Add: deadlock check

README.md

@@ -2,6 +2,13 @@
 
 Optimization of something in golang sdk
 
+## sync
+
+> deadlock
+
+Copy from [go-deadlock](https://github.com/sasha-s/go-deadlock) and add EnableDeadlock.
+
 ## time
 
 Timer optimization through time-wheel.
+

sync/Readme.md

+# Online deadlock detection in go (golang). [![Try it online](https://img.shields.io/badge/try%20it-online-blue.svg)](https://wandbox.org/permlink/hJc6QCZowxbNm9WW) [![Docs](https://godoc.org/github.com/sasha-s/go-deadlock?status.svg)](https://godoc.org/github.com/sasha-s/go-deadlock) [![Build Status](https://travis-ci.org/sasha-s/go-deadlock.svg?branch=master)](https://travis-ci.org/sasha-s/go-deadlock) [![codecov](https://codecov.io/gh/sasha-s/go-deadlock/branch/master/graph/badge.svg)](https://codecov.io/gh/sasha-s/go-deadlock) [![version](https://badge.fury.io/gh/sasha-s%2Fgo-deadlock.svg)](https://github.com/sasha-s/go-deadlock/releases)  [![Go Report Card](https://goreportcard.com/badge/github.com/sasha-s/go-deadlock)](https://goreportcard.com/report/github.com/sasha-s/go-deadlock) [![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0) 
+
+## Why
+Deadlocks happen and are painful to debug.
+
+## What
+go-deadlock provides (RW)Mutex drop-in replacements for sync.(RW)Mutex.
+It would not work if you create a spaghetti of channels.
+Mutexes only.
+
+## Installation
+```sh
+go get github.com/sasha-s/go-deadlock/...
+```
+
+## Usage
+```go
+import "github.com/sasha-s/go-deadlock"
+var mu deadlock.Mutex
+// Use normally, it works exactly like sync.Mutex does.
+mu.Lock()
+
+defer mu.Unlock()
+// Or
+var rw deadlock.RWMutex
+rw.RLock()
+defer rw.RUnlock()
+```
+
+### Deadlocks
+One of the most common sources of deadlocks is inconsistent lock ordering:
+say, you have two mutexes A and B, and in some goroutines you have
+```go
+A.Lock() // defer A.Unlock() or similar.
+...
+B.Lock() // defer B.Unlock() or similar.
+```
+And in another goroutine the order of locks is reversed:
+```go
+B.Lock() // defer B.Unlock() or similar.
+...
+A.Lock() // defer A.Unlock() or similar.
+```
+
+Another common sources of deadlocs is duplicate take a lock in a goroutine:
+```
+A.Rlock() or lock()
+
+A.lock() or A.RLock()
+```
+
+This does not guarantee a deadlock (maybe the goroutines above can never be running at the same time), but it usually a design flaw at least.
+
+go-deadlock can detect such cases (unless you cross goroutine boundary - say lock A, then spawn a goroutine, block until it is singals, and lock B inside of the goroutine), even if the deadlock itself happens very infrequently and is painful to reproduce!
+
+Each time go-deadlock sees a lock attempt for lock B, it records the order A before B, for each lock that is currently being held in the same goroutine, and it prints (and exits the program by default) when it sees the locking order being violated.
+
+In addition, if it sees that we are waiting on a lock for a long time (opts.DeadlockTimeout, 30 seconds by default), it reports a potential deadlock, also printing the stacktrace for a goroutine that is currently holding the lock we are desperately trying to grab.
+
+
+## Sample output
+#### Inconsistent lock ordering:
+```
+POTENTIAL DEADLOCK: Inconsistent locking. saw this ordering in one goroutine:
+happened before
+inmem.go:623 bttest.(*server).ReadModifyWriteRow { r.mu.Lock() } <<<<<
+inmem_test.go:118 bttest.TestConcurrentMutationsReadModifyAndGC.func4 { _, _ = s.ReadModifyWriteRow(ctx, rmw()) }
+
+happened after
+inmem.go:629 bttest.(*server).ReadModifyWriteRow { tbl.mu.RLock() } <<<<<
+inmem_test.go:118 bttest.TestConcurrentMutationsReadModifyAndGC.func4 { _, _ = s.ReadModifyWriteRow(ctx, rmw()) }
+
+in another goroutine: happened before
