Java并发基础实践--死锁
本文是
Java并发基础实践系列中的一篇，介绍了最简单的死锁场景，并使用jstack产生的thread dump来查找死锁。(2013.12.29最后更新)
1. 死锁为了能够维护线程的安全性，Java提供的锁机制，但不恰当地使用锁则可能产生死锁。死锁是并发编程中一个无法绕开的问题。只要在一个任务中使用了一个以上的锁，那么就存在死锁的风险。
死锁产生的直接原因非常简单，即两个线程在相互等待对方所执有的锁。
2. 锁顺序死锁在死锁场景中，最典型的就是锁顺序死锁，代码清单1就是一个很常见的示例。
清单1
public class DeadLock {
private Object leftLock = new Object();
private Object rightLock = new Object();
public void leftRight() {
synchronized (leftLock) {
try {
TimeUnit.SECONDS.sleep(3);
} catch (InterruptedException e) {
e.printStackTrace();
}
synchronized (rightLock) {
System.out.println("leftRight");
}
}
}
public void rightLeft() {
synchronized (rightLock) {
try {
TimeUnit.SECONDS.sleep(3);
} catch (InterruptedException e) {
e.printStackTrace();
}
synchronized (leftLock) {
System.out.println("leftRight");
}
}
}
public static void main(String[] args) {
final DeadLock deadLock = new DeadLock();
Thread t1 = new Thread(new Runnable() {
@Override
public void run() {
deadLock.leftRight();
}
});
Thread t2 = new Thread(new Runnable() {
@Override
public void run() {
deadLock.rightLeft();
}
});
t1.start();
t2.start();
}
}
3. Thread DumpJDK提供了一组命令行工具，其中就包括jstack。通过jstack可以获取当前正运行的Java进程的java stack和native stack信息。如果Java进程崩溃了，也可以通过它来获取core file中的java stack和native stack信息，以方便我们定位问题。
为了能够使用jstack去输出目标Java进程的thread dump，首先必须要弄清楚在执行清单1的程序时，该程序的进程号。JDK提供的另一个命令行工具jps可以获取系统中所有Java进程的相关信息。
在命令行窗口中执行命令
jps，即可以得到清单2所示的结果
清单2
C:\Documents and Settings\Administrator>jps
2848
4552 DeadLock
5256 Jps
其中
4552就是在笔者机器上执行程序DeadLock时所生成Java进程的进程号。
然后再执行命令
jstack 4552，在笔者的机器上就会得到清单3所示的结果
清单3
C:\Documents and Settings\Administrator>jstack 4552
2013-12-29 18:45:41
Full thread dump Java HotSpot(TM) Client VM (23.25-b01 mixed mode, sharing):
"DestroyJavaVM" prio=6 tid=0x00878800 nid=0xd00 waiting on condition [0x00000000]
java.lang.Thread.State: RUNNABLE
"Thread-1" prio=6 tid=0x02b56c00 nid=0x14ec waiting for monitor entry [0x02fdf000]
java.lang.Thread.State: BLOCKED (on object monitor)
at concurrency.deadlock.DeadLock.rightLeft(DeadLock.java:33)
- waiting to lock <0x22be6598> (a java.lang.Object)
- locked <0x22be65a0> (a java.lang.Object)
at concurrency.deadlock.DeadLock$2.run(DeadLock.java:53)
at java.lang.Thread.run(Thread.java:724)
"Thread-0" prio=6 tid=0x02b55c00 nid=0x354 waiting for monitor entry [0x02f8f000]
java.lang.Thread.State: BLOCKED (on object monitor)
at concurrency.deadlock.DeadLock.leftRight(DeadLock.java:19)
- waiting to lock <0x22be65a0> (a java.lang.Object)
- locked <0x22be6598> (a java.lang.Object)
at concurrency.deadlock.DeadLock$1.run(DeadLock.java:45)
at java.lang.Thread.run(Thread.java:724)
"Service Thread" daemon prio=6 tid=0x02b34800 nid=0x133c runnable [0x00000000]
java.lang.Thread.State: RUNNABLE
"C1 CompilerThread0" daemon prio=10 tid=0x02b13800 nid=0x10fc waiting on condition [0x00000000]
java.lang.Thread.State: RUNNABLE
"Attach Listener" daemon prio=10 tid=0x02b11c00 nid=0x1424 waiting on condition [0x00000000]
java.lang.Thread.State: RUNNABLE
"Signal Dispatcher" daemon prio=10 tid=0x02b10800 nid=0x1100 runnable [0x00000000]
