All Downloads are FREE. Search and download functionalities are using the official Maven repository.

io.undertow.util.FastConcurrentDirectDeque Maven / Gradle / Ivy

Go to download

This artifact provides a single jar that contains all classes required to use remote EJB and JMS, including all dependencies. It is intended for use by those not using maven, maven users should just import the EJB and JMS BOM's instead (shaded JAR's cause lots of problems with maven, as it is very easy to inadvertently end up with different versions on classes on the class path).

There is a newer version: 34.0.0.Final
Show newest version
/*
 * JBoss, Home of Professional Open Source.
 * Copyright 2014 Red Hat, Inc., and individual contributors
 * as indicated by the @author tags.
 *
 * 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.
 */

/*
 * Written by Doug Lea and Martin Buchholz with assistance from members of
 * JCP JSR-166 Expert Group and released to the public domain, as explained
 * at http://creativecommons.org/publicdomain/zero/1.0/
 */

package io.undertow.util;

import static org.wildfly.common.Assert.checkNotNullParamWithNullPointerException;

import java.io.Serializable;
import java.lang.invoke.MethodHandles;
import java.lang.invoke.VarHandle;
import java.util.Arrays;
import java.util.Collection;
import java.util.Deque;
import java.util.Iterator;
import java.util.NoSuchElementException;
import java.util.Objects;
import java.util.Queue;
import java.util.Spliterator;
import java.util.Spliterators;
import java.util.function.Consumer;
import java.util.function.Predicate;

/**
 * A modified version of ConcurrentLinkedDeque which includes direct
 * removal. Like the original, it relies on Unsafe for better performance.
 *
 * More specifically, an unbounded concurrent {@linkplain Deque deque} based on linked nodes.
 * Concurrent insertion, removal, and access operations execute safely
 * across multiple threads.
 * A {@code ConcurrentLinkedDeque} is an appropriate choice when
 * many threads will share access to a common collection.
 * Like most other concurrent collection implementations, this class
 * does not permit the use of {@code null} elements.
 *
 * 

Iterators and spliterators are * weakly consistent. * *

Beware that, unlike in most collections, the {@code size} method * is NOT a constant-time operation. Because of the * asynchronous nature of these deques, determining the current number * of elements requires a traversal of the elements, and so may report * inaccurate results if this collection is modified during traversal. * *

Bulk operations that add, remove, or examine multiple elements, * such as {@link #addAll}, {@link #removeIf} or {@link #forEach}, * are not guaranteed to be performed atomically. * For example, a {@code forEach} traversal concurrent with an {@code * addAll} operation might observe only some of the added elements. * *

This class and its iterator implement all of the optional * methods of the {@link Deque} and {@link Iterator} interfaces. * *

Memory consistency effects: As with other concurrent collections, * actions in a thread prior to placing an object into a * {@code ConcurrentLinkedDeque} * happen-before * actions subsequent to the access or removal of that element from * the {@code ConcurrentLinkedDeque} in another thread. * *

