造个可重入锁的轮子
- 介绍
- 目标
- 正文
- sync.Mutex
- sync.Mutex介绍
- 多协程计数器demo
- 多协程计数器加锁
- 源码剖析
- Mutex数据结构
- Lock()
- 加锁核心逻辑
- UnLock()
- 重入锁与可重入锁
- 魔改 sync.Mutex
- 参考文档
介绍
开新坑啦!!
从这篇文章开始,尝试造轮子,包括一些可能有用、也可能没用的轮子。
温故而知新,我相信时常回顾基础的东西能让我们受益良多,这点我深有体会,每过一段时间我都会把《程序员的自我修养》拿出来翻翻,常翻常新,每次读都能有新的收获,开始吧。
“转向毕竟是一个很长的过程,先做起来吧,给我和别的生命一个活下去的机会。”—《三体》
目标
用go 实现可重入锁;
正文
Golang的sync.Mutex是并发场景下的“灵丹妙药”,但是我们真的了解吗?
本文通过对源码的剖析,让我们重新、更全面的认识sync.Mutex,尤其是其优缺点。
在对sync.Mutex有了深入了解后,我们尝试对其进行魔改,实现可重入锁。
sync.Mutex
点击上方的sync.Mutex进入golang.tour我们可以看到,sync.Mutex的简单介绍;
总结如下:
sync.Mutex介绍
当多个goroutine之间需要通信,尤其需要访问(同一份)数据时,需要互斥(锁)来保证一次只有一个goroutine访问数据。
当然,sync.Mutex作为一个同步原语实现了Locker接口(后面会提到),所以只有Lock()、Unlock()两个接口。
多协程计数器demo
比如大家在许多地方看到的例子,多协程累加计数:
var (counter = 1
)func incrCounter() {var wg sync.WaitGroupwg.Add(11)i := 0for {go func() {defer wg.Done()iter:=0for {counter ++if iter > 4 {break}iter ++}}()if i > 9 {break}i ++}wg.Wait()fmt.Println("incrCounter:",counter)os.Exit(1)
}
本地运行结果是incrCounter: 56;大家可以试试本地运行的结果。
多协程计数器加锁
接下来就说到今天的主角了;在上面的多协程计数器代码上,集成sync.Mutex,加两行代码,分别是Lock、Unlock;
代码如下
var (mtx sync.Mutexcounter = 1
)
func incrCounter() {var wg sync.WaitGroupwg.Add(11)i := 0for {go func() {defer wg.Done()iter:=0for {mtx.Lock() counter ++mtx.Unlock()if iter > 4 {break}iter ++}}()if i > 9 {break}i ++}wg.Wait()fmt.Println("incrCounter:",counter)os.Exit(1)
}
本地运行结果是incrCounter: 67;大家可以试试本地运行的结果。(相信很多人看到67会觉得奇怪,没错我是故意的,就是给粗心的同学卖了一个坑,想想为什么是67而不是66?!)
奇怪哎,为啥加了sync.Mutex不一样了呢?!
这里一定要回顾下开头sync.Mutex的介绍!
源码剖析
点击sync.Mutex,我们可以看到它的数据结构;
Mutex数据结构
简单解释下分别是:
state状态位(如果不是远古版本,分为了4段),以及sema信号量变量;不急,后面会细说,这里先了解基本构成;
type Mutex struct {state int32sema uint32
}
回顾多协程计数器中sync.Mutex的使用例子,核心方法只有两个,为什么只有两个呢?看源码发现原来是实现了Locker interface,因为实现了Locker所以有Lock()、UnLock();
这里需要重点说一下,golang中的同步原语都会实现Locker ,比如RWMutex;所以以后提到Lock、UnLock那么就可以思考是不是实现了Locker interface;
// A Locker represents an object that can be locked and unlocked.
type Locker interface {Lock()Unlock()
}
接下来,看下Lock()的实现;看看golang是如何加锁的。
Lock()
照例,点进去看下源码;
如果没加锁,运气很好,加锁就行然后返回;如果已经加过锁了,那么就进入lockSlow,也是加锁逻辑最复杂的地方;
race是做死锁检查的,先不管,捋主体逻辑先;
这里多提一下,fast path一般用来表示捷径或者幸运case,意思是直接成功,不用再执行复杂的逻辑,如果大家看多了开源项目看到fast path就可以跳过这段代码,因为不用看你也能猜到这段代码的意思;
func (m *Mutex) Lock() {// Fast path: grab unlocked mutex.if atomic.CompareAndSwapInt32(&m.state, 0, mutexLocked) {if race.Enabled {race.Acquire(unsafe.Pointer(m))}return}// Slow path (outlined so that the fast path can be inlined)m.lockSlow()
}
加锁核心逻辑
先看几个变量;
表示饥饿模式的starving,
唤醒状态的标记awoke,
迭代次数统计的iter,
当前的加锁状态old;
var waitStartTime int64starving := falseawoke := falseiter := 0old := m.state
接下来是饥饿模式的自旋逻辑;
if old&(mutexLocked|mutexStarving) == mutexLocked && runtime_canSpin(iter) {// Active spinning makes sense.// Try to set mutexWoken flag to inform Unlock// to not wake other blocked goroutines.if !awoke && old&mutexWoken == 0 && old>>mutexWaiterShift != 0 &&atomic.CompareAndSwapInt32(&m.state, old, old|mutexWoken) {awoke = true}runtime_doSpin()iter++old = m.statecontinue}
如果是饥饿模式,那么直接拿到锁,新到的goroutine会放入等待队列(等待队列数+1);
new := old
if old&mutexStarving == 0 {new |= mutexLocked}
if old&(mutexLocked|mutexStarving) != 0 {new += 1 << mutexWaiterShift}
如果当前协程是饥饿模式,并且Mutex并没有标记为饥饿模式,那么就把Mutex标记为饥饿模式;如果已被唤醒那么就标记为已唤醒状态;
if starving && old&mutexLocked != 0 {new |= mutexStarving}
if awoke {// The goroutine has been woken from sleep,// so we need to reset the flag in either case.if new&mutexWoken == 0 {throw("sync: inconsistent mutex state")}new &^= mutexWoken}
