blob: fbeb72f55c8328c80fefecd591f45230e40e157c [file] [log] [blame]
/*
* Copyright (C) 2016-2017 Apple Inc. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#pragma once
#include "ExceptionHelpers.h"
#include "JSObject.h"
namespace JSC {
JS_EXPORT_PRIVATE const ClassInfo* getHashMapBucketKeyClassInfo();
JS_EXPORT_PRIVATE const ClassInfo* getHashMapBucketKeyValueClassInfo();
JS_EXPORT_PRIVATE const ClassInfo* getHashMapImplKeyClassInfo();
JS_EXPORT_PRIVATE const ClassInfo* getHashMapImplKeyValueClassInfo();
enum class HashTableType {
Key,
KeyValue
};
struct HashMapBucketDataKey {
static const HashTableType Type = HashTableType::Key;
WriteBarrier<Unknown> key;
};
struct HashMapBucketDataKeyValue {
static const HashTableType Type = HashTableType::KeyValue;
WriteBarrier<Unknown> key;
WriteBarrier<Unknown> value;
};
template <typename Data>
class HashMapBucket : public JSCell {
typedef JSCell Base;
template <typename T = Data>
static typename std::enable_if<std::is_same<T, HashMapBucketDataKey>::value, Structure*>::type selectStructure(VM& vm)
{
return vm.hashMapBucketSetStructure.get();
}
template <typename T = Data>
static typename std::enable_if<std::is_same<T, HashMapBucketDataKeyValue>::value, Structure*>::type selectStructure(VM& vm)
{
return vm.hashMapBucketMapStructure.get();
}
public:
static const HashTableType Type = Data::Type;
static const ClassInfo s_info; // This is never accessed directly, since that would break linkage on some compilers.
static const ClassInfo* info()
{
switch (Type) {
case HashTableType::Key:
return getHashMapBucketKeyClassInfo();
case HashTableType::KeyValue:
return getHashMapBucketKeyValueClassInfo();
}
RELEASE_ASSERT_NOT_REACHED();
}
static Structure* createStructure(VM& vm, JSGlobalObject* globalObject, JSValue prototype)
{
return Structure::create(vm, globalObject, prototype, TypeInfo(CellType, StructureFlags), info());
}
static HashMapBucket* create(VM& vm)
{
HashMapBucket* bucket = new (NotNull, allocateCell<HashMapBucket<Data>>(vm.heap)) HashMapBucket(vm, selectStructure(vm));
bucket->finishCreation(vm);
ASSERT(!bucket->next());
ASSERT(!bucket->prev());
return bucket;
}
HashMapBucket(VM& vm, Structure* structure)
: Base(vm, structure)
{ }
ALWAYS_INLINE void setNext(VM& vm, HashMapBucket* bucket)
{
m_next.set(vm, this, bucket);
}
ALWAYS_INLINE void setPrev(VM& vm, HashMapBucket* bucket)
{
m_prev.set(vm, this, bucket);
}
ALWAYS_INLINE void setKey(VM& vm, JSValue key)
{
m_data.key.set(vm, this, key);
}
template <typename T = Data>
ALWAYS_INLINE typename std::enable_if<std::is_same<T, HashMapBucketDataKeyValue>::value>::type setValue(VM& vm, JSValue value)
{
m_data.value.set(vm, this, value);
}
template <typename T = Data>
ALWAYS_INLINE typename std::enable_if<std::is_same<T, HashMapBucketDataKey>::value>::type setValue(VM&, JSValue) { }
ALWAYS_INLINE JSValue key() const { return m_data.key.get(); }
template <typename T = Data>
ALWAYS_INLINE typename std::enable_if<std::is_same<T, HashMapBucketDataKeyValue>::value, JSValue>::type value() const
{
return m_data.value.get();
}
static void visitChildren(JSCell*, SlotVisitor&);
ALWAYS_INLINE HashMapBucket* next() const { return m_next.get(); }
ALWAYS_INLINE HashMapBucket* prev() const { return m_prev.get(); }
ALWAYS_INLINE bool deleted() const { return m_deleted; }
ALWAYS_INLINE void setDeleted(bool deleted) { m_deleted = deleted; }
static ptrdiff_t offsetOfKey()
{
return OBJECT_OFFSETOF(HashMapBucket, m_data) + OBJECT_OFFSETOF(Data, key);
}
template <typename T = Data>
static typename std::enable_if<std::is_same<T, HashMapBucketDataKeyValue>::value, ptrdiff_t>::type offsetOfValue()
{