+inmem.go:799 bttest.(*table).gc { t.mu.RLock() } <<<<<
+inmem_test.go:125 bttest.TestConcurrentMutationsReadModifyAndGC.func5 { tbl.gc() }
+
+happend after
+inmem.go:814 bttest.(*table).gc { r.mu.Lock() } <<<<<
+inmem_test.go:125 bttest.TestConcurrentMutationsReadModifyAndGC.func5 { tbl.gc() }
+```
+
+#### Waiting for a lock for a long time:
+
+```
+POTENTIAL DEADLOCK:
+Previous place where the lock was grabbed
+goroutine 240 lock 0xc820160440
+inmem.go:799 bttest.(*table).gc { t.mu.RLock() } <<<<<
+inmem_test.go:125 bttest.TestConcurrentMutationsReadModifyAndGC.func5 { tbl.gc() }
+
+Have been trying to lock it again for more than 40ms
+goroutine 68 lock 0xc820160440
+inmem.go:785 bttest.(*table).mutableRow { t.mu.Lock() } <<<<<
+inmem.go:428 bttest.(*server).MutateRow { r := tbl.mutableRow(string(req.RowKey)) }
+inmem_test.go:111 bttest.TestConcurrentMutationsReadModifyAndGC.func3 { s.MutateRow(ctx, req) }
+
+
+Here is what goroutine 240 doing now
+goroutine 240 [select]:
+github.com/sasha-s/go-deadlock.lock(0xc82028ca10, 0x5189e0, 0xc82013a9b0)
+        /Users/sasha/go/src/github.com/sasha-s/go-deadlock/deadlock.go:163 +0x1640
+github.com/sasha-s/go-deadlock.(*Mutex).Lock(0xc82013a9b0)
+        /Users/sasha/go/src/github.com/sasha-s/go-deadlock/deadlock.go:54 +0x86
+google.golang.org/cloud/bigtable/bttest.(*table).gc(0xc820160440)
+        /Users/sasha/go/src/google.golang.org/cloud/bigtable/bttest/inmem.go:814 +0x28d
+google.golang.org/cloud/bigtable/bttest.TestConcurrentMutationsReadModifyAndGC.func5(0xc82015c760, 0xc820160440)      /Users/sasha/go/src/google.golang.org/cloud/bigtable/bttest/inmem_test.go:125 +0x48
+created by google.golang.org/cloud/bigtable/bttest.TestConcurrentMutationsReadModifyAndGC
+        /Users/sasha/go/src/google.golang.org/cloud/bigtable/bttest/inmem_test.go:126 +0xb6f
+```
+
+## Used in
+[cockroachdb: Potential deadlock between Gossip.SetStorage and Node.gossipStores](https://github.com/cockroachdb/cockroach/issues/7972)
+
+[bigtable/bttest: A race between GC and row mutations](https://code-review.googlesource.com#/c/5301/)
+
+## Need a mutex that works with net.context?
+I have [one](https://github.com/sasha-s/go-csync).
+
+## Grabbing an RLock twice from the same goroutine
+This is, surprisingly, not a good idea!
+
+From [RWMutex](https://golang.org/pkg/sync/#RWMutex) docs:
+
+>If a goroutine holds a RWMutex for reading and another goroutine might call Lock, no goroutine should expect to be able to acquire a read lock until the initial read lock is released. In particular, this prohibits recursive read locking. This is to ensure that the lock eventually becomes available; a blocked Lock call excludes new readers from acquiring the lock.
+
+
+The following code will deadlock &mdash; [run the example on playground](https://play.golang.org/p/AkL-W63nq5f) or [try it online with go-deadlock on wandbox](https://wandbox.org/permlink/JwnL0GMySBju4SII):
+```go
+package main
+
+import (
+	"fmt"
+	"sync"
+)
+
+func main() {
+	var mu sync.RWMutex
+
+	chrlockTwice := make(chan struct{}) // Used to control rlockTwice
+	rlockTwice := func() {
+		mu.RLock()
+		fmt.Println("first Rlock succeeded")
+		<-chrlockTwice
+		<-chrlockTwice
+		fmt.Println("trying to Rlock again")
+		mu.RLock()
+		fmt.Println("second Rlock succeeded")
+		mu.RUnlock()
+		mu.RUnlock()
+	}
+
+	chLock := make(chan struct{}) // Used to contol lock
+	lock := func() {
+		<-chLock
+		fmt.Println("about to Lock")
+		mu.Lock()
+		fmt.Println("Lock succeeded")
+		mu.Unlock()
+		<-chLock
+	}
+
+	control := func() {
+		chrlockTwice <- struct{}{}
+		chLock <- struct{}{}
+
+		close(chrlockTwice)
+		close(chLock)
+	}
+
+	go control()
+	go lock()
+	rlockTwice()
+}
+```
+

