent code due to low survival "
"rate in the old generation %f\n",
old_generation_survival_rate);
}
}
}
void Heap::VisitExternalResources(v8::ExternalResourceVisitor* visitor) {
DisallowGarbageCollection no_gc;
// All external strings are listed in the external string table.
class ExternalStringTableVisitorAdapter : public RootVisitor {
public:
explicit ExternalStringTableVisitorAdapter(
Isolate* isolate, v8::ExternalResourceVisitor* visitor)
: isolate_(isolate), visitor_(visitor) {}
void VisitRootPointers(Root root, const char* description,
FullObjectSlot start, FullObjectSlot end) override {
for (FullObjectSlot p = start; p < end; ++p) {
DCHECK(IsExternalString(*p));
visitor_->VisitExternalString(
Utils::ToLocal(Handle<String>(String::cast(*p), isolate_)));
}
}
private:
Isolate* isolate_;
v8::ExternalResourceVisitor* visitor_;
} external_string_table_visitor(isolate(), visitor);
external_string_table_.IterateAll(&external_string_table_visitor);
}
static_assert(IsAligned(FixedDoubleArray::kHeaderSize, kDoubleAlignment));
#ifdef V8_COMPRESS_POINTERS
// TODO(ishell, v8:8875): When pointer compression is enabled the kHeaderSize
// is only kTaggedSize aligned but we can keep using unaligned access since
// both x64 and arm64 architectures (where pointer compression supported)
// allow unaligned access to doubles.
static_assert(IsAligned(ByteArray::kHeaderSize, kTaggedSize));
#else
static_assert(IsAligned(ByteArray::kHeaderSize, kDoubleAlignment));
#endif
static_assert(!USE_ALLOCATION_ALIGNMENT_BOOL ||
(HeapNumber::kValueOffset & kDoubleAlignmentMask) == kTaggedSize);
int Heap::GetMaximumFillToAlign(AllocationAlignment alignment) {
if (V8_COMPRESS_POINTERS_8GB_BOOL) return 0;
switch (alignment) {
case kTaggedAligned:
return 0;
case kDoubleAligned:
case kDoubleUnaligned:
return kDoubleSize - kTaggedSize;
default:
UNREACHABLE();
}
}
// static
int Heap::GetFillToAlign(Address address, AllocationAlignment alignment) {
if (V8_COMPRESS_POINTERS_8GB_BOOL) return 0;
if (alignment == kDoubleAligned && (address & kDoubleAlignmentMask) != 0)
return kTaggedSize;
if (alignment == kDoubleUnaligned && (address & kDoubleAlignmentMask) == 0) {
return kDoubleSize - kTaggedSize; // No fill if double is always aligned.
}
return 0;
}
size_t Heap::GetCodeRangeReservedAreaSize() {
return CodeRange::GetWritableReservedAreaSize();
}
Tagged<HeapObject> Heap::PrecedeWithFiller(Tagged<HeapObject> object,
int filler_size) {
CreateFillerObjectAt(object.address(), filler_size);
return HeapObject::FromAddress(object.address() + filler_size);
}
Tagged<HeapObject> Heap::PrecedeWithFillerBackground(Tagged<HeapObject> object,
int filler_size) {
CreateFillerObjectAtBackground(object.address(), filler_size);
return HeapObject::FromAddress(object.address() + filler_size);
}
Tagged<HeapObject> Heap::AlignWithFillerBackground(
Tagged<HeapObject> object, int object_size, int allocation_size,
AllocationAlignment alignment) {
const int filler_size = allocation_size - object_size;
DCHECK_LT(0, filler_size);
const int pre_filler = GetFillToAlign(object.address(), alignment);
if (pre_filler) {
object = PrecedeWithFillerBackground(object, pre_filler);
}
DCHECK_LE(0, filler_size - pre_filler);
const int post_filler = filler_size - pre_filler;
if (post_filler) {
CreateFillerObjectAtBackground(object.address() + object_size, post_filler);
}
return object;
}
void* Heap::AllocateExternalBackingStore(
const std::function<void*(size_t)>& allocate, size_t byte_length) {
if (!always_allocate() && new_space()) {
size_t new_space_backing_store_bytes =
new_space()->ExternalBackingStoreOverallBytes();
if (new_space_backing_store_bytes >= 2 * DefaultMaxSemiSpaceSize() &&
new_space_backing_store_bytes >= byte_length) {
// Performing a young generation GC amortizes over the allocated backing
// store bytes and may free enough external bytes for this allocation.
CollectGarbage(NEW_SPACE,
GarbageCollectionReason::kExternalMemoryPressure);
}
}
void* result = allocate(byte_length);
if (result) return result;
if (!always_allocate()) {
for (int i = 0; i < 2; i++) {
CollectGarbage(OLD_SPACE,
GarbageCollectionReason::kExternalMemoryPressure);
result = allocate(byte_length);
if (result) return result;
}
CollectAllAvailableGarbage(
GarbageCollectionReason::kExternalMemoryPressure);
}
return allocate(byte_length);
}
// When old generation allocation limit is not configured (before the first full
// GC), this method shrinks the initial very large old generation size. This
// method can only shrink allocation limits but not increase it again.
void Heap::ShrinkOldGenerationAllocationLimitIfNotConfigured() {
if (!old_generation_allocation_limit_configured_ &&
tracer()->SurvivalEventsRecorded()) {
const size_t minimum_growing_step =
MemoryController<V8HeapTrait>::MinimumAllocationLimitGrowingStep(
CurrentHeapGrowingMode());
size_t new_old_generation_allocation_limit =
std::max(OldGenerationSizeOfObjects() + minimum_growing_step,
static_cast<size_t>(
static_cast<double>(old_generation_allocation_limit()) *
(tracer()->AverageSurvivalRatio() / 100)));
new_old_generation_allocation_limit = std::min(
new_old_generation_allocation_limit, old_generation_allocation_limit());
size_t new_global_allocation_limit = std::max(
GlobalSizeOfObjects() + minimum_growing_step,
static_cast<size_t>(static_cast<double>(global_allocation_limit_) *
(tracer()->AverageSurvivalRatio() / 100)));
new_global_allocation_limit =
std::min(new_global_allocation_limit, global_allocation_limit_);
SetOldGenerationAndGlobalAllocationLimit(
new_old_generation_allocation_limit, new_global_allocation_limit);
// We need to update limits but still remain in the "not configured" state.
// The first full GC will configure the heap.
old_generation_allocation_limit_configured_ = false;
}
}
void Heap::FlushNumberStringCache() {
// Flush the number to string cache.
int len = number_string_cache()->length();
for (int i = 0; i < len; i++) {
number_string_cache()->set_undefined(i);
}
}
namespace {
void CreateFillerObjectAtImpl(Heap* heap, Address addr, int size,
ClearFreedMemoryMode clear_memory_mode) {
if (size == 0) return;
DCHECK_IMPLIES(V8_COMPRESS_POINTERS_8GB_BOOL,
IsAligned(addr, kObjectAlignment8GbHeap));
DCHECK_IMPLIES(V8_COMPRESS_POINTERS_8GB_BOOL,
IsAligned(size, kObjectAlignment8GbHeap));
CodePageMemoryModificationScope memory_modification_scope(
BasicMemoryChunk::FromAddress(addr));
// TODO(v8:13070): Filler sizes are irrelevant for 8GB+ heaps. Adding them
// should be avoided in this mode.
Tagged<HeapObject> filler = HeapObject::FromAddress(addr);
ReadOnlyRoots roots(heap);
if (size == kTaggedSize) {
filler->set_map_after_allocation(roots.unchecked_one_pointer_filler_map(),
SKIP_WRITE_BARRIER);
// Ensure the filler map is properly initialized.
DCHECK(IsMap(filler->map(heap->isolate())));
} else if (size == 2 * kTaggedSize) {
filler->set_map_after_allocation(roots.unchecked_two_pointer_filler_map(),
SKIP_WRITE_BARRIER);
if (clear_memory_mode == ClearFreedMemoryMode::kClearFreedMemory) {
AtomicSlot slot(ObjectSlot(addr) + 1);
*slot = static_cast<Tagged_t>(kClearedFreeMemoryValue);
}
// Ensure the filler map is properly initialized.
DCHECK(IsMap(filler->map(heap->isolate())));
} else {
DCHECK_GT(size, 2 * kTaggedSize);
filler->set_map_after_allocation(roots.unchecked_free_space_map(),
SKIP_WRITE_BARRIER);
FreeSpace::cast(filler)->set_size(size, kRelaxedStore);
if (clear_memory_mode == ClearFreedMemoryMode::kClearFreedMemory) {
MemsetTagged(ObjectSlot(addr) + 2,
Tagged<Object>(kClearedFreeMemoryValue),
(size / kTaggedSize) - 2);
}
// During bootstrapping we need to create a free space object before its
// map is initialized. In this case we cannot access the map yet, as it
// might be null, or not set up properly yet.
DCHECK_IMPLIES(roots.is_initialized(RootIndex::kFreeSpaceMap),
IsMap(filler->map(heap->isolate())));
}
}
#ifdef DEBUG
void VerifyNoNeedToClearSlots(Address start, Address end) {
BasicMemoryChunk* basic_chunk = BasicMemoryChunk::FromAddress(start);
if (basic_chunk->InReadOnlySpace()) return;
MemoryChunk* chunk = static_cast<MemoryChunk*>(basic_chunk);
if (chunk->InYoungGeneration()) return;
BaseSpace* space = chunk->owner();
space->heap()->VerifySlotRangeHasNoRecordedSlots(start, end);
}
#else
void VerifyNoNeedToClearSlots(Address start, Address end) {}
#endif // DEBUG
} // namespace
void Heap::CreateFillerObjectAtBackground(Address addr, int size) {
// TODO(leszeks): Verify that no slots need to be recorded.
// Do not verify whether slots are cleared here: the concurrent thread is not
// allowed to access the main thread's remembered set.
CreateFillerObjectAtRaw(addr, size,
ClearFreedMemoryMode::kDontClearFreedMemory,
ClearRecordedSlots::kNo, VerifyNoSlotsRecorded::kNo);
}
void Heap::CreateFillerObjectAtSweeper(Address addr, int size) {
// Do not verify whether slots are cleared here: the concurrent sweeper is not
// allowed to access the main thread's remembered set.
CreateFillerObjectAtRaw(addr, size,
ClearFreedMemoryMode::kDontClearFreedMemory,
ClearRecordedSlots::kNo, VerifyNoSlotsRecorded::kNo);
}
void Heap::CreateFillerObjectAt(Address addr, int size,
ClearFreedMemoryMode clear_memory_mode) {
CreateFillerObjectAtRaw(addr, size, clear_memory_mode,
ClearRecordedSlots::kNo, VerifyNoSlotsRecorded::kYes);
}
void Heap::CreateFillerObjectAtRaw(
Address addr, int size, ClearFreedMemoryMode clear_memory_mode,
ClearRecordedSlots clear_slots_mode,
VerifyNoSlotsRecorded verify_no_slots_recorded) {
// TODO(mlippautz): It would be nice to DCHECK that we never call this
// with {addr} pointing into large object space; however we currently do,
// see, e.g., Factory::NewFillerObject and in many tests.
if (size == 0) return;
CreateFillerObjectAtImpl(this, addr, size, clear_memory_mode);
if (!V8_ENABLE_THIRD_PARTY_HEAP_BOOL) {
if (clear_slots_mode == ClearRecordedSlots::kYes) {
ClearRecordedSlotRange(addr, addr + size);
} else if (verify_no_slots_recorded == VerifyNoSlotsRecorded::kYes) {
VerifyNoNeedToClearSlots(addr, addr + size);
}
}
}
bool Heap::CanMoveObjectStart(Tagged<HeapObject> object) {
if (!v8_flags.move_object_start) return false;
// Sampling heap profiler may have a reference to the object.
if (isolate()->heap_profiler()->is_sampling_allocations()) return false;
if (IsLargeObject(object)) return false;
// Compilation jobs may have references to the object.
if (isolate()->concurrent_recompilation_enabled() &&
isolate()->optimizing_compile_dispatcher()->HasJobs()) {
return false;
}
// Concurrent marking does not support moving object starts without snapshot
// protocol.
//
// TODO(v8:13726): This can be improved via concurrently reading the contents
// in the marker at the cost of some complexity.
if (incremental_marking()->IsMarking()) return false;
// Concurrent sweeper does not support moving object starts. It assumes that
// markbits (black regions) and object starts are matching up.
if (!Page::FromHeapObject(object)->SweepingDone()) return false;
return true;
}
bool Heap::IsImmovable(Tagged<HeapObject> object) {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL)
return third_party_heap::Heap::IsImmovable(object);
BasicMemoryChunk* chunk = BasicMemoryChunk::FromHeapObject(object);
return chunk->NeverEvacuate() || IsLargeObject(object);
}
bool Heap::IsLargeObject(Tagged<HeapObject> object) {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL)
return third_party_heap::Heap::InLargeObjectSpace(object.address()) ||
third_party_heap::Heap::InSpace(object.address(), CODE_LO_SPACE);
return BasicMemoryChunk::FromHeapObject(object)->IsLargePage();
}
#ifdef ENABLE_SLOW_DCHECKS
namespace {
class LeftTrimmerVerifierRootVisitor : public RootVisitor {
public:
explicit LeftTrimmerVerifierRootVisitor(Tagged<FixedArrayBase> to_check)
: to_check_(to_check) {}
LeftTrimmerVerifierRootVisitor(const LeftTrimmerVerifierRootVisitor&) =
delete;
LeftTrimmerVerifierRootVisitor& operator=(
const LeftTrimmerVerifierRootVisitor&) = delete;
void VisitRootPointers(Root root, const char* description,
FullObjectSlot start, FullObjectSlot end) override {
for (FullObjectSlot p = start; p < end; ++p) {
// V8_EXTERNAL_CODE_SPACE specific: we might be comparing
// InstructionStream object with non-InstructionStream object here and it
// might produce false positives because operator== for tagged values
// compares only lower 32 bits when pointer compression is enabled.
DCHECK_NE((*p).ptr(), to_check_.ptr());
}
}
void VisitRootPointers(Root root, const char* description,
OffHeapObjectSlot start,
OffHeapObjectSlot end) override {
DCHECK_EQ(root, Root::kStringTable);
// We can skip iterating the string table, it doesn't point to any fixed
// arrays.
}
private:
Tagged<FixedArrayBase> to_check_;
};
} // namespace
#endif // ENABLE_SLOW_DCHECKS
namespace {
bool MayContainRecordedSlots(Tagged<HeapObject> object) {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) return false;
// New space object do not have recorded slots.
if (BasicMemoryChunk::FromHeapObject(object)->InYoungGeneration())
return false;
// Allowlist objects that definitely do not have pointers.
if (IsByteArray(object) || IsFixedDoubleArray(object)) return false;
// Conservatively return true for other objects.
return true;
}
} // namespace
void Heap::OnMoveEvent(Tagged<HeapObject> source, Tagged<HeapObject> target,
int size_in_bytes) {
HeapProfiler* heap_profiler = isolate_->heap_profiler();
if (heap_profiler->is_tracking_object_moves()) {
heap_profiler->ObjectMoveEvent(source.address(), target.address(),
size_in_bytes, /*is_embedder_object=*/false);
}
for (auto& tracker : allocation_trackers_) {
tracker->MoveEvent(source.address(), target.address(), size_in_bytes);
}
if (IsSharedFunctionInfo(target, isolate_)) {
LOG_CODE_EVENT(isolate_, SharedFunctionInfoMoveEvent(source.address(),
target.address()));
} else if (IsNativeContext(target, isolate_)) {
if (isolate_->current_embedder_state() != nullptr) {
isolate_->current_embedder_state()->OnMoveEvent(source.address(),
target.address());
}
PROFILE(isolate_,
NativeContextMoveEvent(source.address(), target.address()));
} else if (IsMap(target, isolate_)) {
LOG(isolate_, MapMoveEvent(Map::cast(source), Map::cast(target)));
}
}
Tagged<FixedArrayBase> Heap::LeftTrimFixedArray(Tagged<FixedArrayBase> object,
int elements_to_trim) {
if (elements_to_trim == 0) {
// This simplifies reasoning in the rest of the function.
return object;
}
CHECK(!object.is_null());
DCHECK(CanMoveObjectStart(object));
// Add custom visitor to concurrent marker if new left-trimmable type
// is added.
DCHECK(IsFixedArray(object) || IsFixedDoubleArray(object));
const int element_size = IsFixedArray(object) ? kTaggedSize : kDoubleSize;
const int bytes_to_trim = elements_to_trim * element_size;
Tagged<Map> map = object->map();
// For now this trick is only applied to fixed arrays which may be in new
// space or old space. In a large object space the object's start must
// coincide with chunk and thus the trick is just not applicable.
DCHECK(!IsLargeObject(object));
DCHECK(object->map() != ReadOnlyRoots(this).fixed_cow_array_map());
static_assert(FixedArrayBase::kMapOffset == 0);
static_assert(FixedArrayBase::kLengthOffset == kTaggedSize);
static_assert(FixedArrayBase::kHeaderSize == 2 * kTaggedSize);
const int len = object->length();
DCHECK(elements_to_trim <= len);
// Calculate location of new array start.
Address old_start = object.address();
Address new_start = old_start + bytes_to_trim;
// Technically in new space this write might be omitted (except for
// debug mode which iterates through the heap), but to play safer
// we still do it.
CreateFillerObjectAtRaw(
old_start, bytes_to_trim, ClearFreedMemoryMode::kClearFreedMemory,
MayContainRecordedSlots(object) ? ClearRecordedSlots::kYes
: ClearRecordedSlots::kNo,
VerifyNoSlotsRecorded::kYes);
// Initialize header of the trimmed array. Since left trimming is only
// performed on pages which are not concurrently swept creating a filler
// object does not require synchronization.
RELAXED_WRITE_FIELD(object, bytes_to_trim,
Tagged<Object>(MapWord::FromMap(map).ptr()));
RELAXED_WRITE_FIELD(object, bytes_to_trim + kTaggedSize,
Smi::FromInt(len - elements_to_trim));
Tagged<FixedArrayBase> new_object =
Tagged<FixedArrayBase>::cast(HeapObject::FromAddress(new_start));
if (isolate()->log_object_relocation()) {
// Notify the heap profiler of change in object layout.
OnMoveEvent(object, new_object, new_object->Size());
}
#ifdef ENABLE_SLOW_DCHECKS
if (v8_flags.enable_slow_asserts) {
// Make sure the stack or other roots (e.g., Handles) don't contain pointers
// to the original FixedArray (which is now the filler object).
base::Optional<IsolateSafepointScope> safepoint_scope;
{
AllowGarbageCollection allow_gc;
safepoint_scope.emplace(this);
}
LeftTrimmerVerifierRootVisitor root_visitor(object);
ReadOnlyRoots(this).Iterate(&root_visitor);
IterateRoots(&root_visitor,
base::EnumSet<SkipRoot>{SkipRoot::kConservativeStack});
}
#endif // ENABLE_SLOW_DCHECKS
return new_object;
}
void Heap::RightTrimFixedArray(Tagged<FixedArrayBase> object,
int elements_to_trim) {
const int len = object->length();
DCHECK_LE(elements_to_trim, len);
DCHECK_GE(elements_to_trim, 0);
int bytes_to_trim;
if (IsByteArray(object)) {
int new_size = ByteArray::SizeFor(len - elements_to_trim);
bytes_to_trim = ByteArray::SizeFor(len) - new_size;
DCHECK_GE(bytes_to_trim, 0);
} else if (IsFixedArray(object)) {
CHECK_NE(elements_to_trim, len);
bytes_to_trim = elements_to_trim * kTaggedSize;
} else {
DCHECK(IsFixedDoubleArray(object));
CHECK_NE(elements_to_trim, len);
bytes_to_trim = elements_to_trim * kDoubleSize;
}
CreateFillerForArray<FixedArrayBase>(object, elements_to_trim, bytes_to_trim);
}
void Heap::RightTrimWeakFixedArray(Tagged<WeakFixedArray> object,
int elements_to_trim) {
// This function is safe to use only at the end of the mark compact
// collection: When marking, we record the weak slots, and shrinking
// invalidates them.
DCHECK_EQ(gc_state(), MARK_COMPACT);
CreateFillerForArray<WeakFixedArray>(object, elements_to_trim,
elements_to_trim * kTaggedSize);
}
template <typename T>
void Heap::CreateFillerForArray(Tagged<T> object, int elements_to_trim,
int bytes_to_trim) {
DCHECK(IsFixedArrayBase(object) || IsByteArray(object) ||
IsWeakFixedArray(object));
// For now this trick is only applied to objects in new and paged space.
