原文:https://www.goinggo.net/2014/02/the-nature-of-channels-in-go.html

The Nature Of Channels In Go

这篇文章关于channel讲解得非常好,深度形象。深入浅出。尤其是这两幅图,太厉害了。非常清楚,一目了然。

----------------------------------------------------------------------------------------------------------------------------------------

Introduction
In my last post called Concurrency, Goroutines and GOMAXPROCS, I set the stage for talking about channels. We discussed what concurrency was and how goroutines played a role. With that foundation in hand, we can now understand the nature of channels and how they can be used to synchronize goroutines to share resources in a safe, less error prone and fun way.

What Are Channels
Channels are type safe message queues that have the intelligence to control the behavior of any goroutine attempting to receive or send on it. A channel acts as a conduit between two goroutines and will synchronize the exchange of any resource that is passed through it. It is the channel’s ability to control the goroutines interaction that creates the synchronization mechanism. When a channel is created with no capacity, it is called an unbuffered channel. In turn, a channel created with capacity is called a buffered channel.

To understand what the synchronization behavior will be for any goroutine interacting with a channel, we need to know the type and state of the channel. The scenarios are a bit different depending on whether we are using an unbuffered or buffered channel, so let’s talk about each one independently.

Unbuffered Channels
Unbuffered channels have no capacity and therefore require both goroutines to be ready to make any exchange. When a goroutine attempts to send a resource to an unbuffered channel and there is no goroutine waiting to receive the resource, the channel will lock the sending goroutine and make it wait. When a goroutine attempts to receive from an unbuffered channel, and there is no goroutine waiting to send a resource, the channel will lock the receiving goroutine and make it wait.

 

In the diagram above, we see an example of two goroutines making an exchange using an unbuffered channel. In step 1, the two goroutines approach the channel and then in step 2, the goroutine on the left sticks his hand into the channel or performs a send. At this point, that goroutine is locked in the channel until the exchange is complete. Then in step 3, the goroutine on the right places his hand into the channel or performs a receive. That goroutine is also locked in the channel until the exchange is complete. In step 4 and 5 the exchange is made and finally in step 6, both goroutines are free to remove their hands and go on their way.

Synchronization is inherent in the interaction between the send and the receive. One can not happen without the other. The nature of an unbuffered channel is guaranteed synchronization.

Buffered Channels
Buffered channels have capacity and therefore can behave a bit differently. When a goroutine attempts to send a resource to a buffered channel and the channel is full, the channel will lock the goroutine and make it wait until a buffer becomes available. If there is room in the channel, the send can take place immediately and the goroutine can move on. When a goroutine attempts to receive from a buffered channel and the buffered channel is empty, the channel will lock the goroutine and make it wait until a resource has been sent.

In the diagram above, we see an example of two goroutines adding and removing items from a buffered channel independently. In step 1, the goroutine on the right is removing a resource from the channel or performing a receive. In step 2, the goroutine on the right can remove the resource independent of the goroutine on the left adding a new resource to the channel. In step 3, both goroutines are adding and removing a resource from the channel at the same time and in step 4 both goroutines are done.

Synchronization still occurs within the interactions of receives and sends, however when the queue has buffer availability, the sends will not lock. Receives will not lock when there is something to receive from the channel. Consequently, if the buffer is full or if there is nothing to receive, a buffered channel will behave very much like an unbuffered channel.

Relay Race
If you have ever watched a track meet you may have seen a relay race. In a relay race there are four athletes who run around the track as fast as they can as a team. The key to the race is that only one runner per team can be running at a time. The runner with the baton is the only one allowed to run, and the exchange of the baton from runner to runner is critical to winning the race.

Let’s build a sample program that uses four goroutines and a channel to simulate a relay race. The goroutines will be the runners in the race and the channel will be used to exchanged the baton between each runner. This is a classic example of how resources can be passed between goroutines and how a channel controls the behavior of the goroutines that interact with it.

package main

import (
    "fmt"
    "time"
)

func main() {
    // Create an unbuffered channel
    baton := make(chan int)

// First runner to his mark
    go Runner(baton)

// Start the race
    baton <- 1

// Give the runners time to race
    time.Sleep(500 * time.Millisecond)
}

func Runner(baton chan int) {
    var newRunner int

// Wait to receive the baton
    runner := <-baton

// Start running around the track
    fmt.Printf("Runner %d Running With Baton\n", runner)