sync/deadlock.go

+package gxsync
+// ref: github.com/sasha-s/go-deadlock
+
+import (
+	"bufio"
+	"bytes"
+	"fmt"
+	"io"
+	"os"
+	"sync"
+	"time"
+
+	"github.com/petermattis/goid"
+)
+
+// Opts control how deadlock detection behaves.
+// Options are supposed to be set once at a startup (say, when parsing flags).
+var Opts = struct {
+	// Mutex/RWMutex would work exactly as their sync counterparts
+	// -- almost no runtime penalty, no deadlock detection if Disable == true.
+	Disable bool
+	// Would disable lock order based deadlock detection if DisableLockOrderDetection == true.
+	DisableLockOrderDetection bool
+	// Waiting for a lock for longer than DeadlockTimeout is considered a deadlock.
+	// Ignored is DeadlockTimeout <= 0.
+	DeadlockTimeout time.Duration
+	// OnPotentialDeadlock is called each time a potential deadlock is detected -- either based on
+	// lock order or on lock wait time.
+	OnPotentialDeadlock func()
+	// Will keep MaxMapSize lock pairs (happens before // happens after) in the map.
+	// The map resets once the threshold is reached.
+	MaxMapSize int
+	// Will dump stacktraces of all goroutines when inconsistent locking is detected.
+	PrintAllCurrentGoroutines bool
+	mu                        *sync.Mutex // Protects the LogBuf.
+	// Will print deadlock info to log buffer.
+	LogBuf io.Writer
+}{
+	DeadlockTimeout: time.Second * 30,
+	OnPotentialDeadlock: func() {
+		os.Exit(2)
+	},
+	MaxMapSize: 1024 * 64,
+	mu:         &sync.Mutex{},
+	LogBuf:     os.Stderr,
+}
+
+// Cond is sync.Cond wrapper
+type Cond struct {
+	sync.Cond
+}
+
+// Locker is sync.Locker wrapper
+type Locker struct {
+	sync.Locker
+}
+
+// Once is sync.Once wrapper
+type Once struct {
+	sync.Once
+}
+
+// Pool is sync.Poll wrapper
+type Pool struct {
+	sync.Pool
+}
+
+// WaitGroup is sync.WaitGroup wrapper
+type WaitGroup struct {
+	sync.WaitGroup
+}
+
+// A Mutex is a drop-in replacement for sync.Mutex.
+// Performs deadlock detection unless disabled in Opts.
+type Mutex struct {
+	mu sync.Mutex
+}
+
+// Lock locks the mutex.
+// If the lock is already in use, the calling goroutine
+// blocks until the mutex is available.
+//
+// Unless deadlock detection is disabled, logs potential deadlocks to Opts.LogBuf,
+// calling Opts.OnPotentialDeadlock on each occasion.
+func (m *Mutex) Lock() {
+	lock(m.mu.Lock, m)
+}
+
+// Unlock unlocks the mutex.
+// It is a run-time error if m is not locked on entry to Unlock.
+//
+// A locked Mutex is not associated with a particular goroutine.
+// It is allowed for one goroutine to lock a Mutex and then
+// arrange for another goroutine to unlock it.
+func (m *Mutex) Unlock() {
+	m.mu.Unlock()
+	if !Opts.Disable {
+		postUnlock(m)
+	}
+}
+
+// An RWMutex is a drop-in replacement for sync.RWMutex.
+// Performs deadlock detection unless disabled in Opts.
+type RWMutex struct {
+	mu sync.RWMutex
+}
+
+// Lock locks rw for writing.
+// If the lock is already locked for reading or writing,
+// Lock blocks until the lock is available.
+// To ensure that the lock eventually becomes available,
+// a blocked Lock call excludes new readers from acquiring
+// the lock.
+//
+// Unless deadlock detection is disabled, logs potential deadlocks to Opts.LogBuf,
