JAVA线程池
1.1Java 线程池之 Executor 框架
为了实现线程池和管理线程池,JDK 给我们提供了基于 Executor 接口的一系列接口、抽象类、实现类,我们把它称作线程池的 Executor 框架,Executor 框架本质上是一个线程池;
Java 线程(java.lang.Thread)被一对一映射为本地操作系统内核线程,Java 线程启动时会创建一个本地操作系统线程,操作系统会调度所有线程并将它们分配给可用的 CPU 执行,当该 Java 线程终止时,这个操作系统线程也会被回收;
实际上这是两层线程调度模型:
(1)上层 Java 线程的调度由 Executor 框架调度;
(2)下层操作系统的线程调度由操作系统调度;
Java 的线程是这么设计的,包含两部分:
1、工作任务;(Runnable 和 Callable)
2、执行机制;(Thread、Executor 框架)
1.2Executor 框架 的接口与类结构
- java.util.concurrent (并发编程的工具) juc
- java.util.concurrent.atomic (变量的线程安全的原子性操作)
- java.util.concurrent.locks (用于锁定和条件等待同步等)
- Executor [ɪɡˈzekjʊtə] 执行人、执行者
1.3线程池的七大参数
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
5,
10,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(5),
Executors.defaultThreadFactory(),
new ThreadPoolExecutor.CallerRunsPolicy()
);
构造方法最多是 7 个参数;
1)int corePoolSize,
指定线程池中的核心线程数量(最少的线程个数),线程池中会维护一个最小的线程数量,即使这些线程处理空闲状态,它们也不会被销毁,除非设置了 allowCoreThreadTimeOut;默认情况下,创建线程池之后,线程池中是没有线程的,需要提交任务之后才会创建线程;在实际中如果需要线程池创建之后立即创建线程,可以通过以下两种方式:
prestartCoreThread():boolean prestartCoreThread(),初始化一个核心线程;
prestartAllCoreThreads():int prestartAllCoreThreads(),初始化所有核心线程;
2) BlockingQueue workQueue,
任务队列,当核心线程全部繁忙时,由 execute/submit 方法提交的 Runnable 任务存放到该任务队列中,等待被核心线程来执行;
3)int maximumPoolSize
指定线程池中允许的最大线程数,当核心线程全部繁忙且任务队列存满之后,线程池会临时追加线程,直到总线程数达到 maximumPoolSize 这个上限;
4)long keepAliveTime,
线程空闲超时时间,如果一个线程处于空闲状态,并且当前的线程数量大于 corePoolSize,那么在指定时间后,这个空闲线程会被销毁;
5) TimeUnit unit
keepAliveTime 的时间单位 (天、小时、分、秒…)
6) ThreadFactory threadFactory,
线程工厂,用于创建线程,一般采用默认的即可,也可以自定义实现;
Executors.defaultThreadFactory(),
Executors.privilegedThreadFactory(),
7) RejectedExecutionHandler handler,
拒绝策略(饱和策略),当任务太多来不及处理时,如何“拒绝”任务?
