[VOL-1035] Initial submission of flow decomposition code.
Additional test cases will follow to test the core of the flow
decomposition functionality

Change-Id: Ie685714ce5ab54ac89501a67f9489613de195c15
diff --git a/rw_core/flow_decomposition/flow_decomposer_test.go b/rw_core/flow_decomposition/flow_decomposer_test.go
new file mode 100644
index 0000000..b2ba777
--- /dev/null
+++ b/rw_core/flow_decomposition/flow_decomposer_test.go
@@ -0,0 +1,665 @@
+/*
+ * Copyright 2018-present Open Networking Foundation
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package flow_decomposition
+
+import (
+	"errors"
+	"github.com/opencord/voltha-go/common/log"
+	ofp "github.com/opencord/voltha-go/protos/openflow_13"
+	"github.com/opencord/voltha-go/protos/voltha"
+	"github.com/opencord/voltha-go/rw_core/graph"
+	fu "github.com/opencord/voltha-go/rw_core/utils"
+	"github.com/stretchr/testify/assert"
+
+	"testing"
+)
+
+const (
+	maxOnuOnPort4 int = 1
+	maxOnuOnPort5 int = 1
+)
+
+func init() {
+	log.AddPackage(log.JSON, log.DebugLevel, nil)
+	log.UpdateAllLoggers(log.Fields{"instanceId": "flow-descomposition"})
+	log.SetAllLogLevel(log.DebugLevel)
+}
+
+type testDeviceManager struct {
+	devices map[string]*voltha.Device
+}
+
+func newTestDeviceManager() *testDeviceManager {
+	var tdm testDeviceManager
+	tdm.devices = make(map[string]*voltha.Device)
+	tdm.devices["olt"] = &voltha.Device{
+		Id:       "olt",
+		Root:     true,
+		ParentId: "logical_device",
+		Ports: []*voltha.Port{
+			&voltha.Port{PortNo: 1, Label: "pon"},
+			&voltha.Port{PortNo: 2, Label: "nni"},
+		},
+	}
+	tdm.devices["onu1"] = &voltha.Device{
+		Id:       "onu1",
+		Root:     false,
+		ParentId: "olt",
+		Ports: []*voltha.Port{
+			&voltha.Port{PortNo: 1, Label: "pon"},
+			&voltha.Port{PortNo: 2, Label: "uni"},
+		},
+	}
+	tdm.devices["onu2"] = &voltha.Device{
+		Id:       "onu2",
+		Root:     false,
+		ParentId: "olt",
+		Ports: []*voltha.Port{
+			&voltha.Port{PortNo: 1, Label: "pon"},
+			&voltha.Port{PortNo: 2, Label: "uni"},
+		},
+	}
+	tdm.devices["onu3"] = &voltha.Device{
+		Id:       "onu3",
+		Root:     false,
+		ParentId: "olt",
+		Ports: []*voltha.Port{
+			&voltha.Port{PortNo: 1, Label: "pon"},
+			&voltha.Port{PortNo: 2, Label: "uni"},
+		},
+	}
+	tdm.devices["onu4"] = &voltha.Device{
+		Id:       "onu4",
+		Root:     false,
+		ParentId: "olt",
+		Ports: []*voltha.Port{
+			&voltha.Port{PortNo: 1, Label: "pon"},
+			&voltha.Port{PortNo: 2, Label: "uni"},
+		},
+	}
+	return &tdm
+}
+
+func (tdm *testDeviceManager) GetDevice(deviceId string) (*voltha.Device, error) {
+	if d, ok := tdm.devices[deviceId]; ok {
+		return d, nil
+	}
+	return nil, errors.New("Absent")
+}
+
+type testFlowDecomposer struct {
+	dMgr         *testDeviceManager
+	logicalPorts map[uint32]*voltha.LogicalPort
+	routes       map[graph.OFPortLink][]graph.RouteHop
+	defaultRules *fu.DeviceRules
+	deviceGraph  *graph.DeviceGraph
+	fd           *FlowDecomposer
+}
+
+func newTestFlowDecomposer(deviceMgr *testDeviceManager) *testFlowDecomposer {
+	var tfd testFlowDecomposer
+	tfd.dMgr = deviceMgr
+
+	tfd.logicalPorts = make(map[uint32]*voltha.LogicalPort)
+	// Go protobuf interpreted absence of a port as 0, so we can't use port #0 as an openflow
+	// port
+	tfd.logicalPorts[10] = &voltha.LogicalPort{Id: "10", DeviceId: "olt", DevicePortNo: 2}
+	tfd.logicalPorts[1] = &voltha.LogicalPort{Id: "1", DeviceId: "onu1", DevicePortNo: 2}
+	tfd.logicalPorts[2] = &voltha.LogicalPort{Id: "2", DeviceId: "onu2", DevicePortNo: 2}
+	tfd.logicalPorts[3] = &voltha.LogicalPort{Id: "3", DeviceId: "onu3", DevicePortNo: 2}
+	tfd.logicalPorts[4] = &voltha.LogicalPort{Id: "4", DeviceId: "onu4", DevicePortNo: 2}
+
+	tfd.routes = make(map[graph.OFPortLink][]graph.RouteHop)
+
+	//DOWNSTREAM ROUTES
+
+	tfd.routes[graph.OFPortLink{Ingress: 10, Egress: 1}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "onu1",
+			Ingress:  tfd.dMgr.devices["onu1"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["onu1"].Ports[1].PortNo,
+		},
+	}
+
+	tfd.routes[graph.OFPortLink{Ingress: 10, Egress: 2}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "onu2",
+			Ingress:  tfd.dMgr.devices["onu2"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["onu2"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 10, Egress: 3}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "onu3",