This class is a member of the * * Java Collections Framework. * * Based on revision 1.88 of ConcurrentLinkedDeque * (see http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/main/java/util/concurrent/ConcurrentLinkedDeque.java?revision=1.88&view=markup) * This is the version used in JDK 9 b156. * * @since 1.7 * @author Doug Lea * @author Martin Buchholz * @author Jason T. Grene * @param the type of elements held in this deque */ public class FastConcurrentDirectDeque extends ConcurrentDirectDeque implements Deque, Serializable { /* * This is an implementation of a concurrent lock-free deque * supporting interior removes but not interior insertions, as * required to support the entire Deque interface. * * We extend the techniques developed for ConcurrentLinkedQueue and * LinkedTransferQueue (see the internal docs for those classes). * Understanding the ConcurrentLinkedQueue implementation is a * prerequisite for understanding the implementation of this class. * * The data structure is a symmetrical doubly-linked "GC-robust" * linked list of nodes. We minimize the number of volatile writes * using two techniques: advancing multiple hops with a single CAS * and mixing volatile and non-volatile writes of the same memory * locations. * * A node contains the expected E ("item") and links to predecessor * ("prev") and successor ("next") nodes: * * class Node { volatile Node prev, next; volatile E item; } * * A node p is considered "live" if it contains a non-null item * (p.item != null). When an item is CASed to null, the item is * atomically logically deleted from the collection. * * At any time, there is precisely one "first" node with a null * prev reference that terminates any chain of prev references * starting at a live node. Similarly there is precisely one * "last" node terminating any chain of next references starting at * a live node. The "first" and "last" nodes may or may not be live. * The "first" and "last" nodes are always mutually reachable. * * A new element is added atomically by CASing the null prev or * next reference in the first or last node to a fresh node * containing the element. The element's node atomically becomes * "live" at that point. * * A node is considered "active" if it is a live node, or the * first or last node. Active nodes cannot be unlinked. * * A "self-link" is a next or prev reference that is the same node: * p.prev == p or p.next == p * Self-links are used in the node unlinking process. Active nodes * never have self-links. * * A node p is active if and only if: * * p.item != null || * (p.prev == null && p.next != p) || * (p.next == null && p.prev != p) * * The deque object has two node references, "head" and "tail". * The head and tail are only approximations to the first and last * nodes of the deque. The first node can always be found by * following prev pointers from head; likewise for tail. However, * it is permissible for head and tail to be referring to deleted * nodes that have been unlinked and so may not be reachable from * any live node. * * There are 3 stages of node deletion; * "logical deletion", "unlinking", and "gc-unlinking". * * 1. "logical deletion" by CASing item to null atomically removes * the element from the collection, and makes the containing node * eligible for unlinking. * * 2. "unlinking" makes a deleted node unreachable from active * nodes, and thus eventually reclaimable by GC. Unlinked nodes * may remain reachable indefinitely from an iterator. * * Physical node unlinking is merely an optimization (albeit a * critical one), and so can be performed at our convenience. At * any time, the set of live nodes maintained by prev and next * links are identical, that is, the live nodes found via next * links from the first node is equal to the elements found via * prev links from the last node. However, this is not true for * nodes that have already been logically deleted - such nodes may * be reachable in one direction only. * * 3. "gc-unlinking" takes unlinking further by making active * nodes unreachable from deleted nodes, making it easier for the * GC to reclaim future deleted nodes. This step makes the data * structure "gc-robust", as first described in detail by Boehm * (http://portal.acm.org/citation.cfm?doid=503272.503282). * * GC-unlinked nodes may remain reachable indefinitely from an * iterator, but unlike unlinked nodes, are never reachable from * head or tail. * * Making the data structure GC-robust will eliminate the risk of * unbounded memory retention with conservative GCs and is likely * to improve performance with generational GCs. * * When a node is dequeued at either end, e.g. via poll(), we would * like to break any references from the node to active nodes. We * develop further the use of self-links that was very effective in * other concurrent collection classes. The idea is to replace * prev and next pointers with special values that are interpreted * to mean off-the-list-at-one-end. These are approximations, but * good enough to preserve the properties we want in our * traversals, e.g. we guarantee that a traversal will never visit * the same element twice, but we don't guarantee whether a * traversal that runs out of elements will be able to see more * elements later after enqueues at that end. Doing gc-unlinking * safely is particularly tricky, since any node can be in use * indefinitely (for example by an iterator). We must ensure that * the nodes pointed at by head/tail never get gc-unlinked, since * head/tail are needed to get "back on track" by other nodes that * are gc-unlinked. gc-unlinking accounts for much of the * implementation complexity. * * Since neither unlinking nor gc-unlinking are necessary for * correctness, there are many implementation choices regarding * frequency (eagerness) of these operations. Since volatile * reads are likely to be much cheaper than CASes, saving CASes by * unlinking multiple adjacent nodes at a time may be a win. * gc-unlinking can be performed rarely and still be effective, * since it is most important that long chains of deleted nodes * are occasionally broken. * * The actual representation we use is that p.next == p means to * goto the first node (which in turn is reached by following prev * pointers from head), and p.next == null && p.prev == p means * that the iteration is at an end and that p is a (static final) * dummy node, NEXT_TERMINATOR, and not the last active node. * Finishing the iteration when encountering such a TERMINATOR is * good enough for read-only traversals, so such traversals can use * p.next == null as the termination condition. When we need to * find the last (active) node, for enqueueing a new node, we need * to check whether we have reached a TERMINATOR node; if so, * restart traversal from tail. * * The implementation is completely directionally symmetrical, * except that most public methods that iterate through the list * follow next pointers ("forward" direction). * * We believe (without full proof) that all single-element deque * operations (e.g., addFirst, peekLast, pollLast) are linearizable * (see