return OBJECT_OFFSETOF(HashMapBucket, m_data) + OBJECT_OFFSETOF(Data, value);
}
private:
Data m_data;
WriteBarrier<HashMapBucket> m_next;
WriteBarrier<HashMapBucket> m_prev;
bool m_deleted { false };
};
template <typename BucketType>
class HashMapBuffer {
public:
HashMapBuffer() = delete;
static size_t allocationSize(uint32_t capacity)
{
return capacity * sizeof(BucketType*);
}
ALWAYS_INLINE BucketType** buffer() const
{
return bitwise_cast<BucketType**>(this);
}
static HashMapBuffer* create(ExecState* exec, VM& vm, JSCell*, uint32_t capacity)
{
auto scope = DECLARE_THROW_SCOPE(vm);
size_t allocationSize = HashMapBuffer::allocationSize(capacity);
void* data = vm.auxiliarySpace.tryAllocate(allocationSize);
if (!data) {
throwOutOfMemoryError(exec, scope);
return nullptr;
}
HashMapBuffer* buffer = static_cast<HashMapBuffer*>(data);
buffer->reset(capacity);
return buffer;
}
ALWAYS_INLINE void reset(uint32_t capacity)
{
memset(this, -1, allocationSize(capacity));
}
};
ALWAYS_INLINE static bool areKeysEqual(ExecState* exec, JSValue a, JSValue b)
{
// We want +0 and -0 to be compared to true here. sameValue() itself doesn't
// guarantee that, however, we normalize all keys before comparing and storing
// them in the map. The normalization will convert -0.0 and 0.0 to the integer
// representation for 0.
return sameValue(exec, a, b);
}
ALWAYS_INLINE JSValue normalizeMapKey(JSValue key)
{
if (!key.isNumber())
return key;
if (key.isInt32())
return key;
double d = key.asDouble();
if (std::isnan(d))
return key;
int i = static_cast<int>(d);
if (i == d) {
// When a key is -0, we convert it to positive zero.
// When a key is the double representation for an integer, we convert it to an integer.
return jsNumber(i);
}
// This means key is definitely not negative zero, and it's definitely not a double representation of an integer.
return key;
}
static ALWAYS_INLINE uint32_t wangsInt64Hash(uint64_t key)
{
key += ~(key << 32);
key ^= (key >> 22);
key += ~(key << 13);
key ^= (key >> 8);
key += (key << 3);
key ^= (key >> 15);
key += ~(key << 27);
key ^= (key >> 31);
return static_cast<unsigned>(key);
}
ALWAYS_INLINE uint32_t jsMapHash(ExecState* exec, VM& vm, JSValue value)
{
ASSERT_WITH_MESSAGE(normalizeMapKey(value) == value, "We expect normalized values flowing into this function.");
if (value.isString()) {
auto scope = DECLARE_THROW_SCOPE(vm);
const String& wtfString = asString(value)->value(exec);
RETURN_IF_EXCEPTION(scope, UINT_MAX);
return wtfString.impl()->hash();
}
return wangsInt64Hash(JSValue::encode(value));
}
ALWAYS_INLINE std::optional<uint32_t> concurrentJSMapHash(JSValue key)
{
key = normalizeMapKey(key);
if (key.isString()) {
JSString* string = asString(key);
if (string->length() > 10 * 1024)
return std::nullopt;
const StringImpl* impl = string->tryGetValueImpl();
if (!impl)
return std::nullopt;
return impl->concurrentHash();
}
uint64_t rawValue = JSValue::encode(key);
return wangsInt64Hash(rawValue);
}
template <typename HashMapBucketType>
class HashMapImpl : public JSCell {
typedef JSCell Base;
typedef HashMapBuffer<HashMapBucketType> HashMapBufferType;
template <typename T = HashMapBucketType>
static typename std::enable_if<std::is_same<T, HashMapBucket<HashMapBucketDataKey>>::value, Structure*>::type selectStructure(VM& vm)
{
return vm.hashMapImplSetStructure.get();
}
template <typename T = HashMapBucketType>
static typename std::enable_if<std::is_same<T, HashMapBucket<HashMapBucketDataKeyValue>>::value, Structure*>::type selectStructure(VM& vm)
{
return vm.hashMapImplMapStructure.get();
}
public:
static const ClassInfo s_info; // This is never accessed directly, since that would break linkage on some compilers.
