/*
* Copyright 2010 Google Inc. All Rights Reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.googlecode.concurrentlinkedhashmap;
import static com.googlecode.concurrentlinkedhashmap.ConcurrentLinkedHashMap.DrainStatus.IDLE;
import static com.googlecode.concurrentlinkedhashmap.ConcurrentLinkedHashMap.DrainStatus.PROCESSING;
import static com.googlecode.concurrentlinkedhashmap.ConcurrentLinkedHashMap.DrainStatus.REQUIRED;
import static java.util.Collections.emptyList;
import static java.util.Collections.unmodifiableMap;
import static java.util.Collections.unmodifiableSet;
import java.io.InvalidObjectException;
import java.io.ObjectInputStream;
import java.io.Serializable;
import java.lang.ref.WeakReference;
import java.util.AbstractCollection;
import java.util.AbstractMap;
import java.util.AbstractQueue;
import java.util.AbstractSet;
import java.util.Collection;
import java.util.Collections;
import java.util.HashMap;
import java.util.Iterator;
import java.util.LinkedHashMap;
import java.util.LinkedHashSet;
import java.util.List;
import java.util.Map;
import java.util.Queue;
import java.util.Set;
import java.util.concurrent.AbstractExecutorService;
import java.util.concurrent.CancellationException;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentLinkedQueue;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.RejectedExecutionException;
import java.util.concurrent.ScheduledExecutorService;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicIntegerArray;
import java.util.concurrent.atomic.AtomicReference;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReentrantLock;
/**
* A hash table supporting full concurrency of retrievals, adjustable expected
* concurrency for updates, and a maximum capacity to bound the map by. This
* implementation differs from {@link ConcurrentHashMap} in that it maintains a
* page replacement algorithm that is used to evict an entry when the map has
* exceeded its capacity. Unlike the <tt>Java Collections Framework</tt>, this
* map does not have a publicly visible constructor and instances are created
* through a {@link Builder}.
* <p>
* An entry is evicted from the map when the <tt>weighted capacity</tt> exceeds
* its <tt>maximum weighted capacity</tt> threshold. A {@link Weigher} instance
* determines how many units of capacity that a value consumes. The default
* weigher assigns each value a weight of <tt>1</tt> to bound the map by the
* total number of key-value pairs. A map that holds collections may choose to
* weigh values by the number of elements in the collection and bound the map
* by the total number of elements that it contains. A change to a value that
* modifies its weight requires that an update operation is performed on the
* map.
* <p>
* An {@link EvictionListener} may be supplied for notification when an entry
* is evicted from the map. This listener is invoked on a caller's thread and
* will not block other threads from operating on the map. An implementation
* should be aware that the caller's thread will not expect long execution
* times or failures as a side effect of the listener being notified. Execution
* safety and a fast turn around time can be achieved by performing the
* operation asynchronously, such as by submitting a task to an
* {@link java.util.concurrent.ExecutorService}.
* <p>
* The <tt>concurrency level</tt> determines the number of threads that can
* concurrently modify the table. Using a significantly higher or lower value
* than needed can waste space or lead to thread contention, but an estimate
* within an order of magnitude of the ideal value does not usually have a
* noticeable impact. Because placement in hash tables is essentially random,
* the actual concurrency will vary.
* <p>
* This class and its views and iterators implement all of the
* <em>optional</em> methods of the {@link Map} and {@link Iterator}
* interfaces.
* <p>
* Like {@link java.util.Hashtable} but unlike {@link HashMap}, this class
* does <em>not</em> allow <tt>null</tt> to be used as a key or value. Unlike
* {@link java.util.LinkedHashMap}, this class does <em>not</em> provide
* predictable iteration order. A snapshot of the keys and entries may be
* obtained in ascending and descending order of retention.
*
* @author ben.manes@gmail.com (Ben Manes)
* @param <K> the type of keys maintained by this map
* @param <V> the type of mapped values
* @see <a href="http://code.google.com/p/concurrentlinkedhashmap/">
* http://code.google.com/p/concurrentlinkedhashmap/</a>
*/
@ThreadSafe
public final class ConcurrentLinkedHashMap<K, V> extends AbstractMap<K, V>
implements ConcurrentMap<K, V>, Serializable {
/*
* This class performs a best-effort bounding of a ConcurrentHashMap using a
* page-replacement algorithm to determine which entries to evict when the
* capacity is exceeded.
*
* The page replacement algorithm's data structures are kept eventually
* consistent with the map. An update to the map and recording of reads may
* not be immediately reflected on the algorithm's data structures. These
* structures are guarded by a lock and operations are applied in batches to
* avoid lock contention. The penalty of applying the batches is spread across
* threads so that the amortized cost is slightly higher than performing just
* the ConcurrentHashMap operation.
*
* A memento of the reads and writes that were performed on the map are
* recorded in a buffer. These buffers are drained at the first opportunity
* after a write or when a buffer exceeds a threshold size. A mostly strict
* ordering is achieved by observing that each buffer is in a weakly sorted
* order relative to the last drain. This allows the buffers to be merged in
* O(n) time so that the operations are run in the expected order.
*
* Due to a lack of a strict ordering guarantee, a task can be executed
* out-of-order, such as a removal followed by its addition. The state of the
* entry is encoded within the value's weight.
*
* Alive: The entry is in both the hash-table and the page replacement policy.
* This is represented by a positive weight.
*
* Retired: The entry is not in the hash-table and is pending removal from the
* page replacement policy. This is represented by a negative weight.
*
* Dead: The entry is not in the hash-table and is not in the page replacement
* policy. This is represented by a weight of zero.
*
* The Least Recently Used page replacement algorithm was chosen due to its
* simplicity, high hit rate, and ability to be implemented with O(1) time
* complexity.