Herlihy and Shavit's book). However, some combinations of * operations are known not to be linearizable. In particular, * when an addFirst(A) is racing with pollFirst() removing B, it is * possible for an observer iterating over the elements to observe * A B C and subsequently observe A C, even though no interior * removes are ever performed. Nevertheless, iterators behave * reasonably, providing the "weakly consistent" guarantees. * * Empirically, microbenchmarks suggest that this class adds about * 40% overhead relative to ConcurrentLinkedQueue, which feels as * good as we can hope for. */ private static final long serialVersionUID = 876323262645176354L; /** * A node from which the first node on list (that is, the unique node p * with p.prev == null && p.next != p) can be reached in O(1) time. * Invariants: * - the first node is always O(1) reachable from head via prev links * - all live nodes are reachable from the first node via succ() * - head != null * - (tmp = head).next != tmp || tmp != head * - head is never gc-unlinked (but may be unlinked) * Non-invariants: * - head.item may or may not be null * - head may not be reachable from the first or last node, or from tail */ private transient volatile Node head; /** * A node from which the last node on list (that is, the unique node p * with p.next == null && p.prev != p) can be reached in O(1) time. * Invariants: * - the last node is always O(1) reachable from tail via next links * - all live nodes are reachable from the last node via pred() * - tail != null * - tail is never gc-unlinked (but may be unlinked) * Non-invariants: * - tail.item may or may not be null * - tail may not be reachable from the first or last node, or from head */ private transient volatile Node tail; private static final Node PREV_TERMINATOR, NEXT_TERMINATOR; @SuppressWarnings("unchecked") Node prevTerminator() { return (Node) PREV_TERMINATOR; } @SuppressWarnings("unchecked") Node nextTerminator() { return (Node) NEXT_TERMINATOR; } static final class Node { volatile Node prev; volatile E item; volatile Node next; } /** * Returns a new node holding item. Uses relaxed write because item * can only be seen after piggy-backing publication via CAS. */ static Node newNode(E item) { Node node = new Node<>(); ITEM.set(node, item); return node; } /** * Links e as first element. */ private Node linkFirst(E e) { final Node newNode = newNode(Objects.requireNonNull(e)); restartFromHead: for (;;) for (Node h = head, p = h, q;;) { if ((q = p.prev) != null && (q = (p = q).prev) != null) // Check for head updates every other hop. // If p == q, we are sure to follow head instead. p = (h != (h = head)) ? h : q; else if (p.next == p) // PREV_TERMINATOR continue restartFromHead; else { // p is first node NEXT.set(newNode, p); // CAS piggyback if (PREV.compareAndSet(p, null, newNode)) { // Successful CAS is the linearization point // for e to become an element of this deque, // and for newNode to become "live". if (p != h) // hop two nodes at a time; failure is OK HEAD.weakCompareAndSet(this, h, newNode); return newNode; } // Lost CAS race to another thread; re-read prev } } } /** * Links e as last element. */ private Node linkLast(E e) { final Node newNode = newNode(Objects.requireNonNull(e)); restartFromTail: for (;;) for (Node t = tail, p = t, q;;) { if ((q = p.next) != null && (q = (p = q).next) != null) // Check for tail updates every other hop. // If p == q, we are sure to follow tail instead. p = (t != (t = tail)) ? t : q; else if (p.prev == p) // NEXT_TERMINATOR continue restartFromTail; else { // p is last node PREV.set(newNode, p); // CAS piggyback if (NEXT.compareAndSet(p, null, newNode)) { // Successful CAS is the linearization point // for e to become an element of this deque, // and for newNode to become "live". if (p != t) // hop two nodes at a time; failure is OK TAIL.weakCompareAndSet(this, t, newNode); return newNode; } // Lost CAS race to another thread; re-read next } } } private static final int HOPS = 2; /** * Unlinks non-null node x. */ void unlink(Node x) { // assert x != null; // assert x.item == null; // assert x != PREV_TERMINATOR; // assert x != NEXT_TERMINATOR; final Node prev = x.prev; final Node next = x.next; if (prev == null) { unlinkFirst(x, next); } else if (next == null) { unlinkLast(x, prev); } else { // Unlink interior node. // // This is the common case, since a series of polls at the // same end will be "interior" removes, except perhaps for // the first one, since end nodes cannot be unlinked. // // At any time, all active nodes are mutually reachable by // following a sequence of either next or prev pointers. // // Our strategy is to find the unique active predecessor // and successor of x. Try to fix up their links so that // they point to each other, leaving x unreachable from // active nodes. If successful, and if x has no live // predecessor/successor, we additionally try to gc-unlink, // leaving active nodes unreachable from x, by rechecking // that the status of predecessor and successor are // unchanged and ensuring that x is not reachable from // tail/head, before setting x's prev/next links to their // logical approximate replacements, self/TERMINATOR. Node activePred, activeSucc; boolean isFirst, isLast; int hops = 1; // Find active predecessor for (Node p = prev; ; ++hops) { if (p.item != null) { activePred = p; isFirst = false; break; } Node q = p.prev; if (q == null) { if (p.next == p) return; activePred = p; isFirst = true; break; } else if (p == q) return; else p = q; } // Find active successor for (Node p = next; ; ++hops) { if (p.item != null) { activeSucc = p; isLast = false; break; } Node q = p.next; if (q == null) { if (p.prev == p) return; activeSucc = p; isLast = true; break; } else if (p == q) return; else p = q; } // TODO: better HOP heuristics if (hops < HOPS // always squeeze out interior deleted nodes && (isFirst | isLast)) return; // Squeeze out deleted nodes between activePred and // activeSucc, including x. skipDeletedSuccessors(activePred); skipDeletedPredecessors(activeSucc); // Try to gc-unlink, if possible if ((isFirst | isLast) && // Recheck expected state of predecessor and successor (activePred.next == activeSucc) && (activeSucc.prev == activePred) && (isFirst ? activePred.prev == null : activePred.item != null) && (isLast ? activeSucc.next == null : activeSucc.item != null)) { updateHead(); // Ensure x is not reachable from head updateTail(); // Ensure x is not reachable from tail // Finally, actually gc-unlink PREV.setRelease(x, isFirst ? prevTerminator() : x); NEXT.setRelease(x, isLast ? nextTerminator() : x); } } } /** * Unlinks non-null first node. */ private void unlinkFirst(Node first, Node next) { // assert first != null; // assert next != null; // assert first.item == null; for (Node o = null, p = next, q;;) { if (p.item != null || (q = p.next) == null) { if (o != null && p.prev != p && NEXT.compareAndSet(first, next, p)) { skipDeletedPredecessors(p); if (first.prev == null && (p.next == null || p.item != null) && p.prev == first) { updateHead(); // Ensure o is not reachable