static const ClassInfo* info()
{
switch (HashMapBucketType::Type) {
case HashTableType::Key:
return getHashMapImplKeyClassInfo();
case HashTableType::KeyValue:
return getHashMapImplKeyValueClassInfo();
}
RELEASE_ASSERT_NOT_REACHED();
}
static Structure* createStructure(VM& vm, JSGlobalObject* globalObject, JSValue prototype)
{
return Structure::create(vm, globalObject, prototype, TypeInfo(CellType, StructureFlags), info());
}
static HashMapImpl* create(ExecState* exec, VM& vm)
{
ASSERT_WITH_MESSAGE(HashMapBucket<HashMapBucketDataKey>::offsetOfKey() == HashMapBucket<HashMapBucketDataKeyValue>::offsetOfKey(), "We assume this to be true in the DFG and FTL JIT.");
HashMapImpl* impl = new (NotNull, allocateCell<HashMapImpl>(vm.heap)) HashMapImpl(vm, selectStructure(vm));
impl->finishCreation(exec, vm);
return impl;
}
static void visitChildren(JSCell*, SlotVisitor&);
HashMapImpl(VM& vm, Structure* structure)
: Base(vm, structure)
, m_keyCount(0)
, m_deleteCount(0)
, m_capacity(4)
{
}
ALWAYS_INLINE HashMapBucketType** buffer() const
{
return m_buffer.get()->buffer();
}
void finishCreation(ExecState* exec, VM& vm)
{
auto scope = DECLARE_THROW_SCOPE(vm);
Base::finishCreation(vm);
makeAndSetNewBuffer(exec, vm);
RETURN_IF_EXCEPTION(scope, void());
m_head.set(vm, this, HashMapBucketType::create(vm));
m_tail.set(vm, this, HashMapBucketType::create(vm));
m_head->setNext(vm, m_tail.get());
m_tail->setPrev(vm, m_head.get());
m_head->setDeleted(true);
m_tail->setDeleted(true);
}
static HashMapBucketType* emptyValue()
{
return bitwise_cast<HashMapBucketType*>(static_cast<uintptr_t>(-1));
}
static ALWAYS_INLINE bool isEmpty(HashMapBucketType* bucket)
{
return bucket == emptyValue();
}
static HashMapBucketType* deletedValue()
{
return bitwise_cast<HashMapBucketType*>(static_cast<uintptr_t>(-3));
}
static ALWAYS_INLINE bool isDeleted(HashMapBucketType* bucket)
{
return bucket == deletedValue();
}
ALWAYS_INLINE HashMapBucketType** findBucket(ExecState* exec, JSValue key)
{
VM& vm = exec->vm();
auto scope = DECLARE_THROW_SCOPE(vm);
key = normalizeMapKey(key);
uint32_t hash = jsMapHash(exec, vm, key);
RETURN_IF_EXCEPTION(scope, nullptr);
return findBucket(exec, key, hash);
}
ALWAYS_INLINE HashMapBucketType** findBucket(ExecState* exec, JSValue key, uint32_t hash)
{
ASSERT_WITH_MESSAGE(normalizeMapKey(key) == key, "We expect normalized values flowing into this function.");
return findBucketAlreadyHashedAndNormalized(exec, key, hash);
}
ALWAYS_INLINE JSValue get(ExecState* exec, JSValue key)
{
if (HashMapBucketType** bucket = findBucket(exec, key))
return (*bucket)->value();
return jsUndefined();
}
ALWAYS_INLINE bool has(ExecState* exec, JSValue key)
{
return !!findBucket(exec, key);