+// calling Opts.OnPotentialDeadlock on each occasion.
+func (m *RWMutex) Lock() {
+	lock(m.mu.Lock, m)
+}
+
+// Unlock unlocks the mutex for writing.  It is a run-time error if rw is
+// not locked for writing on entry to Unlock.
+//
+// As with Mutexes, a locked RWMutex is not associated with a particular
+// goroutine.  One goroutine may RLock (Lock) an RWMutex and then
+// arrange for another goroutine to RUnlock (Unlock) it.
+func (m *RWMutex) Unlock() {
+	m.mu.Unlock()
+	if !Opts.Disable {
+		postUnlock(m)
+	}
+}
+
+// RLock locks the mutex for reading.
+//
+// Unless deadlock detection is disabled, logs potential deadlocks to Opts.LogBuf,
+// calling Opts.OnPotentialDeadlock on each occasion.
+func (m *RWMutex) RLock() {
+	lock(m.mu.RLock, m)
+}
+
+// RUnlock undoes a single RLock call;
+// it does not affect other simultaneous readers.
+// It is a run-time error if rw is not locked for reading
+// on entry to RUnlock.
+func (m *RWMutex) RUnlock() {
+	m.mu.RUnlock()
+	if !Opts.Disable {
+		postUnlock(m)
+	}
+}
+
+// RLocker returns a Locker interface that implements
+// the Lock and Unlock methods by calling RLock and RUnlock.
+func (m *RWMutex) RLocker() sync.Locker {
+	return (*rlocker)(m)
+}
+
+func preLock(skip int, p interface{}) {
+	lo.preLock(skip, p)
+}
+
+func postLock(skip int, p interface{}) {
+	lo.postLock(skip, p)
+}
+
+func postUnlock(p interface{}) {
+	lo.postUnlock(p)
+}
+
+func lock(lockFn func(), ptr interface{}) {
+	if Opts.Disable {
+		lockFn()
+		return
+	}
+	preLock(4, ptr)
+	if Opts.DeadlockTimeout <= 0 {
+		lockFn()
+	} else {
+		ch := make(chan struct{})
+		go func() {
+			for {
+				t := time.NewTimer(Opts.DeadlockTimeout)
+				defer t.Stop() // This runs after the losure finishes, but it's OK.
+				select {
+				case <-t.C:
+					lo.mu.Lock()
+					prev, ok := lo.cur[ptr]
+					if !ok {
+						lo.mu.Unlock()
+						break // Nobody seems to be holding the lock, try again.
+					}
+					Opts.mu.Lock()
+					fmt.Fprintln(Opts.LogBuf, header)
+					fmt.Fprintln(Opts.LogBuf, "Previous place where the lock was grabbed")
+					fmt.Fprintf(Opts.LogBuf, "goroutine %v lock %p\n", prev.gid, ptr)
+					printStack(Opts.LogBuf, prev.stack)
+					fmt.Fprintln(Opts.LogBuf, "Have been trying to lock it again for more than", Opts.DeadlockTimeout)
+					fmt.Fprintf(Opts.LogBuf, "goroutine %v lock %p\n", goid.Get(), ptr)
+					printStack(Opts.LogBuf, callers(2))
+					stacks := stacks()
+					grs := bytes.Split(stacks, []byte("\n\n"))
+					for _, g := range grs {
+						if goid.ExtractGID(g) == prev.gid {
+							fmt.Fprintln(Opts.LogBuf, "Here is what goroutine", prev.gid, "doing now")
+							Opts.LogBuf.Write(g)
+							fmt.Fprintln(Opts.LogBuf)
+						}
+					}
+					lo.other(ptr)
+					if Opts.PrintAllCurrentGoroutines {
+						fmt.Fprintln(Opts.LogBuf, "All current goroutines:")
+						Opts.LogBuf.Write(stacks)
+					}
+					fmt.Fprintln(Opts.LogBuf)
+					if buf, ok := Opts.LogBuf.(*bufio.Writer); ok {
+						buf.Flush()
+					}
+					Opts.mu.Unlock()
+					lo.mu.Unlock()
+					Opts.OnPotentialDeadlock()
+					<-ch
+					return
+				case <-ch:
+					return
+				}
+			}
+		}()
+		lockFn()
+		postLock(4, ptr)
+		close(ch)
+		return
+	}
