Channel实现原理分析

文章目录

  • 什么是channel
  • channel的实现
  • 问题
  • 参考文献

什么是channel

我们来看《Go语言编程》中的一段话

channel是Go语言在语言级别提供的goroutine间的通信方式,是一种进程内的通信方式。

通俗点儿解释就是channel可以在两个或者多个goroutine之间传递消息。在Go中,goroutine和channel是并发编程的两大基石,goroutine用来执行并发任务,channel用来在goroutine之间来传递消息。

Do not communicate by sharing memory; instead, share memory by communicating.
不要通过共享内存来通信,而要通过通信来实现共享内存。

推荐一本书《Concurrency in Go》,这本书对golang中的并发做了深入的讲解。

channel的实现

1、引入概念

首先我们来看两个例子来简单看下golang中channel是如何使用的:


package main

import (
  "fmt"
)

func goroutineA(a <-chan int) {
  for {
    select {
    case val := <-a:
      fmt.Println(val)
    }
  }
}

func main() {
  ch := make(chan int)
  go goroutineA(ch)
  ch <- 3
  ch <- 5
}

很简单的一段程序,初始化了一个非缓冲的channel,然后并发一个协程去接收channel中的数据,然后往channel中连续发送两个值,首先大家先理解一组概念,什么是非缓冲型channel和缓冲型channel?对,其实很简单,make时如果channel空间不为0,就是缓冲型的channel。

ch := make(chan int)//非缓冲型
ch := make(chan int1024)//缓冲型

如果我们将go goroutineA(ch)这行代码往下移,会发生什么?对,会报

fatal error: all goroutines are asleep - deadlock!

因为channel没有缓冲,也没有正在等待接收的goroutine,这个概念接下来我会讲到。

另外,我们会看到goroutineA的入参是a <-chan int,代表一个只能用于接收的channel,也就是单向channel。

var a <-chan int//单向channel,只用于接收,也就是从channel读取数据
var a chan<- int//单向channel,只用于发送,也就是往channel写入数据

大家一定要清楚接收和发送的概念

接收代表从channel读取数据
发送代表往channel写入数据

再看一个复杂点儿的例子

package main

import (
  "fmt"
  "os"
  "os/signal"
  "syscall"
  "time"
)

var exit = make(chan string, 1)

func main() {
  go dealSignal()
  exited := make(chan struct{}, 1)
  go channel1(exited)
  count := 0
  t := time.Tick(time.Second)
Loop:
  for {
    select {
    case <-t:
      count++
      fmt.Printf("main run %d\n", count)
    case <-exited:
      fmt.Println("main exit begin")
      break Loop
    }
  }
  fmt.Println("main exit end")
}

func dealSignal() {
  c := make(chan os.Signal, 1)
  signal.Notify(c, os.Interrupt, syscall.SIGTERM)
  go func() {
    <-c
    exit <- "shutdown"
  }()
}

func channel1(exited chan<- struct{}) {
  t := time.Tick(time.Second)
  count := 0
  for {
    select {
    case <-t:
      count++
      fmt.Printf("channel1 run %d\n", count)
    case <-exit:
      fmt.Println("channel1 exit")
      close(exited)
      return
    }
  }
}

这个例子首先并发出一个dealsign方法,用来接收关闭信号,如果接收到关闭信号后往exit channel发送一条消息,然后并发运行channel1,channel1中定了一个ticker,正常情况下channel1每秒打印第一个case语句,如果接收到exit的信号,进入第二个case,然后关闭传入的exited channel,那么main中的Loop,接收到exited关闭的信号后,打印“main exit begin”, 然后退出循环,进程成功退出。这个例子演示了channel在goroutine中起到的传递消息的作用。这个例子是为了向大家展示channel在多个goroutine之间进行通信。

2、数据结构

channel为什么会天生具备这种传递消息的特性呢,我们不禁对其底层的数据结构产生兴趣,我们来看下runtime/chan.go文件,有关channel的一切底层操作都在这个文件,我们首先来看下数据结构:

type hchan struct {
  qcount   uint           // total data in the queue;chan中的元素总数
  dataqsiz uint           // size of the circular queue;底层循环数组的size
  buf      unsafe.Pointer // points to an array of dataqsiz elements,指向底层循环数组的指针,只针对有缓冲的channel
  elemsize uint16  //chan中元素的大小
  closed   uint32  //chan是否关闭
  elemtype *_type // element type;元素类型
  sendx    uint   // send index;已发送元素在循环数组中的索引
  recvx    uint   // receive index;已接收元素在循环数组中的索引
  recvq    waitq  // list of recv waiters,等待接收消息的goroutine队列
  sendq    waitq  // list of send waiters,等待发送消息的goroutine队列

