diff --git a/trace/icmp_ipv4.go b/trace/icmp_ipv4.go
index d6d61262..0d796886 100644
--- a/trace/icmp_ipv4.go
+++ b/trace/icmp_ipv4.go
@@ -10,27 +10,17 @@ import (
 	"golang.org/x/net/context"
 	"golang.org/x/net/icmp"
 	"golang.org/x/net/ipv4"
-	"golang.org/x/sync/semaphore"
 )
 
 type ICMPTracer struct {
 	Config
-	wg                  sync.WaitGroup
-	res                 Result
-	ctx                 context.Context
-	inflightRequest     map[int]chan Hop
-	inflightRequestLock sync.Mutex
-	icmpListen          net.PacketConn
-	workFork            workFork
-	final               int
-	finalLock           sync.Mutex
-
-	sem *semaphore.Weighted
-}
-
-type workFork struct {
-	ttl int
-	num int
+	wg         sync.WaitGroup
+	res        Result
+	ctx        context.Context
+	resCh      chan Hop
+	icmpListen net.PacketConn
+	final      int
+	finalLock  sync.Mutex
 }
 
 func (t *ICMPTracer) Execute() (*Result, error) {
@@ -49,24 +39,24 @@ func (t *ICMPTracer) Execute() (*Result, error) {
 	var cancel context.CancelFunc
 	t.ctx, cancel = context.WithCancel(context.Background())
 	defer cancel()
-	t.inflightRequest = make(map[int]chan Hop)
+	t.resCh = make(chan Hop)
 	t.final = -1
 
 	go t.listenICMP()
 
-	t.sem = semaphore.NewWeighted(int64(t.ParallelRequests))
-
-	for t.workFork.ttl = 1; t.workFork.ttl <= t.MaxHops; t.workFork.ttl++ {
+	for ttl := 1; ttl <= t.MaxHops; ttl++ {
+		if t.final != -1 && ttl > t.final {
+			break
+		}
 		for i := 0; i < t.NumMeasurements; i++ {
 			t.wg.Add(1)
-			go t.send(workFork{t.workFork.ttl, i})
+			go t.send(ttl)
 		}
 		// 一组TTL全部退出（收到应答或者超时终止）以后，再进行下一个TTL的包发送
 		t.wg.Wait()
 		if t.RealtimePrinter != nil {
-			t.RealtimePrinter(&t.res, t.workFork.ttl-1)
+			t.RealtimePrinter(&t.res, ttl-1)
 		}
-		t.workFork.num = 0
 	}
 	t.res.reduce(t.final)
 
@@ -103,29 +93,15 @@ func (t *ICMPTracer) listenICMP() {
 }
 
 func (t *ICMPTracer) handleICMPMessage(msg ReceivedMessage, icmpType int8, data []byte) {
-
-	t.inflightRequestLock.Lock()
-	defer t.inflightRequestLock.Unlock()
-	ch, ok := t.inflightRequest[t.workFork.num]
-	t.workFork.num += 1
-	if !ok {
-		return
-	}
-	ch <- Hop{
+	t.resCh <- Hop{
 		Success: true,
 		Address: msg.Peer,
 	}
 }
 
-func (t *ICMPTracer) send(fork workFork) error {
-	err := t.sem.Acquire(context.Background(), 1)
-	if err != nil {
-		return err
-	}
-	defer t.sem.Release(1)
-
+func (t *ICMPTracer) send(ttl int) error {
 	defer t.wg.Done()
-	if t.final != -1 && fork.ttl > t.final {
+	if t.final != -1 && ttl > t.final {
 		return nil
 	}
 
@@ -137,7 +113,7 @@ func (t *ICMPTracer) send(fork workFork) error {
 		},
 	}
 
-	ipv4.NewPacketConn(t.icmpListen).SetTTL(fork.ttl)
+	ipv4.NewPacketConn(t.icmpListen).SetTTL(ttl)
 
 	wb, err := icmpHeader.Marshal(nil)
 	if err != nil {
@@ -151,41 +127,30 @@ func (t *ICMPTracer) send(fork workFork) error {
 	if err := t.icmpListen.SetReadDeadline(time.Now().Add(3 * time.Second)); err != nil {
 		log.Fatal(err)
 	}
-	t.inflightRequestLock.Lock()
-	hopCh := make(chan Hop)
-	t.inflightRequest[fork.num] = hopCh
-	t.inflightRequestLock.Unlock()
-
-	// defer func() {
-	// 	t.inflightRequestLock.Lock()
-	// 	close(hopCh)
-	// 	delete(t.inflightRequest, fork.ttl)
-	// 	t.inflightRequestLock.Unlock()
-	// }()
 