任务拒绝是线程池的保护措施,当核心线程 corePoolSize 正在执行任务、线程池的任务队列
workQueue 已满、并且线程池中的线程数达到 maximumPoolSize 时,就需要“拒绝”掉新提交
过来的任务;
示例:
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
public class Test01 {
public static void main(String[] args) {
Thread t=new Thread(){
@Override
public void run() {
System.out.println("Runnable任务1");
}
};
t.start();
Thread t2=new Thread(new Runnable() {
@Override
public void run() {
System.out.println("Runnable任务2");
}
});
t2.start();
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor=new ThreadPoolExecutor(
5,
10,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(5),
new ThreadPoolExecutor.CallerRunsPolicy());
threadPoolExecutor.execute(()->{
System.out.println("工作任务2");
});
//线程池关闭
threadPoolExecutor.shutdown();
}
}
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
public class Test02 {
//基于Executor框架实现线程池 (此时线程池中一个线程也没有)
public static void main(String[] args) {
ThreadPoolExecutor threadPoolExecutor=new ThreadPoolExecutor(
5,
10,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(5),
Executors.defaultThreadFactory(),
new ThreadPoolExecutor.DiscardPolicy()
);
//当提交了一个工作任务,此时线程池中就有一个线程
threadPoolExecutor.execute(()->{
System.out.println(Thread.currentThread().getName());
});
//关闭线程池
//threadPoolExecutor.shutdown();
//当核心线程处于空闲状态时候,允许销毁这些空闲的核心线程,默认是不允许销毁核心线程的
//threadPoolExecutor.allowCoreThreadTimeOut(true);
//如果想创建线程池后,立刻就创建好线程,那么执行:
threadPoolExecutor.prestartCoreThread();//初始化/创建一个核心线程
threadPoolExecutor.prestartAllCoreThreads();//初始化/创建所有的核心线程
for (int i=0;i<50;i++){
threadPoolExecutor.execute(()->{
System.out.println(Thread.currentThread().getName());
});
}
threadPoolExecutor.shutdown();
}
}
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.ThreadFactory;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
public class Test03 {
public static void main(String[] args) {
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor=new ThreadPoolExecutor(
5,
12,
5,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(5),
new MyThreadFactory(),
new ThreadPoolExecutor.DiscardPolicy()
);
threadPoolExecutor.execute(()->{
System.out.println(Thread.currentThread().getName());
});
}
/**
* 自己实现线程工厂
*/
static class MyThreadFactory implements ThreadFactory {
@Override
public Thread newThread(Runnable r) {
return new Thread(r, "my-thread");
}
}
}
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
/**
* ThreadPoolExecutor线程池
*
* @author Cat老师,关注我,抖音搜索:java512
*/
public class Test04 {
public static void main(String[] args) {
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
1,
1,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(2),
Executors.defaultThreadFactory(),
//Executors.privilegedThreadFactory(),
new ThreadPoolExecutor.CallerRunsPolicy()
);
//同时提交4个任务
threadPoolExecutor.execute(new MyRunnable(1));
threadPoolExecutor.execute(new MyRunnable(2));
threadPoolExecutor.execute(new MyRunnable(3));
threadPoolExecutor.execute(new MyRunnable(4));
threadPoolExecutor.shutdown();
}
static class MyRunnable implements Runnable {
private int i;
public MyRunnable(int i) {
this.i = i;
}
@Override
public void run() {
System.out.println(Thread.currentThread().getName() + ": " + this.i);
}
}
}
运行结果:
main: 4
pool-1-thread-1: 1