  // lock protects all fields in hchan, as well as several
  // fields in sudogs blocked on this channel.
  //
  // Do not change another G's status while holding this lock
  // (in particular, do not ready a G), as this can deadlock
  // with stack shrinking.
  lock mutex
}

type waitq struct {
  first *sudog
  last  *sudog
}

创建一个底层数组容量为5,元素类型为int,那么channel的数据结构如下图所示:

Channel实现原理分析_第1张图片
3、创建

首先我们先来了解一下 Channel 在 Go 语言中是如何创建的,Go 语言 Channel 的创建都是由 make 关键字完成的,我们在前面介绍slice和map的创建时都介绍了使用 make 关键字初始化数据结构,那么一个问题,那么Go语言是如何实现通过make方式来创建不同的数据结构的呢?

Golang 中所有形如 make(chan int, 10) 在编译期间会先被转换成 OMAKE 类型的节点,随后的类型检查阶段在发现 make 的第一个参数是 Channel 类型时会将 OMAKE 类型的节点转换成 OMAKECHAN:


func typecheck1(n *Node, top int) (res *Node) {
    switch n.Op {
    case OMAKE:
        // ...
        switch t.Etype {
        case TCHAN:
            l = nil
            if i < len(args) {
                l = args[i]
                i++
                l = typecheck(l, ctxExpr)
                l = defaultlit(l, types.Types[TINT])
                if l.Type == nil {
                    n.Type = nil
                    return n
                }
                if !checkmake(t, "buffer", l) {
                    n.Type = nil
                    return n
                }
                n.Left = l
            } else {
                n.Left = nodintconst(0)
            }
            n.Op = OMAKECHAN
        }
    }

OMAKECHAN 类型的节点最终都会在SSA中间代码生成阶段之前被转换成makechan 或者 makechan64 的函数调用:


func walkexpr(n *Node, init *Nodes) *Node {
    switch n.Op {
    case OMAKECHAN:
        size := n.Left
        fnname := "makechan64"
        argtype := types.Types[TINT64]

        if size.Type.IsKind(TIDEAL) || maxintval[size.Type.Etype].Cmp(maxintval[TUINT]) <= 0 {
            fnname = "makechan"
            argtype = types.Types[TINT]
        }

        n = mkcall1(chanfn(fnname, 1, n.Type), n.Type, init, typename(n.Type), conv(size, argtype))
    }
}

创建channel的时候,其实底层是调用makechan方法,我们来看下源码:


func makechan(t *chantype, size int) *hchan {
  elem := t.elem

  // compiler checks this but be safe.
  if elem.size >= 1<<16 {
    throw("makechan: invalid channel element type")
  }
  if hchanSize%maxAlign != 0 || elem.align > maxAlign {
    throw("makechan: bad alignment")
  }

  mem, overflow := math.MulUintptr(elem.size, uintptr(size))
  if overflow || mem > maxAlloc-hchanSize || size < 0 {
    panic(plainError("makechan: size out of range"))
  }

  // Hchan does not contain pointers interesting for GC when elements stored in buf do not contain pointers.
  // buf points into the same allocation, elemtype is persistent.
  // SudoG's are referenced from their owning thread so they can't be collected.
  // TODO(dvyukov,rlh): Rethink when collector can move allocated objects.
  var c *hchan
  switch {
  case mem == 0:
    // Queue or element size is zero.
    c = (*hchan)(mallocgc(hchanSize, nil, true))
    // Race detector uses this location for synchronization.
    c.buf = c.raceaddr()
  case elem.ptrdata == 0:
    // Elements do not contain pointers.
    // Allocate hchan and buf in one call.
    c = (*hchan)(mallocgc(hchanSize+mem, nil, true))
    c.buf = add(unsafe.Pointer(c), hchanSize)
  default:
    // Elements contain pointers.
    c = new(hchan)
    c.buf = mallocgc(mem, elem, true)
  }

  c.elemsize = uint16(elem.size)
  c.elemtype = elem
  c.dataqsiz = uint(size)

  if debugChan {
    print("makechan: chan=", c, "; elemsize=", elem.size, "; elemalg=", elem.alg, "; dataqsiz=", size, "\n")
  }
  return c
}