*/
/** The maximum weighted capacity of the map. */
static final int MAXIMUM_CAPACITY = 1 << 30;
/** The maximum weight of a value. */
static final int MAXIMUM_WEIGHT = 1 << 29;
/** The maximum number of pending operations per buffer. */
static final int MAXIMUM_BUFFER_SIZE = 1 << 20;
/** The number of pending operations per buffer before attempting to drain. */
static final int BUFFER_THRESHOLD = 16;
/** The number of buffers to use. */
static final int NUMBER_OF_BUFFERS;
/** Mask value for indexing into the buffers. */
static final int BUFFER_MASK;
/** The maximum number of operations to perform per amortized drain. */
static final int AMORTIZED_DRAIN_THRESHOLD;
/** A queue that discards all entries. */
static final Queue<?> DISCARDING_QUEUE = new DiscardingQueue();
static {
int buffers = ceilingNextPowerOfTwo(Runtime.getRuntime().availableProcessors());
AMORTIZED_DRAIN_THRESHOLD = (1 + buffers) * BUFFER_THRESHOLD;
NUMBER_OF_BUFFERS = buffers;
BUFFER_MASK = buffers - 1;
}
static final int ceilingNextPowerOfTwo(int x) {
// From Hacker's Delight, Chapter 3, Harry S. Warren Jr.
return 1 << (Integer.SIZE - Integer.numberOfLeadingZeros(x - 1));
}
/** The draining status of the buffers. */
enum DrainStatus {
/** A drain is not taking place. */
IDLE,
/** A drain is required due to a pending write modification. */
REQUIRED,
/** A drain is in progress. */
PROCESSING
}
// The backing data store holding the key-value associations
final ConcurrentMap<K, Node> data;
final int concurrencyLevel;
// These fields provide support to bound the map by a maximum capacity
@GuardedBy("evictionLock")
final LinkedDeque<Node> evictionDeque;
@GuardedBy("evictionLock") // must write under lock
volatile int weightedSize;
@GuardedBy("evictionLock") // must write under lock
volatile int capacity;
volatile int nextOrder;
@GuardedBy("evictionLock")
int drainedOrder;
final Lock evictionLock;
final Queue<Task>[] buffers;
final ExecutorService executor;
final Weigher<? super V> weigher;
final AtomicIntegerArray bufferLengths;
final AtomicReference<DrainStatus> drainStatus;
// These fields provide support for notifying a listener.
final Queue<Node> pendingNotifications;
final EvictionListener<K, V> listener;
transient Set<K> keySet;
transient Collection<V> values;
transient Set<Entry<K, V>> entrySet;
/**
* Creates an instance based on the builder's configuration.
*/
@SuppressWarnings({"unchecked", "cast"})
private ConcurrentLinkedHashMap(Builder<K, V> builder) {
// The data store and its maximum capacity
concurrencyLevel = builder.concurrencyLevel;
capacity = Math.min(builder.capacity, MAXIMUM_CAPACITY);
data = new ConcurrentHashMap<K, Node>(builder.initialCapacity, 0.75f, concurrencyLevel);
// The eviction support
weigher = builder.weigher;
executor = builder.executor;
nextOrder = Integer.MIN_VALUE;
drainedOrder = Integer.MIN_VALUE;
evictionLock = new ReentrantLock();
evictionDeque = new LinkedDeque<Node>();
drainStatus = new AtomicReference<DrainStatus>(IDLE);
buffers = (Queue<Task>[]) new Queue[NUMBER_OF_BUFFERS];
bufferLengths = new AtomicIntegerArray(NUMBER_OF_BUFFERS);
for (int i = 0; i < NUMBER_OF_BUFFERS; i++) {
buffers[i] = new ConcurrentLinkedQueue<Task>();
}
// The notification queue and listener
listener = builder.listener;
pendingNotifications = (listener == DiscardingListener.INSTANCE)
? (Queue<Node>) DISCARDING_QUEUE
: new ConcurrentLinkedQueue<Node>();
}
/** Asserts that the object is not null. */
static void checkNotNull(Object o) {
if (o == null) {
throw new NullPointerException();
}
}
/* ---------------- Eviction Support -------------- */
/**
* Retrieves the maximum weighted capacity of the map.
*
* @return the maximum weighted capacity
*/
public int capacity() {
return capacity;
}
/**
* Sets the maximum weighted capacity of the map and eagerly evicts entries
* until it shrinks to the appropriate size.
*
* @param capacity the maximum weighted capacity of the map
* @throws IllegalArgumentException if the capacity is negative
*/
public void setCapacity(int capacity) {
if (capacity < 0) {
throw new IllegalArgumentException();
}
evictionLock.lock();
try {
this.capacity = Math.min(capacity, MAXIMUM_CAPACITY);
drainBuffers(AMORTIZED_DRAIN_THRESHOLD);
evict();
} finally {
evictionLock.unlock();
}
notifyListener();
}
/** Determines whether the map has exceeded its capacity. */
boolean hasOverflowed() {
return weightedSize > capacity;
}
/**
* Evicts entries from the map while it exceeds the capacity and appends
* evicted entries to the notification queue for processing.
*/
@GuardedBy("evictionLock")
void evict() {
// Attempts to evict entries from the map if it exceeds the maximum
// capacity. If the eviction fails due to a concurrent removal of the
// victim, that removal may cancel out the addition that triggered this
// eviction. The victim is eagerly unlinked before the removal task so
// that if an eviction is still required then a new victim will be chosen
// for removal.
while (hasOverflowed()) {
Node node = evictionDeque.poll();
// If weighted values are used, then the pending operations will adjust
// the size to reflect the correct weight
if (node == null) {
return;
}
// Notify the listener only if the entry was evicted
if (data.remove(node.key, node)) {
pendingNotifications.add(node);
}
node.makeDead();
}
}
/**
* Performs the post-processing work required after the map operation.
*
* @param task the pending operation to be applied
*/
void afterCompletion(Task task) {
boolean delayable = schedule(task);
if (shouldDrainBuffers(delayable)) {
tryToDrainBuffers(AMORTIZED_DRAIN_THRESHOLD);
}
notifyListener();
}
/**
* Schedules the task to be applied to the page replacement policy.