}
ALWAYS_INLINE void add(ExecState* exec, JSValue key, JSValue value = JSValue())
{
key = normalizeMapKey(key);
VM& vm = exec->vm();
auto scope = DECLARE_THROW_SCOPE(vm);
const uint32_t mask = m_capacity - 1;
uint32_t index = jsMapHash(exec, vm, key) & mask;
RETURN_IF_EXCEPTION(scope, void());
HashMapBucketType** buffer = this->buffer();
HashMapBucketType* bucket = buffer[index];
while (!isEmpty(bucket)) {
if (!isDeleted(bucket) && areKeysEqual(exec, key, bucket->key())) {
bucket->setValue(vm, value);
return;
}
index = (index + 1) & mask;
bucket = buffer[index];
}
HashMapBucketType* newEntry = m_tail.get();
buffer[index] = newEntry;
newEntry->setKey(vm, key);
newEntry->setValue(vm, value);
newEntry->setDeleted(false);
HashMapBucketType* newTail = HashMapBucketType::create(vm);
m_tail.set(vm, this, newTail);
newTail->setPrev(vm, newEntry);
newTail->setDeleted(true);
newEntry->setNext(vm, newTail);
++m_keyCount;
if (shouldRehashAfterAdd())
rehash(exec);
}
ALWAYS_INLINE bool remove(ExecState* exec, JSValue key)
{
HashMapBucketType** bucket = findBucket(exec, key);
if (!bucket)
return false;
VM& vm = exec->vm();
HashMapBucketType* impl = *bucket;
impl->next()->setPrev(vm, impl->prev());
impl->prev()->setNext(vm, impl->next());
impl->setDeleted(true);
*bucket = deletedValue();
++m_deleteCount;
ASSERT(m_keyCount > 0);
--m_keyCount;
if (shouldShrink())
rehash(exec);
return true;
}
ALWAYS_INLINE uint32_t size() const
{
return m_keyCount;
}
ALWAYS_INLINE void clear(ExecState* exec)
{
VM& vm = exec->vm();
m_keyCount = 0;
m_deleteCount = 0;
HashMapBucketType* head = m_head.get();
HashMapBucketType* bucket = m_head->next();
HashMapBucketType* tail = m_tail.get();
while (bucket != tail) {
HashMapBucketType* next = bucket->next();
// We restart each iterator by pointing it to the head of the list.
bucket->setNext(vm, head);
bucket->setDeleted(true);
bucket = next;
}
m_head->setNext(vm, m_tail.get());
m_tail->setPrev(vm, m_head.get());
m_capacity = 4;
makeAndSetNewBuffer(exec, vm);
checkConsistency();
}
ALWAYS_INLINE size_t bufferSizeInBytes() const
{
return m_capacity * sizeof(HashMapBucketType*);
}
static ptrdiff_t offsetOfBuffer()
{
return OBJECT_OFFSETOF(HashMapImpl<HashMapBucketType>, m_buffer);
}
static ptrdiff_t offsetOfCapacity()
{
return OBJECT_OFFSETOF(HashMapImpl<HashMapBucketType>, m_capacity);
}
HashMapBucketType* head() { return m_head.get(); }
HashMapBucketType* tail() { return m_tail.get(); }
size_t approximateSize() const
{
size_t size = sizeof(HashMapImpl);
size += bufferSizeInBytes();
size += 2 * sizeof(HashMapBucketType); // Head and tail members.
size += m_keyCount * sizeof(HashMapBucketType); // Number of members that are on the list.
