diff --git a/event.go b/event.go
index c21ef6cbb..540dac213 100644
--- a/event.go
+++ b/event.go
@@ -63,3 +63,22 @@ type Event struct {
 type Handler interface {
 	Handle(*Event) error
 }
+
+type commonEvent struct {
+	data *Event
+	sv   *Supervisor
+}
+
+func (e *commonEvent) Handle() {
+	h, ok := e.sv.handlers[e.data.Type]
+	if !ok {
+		e.data.Err <- ErrUnknownEvent
+		return
+	}
+	err := h.Handle(e.data)
+	if err != errDeferedResponse {
+		e.data.Err <- err
+		close(e.data.Err)
+		return
+	}
+}
diff --git a/eventloop/eventloop.go b/eventloop/eventloop.go
new file mode 100644
index 000000000..7a80a273e
--- /dev/null
+++ b/eventloop/eventloop.go
@@ -0,0 +1,51 @@
+package eventloop
+
+import (
+	"runtime"
+	"sync"
+)
+
+// Event is receiving notification from loop with Handle() call.
+type Event interface {
+	Handle()
+}
+
+// EventLoop is interface for event loops.
+// Start starting events processing
+// Send adding event to loop
+type EventLoop interface {
+	Start() error
+	Send(Event) error
+}
+
+// ChanLoop is implementation of EventLoop based on channels.
+type ChanLoop struct {
+	events chan Event
+	once   sync.Once
+}
+
+// NewChanLoop returns ChanLoop with internal channel buffer set to q.
+func NewChanLoop(q int) EventLoop {
+	return &ChanLoop{
+		events: make(chan Event, q),
+	}
+}
+
+// Start starting to read events from channel in separate goroutines.
+// All calls after first is no-op.
+func (el *ChanLoop) Start() error {
+	go el.once.Do(func() {
+		// allocate whole OS thread, so nothing can get scheduled over eventloop
+		runtime.LockOSThread()
+		for ev := range el.events {
+			ev.Handle()
+		}
+	})
+	return nil
+}
+
+// Send sends event to channel. Will block if buffer is full.
+func (el *ChanLoop) Send(ev Event) error {
+	el.events <- ev
+	return nil
+}
diff --git a/eventloop/eventloop_test.go b/eventloop/eventloop_test.go
new file mode 100644
index 000000000..a4fda4708
--- /dev/null
+++ b/eventloop/eventloop_test.go
@@ -0,0 +1,73 @@
+package eventloop
+
+import (
+	"sync"
+	"testing"
+	"time"
+)
+
+type racyEvent struct {
+	m  map[int]struct{}
+	wg *sync.WaitGroup
+}
+
+func (e *racyEvent) Handle() {
+	e.m[0] = struct{}{}
+	e.wg.Done()
+}
+
+func simulateRacyEvents(el EventLoop) {
+	wg := &sync.WaitGroup{}
+	raceMap := make(map[int]struct{})
+	var evs []*racyEvent
+	for i := 0; i < 1024; i++ {
+		wg.Add(1)
+		evs = append(evs, &racyEvent{m: raceMap, wg: wg})
+	}
+	for _, ev := range evs {
+		el.Send(ev)
+	}
+	wg.Wait()
+}
+
+// run with -race
+func TestChanRace(t *testing.T) {
+	e := NewChanLoop(1024)
+	e.Start()
+	simulateRacyEvents(e)
+}
+
+// run with -race
+func TestChanStartTwiceRace(t *testing.T) {
+	e := NewChanLoop(1024)
+	e.Start()
+	e.Start()
+	simulateRacyEvents(e)
+}
+
+type testEvent struct {
+	wg *sync.WaitGroup
+}
+
+func (e *testEvent) Handle() {
+	e.wg.Done()
+}
+
+func TestChanEventSpawn(t *testing.T) {
+	e := NewChanLoop(1024)
+	e.Start()
+	wg := &sync.WaitGroup{}
+	wg.Add(2)
+	e.Send(&testEvent{wg: wg})
+	e.Send(&testEvent{wg: wg})
+	waitCh := make(chan struct{})
+	go func() {
+		wg.Wait()
+		close(waitCh)
+	}()
+	select {
+	case <-waitCh:
+	case <-time.After(1 * time.Second):
+		t.Fatal("Events was not handled in loop")
+	}
+}
diff --git a/supervisor.go b/supervisor.go
index 2ac9a8c39..9406899eb 100644
--- a/supervisor.go
+++ b/supervisor.go
@@ -4,12 +4,12 @@ import (
 	"os"
 	"os/signal"
 	"path/filepath"
-	goruntime "runtime"
 	"sync"
 	"syscall"
 	"time"
 
 	"github.com/Sirupsen/logrus"
+	"github.com/docker/containerd/eventloop"
 	"github.com/docker/containerd/runtime"
 	"github.com/opencontainers/runc/libcontainer"
 )
@@ -36,10 +36,10 @@ func NewSupervisor(id, stateDir string, tasks chan *StartTask, oom bool) (*Super
 		processes:      make(map[int]*containerInfo),
 		runtime:        r,
 		tasks:          tasks,
-		events:         make(chan *Event, DefaultBufferSize),
 		machine:        machine,
 		subscribers:    make(map[chan *Event]struct{}),
 		statsCollector: newStatsCollector(statsInterval),
+		el:             eventloop.NewChanLoop(DefaultBufferSize),
 	}
 	if oom {
 		s.notifier = newNotifier(s)
@@ -86,6 +86,7 @@ type Supervisor struct {
 	containerGroup sync.WaitGroup
 	statsCollector *statsCollector
 	notifier       *notifier
+	el             eventloop.EventLoop
 }
 
 // Stop closes all tasks and sends a SIGTERM to each container's pid1 then waits for they to
@@ -174,32 +175,11 @@ func (s *Supervisor) notifySubscribers(e *Event) {
 // therefore it is save to do operations in the handlers that modify state of the system or
 // state of the Supervisor
 func (s *Supervisor) Start() error {
-	go func() {
-		// allocate an entire thread to this goroutine for the main event loop
-		// so that nothing else is scheduled over the top of it.
-		goruntime.LockOSThread()
-		for e := range s.events {
-			EventsCounter.Inc(1)
-			h, ok := s.handlers[e.Type]
-			if !ok {
-				e.Err <- ErrUnknownEvent
-				continue
-			}
-			if err := h.Handle(e); err != nil {
-				if err != errDeferedResponse {
-					e.Err <- err
-					close(e.Err)
-				}
-				continue
-			}
-			close(e.Err)
-		}
-	}()
 	logrus.WithFields(logrus.Fields{
 		"runtime":  s.runtime.Type(),
 		"stateDir": s.stateDir,
 	}).Debug("Supervisor started")
-	return nil
+	return s.el.Start()
 }
 
 // Machine returns the machine information for which the
@@ -231,7 +211,8 @@ func (s *Supervisor) getContainerForPid(pid int) (runtime.Container, error) {
 
 // SendEvent sends the provided event the the supervisors main event loop
 func (s *Supervisor) SendEvent(evt *Event) {
-	s.events <- evt
+	EventsCounter.Inc(1)
+	s.el.Send(&commonEvent{data: evt, sv: s})
 }
 
 func (s *Supervisor) copyIO(stdin, stdout, stderr string, i *runtime.IO) (*copier, error) {