*
* @param task the pending operation
* @return if the draining of the buffers can be delayed
*/
private boolean schedule(Task task) {
int index = bufferIndex();
int buffered = bufferLengths.incrementAndGet(index);
if (task.isWrite()) {
buffers[index].add(task);
drainStatus.set(REQUIRED);
return false;
}
// A buffer may discard a read task if its length exceeds a tolerance level
if (buffered <= MAXIMUM_BUFFER_SIZE) {
buffers[index].add(task);
return (buffered <= BUFFER_THRESHOLD);
} else { // not optimized for fail-safe scenario
bufferLengths.decrementAndGet(index);
return false;
}
}
/** Returns the index to the buffer that the task should be scheduled on. */
static int bufferIndex() {
// A buffer is chosen by the thread's id so that tasks are distributed in a
// pseudo evenly manner. This helps avoid hot entries causing contention due
// to other threads trying to append to the same buffer.
return (int) Thread.currentThread().getId() & BUFFER_MASK;
}
/** Returns the ordering value to assign to a task. */
int nextOrdering() {
// The next ordering is acquired in a racy fashion as the increment is not
// atomic with the insertion into a buffer. This means that concurrent tasks
// can have the same ordering and the buffers are in a weakly sorted order.
return nextOrder++;
}
/**
* Determines whether the buffers should be drained.
*
* @param delayable if a drain should be delayed until required
* @return if a drain should be attempted
*/
boolean shouldDrainBuffers(boolean delayable) {
if (executor.isShutdown()) {
DrainStatus status = drainStatus.get();
return (status != PROCESSING) & (!delayable | (status == REQUIRED));
}
return false;
}
/**
* Attempts to acquire the eviction lock and apply the pending operations to
* the page replacement policy.
*
* @param maxToDrain the maximum number of operations to drain
*/
void tryToDrainBuffers(int maxToDrain) {
if (evictionLock.tryLock()) {
try {
drainStatus.set(PROCESSING);
drainBuffers(maxToDrain);
} finally {
drainStatus.compareAndSet(PROCESSING, IDLE);
evictionLock.unlock();
}
}
}
/**
* Drains the buffers and applies the pending operations.
*
* @param maxToDrain the maximum number of operations to drain
*/
@GuardedBy("evictionLock")
void drainBuffers(int maxToDrain) {
// A mostly strict ordering is achieved by observing that each buffer
// contains tasks in a weakly sorted order starting from the last drain.
// The buffers can be merged into a sorted list in O(n) time by using
// counting sort and chaining on a collision.
// The output is capped to the expected number of tasks plus additional
// slack to optimistically handle the concurrent additions to the buffers.
Task[] tasks = new Task[maxToDrain];
// Moves the tasks into the output array, applies them, and updates the
// marker for the starting order of the next drain.
int maxTaskIndex = moveTasksFromBuffers(tasks);
runTasks(tasks, maxTaskIndex);
updateDrainedOrder(tasks, maxTaskIndex);
}
/**
* Moves the tasks from the buffers into the output array.
*
* @param tasks the ordered array of the pending operations
* @return the highest index location of a task that was added to the array
*/
@GuardedBy("evictionLock")
int moveTasksFromBuffers(Task[] tasks) {
int maxTaskIndex = -1;
for (int i = 0; i < buffers.length; i++) {
int maxIndex = moveTasksFromBuffer(tasks, i);
maxTaskIndex = Math.max(maxIndex, maxTaskIndex);
}
return maxTaskIndex;
}
/**
* Moves the tasks from the specified buffer into the output array.
*
* @param tasks the ordered array of the pending operations
* @param bufferIndex the buffer to drain into the tasks array
* @return the highest index location of a task that was added to the array
*/
@GuardedBy("evictionLock")
int moveTasksFromBuffer(Task[] tasks, int bufferIndex) {
// While a buffer is being drained it may be concurrently appended to. The
// number of tasks removed are tracked so that the length can be decremented
// by the delta rather than set to zero.
Queue<Task> buffer = buffers[bufferIndex];
int removedFromBuffer = 0;
Task task;
int maxIndex = -1;
while ((task = buffer.poll()) != null) {
removedFromBuffer++;
// The index into the output array is determined by calculating the offset
// since the last drain
int index = task.getOrder() - drainedOrder;
if (index < 0) {
// The task was missed by the last drain and can be run immediately
task.run();
} else if (index >= tasks.length) {
// Due to concurrent additions, the order exceeds the capacity of the
// output array. It is added to the end as overflow and the remaining
// tasks in the buffer will be handled by the next drain.
maxIndex = tasks.length - 1;
addTaskToChain(tasks, task, maxIndex);
break;
} else {
maxIndex = Math.max(index, maxIndex);
addTaskToChain(tasks, task, index);
}
}
bufferLengths.addAndGet(bufferIndex, -removedFromBuffer);
return maxIndex;
}
/**
* Adds the task as the head of the chain at the index location.
*
* @param tasks the ordered array of the pending operations
* @param task the pending operation to add
* @param index the array location
*/
@GuardedBy("evictionLock")
void addTaskToChain(Task[] tasks, Task task, int index) {
task.setNext(tasks[index]);
tasks[index] = task;
}
/**
* Runs the pending page replacement policy operations.
*
* @param tasks the ordered array of the pending operations
* @param maxTaskIndex the maximum index of the array
*/
@GuardedBy("evictionLock")
void runTasks(Task[] tasks, int maxTaskIndex) {
for (int i = 0; i <= maxTaskIndex; i++) {
runTasksInChain(tasks[i]);
}
}
/**
* Runs the pending operations on the linked chain.
*
* @param task the first task in the chain of operations
*/
@GuardedBy("evictionLock")
void runTasksInChain(Task task) {
while (task != null) {
Task current = task;
task = task.getNext();
current.setNext(null);
current.run();
}
}
/**
* Updates the order to start the next drain from.