+			Ingress:  tfd.dMgr.devices["onu3"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["onu3"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 10, Egress: 4}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "onu4",
+			Ingress:  tfd.dMgr.devices["onu4"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["onu4"].Ports[1].PortNo,
+		},
+	}
+
+	//UPSTREAM DATA PLANE
+
+	tfd.routes[graph.OFPortLink{Ingress: 1, Egress: 10}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu1",
+			Ingress:  tfd.dMgr.devices["onu1"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu1"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 2, Egress: 10}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu2",
+			Ingress:  tfd.dMgr.devices["onu2"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu2"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 3, Egress: 10}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu3",
+			Ingress:  tfd.dMgr.devices["onu3"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu3"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 4, Egress: 10}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu4",
+			Ingress:  tfd.dMgr.devices["onu4"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu4"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+
+	//UPSTREAM NEXT TABLE BASED
+
+	// openflow port 0 means absence of a port - go/protobuf interpretation
+	tfd.routes[graph.OFPortLink{Ingress: 1, Egress: 0}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu1",
+			Ingress:  tfd.dMgr.devices["onu1"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu1"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 2, Egress: 0}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu2",
+			Ingress:  tfd.dMgr.devices["onu2"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu2"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 3, Egress: 0}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu3",
+			Ingress:  tfd.dMgr.devices["onu3"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu3"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+	tfd.routes[graph.OFPortLink{Ingress: 4, Egress: 0}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "onu4",
+			Ingress:  tfd.dMgr.devices["onu4"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["onu4"].Ports[0].PortNo,
+		},
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+
+	// DOWNSTREAM NEXT TABLE BASED
+
+	tfd.routes[graph.OFPortLink{Ingress: 10, Egress: 0}] = []graph.RouteHop{
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[1].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[0].PortNo,
+		},
+		graph.RouteHop{}, // 2nd hop is not known yet
+	}
+
+	tfd.routes[graph.OFPortLink{Ingress: 0, Egress: 10}] = []graph.RouteHop{
+		graph.RouteHop{}, // 1st hop is wildcard
+		graph.RouteHop{
+			DeviceID: "olt",
+			Ingress:  tfd.dMgr.devices["olt"].Ports[0].PortNo,
+			Egress:   tfd.dMgr.devices["olt"].Ports[1].PortNo,
+		},
+	}
+
+	// DEFAULT RULES
+
+	tfd.defaultRules = fu.NewDeviceRules()
+	fg := fu.NewFlowsAndGroups()
+	var fa *fu.FlowArgs
+	fa = &fu.FlowArgs{
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+		},
+		Actions: []*ofp.OfpAction{
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101)),
+			Output(1),
+		},
+	}
+	fg.AddFlow(MkFlowStat(fa))
+	tfd.defaultRules.AddFlowsAndGroup("onu1", fg)
+
+	fg = fu.NewFlowsAndGroups()
+	fa = &fu.FlowArgs{
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+		},
+		Actions: []*ofp.OfpAction{
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 102)),
+			Output(1),
+		},
+	}
+	fg.AddFlow(MkFlowStat(fa))
+	tfd.defaultRules.AddFlowsAndGroup("onu2", fg)
+
+	fg = fu.NewFlowsAndGroups()
+	fa = &fu.FlowArgs{
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+		},
+		Actions: []*ofp.OfpAction{
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 103)),
+			Output(1),
+		},
+	}
+	fg.AddFlow(MkFlowStat(fa))
+	tfd.defaultRules.AddFlowsAndGroup("onu3", fg)
+
+	fg = fu.NewFlowsAndGroups()
+	fa = &fu.FlowArgs{
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+		},
+		Actions: []*ofp.OfpAction{
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 104)),
+			Output(1),
+		},
+	}
+	fg.AddFlow(MkFlowStat(fa))
+	tfd.defaultRules.AddFlowsAndGroup("onu4", fg)
+
+	//Set up the device graph - flow decomposer uses it only to verify whether a port is a root port.