pool-1-thread-1: 2
pool-1-thread-1: 3
Process finished with exit code 0
1.4线程池的拒绝策略
JDK 提供了 4 种内置的拒绝策略:AbortPolicy、CallerRunsPolicy、DiscardOldestPolicy 和DiscardPolicy;
1、AbortPolicy(默认):丢弃任务并抛出 RejectedExecutionException 异常,这是线程池默认、的拒绝策略,在任务不能再提交的时候抛出异常,让开发人员及时知道程序运行状态,这样能在系统不能承载更大的并发量时,及时通过异常信息发现;
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
public class Test05 {
public static void main(String[] args) {
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
8,
16,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(10),
Executors.defaultThreadFactory(),
//Executors.privilegedThreadFactory(),
new ThreadPoolExecutor.AbortPolicy()
//new MyRejectedExecutionHandler()
);
for (int i = 0; i < 40000; i++) {
threadPoolExecutor.execute(new MyRunnable(i));
}
threadPoolExecutor.shutdown();
}
static class MyRunnable implements Runnable {
private int i;
public MyRunnable(int i) {
this.i = i;
}
@Override
public void run() {
System.out.println(Thread.currentThread().getName() + ": " + this.i);
}
}
}
运行结果
Exception in thread "main" java.util.concurrent.RejectedExecutionException: Task com.lisus.threadpool.Test05$MyRunnable@61bbe9ba rejected from java.util.concurrent.ThreadPoolExecutor@610455d6[Running, pool size = 16, active threads = 16, queued tasks = 10, completed tasks = 0]
at java.util.concurrent.ThreadPoolExecutor$AbortPolicy.rejectedExecution(ThreadPoolExecutor.java:2063)
at java.util.concurrent.ThreadPoolExecutor.reject(ThreadPoolExecutor.java:830)
at java.util.concurrent.ThreadPoolExecutor.execute(ThreadPoolExecutor.java:1379)
at com.lisus.threadpool.Test05.main(Test05.java:24)
pool-1-thread-1: 0
pool-1-thread-1: 8
pool-1-thread-1: 9
pool-1-thread-1: 10
pool-1-thread-1: 11
pool-1-thread-1: 12
pool-1-thread-1: 13
pool-1-thread-1: 14
pool-1-thread-1: 15
pool-1-thread-1: 16
pool-1-thread-1: 17
pool-1-thread-2: 1
pool-1-thread-3: 2
pool-1-thread-4: 3
pool-1-thread-5: 4
pool-1-thread-6: 5
pool-1-thread-7: 6
pool-1-thread-8: 7
pool-1-thread-9: 18
pool-1-thread-10: 19
pool-1-thread-11: 20
pool-1-thread-12: 21
pool-1-thread-13: 22
pool-1-thread-14: 23
pool-1-thread-15: 24
pool-1-thread-16: 25
2、DiscardPolicy:直接丢弃任务,不抛出异常,使用此策略可能会使我们无法发现系统的异、常状态,建议一些无关紧要的业务采用此策略;
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
public class Test05 {
public static void main(String[] args) {
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
8,
16,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(10),
Executors.defaultThreadFactory(),
//Executors.privilegedThreadFactory(),
new ThreadPoolExecutor.DiscardPolicy()
//new MyRejectedExecutionHandler()
);
for (int i = 0; i < 40000; i++) {
threadPoolExecutor.execute(new MyRunnable(i));
}
threadPoolExecutor.shutdown();
}
static class MyRunnable implements Runnable {
private int i;
public MyRunnable(int i) {
this.i = i;
}
@Override
public void run() {
System.out.println(Thread.currentThread().getName() + ": " + this.i);
}
}
}
运行结果
pool-1-thread-1: 0
pool-1-thread-1: 8
pool-1-thread-1: 9
pool-1-thread-1: 10
pool-1-thread-1: 11
pool-1-thread-1: 12
pool-1-thread-1: 13
pool-1-thread-1: 14
pool-1-thread-1: 15
pool-1-thread-1: 16
pool-1-thread-1: 17
pool-1-thread-2: 1
pool-1-thread-3: 2
pool-1-thread-4: 3
pool-1-thread-5: 4