return size;
}
private:
ALWAYS_INLINE uint32_t shouldRehashAfterAdd() const
{
return 2 * (m_keyCount + m_deleteCount) >= m_capacity;
}
ALWAYS_INLINE uint32_t shouldShrink() const
{
return 8 * m_keyCount <= m_capacity && m_capacity > 4;
}
ALWAYS_INLINE HashMapBucketType** findBucketAlreadyHashedAndNormalized(ExecState* exec, JSValue key, uint32_t hash)
{
const uint32_t mask = m_capacity - 1;
uint32_t index = hash & mask;
HashMapBucketType** buffer = this->buffer();
HashMapBucketType* bucket = buffer[index];
while (!isEmpty(bucket)) {
if (!isDeleted(bucket) && areKeysEqual(exec, key, bucket->key()))
return buffer + index;
index = (index + 1) & mask;
bucket = buffer[index];
}
return nullptr;
}
void rehash(ExecState* exec)
{
VM& vm = exec->vm();
auto scope = DECLARE_THROW_SCOPE(vm);
uint32_t oldCapacity = m_capacity;
if (shouldShrink()) {
m_capacity = m_capacity / 2;
ASSERT(m_capacity >= 4);
} else if (3 * m_keyCount <= m_capacity && m_capacity > 64) {
// We stay at the same size if rehashing would cause us to be no more than
// 1/3rd full. This comes up for programs like this:
// Say the hash table grew to a key count of 64, causing it to grow to a capacity of 256.
// Then, the table added 63 items. The load is now 127. Then, 63 items are deleted.
// The load is still 127. Then, another item is added. The load is now 128, and we
// decide that we need to rehash. The key count is 65, almost exactly what it was
// when we grew to a capacity of 256. We don't really need to grow to a capacity
// of 512 in this situation. Instead, we choose to rehash at the same size. This
// will bring the load down to 65. We rehash into the same size when we determine
// that the new load ratio will be under 1/3rd. (We also pick a minumum capacity
// at which this rule kicks in because otherwise we will be too sensitive to rehashing
// at the same capacity).
} else
m_capacity = (Checked<uint32_t>(m_capacity) * 2).unsafeGet();
if (m_capacity != oldCapacity) {
makeAndSetNewBuffer(exec, vm);
RETURN_IF_EXCEPTION(scope, void());
} else {
m_buffer.get()->reset(m_capacity);
assertBufferIsEmpty();
}
HashMapBucketType* iter = m_head->next();
HashMapBucketType* end = m_tail.get();
const uint32_t mask = m_capacity - 1;
RELEASE_ASSERT(!(m_capacity & (m_capacity - 1)));
HashMapBucketType** buffer = this->buffer();
while (iter != end) {
uint32_t index = jsMapHash(exec, vm, iter->key()) & mask;
ASSERT_WITH_MESSAGE(!scope.exception(), "All keys should already be hashed before, so this should not throw because it won't resolve ropes.");
{
HashMapBucketType* bucket = buffer[index];
while (!isEmpty(bucket)) {
index = (index + 1) & mask;
bucket = buffer[index];
}
}
buffer[index] = iter;
iter = iter->next();
}
m_deleteCount = 0;
checkConsistency();
}
ALWAYS_INLINE void checkConsistency() const
{
if (!ASSERT_DISABLED) {
HashMapBucketType* iter = m_head->next();
HashMapBucketType* end = m_tail.get();
uint32_t size = 0;
while (iter != end) {
++size;
iter = iter->next();
}
ASSERT(size == m_keyCount);
}
}
void makeAndSetNewBuffer(ExecState* exec, VM& vm)
{
ASSERT(!(m_capacity & (m_capacity - 1)));
HashMapBufferType* buffer = HashMapBufferType::create(exec, vm, this, m_capacity);
if (UNLIKELY(!buffer))
return;
m_buffer.set(vm, this, buffer);
assertBufferIsEmpty();
}
ALWAYS_INLINE void assertBufferIsEmpty() const
{
if (!ASSERT_DISABLED) {
for (unsigned i = 0; i < m_capacity; i++)
ASSERT(isEmpty(buffer()[i]));
}
}
WriteBarrier<HashMapBucketType> m_head;
WriteBarrier<HashMapBucketType> m_tail;
AuxiliaryBarrier<HashMapBufferType*> m_buffer;
uint32_t m_keyCount;
uint32_t m_deleteCount;
uint32_t m_capacity;
};
} // namespace JSC