+	tfd.deviceGraph = graph.NewDeviceGraph(tfd.getDeviceHelper)
+	tfd.deviceGraph.RootPorts = make(map[uint32]uint32)
+	tfd.deviceGraph.RootPorts[10] = 10
+
+	tfd.fd = NewFlowDecomposer(tfd.dMgr)
+
+	return &tfd
+}
+
+func (tfd *testFlowDecomposer) getDeviceHelper(deviceId string) (*voltha.Device, error) {
+	return tfd.dMgr.GetDevice(deviceId)
+}
+
+func (tfd *testFlowDecomposer) GetDeviceLogicalId() string {
+	return ""
+}
+
+func (tfd *testFlowDecomposer) GetLogicalDevice() *voltha.LogicalDevice {
+	return nil
+}
+
+func (tfd *testFlowDecomposer) GetDeviceGraph() *graph.DeviceGraph {
+	return tfd.deviceGraph
+}
+
+func (tfd *testFlowDecomposer) GetAllDefaultRules() *fu.DeviceRules {
+	return tfd.defaultRules
+}
+
+func (tfd *testFlowDecomposer) GetWildcardInputPorts(excludePort ...uint32) []uint32 {
+	lPorts := make([]uint32, 0)
+	var exclPort uint32
+	if len(excludePort) == 1 {
+		exclPort = excludePort[0]
+	}
+	for portno, _ := range tfd.logicalPorts {
+		if portno != exclPort {
+			lPorts = append(lPorts, portno)
+		}
+	}
+	return lPorts
+}
+
+func (tfd *testFlowDecomposer) GetRoute(ingressPortNo *uint32, egressPortNo *uint32) []graph.RouteHop {
+	var portLink graph.OFPortLink
+	if egressPortNo == nil {
+		portLink.Egress = 0
+	} else if *egressPortNo&0x7fffffff == uint32(ofp.OfpPortNo_OFPP_CONTROLLER) {
+		portLink.Egress = 10
+	} else {
+		portLink.Egress = *egressPortNo
+	}
+	if ingressPortNo == nil {
+		portLink.Ingress = 0
+	} else {
+		portLink.Ingress = *ingressPortNo
+	}
+	for key, val := range tfd.routes {
+		if key.Ingress == portLink.Ingress && key.Egress == portLink.Egress {
+			return val
+		}
+	}
+	return nil
+}
+
+func TestEapolReRouteRuleDecomposition(t *testing.T) {
+
+	var fa *fu.FlowArgs
+	fa = &fu.FlowArgs{
+		KV: fu.OfpFlowModArgs{"priority": 1000},
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(1),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+			EthType(0x888e),
+		},
+		Actions: []*ofp.OfpAction{
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101)),
+			Output(uint32(ofp.OfpPortNo_OFPP_CONTROLLER)),
+		},
+	}
+
+	flows := ofp.Flows{Items: []*ofp.OfpFlowStats{MkFlowStat(fa)}}
+	groups := ofp.FlowGroups{}
+	tfd := newTestFlowDecomposer(newTestDeviceManager())
+
+	device_rules := tfd.fd.DecomposeRules(tfd, flows, groups)
+	onu1FlowAndGroup := device_rules.Rules["onu1"]
+	oltFlowAndGroup := device_rules.Rules["olt"]
+	assert.Equal(t, 1, onu1FlowAndGroup.Flows.Len())
+	assert.Equal(t, 0, onu1FlowAndGroup.Groups.Len())
+	assert.Equal(t, 2, oltFlowAndGroup.Flows.Len())
+	assert.Equal(t, 0, oltFlowAndGroup.Groups.Len())
+
+	fa = &fu.FlowArgs{
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+		},
+		Actions: []*ofp.OfpAction{
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101)),
+			Output(1),
+		},
+	}
+	expectedOnu1Flow := MkFlowStat(fa)
+	derivedFlow := onu1FlowAndGroup.GetFlow(0)