DCHECK(object->map() != ReadOnlyRoots(this).fixed_cow_array_map());
if (bytes_to_trim == 0) {
DCHECK_EQ(elements_to_trim, 0);
// No need to create filler and update live bytes counters.
return;
}
// Calculate location of new array end.
int old_size = object->Size();
Address old_end = object.address() + old_size;
Address new_end = old_end - bytes_to_trim;
bool clear_slots = MayContainRecordedSlots(object);
// Technically in new space this write might be omitted (except for
// debug mode which iterates through the heap), but to play safer
// we still do it.
// We do not create a filler for objects in a large object space.
if (!IsLargeObject(object)) {
NotifyObjectSizeChange(
object, old_size, old_size - bytes_to_trim,
clear_slots ? ClearRecordedSlots::kYes : ClearRecordedSlots::kNo);
Tagged<HeapObject> filler = HeapObject::FromAddress(new_end);
// Clear the mark bits of the black area that belongs now to the filler.
// This is an optimization. The sweeper will release black fillers anyway.
if (incremental_marking()->black_allocation() &&
marking_state()->IsMarked(filler)) {
Page* page = Page::FromAddress(new_end);
page->marking_bitmap()->ClearRange<AccessMode::ATOMIC>(
MarkingBitmap::AddressToIndex(new_end),
MarkingBitmap::LimitAddressToIndex(new_end + bytes_to_trim));
}
} else if (clear_slots) {
// Large objects are not swept, so it is not necessary to clear the
// recorded slot.
MemsetTagged(ObjectSlot(new_end), Tagged<Object>(kClearedFreeMemoryValue),
(old_end - new_end) / kTaggedSize);
}
// Initialize header of the trimmed array. We are storing the new length
// using release store after creating a filler for the left-over space to
// avoid races with the sweeper thread.
object->set_length(object->length() - elements_to_trim, kReleaseStore);
// Notify the heap object allocation tracker of change in object layout. The
// array may not be moved during GC, and size has to be adjusted nevertheless.
for (auto& tracker : allocation_trackers_) {
tracker->UpdateObjectSizeEvent(object.address(), object->Size());
}
}
void Heap::MakeHeapIterable() {
EnsureSweepingCompleted(SweepingForcedFinalizationMode::kV8Only);
safepoint()->IterateLocalHeaps([](LocalHeap* local_heap) {
local_heap->MakeLinearAllocationAreaIterable();
});
if (isolate()->is_shared_space_isolate()) {
isolate()->global_safepoint()->IterateSharedSpaceAndClientIsolates(
[](Isolate* client) {
client->heap()->MakeSharedLinearAllocationAreasIterable();
});
}
allocator()->MakeLinearAllocationAreaIterable();
if (shared_space_allocator_) {
shared_space_allocator_->MakeLinearAllocationAreaIterable();
}
}
void Heap::FreeLinearAllocationAreas() {
FreeMainThreadLinearAllocationAreas();
safepoint()->IterateLocalHeaps(
[](LocalHeap* local_heap) { local_heap->FreeLinearAllocationArea(); });
if (isolate()->is_shared_space_isolate()) {
isolate()->global_safepoint()->IterateSharedSpaceAndClientIsolates(
[](Isolate* client) {
client->heap()->FreeSharedLinearAllocationAreas();
});
}
}
void Heap::FreeMainThreadLinearAllocationAreas() {
PagedSpaceIterator spaces(this);
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
base::MutexGuard guard(space->mutex());
space->FreeLinearAllocationArea();
}
if (shared_space_allocator_) {
shared_space_allocator_->FreeLinearAllocationArea();
}
if (new_space()) {
allocator()->new_space_allocator()->FreeLinearAllocationArea();
}
}
void Heap::FreeSharedLinearAllocationAreas() {
if (!isolate()->has_shared_space()) return;
safepoint()->IterateLocalHeaps([](LocalHeap* local_heap) {
local_heap->FreeSharedLinearAllocationArea();
});
FreeMainThreadSharedLinearAllocationAreas();
}
void Heap::FreeMainThreadSharedLinearAllocationAreas() {
if (!isolate()->has_shared_space()) return;
shared_space_allocator_->FreeLinearAllocationArea();
main_thread_local_heap()->FreeSharedLinearAllocationArea();
}
void Heap::MakeSharedLinearAllocationAreasIterable() {
if (!isolate()->has_shared_space()) return;
safepoint()->IterateLocalHeaps([](LocalHeap* local_heap) {
local_heap->MakeSharedLinearAllocationAreaIterable();
});
if (shared_space_allocator_) {
shared_space_allocator_->MakeLinearAllocationAreaIterable();
}
main_thread_local_heap()->MakeSharedLinearAllocationAreaIterable();
}
void Heap::MarkSharedLinearAllocationAreasBlack() {
if (shared_space_allocator_) {
shared_space_allocator_->MarkLinearAllocationAreaBlack();
}
safepoint()->IterateLocalHeaps([](LocalHeap* local_heap) {
local_heap->MarkSharedLinearAllocationAreaBlack();
});
main_thread_local_heap()->MarkSharedLinearAllocationAreaBlack();
}
void Heap::UnmarkSharedLinearAllocationAreas() {
if (shared_space_allocator_) {
shared_space_allocator_->UnmarkLinearAllocationArea();
}
safepoint()->IterateLocalHeaps([](LocalHeap* local_heap) {
local_heap->UnmarkSharedLinearAllocationArea();
});
main_thread_local_heap()->UnmarkSharedLinearAllocationArea();
}
namespace {
double ComputeMutatorUtilizationImpl(double mutator_speed, double gc_speed) {
constexpr double kMinMutatorUtilization = 0.0;
constexpr double kConservativeGcSpeedInBytesPerMillisecond = 200000;
if (mutator_speed == 0) return kMinMutatorUtilization;
if (gc_speed == 0) gc_speed = kConservativeGcSpeedInBytesPerMillisecond;
// Derivation:
// mutator_utilization = mutator_time / (mutator_time + gc_time)
// mutator_time = 1 / mutator_speed
// gc_time = 1 / gc_speed
// mutator_utilization = (1 / mutator_speed) /
// (1 / mutator_speed + 1 / gc_speed)
// mutator_utilization = gc_speed / (mutator_speed + gc_speed)
return gc_speed / (mutator_speed + gc_speed);
}
} // namespace
double Heap::ComputeMutatorUtilization(const char* tag, double mutator_speed,
double gc_speed) {
double result = ComputeMutatorUtilizationImpl(mutator_speed, gc_speed);
if (v8_flags.trace_mutator_utilization) {
isolate()->PrintWithTimestamp(
"%s mutator utilization = %.3f ("
"mutator_speed=%.f, gc_speed=%.f)\n",
tag, result, mutator_speed, gc_speed);
}
return result;
}
bool Heap::HasLowYoungGenerationAllocationRate() {
double mu = ComputeMutatorUtilization(
"Young generation",
tracer()->NewSpaceAllocationThroughputInBytesPerMillisecond(),
tracer()->ScavengeSpeedInBytesPerMillisecond(kForSurvivedObjects));
constexpr double kHighMutatorUtilization = 0.993;
return mu > kHighMutatorUtilization;
}
bool Heap::HasLowOldGenerationAllocationRate() {
double mu = ComputeMutatorUtilization(
"Old generation",
tracer()->OldGenerationAllocationThroughputInBytesPerMillisecond(),
tracer()->CombinedMarkCompactSpeedInBytesPerMillisecond());
const double kHighMutatorUtilization = 0.993;
return mu > kHighMutatorUtilization;
}
bool Heap::HasLowEmbedderAllocationRate() {
double mu = ComputeMutatorUtilization(
"Embedder",
tracer()->CurrentEmbedderAllocationThroughputInBytesPerMillisecond(),
tracer()->EmbedderSpeedInBytesPerMillisecond());
const double kHighMutatorUtilization = 0.993;
return mu > kHighMutatorUtilization;
}
bool Heap::HasLowAllocationRate() {
return HasLowYoungGenerationAllocationRate() &&
HasLowOldGenerationAllocationRate() && HasLowEmbedderAllocationRate();
}
bool Heap::IsIneffectiveMarkCompact(size_t old_generation_size,
double mutator_utilization) {
const double kHighHeapPercentage = 0.8;
const double kLowMutatorUtilization = 0.4;
return old_generation_size >=
kHighHeapPercentage * max_old_generation_size() &&
mutator_utilization < kLowMutatorUtilization;
}
void Heap::CheckIneffectiveMarkCompact(size_t old_generation_size,
double mutator_utilization) {
const int kMaxConsecutiveIneffectiveMarkCompacts = 4;
if (!v8_flags.detect_ineffective_gcs_near_heap_limit) return;
if (!IsIneffectiveMarkCompact(old_generation_size, mutator_utilization)) {
consecutive_ineffective_mark_compacts_ = 0;
return;
}
++consecutive_ineffective_mark_compacts_;
if (consecutive_ineffective_mark_compacts_ ==
kMaxConsecutiveIneffectiveMarkCompacts) {
if (InvokeNearHeapLimitCallback()) {
// The callback increased the heap limit.
consecutive_ineffective_mark_compacts_ = 0;
return;
}
FatalProcessOutOfMemory("Ineffective mark-compacts near heap limit");
}
}
bool Heap::HasHighFragmentation() {
const size_t used = OldGenerationSizeOfObjects();
const size_t committed = CommittedOldGenerationMemory();
// Background thread allocation could result in committed memory being less
// than used memory in some situations.
if (committed < used) return false;
constexpr size_t kSlack = 16 * MB;
// Fragmentation is high if committed > 2 * used + kSlack.
// Rewrite the expression to avoid overflow.
return committed - used > used + kSlack;
}
bool Heap::ShouldOptimizeForMemoryUsage() {
const size_t kOldGenerationSlack = max_old_generation_size() / 8;
return v8_flags.optimize_for_size || isolate()->IsIsolateInBackground() ||
HighMemoryPressure() || !CanExpandOldGeneration(kOldGenerationSlack);
}
class ActivateMemoryReducerTask : public CancelableTask {
public:
explicit ActivateMemoryReducerTask(Heap* heap)
: CancelableTask(heap->isolate()), heap_(heap) {}
~ActivateMemoryReducerTask() override = default;
ActivateMemoryReducerTask(const ActivateMemoryReducerTask&) = delete;
ActivateMemoryReducerTask& operator=(const ActivateMemoryReducerTask&) =
delete;
private:
// v8::internal::CancelableTask overrides.
void RunInternal() override {
heap_->ActivateMemoryReducerIfNeededOnMainThread();
}
Heap* heap_;
};
void Heap::ActivateMemoryReducerIfNeeded() {
if (memory_reducer_ == nullptr) return;
// This method may be called from any thread. Post a task to run it on the
// isolate's main thread to avoid synchronization.
task_runner_->PostTask(std::make_unique<ActivateMemoryReducerTask>(this));
}
void Heap::ActivateMemoryReducerIfNeededOnMainThread() {
// Activate memory reducer when switching to background if
// - there was no mark compact since the start.
// - the committed memory can be potentially reduced.
// 2 pages for the old, code, and map space + 1 page for new space.
const int kMinCommittedMemory = 7 * Page::kPageSize;
if (ms_count_ == 0 && CommittedMemory() > kMinCommittedMemory &&
isolate()->IsIsolateInBackground()) {
memory_reducer_->NotifyPossibleGarbage();
}
}
Heap::ResizeNewSpaceMode Heap::ShouldResizeNewSpace() {
if (ShouldReduceMemory()) {
return (v8_flags.predictable) ? ResizeNewSpaceMode::kNone
: ResizeNewSpaceMode::kShrink;
}
static const size_t kLowAllocationThroughput = 1000;
const double allocation_throughput =
tracer_->CurrentAllocationThroughputInBytesPerMillisecond();
const bool should_shrink = !v8_flags.predictable &&
(allocation_throughput != 0) &&
(allocation_throughput < kLowAllocationThroughput);
const bool should_grow =
(new_space_->TotalCapacity() < new_space_->MaximumCapacity()) &&
(survived_since_last_expansion_ > new_space_->TotalCapacity());
if (should_grow) survived_since_last_expansion_ = 0;
if (should_grow == should_shrink) return ResizeNewSpaceMode::kNone;
return should_grow ? ResizeNewSpaceMode::kGrow : ResizeNewSpaceMode::kShrink;
}
void Heap::ExpandNewSpaceSize() {
// Grow the size of new space if there is room to grow, and enough data
// has survived scavenge since the last expansion.
new_space_->Grow();
new_lo_space()->SetCapacity(new_space()->TotalCapacity());
}
void Heap::ReduceNewSpaceSize() {
// MinorMS shrinks new space as part of sweeping.
if (!v8_flags.minor_ms) {
SemiSpaceNewSpace::From(new_space())->Shrink();
} else {
paged_new_space()->FinishShrinking();
}
new_lo_space_->SetCapacity(new_space()->TotalCapacity());
}
size_t Heap::NewSpaceSize() { return new_space() ? new_space()->Size() : 0; }
size_t Heap::NewSpaceCapacity() const {
return new_space() ? new_space()->Capacity() : 0;
}
size_t Heap::NewSpaceTargetCapacity() const {
return new_space() ? new_space()->TotalCapacity() : 0;
}
void Heap::FinalizeIncrementalMarkingAtomically(
GarbageCollectionReason gc_reason) {
DCHECK(!incremental_marking()->IsStopped());
CollectAllGarbage(current_gc_flags_, gc_reason, current_gc_callback_flags_);
}
void Heap::InvokeIncrementalMarkingPrologueCallbacks() {
AllowGarbageCollection allow_allocation;
VMState<EXTERNAL> state(isolate_);
CallGCPrologueCallbacks(kGCTypeIncrementalMarking, kNoGCCallbackFlags,
GCTracer::Scope::MC_INCREMENTAL_EXTERNAL_PROLOGUE);
}
void Heap::InvokeIncrementalMarkingEpilogueCallbacks() {
AllowGarbageCollection allow_allocation;
VMState<EXTERNAL> state(isolate_);
CallGCEpilogueCallbacks(kGCTypeIncrementalMarking, kNoGCCallbackFlags,
GCTracer::Scope::MC_INCREMENTAL_EXTERNAL_EPILOGUE);
}
namespace {
thread_local Address pending_layout_change_object_address = kNullAddress;
} // namespace
void Heap::NotifyObjectLayoutChange(
Tagged<HeapObject> object, const DisallowGarbageCollection&,
InvalidateRecordedSlots invalidate_recorded_slots, int new_size) {
if (invalidate_recorded_slots == InvalidateRecordedSlots::kYes) {
const bool may_contain_recorded_slots = MayContainRecordedSlots(object);
MemoryChunk* const chunk = MemoryChunk::FromHeapObject(object);
// Do not remove the recorded slot in the map word as this one can never be
// invalidated.
const Address clear_range_start = object.address() + kTaggedSize;
// Only slots in the range of the new object size (which is potentially
// smaller than the original one) can be invalidated. Clearing of recorded
// slots up to the original object size even conflicts with concurrent
// sweeping.
const Address clear_range_end = object.address() + new_size;
if (incremental_marking()->IsMarking()) {
ExclusiveObjectLock::Lock(object);
DCHECK_EQ(pending_layout_change_object_address, kNullAddress);
pending_layout_change_object_address = object.address();
if (may_contain_recorded_slots && incremental_marking()->IsCompacting()) {
RememberedSet<OLD_TO_OLD>::RemoveRange(
chunk, clear_range_start, clear_range_end,
SlotSet::EmptyBucketMode::KEEP_EMPTY_BUCKETS);
}
}
if (may_contain_recorded_slots) {
RememberedSet<OLD_TO_NEW>::RemoveRange(
chunk, clear_range_start, clear_range_end,
SlotSet::EmptyBucketMode::KEEP_EMPTY_BUCKETS);
RememberedSet<OLD_TO_NEW_BACKGROUND>::RemoveRange(
chunk, clear_range_start, clear_range_end,
SlotSet::EmptyBucketMode::KEEP_EMPTY_BUCKETS);
RememberedSet<OLD_TO_SHARED>::RemoveRange(
chunk, clear_range_start, clear_range_end,
SlotSet::EmptyBucketMode::KEEP_EMPTY_BUCKETS);
}
}
#ifdef VERIFY_HEAP
if (v8_flags.verify_heap) {
HeapVerifier::SetPendingLayoutChangeObject(this, object);
}
#endif
}
// static
void Heap::NotifyObjectLayoutChangeDone(Tagged<HeapObject> object) {
if (pending_layout_change_object_address != kNullAddress) {
DCHECK_EQ(pending_layout_change_object_address, object.address());
ExclusiveObjectLock::Unlock(object);
pending_layout_change_object_address = kNullAddress;
}
}
void Heap::NotifyObjectSizeChange(Tagged<HeapObject> object, int old_size,
int new_size,
ClearRecordedSlots clear_recorded_slots) {
old_size = ALIGN_TO_ALLOCATION_ALIGNMENT(old_size);
new_size = ALIGN_TO_ALLOCATION_ALIGNMENT(new_size);
DCHECK_LE(new_size, old_size);
DCHECK(!IsLargeObject(object));
if (new_size == old_size) return;
const bool is_main_thread = LocalHeap::Current() == nullptr;
DCHECK_IMPLIES(!is_main_thread,
clear_recorded_slots == ClearRecordedSlots::kNo);
const auto verify_no_slots_recorded =
is_main_thread ? VerifyNoSlotsRecorded::kYes : VerifyNoSlotsRecorded::kNo;
const auto clear_memory_mode = ClearFreedMemoryMode::kDontClearFreedMemory;
const Address filler = object.address() + new_size;
const int filler_size = old_size - new_size;
CreateFillerObjectAtRaw(filler, filler_size, clear_memory_mode,
clear_recorded_slots, verify_no_slots_recorded);
}
GCIdleTimeHeapState Heap::ComputeHeapState() {
GCIdleTimeHeapState heap_state;
heap_state.size_of_objects = static_cast<size_t>(SizeOfObjects());
heap_state.incremental_marking_stopped = incremental_marking()->IsStopped();
return heap_state;
}
bool Heap::PerformIdleTimeAction(GCIdleTimeAction action,
GCIdleTimeHeapState heap_state,
double deadline_in_ms) {
bool result = false;
switch (action) {
case GCIdleTimeAction::kDone:
result = true;
break;
case GCIdleTimeAction::kIncrementalStep: {
incremental_marking()->AdvanceAndFinalizeIfComplete();
result = incremental_marking()->IsStopped();
break;
}
}
return result;
}
void Heap::IdleNotificationEpilogue(GCIdleTimeAction action,
GCIdleTimeHeapState heap_state,
double start_ms, double deadline_in_ms) {
const double idle_time_in_ms = deadline_in_ms - start_ms;
const double deadline_difference =
deadline_in_ms - MonotonicallyIncreasingTimeInMs();
if (v8_flags.trace_idle_notification) {
isolate_->PrintWithTimestamp(
"Idle notification: requested idle time %.2f ms, used idle time %.2f "
"ms, deadline usage %.2f ms [",
idle_time_in_ms, idle_time_in_ms - deadline_difference,
deadline_difference);
switch (action) {
case GCIdleTimeAction::kDone:
PrintF("done");
break;
case GCIdleTimeAction::kIncrementalStep:
PrintF("incremental step");
break;
}
PrintF("]");
if (v8_flags.trace_idle_notification_verbose) {
PrintF("[");
heap_state.Print();
PrintF("]");
}
PrintF("\n");
}
}
double Heap::MonotonicallyIncreasingTimeInMs() const {
return V8::GetCurrentPlatform()->MonotonicallyIncreasingTime() *
static_cast<double>(base::Time::kMillisecondsPerSecond);
}
#if DEBUG
void Heap::VerifyNewSpaceTop() {
if (!new_space()) return;
allocator()->new_space_allocator()->Verify();
}
#endif // DEBUG
bool Heap::IdleNotification(int idle_time_in_ms) {
return IdleNotification(
V8::GetCurrentPlatform()->MonotonicallyIncreasingTime() +
(static_cast<double>(idle_time_in_ms) /
static_cast<double>(base::Time::kMillisecondsPerSecond)));
}
bool Heap::IdleNotification(double deadline_in_seconds) {
CHECK(HasBeenSetUp());
double deadline_in_ms =
deadline_in_seconds *
static_cast<double>(base::Time::kMillisecondsPerSecond);
NestedTimedHistogramScope idle_notification_scope(
isolate_->counters()->gc_idle_notification());
TRACE_EVENT0("v8", "V8.GCIdleNotification");
double start_ms = MonotonicallyIncreasingTimeInMs();
double idle_time_in_ms = deadline_in_ms - start_ms;
tracer()->SampleAllocation(
base::TimeTicks::Now(), NewSpaceAllocationCounter(),
OldGenerationAllocationCounter(), EmbedderAllocationCounter());
GCIdleTimeHeapState heap_state = ComputeHeapState();
GCIdleTimeAction action =
gc_idle_time_handler_->Compute(idle_time_in_ms, heap_state);
bool result = PerformIdleTimeAction(action, heap_state, deadline_in_ms);
IdleNotificationEpilogue(action, heap_state, start_ms, deadline_in_ms);
return result;
}
class MemoryPressureInterruptTask : public CancelableTask {
public:
explicit MemoryPressureInterruptTask(Heap* heap)
: CancelableTask(heap->isolate()), heap_(heap) {}
~MemoryPressureInterruptTask() override = default;
MemoryPressureInterruptTask(const MemoryPressureInterruptTask&) = delete;
MemoryPressureInterruptTask& operator=(const MemoryPressureInterruptTask&) =
delete;
private:
// v8::internal::CancelableTask overrides.
void RunInternal() override { heap_->CheckMemoryPressure(); }
Heap* heap_;
};
void Heap::CheckMemoryPressure() {
if (HighMemoryPressure()) {
// The optimizing compiler may be unnecessarily holding on to memory.
isolate()->AbortConcurrentOptimization(BlockingBehavior::kDontBlock);
}
// Reset the memory pressure level to avoid recursive GCs triggered by
// CheckMemoryPressure from AdjustAmountOfExternalMemory called by
// the finalizers.