*
* @param tasks the ordered array of operations
* @param maxTaskIndex the maximum index of the array
*/
@GuardedBy("evictionLock")
void updateDrainedOrder(Task[] tasks, int maxTaskIndex) {
if (maxTaskIndex >= 0) {
Task task = tasks[maxTaskIndex];
drainedOrder = task.getOrder() + 1;
}
}
/** Notifies the listener of entries that were evicted. */
void notifyListener() {
Node node;
while ((node = pendingNotifications.poll()) != null) {
listener.onEviction(node.key, node.getValue());
}
}
/** Updates the node's location in the page replacement policy. */
class ReadTask extends AbstractTask {
final Node node;
ReadTask(Node node) {
this.node = node;
}
@Override
@GuardedBy("evictionLock")
public void run() {
// An entry may scheduled for reordering despite having been previously
// removed. This can occur when the entry was concurrently read while a
// writer was removing it. If the entry is no longer linked then it does
// not need to be processed.
if (evictionDeque.contains(node)) {
evictionDeque.moveToBack(node);
}
}
@Override
public boolean isWrite() {
return false;
}
}
/** Adds the node to the page replacement policy. */
final class AddTask extends AbstractTask {
final Node node;
final int weight;
AddTask(Node node, int weight) {
this.weight = weight;
this.node = node;
}
@Override
@GuardedBy("evictionLock")
public void run() {
weightedSize += weight;
// ignore out-of-order write operations
if (node.get().isAlive()) {
evictionDeque.add(node);
evict();
}
}
@Override
public boolean isWrite() {
return true;
}
}
/** Removes a node from the page replacement policy. */
final class RemovalTask extends AbstractTask {
final Node node;
RemovalTask(Node node) {
this.node = node;
}
@Override
@GuardedBy("evictionLock")
public void run() {
// add may not have been processed yet
evictionDeque.remove(node);
node.makeDead();
}
@Override
public boolean isWrite() {
return true;
}
}
/** Updates the weighted size and evicts an entry on overflow. */
final class UpdateTask extends ReadTask {
final int weightDifference;
public UpdateTask(Node node, int weightDifference) {
super(node);
this.weightDifference = weightDifference;
}
@Override
@GuardedBy("evictionLock")
public void run() {
super.run();
weightedSize += weightDifference;
evict();
}
@Override
public boolean isWrite() {
return true;
}
}
/* ---------------- Concurrent Map Support -------------- */
@Override
public boolean isEmpty() {
return data.isEmpty();
}
@Override
public int size() {
return data.size();
}
/**
* Returns the weighted size of this map.
*
* @return the combined weight of the values in this map
*/
public int weightedSize() {
return Math.max(0, weightedSize);
}
@Override
public void clear() {
// The alternative is to iterate through the keys and call #remove(), which
// adds unnecessary contention on the eviction lock and buffers.
evictionLock.lock();
try {
Node node;
while ((node = evictionDeque.poll()) != null) {
data.remove(node.key, node);
node.makeDead();
}
// Drain the buffers and run only the write tasks
for (int i = 0; i < buffers.length; i++) {
Queue<Task> buffer = buffers[i];
int removed = 0;
Task task;
while ((task = buffer.poll()) != null) {
if (task.isWrite()) {
task.run();
}
removed++;
}
bufferLengths.addAndGet(i, -removed);
}
} finally {
evictionLock.unlock();
}
}
@Override
public boolean containsKey(Object key) {
return data.containsKey(key);
}
@Override
public boolean containsValue(Object value) {
checkNotNull(value);
for (Node node : data.values()) {
if (node.getValue().equals(value)) {
return true;
}
}
return false;
}
@Override
public V get(Object key) {
final Node node = data.get(key);
if (node == null) {
return null;
}
afterCompletion(new ReadTask(node));
return node.getValue();
}
@Override
public V put(K key, V value) {
return put(key, value, false);
}
@Override
public V putIfAbsent(K key, V value) {
return put(key, value, true);
}
/**
* Adds a node to the list and the data store. If an existing node is found,
* then its value is updated if allowed.
*
* @param key key with which the specified value is to be associated
* @param value value to be associated with the specified key
* @param onlyIfAbsent a write is performed only if the key is not already
* associated with a value
* @return the prior value in the data store or null if no mapping was found
*/
V put(K key, V value, boolean onlyIfAbsent) {
checkNotNull(value);
final int weight = weigher.weightOf(value);
final WeightedValue<V> weightedValue = new WeightedValue<V>(value, weight);
final Node node = new Node(key, weightedValue);
for (;;) {
final Node prior = data.putIfAbsent(node.key, node);
if (prior == null) {
afterCompletion(new AddTask(node, weight));
return null;
} else if (onlyIfAbsent) {
afterCompletion(new ReadTask(prior));
return prior.getValue();
}
for (;;) {
final WeightedValue<V> oldWeightedValue = prior.get();
if (!oldWeightedValue.isAlive()) {
break;
}
if (prior.compareAndSet(oldWeightedValue, weightedValue)) {
final int weightedDifference = weight - oldWeightedValue.weight;
final Task task = (weightedDifference == 0)
? new ReadTask(prior)
: new UpdateTask(prior, weightedDifference);
afterCompletion(task);
return oldWeightedValue.value;
}
}
}
}
@Override
public V remove(Object key) {
final Node node = data.remove(key);
if (node == null) {
return null;
}
node.makeRetired();
afterCompletion(new RemovalTask(node));
return node.getValue();
}
@Override
public boolean remove(Object key, Object value) {
Node node = data.get(key);
if ((node == null) || (value == null)) {
return false;
}
WeightedValue<V> weightedValue = node.get();
for (;;) {
if (weightedValue.hasValue(value)) {
if (node.tryToRetire(weightedValue)) {
if (data.remove(key, node)) {
afterCompletion(new RemovalTask(node));
return true;
}
} else {
weightedValue = node.get();
if (weightedValue.isAlive()) {
// retry as an intermediate update may have replaced the value with
// an equal instance that has a different reference identity
continue;
}
}
}
return false;
}
}
@Override
public V replace(K key, V value) {
checkNotNull(value);
final int weight = weigher.weightOf(value);
final WeightedValue<V> weightedValue = new WeightedValue<V>(value, weight);
final Node node = data.get(key);
if (node == null) {
return null;
}
for (;;) {
WeightedValue<V> oldWeightedValue = node.get();
if (!oldWeightedValue.isAlive()) {
return null;
}
if (node.compareAndSet(oldWeightedValue, weightedValue)) {
int weightedDifference = weight - oldWeightedValue.weight;
final Task task = (weightedDifference == 0)
? new ReadTask(node)
: new UpdateTask(node, weightedDifference);
afterCompletion(task);
return oldWeightedValue.value;
}
}
}
@Override
public boolean replace(K key, V oldValue, V newValue) {
checkNotNull(oldValue);
checkNotNull(newValue);
final int weight = weigher.weightOf(newValue);
final WeightedValue<V> newWeightedValue = new WeightedValue<V>(newValue, weight);
final Node node = data.get(key);
if (node == null) {
return false;
}
for (;;) {
final WeightedValue<V> weightedValue = node.get();
if (!weightedValue.isAlive() || !weightedValue.hasValue(oldValue)) {
return false;
}
if (node.compareAndSet(weightedValue, newWeightedValue)) {
int weightedDifference = weight - weightedValue.weight;
final Task task = (weightedDifference == 0)
? new ReadTask(node)
: new UpdateTask(node, weightedDifference);
afterCompletion(task);
return true;
}
}
}
@Override
public Set<K> keySet() {
Set<K> ks = keySet;
return (ks == null) ? (keySet = new KeySet()) : ks;
}
/**
* Returns a unmodifiable snapshot {@link Set} view of the keys contained in
* this map. The set's iterator returns the keys whose order of iteration is
* the ascending order in which its entries are considered eligible for
* retention, from the least-likely to be retained to the most-likely.