pool-1-thread-6: 5
pool-1-thread-7: 6
pool-1-thread-8: 7
pool-1-thread-9: 18
pool-1-thread-10: 19
pool-1-thread-11: 20
pool-1-thread-12: 21
pool-1-thread-13: 22
pool-1-thread-14: 23
pool-1-thread-15: 24
pool-1-thread-16: 25
Process finished with exit code 0
3、DiscardOldestPolicy:丢弃任务队列中靠最前的任务,并执行当前任务,是否要采用此拒绝策略,根据实际业务是否允许丢弃老任务来评估和衡量;
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
public class Test05 {
public static void main(String[] args) {
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
8,
16,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(10),
Executors.defaultThreadFactory(),
//Executors.privilegedThreadFactory(),
new ThreadPoolExecutor.DiscardOldestPolicy()
//new MyRejectedExecutionHandler()
);
for (int i = 0; i < 40000; i++) {
threadPoolExecutor.execute(new MyRunnable(i));
}
threadPoolExecutor.shutdown();
}
static class MyRunnable implements Runnable {
private int i;
public MyRunnable(int i) {
this.i = i;
}
@Override
public void run() {
System.out.println(Thread.currentThread().getName() + ": " + this.i);
}
}
}
4、CallerRunsPolicy: 交由任务的调用线程(提交任务的线程)来执行当前任务;这种拒绝策略会让所有任务都能得到执行,适合大量计算类型的任务执行,使用这种策略的最终目标是要、让每个任务都能执行完毕,而使用多线程执行计算任务只是作为增大吞吐量的手段;
新来的任务可以用 main 线程去执行,不用线程池里面的线程执行;
package com.lisus.threadpool;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.Executors;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
public class Test05 {
public static void main(String[] args) {
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
8,
16,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(10),
Executors.defaultThreadFactory(),
//Executors.privilegedThreadFactory(),
new ThreadPoolExecutor.CallerRunsPolicy()
//new MyRejectedExecutionHandler()
);
for (int i = 0; i < 40000; i++) {
threadPoolExecutor.execute(new MyRunnable(i));
}
threadPoolExecutor.shutdown();
}
static class MyRunnable implements Runnable {
private int i;
public MyRunnable(int i) {
this.i = i;
}
@Override
public void run() {
System.out.println(Thread.currentThread().getName() + ": " + this.i);
}
}
}
除了上面的四种拒绝策略,还可以通过实现 RejectedExecutionHandler 接口,实现自定义的拒绝策略;
package com.lisus.threadpool;
import java.util.concurrent.*;
public class Test05 {
public static void main(String[] args) {
//基于Executor框架实现线程池
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(
8,
16,
15,
TimeUnit.SECONDS,
new ArrayBlockingQueue<Runnable>(10),
Executors.defaultThreadFactory(),
//Executors.privilegedThreadFactory(),
//new ThreadPoolExecutor.CallerRunsPolicy()
new MyRejectedExecutionHandler()
);
for (int i = 0; i < 40000; i++) {
threadPoolExecutor.execute(new MyRunnable(i));
}
threadPoolExecutor.shutdown();
}
static class MyRejectedExecutionHandler implements RejectedExecutionHandler {
@Override
public void rejectedExecution(Runnable r, ThreadPoolExecutor executor) {
//如果任务队列满了,就超时等待,可以设置一个时间
try {
executor.getQueue().offer(r, 60, TimeUnit.SECONDS);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
}
static class MyRunnable implements Runnable {
private int i;
public MyRunnable(int i) {
this.i = i;
}
@Override
public void run() {
System.out.println(Thread.currentThread().getName() + ": " + this.i);
}
}
}
AbortPolicy 异常中止策略:异常中止,无特殊场景;
DiscardPolicy 丢弃策略:无关紧要的任务(文章点击量、商品浏览量等);