+	assert.Equal(t, expectedOnu1Flow.String(), derivedFlow.String())
+
+	fa = &fu.FlowArgs{
+		KV: fu.OfpFlowModArgs{"priority": 1000},
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(1),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 1),
+			EthType(0x888e),
+		},
+		Actions: []*ofp.OfpAction{
+			PushVlan(0x8100),
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 4000)),
+			Output(2),
+		},
+	}
+	expectedOltFlow := MkFlowStat(fa)
+	derivedFlow = oltFlowAndGroup.GetFlow(0)
+	assert.Equal(t, expectedOltFlow.String(), derivedFlow.String())
+
+	fa = &fu.FlowArgs{
+		KV: fu.OfpFlowModArgs{"priority": 1000},
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 4000),
+			VlanPcp(0),
+			Metadata_ofp(1),
+		},
+		Actions: []*ofp.OfpAction{
+			PopVlan(),
+			Output(1),
+		},
+	}
+	expectedOltFlow = MkFlowStat(fa)
+	derivedFlow = oltFlowAndGroup.GetFlow(1)
+	assert.Equal(t, expectedOltFlow.String(), derivedFlow.String())
+}
+
+func TestDhcpReRouteRuleDecomposition(t *testing.T) {
+
+	var fa *fu.FlowArgs
+	fa = &fu.FlowArgs{
+		KV: fu.OfpFlowModArgs{"priority": 1000},
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(1),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+			EthType(0x0800),
+			Ipv4Dst(0xffffffff),
+			IpProto(17),
+			UdpSrc(68),
+			UdpDst(67),
+		},
+		Actions: []*ofp.OfpAction{
+			Output(uint32(ofp.OfpPortNo_OFPP_CONTROLLER)),
+		},
+	}
+
+	flows := ofp.Flows{Items: []*ofp.OfpFlowStats{MkFlowStat(fa)}}
+	groups := ofp.FlowGroups{}
+	tfd := newTestFlowDecomposer(newTestDeviceManager())
+
+	device_rules := tfd.fd.DecomposeRules(tfd, flows, groups)
+	onu1FlowAndGroup := device_rules.Rules["onu1"]
+	oltFlowAndGroup := device_rules.Rules["olt"]
+	assert.Equal(t, 1, onu1FlowAndGroup.Flows.Len())
+	assert.Equal(t, 0, onu1FlowAndGroup.Groups.Len())
+	assert.Equal(t, 2, oltFlowAndGroup.Flows.Len())
+	assert.Equal(t, 0, oltFlowAndGroup.Groups.Len())
+
+	fa = &fu.FlowArgs{
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+		},
+		Actions: []*ofp.OfpAction{
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101)),
+			Output(1),
+		},
+	}
+	expectedOnu1Flow := MkFlowStat(fa)
+	derivedFlow := onu1FlowAndGroup.GetFlow(0)
+	assert.Equal(t, expectedOnu1Flow.String(), derivedFlow.String())
+
+	fa = &fu.FlowArgs{
+		KV: fu.OfpFlowModArgs{"priority": 1000},
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(1),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 1),
+			EthType(0x0800),
+			Ipv4Dst(0xffffffff),
+			IpProto(17),
+			UdpSrc(68),
+			UdpDst(67),
+		},
+		Actions: []*ofp.OfpAction{
+			PushVlan(0x8100),
+			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 4000)),
+			Output(2),
+		},
+	}
+	expectedOltFlow := MkFlowStat(fa)
+	derivedFlow = oltFlowAndGroup.GetFlow(0)
+	assert.Equal(t, expectedOltFlow.String(), derivedFlow.String())
+
+	fa = &fu.FlowArgs{
+		KV: fu.OfpFlowModArgs{"priority": 1000},
