本文主要是介绍Golang sync.WaitGroup源码详细分析,希望对大家解决编程问题提供一定的参考价值,需要的开发者们随着小编来一起学习吧!
一、介绍
WaitGroup是多个goroutine之间协作的一种实现方式,主要功能就是阻塞等待一组goroutine执行完成。
常用的使用场景:主goroutine调用Add函数设置需要等待的goroutine的数量,当每个goroutine执行完成后调用Done函数(将counter减1),Wait函数用于阻塞等待直到该组中的所有goroutine都执行完成。
源码中主要设计到的三个概念:counter、waiter和semaphore
counter: 当前还未执行结束的goroutine计数器
waiter : 等待goroutine-group结束的goroutine数量,即有多少个等候者
semaphore: 信号量
信号量是Unix系统提供的一种保护共享资源的机制,用于防止多个线程同时访问某个资源。
可简单理解为信号量为一个数值:
当信号量>0时,表示资源可用,获取信号量时系统自动将信号量减1;
当信号量=0时,表示资源暂不可用,获取信号量时,当前线程会进入睡眠,当信号量为正时被唤醒。
二、源码分析
Golang源码版本 :1.10.3
1.结构体
type WaitGroup struct {noCopy noCopy //该WaitGroup对象不允许拷贝使用,只能用指针传递// 64-bit value: high 32 bits are counter, low 32 bits are waiter count.// 64-bit atomic operations require 64-bit alignment, but 32-bit// compilers do not ensure it. So we allocate 12 bytes and then use// the aligned 8 bytes in them as state.//用于存储计数器(counter)和waiter的值// 只需要64位,即8个字节,其中高32位是counter值,低32位值是waiter值// 不直接使用uint64,是因为uint64的原子操作需要64位系统,而32位系统下,可能会出现崩溃// 所以这里用byte数组来实现,32位系统下4字节对齐,64位系统下8字节对齐,所以申请12个字节,其中必定有8个字节是符合8字节对齐的,下面的state()函数中有进行判断state1 [12]bytesema uint32 //信号量
}
从结构体中我们看到
state1是一个12位长度的byte数组,用于存储counter和waiter的值
sema就是传说中的信号量
2.state函数
state是一个内部函数,用于获取counter和 waiter的值
//获取counter 、 waiter的值 (counter是uint64的高32位,waiter是uint64的低32位)
func (wg *WaitGroup) state() *uint64 {// 根据state1的起始地址分析,若是8字节对齐的,则直接用前8个字节作为*uint64类型// 若不是,说明是4字节对齐,则后移4个字节后,这样必为8字节对齐,然后取后面8个字节作为*uint64类型if uintptr(unsafe.Pointer(&wg.state1))%8 == 0 {return (*uint64)(unsafe.Pointer(&wg.state1))} else {return (*uint64)(unsafe.Pointer(&wg.state1[4]))}
}
3.Add方法
//用于增加或减少计数器(counter)的值
//如果计数器为0,则释放调用Wait方法时的阻塞,如果计数器为负,则panic
//Add()方法应该在Wait()方法调用之前
func (wg *WaitGroup) Add(delta int) {//获取当前counter和 waiter的值statep := wg.state()if race.Enabled {_ = *statep // trigger nil deref earlyif delta < 0 {// Synchronize decrements with Wait.race.ReleaseMerge(unsafe.Pointer(wg))}race.Disable()defer race.Enable()}//将delta的值添加到counter上state := atomic.AddUint64(statep, uint64(delta)<<32)v := int32(state >> 32) //counter值w := uint32(state) //waiter值if race.Enabled && delta > 0 && v == int32(delta) {// The first increment must be synchronized with Wait.// Need to model this as a read, because there can be// several concurrent wg.counter transitions from 0.race.Read(unsafe.Pointer(&wg.sema))}//counter为负数,则触发panicif v < 0 {panic("sync: negative WaitGroup counter")}// waiter值不为0,累加后的counter值和delta相等,说明Wait()方法没有在Add()方法之后调用,触发panic,因为正确的做法是先Add()后Wait()if w != 0 && delta > 0 && v == int32(delta) {panic("sync: WaitGroup misuse: Add called concurrently with Wait")}//Add()添加正常返回//1.counter > 0,说明还不需要释放信号量,可以直接返回//2. waiter = 0 ,说明没有等待的goroutine,也不需要释放信号量,可以直接返回if v > 0 || w == 0 {return}// This goroutine has set counter to 0 when waiters > 0.// Now there can't be concurrent mutations of state:// - Adds must not happen concurrently with Wait,// - Wait does not increment waiters if it sees counter == 0.// Still do a cheap sanity check to detect WaitGroup misuse.//下面是 counter == 0 并且 waiter > 0的情况//现在若原state和新的state不等,则有以下两种可能//1. Add 和 Wait方法同时调用//2. counter已经为0,但waiter值有增加,这种情况永远不会触发信号量了// 以上两种情况都是错误的,所以触发异常//注:state := atomic.AddUint64(statep, uint64(delta)<<32) 这一步调用之后,state和*statep的值应该是相等的,除非有以上两种情况发生if *statep != state {panic("sync: WaitGroup misuse: Add called concurrently with Wait")}// Reset waiters count to 0.//将waiter 和 counter都置为0*statep = 0//原子递减信号量,并通知等待的goroutinefor ; w != 0; w-- {runtime_Semrelease(&wg.sema, false)}
}
4.Done方法
// Done decrements the WaitGroup counter by one.