可以看出makechan中其实主要的代码就是一个switch,针对不同的情况

1、case mem == 0代表无缓冲型channel,只分配hchan本身结构体大小的内存
2、case elem.ptrdata==0 代表元素类型不含指针,只分配hchan本身结构体大小+元素大小*个数的内存,是连续的内存空间
3、default元素类型包括指针,两次分配内存的操作

然后将buf指向对应的地址,然后是hchan中其他变量的赋值。

4、接收

接下来我们来讲channel的接收和发送,我们使用一段程序来进行讲解


func goroutineA(a <-chan int) {
    val := <- a
    fmt.Println("G1 received data: ", val)
    return
}

func goroutineB(b <-chan int) {
    val := <- b
    fmt.Println("G2 received data: ", val)
    return
}

func main() {
    ch := make(chan int)
    go goroutineA(ch)
    go goroutineB(ch)
    ch <- 3
    time.Sleep(time.Second)
}

首先创建了一个无缓冲型的channel,然后启动两个goroutine去消费channel的数据,紧接着向channel中发送数据。我们一步一步来分析channel是如何接收和发送数据的,首先来看接收,golang中接收channel数据有两种方式:

i <- ch
i, ok <- ch

这两种不同的方法经过编译器的处理都会变成 ORECV 类型的节点,但是后者会在类型检查阶段被转换成 OAS2RECV 节点,我们可以简单看一下这里转换的路线图:
Channel实现原理分析_第2张图片


// entry points for <- c from compiled code
//go:nosplit
func chanrecv1(c *hchan, elem unsafe.Pointer) {
  chanrecv(c, elem, true)
}

//go:nosplit
func chanrecv2(c *hchan, elem unsafe.Pointer) (received bool) {
  _, received = chanrecv(c, elem, true)
  return
}

两者用法不同,chanrecv2可以返回channel是否关闭,但是最终调用方法都是chanrecv,我们来看下源码:

func chanrecv(c *hchan, ep unsafe.Pointer, block bool) (selected, received bool) {
  if debugChan {
    print("chanrecv: chan=", c, "\n")
  }
  //##################step1####################
  if c == nil {
    if !block {
      return
    }
    gopark(nil, nil, waitReasonChanReceiveNilChan, traceEvGoStop, 2)
    throw("unreachable")
  }

  //##################step2####################
  if !block && (c.dataqsiz == 0 && c.sendq.first == nil ||
    c.dataqsiz > 0 && atomic.Loaduint(&c.qcount) == 0) &&
    atomic.Load(&c.closed) == 0 {
    return
  }

  var t0 int64
  if blockprofilerate > 0 {
    t0 = cputicks()
  }

  lock(&c.lock)

  if c.closed != 0 && c.qcount == 0 {
    if raceenabled {
      raceacquire(c.raceaddr())
    }
    unlock(&c.lock)
    if ep != nil {
      typedmemclr(c.elemtype, ep)
    }
    return true, false
  }

  if sg := c.sendq.dequeue(); sg != nil {
    // Found a waiting sender. If buffer is size 0, receive value
    // directly from sender. Otherwise, receive from head of queue
    // and add sender's value to the tail of the queue (both map to
    // the same buffer slot because the queue is full).
    recv(c, sg, ep, func() { unlock(&c.lock) }, 3)
    return true, true
  }

  if c.qcount > 0 {
    // Receive directly from queue
    qp := chanbuf(c, c.recvx)
    if raceenabled {
      raceacquire(qp)
      racerelease(qp)
    }
    if ep != nil {
      typedmemmove(c.elemtype, ep, qp)
    }
    typedmemclr(c.elemtype, qp)
    c.recvx++
    if c.recvx == c.dataqsiz {
      c.recvx = 0
    }
    c.qcount--
    unlock(&c.lock)
    return true, true
  }

  if !block {
    unlock(&c.lock)
    return false, false
  }

  // no sender available: block on this channel.
  gp := getg()
  mysg := acquireSudog()
  mysg.releasetime = 0
  if t0 != 0 {
    mysg.releasetime = -1
  }
  // No stack splits between assigning elem and enqueuing mysg
  // on gp.waiting where copystack can find it.
  mysg.elem = ep
  mysg.waitlink = nil
  gp.waiting = mysg
  mysg.g = gp
  mysg.isSelect = false
  mysg.c = c
  gp.param = nil
  c.recvq.enqueue(mysg)
  goparkunlock(&c.lock, waitReasonChanReceive, traceEvGoBlockRecv, 3)