* <p>
* Beware that, unlike in {@link #keySet()}, obtaining the set is <em>NOT</em>
* a constant-time operation. Because of the asynchronous nature of the page
* replacement policy, determining the retention ordering requires a traversal
* of the keys.
*
* @return an ascending snapshot view of the keys in this map
*/
public Set<K> ascendingKeySet() {
return orderedKeySet(true, Integer.MAX_VALUE);
}
/**
* Returns an unmodifiable snapshot {@link Set} view of the keys contained in
* this map. The set's iterator returns the keys whose order of iteration is
* the ascending order in which its entries are considered eligible for
* retention, from the least-likely to be retained to the most-likely.
* <p>
* Beware that, unlike in {@link #keySet()}, obtaining the set is <em>NOT</em>
* a constant-time operation. Because of the asynchronous nature of the page
* replacement policy, determining the retention ordering requires a traversal
* of the keys.
*
* @param limit the maximum size of the returned set
* @return a ascending snapshot view of the keys in this map
* @throws IllegalArgumentException if the limit is negative
*/
public Set<K> ascendingKeySetWithLimit(int limit) {
return orderedKeySet(true, limit);
}
/**
* Returns an unmodifiable snapshot {@link Set} view of the keys contained in
* this map. The set's iterator returns the keys whose order of iteration is
* the descending order in which its entries are considered eligible for
* retention, from the most-likely to be retained to the least-likely.
* <p>
* Beware that, unlike in {@link #keySet()}, obtaining the set is <em>NOT</em>
* a constant-time operation. Because of the asynchronous nature of the page
* replacement policy, determining the retention ordering requires a traversal
* of the keys.
*
* @return a descending snapshot view of the keys in this map
*/
public Set<K> descendingKeySet() {
return orderedKeySet(false, Integer.MAX_VALUE);
}
/**
* Returns an unmodifiable snapshot {@link Set} view of the keys contained in
* this map. The set's iterator returns the keys whose order of iteration is
* the descending order in which its entries are considered eligible for
* retention, from the most-likely to be retained to the least-likely.
* <p>
* Beware that, unlike in {@link #keySet()}, obtaining the set is <em>NOT</em>
* a constant-time operation. Because of the asynchronous nature of the page
* replacement policy, determining the retention ordering requires a traversal
* of the keys.
*
* @param limit the maximum size of the returned set
* @return a descending snapshot view of the keys in this map
* @throws IllegalArgumentException if the limit is negative
*/
public Set<K> descendingKeySetWithLimit(int limit) {
return orderedKeySet(false, limit);
}
Set<K> orderedKeySet(boolean ascending, int limit) {
if (limit < 0) {
throw new IllegalArgumentException();
}
evictionLock.lock();
try {
drainBuffers(AMORTIZED_DRAIN_THRESHOLD);
int initialCapacity = (weigher == Weighers.singleton())
? Math.min(limit, weightedSize())
: 16;
Set<K> keys = new LinkedHashSet<K>(initialCapacity);
Iterator<Node> iterator = ascending
? evictionDeque.iterator()
: evictionDeque.descendingIterator();
while (iterator.hasNext() && (limit > keys.size())) {
keys.add(iterator.next().key);
}
return unmodifiableSet(keys);
} finally {
evictionLock.unlock();
}
}
@Override
public Collection<V> values() {
Collection<V> vs = values;
return (vs == null) ? (values = new Values()) : vs;
}
@Override
public Set<Entry<K, V>> entrySet() {
Set<Entry<K, V>> es = entrySet;
return (es == null) ? (entrySet = new EntrySet()) : es;
}
/**
* Returns an unmodifiable snapshot {@link Map} view of the mappings contained
* in this map. The map's collections return the mappings whose order of
* iteration is the ascending order in which its entries are considered
* eligible for retention, from the least-likely to be retained to the
* most-likely.
* <p>
* Beware that obtaining the mappings is <em>NOT</em> a constant-time
* operation. Because of the asynchronous nature of the page replacement
* policy, determining the retention ordering requires a traversal of the
* entries.
*
* @return a ascending snapshot view of this map
*/
public Map<K, V> ascendingMap() {
return orderedMap(true, Integer.MAX_VALUE);
}
/**
* Returns an unmodifiable snapshot {@link Map} view of the mappings contained
* in this map. The map's collections return the mappings whose order of
* iteration is the ascending order in which its entries are considered
* eligible for retention, from the least-likely to be retained to the
* most-likely.
* <p>
* Beware that obtaining the mappings is <em>NOT</em> a constant-time
* operation. Because of the asynchronous nature of the page replacement
* policy, determining the retention ordering requires a traversal of the
* entries.
*
* @param limit the maximum size of the returned map
* @return a ascending snapshot view of this map
* @throws IllegalArgumentException if the limit is negative
*/
public Map<K, V> ascendingMapWithLimit(int limit) {
return orderedMap(true, limit);
}
/**
* Returns an unmodifiable snapshot {@link Map} view of the mappings contained
* in this map. The map's collections return the mappings whose order of
* iteration is the descending order in which its entries are considered
* eligible for retention, from the most-likely to be retained to the
* least-likely.
* <p>
* Beware that obtaining the mappings is <em>NOT</em> a constant-time
* operation. Because of the asynchronous nature of the page replacement
* policy, determining the retention ordering requires a traversal of the
* entries.