from head updateTail(); // Ensure o is not reachable from tail // Finally, actually gc-unlink NEXT.setRelease(o, o); PREV.setRelease(o, prevTerminator()); } } return; } else if (p == q) return; else { o = p; p = q; } } } /** * Unlinks non-null last node. */ private void unlinkLast(Node last, Node prev) { // assert last != null; // assert prev != null; // assert last.item == null; for (Node o = null, p = prev, q;;) { if (p.item != null || (q = p.prev) == null) { if (o != null && p.next != p && PREV.compareAndSet(last, prev, p)) { skipDeletedSuccessors(p); if (last.next == null && (p.prev == null || p.item != null) && p.next == last) { updateHead(); // Ensure o is not reachable from head updateTail(); // Ensure o is not reachable from tail // Finally, actually gc-unlink PREV.setRelease(o, o); NEXT.setRelease(o, nextTerminator()); } } return; } else if (p == q) return; else { o = p; p = q; } } } /** * Guarantees that any node which was unlinked before a call to * this method will be unreachable from head after it returns. * Does not guarantee to eliminate slack, only that head will * point to a node that was active while this method was running. */ private void updateHead() { // Either head already points to an active node, or we keep // trying to cas it to the first node until it does. Node h, p, q; restartFromHead: while ((h = head).item == null && (p = h.prev) != null) { for (;;) { if ((q = p.prev) == null || (q = (p = q).prev) == null) { // It is possible that p is PREV_TERMINATOR, // but if so, the CAS is guaranteed to fail. if (HEAD.compareAndSet(this, h, p)) return; else continue restartFromHead; } else if (h != head) continue restartFromHead; else p = q; } } } /** * Guarantees that any node which was unlinked before a call to * this method will be unreachable from tail after it returns. * Does not guarantee to eliminate slack, only that tail will * point to a node that was active while this method was running. */ private void updateTail() { // Either tail already points to an active node, or we keep // trying to cas it to the last node until it does. Node t, p, q; restartFromTail: while ((t = tail).item == null && (p = t.next) != null) { for (;;) { if ((q = p.next) == null || (q = (p = q).next) == null) { // It is possible that p is NEXT_TERMINATOR, // but if so, the CAS is guaranteed to fail. if (TAIL.compareAndSet(this, t, p)) return; else continue restartFromTail; } else if (t != tail) continue restartFromTail; else p = q; } } } private void skipDeletedPredecessors(Node x) { whileActive: do { Node prev = x.prev; // assert prev != null; // assert x != NEXT_TERMINATOR; // assert x != PREV_TERMINATOR; Node p = prev; findActive: for (;;) { if (p.item != null) break findActive; Node q = p.prev; if (q == null) { if (p.next == p) continue whileActive; break findActive; } else if (p == q) continue whileActive; else p = q; } // found active CAS target if (prev == p || PREV.compareAndSet(x, prev, p)) return; } while (x.item != null || x.next == null); } private void skipDeletedSuccessors(Node x) { whileActive: do { Node next = x.next; // assert next != null; // assert x != NEXT_TERMINATOR; // assert x != PREV_TERMINATOR; Node p = next; findActive: for (;;) { if (p.item != null) break findActive; Node q = p.next; if (q == null) { if (p.prev == p) continue whileActive; break findActive; } else if (p == q) continue whileActive; else p = q; } // found active CAS target if (next == p || NEXT.compareAndSet(x, next, p)) return; } while (x.item != null || x.prev == null); } /** * Returns the successor of p, or the first node if p.next has been * linked to self, which will only be true if traversing with a * stale pointer that is now off the list. */ final Node succ(Node p) { // TODO: should we skip deleted nodes here? if (p == (p = p.next)) p = first(); return p; } /** * Returns the predecessor of p, or the last node if p.prev has been * linked to self, which will only be true if traversing with a * stale pointer that is now off the list. */ final Node pred(Node p) { Node q = p.prev; return (p == q) ? last() : q; } /** * Returns the first node, the unique node p for which: * p.prev == null && p.next != p * The returned node may or may not be logically deleted. * Guarantees that head is set to the returned node. */ Node first() { restartFromHead: for (;;) for (Node h = head, p = h, q;;) { if ((q = p.prev) != null && (q = (p = q).prev) != null) // Check for head updates every other hop. // If p == q, we are sure to follow head instead. p = (h != (h = head)) ? h : q; else if (p == h // It is possible that p is PREV_TERMINATOR, // but if so, the CAS is guaranteed to fail. || HEAD.compareAndSet(this, h, p)) return p; else continue restartFromHead; } } /** * Returns the last node, the unique node p for which: * p.next == null && p.prev != p * The returned node may or may not be logically deleted. * Guarantees that tail is set to the returned node. */ Node last() { restartFromTail: for (;;) for (Node t = tail, p = t, q;;) { if ((q = p.next) != null && (q = (p = q).next) != null) // Check for tail updates every other hop. // If p == q, we are sure to follow tail instead. p = (t != (t = tail)) ? t : q; else if (p == t // It is possible that p is NEXT_TERMINATOR, // but if so, the CAS is guaranteed to fail. || TAIL.compareAndSet(this, t, p)) return p; else continue restartFromTail; } } // Minor convenience utilities /** * Returns element unless it is null, in which case throws * NoSuchElementException. * * @param v the element * @return the element */ private E screenNullResult(E v) { if (v == null) throw new NoSuchElementException(); return v; } /** * Constructs an empty deque. */ public FastConcurrentDirectDeque() { head = tail = new Node<>(); } /** * Constructs a deque initially containing the elements of * the given collection, added in traversal order of the * collection's iterator. * * @param c the collection of elements to initially contain * @throws NullPointerException if the specified collection or any * of its elements are null */ public FastConcurrentDirectDeque(Collection c) { // Copy c into a private chain of Nodes Node h = null, t = null; for (E e : c) { Node newNode = newNode(Objects.requireNonNull(e)); if (h == null) h = t = newNode; else { NEXT.set(t, newNode); PREV.set(newNode, t); t = newNode; } } initHeadTail(h, t); } /** * Initializes head and tail, ensuring invariants hold. */ private void initHeadTail(Node h, Node t) { if (h == t) { if (h == null) h = t = new Node<>(); else { // Avoid edge case of a single Node with non-null item. Node newNode = new Node<>(); NEXT.set(t, newNode); PREV.set(newNode, t); t = newNode; } } head = h; tail = t; } /** * Inserts the specified element at the front of this deque. * As the deque is unbounded, this method will never throw * {@link IllegalStateException}. * * @throws NullPointerException if the specified element is null */ public void addFirst(E e) { linkFirst(e); } /** * Inserts the specified element at the end of this deque. * As the deque is unbounded, this method will never throw * {@link IllegalStateException}. * *