+	postLock(4, ptr)
+}
+
+type lockOrder struct {
+	mu    sync.Mutex
+	cur   map[interface{}]stackGID // stacktraces + gids for the locks currently taken.
+	order map[beforeAfter]ss       // expected order of locks.
+}
+
+type stackGID struct {
+	stack []uintptr
+	gid   int64
+}
+
+type beforeAfter struct {
+	before interface{}
+	after  interface{}
+}
+
+type ss struct {
+	before []uintptr
+	after  []uintptr
+}
+
+var lo = newLockOrder()
+
+func newLockOrder() *lockOrder {
+	return &lockOrder{
+		cur:   map[interface{}]stackGID{},
+		order: map[beforeAfter]ss{},
+	}
+}
+
+func (l *lockOrder) postLock(skip int, p interface{}) {
+	stack := callers(skip)
+	gid := goid.Get()
+	l.mu.Lock()
+	l.cur[p] = stackGID{stack, gid}
+	l.mu.Unlock()
+}
+
+func (l *lockOrder) preLock(skip int, p interface{}) {
+	if Opts.DisableLockOrderDetection {
+		return
+	}
+	stack := callers(skip)
+	gid := goid.Get()
+	l.mu.Lock()
+	for b, bs := range l.cur {
+		if b == p {
+			if bs.gid == gid {
+				Opts.mu.Lock()
+				fmt.Fprintln(Opts.LogBuf, header, "Recursive locking:")
+				fmt.Fprintf(Opts.LogBuf, "current goroutine %d lock %p\n", gid, b)
+				printStack(Opts.LogBuf, stack)
+				fmt.Fprintln(Opts.LogBuf, "Previous place where the lock was grabbed (same goroutine)")
+				printStack(Opts.LogBuf, bs.stack)
+				l.other(p)
+				if buf, ok := Opts.LogBuf.(*bufio.Writer); ok {
+					buf.Flush()
+				}
+				Opts.mu.Unlock()
+				Opts.OnPotentialDeadlock()
+			}
+			continue
+		}
+		if bs.gid != gid { // We want locks taken in the same goroutine only.
+			continue
+		}
+		if s, ok := l.order[beforeAfter{p, b}]; ok {
+			Opts.mu.Lock()
+			fmt.Fprintln(Opts.LogBuf, header, "Inconsistent locking. saw this ordering in one goroutine:")
+			fmt.Fprintln(Opts.LogBuf, "happened before")
+			printStack(Opts.LogBuf, s.before)
+			fmt.Fprintln(Opts.LogBuf, "happened after")
+			printStack(Opts.LogBuf, s.after)
+			fmt.Fprintln(Opts.LogBuf, "in another goroutine: happened before")
+			printStack(Opts.LogBuf, bs.stack)
+			fmt.Fprintln(Opts.LogBuf, "happened after")
+			printStack(Opts.LogBuf, stack)
+			l.other(p)
+			fmt.Fprintln(Opts.LogBuf)
+			if buf, ok := Opts.LogBuf.(*bufio.Writer); ok {
+				buf.Flush()
+			}
+			Opts.mu.Unlock()
+			Opts.OnPotentialDeadlock()
+		}
+		l.order[beforeAfter{b, p}] = ss{bs.stack, stack}
+		if len(l.order) == Opts.MaxMapSize { // Reset the map to keep memory footprint bounded.
+			l.order = map[beforeAfter]ss{}
+		}
+	}
+	l.mu.Unlock()
+}
+
+func (l *lockOrder) postUnlock(p interface{}) {
+	l.mu.Lock()
+	delete(l.cur, p)
+	l.mu.Unlock()
+}
+
+type rlocker RWMutex
+
+func (r *rlocker) Lock()   { (*RWMutex)(r).RLock() }
+func (r *rlocker) Unlock() { (*RWMutex)(r).RUnlock() }
+
+// Under lo.mu Locked.
+func (l *lockOrder) other(ptr interface{}) {
+	empty := true
+	for k := range l.cur {
+		if k == ptr {
+			continue
+		}
+		empty = false
+	}
+	if empty {
+		return
+	}
+	fmt.Fprintln(Opts.LogBuf, "Other goroutines holding locks:")
+	for k, pp := range l.cur {
+		if k == ptr {
+			continue
+		}
+		fmt.Fprintf(Opts.LogBuf, "goroutine %v lock %p\n", pp.gid, k)
+		printStack(Opts.LogBuf, pp.stack)
+	}
+	fmt.Fprintln(Opts.LogBuf)
+}
+
+const header = "POTENTIAL DEADLOCK:"