java.lang.Thread.State: RUNNABLE
"Finalizer" daemon prio=8 tid=0x02af4c00 nid=0x1238 in Object.wait() [0x02daf000]
java.lang.Thread.State: WAITING (on object monitor)
at java.lang.Object.wait(Native Method)
- waiting on <0x22b60fb8> (a java.lang.ref.ReferenceQueue$Lock)
at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:135)
- locked <0x22b60fb8> (a java.lang.ref.ReferenceQueue$Lock)
at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:151)
at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:189)
"Reference Handler" daemon prio=10 tid=0x02af0000 nid=0x12e8 in Object.wait() [0x02d5f000]
java.lang.Thread.State: WAITING (on object monitor)
at java.lang.Object.wait(Native Method)
- waiting on <0x22b60da0> (a java.lang.ref.Reference$Lock)
at java.lang.Object.wait(Object.java:503)
at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:133)
- locked <0x22b60da0> (a java.lang.ref.Reference$Lock)
"VM Thread" prio=10 tid=0x02aee400 nid=0x129c runnable
"VM Periodic Task Thread" prio=10 tid=0x02b48000 nid=0x89c waiting on condition
JNI global references: 117
Found one Java-level deadlock:
=============================
"Thread-1":
waiting to lock monitor 0x02af4a3c (object 0x22be6598, a java.lang.Object),
which is held by "Thread-0"
"Thread-0":
waiting to lock monitor 0x02af310c (object 0x22be65a0, a java.lang.Object),
which is held by "Thread-1"
Java stack information for the threads listed above:
===================================================
"Thread-1":
at concurrency.deadlock.DeadLock.rightLeft(DeadLock.java:33)
- waiting to lock <0x22be6598> (a java.lang.Object)
- locked <0x22be65a0> (a java.lang.Object)
at concurrency.deadlock.DeadLock$2.run(DeadLock.java:53)
at java.lang.Thread.run(Thread.java:724)
"Thread-0":
at concurrency.deadlock.DeadLock.leftRight(DeadLock.java:19)
- waiting to lock <0x22be65a0> (a java.lang.Object)
- locked <0x22be6598> (a java.lang.Object)
at concurrency.deadlock.DeadLock$1.run(DeadLock.java:45)
at java.lang.Thread.run(Thread.java:724)
Found 1 deadlock.
在上述输出中，我们可以很明确地看到一个死锁
"Thread-1":
waiting to lock monitor 0x02af4a3c (object 0x22be6598, a java.lang.Object),
which is held by "Thread-0"
"Thread-0":
waiting to lock monitor 0x02af310c (object 0x22be65a0, a java.lang.Object),
which is held by "Thread-1"
并且它还标明了程序是在哪个地方时发现了上述死锁
"Thread-1":
at concurrency.deadlock.DeadLock.rightLeft(DeadLock.java:33)
- waiting to lock <0x22be6598> (a java.lang.Object)
- locked <0x22be65a0> (a java.lang.Object)
at concurrency.deadlock.DeadLock$2.run(DeadLock.java:53)
at java.lang.Thread.run(Thread.java:724)
"Thread-0":
at concurrency.deadlock.DeadLock.leftRight(DeadLock.java:19)
- waiting to lock <0x22be65a0> (a java.lang.Object)
- locked <0x22be6598> (a java.lang.Object)
at concurrency.deadlock.DeadLock$1.run(DeadLock.java:45)
at java.lang.Thread.run(Thread.java:724)
4. 小结死锁产生的直接原因非常简单，即两个线程在相互等待对方所执有的锁。锁顺序死锁是其中最经典的场景，此外还有动态的锁顺序死锁。虽然表现形式有所不同，但本质上都是两个线程在以不同的顺序来获取相同锁时，发生了死锁问题。
使用thread dump可以帮助我们分析死锁产生的原因。除了直接使用jstack命令来获取thread dump输出以外，JDK还提供了jvisualvm工具，它能以可视化的方式展示Java程序的进程号并导出thread dump。