 	select {
 	case <-t.ctx.Done():
 		return nil
-	case h := <-hopCh:
+	case h := <-t.resCh:
 		rtt := time.Since(start)
-		if t.final != -1 && fork.ttl > t.final {
+		if t.final != -1 && ttl > t.final {
 			return nil
 		}
 		if addr, ok := h.Address.(*net.IPAddr); ok && addr.IP.Equal(t.DestIP) {
 			t.finalLock.Lock()
-			if t.final == -1 || fork.ttl < t.final {
-				t.final = fork.ttl
+			if t.final == -1 || ttl < t.final {
+				t.final = ttl
 			}
 			t.finalLock.Unlock()
 		} else if addr, ok := h.Address.(*net.TCPAddr); ok && addr.IP.Equal(t.DestIP) {
 			t.finalLock.Lock()
-			if t.final == -1 || fork.ttl < t.final {
-				t.final = fork.ttl
+			if t.final == -1 || ttl < t.final {
+				t.final = ttl
 			}
 			t.finalLock.Unlock()
 		}
 
-		h.TTL = fork.ttl
+		h.TTL = ttl
 		h.RTT = rtt
 
 		h.fetchIPData(t.Config)
@@ -193,14 +158,14 @@ func (t *ICMPTracer) send(fork workFork) error {
 		t.res.add(h)
 
 	case <-time.After(t.Timeout):
-		if t.final != -1 && fork.ttl > t.final {
+		if t.final != -1 && ttl > t.final {
 			return nil
 		}
 
 		t.res.add(Hop{
 			Success: false,
 			Address: nil,
-			TTL:     fork.ttl,
+			TTL:     ttl,
 			RTT:     0,
 			Error:   ErrHopLimitTimeout,
 		})
diff --git a/trace/icmp_ipv6.go b/trace/icmp_ipv6.go
index 0f365777..d6849d44 100644
--- a/trace/icmp_ipv6.go
+++ b/trace/icmp_ipv6.go
@@ -10,22 +10,17 @@ import (
 	"golang.org/x/net/context"
 	"golang.org/x/net/icmp"
 	"golang.org/x/net/ipv6"
-	"golang.org/x/sync/semaphore"
 )
 
 type ICMPTracerv6 struct {
 	Config
-	wg                  sync.WaitGroup
-	res                 Result
-	ctx                 context.Context
-	inflightRequest     map[int]chan Hop
-	inflightRequestLock sync.Mutex
-	icmpListen          net.PacketConn
-	workFork            workFork
-	final               int
-	finalLock           sync.Mutex
-
-	sem *semaphore.Weighted
+	wg         sync.WaitGroup
+	res        Result
+	ctx        context.Context
+	resCh      chan Hop
+	icmpListen net.PacketConn
+	final      int
+	finalLock  sync.Mutex
 }
 
 func (t *ICMPTracerv6) Execute() (*Result, error) {
@@ -44,21 +39,24 @@ func (t *ICMPTracerv6) Execute() (*Result, error) {
 	var cancel context.CancelFunc
 	t.ctx, cancel = context.WithCancel(context.Background())
 	defer cancel()
-	t.inflightRequest = make(map[int]chan Hop)
+	t.resCh = make(chan Hop)
 	t.final = -1
 
 	go t.listenICMP()
 
-	t.sem = semaphore.NewWeighted(int64(t.ParallelRequests))
-
-	for t.workFork.ttl = 1; t.workFork.ttl <= t.MaxHops; t.workFork.ttl++ {
+	for ttl := 1; ttl <= t.MaxHops; ttl++ {
+		if t.final != -1 && ttl > t.final {
+			break
+		}
 		for i := 0; i < t.NumMeasurements; i++ {
 			t.wg.Add(1)
-			go t.send(workFork{t.workFork.ttl, i})
+			go t.send(ttl)
 		}
 		// 一组TTL全部退出（收到应答或者超时终止）以后，再进行下一个TTL的包发送
 		t.wg.Wait()
-		t.workFork.num = 0
+		if t.RealtimePrinter != nil {
+			t.RealtimePrinter(&t.res, ttl-1)
+		}
 	}
 	t.res.reduce(t.final)
 