MemoryPressureLevel memory_pressure_level = memory_pressure_level_.exchange(
MemoryPressureLevel::kNone, std::memory_order_relaxed);
if (memory_pressure_level == MemoryPressureLevel::kCritical) {
TRACE_EVENT0("devtools.timeline,v8", "V8.CheckMemoryPressure");
CollectGarbageOnMemoryPressure();
} else if (memory_pressure_level == MemoryPressureLevel::kModerate) {
if (v8_flags.incremental_marking && incremental_marking()->IsStopped()) {
TRACE_EVENT0("devtools.timeline,v8", "V8.CheckMemoryPressure");
StartIncrementalMarking(GCFlag::kReduceMemoryFootprint,
GarbageCollectionReason::kMemoryPressure);
}
}
}
void Heap::CollectGarbageOnMemoryPressure() {
const int kGarbageThresholdInBytes = 8 * MB;
const double kGarbageThresholdAsFractionOfTotalMemory = 0.1;
// This constant is the maximum response time in RAIL performance model.
const double kMaxMemoryPressurePauseMs = 100;
double start = MonotonicallyIncreasingTimeInMs();
CollectAllGarbage(GCFlag::kReduceMemoryFootprint,
GarbageCollectionReason::kMemoryPressure,
kGCCallbackFlagCollectAllAvailableGarbage);
EagerlyFreeExternalMemory();
double end = MonotonicallyIncreasingTimeInMs();
// Estimate how much memory we can free.
int64_t potential_garbage =
(CommittedMemory() - SizeOfObjects()) + external_memory_.total();
// If we can potentially free large amount of memory, then start GC right
// away instead of waiting for memory reducer.
if (potential_garbage >= kGarbageThresholdInBytes &&
potential_garbage >=
CommittedMemory() * kGarbageThresholdAsFractionOfTotalMemory) {
// If we spent less than half of the time budget, then perform full GC
// Otherwise, start incremental marking.
if (end - start < kMaxMemoryPressurePauseMs / 2) {
CollectAllGarbage(GCFlag::kReduceMemoryFootprint,
GarbageCollectionReason::kMemoryPressure,
kGCCallbackFlagCollectAllAvailableGarbage);
} else {
if (v8_flags.incremental_marking && incremental_marking()->IsStopped()) {
StartIncrementalMarking(GCFlag::kReduceMemoryFootprint,
GarbageCollectionReason::kMemoryPressure);
}
}
}
}
void Heap::MemoryPressureNotification(MemoryPressureLevel level,
bool is_isolate_locked) {
TRACE_EVENT1("devtools.timeline,v8", "V8.MemoryPressureNotification", "level",
static_cast<int>(level));
MemoryPressureLevel previous =
memory_pressure_level_.exchange(level, std::memory_order_relaxed);
if ((previous != MemoryPressureLevel::kCritical &&
level == MemoryPressureLevel::kCritical) ||
(previous == MemoryPressureLevel::kNone &&
level == MemoryPressureLevel::kModerate)) {
if (is_isolate_locked) {
CheckMemoryPressure();
} else {
ExecutionAccess access(isolate());
isolate()->stack_guard()->RequestGC();
task_runner_->PostTask(
std::make_unique<MemoryPressureInterruptTask>(this));
}
}
}
void Heap::EagerlyFreeExternalMemory() {
CompleteArrayBufferSweeping(this);
memory_allocator()->unmapper()->EnsureUnmappingCompleted();
}
void Heap::AddNearHeapLimitCallback(v8::NearHeapLimitCallback callback,
void* data) {
const size_t kMaxCallbacks = 100;
CHECK_LT(near_heap_limit_callbacks_.size(), kMaxCallbacks);
for (auto callback_data : near_heap_limit_callbacks_) {
CHECK_NE(callback_data.first, callback);
}
near_heap_limit_callbacks_.push_back(std::make_pair(callback, data));
}
void Heap::RemoveNearHeapLimitCallback(v8::NearHeapLimitCallback callback,
size_t heap_limit) {
for (size_t i = 0; i < near_heap_limit_callbacks_.size(); i++) {
if (near_heap_limit_callbacks_[i].first == callback) {
near_heap_limit_callbacks_.erase(near_heap_limit_callbacks_.begin() + i);
if (heap_limit) {
RestoreHeapLimit(heap_limit);
}
return;
}
}
UNREACHABLE();
}
void Heap::AppendArrayBufferExtension(Tagged<JSArrayBuffer> object,
ArrayBufferExtension* extension) {
// ArrayBufferSweeper is managing all counters and updating Heap counters.
array_buffer_sweeper_->Append(object, extension);
}
void Heap::DetachArrayBufferExtension(Tagged<JSArrayBuffer> object,
ArrayBufferExtension* extension) {
// ArrayBufferSweeper is managing all counters and updating Heap counters.
return array_buffer_sweeper_->Detach(object, extension);
}
void Heap::AutomaticallyRestoreInitialHeapLimit(double threshold_percent) {
initial_max_old_generation_size_threshold_ =
initial_max_old_generation_size_ * threshold_percent;
}
bool Heap::InvokeNearHeapLimitCallback() {
if (near_heap_limit_callbacks_.size() > 0) {
AllowGarbageCollection allow_gc;
TRACE_GC(tracer(), GCTracer::Scope::HEAP_EXTERNAL_NEAR_HEAP_LIMIT);
VMState<EXTERNAL> callback_state(isolate());
HandleScope scope(isolate());
v8::NearHeapLimitCallback callback =
near_heap_limit_callbacks_.back().first;
void* data = near_heap_limit_callbacks_.back().second;
size_t heap_limit = callback(data, max_old_generation_size(),
initial_max_old_generation_size_);
if (heap_limit > max_old_generation_size()) {
SetOldGenerationAndGlobalMaximumSize(
std::min(heap_limit, AllocatorLimitOnMaxOldGenerationSize()));
return true;
}
}
return false;
}
bool Heap::MeasureMemory(std::unique_ptr<v8::MeasureMemoryDelegate> delegate,
v8::MeasureMemoryExecution execution) {
HandleScope handle_scope(isolate());
std::vector<Handle<NativeContext>> contexts = FindAllNativeContexts();
std::vector<Handle<NativeContext>> to_measure;
for (auto& current : contexts) {
if (delegate->ShouldMeasure(v8::Utils::ToLocal(current))) {
to_measure.push_back(current);
}
}
return memory_measurement_->EnqueueRequest(std::move(delegate), execution,
to_measure);
}
std::unique_ptr<v8::MeasureMemoryDelegate> Heap::MeasureMemoryDelegate(
Handle<NativeContext> context, Handle<JSPromise> promise,
v8::MeasureMemoryMode mode) {
return i::MemoryMeasurement::DefaultDelegate(isolate_, context, promise,
mode);
}
void Heap::CollectCodeStatistics() {
TRACE_EVENT0("v8", "Heap::CollectCodeStatistics");
IsolateSafepointScope safepoint_scope(this);
MakeHeapIterable();
CodeStatistics::ResetCodeAndMetadataStatistics(isolate());
// We do not look for code in new space, or map space. If code
// somehow ends up in those spaces, we would miss it here.
CodeStatistics::CollectCodeStatistics(code_space_, isolate());
CodeStatistics::CollectCodeStatistics(old_space_, isolate());
CodeStatistics::CollectCodeStatistics(code_lo_space_, isolate());
CodeStatistics::CollectCodeStatistics(trusted_space_, isolate());
CodeStatistics::CollectCodeStatistics(trusted_lo_space_, isolate());
}
#ifdef DEBUG
void Heap::Print() {
if (!HasBeenSetUp()) return;
isolate()->PrintStack(stdout);
for (SpaceIterator it(this); it.HasNext();) {
it.Next()->Print();
}
}
void Heap::ReportCodeStatistics(const char* title) {
PrintF("###### Code Stats (%s) ######\n", title);
CollectCodeStatistics();
CodeStatistics::ReportCodeStatistics(isolate());
}
#endif // DEBUG
bool Heap::Contains(Tagged<HeapObject> value) const {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) {
return true;
}
if (ReadOnlyHeap::Contains(value)) {
return false;
}
if (memory_allocator()->IsOutsideAllocatedSpace(value.address())) {
return false;
}
if (!HasBeenSetUp()) return false;
return (new_space_ && new_space_->Contains(value)) ||
old_space_->Contains(value) || code_space_->Contains(value) ||
(shared_space_ && shared_space_->Contains(value)) ||
lo_space_->Contains(value) || code_lo_space_->Contains(value) ||
(new_lo_space_ && new_lo_space_->Contains(value)) ||
trusted_space_->Contains(value) ||
trusted_lo_space_->Contains(value) ||
(shared_lo_space_ && shared_lo_space_->Contains(value));
}
bool Heap::ContainsCode(Tagged<HeapObject> value) const {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) {
return true;
}
// TODO(v8:11880): support external code space.
if (memory_allocator()->IsOutsideAllocatedSpace(value.address(),
EXECUTABLE)) {
return false;
}
return HasBeenSetUp() &&
(code_space_->Contains(value) || code_lo_space_->Contains(value));
}
bool Heap::SharedHeapContains(Tagged<HeapObject> value) const {
if (shared_allocation_space_) {
if (shared_allocation_space_->Contains(value)) return true;
if (shared_lo_allocation_space_->Contains(value)) return true;
}
return false;
}
bool Heap::MustBeInSharedOldSpace(Tagged<HeapObject> value) {
if (isolate()->OwnsStringTables()) return false;
if (ReadOnlyHeap::Contains(value)) return false;
if (Heap::InYoungGeneration(value)) return false;
if (IsExternalString(value)) return false;
if (IsInternalizedString(value)) return true;
return false;
}
bool Heap::InSpace(Tagged<HeapObject> value, AllocationSpace space) const {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL)
return third_party_heap::Heap::InSpace(value.address(), space);
if (memory_allocator()->IsOutsideAllocatedSpace(
value.address(),
IsAnyCodeSpace(space) ? EXECUTABLE : NOT_EXECUTABLE)) {
return false;
}
if (!HasBeenSetUp()) return false;
switch (space) {
case NEW_SPACE:
return new_space_->Contains(value);
case OLD_SPACE:
return old_space_->Contains(value);
case CODE_SPACE:
return code_space_->Contains(value);
case SHARED_SPACE:
return shared_space_->Contains(value);
case TRUSTED_SPACE:
return trusted_space_->Contains(value);
case LO_SPACE:
return lo_space_->Contains(value);
case CODE_LO_SPACE:
return code_lo_space_->Contains(value);
case NEW_LO_SPACE:
return new_lo_space_->Contains(value);
case SHARED_LO_SPACE:
return shared_lo_space_->Contains(value);
case TRUSTED_LO_SPACE:
return trusted_lo_space_->Contains(value);
case RO_SPACE:
return ReadOnlyHeap::Contains(value);
}
UNREACHABLE();
}
bool Heap::InSpaceSlow(Address addr, AllocationSpace space) const {
if (memory_allocator()->IsOutsideAllocatedSpace(
addr, IsAnyCodeSpace(space) ? EXECUTABLE : NOT_EXECUTABLE)) {
return false;
}
if (!HasBeenSetUp()) return false;
switch (space) {
case NEW_SPACE:
return new_space_->ContainsSlow(addr);
case OLD_SPACE:
return old_space_->ContainsSlow(addr);
case CODE_SPACE:
return code_space_->ContainsSlow(addr);
case SHARED_SPACE:
return shared_space_->ContainsSlow(addr);
case TRUSTED_SPACE:
return trusted_space_->ContainsSlow(addr);
case LO_SPACE:
return lo_space_->ContainsSlow(addr);
case CODE_LO_SPACE:
return code_lo_space_->ContainsSlow(addr);
case NEW_LO_SPACE:
return new_lo_space_->ContainsSlow(addr);
case SHARED_LO_SPACE:
return shared_lo_space_->ContainsSlow(addr);
case TRUSTED_LO_SPACE:
return trusted_lo_space_->ContainsSlow(addr);
case RO_SPACE:
return read_only_space_->ContainsSlow(addr);
}
UNREACHABLE();
}
bool Heap::IsValidAllocationSpace(AllocationSpace space) {
switch (space) {
case NEW_SPACE:
case OLD_SPACE:
case CODE_SPACE:
case SHARED_SPACE:
case LO_SPACE:
case NEW_LO_SPACE:
case CODE_LO_SPACE:
case SHARED_LO_SPACE:
case TRUSTED_SPACE:
case TRUSTED_LO_SPACE:
case RO_SPACE:
return true;
default:
return false;
}
}
#ifdef DEBUG
void Heap::VerifyCountersAfterSweeping() {
MakeHeapIterable();
PagedSpaceIterator spaces(this);
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
space->VerifyCountersAfterSweeping(this);
}
}
void Heap::VerifyCountersBeforeConcurrentSweeping(GarbageCollector collector) {
if (v8_flags.minor_ms && new_space()) {
PagedSpaceBase* space = paged_new_space()->paged_space();
space->RefillFreeList();
space->VerifyCountersBeforeConcurrentSweeping();
}
if (collector != GarbageCollector::MARK_COMPACTOR) return;
PagedSpaceIterator spaces(this);
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
// We need to refine the counters on pages that are already swept and have
// not been moved over to the actual space. Otherwise, the AccountingStats
// are just an over approximation.
space->RefillFreeList();
space->VerifyCountersBeforeConcurrentSweeping();
}
}
void Heap::VerifyCommittedPhysicalMemory() {
PagedSpaceIterator spaces(this);
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
space->VerifyCommittedPhysicalMemory();
}
if (v8_flags.minor_ms && new_space()) {
paged_new_space()->paged_space()->VerifyCommittedPhysicalMemory();
}
}
#endif // DEBUG
void Heap::IterateWeakRoots(RootVisitor* v, base::EnumSet<SkipRoot> options) {
DCHECK(!options.contains(SkipRoot::kWeak));
if (!options.contains(SkipRoot::kOldGeneration) &&
!options.contains(SkipRoot::kUnserializable) &&
isolate()->OwnsStringTables()) {
// Do not visit for the following reasons.
// - Serialization, since the string table is custom serialized.
// - If we are skipping old generation, since all internalized strings
// are in old space.
// - If the string table is shared and this is not the shared heap,
// since all internalized strings are in the shared heap.
isolate()->string_table()->IterateElements(v);
}
v->Synchronize(VisitorSynchronization::kStringTable);
if (!options.contains(SkipRoot::kExternalStringTable) &&
!options.contains(SkipRoot::kUnserializable)) {
// Scavenge collections have special processing for this.
// Do not visit for serialization, since the external string table will
// be populated from scratch upon deserialization.
external_string_table_.IterateAll(v);
}
v->Synchronize(VisitorSynchronization::kExternalStringsTable);
}
void Heap::IterateSmiRoots(RootVisitor* v) {
// Acquire execution access since we are going to read stack limit values.
ExecutionAccess access(isolate());
v->VisitRootPointers(Root::kSmiRootList, nullptr,
roots_table().smi_roots_begin(),
roots_table().smi_roots_end());
v->Synchronize(VisitorSynchronization::kSmiRootList);
}
// We cannot avoid stale handles to left-trimmed objects, but can only make
// sure all handles still needed are updated. Filter out a stale pointer
// and clear the slot to allow post processing of handles (needed because
// the sweeper might actually free the underlying page).
class ClearStaleLeftTrimmedHandlesVisitor : public RootVisitor {
public:
explicit ClearStaleLeftTrimmedHandlesVisitor(Heap* heap)
: heap_(heap)
#if V8_COMPRESS_POINTERS
,
cage_base_(heap->isolate())
#endif // V8_COMPRESS_POINTERS
{
USE(heap_);
}
void VisitRootPointer(Root root, const char* description,
FullObjectSlot p) override {
FixHandle(p);
}
void VisitRootPointers(Root root, const char* description,
FullObjectSlot start, FullObjectSlot end) override {
for (FullObjectSlot p = start; p < end; ++p) {
FixHandle(p);
}
}
// The pointer compression cage base value used for decompression of all
// tagged values except references to InstructionStream objects.
PtrComprCageBase cage_base() const {
#if V8_COMPRESS_POINTERS
return cage_base_;
#else
return PtrComprCageBase{};
#endif // V8_COMPRESS_POINTERS
}
private:
inline void FixHandle(FullObjectSlot p) {
if (!IsHeapObject(*p)) return;
Tagged<HeapObject> current = HeapObject::cast(*p);
if (!current->map_word(cage_base(), kRelaxedLoad).IsForwardingAddress() &&
IsFreeSpaceOrFiller(current, cage_base())) {
#ifdef DEBUG
// We need to find a FixedArrayBase map after walking the fillers.
while (
!current->map_word(cage_base(), kRelaxedLoad).IsForwardingAddress() &&
IsFreeSpaceOrFiller(current, cage_base())) {
Address next = current.ptr();
if (current->map(cage_base()) ==
ReadOnlyRoots(heap_).one_pointer_filler_map()) {
next += kTaggedSize;
} else if (current->map(cage_base()) ==
ReadOnlyRoots(heap_).two_pointer_filler_map()) {
next += 2 * kTaggedSize;
} else {
next += current->Size();
}
current = HeapObject::cast(Tagged<Object>(next));
}
DCHECK(
current->map_word(cage_base(), kRelaxedLoad).IsForwardingAddress() ||
IsFixedArrayBase(current, cage_base()));
#endif // DEBUG
p.store(Smi::zero());
}
}
Heap* heap_;
#if V8_COMPRESS_POINTERS
const PtrComprCageBase cage_base_;
#endif // V8_COMPRESS_POINTERS
};
void Heap::IterateRoots(RootVisitor* v, base::EnumSet<SkipRoot> options,
IterateRootsMode roots_mode) {
v->VisitRootPointers(Root::kStrongRootList, nullptr,
roots_table().strong_roots_begin(),
roots_table().strong_roots_end());
v->Synchronize(VisitorSynchronization::kStrongRootList);
isolate_->bootstrapper()->Iterate(v);
v->Synchronize(VisitorSynchronization::kBootstrapper);
Relocatable::Iterate(isolate_, v);
v->Synchronize(VisitorSynchronization::kRelocatable);
isolate_->debug()->Iterate(v);
v->Synchronize(VisitorSynchronization::kDebug);
isolate_->compilation_cache()->Iterate(v);
v->Synchronize(VisitorSynchronization::kCompilationCache);
const bool skip_iterate_builtins =
options.contains(SkipRoot::kOldGeneration) ||
(Builtins::kCodeObjectsAreInROSpace &&
options.contains(SkipRoot::kReadOnlyBuiltins) &&
// Prior to ReadOnlyPromotion, builtins may be on the mutable heap.
!isolate_->serializer_enabled());
if (!skip_iterate_builtins) {
IterateBuiltins(v);
v->Synchronize(VisitorSynchronization::kBuiltins);
}
// Iterate over pointers being held by inactive threads.
isolate_->thread_manager()->Iterate(v);
v->Synchronize(VisitorSynchronization::kThreadManager);
// Visitors in this block only run when not serializing. These include:
//
// - Thread-local and stack.
// - Handles.
// - Microtasks.
// - The startup object cache.