This method is equivalent to {@link #add}. * * @throws NullPointerException if the specified element is null */ public void addLast(E e) { linkLast(e); } /** * Inserts the specified element at the front of this deque. * As the deque is unbounded, this method will never return {@code false}. * * @return {@code true} (as specified by {@link Deque#offerFirst}) * @throws NullPointerException if the specified element is null */ public boolean offerFirst(E e) { linkFirst(e); return true; } public Object offerFirstAndReturnToken(E e) { return linkFirst(e); } public Object offerLastAndReturnToken(E e) { return linkLast(e); } public void removeToken(Object token) { if (!(token instanceof Node)) { throw new IllegalArgumentException(); } Node node = (Node) (token); while (! ITEM.compareAndSet(node, node.item, null)) {} unlink(node); } /** * Inserts the specified element at the end of this deque. * As the deque is unbounded, this method will never return {@code false}. * *

This method is equivalent to {@link #add}. * * @return {@code true} (as specified by {@link Deque#offerLast}) * @throws NullPointerException if the specified element is null */ public boolean offerLast(E e) { linkLast(e); return true; } public E peekFirst() { for (Node p = first(); p != null; p = succ(p)) { final E item; if ((item = p.item) != null) return item; } return null; } public E peekLast() { for (Node p = last(); p != null; p = pred(p)) { final E item; if ((item = p.item) != null) return item; } return null; } /** * @throws NoSuchElementException {@inheritDoc} */ public E getFirst() { return screenNullResult(peekFirst()); } /** * @throws NoSuchElementException {@inheritDoc} */ public E getLast() { return screenNullResult(peekLast()); } public E pollFirst() { for (Node p = first(); p != null; p = succ(p)) { final E item; if ((item = p.item) != null && ITEM.compareAndSet(p, item, null)) { unlink(p); return item; } } return null; } public E pollLast() { for (Node p = last(); p != null; p = pred(p)) { final E item; if ((item = p.item) != null && ITEM.compareAndSet(p, item, null)) { unlink(p); return item; } } return null; } /** * @throws NoSuchElementException {@inheritDoc} */ public E removeFirst() { return screenNullResult(pollFirst()); } /** * @throws NoSuchElementException {@inheritDoc} */ public E removeLast() { return screenNullResult(pollLast()); } // *** Queue and stack methods *** /** * Inserts the specified element at the tail of this deque. * As the deque is unbounded, this method will never return {@code false}. * * @return {@code true} (as specified by {@link Queue#offer}) * @throws NullPointerException if the specified element is null */ public boolean offer(E e) { return offerLast(e); } /** * Inserts the specified element at the tail of this deque. * As the deque is unbounded, this method will never throw * {@link IllegalStateException} or return {@code false}. * * @return {@code true} (as specified by {@link Collection#add}) * @throws NullPointerException if the specified element is null */ public boolean add(E e) { return offerLast(e); } public E poll() { return pollFirst(); } public E peek() { return peekFirst(); } /** * @throws NoSuchElementException {@inheritDoc} */ public E remove() { return removeFirst(); } /** * @throws NoSuchElementException {@inheritDoc} */ public E pop() { return removeFirst(); } /** * @throws NoSuchElementException {@inheritDoc} */ public E element() { return getFirst(); } /** * @throws NullPointerException {@inheritDoc} */ public void push(E e) { addFirst( e ); } /** * Removes the first occurrence of the specified element from this deque. * If the deque does not contain the element, it is unchanged. * More formally, removes the first element {@code e} such that * {@code o.equals(e)} (if such an element exists). * Returns {@code true} if this deque contained the specified element * (or equivalently, if this deque changed as a result of the call). * * @param o element to be removed from this deque, if present * @return {@code true} if the deque contained the specified element * @throws NullPointerException if the specified element is null */ public boolean removeFirstOccurrence(Object o) { Objects.requireNonNull(o); for (Node p = first(); p != null; p = succ(p)) { final E item; if ((item = p.item) != null && o.equals(item) && ITEM.compareAndSet(p, item, null)) { unlink(p); return true; } } return false; } /** * Removes the last occurrence of the specified element from this deque. * If the deque does not contain the element, it is unchanged. * More formally, removes the last element {@code e} such that * {@code o.equals(e)} (if such an element exists). * Returns {@code true} if this deque contained the specified element * (or equivalently, if this deque changed as a result of the call). * * @param o element to be removed from this deque, if present * @return {@code true} if the deque contained the specified element * @throws NullPointerException if the specified element is null */ public boolean removeLastOccurrence(Object o) { Objects.requireNonNull(o); for (Node p = last(); p != null; p = pred(p)) { final E item; if ((item = p.item) != null && o.equals(item) && ITEM.compareAndSet(p, item, null)) { unlink(p); return true; } } return false; } /** * Returns {@code true} if this deque contains the specified element. * More formally, returns {@code true} if and only if this deque contains * at least one element {@code e} such that {@code o.equals(e)}. * * @param o element whose presence in this deque is to be tested * @return {@code true} if this deque contains the specified element */ public boolean contains(Object o) { if (o != null) { for (Node p = first(); p != null; p = succ(p)) { final E item; if ((item = p.item) != null && o.equals(item)) return true; } } return false; } /** * Returns {@code true} if this collection contains no elements. * * @return {@code true} if this collection contains no elements */ public boolean isEmpty() { return peekFirst() == null; } /** * Returns the number of elements in this deque. If this deque * contains more than {@code Integer.MAX_VALUE} elements, it * returns {@code Integer.MAX_VALUE}. * *