//将计数器(counter)的值减1
func (wg *WaitGroup) Done() {wg.Add(-1)
}
5.Wait方法
// Wait blocks until the WaitGroup counter is zero.
//调用Wait方法会阻塞当前调用的goroutine直到 counter的值为0
//也会增加waiter的值
func (wg *WaitGroup) Wait() {//获取当前counter和 waiter的值statep := wg.state()if race.Enabled {_ = *statep // trigger nil deref earlyrace.Disable()}//一直等待,直到无需等待或信号量触发,才返回for {state := atomic.LoadUint64(statep)v := int32(state >> 32) //counter值w := uint32(state) //waiter值//如果counter值为0,则说明所有goroutine都退出了,无需等待,直接退出if v == 0 {// Counter is 0, no need to wait.if race.Enabled {race.Enable()race.Acquire(unsafe.Pointer(wg))}return}// Increment waiters count.//原子增加waiter的值,CAS方法,外面for循环会一直尝试,保证多个goroutine同时调用Wait()也能正常累加waiterif atomic.CompareAndSwapUint64(statep, state, state+1) {if race.Enabled && w == 0 {// Wait must be synchronized with the first Add.// Need to model this is as a write to race with the read in Add.// As a consequence, can do the write only for the first waiter,// otherwise concurrent Waits will race with each other.race.Write(unsafe.Pointer(&wg.sema))}//一直等待信号量sema,直到信号量触发,runtime_Semacquire(&wg.sema)//从上面的Add()方法看到,触发信号量之前会将seatep置为0(即counter和waiter都置为0),所以此时应该也为0//如果不为0,说明WaitGroup此时又执行了Add()或者Wait()操作,所以会触发panicif *statep != 0 {panic("sync: WaitGroup is reused before previous Wait has returned")}if race.Enabled {race.Enable()race.Acquire(unsafe.Pointer(wg))}return}}
}
三、注意点
1.Add()必须在Wait()前调用
2.Add()设置的值必须与实际等待的goroutine个数一致,如果设置的值大于实际的goroutine数量,可能会一直阻塞。如果小于会触发panic
3. WaitGroup不可拷贝,可以通过指针传递,否则很容易造成BUG
以下为值拷贝引起的Bug示例
demo1:因为值拷贝引起的死锁
func main() {var wg sync.WaitGroupwg.Add(5)for i := 0 ; i < 5 ; i++ {test(wg)}wg.Wait()
}func test(wg sync.WaitGroup) {go func() {fmt.Println("hello")wg.Done()}()
}
demo2:因为值拷贝引起的不会阻塞等待现象
func main() {var wg sync.WaitGroupfor i := 0 ; i < 5 ; i++ {test(wg)}wg.Wait()
}func test(wg sync.WaitGroup) {go func() {wg.Add(1)fmt.Println("hello")time.Sleep(time.Second*5)wg.Done()}()
}
demo3:因为值拷贝引发的panic
type person struct {wg sync.WaitGroup
}func (t *person) say() {go func() {fmt.Println("say Hello!")time.Sleep(time.Second*5)t.wg.Done()}()
}func main() {var wg sync.WaitGroupt := person{wg:wg}wg.Add(5)for i := 0 ; i< 5 ;i++ {t.say()}wg.Wait()
}
感谢:https://blog.csdn.net/yzf279533105/article/details/97302666
这篇关于Golang sync.WaitGroup源码详细分析的文章就介绍到这儿,希望我们推荐的文章对编程师们有所帮助!