debugmutex

README

debugmutex

Mutex for debugging deadlocks. It can find non-obvious deadlocks in systems with heavy sync.Mutex usage. I found many deadlocks in Docker with it.

Usage

type Struct struct {
    sync.Locker
}

func New() *Struct {
    locker := &sync.Mutex{}
    if os.Getenv("DEBUG") != "" {
        // will crash program with traceback and file:line where deadlock is
        // occured after five tries to acquire mutex with 1 second gap between.
        locker = debugmutex.New(5, true)
    }
    return &Struct{Locker: locker}
}

For logging github.com/sirupsen/logrus is used. You can set debug logging with

logrus.SetLevel(logrus.DebugLevel)

Documentation

Index

• func New(retries int, fatal bool) sync.Locker
• type Mutex
  • func (m *Mutex) Lock()
  • func (m *Mutex) Unlock()

Functions

func New

func New(retries int, fatal bool) sync.Locker

New returns new debug mutex. retries specify the number of retries acquiring Mutex before error message, if <=0, then retry forever. If fatal is true, then program will exit by os.Exit(1).

Types

type Mutex

type Mutex struct {
	// contains filtered or unexported fields
}

Mutex is debugging mutex, which prints logs and traces for locks. Mutex implements sync.Locker interface.

func (*Mutex) Lock

func (m *Mutex) Lock()

Lock tries to lock mutex. Interval between attempts is 1 second. On each attempt stack trace and file:line of a previous Lock printed. Lock does os.Exit(1) after last attempt if fatal parameter was true.

func (*Mutex) Unlock

func (m *Mutex) Unlock()

Unlock unlocks the mutex. It prints place in the code where it was called and where mutex was locked.