DiscardOldestPolicy 弃老策略:允许丢掉老数据的场景;
CallerRunsPolicy 调用者运行策略:不允许失败场景(对性能要求不高、并发量较小的场景);
1.5线程池的原理
1.6线程池底层源码实现
1.6.1线程池构造方法
1.6.2线程池源码-控制变量
COUNT_BITS = 29
CAPACITY = (1 << COUNT_BITS) - 1
int 类型的数是占用 4 字节,32 位,所以前面填了一堆 0;
原码:00000000 00000000 00000000 00000001
左移:00100000 00000000 00000000 00000000
减一:00011111 11111111 11111111 11111111 (536870911 = 5 亿多)
1.6.3线程池源码-线程池状态值
1.6.7线程池源码- 核心源码解读-execute
public void execute(Runnable command) {
if (command == null)
throw new NullPointerException();
//获取 clt 控制变量的值,clt 控制变量记录着 runState 和 workerCount 的值;
int c = ctl.get();
/**
* workerCountOf方法获取控制变量ctl低29位值
* 如果当前活动线程小于核心线程corePoolSize,,则新建一个线程放入线程池中,并把任务添加到该线程中运行
*/
if (workerCountOf(c) < corePoolSize) {
/**
* addWorker 方法
* 第一参数是要提交的工作任务
* 第二个参数:
* 如果是true,根据corePoolSize来判断表示添加核心线程;(保持稳定的线程数来处理任务)
* 如果是 false,根据 maximumPoolSize 来判断,表示添加非核心线程;(应对突发的任务处理)
*/
//addWorker()方法会检查运行状态和工作线程数,如果返回 false 则说明线程没有创建成功;
if (addWorker(command, true))
//添加成功则返回;
return;
//如果添加失败,则重新获取控制变量 ctl 的值;
c = ctl.get();
}
//到这里了,说明 workerCountOf(c) >= corePoolSize,并且如果当前线程池是运行状态并且工作任务添加到任务队列成功
if (isRunning(c) && workQueue.offer(command)) {
// 重新获取 ctl 值
int recheck = ctl.get();
//再次判断线程池是否是运行状态,如果不是运行状态,由于之前已经把 command 添加到 workQueue 中了,此时需要移除该 command;
if (! isRunning(recheck) && remove(command))
reject(command);
/**
* 线程池是运行状态,获取一下线程池中的有效线程数,如果是 0,则执行 addWorker()方法;
* addWorker()方法:
* 第一个参数为 null,表示在线程池中创建一个线程,但不启动;
* 第二个参数为 false,表示是非核心线程;
*
* 接下来这里没有写 else,表示如果判断 workerCount 大于 0,则不需要做什么处理,直接返回,
* 加入到 workQueue 中的 command 会在将来的某个时刻被执行;
*/
else if (workerCountOf(recheck) == 0)
//此处是创建一个线程,但并没有传入任务,因为任务已经被添加到 workQueue 中了,到时候线程会从从 workQueue 中获取任务来执行;
//所以当 workerCountOf(recheck) == 0 时执行 addWorker(null, false);
//是为了保证线程池在 RUNNING 状态下必须要有一个线程来执行任务;
addWorker(null, false);
}
/**
* 如果执行到这里,有两种情况:
* 1. 线程池已经不是 RUNNING 状态;
* 2. 线程池是 RUNNING 状态,但往 workQueue 已经放不进去,即 workerCount >= corePoolSize,并且 workQueue 已满;
* 此时再次调用 addWorker()方法,第二个参数为 false,表示非核心线程,如果失败则拒绝该任务;
*/
else if (!addWorker(command, false))
reject(command);
}
1.6.8线程池源码-核心源码解读-addWorker
private boolean addWorker(Runnable firstTask, boolean core) {
retry:
for (;;) {
// 获取线程池控制变量的值
int c = ctl.get();
//线程运行状态
int rs = runStateOf(c);
// if判断,如果rs>=SHUTDOWN,并且(判断3个条件,只要有1个不满足)返回false
if (rs >= SHUTDOWN &&
! (rs == SHUTDOWN &&
firstTask == null &&
! workQueue.isEmpty()))
return false;
for (;;) {
//获取线程数
int wc = workerCountOf(c);
// 如果 wc 超过 CAPACITY,也就是 ctl 的低 29 位的最大值(二进制是 29 个 1),返回 false;
if (wc >= CAPACITY ||
wc >= (core ? corePoolSize : maximumPoolSize))
return false;
// 尝试增加 workerCount,如果成功,则跳出外层 for 循环
if (compareAndIncrementWorkerCount(c))
break retry;
// 如果增加 workerCount 失败,则重新获取控制变量 ctl 的值
c = ctl.get(); // Re-read ctl
// 如果当前线程池的运行状态不等于 rs,说明线程池运行状态已被改变,返回外层 for 循环继续执行
if (runStateOf(c) != rs)
continue retry;
// else CAS failed due to workerCount change; retry inner loop
}
}
// Worker 线程是否启动
boolean workerStarted = false;
// Worker 线程是否添加
boolean workerAdded = false;
Worker w = null;
try {
// 根据 firstTask 来创建 Worker 对象
w = new Worker(firstTask);
final Thread t = w.thread;
if (t != null) {
final ReentrantLock mainLock = this.mainLock;
mainLock.lock();
try {
// 检查线程池运行状态
int rs = runStateOf(ctl.get());