//
// When creating real startup snapshot, these areas are expected to be empty.
// It is also possible to create a snapshot of a *running* isolate for testing
// purposes. In this case, these areas are likely not empty and will simply be
// skipped.
//
// The general guideline for adding visitors to this section vs. adding them
// above is that non-transient heap state is always visited, transient heap
// state is visited only when not serializing.
if (!options.contains(SkipRoot::kUnserializable)) {
if (!options.contains(SkipRoot::kTracedHandles)) {
// Young GCs always skip traced handles and visit them manually.
DCHECK(!options.contains(SkipRoot::kOldGeneration));
isolate_->traced_handles()->Iterate(v);
}
if (!options.contains(SkipRoot::kGlobalHandles)) {
// Young GCs always skip global handles and visit them manually.
DCHECK(!options.contains(SkipRoot::kOldGeneration));
if (options.contains(SkipRoot::kWeak)) {
isolate_->global_handles()->IterateStrongRoots(v);
} else {
isolate_->global_handles()->IterateAllRoots(v);
}
}
v->Synchronize(VisitorSynchronization::kGlobalHandles);
if (!options.contains(SkipRoot::kStack)) {
IterateStackRoots(v);
if (!options.contains(SkipRoot::kConservativeStack)) {
IterateConservativeStackRoots(v, roots_mode);
}
v->Synchronize(VisitorSynchronization::kStackRoots);
}
// Iterate over main thread handles in handle scopes.
if (!options.contains(SkipRoot::kMainThreadHandles)) {
// Clear main thread handles with stale references to left-trimmed
// objects. The GC would crash on such stale references.
ClearStaleLeftTrimmedHandlesVisitor left_trim_visitor(this);
isolate_->handle_scope_implementer()->Iterate(&left_trim_visitor);
isolate_->handle_scope_implementer()->Iterate(v);
}
// Iterate local handles for all local heaps.
safepoint_->Iterate(v);
// Iterates all persistent handles.
isolate_->persistent_handles_list()->Iterate(v, isolate_);
v->Synchronize(VisitorSynchronization::kHandleScope);
if (options.contains(SkipRoot::kOldGeneration)) {
isolate_->eternal_handles()->IterateYoungRoots(v);
} else {
isolate_->eternal_handles()->IterateAllRoots(v);
}
v->Synchronize(VisitorSynchronization::kEternalHandles);
// Iterate over pending Microtasks stored in MicrotaskQueues.
MicrotaskQueue* default_microtask_queue =
isolate_->default_microtask_queue();
if (default_microtask_queue) {
MicrotaskQueue* microtask_queue = default_microtask_queue;
do {
microtask_queue->IterateMicrotasks(v);
microtask_queue = microtask_queue->next();
} while (microtask_queue != default_microtask_queue);
}
v->Synchronize(VisitorSynchronization::kMicroTasks);
// Iterate over other strong roots (currently only identity maps and
// deoptimization entries).
for (StrongRootsEntry* current = strong_roots_head_; current;
current = current->next) {
v->VisitRootPointers(Root::kStrongRoots, current->label, current->start,
current->end);
}
v->Synchronize(VisitorSynchronization::kStrongRoots);
// Iterate over the startup and shared heap object caches unless
// serializing or deserializing.
SerializerDeserializer::IterateStartupObjectCache(isolate_, v);
v->Synchronize(VisitorSynchronization::kStartupObjectCache);
// Iterate over shared heap object cache when the isolate owns this data
// structure. Isolates which own the shared heap object cache are:
// * Shared isolate
// * Shared space/main isolate
// * All isolates which do not use the shared heap feature.
//
// However, worker/client isolates do not own the shared heap object cache
// and should not iterate it.
if (isolate_->is_shared_space_isolate() || !isolate_->has_shared_space()) {
SerializerDeserializer::IterateSharedHeapObjectCache(isolate_, v);
v->Synchronize(VisitorSynchronization::kSharedHeapObjectCache);
}
}
if (!options.contains(SkipRoot::kWeak)) {
IterateWeakRoots(v, options);
}
}
void Heap::IterateRootsIncludingClients(RootVisitor* v,
base::EnumSet<SkipRoot> options) {
IterateRoots(v, options, IterateRootsMode::kMainIsolate);
if (isolate()->is_shared_space_isolate()) {
ClientRootVisitor<> client_root_visitor(v);
isolate()->global_safepoint()->IterateClientIsolates(
[v = &client_root_visitor, options](Isolate* client) {
client->heap()->IterateRoots(v, options,
IterateRootsMode::kClientIsolate);
});
}
}
void Heap::IterateWeakGlobalHandles(RootVisitor* v) {
isolate_->global_handles()->IterateWeakRoots(v);
isolate_->traced_handles()->Iterate(v);
}
void Heap::IterateBuiltins(RootVisitor* v) {
Builtins* builtins = isolate()->builtins();
for (Builtin builtin = Builtins::kFirst; builtin <= Builtins::kLast;
++builtin) {
const char* name = Builtins::name(builtin);
v->VisitRootPointer(Root::kBuiltins, name, builtins->builtin_slot(builtin));
}
for (Builtin builtin = Builtins::kFirst; builtin <= Builtins::kLastTier0;
++builtin) {
v->VisitRootPointer(Root::kBuiltins, Builtins::name(builtin),
builtins->builtin_tier0_slot(builtin));
}
// The entry table doesn't need to be updated since all builtins are embedded.
static_assert(Builtins::AllBuiltinsAreIsolateIndependent());
}
void Heap::IterateStackRoots(RootVisitor* v) { isolate_->Iterate(v); }
void Heap::IterateConservativeStackRoots(RootVisitor* v,
IterateRootsMode roots_mode) {
#ifdef V8_ENABLE_CONSERVATIVE_STACK_SCANNING
if (!IsGCWithStack()) return;
TRACE_GC(tracer(), GCTracer::Scope::CONSERVATIVE_STACK_SCANNING);
// In case of a shared GC, we're interested in the main isolate for CSS.
Isolate* main_isolate = roots_mode == IterateRootsMode::kClientIsolate
? isolate()->shared_space_isolate()
: isolate();
ConservativeStackVisitor stack_visitor(main_isolate, v);
stack().IteratePointersUntilMarker(&stack_visitor);
#endif // V8_ENABLE_CONSERVATIVE_STACK_SCANNING
}
// static
size_t Heap::DefaultMinSemiSpaceSize() {
#if ENABLE_HUGEPAGE
static constexpr size_t kMinSemiSpaceSize =
kHugePageSize * kPointerMultiplier;
#else
static constexpr size_t kMinSemiSpaceSize = 512 * KB * kPointerMultiplier;
#endif
static_assert(kMinSemiSpaceSize % (1 << kPageSizeBits) == 0);
return kMinSemiSpaceSize;
}
// static
size_t Heap::DefaultMaxSemiSpaceSize() {
#if ENABLE_HUGEPAGE
static constexpr size_t kMaxSemiSpaceCapacityBaseUnit =
kHugePageSize * 2 * kPointerMultiplier;
#else
static constexpr size_t kMaxSemiSpaceCapacityBaseUnit =
MB * kPointerMultiplier;
#endif
static_assert(kMaxSemiSpaceCapacityBaseUnit % (1 << kPageSizeBits) == 0);
size_t max_semi_space_size =
(v8_flags.minor_ms ? v8_flags.minor_ms_max_new_space_capacity_mb
: v8_flags.scavenger_max_new_space_capacity_mb) *
kMaxSemiSpaceCapacityBaseUnit;
DCHECK_EQ(0, max_semi_space_size % (1 << kPageSizeBits));
return max_semi_space_size;
}
// static
size_t Heap::OldGenerationToSemiSpaceRatio() {
DCHECK(!v8_flags.minor_ms);
static constexpr size_t kOldGenerationToSemiSpaceRatio =
128 * kHeapLimitMultiplier / kPointerMultiplier;
return kOldGenerationToSemiSpaceRatio;
}
// static
size_t Heap::OldGenerationToSemiSpaceRatioLowMemory() {
static constexpr size_t kOldGenerationToSemiSpaceRatioLowMemory =
256 * kHeapLimitMultiplier / kPointerMultiplier;
return kOldGenerationToSemiSpaceRatioLowMemory / (v8_flags.minor_ms ? 2 : 1);
}
void Heap::ConfigureHeap(const v8::ResourceConstraints& constraints) {
// Initialize max_semi_space_size_.
{
max_semi_space_size_ = DefaultMaxSemiSpaceSize();
if (constraints.max_young_generation_size_in_bytes() > 0) {
max_semi_space_size_ = SemiSpaceSizeFromYoungGenerationSize(
constraints.max_young_generation_size_in_bytes());
}
if (v8_flags.max_semi_space_size > 0) {
max_semi_space_size_ =
static_cast<size_t>(v8_flags.max_semi_space_size) * MB;
} else if (v8_flags.max_heap_size > 0) {
size_t max_heap_size = static_cast<size_t>(v8_flags.max_heap_size) * MB;
size_t young_generation_size, old_generation_size;
if (v8_flags.max_old_space_size > 0) {
old_generation_size =
static_cast<size_t>(v8_flags.max_old_space_size) * MB;
young_generation_size = max_heap_size > old_generation_size
? max_heap_size - old_generation_size
: 0;
} else {
GenerationSizesFromHeapSize(max_heap_size, &young_generation_size,
&old_generation_size);
}
max_semi_space_size_ =
SemiSpaceSizeFromYoungGenerationSize(young_generation_size);
}
if (v8_flags.stress_compaction) {
// This will cause more frequent GCs when stressing.
max_semi_space_size_ = MB;
}
if (!v8_flags.minor_ms) {
// TODO(dinfuehr): Rounding to a power of 2 is technically no longer
// needed but yields best performance on Pixel2.
max_semi_space_size_ =
static_cast<size_t>(base::bits::RoundUpToPowerOfTwo64(
static_cast<uint64_t>(max_semi_space_size_)));
}
max_semi_space_size_ =
std::max(max_semi_space_size_, DefaultMinSemiSpaceSize());
max_semi_space_size_ = RoundDown<Page::kPageSize>(max_semi_space_size_);
}
// Initialize max_old_generation_size_ and max_global_memory_.
{
size_t max_old_generation_size = 700ul * (kSystemPointerSize / 4) * MB;
if (constraints.max_old_generation_size_in_bytes() > 0) {
max_old_generation_size = constraints.max_old_generation_size_in_bytes();
}
if (v8_flags.max_old_space_size > 0) {
max_old_generation_size =
static_cast<size_t>(v8_flags.max_old_space_size) * MB;
} else if (v8_flags.max_heap_size > 0) {
size_t max_heap_size = static_cast<size_t>(v8_flags.max_heap_size) * MB;
size_t young_generation_size =
YoungGenerationSizeFromSemiSpaceSize(max_semi_space_size_);
max_old_generation_size = max_heap_size > young_generation_size
? max_heap_size - young_generation_size
: 0;
}
max_old_generation_size =
std::max(max_old_generation_size, MinOldGenerationSize());
max_old_generation_size = std::min(max_old_generation_size,
AllocatorLimitOnMaxOldGenerationSize());
max_old_generation_size =
RoundDown<Page::kPageSize>(max_old_generation_size);
SetOldGenerationAndGlobalMaximumSize(max_old_generation_size);
}
CHECK_IMPLIES(
v8_flags.max_heap_size > 0,
v8_flags.max_semi_space_size == 0 || v8_flags.max_old_space_size == 0);
// Initialize initial_semispace_size_.
{
initial_semispace_size_ = DefaultMinSemiSpaceSize();
if (max_semi_space_size_ == DefaultMaxSemiSpaceSize()) {
// Start with at least 1*MB semi-space on machines with a lot of memory.
initial_semispace_size_ =
std::max(initial_semispace_size_, static_cast<size_t>(1 * MB));
}
if (constraints.initial_young_generation_size_in_bytes() > 0) {
initial_semispace_size_ = SemiSpaceSizeFromYoungGenerationSize(
constraints.initial_young_generation_size_in_bytes());
}
if (v8_flags.initial_heap_size > 0) {
size_t young_generation, old_generation;
Heap::GenerationSizesFromHeapSize(
static_cast<size_t>(v8_flags.initial_heap_size) * MB,
&young_generation, &old_generation);
initial_semispace_size_ =
SemiSpaceSizeFromYoungGenerationSize(young_generation);
}
if (v8_flags.min_semi_space_size > 0) {
initial_semispace_size_ =
static_cast<size_t>(v8_flags.min_semi_space_size) * MB;
}
initial_semispace_size_ =
std::min(initial_semispace_size_, max_semi_space_size_);
initial_semispace_size_ =
RoundDown<Page::kPageSize>(initial_semispace_size_);
}
if (v8_flags.lazy_new_space_shrinking) {
initial_semispace_size_ = max_semi_space_size_;
}
// Initialize initial_old_space_size_.
{
initial_old_generation_size_ = kMaxInitialOldGenerationSize;
if (constraints.initial_old_generation_size_in_bytes() > 0) {
initial_old_generation_size_ =
constraints.initial_old_generation_size_in_bytes();
old_generation_allocation_limit_configured_ = true;
}
if (v8_flags.initial_heap_size > 0) {
size_t initial_heap_size =
static_cast<size_t>(v8_flags.initial_heap_size) * MB;
size_t young_generation_size =
YoungGenerationSizeFromSemiSpaceSize(initial_semispace_size_);
initial_old_generation_size_ =
initial_heap_size > young_generation_size
? initial_heap_size - young_generation_size
: 0;
old_generation_allocation_limit_configured_ = true;
}
if (v8_flags.initial_old_space_size > 0) {
initial_old_generation_size_ =
static_cast<size_t>(v8_flags.initial_old_space_size) * MB;
old_generation_allocation_limit_configured_ = true;
}
initial_old_generation_size_ =
std::min(initial_old_generation_size_, max_old_generation_size() / 2);
initial_old_generation_size_ =
RoundDown<Page::kPageSize>(initial_old_generation_size_);
}
if (old_generation_allocation_limit_configured_) {
// If the embedder pre-configures the initial old generation size,
// then allow V8 to skip full GCs below that threshold.
min_old_generation_size_ = initial_old_generation_size_;
min_global_memory_size_ =
GlobalMemorySizeFromV8Size(min_old_generation_size_);
}
if (v8_flags.semi_space_growth_factor < 2) {
v8_flags.semi_space_growth_factor = 2;
}
initial_max_old_generation_size_ = max_old_generation_size();
ResetOldGenerationAndGlobalAllocationLimit();
// We rely on being able to allocate new arrays in paged spaces.
DCHECK(kMaxRegularHeapObjectSize >=
(JSArray::kHeaderSize +
FixedArray::SizeFor(JSArray::kInitialMaxFastElementArray) +
ALIGN_TO_ALLOCATION_ALIGNMENT(AllocationMemento::kSize)));
code_range_size_ = constraints.code_range_size_in_bytes();
configured_ = true;
}
void Heap::AddToRingBuffer(const char* string) {
size_t first_part =
std::min(strlen(string), kTraceRingBufferSize - ring_buffer_end_);
memcpy(trace_ring_buffer_ + ring_buffer_end_, string, first_part);
ring_buffer_end_ += first_part;
if (first_part < strlen(string)) {
ring_buffer_full_ = true;
size_t second_part = strlen(string) - first_part;
memcpy(trace_ring_buffer_, string + first_part, second_part);
ring_buffer_end_ = second_part;
}
}
void Heap::GetFromRingBuffer(char* buffer) {
size_t copied = 0;
if (ring_buffer_full_) {
copied = kTraceRingBufferSize - ring_buffer_end_;
memcpy(buffer, trace_ring_buffer_ + ring_buffer_end_, copied);
}
memcpy(buffer + copied, trace_ring_buffer_, ring_buffer_end_);
}
void Heap::ConfigureHeapDefault() {
v8::ResourceConstraints constraints;
ConfigureHeap(constraints);
}
void Heap::RecordStats(HeapStats* stats, bool take_snapshot) {
*stats->start_marker = HeapStats::kStartMarker;
*stats->end_marker = HeapStats::kEndMarker;
*stats->ro_space_size = read_only_space_->Size();
*stats->ro_space_capacity = read_only_space_->Capacity();
*stats->new_space_size = NewSpaceSize();
*stats->new_space_capacity = NewSpaceCapacity();
*stats->old_space_size = old_space_->SizeOfObjects();
*stats->old_space_capacity = old_space_->Capacity();
*stats->code_space_size = code_space_->SizeOfObjects();
*stats->code_space_capacity = code_space_->Capacity();
*stats->map_space_size = 0;
*stats->map_space_capacity = 0;
*stats->lo_space_size = lo_space_->Size();
*stats->code_lo_space_size = code_lo_space_->Size();
isolate_->global_handles()->RecordStats(stats);
*stats->memory_allocator_size = memory_allocator()->Size();
*stats->memory_allocator_capacity =
memory_allocator()->Size() + memory_allocator()->Available();
*stats->os_error = base::OS::GetLastError();
// TODO(leszeks): Include the string table in both current and peak usage.
*stats->malloced_memory = isolate_->allocator()->GetCurrentMemoryUsage();
*stats->malloced_peak_memory = isolate_->allocator()->GetMaxMemoryUsage();
if (take_snapshot) {
HeapObjectIterator iterator(this);
for (Tagged<HeapObject> obj = iterator.Next(); !obj.is_null();
obj = iterator.Next()) {
InstanceType type = obj->map()->instance_type();
DCHECK(0 <= type && type <= LAST_TYPE);
stats->objects_per_type[type]++;
stats->size_per_type[type] += obj->Size();
}
}
if (stats->last_few_messages != nullptr)
GetFromRingBuffer(stats->last_few_messages);
}
size_t Heap::OldGenerationSizeOfObjects() const {
PagedSpaceIterator spaces(this);
size_t total = 0;
for (PagedSpace* space = spaces.Next(); space != nullptr;
space = spaces.Next()) {
total += space->SizeOfObjects();
}
if (shared_lo_space_) {
total += shared_lo_space_->SizeOfObjects();
}
return total + lo_space_->SizeOfObjects() + code_lo_space_->SizeOfObjects();
}
size_t Heap::EmbedderSizeOfObjects() const {
return cpp_heap_ ? CppHeap::From(cpp_heap_)->used_size() : 0;
}
size_t Heap::GlobalSizeOfObjects() const {
return OldGenerationSizeOfObjects() + EmbedderSizeOfObjects();
}
uint64_t Heap::AllocatedExternalMemorySinceMarkCompact() const {
return external_memory_.AllocatedSinceMarkCompact();
}
bool Heap::AllocationLimitOvershotByLargeMargin() const {
// This guards against too eager finalization in small heaps.
// The number is chosen based on v8.browsing_mobile on Nexus 7v2.
constexpr size_t kMarginForSmallHeaps = 32u * MB;
uint64_t size_now =
OldGenerationSizeOfObjects() + AllocatedExternalMemorySinceMarkCompact();
const size_t v8_overshoot = old_generation_allocation_limit() < size_now
? size_now - old_generation_allocation_limit()
: 0;
const size_t global_overshoot =
global_allocation_limit_ < GlobalSizeOfObjects()
? GlobalSizeOfObjects() - global_allocation_limit_
: 0;
// Bail out if the V8 and global sizes are still below their respective
// limits.
if (v8_overshoot == 0 && global_overshoot == 0) {
return false;
}
// Overshoot margin is 50% of allocation limit or half-way to the max heap
// with special handling of small heaps.
const size_t v8_margin = std::min(
std::max(old_generation_allocation_limit() / 2, kMarginForSmallHeaps),
(max_old_generation_size() - old_generation_allocation_limit()) / 2);
const size_t global_margin =
std::min(std::max(global_allocation_limit_ / 2, kMarginForSmallHeaps),
(max_global_memory_size_ - global_allocation_limit_) / 2);
return v8_overshoot >= v8_margin || global_overshoot >= global_margin;
}
bool Heap::ShouldOptimizeForLoadTime() {
return isolate()->rail_mode() == PERFORMANCE_LOAD &&
!AllocationLimitOvershotByLargeMargin() &&
MonotonicallyIncreasingTimeInMs() <
isolate()->LoadStartTimeMs() + kMaxLoadTimeMs;
}
// This predicate is called when an old generation space cannot allocated from
// the free list and is about to add a new page. Returning false will cause a
// major GC. It happens when the old generation allocation limit is reached and
// - either we need to optimize for memory usage,
// - or the incremental marking is not in progress and we cannot start it.
bool Heap::ShouldExpandOldGenerationOnSlowAllocation(LocalHeap* local_heap,
AllocationOrigin origin) {
if (always_allocate() || OldGenerationSpaceAvailable() > 0) return true;
// We reached the old generation allocation limit.