sync/deadlock_map.go

+// +build go1.9
+
+package gxsync
+
+
+// ref: github.com/sasha-s/go-deadlock
+
+import "sync"
+
+// Map is sync.Map wrapper
+type Map struct {
+	sync.Map
+}

sync/deadlock_test.go

+package gxsync
+
+// ref: github.com/sasha-s/go-deadlock
+
+import (
+	"log"
+	"math/rand"
+	"sync"
+	"sync/atomic"
+	"testing"
+	"time"
+)
+
+func TestNoDeadlocks(t *testing.T) {
+	defer restore()()
+	Opts.DeadlockTimeout = time.Millisecond * 5000
+	var a RWMutex
+	var b Mutex
+	var c RWMutex
+	var wg sync.WaitGroup
+	for i := 0; i < 10; i++ {
+		wg.Add(1)
+		go func() {
+			defer wg.Done()
+			for k := 0; k < 5; k++ {
+				func() {
+					a.Lock()
+					defer a.Unlock()
+					func() {
+						b.Lock()
+						defer b.Unlock()
+						func() {
+							c.RLock()
+							defer c.RUnlock()
+							time.Sleep(time.Duration((1000 + rand.Intn(1000))) * time.Millisecond / 200)
+						}()
+					}()
+				}()
+			}
+		}()
+		wg.Add(1)
+		go func() {
+			defer wg.Done()
+			for k := 0; k < 5; k++ {
+				func() {
+					a.RLock()
+					defer a.RUnlock()
+					func() {
+						b.Lock()
+						defer b.Unlock()
+						func() {
+							c.Lock()
+							defer c.Unlock()
+							time.Sleep(time.Duration((1000 + rand.Intn(1000))) * time.Millisecond / 200)
+						}()
+					}()
+				}()
+			}
+		}()
+	}
+	wg.Wait()
+}
+
+func TestLockOrder(t *testing.T) {
+	defer restore()()
+	Opts.DeadlockTimeout = 0
+	var deadlocks uint32
+	Opts.OnPotentialDeadlock = func() {
+		atomic.AddUint32(&deadlocks, 1)
+	}
+	var a RWMutex
+	var b Mutex
+	var wg sync.WaitGroup
+	wg.Add(1)
+	go func() {
+		defer wg.Done()
+		a.Lock()
+		b.Lock()
+		b.Unlock()
+		a.Unlock()
+	}()
+	wg.Wait()
+	wg.Add(1)
+	go func() {
+		defer wg.Done()
+		b.Lock()
+		a.RLock()
+		a.RUnlock()
+		b.Unlock()
+	}()
+	wg.Wait()
+	if atomic.LoadUint32(&deadlocks) != 1 {
+		t.Fatalf("expected 1 deadlock, detected %d", deadlocks)
+	}
+}
+
+func TestHardDeadlock(t *testing.T) {