Beware that, unlike in most collections, this method is * NOT a constant-time operation. Because of the * asynchronous nature of these deques, determining the current * number of elements requires traversing them all to count them. * Additionally, it is possible for the size to change during * execution of this method, in which case the returned result * will be inaccurate. Thus, this method is typically not very * useful in concurrent applications. * * @return the number of elements in this deque */ public int size() { restartFromHead: for (;;) { int count = 0; for (Node p = first(); p != null;) { if (p.item != null) if (++count == Integer.MAX_VALUE) break; // @see Collection.size() if (p == (p = p.next)) continue restartFromHead; } return count; } } /** * Removes the first occurrence of the specified element from this deque. * If the deque does not contain the element, it is unchanged. * More formally, removes the first element {@code e} such that * {@code o.equals(e)} (if such an element exists). * Returns {@code true} if this deque contained the specified element * (or equivalently, if this deque changed as a result of the call). * *

This method is equivalent to {@link #removeFirstOccurrence(Object)}. * * @param o element to be removed from this deque, if present * @return {@code true} if the deque contained the specified element * @throws NullPointerException if the specified element is null */ public boolean remove(Object o) { return removeFirstOccurrence(o); } /** * Appends all of the elements in the specified collection to the end of * this deque, in the order that they are returned by the specified * collection's iterator. Attempts to {@code addAll} of a deque to * itself result in {@code IllegalArgumentException}. * * @param c the elements to be inserted into this deque * @return {@code true} if this deque changed as a result of the call * @throws NullPointerException if the specified collection or any * of its elements are null * @throws IllegalArgumentException if the collection is this deque */ public boolean addAll(Collection c) { if (c == this) // As historically specified in AbstractQueue#addAll throw new IllegalArgumentException(); // Copy c into a private chain of Nodes Node beginningOfTheEnd = null, last = null; for (E e : c) { Node newNode = newNode(Objects.requireNonNull(e)); if (beginningOfTheEnd == null) beginningOfTheEnd = last = newNode; else { NEXT.set(last, newNode); PREV.set(newNode, last); last = newNode; } } if (beginningOfTheEnd == null) return false; // Atomically append the chain at the tail of this collection restartFromTail: for (;;) for (Node t = tail, p = t, q;;) { if ((q = p.next) != null && (q = (p = q).next) != null) // Check for tail updates every other hop. // If p == q, we are sure to follow tail instead. p = (t != (t = tail)) ? t : q; else if (p.prev == p) // NEXT_TERMINATOR continue restartFromTail; else { // p is last node PREV.set(beginningOfTheEnd, p); // CAS piggyback if (NEXT.compareAndSet(p, null, beginningOfTheEnd)) { // Successful CAS is the linearization point // for all elements to be added to this deque. if (!TAIL.weakCompareAndSet(this, t, last)) { // Try a little harder to update tail, // since we may be adding many elements. t = tail; if (last.next == null) TAIL.weakCompareAndSet(this, t, last); } return true; } // Lost CAS race to another thread; re-read next } } } /** * Removes all of the elements from this deque. */ public void clear() { while (pollFirst() != null) {} } public String toString() { String[] a = null; restartFromHead: for (;;) { int charLength = 0; int size = 0; for (Node p = first(); p != null;) { final E item; if ((item = p.item) != null) { if (a == null) a = new String[4]; else if (size == a.length) a = Arrays.copyOf(a, 2 * size); String s = item.toString(); a[size++] = s; charLength += s.length(); } if (p == (p = p.next)) continue restartFromHead; } if (size == 0) return "[]"; return toString(a, size, charLength); } } private Object[] toArrayInternal(Object[] a) { Object[] x = a; restartFromHead: for (;;) { int size = 0; for (Node p = first(); p != null;) { final E item; if ((item = p.item) != null) { if (x == null) x = new Object[4]; else if (size == x.length) x = Arrays.copyOf(x, 2 * (size + 4)); x[size++] = item; } if (p == (p = p.next)) continue restartFromHead; } if (x == null) return new Object[0]; else if (a != null && size <= a.length) { if (a != x) System.arraycopy(x, 0, a, 0, size); if (size < a.length) a[size] = null; return a; } return (size == x.length) ? x : Arrays.copyOf(x, size); } } /** * Returns an array containing all of the elements in this deque, in * proper sequence (from first to last element). * *