  // someone woke us up
  if mysg != gp.waiting {
    throw("G waiting list is corrupted")
  }
  gp.waiting = nil
  if mysg.releasetime > 0 {
    blockevent(mysg.releasetime-t0, 2)
  }
  closed := gp.param == nil
  gp.param = nil
  mysg.c = nil
  releaseSudog(mysg)
  return true, !closed
}

由于源码较多,我们逐个step进行讲解;
(1)step1

如果channel是nil:如果是非阻塞模式,直接返回(false,false);如果是阻塞模式,调用goprak挂起goroutine,会阻塞下去。

//##################step1####################
  if c == nil {
    if !block {
      return
    }
    gopark(nil, nil, waitReasonChanReceiveNilChan, traceEvGoStop, 2)
    throw("unreachable")
  }

(2)step2

快速操作(不用获取锁,快速返回),三组条件全部满足,快速返(false,false)

条件1:首先是在非阻塞模式下
条件2:如果是非缓冲型(datasiz=0)并且等待发送goroutine队列为空(sendq.first=nil,就是没人往channel写数据),或者缓冲型channel(datasiz>0)并且buf中没有数据;
条件3:channel未关闭

//##################step2####################
  if !block && (c.dataqsiz == 0 && c.sendq.first == nil ||
    c.dataqsiz > 0 && atomic.Loaduint(&c.qcount) == 0) &&
    atomic.Load(&c.closed) == 0 {
    return
  }

(3)step3

首先加锁,如果channel已经关闭,并且buf中没有元素,返回对应类型的0值,但是received为false;两种情况

情形1:非缓冲型,channel已关闭
情形2:缓冲型,channel已关闭,并且buf无元素

//##################step3####################
  lock(&c.lock)

  if c.closed != 0 && c.qcount == 0 {
    if raceenabled {
      raceacquire(c.raceaddr())
    }
    unlock(&c.lock)
    if ep != nil {
      typedmemclr(c.elemtype, ep)
    }
    return true, false
  }

(4)step4

如果等待发送队列中有元素,证明channel已经满了,两种情形

情形1:非缓冲型,无buf
情形2:缓冲型,buf满了

//##################step4####################
if sg := c.sendq.dequeue(); sg != nil {
    recv(c, sg, ep, func() { unlock(&c.lock) }, 3)
    return true, true
  }

两种情形都正常进入recv方法,我们来看下源码:

func recv(c *hchan, sg *sudog, ep unsafe.Pointer, unlockf func(), skip int) {
  //##################step4-1####################
  if c.dataqsiz == 0 {
    if raceenabled {
      racesync(c, sg)
    }
    if ep != nil {
      // copy data from sender
      recvDirect(c.elemtype, sg, ep)
    }
  } else {
     //##################step4-2####################
    // Queue is full. Take the item at the
    // head of the queue. Make the sender enqueue
    // its item at the tail of the queue. Since the
    // queue is full, those are both the same slot.
    qp := chanbuf(c, c.recvx)
    if raceenabled {
      raceacquire(qp)
      racerelease(qp)
      raceacquireg(sg.g, qp)
      racereleaseg(sg.g, qp)
    }
    // copy data from queue to receiver
    if ep != nil {
      typedmemmove(c.elemtype, ep, qp)
    }
    // copy data from sender to queue
    typedmemmove(c.elemtype, qp, sg.elem)
    c.recvx++
    if c.recvx == c.dataqsiz {
      c.recvx = 0
    }
    c.sendx = c.recvx // c.sendx = (c.sendx+1) % c.dataqsiz
  }
  sg.elem = nil
  gp := sg.g
  unlockf()
  gp.param = unsafe.Pointer(sg)
  if sg.releasetime != 0 {
    sg.releasetime = cputicks()
  }
  goready(gp, skip+1)
}

step4-1:如果是非缓冲型,那么直接从发送者的栈copy到接收者的栈

step4-2:缓冲型的,但是buf已经满了,此时recvx和sendx是重合的,如下图
Channel实现原理分析_第3张图片
首先将recvx处的元素0拷贝到接收地址,然后将下一个元素5拷贝到sendx,然后recvx和sendx分别加1。