// rs < SHUTDOWN 表示是 RUNNING 状态;
// 如果 rs 是 RUNNING 状态或者 rs 是 SHUTDOWN 状态并且 firstTask 为 null,向线程池中添加线程。
// 因为在 SHUTDOWN 时不会在添加新的任务,但还是会执行 workQueue 中的任务
if (rs < SHUTDOWN ||
(rs == SHUTDOWN && firstTask == null)) {
// 检查线程已经是运行状态,抛出非法线程状态异常
if (t.isAlive()) // precheck that t is startable
throw new IllegalThreadStateException();
// workers 是一个 HashSet
workers.add(w);
// largestPoolSize 记录着线程池中出现过的最大线程数量
int s = workers.size();
if (s > largestPoolSize)
// 把历史上出现过的最大线程数的值更新一下
largestPoolSize = s;
// Worker 线程添加成功
workerAdded = true;
}
} finally {
// 释放 ReentrantLock 锁
mainLock.unlock();
}
if (workerAdded) {
t.start();
// Worker 线程已经启动
workerStarted = true;
}
}
} finally {
if (! workerStarted)
// Worker 线程没有启动成功
addWorkerFailed(w);
}
// 返回 Worker 线程是否启动成功
return workerStarted;
}
1.6.9线程池源码-核心源码解读-runWorker 方法
final void runWorker(Worker w) {
Thread wt = Thread.currentThread();
Runnable task = w.firstTask;
w.firstTask = null;
//允许响应中断
w.unlock(); // allow interrupts
// 线程退出的原因,true 是任务导致,false 是线程正常退出
boolean completedAbruptly = true;
try {
// 当前任务为空,且当前任务队列为空,停止循环
while (task != null || (task = getTask()) != null) {
// 上锁处理并发问题,防止在 shutdown()时终止正在运行的 worker
w.lock();
// 如果线程池是 stop 状态,并且线程没有被中断,就要确保线程被中断,如果线程池不是,确保线程池没有被中断;
// 清除当前线程的中断标志,做一个 recheck 来应对 shutdownNow 方法
if ((runStateAtLeast(ctl.get(), STOP) ||
(Thread.interrupted() &&
runStateAtLeast(ctl.get(), STOP))) &&
!wt.isInterrupted())
wt.interrupt();
try {
// 执行前(空方法,由子类重写实现)
beforeExecute(wt, task);
Throwable thrown = null;
try {
// 执行 Runnable 类的 run()方法
task.run();
} catch (RuntimeException x) {
thrown = x; throw x;
} catch (Error x) {
thrown = x; throw x;
} catch (Throwable x) {
thrown = x; throw new Error(x);
} finally {
// 执行后(空方法,由子类重写实现)
afterExecute(task, thrown);
}
} finally {
task = null;
// 完成的任务数+1
w.completedTasks++;
// 释放锁
w.unlock();
}
}
// 到此,线程是正常退出
completedAbruptly = false;
} finally {
// 处理 worker 的退出
processWorkerExit(w, completedAbruptly);
}
}
1.6.10线程池源码-核心源码解读-getTask 方法
private Runnable getTask() {
// 表示上一次从任务队列中取任务时是否超时
boolean timedOut = false; // Did the last poll() time out?
for (;;) {
int c = ctl.get();
int rs = runStateOf(c);
// Check if queue empty only if necessary.
/**
如果线程池为`SHUTDOWN`状态且任务队列为空(线程池 shutdown 状态可以处理任务队列中的任务,不再接受新任务)
或者
线程池状态>=STOP,则意味着线程池不必再获取任务了,
将当前工作线程数量-1 并返回 null;
*/
if (rs >= SHUTDOWN && (rs >= STOP || workQueue.isEmpty())) {
decrementWorkerCount();
return null;
}
int wc = workerCountOf(c);
/**
timed 变量用于判断是否需要进行超时控制;
allowCoreThreadTimeOut 默认是 false,也就是核心线程不允许进行超时;
wc > corePoolSize,表示当前线程池中的线程数量大于核心线程数量;
表示对于超过核心线程数量的这些线程,需要进行超时控制(默认情况)
*/
boolean timed = allowCoreThreadTimeOut || wc > corePoolSize;
/**
* 两个条件全部为 true,则通过 CAS 使工作线程数-1,即去除工作线程:
* 条件 1:工作线程数大于 maximumPoolSize,或(工作线程需要超时控制且上次在任务队列拉取任务超时)
* 条件 2:wc > 1 或任务队列为空
* 如果减 1 失败,则返回重试;
*/
if ((wc > maximumPoolSize || (timed && timedOut))
&& (wc > 1 || workQueue.isEmpty())) {
if (compareAndDecrementWorkerCount(c))
return null;
continue;
}
try {
/**
* 执行到这里,说明已经经过前面的校验,开始真正获取 task;
* 根据 timed 来判断,如果工作线程有超时时间,则通过任务队列的 poll 方法进行超时等待方式获取任务 ,
* 如果在 keepAliveTime 时间内没有获取到任务,则返回 null,否则通过 take 方法;
* take 方法表示如果这时任务队列为空,则会阻塞直到任务队列不为空;
* 一般 poll()用于普通线程、take()用于核心线程
*/
Runnable r = timed ?