The returned array will be "safe" in that no references to it are * maintained by this deque. (In other words, this method must allocate * a new array). The caller is thus free to modify the returned array. * *

This method acts as bridge between array-based and collection-based * APIs. * * @return an array containing all of the elements in this deque */ public Object[] toArray() { return toArrayInternal(null); } /** * Returns an array containing all of the elements in this deque, * in proper sequence (from first to last element); the runtime * type of the returned array is that of the specified array. If * the deque fits in the specified array, it is returned therein. * Otherwise, a new array is allocated with the runtime type of * the specified array and the size of this deque. * *

If this deque fits in the specified array with room to spare * (i.e., the array has more elements than this deque), the element in * the array immediately following the end of the deque is set to * {@code null}. * *

Like the {@link #toArray()} method, this method acts as * bridge between array-based and collection-based APIs. Further, * this method allows precise control over the runtime type of the * output array, and may, under certain circumstances, be used to * save allocation costs. * *

Suppose {@code x} is a deque known to contain only strings. * The following code can be used to dump the deque into a newly * allocated array of {@code String}: * *

 {@code String[] y = x.toArray(new String[0]);}
* * Note that {@code toArray(new Object[0])} is identical in function to * {@code toArray()}. * * @param a the array into which the elements of the deque are to * be stored, if it is big enough; otherwise, a new array of the * same runtime type is allocated for this purpose * @return an array containing all of the elements in this deque * @throws ArrayStoreException if the runtime type of the specified array * is not a supertype of the runtime type of every element in * this deque * @throws NullPointerException if the specified array is null */ @SuppressWarnings("unchecked") public T[] toArray(T[] a) { return (T[]) toArrayInternal(checkNotNullParamWithNullPointerException("a", a)); } /** * Returns an iterator over the elements in this deque in proper sequence. * The elements will be returned in order from first (head) to last (tail). * *

The returned iterator is * weakly consistent. * * @return an iterator over the elements in this deque in proper sequence */ public Iterator iterator() { return new Itr(); } /** * Returns an iterator over the elements in this deque in reverse * sequential order. The elements will be returned in order from * last (tail) to first (head). * *

The returned iterator is * weakly consistent. * * @return an iterator over the elements in this deque in reverse order */ public Iterator descendingIterator() { return new DescendingItr(); } // From Helpers 1.2 // http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/main/java/util/concurrent/Helpers.java?revision=1.2 /** * Like Arrays.toString(), but caller guarantees that size > 0, * each element with index 0 <= i < size is a non-null String, * and charLength is the sum of the lengths of the input Strings. */ private String toString(Object[] a, int size, int charLength) { // assert a != null; // assert size > 0; // Copy each string into a perfectly sized char[] // Length of [ , , , ] == 2 * size final char[] chars = new char[charLength + 2 * size]; chars[0] = '['; int j = 1; for (int i = 0; i < size; i++) { if (i > 0) { chars[j++] = ','; chars[j++] = ' '; } String s = (String) a[i]; int len = s.length(); s.getChars(0, len, chars, j); j += len; } chars[j] = ']'; // assert j == chars.length - 1; return new String(chars); } private abstract class AbstractItr implements Iterator { /** * Next node to return item for. */ private Node nextNode; /** * nextItem holds on to item fields because once we claim * that an element exists in hasNext(), we must return it in * the following next() call even if it was in the process of * being removed when hasNext() was called. */ private E nextItem; /** * Node returned by most recent call to next. Needed by remove. * Reset to null if this element is deleted by a call to remove. */ private Node lastRet; abstract Node startNode(); abstract Node nextNode(Node p); AbstractItr() { advance(); } /** * Sets nextNode and nextItem to next valid node, or to null * if no such. */ private void advance() { lastRet = nextNode; Node p = (nextNode == null) ? startNode() : nextNode(nextNode); for (;; p = nextNode(p)) { if (p == null) { // might be at active end or TERMINATOR node; both are OK nextNode = null; nextItem = null; break; } final E item; if ((item = p.item) != null) { nextNode = p; nextItem = item; break; } } } public boolean hasNext() { return nextItem != null; } public E next() { E item = nextItem; if (item == null) throw new NoSuchElementException(); advance(); return item; } public void remove() { Node l = lastRet; if (l == null) throw new IllegalStateException(); l.item = null; unlink(l); lastRet = null; } } /** * Forward iterator */ private class Itr extends AbstractItr { Itr() {} // prevent access constructor creation Node startNode() { return first(); } Node nextNode(Node p) { return succ( p ); } } /** * Descending iterator */ private class DescendingItr extends AbstractItr { DescendingItr() {} // prevent access constructor creation Node startNode() { return last(); } Node nextNode(Node p) { return pred( p ); } } /** A customized variant of Spliterators.IteratorSpliterator */ final class CLDSpliterator implements Spliterator { static final int MAX_BATCH = 1 << 25; // max batch array size; Node current; // current node; null until initialized int batch; // batch size for splits boolean exhausted; // true when no more nodes public Spliterator trySplit() { Node p, q; if ((p = current()) == null || (q = p.next) == null) return null; int i = 0, n = batch = Math.min(batch + 1, MAX_BATCH); Object[] a = null; do { final E e; if ((e = p.item) != null) { if (a == null) a = new Object[n]; a[i++] = e; } if (p == (p = q)) p = first(); } while (p != null && (q = p.next) != null && i < n); setCurrent(p); return (i == 0) ? null : Spliterators.spliterator(a, 0, i, (Spliterator.ORDERED | Spliterator.NONNULL | Spliterator.CONCURRENT)); } public void forEachRemaining(Consumer action) { Objects.requireNonNull(action); Node p; if ((p = current()) != null) { current = null; exhausted = true; do { final E e; if ((e = p.item) != null) action.accept(e); if (p == (p = p.next)) p = first(); } while (p != null); } } public boolean tryAdvance(Consumer action) { Objects.requireNonNull(action); Node p; if ((p = current()) != null) { E e; do { e = p.item; if (p == (p = p.next)) p = first(); } while (e == null && p != null); setCurrent(p); if (e != null) { action.accept(e); return true; } } return false; } private void setCurrent(Node p) { if ((current = p) == null) exhausted = true; } private Node current() { Node p; if ((p = current) == null && !exhausted) setCurrent(p = first()); return p; } public long estimateSize() { return Long.MAX_VALUE; } public int characteristics() { return (Spliterator.ORDERED | Spliterator.NONNULL | Spliterator.CONCURRENT); } } /** * Returns a {@link Spliterator} over the elements in this deque. * *