step4-3:然后唤醒等待发送队列中的goroutine,等待调度器调度。

(5)step5

没有等待发送的队列,并且buf中有元素,直接把接收游标处的数据copy到接收数据的地址,然后改变hchan中元素数据。

if c.qcount > 0 {
    // Receive directly from queue
    qp := chanbuf(c, c.recvx)
    if raceenabled {
      raceacquire(qp)
      racerelease(qp)
    }
    if ep != nil {
      typedmemmove(c.elemtype, ep, qp)
    }
    typedmemclr(c.elemtype, qp)
    c.recvx++
    if c.recvx == c.dataqsiz {
      c.recvx = 0
    }
    c.qcount--
    unlock(&c.lock)
    return true, true
  }

(6)step6

如果是非阻塞,那么直接返回;如果是阻塞的,构造sudog,保存各种值;将sudog保存到channel的recvq中,调用goparkunlock将goroutine挂起

if !block {
    unlock(&c.lock)
    return false, false
  }
// no sender available: block on this channel.
  gp := getg()
  mysg := acquireSudog()
  mysg.releasetime = 0
  if t0 != 0 {
    mysg.releasetime = -1
  }
  // No stack splits between assigning elem and enqueuing mysg
  // on gp.waiting where copystack can find it.
  mysg.elem = ep
  mysg.waitlink = nil
  gp.waiting = mysg
  mysg.g = gp
  mysg.isSelect = false
  mysg.c = c
  gp.param = nil
  c.recvq.enqueue(mysg)
  goparkunlock(&c.lock, waitReasonChanReceive, traceEvGoBlockRecv, 3)

  // someone woke us up
  if mysg != gp.waiting {
    throw("G waiting list is corrupted")
  }
  gp.waiting = nil
  if mysg.releasetime > 0 {
    blockevent(mysg.releasetime-t0, 2)
  }
  closed := gp.param == nil
  gp.param = nil
  mysg.c = nil
  releaseSudog(mysg)
  return true, !closed

我们用本节一开始的例子来讲解下,再贴一遍程序


func goroutineA(a <-chan int) {
    val := <- a
    fmt.Println("G1 received data: ", val)
    return
}

func goroutineB(b <-chan int) {
    val := <- b
    fmt.Println("G2 received data: ", val)
    return
}

func main() {
    ch := make(chan int)
    go goroutineA(ch)
    go goroutineB(ch)
    ch <- 3
    time.Sleep(time.Second)
}

由于我们创建的channel是无缓冲型的,所以两个goroutine启动的G1和G2会被阻塞,G1和G2被加入到recvq中,状态为waiting,等待被唤醒。此时此刻ch如下图:
Channel实现原理分析_第4张图片
问题:当一个channel关闭后,我们是否还能从中读出数据?


package main

import "fmt"

func main() {
  ch := make(chan int, 6)

  ch <- 1
  ch <- 2

  close(ch)

  a := <-ch
  fmt.Println(a)

  b := <-ch
  fmt.Println(b)

  c := <-ch
  fmt.Println(c)
}

输出会是什么?

1
2
0

我们可以看出,当一个channel关闭后,我们依然可以从中读出数据,如果chan的buf中有元素,则读出的是chan中buf的数据,如果buf为空,则输出对应元素类型的零值。那么我们来看下如下的一段程序:


package main

import (
  "fmt"
  "os"
  "os/signal"
  "syscall"
  "time"
)

var exit1 = make(chan struct{}, 1)

func main() {
  go dealSignal1()
  count := 0
  t := time.Tick(time.Second)
  for {
    select {
    case <-t:
      count++
      fmt.Printf("main run %d\n", count)
    case <-exit1:
      fmt.Println("main exit begin")
    }
  }
  fmt.Println("main exit over")
}

func dealSignal1() {
  c := make(chan os.Signal, 2)
  signal.Notify(c, os.Interrupt, syscall.SIGTERM)
  go func() {
    <-c
    close(exit1)
  }()
}

上面这段程序会有什么问题?