workQueue.poll(keepAliveTime, TimeUnit.NANOSECONDS) :
workQueue.take();
if (r != null)
return r;
// 如果 r == null,说明已经超时得不到任务,timedOut 设置为 true
timedOut = true;
} catch (InterruptedException retry) {
// 如果获取任务时当前线程发生了中断,则设置 timedOut 为 false 并返回循环重试
timedOut = false;
}
}
}
1.6.11线程池源码-核心源码解读- 线程池复用
1、threadPoolExecutor.execute(runnable)
2、addWorker(command, boolean)
3、Worker w = new Worker(firstTask); //已经创建了 Thread
4、HashSet workers.add(w);
5、t.start(); //w.thread.start();
6、worker.run();
7、runWorker(this)
8、task = w.firstTask 或者 task = getTask()
9、task.run();
1.6.12线程池源码-核心源码解读- 线程池大小变化
private void processWorkerExit(Worker w, boolean completedAbruptly) {
//completedAbruptly 为 true 表示线程异常执行结束
//completedAbruptly 为 false 表示线程正常执行结束
if (completedAbruptly) // If abrupt, then workerCount wasn't adjusted
decrementWorkerCount();
//从线程 set 集合中移除工作线程,该过程需要加锁,因为 HashSet 是线程不安全的集合
final ReentrantLock mainLock = this.mainLock;
mainLock.lock();
try {
//统计完成的任务数:将该 worker 已完成的任务数追加到线程池已完成的任务数
completedTaskCount += w.completedTasks;
//从 HashSet<Worker>中移除该 worker
workers.remove(w);
} finally {
//释放锁
mainLock.unlock();
}
//根据线程池状态进行判断是否结束线程池
tryTerminate();
int c = ctl.get();
//当线程池是 RUNNING 或 SHUTDOWN 状态时
if (runStateLessThan(c, STOP)) {
//如果 worker 不是异常结束
if (!completedAbruptly) {
//如果 allowCoreThreadTimeOut=true,最小线程个数就可以变为 0;
int min = allowCoreThreadTimeOut ? 0 : corePoolSize;
//但是,如果等待队列有任务,至少保留一个 worker 来处理任务
if (min == 0 && ! workQueue.isEmpty())
min = 1;
//如果工作线程大于等于核心线程,直接 return 就行了,否则就需要添加一个线程;
if (workerCountOf(c) >= min)
return; // replacement not needed
}
//是异常执行结束的,添加一个线程去执行任务
addWorker(null, false);
}
}
try {
//统计完成的任务数:将该 worker 已完成的任务数追加到线程池已完成的任务数
completedTaskCount += w.completedTasks;
//从 HashSet<Worker>中移除该 worker
workers.remove(w);
} finally {
//释放锁
mainLock.unlock();
}
//根据线程池状态进行判断是否结束线程池
tryTerminate();
int c = ctl.get();
//当线程池是 RUNNING 或 SHUTDOWN 状态时
if (runStateLessThan(c, STOP)) {
//如果 worker 不是异常结束
if (!completedAbruptly) {
//如果 allowCoreThreadTimeOut=true,最小线程个数就可以变为 0;
int min = allowCoreThreadTimeOut ? 0 : corePoolSize;
//但是,如果等待队列有任务,至少保留一个 worker 来处理任务
if (min == 0 && ! workQueue.isEmpty())
min = 1;
//如果工作线程大于等于核心线程,直接 return 就行了,否则就需要添加一个线程;
if (workerCountOf(c) >= min)
return; // replacement not needed
}
//是异常执行结束的,添加一个线程去执行任务
addWorker(null, false);
}
}