紧接着,将改变的状态同步到Mutex.state字段;
如果到目前为止当前协程没有获取到锁也没有进入饥饿模式,就可以提前结束当前流程(等待下一次唤醒);
判断运行时间,并调用runtime_SemacquireMutex休眠,并尝试获取信号量;
运行时间超过1ms就自动进入饥饿模式(starving = 1);
一旦当前Mutex被标记为饥饿模式,将状态保存到Mutex.state中;
这里需要注意state(int32)中各段的:
第一段(最左边29位)为等待协程的数量;
第二段(1位)饥饿模式标记;
第三段(1位)唤醒标记;
第四段(1位)是否加锁;
如果没有进入饥饿模式,那么将唤醒标记为true,并且重新开始(继续尝试获取锁);
if atomic.CompareAndSwapInt32(&m.state, old, new) {if old&(mutexLocked|mutexStarving) == 0 {break // locked the mutex with CAS}// If we were already waiting before, queue at the front of the queue.queueLifo := waitStartTime != 0if waitStartTime == 0 {waitStartTime = runtime_nanotime()}runtime_SemacquireMutex(&m.sema, queueLifo, 1)starving = starving || runtime_nanotime()-waitStartTime > starvationThresholdNsold = m.stateif old&mutexStarving != 0 {// If this goroutine was woken and mutex is in starvation mode,// ownership was handed off to us but mutex is in somewhat// inconsistent state: mutexLocked is not set and we are still// accounted as waiter. Fix that.if old&(mutexLocked|mutexWoken) != 0 || old>>mutexWaiterShift == 0 {throw("sync: inconsistent mutex state")}delta := int32(mutexLocked - 1<<mutexWaiterShift)if !starving || old>>mutexWaiterShift == 1 {// Exit starvation mode.// Critical to do it here and consider wait time.// Starvation mode is so inefficient, that two goroutines// can go lock-step infinitely once they switch mutex// to starvation mode.delta -= mutexStarving}atomic.AddInt32(&m.state, delta)break}awoke = trueiter = 0} else {old = m.state}
UnLock()
记得前面的fast path这个case吗;如果state为1就直接释放锁然后就结束了;
这里需要结合加锁逻辑去看;
回顾下组成state的四个部分:
第一段(最左边29位)为等待协程的数量;
第二段(1位)饥饿模式标记;
第三段(1位)唤醒标记;
第四段(1位)是否加锁;
那么state为,说明:没有等待的协程,没有饥饿模式和唤醒标记,仅仅Mutex被加锁了;
func (m *Mutex) Unlock() {if race.Enabled {_ = m.staterace.Release(unsafe.Pointer(m))}// Fast path: drop lock bit.new := atomic.AddInt32(&m.state, -mutexLocked)if new != 0 {// Outlined slow path to allow inlining the fast path.// To hide unlockSlow during tracing we skip one extra frame when tracing GoUnblock.m.unlockSlow(new)}
}
否则,进入unlockSlow逻辑;
入口时异常判断,如果释放一个没有加锁的锁则抛出异常;
如果是饥饿模式,将锁直接给饥饿模式的协程,注意是饥饿模式的协程不是等待队列中的等待协程;
不是饥饿模式(比如正常的等待协程)是正常模式,判断锁是否已被锁定或者是否存在唤醒或者是否是饥饿模式,则直接放回,并不释放锁;否则唤醒等待队列中的协程,直接移交给等待者;
func (m *Mutex) unlockSlow(new int32) {if (new+mutexLocked)&mutexLocked == 0 {fatal("sync: unlock of unlocked mutex")}if new&mutexStarving == 0 {old := newfor {// If there are no waiters or a goroutine has already// been woken or grabbed the lock, no need to wake anyone.// In starvation mode ownership is directly handed off from unlocking// goroutine to the next waiter. We are not part of this chain,// since we did not observe mutexStarving when we unlocked the mutex above.// So get off the way.if old>>mutexWaiterShift == 0 || old&(mutexLocked|mutexWoken|mutexStarving) != 0 {return}// Grab the right to wake someone.new = (old - 1<<mutexWaiterShift) | mutexWokenif atomic.CompareAndSwapInt32(&m.state, old, new) {runtime_Semrelease(&m.sema, false, 1)return}old = m.state}} else {// Starving mode: handoff mutex ownership to the next waiter, and yield// our time slice so that the next waiter can start to run immediately.// Note: mutexLocked is not set, the waiter will set it after wakeup.// But mutex is still considered locked if mutexStarving is set,// so new coming goroutines won't acquire it.runtime_Semrelease(&m.sema, true, 1)}
}
重入锁与可重入锁
未完待续
魔改 sync.Mutex
未完待续
参考文档
认识可重入锁
Mutex
饥饿模式