+	defer restore()()
+	Opts.DisableLockOrderDetection = true
+	Opts.DeadlockTimeout = time.Millisecond * 20
+	var deadlocks uint32
+	Opts.OnPotentialDeadlock = func() {
+		atomic.AddUint32(&deadlocks, 1)
+	}
+	var mu Mutex
+	mu.Lock()
+	ch := make(chan struct{})
+	go func() {
+		defer close(ch)
+		mu.Lock()
+		defer mu.Unlock()
+	}()
+	select {
+	case <-ch:
+	case <-time.After(time.Millisecond * 100):
+	}
+	if atomic.LoadUint32(&deadlocks) != 1 {
+		t.Fatalf("expected 1 deadlock, detected %d", deadlocks)
+	}
+	log.Println("****************")
+	mu.Unlock()
+	<-ch
+}
+
+func TestRWMutex(t *testing.T) {
+	defer restore()()
+	Opts.DeadlockTimeout = time.Millisecond * 20
+	var deadlocks uint32
+	Opts.OnPotentialDeadlock = func() {
+		atomic.AddUint32(&deadlocks, 1)
+	}
+	var a RWMutex
+	a.RLock()
+	go func() {
+		// We detect a potential deadlock here.
+		a.Lock()
+		defer a.Unlock()
+	}()
+	time.Sleep(time.Millisecond * 100) // We want the Lock call to happen.
+	ch := make(chan struct{})
+	go func() {
+		// We detect a potential deadlock here.
+		defer close(ch)
+		a.RLock()
+		defer a.RUnlock()
+	}()
+	select {
+	case <-ch:
+		t.Fatal("expected a timeout")
+	case <-time.After(time.Millisecond * 50):
+	}
+	a.RUnlock()
+	if atomic.LoadUint32(&deadlocks) != 2 {
+		t.Fatalf("expected 2 deadlocks, detected %d", deadlocks)
+	}
+	<-ch
+}
+
+func restore() func() {
+	opts := Opts
+	return func() {
+		Opts = opts
+	}
+}
+
+func TestLockDuplicate(t *testing.T) {
+	defer restore()()
+	Opts.DeadlockTimeout = 0
+	var deadlocks uint32
+	Opts.OnPotentialDeadlock = func() {
+		atomic.AddUint32(&deadlocks, 1)
+	}
+	var a RWMutex
+	var b Mutex
+	go func() {
+		a.RLock()
+		a.Lock()
+		a.RUnlock()
+		a.Unlock()
+	}()
+	go func() {
+		b.Lock()
+		b.Lock()
+		b.Unlock()
+		b.Unlock()
+	}()
+	time.Sleep(time.Second * 1)
+	if atomic.LoadUint32(&deadlocks) != 2 {
+		t.Fatalf("expected 2 deadlocks, detected %d", deadlocks)
+	}
+}