// Allocations in the GC should always succeed if possible.
if (origin == AllocationOrigin::kGC) return true;
// Background threads need to be allowed to allocate without GC after teardown
// was initiated.
if (gc_state() == TEAR_DOWN) return true;
// If main thread is parked, it can't perform the GC. Fix the deadlock by
// allowing the allocation.
if (IsMainThreadParked(local_heap)) return true;
// If allocating isolate is deserialized at the moment then always allow
// allocation.
if (IsIsolateDeserializationActive(local_heap)) return true;
// Make it more likely that retry of allocation on background thread succeeds
if (IsRetryOfFailedAllocation(local_heap)) return true;
// Background thread requested GC, allocation should fail
if (CollectionRequested()) return false;
if (ShouldOptimizeForMemoryUsage()) return false;
if (ShouldOptimizeForLoadTime()) return true;
if (IsMajorMarkingComplete(local_heap)) {
return !AllocationLimitOvershotByLargeMargin();
}
if (incremental_marking()->IsStopped() &&
IncrementalMarkingLimitReached() == IncrementalMarkingLimit::kNoLimit) {
// We cannot start incremental marking.
return false;
}
return true;
}
bool Heap::IsRetryOfFailedAllocation(LocalHeap* local_heap) {
if (!local_heap) return false;
return local_heap->allocation_failed_;
}
bool Heap::IsMainThreadParked(LocalHeap* local_heap) {
if (!local_heap) return false;
return local_heap->main_thread_parked_;
}
bool Heap::IsMajorMarkingComplete(LocalHeap* local_heap) {
// Only check this on the main thread.
if (!local_heap || !local_heap->is_main_thread()) return false;
// But also ignore main threads of client isolates.
if (local_heap->heap() != this) {
DCHECK(isolate()->is_shared_space_isolate());
return false;
}
return incremental_marking()->IsMajorMarkingComplete();
}
Heap::HeapGrowingMode Heap::CurrentHeapGrowingMode() {
if (ShouldReduceMemory() || v8_flags.stress_compaction) {
return Heap::HeapGrowingMode::kMinimal;
}
if (ShouldOptimizeForMemoryUsage()) {
return Heap::HeapGrowingMode::kConservative;
}
if (memory_reducer() != nullptr && memory_reducer()->ShouldGrowHeapSlowly()) {
return Heap::HeapGrowingMode::kSlow;
}
return Heap::HeapGrowingMode::kDefault;
}
base::Optional<size_t> Heap::GlobalMemoryAvailable() {
size_t global_size = GlobalSizeOfObjects();
if (global_size < global_allocation_limit_)
return global_allocation_limit_ - global_size;
return 0;
}
double Heap::PercentToOldGenerationLimit() {
double size_at_gc = old_generation_size_at_last_gc_;
double size_now =
OldGenerationSizeOfObjects() + AllocatedExternalMemorySinceMarkCompact();
double current_bytes = size_now - size_at_gc;
double total_bytes = old_generation_allocation_limit() - size_at_gc;
return total_bytes > 0 ? (current_bytes / total_bytes) * 100.0 : 0;
}
double Heap::PercentToGlobalMemoryLimit() {
double size_at_gc = old_generation_size_at_last_gc_;
double size_now =
OldGenerationSizeOfObjects() + AllocatedExternalMemorySinceMarkCompact();
double current_bytes = size_now - size_at_gc;
double total_bytes = old_generation_allocation_limit() - size_at_gc;
return total_bytes > 0 ? (current_bytes / total_bytes) * 100.0 : 0;
}
// - kNoLimit means that either incremental marking is disabled or it is too
// early to start incremental marking.
// - kSoftLimit means that incremental marking should be started soon.
// - kHardLimit means that incremental marking should be started immediately.
// - kFallbackForEmbedderLimit means that incremental marking should be
// started as soon as the embedder does not allocate with high throughput
// anymore.
Heap::IncrementalMarkingLimit Heap::IncrementalMarkingLimitReached() {
// InstructionStream using an AlwaysAllocateScope assumes that the GC state
// does not change; that implies that no marking steps must be performed.
if (!incremental_marking()->CanBeStarted() || always_allocate()) {
// Incremental marking is disabled or it is too early to start.
return IncrementalMarkingLimit::kNoLimit;
}
if (v8_flags.stress_incremental_marking) {
return IncrementalMarkingLimit::kHardLimit;
}
if (incremental_marking()->IsBelowActivationThresholds()) {
// Incremental marking is disabled or it is too early to start.
return IncrementalMarkingLimit::kNoLimit;
}
if (ShouldStressCompaction() || HighMemoryPressure()) {
// If there is high memory pressure or stress testing is enabled, then
// start marking immediately.
return IncrementalMarkingLimit::kHardLimit;
}
if (v8_flags.stress_marking > 0) {
int current_percent = static_cast<int>(
std::max(PercentToOldGenerationLimit(), PercentToGlobalMemoryLimit()));
if (current_percent > 0) {
if (v8_flags.trace_stress_marking) {
isolate()->PrintWithTimestamp(
"[IncrementalMarking] %d%% of the memory limit reached\n",
current_percent);
}
if (v8_flags.fuzzer_gc_analysis) {
// Skips values >=100% since they already trigger marking.
if (current_percent < 100) {
max_marking_limit_reached_ =
std::max<double>(max_marking_limit_reached_, current_percent);
}
} else if (current_percent >= stress_marking_percentage_) {
return IncrementalMarkingLimit::kHardLimit;
}
}
}
if (v8_flags.incremental_marking_soft_trigger > 0 ||
v8_flags.incremental_marking_hard_trigger > 0) {
int current_percent = static_cast<int>(
std::max(PercentToOldGenerationLimit(), PercentToGlobalMemoryLimit()));
if (current_percent > v8_flags.incremental_marking_hard_trigger &&
v8_flags.incremental_marking_hard_trigger > 0) {
return IncrementalMarkingLimit::kHardLimit;
}
if (current_percent > v8_flags.incremental_marking_soft_trigger &&
v8_flags.incremental_marking_soft_trigger > 0) {
return IncrementalMarkingLimit::kSoftLimit;
}
return IncrementalMarkingLimit::kNoLimit;
}
size_t old_generation_space_available = OldGenerationSpaceAvailable();
const base::Optional<size_t> global_memory_available =
GlobalMemoryAvailable();
if (old_generation_space_available > NewSpaceTargetCapacity() &&
(!global_memory_available ||
global_memory_available > NewSpaceTargetCapacity())) {
if (cpp_heap() && !old_generation_allocation_limit_configured_ &&
gc_count_ == 0) {
// At this point the embedder memory is above the activation
// threshold. No GC happened so far and it's thus unlikely to get a
// configured heap any time soon. Start a memory reducer in this case
// which will wait until the allocation rate is low to trigger garbage
// collection.
return IncrementalMarkingLimit::kFallbackForEmbedderLimit;
}
return IncrementalMarkingLimit::kNoLimit;
}
if (ShouldOptimizeForMemoryUsage()) {
return IncrementalMarkingLimit::kHardLimit;
}
if (ShouldOptimizeForLoadTime()) {
return IncrementalMarkingLimit::kNoLimit;
}
if (old_generation_space_available == 0) {
return IncrementalMarkingLimit::kHardLimit;
}
if (global_memory_available && *global_memory_available == 0) {
return IncrementalMarkingLimit::kHardLimit;
}
return IncrementalMarkingLimit::kSoftLimit;
}
bool Heap::ShouldStressCompaction() const {
return v8_flags.stress_compaction && (gc_count_ & 1) != 0;
}
void Heap::EnableInlineAllocation() { inline_allocation_enabled_ = true; }
void Heap::DisableInlineAllocation() {
inline_allocation_enabled_ = false;
FreeMainThreadLinearAllocationAreas();
}
void Heap::SetUp(LocalHeap* main_thread_local_heap) {
DCHECK_NULL(main_thread_local_heap_);
main_thread_local_heap_ = main_thread_local_heap;
#ifdef V8_ENABLE_ALLOCATION_TIMEOUT
heap_allocator_.UpdateAllocationTimeout();
#endif // V8_ENABLE_ALLOCATION_TIMEOUT
#ifdef V8_ENABLE_THIRD_PARTY_HEAP
tp_heap_ = third_party_heap::Heap::New(isolate());
#endif
// Initialize heap spaces and initial maps and objects.
//
// If the heap is not yet configured (e.g. through the API), configure it.
// Configuration is based on the flags new-space-size (really the semispace
// size) and old-space-size if set or the initial values of semispace_size_
// and old_generation_size_ otherwise.
if (!configured_) ConfigureHeapDefault();
mmap_region_base_ =
reinterpret_cast<uintptr_t>(v8::internal::GetRandomMmapAddr()) &
~kMmapRegionMask;
v8::PageAllocator* code_page_allocator;
if (isolate_->RequiresCodeRange() || code_range_size_ != 0) {
const size_t requested_size =
code_range_size_ == 0 ? kMaximalCodeRangeSize : code_range_size_;
// When a target requires the code range feature, we put all code objects in
// a contiguous range of virtual address space, so that they can call each
// other with near calls.
#ifdef V8_COMPRESS_POINTERS_IN_SHARED_CAGE
// When sharing a pointer cage among Isolates, also share the
// CodeRange. isolate_->page_allocator() is the process-wide pointer
// compression cage's PageAllocator.
code_range_ = CodeRange::EnsureProcessWideCodeRange(
isolate_->page_allocator(), requested_size);
#else
code_range_ = std::make_unique<CodeRange>();
if (!code_range_->InitReservation(isolate_->page_allocator(),
requested_size)) {
V8::FatalProcessOutOfMemory(
isolate_, "Failed to reserve virtual memory for CodeRange");
}
#endif // V8_COMPRESS_POINTERS_IN_SHARED_CAGE
LOG(isolate_,
NewEvent("CodeRange",
reinterpret_cast<void*>(code_range_->reservation()->address()),
code_range_size_));
isolate_->AddCodeRange(code_range_->reservation()->region().begin(),
code_range_->reservation()->region().size());
code_page_allocator = code_range_->page_allocator();
} else {
code_page_allocator = isolate_->page_allocator();
}
task_runner_ = V8::GetCurrentPlatform()->GetForegroundTaskRunner(
reinterpret_cast<v8::Isolate*>(isolate()));
collection_barrier_.reset(new CollectionBarrier(this, this->task_runner_));
// Set up memory allocator.
memory_allocator_.reset(
new MemoryAllocator(isolate_, code_page_allocator, MaxReserved()));
sweeper_.reset(new Sweeper(this));
mark_compact_collector_.reset(new MarkCompactCollector(this));
scavenger_collector_.reset(new ScavengerCollector(this));
minor_mark_sweep_collector_.reset(new MinorMarkSweepCollector(this));
ephemeron_remembered_set_.reset(new EphemeronRememberedSet());
incremental_marking_.reset(
new IncrementalMarking(this, mark_compact_collector_->weak_objects()));
if (v8_flags.concurrent_marking || v8_flags.parallel_marking) {
concurrent_marking_.reset(
new ConcurrentMarking(this, mark_compact_collector_->weak_objects()));
} else {
concurrent_marking_.reset(new ConcurrentMarking(this, nullptr));
}
// Set up layout tracing callback.
if (V8_UNLIKELY(v8_flags.trace_gc_heap_layout)) {
v8::GCType gc_type = kGCTypeMarkSweepCompact;
if (V8_UNLIKELY(!v8_flags.trace_gc_heap_layout_ignore_minor_gc)) {
gc_type = static_cast<v8::GCType>(gc_type | kGCTypeScavenge |
kGCTypeMinorMarkSweep);
}
AddGCPrologueCallback(HeapLayoutTracer::GCProloguePrintHeapLayout, gc_type,
nullptr);
AddGCEpilogueCallback(HeapLayoutTracer::GCEpiloguePrintHeapLayout, gc_type,
nullptr);
}
}
void Heap::SetUpFromReadOnlyHeap(ReadOnlyHeap* ro_heap) {
DCHECK_NOT_NULL(ro_heap);
DCHECK_IMPLIES(read_only_space_ != nullptr,
read_only_space_ == ro_heap->read_only_space());
DCHECK_NULL(space_[RO_SPACE].get());
read_only_space_ = ro_heap->read_only_space();
heap_allocator_.SetReadOnlySpace(read_only_space_);
}
void Heap::ReplaceReadOnlySpace(SharedReadOnlySpace* space) {
CHECK(V8_SHARED_RO_HEAP_BOOL);
if (read_only_space_) {
read_only_space_->TearDown(memory_allocator());
delete read_only_space_;
}
read_only_space_ = space;
heap_allocator_.SetReadOnlySpace(read_only_space_);
}
class StressConcurrentAllocationObserver : public AllocationObserver {
public:
explicit StressConcurrentAllocationObserver(Heap* heap)
: AllocationObserver(1024), heap_(heap) {}
void Step(int bytes_allocated, Address, size_t) override {
DCHECK(heap_->deserialization_complete());
if (v8_flags.stress_concurrent_allocation) {
// Only schedule task if --stress-concurrent-allocation is enabled. This
// allows tests to disable flag even when Isolate was already initialized.
StressConcurrentAllocatorTask::Schedule(heap_->isolate());
}
heap_->RemoveAllocationObserversFromAllSpaces(this, this);
heap_->need_to_remove_stress_concurrent_allocation_observer_ = false;
}
private:
Heap* heap_;
};
namespace {
size_t ReturnNull() { return 0; }
} // namespace
void Heap::SetUpSpaces(LinearAllocationArea& new_allocation_info,
LinearAllocationArea& old_allocation_info) {
// Ensure SetUpFromReadOnlySpace has been ran.
DCHECK_NOT_NULL(read_only_space_);
if (!v8_flags.single_generation) {
if (v8_flags.minor_ms) {
space_[NEW_SPACE] = std::make_unique<PagedNewSpace>(
this, initial_semispace_size_, max_semi_space_size_,
new_allocation_info);
} else {
space_[NEW_SPACE] = std::make_unique<SemiSpaceNewSpace>(
this, initial_semispace_size_, max_semi_space_size_,
new_allocation_info);
}
new_space_ = static_cast<NewSpace*>(space_[NEW_SPACE].get());
space_[NEW_LO_SPACE] =
std::make_unique<NewLargeObjectSpace>(this, NewSpaceCapacity());
new_lo_space_ =
static_cast<NewLargeObjectSpace*>(space_[NEW_LO_SPACE].get());
}
space_[OLD_SPACE] = std::make_unique<OldSpace>(this, old_allocation_info);
old_space_ = static_cast<OldSpace*>(space_[OLD_SPACE].get());
space_[CODE_SPACE] = std::make_unique<CodeSpace>(this);
code_space_ = static_cast<CodeSpace*>(space_[CODE_SPACE].get());
if (isolate()->is_shared_space_isolate()) {
space_[SHARED_SPACE] = std::make_unique<SharedSpace>(this);
shared_space_ = static_cast<SharedSpace*>(space_[SHARED_SPACE].get());
}
space_[LO_SPACE] = std::make_unique<OldLargeObjectSpace>(this);
lo_space_ = static_cast<OldLargeObjectSpace*>(space_[LO_SPACE].get());
space_[CODE_LO_SPACE] = std::make_unique<CodeLargeObjectSpace>(this);
code_lo_space_ =
static_cast<CodeLargeObjectSpace*>(space_[CODE_LO_SPACE].get());
if (isolate()->is_shared_space_isolate()) {
space_[SHARED_LO_SPACE] = std::make_unique<SharedLargeObjectSpace>(this);
shared_lo_space_ =
static_cast<SharedLargeObjectSpace*>(space_[SHARED_LO_SPACE].get());
}
space_[TRUSTED_SPACE] = std::make_unique<TrustedSpace>(this);
trusted_space_ = static_cast<TrustedSpace*>(space_[TRUSTED_SPACE].get());
space_[TRUSTED_LO_SPACE] = std::make_unique<TrustedLargeObjectSpace>(this);
trusted_lo_space_ =
static_cast<TrustedLargeObjectSpace*>(space_[TRUSTED_LO_SPACE].get());
if (isolate()->has_shared_space()) {
Heap* heap = isolate()->shared_space_isolate()->heap();
shared_space_allocator_ = std::make_unique<ConcurrentAllocator>(
main_thread_local_heap(), heap->shared_space_,
ConcurrentAllocator::Context::kNotGC);
shared_allocation_space_ = heap->shared_space_;
shared_lo_allocation_space_ = heap->shared_lo_space_;
}
heap_allocator_.Setup();
main_thread_local_heap()->SetUpMainThread();
for (int i = 0; i < static_cast<int>(v8::Isolate::kUseCounterFeatureCount);
i++) {
deferred_counters_[i] = 0;
}
base::TimeTicks startup_time = base::TimeTicks::Now();
tracer_.reset(new GCTracer(this, startup_time));
array_buffer_sweeper_.reset(new ArrayBufferSweeper(this));
gc_idle_time_handler_.reset(new GCIdleTimeHandler());
memory_measurement_.reset(new MemoryMeasurement(isolate()));
if (v8_flags.memory_reducer) memory_reducer_.reset(new MemoryReducer(this));
if (V8_UNLIKELY(TracingFlags::is_gc_stats_enabled())) {
live_object_stats_.reset(new ObjectStats(this));
dead_object_stats_.reset(new ObjectStats(this));
}
if (Heap::AllocationTrackerForDebugging::IsNeeded()) {
allocation_tracker_for_debugging_ =
std::make_unique<Heap::AllocationTrackerForDebugging>(this);
}
LOG(isolate_, IntPtrTEvent("heap-capacity", Capacity()));
LOG(isolate_, IntPtrTEvent("heap-available", Available()));
SetGetExternallyAllocatedMemoryInBytesCallback(ReturnNull);
write_protect_code_memory_ = v8_flags.write_protect_code_memory;
#if V8_HEAP_USE_PKU_JIT_WRITE_PROTECT
if (RwxMemoryWriteScope::IsSupported()) {
// If PKU machinery is available then use it instead of conventional
// mprotect.
write_protect_code_memory_ = false;
}
#endif // V8_HEAP_USE_PKU_JIT_WRITE_PROTECT
if (new_space()) {
minor_gc_job_.reset(new MinorGCJob(this));
minor_gc_task_observer_.reset(new ScheduleMinorGCTaskObserver(this));
}
if (v8_flags.stress_marking > 0) {
stress_marking_percentage_ = NextStressMarkingLimit();
}
if (IsStressingScavenge()) {
stress_scavenge_observer_ = new StressScavengeObserver(this);
allocator()->new_space_allocator()->AddAllocationObserver(
stress_scavenge_observer_);
}
if (v8_flags.memory_balancer) {
mb_.reset(new MemoryBalancer(this, startup_time));
}
}
void Heap::InitializeHashSeed() {
DCHECK(!deserialization_complete_);
uint64_t new_hash_seed;
if (v8_flags.hash_seed == 0) {
int64_t rnd = isolate()->random_number_generator()->NextInt64();
new_hash_seed = static_cast<uint64_t>(rnd);
} else {
new_hash_seed = static_cast<uint64_t>(v8_flags.hash_seed);
}
ReadOnlyRoots(this).hash_seed()->copy_in(
0, reinterpret_cast<uint8_t*>(&new_hash_seed), kInt64Size);
}
std::shared_ptr<v8::TaskRunner> Heap::GetForegroundTaskRunner() const {
return task_runner_;
}
// static
void Heap::InitializeOncePerProcess() {
#ifdef V8_ENABLE_ALLOCATION_TIMEOUT
HeapAllocator::InitializeOncePerProcess();
#endif
MemoryAllocator::InitializeOncePerProcess();
if (v8_flags.predictable) {
::heap::base::WorklistBase::EnforcePredictableOrder();
}
}
void Heap::PrintMaxMarkingLimitReached() {
PrintF("\n### Maximum marking limit reached = %.02lf\n",
max_marking_limit_reached_);
}
void Heap::PrintMaxNewSpaceSizeReached() {
PrintF("\n### Maximum new space size reached = %.02lf\n",
stress_scavenge_observer_->MaxNewSpaceSizeReached());
}
int Heap::NextStressMarkingLimit() {
return isolate()->fuzzer_rng()->NextInt(v8_flags.stress_marking + 1);
}
void Heap::WeakenDescriptorArrays(
GlobalHandleVector<DescriptorArray> strong_descriptor_arrays) {
if (incremental_marking()->IsMajorMarking()) {
// During incremental/concurrent marking regular DescriptorArray objects are
// treated with custom weakness. This weakness depends on
// DescriptorArray::raw_gc_state() which is not set up properly upon
// deserialization. The strong arrays are transitioned to weak ones at the
// end of the GC.
mark_compact_collector()->RecordStrongDescriptorArraysForWeakening(
std::move(strong_descriptor_arrays));
return;
}
// No GC is running, weaken the arrays right away.