*
* @return a descending snapshot view of this map
*/
public Map<K, V> descendingMap() {
return orderedMap(false, Integer.MAX_VALUE);
}
/**
* Returns an unmodifiable snapshot {@link Map} view of the mappings contained
* in this map. The map's collections return the mappings whose order of
* iteration is the descending order in which its entries are considered
* eligible for retention, from the most-likely to be retained to the
* least-likely.
* <p>
* Beware that obtaining the mappings is <em>NOT</em> a constant-time
* operation. Because of the asynchronous nature of the page replacement
* policy, determining the retention ordering requires a traversal of the
* entries.
*
* @param limit the maximum size of the returned map
* @return a descending snapshot view of this map
* @throws IllegalArgumentException if the limit is negative
*/
public Map<K, V> descendingMapWithLimit(int limit) {
return orderedMap(false, limit);
}
Map<K, V> orderedMap(boolean ascending, int limit) {
if (limit < 0) {
throw new IllegalArgumentException();
}
evictionLock.lock();
try {
drainBuffers(AMORTIZED_DRAIN_THRESHOLD);
int initialCapacity = (weigher == Weighers.singleton())
? Math.min(limit, weightedSize())
: 16;
Map<K, V> map = new LinkedHashMap<K, V>(initialCapacity);
Iterator<Node> iterator = ascending
? evictionDeque.iterator()
: evictionDeque.descendingIterator();
while (iterator.hasNext() && (limit > map.size())) {
Node node = iterator.next();
map.put(node.key, node.getValue());
}
return unmodifiableMap(map);
} finally {
evictionLock.unlock();
}
}
/** A value, its weight, and the entry's status. */
@Immutable
static final class WeightedValue<V> {
final int weight;
final V value;
WeightedValue(V value, int weight) {
this.weight = weight;
this.value = value;
}
boolean hasValue(Object o) {
return (o == value) || value.equals(o);
}
/**
* If the entry is available in the hash-table and page replacement policy.
*/
boolean isAlive() {
return weight > 0;
}
/**
* If the entry was removed from the hash-table and is awaiting removal from
* the page replacement policy.
*/
boolean isRetired() {
return weight < 0;
}
/**
* If the entry was removed from the hash-table and the page replacement
* policy.
*/
boolean isDead() {
return weight == 0;
}
}
/**
* A node contains the key, the weighted value, and the linkage pointers on
* the page-replacement algorithm's data structures.
*/
@SuppressWarnings("serial")
final class Node extends AtomicReference<WeightedValue<V>> implements Linked<Node> {
final K key;
@GuardedBy("evictionLock")
Node prev;
@GuardedBy("evictionLock")
Node next;
/** Creates a new, unlinked node. */
Node(K key, WeightedValue<V> weightedValue) {
super(weightedValue);
this.key = key;
}
@Override
@GuardedBy("evictionLock")
public Node getPrevious() {
return prev;
}
@Override
@GuardedBy("evictionLock")
public void setPrevious(Node prev) {
this.prev = prev;
}
@Override
@GuardedBy("evictionLock")
public Node getNext() {
return next;
}
@Override
@GuardedBy("evictionLock")
public void setNext(Node next) {
this.next = next;
}
/** Retrieves the value held by the current <tt>WeightedValue</tt>. */
V getValue() {
return get().value;
}
/**
* Attempts to transition the node from the <tt>alive</tt> state to the
* <tt>retired</tt> state.
*
* @param expect the expected weighted value
* @return if successful
*/
boolean tryToRetire(WeightedValue<V> expect) {
if (expect.isAlive()) {
WeightedValue<V> retired = new WeightedValue<V>(expect.value, -expect.weight);
return compareAndSet(expect, retired);
}
return false;
}
/**
* Atomically transitions the node from the <tt>alive</tt> state to the
* <tt>retired</tt> state, if a valid transition.
*/
void makeRetired() {
for (;;) {
WeightedValue<V> current = get();
if (!current.isAlive()) {
return;
}
WeightedValue<V> retired = new WeightedValue<V>(current.value, -current.weight);
if (compareAndSet(current, retired)) {
return;
}
}
}
/**
* Atomically transitions the node to the <tt>dead</tt> state and decrements
* the <tt>weightedSize</tt>.
*/
@GuardedBy("evictionLock")
void makeDead() {
for (;;) {
WeightedValue<V> current = get();
WeightedValue<V> dead = new WeightedValue<V>(current.value, 0);
if (compareAndSet(current, dead)) {
weightedSize -= Math.abs(current.weight);
return;
}
}
}
}
/** An adapter to safely externalize the keys. */
final class KeySet extends AbstractSet<K> {
final ConcurrentLinkedHashMap<K, V> map = ConcurrentLinkedHashMap.this;
@Override
public int size() {
return map.size();
}
@Override
public void clear() {
map.clear();
}
@Override
public Iterator<K> iterator() {
return new KeyIterator();
}
@Override
public boolean contains(Object obj) {
return containsKey(obj);
}
@Override
public boolean remove(Object obj) {
return (map.remove(obj) != null);
}
@Override
public Object[] toArray() {
return map.data.keySet().toArray();
}
@Override
public <T> T[] toArray(T[] array) {
return map.data.keySet().toArray(array);
}
}
/** An adapter to safely externalize the key iterator. */
final class KeyIterator implements Iterator<K> {
final Iterator<K> iterator = data.keySet().iterator();
K current;
@Override
public boolean hasNext() {
return iterator.hasNext();
}
@Override
public K next() {
current = iterator.next();
return current;
}
@Override
public void remove() {
if (current == null) {
throw new IllegalStateException();
}
ConcurrentLinkedHashMap.this.remove(current);
current = null;
}
}
/** An adapter to safely externalize the values. */
final class Values extends AbstractCollection<V> {
@Override
public int size() {
return ConcurrentLinkedHashMap.this.size();
}
@Override
public void clear() {
ConcurrentLinkedHashMap.this.clear();
}
@Override
public Iterator<V> iterator() {
return new ValueIterator();
}
@Override
public boolean contains(Object o) {
return containsValue(o);
}
}
/** An adapter to safely externalize the value iterator. */
final class ValueIterator implements Iterator<V> {
final Iterator<Node> iterator = data.values().iterator();