5、发送

我们继续往下走,G1、G2被挂起后,往channel中发送一个数据3,其实调用的是chansend方法,我们还是逐步的去讲解

(1)step1

如果channel=nil,当前goroutine会被挂起

if c == nil {
    if !block {
      return false
    }
    gopark(nil, nil, waitReasonChanSendNilChan, traceEvGoStop, 2)
    throw("unreachable")
  }

(2)step2

依然是一个不加锁的快速操作,三组条件

条件1:非阻塞
条件2:channel未关闭
条件3:channel是非缓冲型,并且等待接收队列为空;或者缓冲型,并且循环数组已经满了

if !block && c.closed == 0 && ((c.dataqsiz == 0 && c.recvq.first == nil) ||
    (c.dataqsiz > 0 && c.qcount == c.dataqsiz)) {
    return false
  }

(3)step3

加锁,如果channel已经关闭,直接panic

lock(&c.lock)

if c.closed != 0 {
    unlock(&c.lock)
    panic(plainError("send on closed channel"))
}

(4)step4

如果等待接收队列不为空,说明什么?

情形1:非缓冲型,等待接收队列不为空
情形2:缓冲型,等待接收队列不为空(说明buf为空)

两种情形,都是直接将待发送数据直接copy到接收处

if sg := c.recvq.dequeue(); sg != nil {
    // Found a waiting receiver. We pass the value we want to send
    // directly to the receiver, bypassing the channel buffer (if any).
    send(c, sg, ep, func() { unlock(&c.lock) }, 3)//直接从ep copy到sg
    return true
}
func send(c *hchan, sg *sudog, ep unsafe.Pointer, unlockf func(), skip int) {
  if raceenabled {
    if c.dataqsiz == 0 {
      racesync(c, sg)
    } else {
      // Pretend we go through the buffer, even though
      // we copy directly. Note that we need to increment
      // the head/tail locations only when raceenabled.
      qp := chanbuf(c, c.recvx)
      raceacquire(qp)
      racerelease(qp)
      raceacquireg(sg.g, qp)
      racereleaseg(sg.g, qp)
      c.recvx++
      if c.recvx == c.dataqsiz {
        c.recvx = 0
      }
      c.sendx = c.recvx // c.sendx = (c.sendx+1) % c.dataqsiz
    }
  }
  if sg.elem != nil {
    sendDirect(c.elemtype, sg, ep)
    sg.elem = nil
  }
  gp := sg.g
  unlockf()
  gp.param = unsafe.Pointer(sg)
  if sg.releasetime != 0 {
    sg.releasetime = cputicks()
  }
  goready(gp, skip+1)
}

两种情形,都直接从一个用一个goroutine操作另一个goroutine的栈,因此在sendDirect方法中会有一次写屏障


func sendDirect(t *_type, sg *sudog, src unsafe.Pointer) {
  // src is on our stack, dst is a slot on another stack.

  // Once we read sg.elem out of sg, it will no longer
  // be updated if the destination's stack gets copied (shrunk).
  // So make sure that no preemption points can happen between read & use.
  dst := sg.elem
  typeBitsBulkBarrier(t, uintptr(dst), uintptr(src), t.size)
  // No need for cgo write barrier checks because dst is always
  // Go memory.
  memmove(dst, src, t.size)
}

(5)step5

如果等待队列为空,并且缓冲区未满,肯定是缓冲型的channel

if c.qcount < c.dataqsiz {
    // Space is available in the channel buffer. Enqueue the element to send.
    qp := chanbuf(c, c.sendx)
    if raceenabled {
      raceacquire(qp)
      racerelease(qp)
    }
    typedmemmove(c.elemtype, qp, ep)
    c.sendx++
    if c.sendx == c.dataqsiz {
      c.sendx = 0
    }
    c.qcount++
    unlock(&c.lock)
    return true
  }

将元素放在sendx处,然后sendx加1,channel总量加1

(6)step6

如果以上情况都没有命中,说明什么?说明channel已经满了,如果是非阻塞的直接返回,否则需要调用gopack将这个goroutine挂起,等待被唤醒。


if !block {
    unlock(&c.lock)
    return false
  }

  // Block on the channel. Some receiver will complete our operation for us.
  gp := getg()
  mysg := acquireSudog()
  mysg.releasetime = 0
  if t0 != 0 {
    mysg.releasetime = -1
  }
  // No stack splits between assigning elem and enqueuing mysg
  // on gp.waiting where copystack can find it.
  mysg.elem = ep
  mysg.waitlink = nil
  mysg.g = gp
  mysg.isSelect = false
  mysg.c = c
  gp.waiting = mysg
  gp.param = nil
  c.sendq.enqueue(mysg)
  goparkunlock(&c.lock, waitReasonChanSend, traceEvGoBlockSend, 3)
  // Ensure the value being sent is kept alive until the
  // receiver copies it out. The sudog has a pointer to the
  // stack object, but sudogs aren't considered as roots of the
  // stack tracer.
  KeepAlive(ep)