DisallowGarbageCollection no_gc;
Tagged<Map> descriptor_array_map =
ReadOnlyRoots(isolate()).descriptor_array_map();
for (auto it = strong_descriptor_arrays.begin();
it != strong_descriptor_arrays.end(); ++it) {
Tagged<DescriptorArray> array = it.raw();
DCHECK(IsStrongDescriptorArray(array));
array->set_map_safe_transition_no_write_barrier(descriptor_array_map);
DCHECK_EQ(array->raw_gc_state(kRelaxedLoad), 0);
}
}
void Heap::NotifyDeserializationComplete() {
PagedSpaceIterator spaces(this);
for (PagedSpace* s = spaces.Next(); s != nullptr; s = spaces.Next()) {
// Shared space is used concurrently and cannot be shrunk.
if (s->identity() == SHARED_SPACE) continue;
if (isolate()->snapshot_available()) s->ShrinkImmortalImmovablePages();
#ifdef DEBUG
// All pages right after bootstrapping must be marked as never-evacuate.
for (Page* p : *s) {
DCHECK(p->NeverEvacuate());
}
#endif // DEBUG
}
if (v8_flags.stress_concurrent_allocation) {
stress_concurrent_allocation_observer_.reset(
new StressConcurrentAllocationObserver(this));
AddAllocationObserversToAllSpaces(
stress_concurrent_allocation_observer_.get(),
stress_concurrent_allocation_observer_.get());
need_to_remove_stress_concurrent_allocation_observer_ = true;
}
// Deserialization will never create objects in new space.
DCHECK_IMPLIES(new_space(), new_space()->Size() == 0);
DCHECK_IMPLIES(new_lo_space(), new_lo_space()->Size() == 0);
deserialization_complete_ = true;
}
void Heap::NotifyBootstrapComplete() {
// This function is invoked for each native context creation. We are
// interested only in the first native context.
if (old_generation_capacity_after_bootstrap_ == 0) {
old_generation_capacity_after_bootstrap_ = OldGenerationCapacity();
}
}
void Heap::NotifyOldGenerationExpansion(AllocationSpace space,
MemoryChunk* chunk) {
// Do the same thing we would have done for background expansion.
NotifyOldGenerationExpansionBackground(space, chunk);
const size_t kMemoryReducerActivationThreshold = 1 * MB;
if (memory_reducer() != nullptr && old_generation_capacity_after_bootstrap_ &&
ms_count_ == 0 &&
OldGenerationCapacity() >= old_generation_capacity_after_bootstrap_ +
kMemoryReducerActivationThreshold &&
v8_flags.memory_reducer_for_small_heaps) {
memory_reducer()->NotifyPossibleGarbage();
}
}
void Heap::NotifyOldGenerationExpansionBackground(AllocationSpace space,
MemoryChunk* chunk) {
// Pages created during bootstrapping may contain immortal immovable objects.
if (!deserialization_complete()) {
DCHECK_NE(NEW_SPACE, chunk->owner()->identity());
chunk->MarkNeverEvacuate();
}
if (IsAnyCodeSpace(space)) {
isolate()->AddCodeMemoryChunk(chunk);
}
}
void Heap::SetEmbedderRootsHandler(EmbedderRootsHandler* handler) {
embedder_roots_handler_ = handler;
}
EmbedderRootsHandler* Heap::GetEmbedderRootsHandler() const {
return embedder_roots_handler_;
}
void Heap::AttachCppHeap(v8::CppHeap* cpp_heap) {
CHECK(!incremental_marking()->IsMarking());
CppHeap::From(cpp_heap)->AttachIsolate(isolate());
cpp_heap_ = cpp_heap;
}
void Heap::DetachCppHeap() {
CppHeap::From(cpp_heap_)->DetachIsolate();
cpp_heap_ = nullptr;
}
const cppgc::EmbedderStackState* Heap::overriden_stack_state() const {
const auto* cpp_heap = CppHeap::From(cpp_heap_);
return cpp_heap ? cpp_heap->override_stack_state() : nullptr;
}
void Heap::SetStackStart(void* stack_start) {
stack().SetStackStart(stack_start);
}
::heap::base::Stack& Heap::stack() { return isolate_->stack(); }
void Heap::StartTearDown() {
// Finish any ongoing sweeping to avoid stray background tasks still accessing
// the heap during teardown.
CompleteSweepingFull();
memory_allocator()->unmapper()->EnsureUnmappingCompleted();
if (v8_flags.concurrent_marking) {
concurrent_marking()->Pause();
}
SetGCState(TEAR_DOWN);
// Background threads may allocate and block until GC is performed. However
// this might never happen when the main thread tries to quit and doesn't
// process the event queue anymore. Avoid this deadlock by allowing all
// allocations after tear down was requested to make sure all background
// threads finish.
collection_barrier_->NotifyShutdownRequested();
// Main thread isn't going to allocate anymore.
main_thread_local_heap()->FreeLinearAllocationArea();
FreeMainThreadSharedLinearAllocationAreas();
}
void Heap::TearDownWithSharedHeap() {
DCHECK_EQ(gc_state(), TEAR_DOWN);
// Assert that there are no background threads left and no executable memory
// chunks are unprotected.
safepoint()->AssertMainThreadIsOnlyThread();
// Now that all threads are stopped, verify the heap before tearing down the
// heap/isolate.
HeapVerifier::VerifyHeapIfEnabled(this);
// Might use the external pointer which might be in the shared heap.
external_string_table_.TearDown();
// Publish shared object worklist for the main thread if incremental marking
// is enabled for the shared heap.
main_thread_local_heap()->marking_barrier()->PublishSharedIfNeeded();
}
void Heap::TearDown() {
DCHECK_EQ(gc_state(), TEAR_DOWN);
// Assert that there are no background threads left and no executable memory
// chunks are unprotected.
safepoint()->AssertMainThreadIsOnlyThread();
DCHECK(concurrent_marking()->IsStopped());
// It's too late for Heap::Verify() here, as parts of the Isolate are
// already gone by the time this is called.
UpdateMaximumCommitted();
if (v8_flags.fuzzer_gc_analysis) {
if (v8_flags.stress_marking > 0) {
PrintMaxMarkingLimitReached();
}
if (IsStressingScavenge()) {
PrintMaxNewSpaceSizeReached();
}
}
minor_gc_task_observer_.reset();
minor_gc_job_.reset();
if (need_to_remove_stress_concurrent_allocation_observer_) {
RemoveAllocationObserversFromAllSpaces(
stress_concurrent_allocation_observer_.get(),
stress_concurrent_allocation_observer_.get());
}
stress_concurrent_allocation_observer_.reset();
if (IsStressingScavenge()) {
allocator()->new_space_allocator()->RemoveAllocationObserver(
stress_scavenge_observer_);
delete stress_scavenge_observer_;
stress_scavenge_observer_ = nullptr;
}
if (mark_compact_collector_) {
mark_compact_collector_->TearDown();
mark_compact_collector_.reset();
}
if (minor_mark_sweep_collector_) {
minor_mark_sweep_collector_->TearDown();
minor_mark_sweep_collector_.reset();
}
sweeper_->TearDown();
sweeper_.reset();
scavenger_collector_.reset();
array_buffer_sweeper_.reset();
incremental_marking_.reset();
concurrent_marking_.reset();
gc_idle_time_handler_.reset();
memory_measurement_.reset();
allocation_tracker_for_debugging_.reset();
ephemeron_remembered_set_.reset();
if (memory_reducer_ != nullptr) {
memory_reducer_->TearDown();
memory_reducer_.reset();
}
live_object_stats_.reset();
dead_object_stats_.reset();
embedder_roots_handler_ = nullptr;
if (cpp_heap_) {
CppHeap::From(cpp_heap_)->DetachIsolate();
cpp_heap_ = nullptr;
}
tracer_.reset();
pretenuring_handler_.reset();
shared_space_allocator_.reset();
{
CodePageHeaderModificationScope rwx_write_scope(
"Deletion of CODE_SPACE and CODE_LO_SPACE requires write access to "
"Code page headers");
for (int i = FIRST_MUTABLE_SPACE; i <= LAST_MUTABLE_SPACE; i++) {
space_[i].reset();
}
}
isolate()->read_only_heap()->OnHeapTearDown(this);
read_only_space_ = nullptr;
memory_allocator()->TearDown();
StrongRootsEntry* next = nullptr;
for (StrongRootsEntry* current = strong_roots_head_; current;
current = next) {
next = current->next;
delete current;
}
strong_roots_head_ = nullptr;
memory_allocator_.reset();
}
void Heap::AddGCPrologueCallback(v8::Isolate::GCCallbackWithData callback,
GCType gc_type, void* data) {
gc_prologue_callbacks_.Add(
callback, reinterpret_cast<v8::Isolate*>(isolate()), gc_type, data);
}
void Heap::RemoveGCPrologueCallback(v8::Isolate::GCCallbackWithData callback,
void* data) {
gc_prologue_callbacks_.Remove(callback, data);
}
void Heap::AddGCEpilogueCallback(v8::Isolate::GCCallbackWithData callback,
GCType gc_type, void* data) {
gc_epilogue_callbacks_.Add(
callback, reinterpret_cast<v8::Isolate*>(isolate()), gc_type, data);
}
void Heap::RemoveGCEpilogueCallback(v8::Isolate::GCCallbackWithData callback,
void* data) {
gc_epilogue_callbacks_.Remove(callback, data);
}
namespace {
Handle<WeakArrayList> CompactWeakArrayList(Heap* heap,
Handle<WeakArrayList> array,
AllocationType allocation) {
if (array->length() == 0) {
return array;
}
int new_length = array->CountLiveWeakReferences();
if (new_length == array->length()) {
return array;
}
Handle<WeakArrayList> new_array = WeakArrayList::EnsureSpace(
heap->isolate(),
handle(ReadOnlyRoots(heap).empty_weak_array_list(), heap->isolate()),
new_length, allocation);
// Allocation might have caused GC and turned some of the elements into
// cleared weak heap objects. Count the number of live references again and
// fill in the new array.
int copy_to = 0;
for (int i = 0; i < array->length(); i++) {
MaybeObject element = array->Get(i);
if (element->IsCleared()) continue;
new_array->Set(copy_to++, element);
}
new_array->set_length(copy_to);
return new_array;
}
} // anonymous namespace
void Heap::CompactWeakArrayLists() {
// Find known PrototypeUsers and compact them.
std::vector<Handle<PrototypeInfo>> prototype_infos;
{
HeapObjectIterator iterator(this);
for (Tagged<HeapObject> o = iterator.Next(); !o.is_null();
o = iterator.Next()) {
if (IsPrototypeInfo(*o)) {
Tagged<PrototypeInfo> prototype_info = Tagged<PrototypeInfo>::cast(o);
if (IsWeakArrayList(prototype_info->prototype_users())) {
prototype_infos.emplace_back(handle(prototype_info, isolate()));
}
}
}
}
for (auto& prototype_info : prototype_infos) {
Handle<WeakArrayList> array(
WeakArrayList::cast(prototype_info->prototype_users()), isolate());
DCHECK(InOldSpace(*array) ||
*array == ReadOnlyRoots(this).empty_weak_array_list());
Tagged<WeakArrayList> new_array = PrototypeUsers::Compact(
array, this, JSObject::PrototypeRegistryCompactionCallback,
AllocationType::kOld);
prototype_info->set_prototype_users(new_array);
}
// Find known WeakArrayLists and compact them.
Handle<WeakArrayList> scripts(script_list(), isolate());
DCHECK_IMPLIES(!V8_ENABLE_THIRD_PARTY_HEAP_BOOL, InOldSpace(*scripts));
scripts = CompactWeakArrayList(this, scripts, AllocationType::kOld);
set_script_list(*scripts);
}
void Heap::AddRetainedMaps(Handle<NativeContext> context,
GlobalHandleVector<Map> maps) {
Handle<WeakArrayList> array(WeakArrayList::cast(context->retained_maps()),
isolate());
if (array->IsFull()) {
CompactRetainedMaps(*array);
}
int cur_length = array->length();
array = WeakArrayList::EnsureSpace(
isolate(), array, cur_length + static_cast<int>(maps.size()) * 2);
if (*array != context->retained_maps()) {
context->set_retained_maps(*array);
}
{
DisallowGarbageCollection no_gc;
Tagged<WeakArrayList> raw_array = *array;
for (Handle<Map> map : maps) {
DCHECK(!map->InAnySharedSpace());
if (map->is_in_retained_map_list()) {
continue;
}
raw_array->Set(cur_length, HeapObjectReference::Weak(*map));
raw_array->Set(cur_length + 1,
Smi::FromInt(v8_flags.retain_maps_for_n_gc));
cur_length += 2;
raw_array->set_length(cur_length);
map->set_is_in_retained_map_list(true);
}
}
}
void Heap::CompactRetainedMaps(Tagged<WeakArrayList> retained_maps) {
int length = retained_maps->length();
int new_length = 0;
// This loop compacts the array by removing cleared weak cells.
for (int i = 0; i < length; i += 2) {
MaybeObject maybe_object = retained_maps->Get(i);
if (maybe_object->IsCleared()) {
continue;
}
DCHECK(maybe_object->IsWeak());
MaybeObject age = retained_maps->Get(i + 1);
DCHECK(IsSmi(age));
if (i != new_length) {
retained_maps->Set(new_length, maybe_object);
retained_maps->Set(new_length + 1, age);
}
new_length += 2;
}
Tagged<HeapObject> undefined = ReadOnlyRoots(this).undefined_value();
for (int i = new_length; i < length; i++) {
retained_maps->Set(i, HeapObjectReference::Strong(undefined));
}
if (new_length != length) retained_maps->set_length(new_length);
}
void Heap::FatalProcessOutOfMemory(const char* location) {
V8::FatalProcessOutOfMemory(isolate(), location, V8::kHeapOOM);
}
#ifdef DEBUG
class PrintHandleVisitor : public RootVisitor {
public:
void VisitRootPointers(Root root, const char* description,
FullObjectSlot start, FullObjectSlot end) override {
for (FullObjectSlot p = start; p < end; ++p)
PrintF(" handle %p to %p\n", p.ToVoidPtr(),
reinterpret_cast<void*>((*p).ptr()));
}
};
void Heap::PrintHandles() {
PrintF("Handles:\n");
PrintHandleVisitor v;
isolate_->handle_scope_implementer()->Iterate(&v);
}
#endif
class CheckHandleCountVisitor : public RootVisitor {
public:
CheckHandleCountVisitor() : handle_count_(0) {}
~CheckHandleCountVisitor() override {
CHECK_GT(HandleScope::kCheckHandleThreshold, handle_count_);
}
void VisitRootPointers(Root root, const char* description,
FullObjectSlot start, FullObjectSlot end) override {
handle_count_ += end - start;
}
private:
ptrdiff_t handle_count_;
};
void Heap::CheckHandleCount() {
CheckHandleCountVisitor v;
isolate_->handle_scope_implementer()->Iterate(&v);
}
void Heap::ClearRecordedSlot(Tagged<HeapObject> object, ObjectSlot slot) {
#ifndef V8_DISABLE_WRITE_BARRIERS
DCHECK(!IsLargeObject(object));
Page* page = Page::FromAddress(slot.address());
if (!page->InYoungGeneration()) {
DCHECK_EQ(page->owner_identity(), OLD_SPACE);
// We only need to remove that slot when sweeping is still in progress.
// Because in that case, a concurrent sweeper could find that memory and
// reuse it for subsequent allocations. The runtime could install another
// property at this slot but without unboxed doubles this will always be a
// tagged pointer.
if (!page->SweepingDone()) {
// No need to update old-to-old here since that remembered set is gone
// after a full GC and not re-recorded until sweeping is finished.
RememberedSet<OLD_TO_NEW>::Remove(page, slot.address());
RememberedSet<OLD_TO_NEW_BACKGROUND>::Remove(page, slot.address());
RememberedSet<OLD_TO_SHARED>::Remove(page, slot.address());
}
}
#endif
}
// static
int Heap::InsertIntoRememberedSetFromCode(MemoryChunk* chunk, Address slot) {
// This is called during runtime by a builtin, therefore it is run in the main
// thread.
DCHECK_NULL(LocalHeap::Current());
RememberedSet<OLD_TO_NEW>::Insert<AccessMode::NON_ATOMIC>(chunk, slot);
return 0;
}
#ifdef DEBUG
void Heap::VerifySlotRangeHasNoRecordedSlots(Address start, Address end) {
#ifndef V8_DISABLE_WRITE_BARRIERS
Page* page = Page::FromAddress(start);
RememberedSet<OLD_TO_NEW>::CheckNoneInRange(page, start, end);
RememberedSet<OLD_TO_NEW_BACKGROUND>::CheckNoneInRange(page, start, end);
RememberedSet<OLD_TO_SHARED>::CheckNoneInRange(page, start, end);
#endif
}
#endif
void Heap::ClearRecordedSlotRange(Address start, Address end) {
#ifndef V8_DISABLE_WRITE_BARRIERS
Page* page = Page::FromAddress(start);
DCHECK(!page->IsLargePage());
if (!page->InYoungGeneration()) {
// This method will be invoked on objects in shared space for
// internalization and string forwarding during GC.
DCHECK(page->owner_identity() == OLD_SPACE ||
page->owner_identity() == SHARED_SPACE);
if (!page->SweepingDone()) {
RememberedSet<OLD_TO_NEW>::RemoveRange(page, start, end,
SlotSet::KEEP_EMPTY_BUCKETS);
RememberedSet<OLD_TO_NEW_BACKGROUND>::RemoveRange(
page, start, end, SlotSet::KEEP_EMPTY_BUCKETS);
RememberedSet<OLD_TO_SHARED>::RemoveRange(page, start, end,
SlotSet::KEEP_EMPTY_BUCKETS);
}
}
#endif
}
PagedSpace* PagedSpaceIterator::Next() {
DCHECK_GE(counter_, FIRST_GROWABLE_PAGED_SPACE);
while (counter_ <= LAST_GROWABLE_PAGED_SPACE) {
PagedSpace* space = heap_->paged_space(counter_++);
if (space) return space;
}
return nullptr;
}
class HeapObjectsFilter {
public:
virtual ~HeapObjectsFilter() = default;
virtual bool SkipObject(Tagged<HeapObject> object) = 0;
};
class UnreachableObjectsFilter : public HeapObjectsFilter {
public:
explicit UnreachableObjectsFilter(Heap* heap) : heap_(heap) {
MarkReachableObjects();
}
~UnreachableObjectsFilter() override = default;
bool SkipObject(Tagged<HeapObject> object) override {
// Space object iterators should skip free space or filler objects.
DCHECK(!IsFreeSpaceOrFiller(object));
// If the bucket corresponding to the object's chunk does not exist, or the
// object is not found in the bucket, return true.
BasicMemoryChunk* chunk = BasicMemoryChunk::FromHeapObject(object);
if (reachable_.count(chunk) == 0) return true;
return reachable_[chunk]->count(object) == 0;
}
private:
using BucketType = std::unordered_set<Tagged<HeapObject>, Object::Hasher>;
bool MarkAsReachable(Tagged<HeapObject> object) {
// If the bucket corresponding to the object's chunk does not exist, then
// create an empty bucket.