Node current;
@Override
public boolean hasNext() {
return iterator.hasNext();
}
@Override
public V next() {
current = iterator.next();
return current.getValue();
}
@Override
public void remove() {
if (current == null) {
throw new IllegalStateException();
}
ConcurrentLinkedHashMap.this.remove(current.key);
current = null;
}
}
/** An adapter to safely externalize the entries. */
final class EntrySet extends AbstractSet<Entry<K, V>> {
final ConcurrentLinkedHashMap<K, V> map = ConcurrentLinkedHashMap.this;
@Override
public int size() {
return map.size();
}
@Override
public void clear() {
map.clear();
}
@Override
public Iterator<Entry<K, V>> iterator() {
return new EntryIterator();
}
@Override
public boolean contains(Object obj) {
if (!(obj instanceof Entry<?, ?>)) {
return false;
}
Entry<?, ?> entry = (Entry<?, ?>) obj;
Node node = map.data.get(entry.getKey());
return (node != null) && (node.getValue().equals(entry.getValue()));
}
@Override
public boolean add(Entry<K, V> entry) {
return (map.putIfAbsent(entry.getKey(), entry.getValue()) == null);
}
@Override
public boolean remove(Object obj) {
if (!(obj instanceof Entry<?, ?>)) {
return false;
}
Entry<?, ?> entry = (Entry<?, ?>) obj;
return map.remove(entry.getKey(), entry.getValue());
}
}
/** An adapter to safely externalize the entry iterator. */
final class EntryIterator implements Iterator<Entry<K, V>> {
final Iterator<Node> iterator = data.values().iterator();
Node current;
@Override
public boolean hasNext() {
return iterator.hasNext();
}
@Override
public Entry<K, V> next() {
current = iterator.next();
return new WriteThroughEntry(current);
}
@Override
public void remove() {
if (current == null) {
throw new IllegalStateException();
}
ConcurrentLinkedHashMap.this.remove(current.key);
current = null;
}
}
/** An entry that allows updates to write through to the map. */
final class WriteThroughEntry extends SimpleEntry<K, V> {
static final long serialVersionUID = 1;
WriteThroughEntry(Node node) {
super(node.key, node.getValue());
}
@Override
public V setValue(V value) {
put(getKey(), value);
return super.setValue(value);
}
Object writeReplace() {
return new SimpleEntry<K, V>(this);
}
}
/** A weigher that enforces that the weight falls within a valid range. */
static final class BoundedWeigher<V> implements Weigher<V>, Serializable {
static final long serialVersionUID = 1;
final Weigher<? super V> weigher;
BoundedWeigher(Weigher<? super V> weigher) {
checkNotNull(weigher);
this.weigher = weigher;
}
@Override
public int weightOf(V value) {
int weight = weigher.weightOf(value);
if ((weight < 1) || (weight > MAXIMUM_WEIGHT)) {
throw new IllegalArgumentException("invalid weight");
}
return weight;
}
Object writeReplace() {
return weigher;
}
}
/** A task that catches up the page replacement policy. */
static final class CatchUpTask implements Runnable {
final WeakReference<ConcurrentLinkedHashMap<?, ?>> mapRef;
CatchUpTask(ConcurrentLinkedHashMap<?, ?> map) {
this.mapRef = new WeakReference<ConcurrentLinkedHashMap<?, ?>>(map);
}
@Override
public void run() {
ConcurrentLinkedHashMap<?, ?> map = mapRef.get();
if (map == null) {
throw new CancellationException();
}
int pendingTasks = 0;
for (int i = 0; i < map.buffers.length; i++) {
pendingTasks += map.bufferLengths.get(i);
}
if (pendingTasks != 0) {
map.tryToDrainBuffers(pendingTasks + BUFFER_THRESHOLD);
}
}
}
/** An executor that is always terminated. */
static final class DisabledExecutorService extends AbstractExecutorService {
@Override public boolean isShutdown() { return true; }
@Override public boolean isTerminated() { return true; }
@Override public void shutdown() {}
@Override public List<Runnable> shutdownNow() { return Collections.emptyList(); }
@Override public boolean awaitTermination(long timeout, TimeUnit unit) { return true; }
@Override public void execute(Runnable command) { throw new RejectedExecutionException(); }
}
/** A queue that discards all additions and is always empty. */
static final class DiscardingQueue extends AbstractQueue<Object> {
@Override public boolean add(Object e) { return true; }
@Override public boolean offer(Object e) { return true; }
@Override public Object poll() { return null; }
@Override public Object peek() { return null; }
@Override public int size() { return 0; }
@Override public Iterator<Object> iterator() { return emptyList().iterator(); }
}
/** A listener that ignores all notifications. */
enum DiscardingListener implements EvictionListener<Object, Object> {
INSTANCE;
@Override public void onEviction(Object key, Object value) {}
}
/** An operation that can be lazily applied to the page replacement policy. */
interface Task extends Runnable {
/** The priority order. */
int getOrder();
/** If the task represents an add, modify, or remove operation. */
boolean isWrite();
/** Returns the next task on the link chain. */
Task getNext();
/** Sets the next task on the link chain. */
void setNext(Task task);
}
/** A skeletal implementation of the <tt>Task</tt> interface. */
abstract class AbstractTask implements Task {
final int order;
Task task;
AbstractTask() {
order = nextOrdering();
}
@Override
public int getOrder() {
return order;
}
@Override
public Task getNext() {
return task;
}
@Override
public void setNext(Task task) {
this.task = task;
}
}
/* ---------------- Serialization Support -------------- */
static final long serialVersionUID = 1;
Object writeReplace() {
return new SerializationProxy<K, V>(this);
}
private void readObject(ObjectInputStream stream) throws InvalidObjectException {
throw new InvalidObjectException("Proxy required");
}
/**
* A proxy that is serialized instead of the map. The page-replacement
* algorithm's data structures are not serialized so the deserialized
* instance contains only the entries. This is acceptable as caches hold
* transient data that is recomputable and serialization would tend to be
* used as a fast warm-up process.
