原文摘自：https://www.jianshu.com/p/d48f854ead85
1.中斷后LockSupport.park()直接返回
public class MultInterruptParkDemo {
public static volatile boolean flag = true;
public static void main(String[] args) {
ThreadDemo04 t4 = new ThreadDemo04();
t4.start();
t4.interrupt();
flag = false;
}
public static class ThreadDemo04 extends Thread {
@Override
public void run() {
while (flag) {
}
LockSupport.park();
System.out.println("本打印出現在第一個park()之后");
LockSupport.park();
System.out.println("本打印出現在第二個park()之后");
}
}
}
結果：
本打印出現在第一個park()之后 本打印出現在第二個park()之后
1.1 分析
Thread中斷方法分析（包括HotSpot源碼）中interrupt源碼：
void os::interrupt(Thread* thread) {
assert(Thread::current() == thread || Threads_lock->owned_by_self(),
"possibility of dangling Thread pointer");
OSThread* osthread = thread->osthread();
if (!osthread->interrupted()) {
osthread->set_interrupted(true);
如果線程未設置中斷標志位，會進行設置。
LockSupport及HotSpot層Parker::park/unpark分析中park源碼：
void Parker::park(bool isAbsolute, jlong time) {
// Ideally we'd do something useful while spinning, such
// as calling unpackTime().
// Optional fast-path check:
// Return immediately if a permit is available.
// We depend on Atomic::xchg() having full barrier semantics
// since we are doing a lock-free update to _counter.
if (Atomic::xchg(0, &_counter) > 0) return;
Thread* thread = Thread::current();
assert(thread->is_Java_thread(), "Must be JavaThread");
JavaThread *jt = (JavaThread *)thread;
// Optional optimization -- avoid state transitions if there's an interrupt pending.
// Check interrupt before trying to wait
if (Thread::is_interrupted(thread, false)) {
return;
}
只要線程設置了中斷標志位，就直接返回，且這里的Thread::is_interrupted(thread, false)因為是false，所以不會清除中斷標志位。
1.2 總結
- t4.interrupt()會設置線程的中斷標志位
- LockSupport.park()會檢查線程是否設置了中斷標志位，如果設置了，則返回（這里並不會清除中斷標志位）
2.unpark不會累積許可（最多為1）
public class MultInterruptParkDemo2 {
public static volatile boolean flag = true;
public static void main(String[] args) {
ThreadDemo04 t4 = new ThreadDemo04();
t4.start();
LockSupport.unpark(t4);
LockSupport.unpark(t4);
LockSupport.unpark(t4);
flag = false;
}
public static class ThreadDemo04 extends Thread {
@Override
public void run() {
while (flag) {
}
LockSupport.park();
System.out.println("本打印出現在第一個park()之后");
LockSupport.park();
System.out.println("本打印出現在第二個park()之后");
}
}
}
結果：
本打印出現在第一個park()之后
2.1 分析
LockSupport及HotSpot層Parker::park/unpark分析中park源碼：
void Parker::park(bool isAbsolute, jlong time) {
// Ideally we'd do something useful while spinning, such
// as calling unpackTime().
// Optional fast-path check:
// Return immediately if a permit is available.
// We depend on Atomic::xchg() having full barrier semantics
// since we are doing a lock-free update to _counter.
if (Atomic::xchg(0, &_counter) > 0) return;
每次調用park都會將_counter直接置為0。
void Parker::unpark() {
int s, status ;
status = pthread_mutex_lock(_mutex);
assert (status == 0, "invariant") ;
s = _counter;
_counter = 1;
每次調用unpark都會將_counter直接置為1.