我们对照程序分析下,在前一个小节G1、G2被挂起来了,等待sender的解救;这时候往ch中发送了一个3,(step4)这时sender发现ch的等待接收队列recvq中有receiver,就会出队一个sudog,然后将元素直接copy到sudog的elem处,然后调用goready将G1唤醒,继续执行G1原来的代码,打印出结果。如下图:
Channel实现原理分析_第5张图片
6、关闭

close一个channel会调用closechan方法,比较简单,我们也来看下

(1)step1

如果channel为nil,会直接panic

if c == nil {
    panic(plainError("close of nil channel"))
  }

(2)step2

加锁,如果channel已经关闭,再次关闭会panic

lock(&c.lock)
  if c.closed != 0 {
    unlock(&c.lock)
    panic(plainError("close of closed channel"))
  }

(3)step3

首选将hchan对应close标志置为1,然后声明一个链表;将等待接收队列中的所有sudog加入到链表,并将其elem赋予一个相应类型的0值;


c.closed = 1

  var glist gList

  // release all readers
  for {
    sg := c.recvq.dequeue()
    if sg == nil {
      break
    }
    if sg.elem != nil {
      typedmemclr(c.elemtype, sg.elem)
      sg.elem = nil
    }
    if sg.releasetime != 0 {
      sg.releasetime = cputicks()
    }
    gp := sg.g
    gp.param = nil
    if raceenabled {
      raceacquireg(gp, c.raceaddr())
    }
    glist.push(gp)
  }

(4)step4

将所有等待发送队列的sudog加入链表

// release all writers (they will panic)
  for {
    sg := c.sendq.dequeue()
    if sg == nil {
      break
    }
    sg.elem = nil
    if sg.releasetime != 0 {
      sg.releasetime = cputicks()
    }
    gp := sg.g
    gp.param = nil
    if raceenabled {
      raceacquireg(gp, c.raceaddr())
    }
    glist.push(gp)
  }
  unlock(&c.lock)

(5)step5

唤醒sudog所有goroutine

for !glist.empty() {
    gp := glist.pop()
    gp.schedlink = 0
    goready(gp, 3)
  }

问题

问题1:哪些操作会使channel发生panic?

三种情况

情况1:往一个已经关闭的channel写数据
情况2:关闭一个nil的channel
情况3:关闭一个已经关闭的channel

问题2:channel是并发安全的吗?

问题3:当一个channel关闭后,我们是否还能从channel读到数据?

可以,只不过接收的是无效数据

问题4:channel发送和接收元素的本质是什么?

值的拷贝

看一段示例


package main
import (
  "fmt"
  "time"
)

func print(u <-chan int) {
  time.Sleep(2 * time.Second)
  fmt.Println("print int", <-u)
}

func main() {
  c := make(chan int, 5)
  a := 0

  c <- a
  fmt.Println(a)
  // modify g
  a = 1

  go print(c)
  time.Sleep(5 * time.Second)
  fmt.Println(a)
}

再看一段复杂一点的


package main

import (
  "fmt"
  "time"
)

type people struct {
  name string
}

var u = people{name: "A"}

func printPeople(u <-chan *people) {
  time.Sleep(2 * time.Second)
  fmt.Println("printPeople", <-u)
}

func main() {
  c := make(chan *people, 5)
  var a = &u

  c <- a
  fmt.Println(a)
  // modify g
  a = &people{name:"B"}

  go printPeople(c)
  time.Sleep(5 * time.Second)
  fmt.Println(a)
}

输出会是什么

&{A}
printPeople &{A}
&{B}

因为chan中保存的是u的地址的值的拷贝,这个地址未发生改变,虽然调用a = &people{name:“B”}重新赋予了a新的地址,但是channel中的未改变。

参考文献

【1】《深度解密Go语言之channel》出处:博客园 作者:Stefno

【2】《浅谈Go语言编译原理》出处:Draveness’s Blog(draveness.me)

【3】《Concurrency in Go》

【4】https://mp.weixin.qq.com/s/MF3o5Jr7h-EiSN8amKW0QA

你可能感兴趣的:(go)