*/
static final class SerializationProxy<K, V> implements Serializable {
final EvictionListener<K, V> listener;
final Weigher<? super V> weigher;
final int concurrencyLevel;
final Map<K, V> data;
final int capacity;
SerializationProxy(ConcurrentLinkedHashMap<K, V> map) {
concurrencyLevel = map.concurrencyLevel;
data = new HashMap<K, V>(map);
capacity = map.capacity;
listener = map.listener;
weigher = map.weigher;
}
Object readResolve() {
ConcurrentLinkedHashMap<K, V> map = new Builder<K, V>()
.concurrencyLevel(concurrencyLevel)
.maximumWeightedCapacity(capacity)
.listener(listener)
.weigher(weigher)
.build();
map.putAll(data);
return map;
}
static final long serialVersionUID = 1;
}
/* ---------------- Builder -------------- */
/**
* A builder that creates {@link ConcurrentLinkedHashMap} instances. It
* provides a flexible approach for constructing customized instances with
* a named parameter syntax. It can be used in the following manner:
* <pre>{@code
* ConcurrentMap<Vertex, Set<Edge>> graph = new Builder<Vertex, Set<Edge>>()
* .maximumWeightedCapacity(5000)
* .weigher(Weighers.<Edge>set())
* .build();
* }</pre>
*/
public static final class Builder<K, V> {
static final ExecutorService DEFAULT_EXECUTOR = new DisabledExecutorService();
static final int DEFAULT_CONCURRENCY_LEVEL = 16;
static final int DEFAULT_INITIAL_CAPACITY = 16;
EvictionListener<K, V> listener;
Weigher<? super V> weigher;
ExecutorService executor;
TimeUnit unit;
long delay;
int concurrencyLevel;
int initialCapacity;
int capacity;
@SuppressWarnings("unchecked")
public Builder() {
capacity = -1;
executor = DEFAULT_EXECUTOR;
weigher = Weighers.singleton();
initialCapacity = DEFAULT_INITIAL_CAPACITY;
concurrencyLevel = DEFAULT_CONCURRENCY_LEVEL;
listener = (EvictionListener<K, V>) DiscardingListener.INSTANCE;
}
/**
* Specifies the initial capacity of the hash table (default <tt>16</tt>).
* This is the number of key-value pairs that the hash table can hold
* before a resize operation is required.
*
* @param initialCapacity the initial capacity used to size the hash table
* to accommodate this many entries.
* @throws IllegalArgumentException if the initialCapacity is negative
*/
public Builder<K, V> initialCapacity(int initialCapacity) {
if (initialCapacity < 0) {
throw new IllegalArgumentException();
}
this.initialCapacity = initialCapacity;
return this;
}
/**
* Specifies the maximum weighted capacity to coerce the map to and may
* exceed it temporarily.
*
* @param capacity the weighted threshold to bound the map by
* @throws IllegalArgumentException if the maximumWeightedCapacity is
* negative
*/
public Builder<K, V> maximumWeightedCapacity(int capacity) {
if (capacity < 0) {
throw new IllegalArgumentException();
}
this.capacity = capacity;
return this;
}
/**
* Specifies the estimated number of concurrently updating threads. The
* implementation performs internal sizing to try to accommodate this many
* threads (default <tt>16</tt>).
*
* @param concurrencyLevel the estimated number of concurrently updating
* threads
* @throws IllegalArgumentException if the concurrencyLevel is less than or
* equal to zero
*/
public Builder<K, V> concurrencyLevel(int concurrencyLevel) {
if (concurrencyLevel <= 0) {
throw new IllegalArgumentException();
}
this.concurrencyLevel = concurrencyLevel;
return this;
}
/**
* Specifies an optional listener that is registered for notification when
* an entry is evicted.
*
* @param listener the object to forward evicted entries to
* @throws NullPointerException if the listener is null
*/
public Builder<K, V> listener(EvictionListener<K, V> listener) {
checkNotNull(listener);
this.listener = listener;
return this;
}
/**
* Specifies an algorithm to determine how many the units of capacity a
* value consumes. The default algorithm bounds the map by the number of
* key-value pairs by giving each entry a weight of <tt>1</tt>.
*
* @param weigher the algorithm to determine a value's weight
* @throws NullPointerException if the weigher is null
*/
public Builder<K, V> weigher(Weigher<? super V> weigher) {
this.weigher = (weigher == Weighers.singleton())
? Weighers.<V>singleton()
: new BoundedWeigher<V>(weigher);
return this;
}
/**
* Specifies an executor for use in catching up the page replacement policy.
* The catch-up phase processes both updates to the retention ordering and
* writes that may trigger an eviction. The delay should be chosen carefully
* as the map will not automatically evict between executions.
* <p>
* If unspecified or the executor is shutdown, the catching up will be
* amortized on user threads during write operations (or during read
* operations, in the absence of writes).
* <p>
* A single-threaded {@link ScheduledExecutorService} should be sufficient
* for catching up the page replacement policy in many maps.
*
* @param executor the executor to schedule on
* @param delay the delay between executions
* @param unit the time unit of the delay parameter
* @throws NullPointerException if the executor or time unit is null
* @throws IllegalArgumentException if the delay is less than or equal to
* zero
*/
public Builder<K, V> catchup(ScheduledExecutorService executor, long delay, TimeUnit unit) {
if (delay <= 0) {
throw new IllegalArgumentException();
}
checkNotNull(executor);
checkNotNull(unit);
this.executor = executor;
this.delay = delay;
this.unit = unit;
return this;
}
/**
* Creates a new {@link ConcurrentLinkedHashMap} instance.
*
* @throws IllegalStateException if the maximum weighted capacity was
* not set
* @throws RejectedExecutionException if an executor was specified and the
* catch-up task cannot be scheduled for execution
*/
public ConcurrentLinkedHashMap<K, V> build() {
if (capacity < 0) {
throw new IllegalStateException();
}
ConcurrentLinkedHashMap<K, V> map = new ConcurrentLinkedHashMap<K, V>(this);
if (executor != DEFAULT_EXECUTOR) {
ScheduledExecutorService es = (ScheduledExecutorService) executor;
es.scheduleWithFixedDelay(new CatchUpTask(map), delay, delay, unit);
}
return map;
}
}
}