BasicMemoryChunk* chunk = BasicMemoryChunk::FromHeapObject(object);
if (reachable_.count(chunk) == 0) {
reachable_[chunk] = std::make_unique<BucketType>();
}
// Insert the object if not present; return whether it was indeed inserted.
if (reachable_[chunk]->count(object)) return false;
reachable_[chunk]->insert(object);
return true;
}
class MarkingVisitor : public ObjectVisitorWithCageBases, public RootVisitor {
public:
explicit MarkingVisitor(UnreachableObjectsFilter* filter)
: ObjectVisitorWithCageBases(filter->heap_), filter_(filter) {}
void VisitMapPointer(Tagged<HeapObject> object) override {
MarkHeapObject(Map::unchecked_cast(object->map(cage_base())));
}
void VisitPointers(Tagged<HeapObject> host, ObjectSlot start,
ObjectSlot end) override {
MarkPointers(MaybeObjectSlot(start), MaybeObjectSlot(end));
}
void VisitPointers(Tagged<HeapObject> host, MaybeObjectSlot start,
MaybeObjectSlot end) final {
MarkPointers(start, end);
}
void VisitInstructionStreamPointer(Tagged<Code> host,
InstructionStreamSlot slot) override {
Tagged<Object> maybe_code = slot.load(code_cage_base());
Tagged<HeapObject> heap_object;
if (maybe_code.GetHeapObject(&heap_object)) {
MarkHeapObject(heap_object);
}
}
void VisitCodeTarget(Tagged<InstructionStream> host,
RelocInfo* rinfo) final {
Tagged<InstructionStream> target =
InstructionStream::FromTargetAddress(rinfo->target_address());
MarkHeapObject(target);
}
void VisitEmbeddedPointer(Tagged<InstructionStream> host,
RelocInfo* rinfo) final {
MarkHeapObject(rinfo->target_object(cage_base()));
}
void VisitRootPointers(Root root, const char* description,
FullObjectSlot start, FullObjectSlot end) override {
MarkPointersImpl(start, end);
}
void VisitRootPointers(Root root, const char* description,
OffHeapObjectSlot start,
OffHeapObjectSlot end) override {
MarkPointersImpl(start, end);
}
void TransitiveClosure() {
while (!marking_stack_.empty()) {
Tagged<HeapObject> obj = marking_stack_.back();
marking_stack_.pop_back();
obj->Iterate(cage_base(), this);
}
}
private:
void MarkPointers(MaybeObjectSlot start, MaybeObjectSlot end) {
MarkPointersImpl(start, end);
}
template <typename TSlot>
V8_INLINE void MarkPointersImpl(TSlot start, TSlot end) {
// Treat weak references as strong.
for (TSlot p = start; p < end; ++p) {
typename TSlot::TObject object = p.load(cage_base());
Tagged<HeapObject> heap_object;
if (object.GetHeapObject(&heap_object)) {
MarkHeapObject(heap_object);
}
}
}
V8_INLINE void MarkHeapObject(Tagged<HeapObject> heap_object) {
if (filter_->MarkAsReachable(heap_object)) {
marking_stack_.push_back(heap_object);
}
}
UnreachableObjectsFilter* filter_;
std::vector<Tagged<HeapObject>> marking_stack_;
};
friend class MarkingVisitor;
void MarkReachableObjects() {
MarkingVisitor visitor(this);
heap_->stack().SetMarkerIfNeededAndCallback(
[this, &visitor]() { heap_->IterateRoots(&visitor, {}); });
visitor.TransitiveClosure();
}
Heap* heap_;
DISALLOW_GARBAGE_COLLECTION(no_gc_)
std::unordered_map<BasicMemoryChunk*, std::unique_ptr<BucketType>,
base::hash<BasicMemoryChunk*>>
reachable_;
};
HeapObjectIterator::HeapObjectIterator(
Heap* heap, HeapObjectIterator::HeapObjectsFiltering filtering)
: HeapObjectIterator(
heap,
new SafepointScope(heap->isolate(),
heap->isolate()->is_shared_space_isolate()
? SafepointKind::kGlobal
: SafepointKind::kIsolate),
filtering) {}
HeapObjectIterator::HeapObjectIterator(Heap* heap,
const SafepointScope& safepoint_scope,
HeapObjectsFiltering filtering)
: HeapObjectIterator(heap, nullptr, filtering) {}
HeapObjectIterator::HeapObjectIterator(
Heap* heap, SafepointScope* safepoint_scope_or_nullptr,
HeapObjectsFiltering filtering)
: heap_(heap),
safepoint_scope_(safepoint_scope_or_nullptr),
space_iterator_(heap_) {
heap_->MakeHeapIterable();
switch (filtering) {
case kFilterUnreachable:
filter_ = std::make_unique<UnreachableObjectsFilter>(heap_);
break;
default:
break;
}
// Start the iteration.
CHECK(space_iterator_.HasNext());
object_iterator_ = space_iterator_.Next()->GetObjectIterator(heap_);
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) heap_->tp_heap_->ResetIterator();
}
HeapObjectIterator::~HeapObjectIterator() = default;
Tagged<HeapObject> HeapObjectIterator::Next() {
if (!filter_) return NextObject();
Tagged<HeapObject> obj = NextObject();
while (!obj.is_null() && filter_->SkipObject(obj)) obj = NextObject();
return obj;
}
Tagged<HeapObject> HeapObjectIterator::NextObject() {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) return heap_->tp_heap_->NextObject();
// No iterator means we are done.
if (!object_iterator_) return Tagged<HeapObject>();
Tagged<HeapObject> obj = object_iterator_->Next();
// If the current iterator has more objects we are fine.
if (!obj.is_null()) return obj;
// Go though the spaces looking for one that has objects.
while (space_iterator_.HasNext()) {
object_iterator_ = space_iterator_.Next()->GetObjectIterator(heap_);
obj = object_iterator_->Next();
if (!obj.is_null()) return obj;
}
// Done with the last space.
object_iterator_.reset();
return Tagged<HeapObject>();
}
void Heap::UpdateTotalGCTime(base::TimeDelta duration) {
total_gc_time_ms_ += duration;
}
void Heap::ExternalStringTable::CleanUpYoung() {
int last = 0;
Isolate* isolate = heap_->isolate();
for (size_t i = 0; i < young_strings_.size(); ++i) {
Tagged<Object> o = young_strings_[i];
if (IsTheHole(o, isolate)) {
continue;
}
// The real external string is already in one of these vectors and was or
// will be processed. Re-processing it will add a duplicate to the vector.
if (IsThinString(o)) continue;
DCHECK(IsExternalString(o));
if (InYoungGeneration(o)) {
young_strings_[last++] = o;
} else {
old_strings_.push_back(o);
}
}
young_strings_.resize(last);
}
void Heap::ExternalStringTable::CleanUpAll() {
CleanUpYoung();
int last = 0;
Isolate* isolate = heap_->isolate();
for (size_t i = 0; i < old_strings_.size(); ++i) {
Tagged<Object> o = old_strings_[i];
if (IsTheHole(o, isolate)) {
continue;
}
// The real external string is already in one of these vectors and was or
// will be processed. Re-processing it will add a duplicate to the vector.
if (IsThinString(o)) continue;
DCHECK(IsExternalString(o));
DCHECK(!InYoungGeneration(o));
old_strings_[last++] = o;
}
old_strings_.resize(last);
if (v8_flags.verify_heap && !v8_flags.enable_third_party_heap) {
Verify();
}
}
void Heap::ExternalStringTable::TearDown() {
for (size_t i = 0; i < young_strings_.size(); ++i) {
Tagged<Object> o = young_strings_[i];
// Dont finalize thin strings.
if (IsThinString(o)) continue;
heap_->FinalizeExternalString(ExternalString::cast(o));
}
young_strings_.clear();
for (size_t i = 0; i < old_strings_.size(); ++i) {
Tagged<Object> o = old_strings_[i];
// Dont finalize thin strings.
if (IsThinString(o)) continue;
heap_->FinalizeExternalString(ExternalString::cast(o));
}
old_strings_.clear();
}
void Heap::RememberUnmappedPage(Address page, bool compacted) {
// Tag the page pointer to make it findable in the dump file.
if (compacted) {
page ^= 0xC1EAD & (Page::kPageSize - 1); // Cleared.
} else {
page ^= 0x1D1ED & (Page::kPageSize - 1); // I died.
}
remembered_unmapped_pages_[remembered_unmapped_pages_index_] = page;
remembered_unmapped_pages_index_++;
remembered_unmapped_pages_index_ %= kRememberedUnmappedPages;
}
size_t Heap::YoungArrayBufferBytes() {
return array_buffer_sweeper()->YoungBytes();
}
size_t Heap::OldArrayBufferBytes() {
return array_buffer_sweeper()->OldBytes();
}
StrongRootsEntry* Heap::RegisterStrongRoots(const char* label,
FullObjectSlot start,
FullObjectSlot end) {
// We're either on the main thread, or in a background thread with an active
// local heap.
DCHECK(isolate()->CurrentLocalHeap()->IsRunning());
base::MutexGuard guard(&strong_roots_mutex_);
StrongRootsEntry* entry = new StrongRootsEntry(label);
entry->start = start;
entry->end = end;
entry->prev = nullptr;
entry->next = strong_roots_head_;
if (strong_roots_head_) {
DCHECK_NULL(strong_roots_head_->prev);
strong_roots_head_->prev = entry;
}
strong_roots_head_ = entry;
return entry;
}
void Heap::UpdateStrongRoots(StrongRootsEntry* entry, FullObjectSlot start,
FullObjectSlot end) {
entry->start = start;
entry->end = end;
}
void Heap::UnregisterStrongRoots(StrongRootsEntry* entry) {
// We're either on the main thread, or in a background thread with an active
// local heap.
DCHECK(isolate()->CurrentLocalHeap()->IsRunning());
base::MutexGuard guard(&strong_roots_mutex_);
StrongRootsEntry* prev = entry->prev;
StrongRootsEntry* next = entry->next;
if (prev) prev->next = next;
if (next) next->prev = prev;
if (strong_roots_head_ == entry) {
DCHECK_NULL(prev);
strong_roots_head_ = next;
}
delete entry;
}
void Heap::SetBuiltinsConstantsTable(Tagged<FixedArray> cache) {
set_builtins_constants_table(cache);
}
void Heap::SetDetachedContexts(Tagged<WeakArrayList> detached_contexts) {
set_detached_contexts(detached_contexts);
}
void Heap::PostFinalizationRegistryCleanupTaskIfNeeded() {
// Only one cleanup task is posted at a time.
if (!HasDirtyJSFinalizationRegistries() ||
is_finalization_registry_cleanup_task_posted_) {
return;
}
auto task = std::make_unique<FinalizationRegistryCleanupTask>(this);
task_runner_->PostNonNestableTask(std::move(task));
is_finalization_registry_cleanup_task_posted_ = true;
}
void Heap::EnqueueDirtyJSFinalizationRegistry(
Tagged<JSFinalizationRegistry> finalization_registry,
std::function<void(Tagged<HeapObject> object, ObjectSlot slot,
Tagged<Object> target)>
gc_notify_updated_slot) {
// Add a FinalizationRegistry to the tail of the dirty list.
DCHECK(!HasDirtyJSFinalizationRegistries() ||
IsJSFinalizationRegistry(dirty_js_finalization_registries_list()));
DCHECK(IsUndefined(finalization_registry->next_dirty(), isolate()));
DCHECK(!finalization_registry->scheduled_for_cleanup());
finalization_registry->set_scheduled_for_cleanup(true);
if (IsUndefined(dirty_js_finalization_registries_list_tail(), isolate())) {
DCHECK(IsUndefined(dirty_js_finalization_registries_list(), isolate()));
set_dirty_js_finalization_registries_list(finalization_registry);
// dirty_js_finalization_registries_list_ is rescanned by
// ProcessWeakListRoots.
} else {
Tagged<JSFinalizationRegistry> tail = Tagged<JSFinalizationRegistry>::cast(
dirty_js_finalization_registries_list_tail());
tail->set_next_dirty(finalization_registry);
gc_notify_updated_slot(
tail, tail->RawField(JSFinalizationRegistry::kNextDirtyOffset),
finalization_registry);
}
set_dirty_js_finalization_registries_list_tail(finalization_registry);
// dirty_js_finalization_registries_list_tail_ is rescanned by
// ProcessWeakListRoots.
}
MaybeHandle<JSFinalizationRegistry> Heap::DequeueDirtyJSFinalizationRegistry() {
// Take a FinalizationRegistry from the head of the dirty list for fairness.
if (HasDirtyJSFinalizationRegistries()) {
Handle<JSFinalizationRegistry> head(
JSFinalizationRegistry::cast(dirty_js_finalization_registries_list()),
isolate());
set_dirty_js_finalization_registries_list(head->next_dirty());
head->set_next_dirty(ReadOnlyRoots(this).undefined_value());
if (*head == dirty_js_finalization_registries_list_tail()) {
set_dirty_js_finalization_registries_list_tail(
ReadOnlyRoots(this).undefined_value());
}
return head;
}
return {};
}
void Heap::RemoveDirtyFinalizationRegistriesOnContext(
Tagged<NativeContext> context) {
DisallowGarbageCollection no_gc;
Isolate* isolate = this->isolate();
Tagged<Object> prev = ReadOnlyRoots(isolate).undefined_value();
Tagged<Object> current = dirty_js_finalization_registries_list();
while (!IsUndefined(current, isolate)) {
Tagged<JSFinalizationRegistry> finalization_registry =
Tagged<JSFinalizationRegistry>::cast(current);
if (finalization_registry->native_context() == context) {
if (IsUndefined(prev, isolate)) {
set_dirty_js_finalization_registries_list(
finalization_registry->next_dirty());
} else {
Tagged<JSFinalizationRegistry>::cast(prev)->set_next_dirty(
finalization_registry->next_dirty());
}
finalization_registry->set_scheduled_for_cleanup(false);
current = finalization_registry->next_dirty();
finalization_registry->set_next_dirty(
ReadOnlyRoots(isolate).undefined_value());
} else {
prev = current;
current = finalization_registry->next_dirty();
}
}
set_dirty_js_finalization_registries_list_tail(prev);
}
void Heap::KeepDuringJob(Handle<HeapObject> target) {
DCHECK(IsUndefined(weak_refs_keep_during_job()) ||
IsOrderedHashSet(weak_refs_keep_during_job()));
Handle<OrderedHashSet> table;
if (IsUndefined(weak_refs_keep_during_job(), isolate())) {
table = isolate()->factory()->NewOrderedHashSet();
} else {
table =
handle(OrderedHashSet::cast(weak_refs_keep_during_job()), isolate());
}
table = OrderedHashSet::Add(isolate(), table, target).ToHandleChecked();
set_weak_refs_keep_during_job(*table);
}
void Heap::ClearKeptObjects() {
set_weak_refs_keep_during_job(ReadOnlyRoots(isolate()).undefined_value());
}
size_t Heap::NumberOfTrackedHeapObjectTypes() {
return ObjectStats::OBJECT_STATS_COUNT;
}
size_t Heap::ObjectCountAtLastGC(size_t index) {
if (live_object_stats_ == nullptr || index >= ObjectStats::OBJECT_STATS_COUNT)
return 0;
return live_object_stats_->object_count_last_gc(index);
}
size_t Heap::ObjectSizeAtLastGC(size_t index) {
if (live_object_stats_ == nullptr || index >= ObjectStats::OBJECT_STATS_COUNT)
return 0;
return live_object_stats_->object_size_last_gc(index);
}
bool Heap::GetObjectTypeName(size_t index, const char** object_type,
const char** object_sub_type) {
if (index >= ObjectStats::OBJECT_STATS_COUNT) return false;
switch (static_cast<int>(index)) {
#define COMPARE_AND_RETURN_NAME(name) \
case name: \
*object_type = #name; \
*object_sub_type = ""; \
return true;
INSTANCE_TYPE_LIST(COMPARE_AND_RETURN_NAME)
#undef COMPARE_AND_RETURN_NAME
#define COMPARE_AND_RETURN_NAME(name) \
case ObjectStats::FIRST_VIRTUAL_TYPE + ObjectStats::name: \
*object_type = #name; \
*object_sub_type = ""; \
return true;
VIRTUAL_INSTANCE_TYPE_LIST(COMPARE_AND_RETURN_NAME)
#undef COMPARE_AND_RETURN_NAME
}
return false;
}
size_t Heap::NumberOfNativeContexts() {
int result = 0;
Tagged<Object> context = native_contexts_list();
while (!IsUndefined(context, isolate())) {
++result;
Tagged<Context> native_context = Tagged<Context>::cast(context);
context = native_context->next_context_link();
}
return result;
}
std::vector<Handle<NativeContext>> Heap::FindAllNativeContexts() {
std::vector<Handle<NativeContext>> result;
Tagged<Object> context = native_contexts_list();
while (!IsUndefined(context, isolate())) {
Tagged<NativeContext> native_context = Tagged<NativeContext>::cast(context);
result.push_back(handle(native_context, isolate()));
context = native_context->next_context_link();
}
return result;
}
std::vector<Tagged<WeakArrayList>> Heap::FindAllRetainedMaps() {
std::vector<Tagged<WeakArrayList>> result;
Tagged<Object> context = native_contexts_list();
while (!IsUndefined(context, isolate())) {
Tagged<NativeContext> native_context = Tagged<NativeContext>::cast(context);
result.push_back(WeakArrayList::cast(native_context->retained_maps()));
context = native_context->next_context_link();
}
return result;
}
size_t Heap::NumberOfDetachedContexts() {
// The detached_contexts() array has two entries per detached context.
return detached_contexts()->length() / 2;
}
bool Heap::AllowedToBeMigrated(Tagged<Map> map, Tagged<HeapObject> obj,
AllocationSpace dst) {
// Object migration is governed by the following rules:
//
// 1) Objects in new-space can be migrated to the old space
// that matches their target space or they stay in new-space.
// 2) Objects in old-space stay in the same space when migrating.
// 3) Fillers (two or more words) can migrate due to left-trimming of
// fixed arrays in new-space or old space.
// 4) Fillers (one word) can never migrate, they are skipped by
// incremental marking explicitly to prevent invalid pattern.
//
// Since this function is used for debugging only, we do not place
// asserts here, but check everything explicitly.
if (map == ReadOnlyRoots(this).one_pointer_filler_map()) return false;
InstanceType type = map->instance_type();
MemoryChunk* chunk = MemoryChunk::FromHeapObject(obj);
AllocationSpace src = chunk->owner_identity();
switch (src) {
case NEW_SPACE:
return dst == NEW_SPACE || dst == OLD_SPACE;
case OLD_SPACE:
return dst == OLD_SPACE;
case CODE_SPACE:
return dst == CODE_SPACE && type == INSTRUCTION_STREAM_TYPE;
case SHARED_SPACE:
return dst == SHARED_SPACE;
case TRUSTED_SPACE:
return dst == TRUSTED_SPACE;
case LO_SPACE:
case CODE_LO_SPACE:
case NEW_LO_SPACE:
case SHARED_LO_SPACE:
case TRUSTED_LO_SPACE:
case RO_SPACE:
return false;
}
UNREACHABLE();
}
size_t Heap::EmbedderAllocationCounter() const {
return cpp_heap_ ? CppHeap::From(cpp_heap_)->allocated_size() : 0;
}
void Heap::CreateObjectStats() {
if (V8_LIKELY(!TracingFlags::is_gc_stats_enabled())) return;
if (!live_object_stats_) {
live_object_stats_.reset(new ObjectStats(this));
}
if (!dead_object_stats_) {
dead_object_stats_.reset(new ObjectStats(this));
}
}
Tagged<Map> Heap::GcSafeMapOfHeapObject(Tagged<HeapObject> object) {
PtrComprCageBase cage_base(isolate());
MapWord map_word = object->map_word(cage_base, kRelaxedLoad);
if (map_word.IsForwardingAddress()) {
return map_word.ToForwardingAddress(object)->map(cage_base);
}
return map_word.ToMap();
}
Tagged<GcSafeCode> Heap::GcSafeGetCodeFromInstructionStream(
Tagged<HeapObject> instruction_stream, Address inner_pointer) {
Tagged<InstructionStream> istream =
InstructionStream::unchecked_cast(instruction_stream);
DCHECK(!istream.is_null());
DCHECK(GcSafeInstructionStreamContains(istream, inner_pointer));
return GcSafeCode::unchecked_cast(istream->raw_code(kAcquireLoad));
}
bool Heap::GcSafeInstructionStreamContains(Tagged<InstructionStream> istream,
Address addr) {
Tagged<Map> map = GcSafeMapOfHeapObject(istream);
DCHECK_EQ(map, ReadOnlyRoots(this).instruction_stream_map());
Builtin builtin_lookup_result =
OffHeapInstructionStream::TryLookupCode(isolate(), addr);
if (Builtins::IsBuiltinId(builtin_lookup_result)) {
// Builtins don't have InstructionStream objects.
DCHECK(!Builtins::IsBuiltinId(istream->code(kAcquireLoad)->builtin_id()));
return false;
}
Address start = istream.address();
Address end = start + istream->SizeFromMap(map);
return start <= addr && addr < end;
}
base::Optional<Tagged<InstructionStream>>
Heap::GcSafeTryFindInstructionStreamForInnerPointer(Address inner_pointer) {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) {
Address start = tp_heap_->GetObjectFromInnerPointer(inner_pointer);
return InstructionStream::unchecked_cast(HeapObject::FromAddress(start));
}
base::Optional<Address> start =
ThreadIsolation::StartOfJitAllocationAt(inner_pointer);
if (start.has_value()) {
return InstructionStream::unchecked_cast(HeapObject::FromAddress(*start));
}
return {};
}
base::Optional<Tagged<GcSafeCode>> Heap::GcSafeTryFindCodeForInnerPointer(
Address inner_pointer) {
Builtin maybe_builtin =
OffHeapInstructionStream::TryLookupCode(isolate(), inner_pointer);
if (Builtins::IsBuiltinId(maybe_builtin)) {
return GcSafeCode::cast(isolate()->builtins()->code(maybe_builtin));
}
base::Optional<Tagged<InstructionStream>> maybe_istream =
GcSafeTryFindInstructionStreamForInnerPointer(inner_pointer);
if (!maybe_istream) return {};
return GcSafeGetCodeFromInstructionStream(*maybe_istream, inner_pointer);
}
Tagged<Code> Heap::FindCodeForInnerPointer(Address inner_pointer) {
return GcSafeFindCodeForInnerPointer(inner_pointer)->UnsafeCastToCode();
}
Tagged<GcSafeCode> Heap::GcSafeFindCodeForInnerPointer(Address inner_pointer) {
base::Optional<Tagged<GcSafeCode>> maybe_code =
GcSafeTryFindCodeForInnerPointer(inner_pointer);
// Callers expect that the code object is found.