// New runner to the line
    if runner != 4 {
        newRunner = runner + 1
        fmt.Printf("Runner %d To The Line\n", newRunner)
        go Runner(baton)
    }

// Running around the track
    time.Sleep(100 * time.Millisecond)

// Is the race over
    if runner == 4 {
        fmt.Printf("Runner %d Finished, Race Over\n", runner)
        return
    }

// Exchange the baton for the next runner
    fmt.Printf("Runner %d Exchange With Runner %d\n", runner, newRunner)
    baton <- newRunner
}

When we run the sample program we get the following output:

Runner 1 Running With Baton
Runner 2 To The Line
Runner 1 Exchange With Runner 2
Runner 2 Running With Baton
Runner 3 To The Line
Runner 2 Exchange With Runner 3
Runner 3 Running With Baton
Runner 4 To The Line
Runner 3 Exchange With Runner 4
Runner 4 Running With Baton
Runner 4 Finished, Race Over

The program starts out creating an unbuffered channel:

// Create an unbuffered channel
baton := make(chan int)

Using an unbuffered channel forces the goroutines to be ready at the same time to make the exchange of the baton. This need for both goroutines to be ready creates the guaranteed synchronization.

If we look at the rest of the main function, we see a goroutine created for the first runner in the race and then the baton is handed off to that runner. The baton in this example is an integer value that is being passed between each runner. The sample is using a sleep to let the race complete before main terminates and ends the program:

// Create an unbuffered channel
baton := make(chan int)

// First runner to his mark
go Runner(baton)

// Start the race
baton <- 1

// Give the runners time to race
time.Sleep(500 * time.Millisecond)

If we just focus on the core parts of the Runner function, we can see how the baton exchange takes place until the race is over. The Runner function is launched as a goroutine for each runner in the race. Every time a new goroutine is launched, the channel is passed into the goroutine. The channel is the conduit for the exchange, so the current runner and the one waiting to go next need to reference the channel:

func Runner(baton chan int)

The first thing each runner does is wait for the baton exchange. That is simulated with the receive on the channel. The receive immediately locks the goroutine until the baton is sent into the channel. Once the baton is send into the channel, the receive will release and the goroutine will simulate the next runner sprinting down the track. If the fourth runner is running, no new runner will enter the race. If we are still in the middle of the race, a new goroutine for the next runner is launched.

// Wait to receive the baton
runner := <-baton

// New runner to the line
if runner != 4 {
    newRunner = runner + 1
    go Runner(baton)
}

Then we sleep to simulate some time it takes for the runner to run around the track. If this is the fourth runner, the goroutine terminates after the sleep and the race is complete. If not, the baton exchange takes place with the send into the channel. There is a goroutine already locked and waiting for this exchange. As soon as the baton is sent into the channel, the exchange is made and the race continue:

// Running around the track
time.Sleep(100 * time.Millisecond)

// Is the race over
if runner == 4 {
    return
}

// Exchange the baton for the next runner
baton <- newRunner

Conclusion
The example showcases a real world event, a relay race between runners, being implemented in a way that mimics the actual events. This is one of the beautiful things about channels. The code flows in a way that simulates how these types of exchanges can happen in the real world.

Now that we have an understanding of the nature of unbuffered and buffered channels, we can look at different concurrency patterns we can implement using channels. Concurrency patterns allow us to implement more complex exchanges between goroutines that simulate real world computing problems like semaphores, generators and multiplexers.

golang中channels的本质详解,经典!的更多相关文章

  1. golang中的反射reflect详解

    先重复一遍反射三定律: 1.反射可以将"接口类型变量"转换为"反射类型对象". 2.反射可以将"反射类型对象"转换为"接口类型变量 ...

  2. Scala 深入浅出实战经典 第57讲:Scala中Dependency Injection实战详解

    王家林亲授<DT大数据梦工厂>大数据实战视频 Scala 深入浅出实战经典(1-87讲)完整视频.PPT.代码下载:百度云盘:http://pan.baidu.com/s/1c0noOt6 ...

  3. Scala 深入浅出实战经典 第55讲:Scala中Infix Type实战详解

    王家林亲授<DT大数据梦工厂>大数据实战视频 Scala 深入浅出实战经典(1-64讲)完整视频.PPT.代码下载: 百度云盘:http://pan.baidu.com/s/1c0noOt ...