The returned spliterator is * weakly consistent. * *

The {@code Spliterator} reports {@link Spliterator#CONCURRENT}, * {@link Spliterator#ORDERED}, and {@link Spliterator#NONNULL}. * * @implNote * The {@code Spliterator} implements {@code trySplit} to permit limited * parallelism. * * @return a {@code Spliterator} over the elements in this deque * @since 1.8 */ public Spliterator spliterator() { return new CLDSpliterator(); } /** * Saves this deque to a stream (that is, serializes it). * * @param s the stream * @throws java.io.IOException if an I/O error occurs * @serialData All of the elements (each an {@code E}) in * the proper order, followed by a null */ private void writeObject(java.io.ObjectOutputStream s) throws java.io.IOException { // Write out any hidden stuff s.defaultWriteObject(); // Write out all elements in the proper order. for (Node p = first(); p != null; p = succ(p)) { final E item; if ((item = p.item) != null) s.writeObject(item); } // Use trailing null as sentinel s.writeObject(null); } /** * Reconstitutes this deque from a stream (that is, deserializes it). * @param s the stream * @throws ClassNotFoundException if the class of a serialized object * could not be found * @throws java.io.IOException if an I/O error occurs */ private void readObject(java.io.ObjectInputStream s) throws java.io.IOException, ClassNotFoundException { s.defaultReadObject(); // Read in elements until trailing null sentinel found Node h = null, t = null; for (Object item; (item = s.readObject()) != null; ) { @SuppressWarnings("unchecked") Node newNode = newNode((E) item); if (h == null) h = t = newNode; else { NEXT.set(t, newNode); PREV.set(newNode, t); t = newNode; } } initHeadTail(h, t); } /** * @throws NullPointerException {@inheritDoc} */ public boolean removeIf(Predicate filter) { Objects.requireNonNull(filter); return bulkRemove(filter); } /** * @throws NullPointerException {@inheritDoc} */ public boolean removeAll(Collection c) { Objects.requireNonNull(c); return bulkRemove(e -> c.contains(e)); } /** * @throws NullPointerException {@inheritDoc} */ public boolean retainAll(Collection c) { Objects.requireNonNull(c); return bulkRemove(e -> !c.contains(e)); } /** Implementation of bulk remove methods. */ private boolean bulkRemove(Predicate filter) { boolean removed = false; for (Node p = first(), succ; p != null; p = succ) { succ = succ(p); final E item; if ((item = p.item) != null && filter.test(item) && ITEM.compareAndSet(p, item, null)) { unlink(p); removed = true; } } return removed; } /** * @throws NullPointerException {@inheritDoc} */ public void forEach(Consumer action) { Objects.requireNonNull(action); E item; for (Node p = first(); p != null; p = succ(p)) if ((item = p.item) != null) action.accept(item); } // VarHandle mechanics private static final VarHandle HEAD; private static final VarHandle TAIL; private static final VarHandle PREV; private static final VarHandle NEXT; private static final VarHandle ITEM; static { PREV_TERMINATOR = new Node<>(); PREV_TERMINATOR.next = PREV_TERMINATOR; NEXT_TERMINATOR = new Node<>(); NEXT_TERMINATOR.prev = NEXT_TERMINATOR; try { MethodHandles.Lookup l = MethodHandles.lookup(); HEAD = l.findVarHandle(FastConcurrentDirectDeque.class, "head", Node.class); TAIL = l.findVarHandle(FastConcurrentDirectDeque.class, "tail", Node.class); PREV = l.findVarHandle(Node.class, "prev", Node.class); NEXT = l.findVarHandle(Node.class, "next", Node.class); ITEM = l.findVarHandle(Node.class, "item", Object.class); } catch (ReflectiveOperationException e) { throw new Error(e); } } }