@@ -96,29 +94,15 @@ func (t *ICMPTracerv6) listenICMP() {
 }
 
 func (t *ICMPTracerv6) handleICMPMessage(msg ReceivedMessage, icmpType int8, data []byte) {
-	t.inflightRequestLock.Lock()
-	defer t.inflightRequestLock.Unlock()
-	ch, ok := t.inflightRequest[t.workFork.num]
-	t.workFork.num += 1
-	if !ok {
-		return
-	}
-	ch <- Hop{
+	t.resCh <- Hop{
 		Success: true,
 		Address: msg.Peer,
 	}
 }
 
-func (t *ICMPTracerv6) send(fork workFork) error {
-	err := t.sem.Acquire(context.Background(), 1)
-	if err != nil {
-		return err
-	}
-
-	defer t.sem.Release(1)
-
+func (t *ICMPTracerv6) send(ttl int) error {
 	defer t.wg.Done()
-	if t.final != -1 && fork.ttl > t.final {
+	if t.final != -1 && ttl > t.final {
 		return nil
 	}
 
@@ -132,8 +116,8 @@ func (t *ICMPTracerv6) send(fork workFork) error {
 
 	p := ipv6.NewPacketConn(t.icmpListen)
 
-	icmpHeader.Body.(*icmp.Echo).Seq = fork.ttl
-	p.SetHopLimit(fork.ttl)
+	icmpHeader.Body.(*icmp.Echo).Seq = ttl
+	p.SetHopLimit(ttl)
 
 	wb, err := icmpHeader.Marshal(nil)
 	if err != nil {
@@ -147,41 +131,30 @@ func (t *ICMPTracerv6) send(fork workFork) error {
 	if err := t.icmpListen.SetReadDeadline(time.Now().Add(3 * time.Second)); err != nil {
 		log.Fatal(err)
 	}
-	t.inflightRequestLock.Lock()
-	hopCh := make(chan Hop)
-	t.inflightRequest[fork.num] = hopCh
-	t.inflightRequestLock.Unlock()
-
-	// defer func() {
-	// 	t.inflightRequestLock.Lock()
-	// 	close(hopCh)
-	// 	delete(t.inflightRequest, fork.ttl)
-	// 	t.inflightRequestLock.Unlock()
-	// }()
 
 	select {
 	case <-t.ctx.Done():
 		return nil
-	case h := <-hopCh:
+	case h := <-t.resCh:
 		rtt := time.Since(start)
-		if t.final != -1 && fork.ttl > t.final {
+		if t.final != -1 && ttl > t.final {
 			return nil
 		}
 		if addr, ok := h.Address.(*net.IPAddr); ok && addr.IP.Equal(t.DestIP) {
 			t.finalLock.Lock()
-			if t.final == -1 || fork.ttl < t.final {
-				t.final = fork.ttl
+			if t.final == -1 || ttl < t.final {
+				t.final = ttl
 			}
 			t.finalLock.Unlock()
 		} else if addr, ok := h.Address.(*net.TCPAddr); ok && addr.IP.Equal(t.DestIP) {
 			t.finalLock.Lock()
-			if t.final == -1 || fork.ttl < t.final {
-				t.final = fork.ttl
+			if t.final == -1 || ttl < t.final {
+				t.final = ttl
 			}
 			t.finalLock.Unlock()
 		}
 
-		h.TTL = fork.ttl
+		h.TTL = ttl
 		h.RTT = rtt
 
 		h.fetchIPData(t.Config)
@@ -189,14 +162,14 @@ func (t *ICMPTracerv6) send(fork workFork) error {
 		t.res.add(h)
 
 	case <-time.After(t.Timeout):
-		if t.final != -1 && fork.ttl > t.final {
+		if t.final != -1 && ttl > t.final {
 			return nil
 		}
 
 		t.res.add(Hop{
 			Success: false,
 			Address: nil,
-			TTL:     fork.ttl,
+			TTL:     ttl,
 			RTT:     0,
 			Error:   ErrHopLimitTimeout,
 		})