sync/enable.go

+package gxsync
+
+// ref: github.com/sasha-s/go-deadlock
+
+func EnableDeadlock(enable bool) {
+	Opts.Disable = true
+	Opts.DisableLockOrderDetection = true
+
+	if enable {
+		Opts.Disable = false
+		Opts.DisableLockOrderDetection = false
+	}
+}
+

sync/stacktraces.go

+package gxsync
+
+// ref: github.com/sasha-s/go-deadlock
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"os"
+	"os/user"
+	"path/filepath"
+	"runtime"
+	"strings"
+	"sync"
+)
+
+func callers(skip int) []uintptr {
+	s := make([]uintptr, 50) // Most relevant context seem to appear near the top of the stack.
+	return s[:runtime.Callers(2+skip, s)]
+}
+
+func printStack(w io.Writer, stack []uintptr) {
+	home := os.Getenv("HOME")
+	usr, err := user.Current()
+	if err == nil {
+		home = usr.HomeDir
+	}
+	cwd, _ := os.Getwd()
+
+	for i, pc := range stack {
+		f := runtime.FuncForPC(pc)
+		name := f.Name()
+		pkg := ""
+		if pos := strings.LastIndex(name, "/"); pos >= 0 {
+			name = name[pos+1:]
+		}
+		if pos := strings.Index(name, "."); pos >= 0 {
+			pkg = name[:pos]
+			name = name[pos+1:]
+		}
+		file, line := f.FileLine(pc - 1)
+		if (pkg == "runtime" && name == "goexit") || (pkg == "testing" && name == "tRunner") {
+			fmt.Fprintln(w)
+			return
+		}
+		tail := ""
+		if i == 0 {
+			tail = " <<<<<" // Make the line performing a lock prominent.
+		}
+		// Shorten the file name.
+		clean := file
+		if cwd != "" {
+			cl, err := filepath.Rel(cwd, file)
+			if err == nil {
+				clean = cl
+			}
+		}
+		if home != "" {
+			s2 := strings.Replace(file, home, "~", 1)
+			if len(clean) > len(s2) {
+				clean = s2
+			}
+		}
+		fmt.Fprintf(w, "%s:%d %s.%s %s%s\n", clean, line, pkg, name, code(file, line), tail)
+	}
+	fmt.Fprintln(w)
+}
+
+var fileSources struct {
+	sync.Mutex
+	lines map[string][][]byte
+}
+
+// Reads souce file lines from disk if not cached already.
+func getSourceLines(file string) [][]byte {
+	fileSources.Lock()
+	defer fileSources.Unlock()
+	if fileSources.lines == nil {
+		fileSources.lines = map[string][][]byte{}
+	}
+	if lines, ok := fileSources.lines[file]; ok {
+		return lines
+	}
+	text, _ := ioutil.ReadFile(file)
+	fileSources.lines[file] = bytes.Split(text, []byte{'\n'})
+	return fileSources.lines[file]
+}
+
+func code(file string, line int) string {
+	lines := getSourceLines(file)
+	// lines are 1 based.
+	if line >= len(lines) || line <= 0 {
+		return "???"
+	}
+	return "{ " + string(bytes.TrimSpace(lines[line-1])) + " }"
+}
+
+// Stacktraces for all goroutines.
+func stacks() []byte {
+	buf := make([]byte, 1024*16)
+	for {
+		n := runtime.Stack(buf, true)
+		if n < len(buf) {
+			return buf[:n]
+		}
+		buf = make([]byte, 2*len(buf))
+	}
+}