+		MatchFields: []*ofp.OfpOxmOfbField{
+			InPort(2),
+			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 4000),
+			VlanPcp(0),
+			Metadata_ofp(1),
+		},
+		Actions: []*ofp.OfpAction{
+			PopVlan(),
+			Output(1),
+		},
+	}
+	expectedOltFlow = MkFlowStat(fa)
+	derivedFlow = oltFlowAndGroup.GetFlow(1)
+	assert.Equal(t, expectedOltFlow.String(), derivedFlow.String())
+}
+
+//func TestUnicastUpstreamRuleDecomposition(t *testing.T) {
+//
+//	var fa *fu.FlowArgs
+//	fa = &fu.FlowArgs{
+//		KV: fu.OfpFlowModArgs{"priority": 500, "table_id":1},
+//		MatchFields: []*ofp.OfpOxmOfbField{
+//			InPort(1),
+//			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+//			VlanPcp(0),
+//		},
+//		Actions: []*ofp.OfpAction{
+//			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101)),
+//		},
+//	}
+//
+//	var fa2 *fu.FlowArgs
+//	fa2 = &fu.FlowArgs{
+//		KV: fu.OfpFlowModArgs{"priority": 500},
+//		MatchFields: []*ofp.OfpOxmOfbField{
+//			InPort(1),
+//			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101),
+//			VlanPcp(0),
+//		},
+//		Actions: []*ofp.OfpAction{
+//			PushVlan(0x8100),
+//			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 1000)),
+//			SetField(VlanPcp(0)),
+//			Output(10),
+//		},
+//	}
+//
+//	flows := ofp.Flows{Items:[]*ofp.OfpFlowStats{MkFlowStat(fa), MkFlowStat(fa2)}}
+//	groups := ofp.FlowGroups{}
+//	tfd := newTestFlowDecomposer(newTestDeviceManager())
+//
+//	device_rules := tfd.fd.DecomposeRules(tfd, flows, groups)
+//	onu1FlowAndGroup := device_rules.Rules["onu1"]
+//	oltFlowAndGroup := device_rules.Rules["olt"]
+//	assert.Equal(t, 2, onu1FlowAndGroup.Flows.Len())
+//	assert.Equal(t, 0, onu1FlowAndGroup.Groups.Len())
+//	assert.Equal(t, 1, oltFlowAndGroup.Flows.Len())
+//	assert.Equal(t, 0, oltFlowAndGroup.Groups.Len())
+//
+//	fa = &fu.FlowArgs{
+//		KV: fu.OfpFlowModArgs{"priority": 500},
+//		MatchFields: []*ofp.OfpOxmOfbField{
+//			InPort(2),
+//			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 0),
+//			VlanPcp(0),
+//		},
+//		Actions: []*ofp.OfpAction{
+//			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101)),
+//			Output(1),
+//		},
+//	}
+//	expectedOnu1Flow := MkFlowStat(fa)
+//	derivedFlow := onu1FlowAndGroup.GetFlow(1)
+//	assert.Equal(t, expectedOnu1Flow.String(), derivedFlow.String())
+//
+//	fa = &fu.FlowArgs{
+//		KV: fu.OfpFlowModArgs{"priority": 500},
+//		MatchFields: []*ofp.OfpOxmOfbField{
+//			InPort(1),
+//			VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 101),
+//			VlanPcp(0),
+//		},
+//		Actions: []*ofp.OfpAction{
+//			PushVlan(0x8100),
+//			SetField(VlanVid(uint32(ofp.OfpVlanId_OFPVID_PRESENT) | 1000)),
+//			SetField(VlanPcp(0)),
+//			Output(2),
+//		},
+//	}
+//	expectedOltFlow := MkFlowStat(fa)
+//	derivedFlow = oltFlowAndGroup.GetFlow(0)
+//	assert.Equal(t, expectedOltFlow.String(), derivedFlow.String())
+//}