CHECK(maybe_code.has_value());
return GcSafeCode::unchecked_cast(maybe_code.value());
}
base::Optional<Tagged<Code>> Heap::TryFindCodeForInnerPointerForPrinting(
Address inner_pointer) {
if (InSpaceSlow(inner_pointer, i::CODE_SPACE) ||
InSpaceSlow(inner_pointer, i::CODE_LO_SPACE) ||
i::OffHeapInstructionStream::PcIsOffHeap(isolate(), inner_pointer)) {
base::Optional<Tagged<GcSafeCode>> maybe_code =
GcSafeTryFindCodeForInnerPointer(inner_pointer);
if (maybe_code.has_value()) {
return maybe_code.value()->UnsafeCastToCode();
}
}
return {};
}
void Heap::CombinedGenerationalAndSharedBarrierSlow(Tagged<HeapObject> object,
Address slot,
Tagged<HeapObject> value) {
MemoryChunk* value_chunk = MemoryChunk::FromHeapObject(value);
if (value_chunk->InYoungGeneration()) {
Heap::GenerationalBarrierSlow(object, slot, value);
} else {
DCHECK(value_chunk->InWritableSharedSpace());
DCHECK(!object->InWritableSharedSpace());
Heap::SharedHeapBarrierSlow(object, slot);
}
}
void Heap::CombinedGenerationalAndSharedEphemeronBarrierSlow(
Tagged<EphemeronHashTable> table, Address slot, Tagged<HeapObject> value) {
MemoryChunk* value_chunk = MemoryChunk::FromHeapObject(value);
if (value_chunk->InYoungGeneration()) {
MemoryChunk* table_chunk = MemoryChunk::FromHeapObject(table);
table_chunk->heap()->RecordEphemeronKeyWrite(table, slot);
} else {
DCHECK(value_chunk->InWritableSharedSpace());
DCHECK(!table->InWritableSharedSpace());
Heap::SharedHeapBarrierSlow(table, slot);
}
}
void Heap::GenerationalBarrierSlow(Tagged<HeapObject> object, Address slot,
Tagged<HeapObject> value) {
MemoryChunk* chunk = MemoryChunk::FromHeapObject(object);
if (LocalHeap::Current() == nullptr) {
RememberedSet<OLD_TO_NEW>::Insert<AccessMode::NON_ATOMIC>(chunk, slot);
} else {
RememberedSet<OLD_TO_NEW_BACKGROUND>::Insert<AccessMode::ATOMIC>(chunk,
slot);
}
}
void Heap::SharedHeapBarrierSlow(Tagged<HeapObject> object, Address slot) {
MemoryChunk* chunk = MemoryChunk::FromHeapObject(object);
DCHECK(!chunk->InWritableSharedSpace());
RememberedSet<OLD_TO_SHARED>::Insert<AccessMode::ATOMIC>(chunk, slot);
}
void Heap::RecordEphemeronKeyWrite(Tagged<EphemeronHashTable> table,
Address slot) {
ephemeron_remembered_set_->RecordEphemeronKeyWrite(table, slot);
}
void Heap::EphemeronKeyWriteBarrierFromCode(Address raw_object,
Address key_slot_address,
Isolate* isolate) {
Tagged<EphemeronHashTable> table =
Tagged<EphemeronHashTable>::cast(Tagged<Object>(raw_object));
ObjectSlot key_slot(key_slot_address);
CombinedEphemeronWriteBarrier(table, key_slot, *key_slot,
UPDATE_WRITE_BARRIER);
}
enum RangeWriteBarrierMode {
kDoGenerationalOrShared = 1 << 0,
kDoMarking = 1 << 1,
kDoEvacuationSlotRecording = 1 << 2,
};
template <int kModeMask, typename TSlot>
void Heap::WriteBarrierForRangeImpl(MemoryChunk* source_page,
Tagged<HeapObject> object, TSlot start_slot,
TSlot end_slot) {
// At least one of generational or marking write barrier should be requested.
static_assert(kModeMask & (kDoGenerationalOrShared | kDoMarking));
// kDoEvacuationSlotRecording implies kDoMarking.
static_assert(!(kModeMask & kDoEvacuationSlotRecording) ||
(kModeMask & kDoMarking));
MarkingBarrier* marking_barrier = nullptr;
if (kModeMask & kDoMarking) {
marking_barrier = WriteBarrier::CurrentMarkingBarrier(object);
}
MarkCompactCollector* collector = this->mark_compact_collector();
for (TSlot slot = start_slot; slot < end_slot; ++slot) {
typename TSlot::TObject value = *slot;
Tagged<HeapObject> value_heap_object;
if (!value.GetHeapObject(&value_heap_object)) continue;
if (kModeMask & kDoGenerationalOrShared) {
if (Heap::InYoungGeneration(value_heap_object)) {
RememberedSet<OLD_TO_NEW>::Insert<AccessMode::NON_ATOMIC>(
source_page, slot.address());
} else if (value_heap_object.InWritableSharedSpace()) {
RememberedSet<OLD_TO_SHARED>::Insert<AccessMode::ATOMIC>(
source_page, slot.address());
}
}
if (kModeMask & kDoMarking) {
marking_barrier->MarkValue(object, value_heap_object);
if (kModeMask & kDoEvacuationSlotRecording) {
collector->RecordSlot(source_page, HeapObjectSlot(slot),
value_heap_object);
}
}
}
}
// Instantiate Heap::WriteBarrierForRange() for ObjectSlot and MaybeObjectSlot.
template void Heap::WriteBarrierForRange<ObjectSlot>(Tagged<HeapObject> object,
ObjectSlot start_slot,
ObjectSlot end_slot);
template void Heap::WriteBarrierForRange<MaybeObjectSlot>(
Tagged<HeapObject> object, MaybeObjectSlot start_slot,
MaybeObjectSlot end_slot);
template <typename TSlot>
void Heap::WriteBarrierForRange(Tagged<HeapObject> object, TSlot start_slot,
TSlot end_slot) {
if (v8_flags.disable_write_barriers) return;
MemoryChunk* source_page = MemoryChunk::FromHeapObject(object);
base::Flags<RangeWriteBarrierMode> mode;
if (!source_page->InYoungGeneration() &&
!source_page->InWritableSharedSpace()) {
mode |= kDoGenerationalOrShared;
}
if (incremental_marking()->IsMarking()) {
mode |= kDoMarking;
if (!source_page->ShouldSkipEvacuationSlotRecording()) {
mode |= kDoEvacuationSlotRecording;
}
}
switch (mode) {
// Nothing to be done.
case 0:
return;
// Generational only.
case kDoGenerationalOrShared:
return WriteBarrierForRangeImpl<kDoGenerationalOrShared>(
source_page, object, start_slot, end_slot);
// Marking, no evacuation slot recording.
case kDoMarking:
return WriteBarrierForRangeImpl<kDoMarking>(source_page, object,
start_slot, end_slot);
// Marking with evacuation slot recording.
case kDoMarking | kDoEvacuationSlotRecording:
return WriteBarrierForRangeImpl<kDoMarking | kDoEvacuationSlotRecording>(
source_page, object, start_slot, end_slot);
// Generational and marking, no evacuation slot recording.
case kDoGenerationalOrShared | kDoMarking:
return WriteBarrierForRangeImpl<kDoGenerationalOrShared | kDoMarking>(
source_page, object, start_slot, end_slot);
// Generational and marking with evacuation slot recording.
case kDoGenerationalOrShared | kDoMarking | kDoEvacuationSlotRecording:
return WriteBarrierForRangeImpl<kDoGenerationalOrShared | kDoMarking |
kDoEvacuationSlotRecording>(
source_page, object, start_slot, end_slot);
default:
UNREACHABLE();
}
}
void Heap::GenerationalBarrierForCodeSlow(Tagged<InstructionStream> host,
RelocInfo* rinfo,
Tagged<HeapObject> object) {
DCHECK(InYoungGeneration(object));
const MarkCompactCollector::RecordRelocSlotInfo info =
MarkCompactCollector::ProcessRelocInfo(host, rinfo, object);
base::MutexGuard write_scope(info.memory_chunk->mutex());
RememberedSet<OLD_TO_NEW>::InsertTyped(info.memory_chunk, info.slot_type,
info.offset);
}
bool Heap::PageFlagsAreConsistent(Tagged<HeapObject> object) {
if (V8_ENABLE_THIRD_PARTY_HEAP_BOOL) {
return true;
}
BasicMemoryChunk* chunk = BasicMemoryChunk::FromHeapObject(object);
heap_internals::MemoryChunk* slim_chunk =
heap_internals::MemoryChunk::FromHeapObject(object);
// Slim chunk flags consistency.
CHECK_EQ(chunk->InYoungGeneration(), slim_chunk->InYoungGeneration());
CHECK_EQ(chunk->IsFlagSet(MemoryChunk::INCREMENTAL_MARKING),
slim_chunk->IsMarking());
AllocationSpace identity = chunk->owner()->identity();
// Generation consistency.
CHECK_EQ(identity == NEW_SPACE || identity == NEW_LO_SPACE,
slim_chunk->InYoungGeneration());
// Read-only consistency.
CHECK_EQ(chunk->InReadOnlySpace(), slim_chunk->InReadOnlySpace());
// Marking consistency.
if (chunk->IsWritable()) {
// RO_SPACE can be shared between heaps, so we can't use RO_SPACE objects to
// find a heap. The exception is when the ReadOnlySpace is writeable, during
// bootstrapping, so explicitly allow this case.
Heap* heap = Heap::FromWritableHeapObject(object);
CHECK_EQ(slim_chunk->IsMarking(), heap->incremental_marking()->IsMarking());
} else {
// Non-writable RO_SPACE must never have marking flag set.
CHECK(!slim_chunk->IsMarking());
}
return true;
}
#ifdef DEBUG
void Heap::IncrementObjectCounters() {
isolate_->counters()->objs_since_last_full()->Increment();
isolate_->counters()->objs_since_last_young()->Increment();
}
#endif // DEBUG
bool Heap::IsStressingScavenge() {
return v8_flags.stress_scavenge > 0 && new_space();
}
void Heap::SetIsMarkingFlag(bool value) {
isolate()->isolate_data()->is_marking_flag_ = value;
}
uint8_t* Heap::IsMarkingFlagAddress() {
return &isolate()->isolate_data()->is_marking_flag_;
}
void Heap::SetIsMinorMarkingFlag(bool value) {
isolate()->isolate_data()->is_minor_marking_flag_ = value;
}
uint8_t* Heap::IsMinorMarkingFlagAddress() {
return &isolate()->isolate_data()->is_minor_marking_flag_;
}
// StrongRootBlocks are allocated as a block of addresses, prefixed with a
// StrongRootsEntry pointer:
//
// | StrongRootsEntry*
// | Address 1
// | ...
// | Address N
//
// The allocate method registers the range "Address 1" to "Address N" with the
// heap as a strong root array, saves that entry in StrongRootsEntry*, and
// returns a pointer to Address 1.
Address* StrongRootBlockAllocator::allocate(size_t n) {
void* block = base::Malloc(sizeof(StrongRootsEntry*) + n * sizeof(Address));
StrongRootsEntry** header = reinterpret_cast<StrongRootsEntry**>(block);
Address* ret = reinterpret_cast<Address*>(reinterpret_cast<char*>(block) +
sizeof(StrongRootsEntry*));
memset(ret, kNullAddress, n * sizeof(Address));
*header = heap_->RegisterStrongRoots(
"StrongRootBlockAllocator", FullObjectSlot(ret), FullObjectSlot(ret + n));
return ret;
}
void StrongRootBlockAllocator::deallocate(Address* p, size_t n) noexcept {
// The allocate method returns a pointer to Address 1, so the deallocate
// method has to offset that pointer back by sizeof(StrongRootsEntry*).
void* block = reinterpret_cast<char*>(p) - sizeof(StrongRootsEntry*);
StrongRootsEntry** header = reinterpret_cast<StrongRootsEntry**>(block);
heap_->UnregisterStrongRoots(*header);
base::Free(block);
}
#ifdef V8_ENABLE_ALLOCATION_TIMEOUT
void Heap::set_allocation_timeout(int allocation_timeout) {
heap_allocator_.SetAllocationTimeout(allocation_timeout);
}
#endif // V8_ENABLE_ALLOCATION_TIMEOUT
void Heap::FinishSweepingIfOutOfWork() {
if (sweeper()->major_sweeping_in_progress() && v8_flags.concurrent_sweeping &&
!sweeper()->AreMajorSweeperTasksRunning()) {
// At this point we know that all concurrent sweeping tasks have run
// out of work and quit: all pages are swept. The main thread still needs
// to complete sweeping though.
EnsureSweepingCompleted(SweepingForcedFinalizationMode::kV8Only);
}
if (cpp_heap()) {
// Ensure that sweeping is also completed for the C++ managed heap, if one
// exists and it's out of work.
CppHeap::From(cpp_heap())->FinishSweepingIfOutOfWork();
}
}
void Heap::EnsureSweepingCompleted(SweepingForcedFinalizationMode mode) {
CompleteArrayBufferSweeping(this);
if (sweeper()->sweeping_in_progress()) {
bool was_minor_sweeping_in_progress = minor_sweeping_in_progress();
bool was_major_sweeping_in_progress = major_sweeping_in_progress();
sweeper()->EnsureMajorCompleted();
if (was_major_sweeping_in_progress) {
TRACE_GC_EPOCH_WITH_FLOW(tracer(), GCTracer::Scope::MC_COMPLETE_SWEEPING,
ThreadKind::kMain,
sweeper_->GetTraceIdForFlowEvent(
GCTracer::Scope::MC_COMPLETE_SWEEPING),
TRACE_EVENT_FLAG_FLOW_IN);
old_space()->RefillFreeList();
{
CodePageHeaderModificationScope rwx_write_scope(
"Updating per-page stats stored in page headers requires write "
"access to Code page headers");
code_space()->RefillFreeList();
}
if (shared_space()) {
shared_space()->RefillFreeList();
}
trusted_space()->RefillFreeList();
}
if (v8_flags.minor_ms && new_space() && was_minor_sweeping_in_progress) {
TRACE_GC_EPOCH_WITH_FLOW(
tracer(), GCTracer::Scope::MINOR_MS_COMPLETE_SWEEPING,
ThreadKind::kMain,
sweeper_->GetTraceIdForFlowEvent(
GCTracer::Scope::MINOR_MS_COMPLETE_SWEEPING),
TRACE_EVENT_FLAG_FLOW_IN | TRACE_EVENT_FLAG_FLOW_OUT);
paged_new_space()->paged_space()->RefillFreeList();
// Refill OLD_SPACE's freelist again for swept promoted pages.
old_space()->RefillFreeList();
}
tracer()->NotifyFullSweepingCompleted();
#ifdef VERIFY_HEAP
if (v8_flags.verify_heap) {
EvacuationVerifier verifier(this);
verifier.Run();
}
#endif
}
if (mode == SweepingForcedFinalizationMode::kUnifiedHeap && cpp_heap()) {
// Ensure that sweeping is also completed for the C++ managed heap, if one
// exists.
CppHeap::From(cpp_heap())->FinishSweepingIfRunning();
DCHECK(!CppHeap::From(cpp_heap())->sweeper().IsSweepingInProgress());
}
DCHECK_IMPLIES(
mode == SweepingForcedFinalizationMode::kUnifiedHeap || !cpp_heap(),
!tracer()->IsSweepingInProgress());
}
void Heap::EnsureYoungSweepingCompleted() {
if (!sweeper()->minor_sweeping_in_progress()) return;
TRACE_GC_EPOCH_WITH_FLOW(
tracer(), GCTracer::Scope::MINOR_MS_COMPLETE_SWEEPING, ThreadKind::kMain,
sweeper_->GetTraceIdForFlowEvent(
GCTracer::Scope::MINOR_MS_COMPLETE_SWEEPING),
TRACE_EVENT_FLAG_FLOW_IN);
sweeper()->EnsureMinorCompleted();
paged_new_space()->paged_space()->RefillFreeList();
old_space()->RefillFreeList();
tracer()->NotifyYoungSweepingCompleted();
}
void Heap::DrainSweepingWorklistForSpace(AllocationSpace space) {
if (!sweeper()->sweeping_in_progress_for_space(space)) return;
sweeper()->DrainSweepingWorklistForSpace(space);
}
EmbedderStackStateScope::EmbedderStackStateScope(Heap* heap, Origin origin,
StackState stack_state)
: heap_(heap), old_stack_state_(heap_->embedder_stack_state_) {
if (origin == kImplicitThroughTask && heap->overriden_stack_state()) {
stack_state = *heap->overriden_stack_state();
}
heap_->embedder_stack_state_ = stack_state;
}
// static
EmbedderStackStateScope EmbedderStackStateScope::ExplicitScopeForTesting(
Heap* heap, StackState stack_state) {
return EmbedderStackStateScope(heap, Origin::kExplicitInvocation,
stack_state);
}
EmbedderStackStateScope::~EmbedderStackStateScope() {
heap_->embedder_stack_state_ = old_stack_state_;
}
CppClassNamesAsHeapObjectNameScope::CppClassNamesAsHeapObjectNameScope(
v8::CppHeap* heap)
: scope_(std::make_unique<cppgc::internal::ClassNameAsHeapObjectNameScope>(
*CppHeap::From(heap))) {}
CppClassNamesAsHeapObjectNameScope::~CppClassNamesAsHeapObjectNameScope() =
default;
#if V8_HEAP_USE_PTHREAD_JIT_WRITE_PROTECT || V8_HEAP_USE_PKU_JIT_WRITE_PROTECT
CodePageMemoryModificationScopeForDebugging::
CodePageMemoryModificationScopeForDebugging(Heap* heap,
VirtualMemory* reservation,
base::AddressRegion region)
: rwx_write_scope_("Write access for zapping.") {
#if !defined(DEBUG) && !defined(VERIFY_HEAP)
UNREACHABLE();
#endif
}
CodePageMemoryModificationScopeForDebugging::
CodePageMemoryModificationScopeForDebugging(BasicMemoryChunk* chunk)
: rwx_write_scope_("Write access for zapping.") {
#if !defined(DEBUG) && !defined(VERIFY_HEAP)
UNREACHABLE();
#endif
}
CodePageMemoryModificationScopeForDebugging::
~CodePageMemoryModificationScopeForDebugging() {}
#else // V8_HEAP_USE_PTHREAD_JIT_WRITE_PROTECT ||
// V8_HEAP_USE_PKU_JIT_WRITE_PROTECT
CodePageMemoryModificationScopeForDebugging::
CodePageMemoryModificationScopeForDebugging(Heap* heap,
VirtualMemory* reservation,
base::AddressRegion region) {
#if !defined(DEBUG) && !defined(VERIFY_HEAP)
UNREACHABLE();
#else
if (heap->write_protect_code_memory()) {
reservation_ = reservation;
region_.emplace(region);
CHECK(reservation_->SetPermissions(
region_->begin(), region_->size(),
MemoryChunk::GetCodeModificationPermission()));
}
#endif
}
CodePageMemoryModificationScopeForDebugging::
CodePageMemoryModificationScopeForDebugging(BasicMemoryChunk* chunk)
: memory_modification_scope_(chunk) {
#if !defined(DEBUG) && !defined(VERIFY_HEAP)
UNREACHABLE();
#endif
}
CodePageMemoryModificationScopeForDebugging::
~CodePageMemoryModificationScopeForDebugging() {
if (reservation_) {
DCHECK(region_.has_value());
CHECK(reservation_->SetPermissions(
region_->begin(), region_->size(),
v8_flags.jitless ? PageAllocator::Permission::kRead
: PageAllocator::Permission::kReadExecute));
}
}
#endif
} // namespace internal
} // namespace v8
Zerion Mini Shell 1.0