  4. Scala 深入浅出实战经典 第54讲:Scala中复合类型实战详解

    王家林亲授<DT大数据梦工厂>大数据实战视频 Scala 深入浅出实战经典(1-64讲)完整视频.PPT.代码下载:百度云盘:http://pan.baidu.com/s/1c0noOt6 ...

  5. Scala 深入浅出实战经典 第53讲:Scala中结构类型实战详解

    王家林亲授<DT大数据梦工厂>大数据实战视频 Scala 深入浅出实战经典(1-64讲)完整视频.PPT.代码下载:百度云盘:http://pan.baidu.com/s/1c0noOt6 ...

  6. Oracle Statspack报告中各项指标含义详解~~学习性能必看!!!

    Oracle Statspack报告中各项指标含义详解~~学习性能必看!!! Data Buffer Hit Ratio#<#90# 数据块在数据缓冲区中的命中率,通常应该在90%以上,否则考虑 ...

  7. 巨人大哥谈Web应用中的Session(session详解)

    巨人大哥谈Web应用中的Session(session详解) 虽然session机制在web应用程序中被采用已经很长时间了,但是仍然有很多人不清楚session机制的本质,以至不能正确的应用这一技术. ...

  8. Java I/O : Java中的进制详解

    作者:李强强 上一篇,泥瓦匠基础地讲了下Java I/O : Bit Operation 位运算.这一讲,泥瓦匠带你走进Java中的进制详解. 一.引子 在Java世界里,99%的工作都是处理这高层. ...

  9. Swift 中的Closures(闭包)详解

    Swift 中的Closures(闭包)详解 在Swift没有发布之前,所有人使用OC语言编写Cocoa上的程序,而其中经常被人们讨论的其中之一 -- Block 一直备受大家的喜爱.在Swift中, ...

随机推荐

  1. webview页面间的通信问题

    前提 记一次多页面开发. 开发需求时会对页面刷新(reload),返回到上一页(用户返回 / history.go()) 页面间的通信 sessionStorage保存本次会话的信息,同步到新页面或上 ...

  2. PendingIntent 显示通知

    安卓显示通知 PendingIntent pendingIntent=PendingIntent.getActivity(Media.this,0, new Intent(Media.this,Med ...

  3. jdbc分页查询

    虽然现在db层的框架很多,用起来也非常的方便,像分页这种非常常用的功能也基本上都有对应的接口可以直接使用.但是有时候数据源不在配置的范围的时候,就必须要使用到jdbc来执行sql,jdbc执行的是原生 ...

  4. JDBC性能优化

    一.使用PreparedStatement的Batch功能 参见本人一下文章:http://blog.csdn.net/lmb55/article/details/50631062 二.选择合适的光标 ...

  5. 【Python-2.7】换行符和制表符

    在Python中换行符“\n”表示接下来的内容将会换到下一行显示,制表符“\t”表示下面的内容显示时在前面留出空白,如打印如下内容: Dear: I love you forever! 上面的一段话分 ...

  6. Linux系统调用--getrusage函数详解

    Linux系统调用--getrusage函数详解 功能描述:     获得进程的相关资源信息.如:用户开销时间,系统开销时间,接收的信号量等等;   用法:    #include <sys/t ...

  7. Windows提高_2.2第二部分:用户区同步

    第二部分:用户区同步 同步和互斥 同步:就是按照一定的顺序执行不同的线程 互斥:当一个线程访问某一资源的时候,其它线程不能同时访问 多线程产生的问题 #include <stdio.h> ...

  8. 【Linux】CentOS安装Jenkins

    sudo wget -O /etc/yum.repos.d/jenkins.repo http://pkg.jenkins-ci.org/redhat/jenkins.repo sudo rpm -- ...

  9. PHP数据乱码

    本文主要总结下PHP数据乱码的解决方案 要点:多个不同文件系统里一定要统一编码 [注意] (1)HTML编码与MySQL编码一致: (2)PHP编码与MySQL编码一致: (3)header头发送字符 ...

  10. 16Oracle Database 系统权限和对象权限

    Oracle Database 系统权限和对象权限 Oracle中的系统权限和对象权限 DCL 数据控制语言 -- 查看对象的权限 grant / revoke 查看登录用户 Show user 查看 ...