2.2 總結
- 因此，總的許可數總是保持在1，無論調用多少次unpark，都只會將_counter置為1。
- 每次park都會將_counter置為0，如果之前為1，則直接返回。后面的park調用就會阻塞。
3.Thread.interrupted()會清除中斷標志
public class MultInterruptParkDemo3 {
public static volatile boolean flag = true;
public static void main(String[] args) {
ThreadDemo04 t4 = new ThreadDemo04();
t4.start();
t4.interrupt();
flag = false;
}
public static class ThreadDemo04 extends Thread {
@Override
public void run() {
while (flag) {
}
LockSupport.park();
System.out.println("本打印出現在第一個park()之后");
System.out.println(Thread.interrupted());
System.out.println(Thread.interrupted());
LockSupport.park();
System.out.println("本打印出現在第二個park()之后");
}
}
}
結果：
本打印出現在第一個park()之后 true false
3.1 分析
Thread中斷方法分析（包括HotSpot源碼）中Thread.interrupted()源碼：
public static boolean interrupted() {
return currentThread().isInterrupted(true);
}
bool os::is_interrupted(Thread* thread, bool clear_interrupted) {
assert(Thread::current() == thread || Threads_lock->owned_by_self(),
"possibility of dangling Thread pointer");
OSThread* osthread = thread->osthread();
bool interrupted = osthread->interrupted();
if (interrupted && clear_interrupted) {
osthread->set_interrupted(false);
// consider thread->_SleepEvent->reset() ... optional optimization
}
return interrupted;
}
可以看到如果中斷了並且clear_interrupted為true，則會清除中斷標志位。
3.2 總結
Thread.interrupted()清除中斷標志位后，第二個LockSupport.park()就不會直接返回，而是正常地阻塞。
4.sleep清除中斷標志位但不會更改_counter
public class MultInterruptParkDemo4 {
public static volatile boolean flag = true;
public static void main(String[] args) {
ThreadDemo04 t4 = new ThreadDemo04();
t4.start();
t4.interrupt(); // 1. 調用unpark()方法，重置counter為1
flag = false;
}
public static class ThreadDemo04 extends Thread {
@Override
public void run() {
while (flag) {
}
try {
Thread.sleep(3000);
} catch (InterruptedException e) {
e.printStackTrace(); // 2. 重置interrupt為false，LockSupport.park()將不會因為interrupt為true而放行！！
}
System.out.println("本打印出現在第一個sleep()之后");
System.out.println(Thread.interrupted());
System.out.println(Thread.interrupted());
LockSupport.park(); // 3. counter為1，不發生阻塞；重置counter為0；
System.out.println("本打印出現在第二個park()之后"); LockSupport.park(); // 4. interrupt為false，並且counter為0；阻塞在這里！！
System.out.println("本打印出現在第三個park()之后");
}
}
}
結果
java.lang.InterruptedException: sleep interrupted
at java.lang.Thread.sleep(Native Method)
at wz.interrupt.MultInterruptParkDemo4$ThreadDemo04.run(MultInterruptParkDemo4.java:19)
本打印出現在第一個sleep()之后
false
false
本打印出現在第二個park()之后
4.1 分析
Thread中斷方法分析（包括HotSpot源碼）中interrupt源碼：
void os::interrupt(Thread* thread) {
assert(Thread::current() == thread || Threads_lock->owned_by_self(),
"possibility of dangling Thread pointer");
OSThread* osthread = thread->osthread();
if (!osthread->interrupted()) {
osthread->set_interrupted(true);
// More than one thread can get here with the same value of osthread,
// resulting in multiple notifications. We do, however, want the store
// to interrupted() to be visible to other threads before we execute unpark().
OrderAccess::fence();
ParkEvent * const slp = thread->_SleepEvent ;
if (slp != NULL) slp->unpark() ;
}
// For JSR166. Unpark even if interrupt status already was set
if (thread->is_Java_thread())
((JavaThread*)thread)->parker()->unpark();
ParkEvent * ev = thread->_ParkEvent ;
if (ev != NULL) ev->unpark() ;
}
會調用Parker::unpark()，那么_counter會設置為1.
參考阻塞線程的相關方法分析（包括HotSpot層源碼）中關於sleep源碼：
JVM_ENTRY(void, JVM_Sleep(JNIEnv* env, jclass threadClass, jlong millis))
JVMWrapper("JVM_Sleep");
if (millis < 0) {
THROW_MSG(vmSymbols::java_lang_IllegalArgumentException(), "timeout value is negative");
}
if (Thread::is_interrupted (THREAD, true) && !HAS_PENDING_EXCEPTION) {
THROW_MSG(vmSymbols::java_lang_InterruptedException(), "sleep interrupted");
}
如果線程已中斷，則清除中斷標記並拋出中斷異常，直接返回，但是並沒有更改_counter。
4.2 總結
- interrupt會調用((JavaThread*)thread)->parker()->unpark()，將_counter設置為1，也即后面調用park不會阻塞
- sleep如果檢測到中斷會直接清除中斷標志，並拋出異常。
- 因此兩個Thread.interrupted()都返回false，且LockSupport.park()不會阻塞。
5.對於LockSupport.park()阻塞總結
如下兩個條件任何一個成立，park()都不會阻塞：
- 中斷標志位存在（wait、join、sleep或Thread.interrupted()都會清除中斷標志位）
- _counter為1（之前調用了unpark或者interrupt）
參考
作者：王偵
鏈接：https://www.jianshu.com/p/d48f854ead85
來源：簡書
著作權歸作者所有。商業轉載請聯系作者獲得授權，非商業轉載請注明出處。