Address comments

Address comments and add more validation UT cases. Signed-off-by: wgrayson <[email protected]>
antrea-io · Aug 2, 2022 · 090b3a6 · 090b3a6
1 parent 8320584
commit 090b3a6
Show file tree

Hide file tree

Showing 9 changed files with 277 additions and 98 deletions.
diff --git a/build/yamls/antrea.yml b/build/yamls/antrea.yml
@@ -2767,7 +2767,7 @@ data:
 
     # Provide the address of Kubernetes apiserver, to override any value provided in kubeconfig or InClusterConfig.
     # Defaults to "". It must be a host string, a host:port pair, or a URL to the base of the apiserver.
-    kubeAPIServerOverride: ""
+    kubeAPIServerOverride: "https://192.168.77.100:6443"
 
     # Provide the address of DNS server, to override the kube-dns service. It's used to resolve hostname in FQDN policy.
     # Defaults to "". It must be a host string or a host:port pair of the DNS server (e.g. 10.96.0.10, 10.96.0.10:53,
@@ -2817,7 +2817,7 @@ data:
       # feature to be enabled.
       # Note that this option is experimental. If kube-proxy is removed, option kubeAPIServerOverride must be used to access
       # apiserver directly.
-      proxyAll: false
+      proxyAll: true
       # A string array of values which specifies the host IPv4/IPv6 addresses for NodePort. Values can be valid IP blocks.
       # (e.g. 1.2.3.0/24, 1.2.3.4/32). An empty string slice is meant to select all host IPv4/IPv6 addresses.
       # Note that the option is only valid when proxyAll is true.
@@ -3712,7 +3712,7 @@ spec:
         kubectl.kubernetes.io/default-container: antrea-agent
         # Automatically restart Pods with a RollingUpdate if the ConfigMap changes
         # See https://helm.sh/docs/howto/charts_tips_and_tricks/#automatically-roll-deployments
-        checksum/config: 0814cc9f3baa94e76e83a108b04d05200485610c7f5950c584503af7151a9e86
+        checksum/config: 5204c8793a312441190994144e04cce286931362a4d92d3d692a945ad333fb65
       labels:
         app: antrea
         component: antrea-agent
@@ -3952,7 +3952,7 @@ spec:
       annotations:
         # Automatically restart Pod if the ConfigMap changes
         # See https://helm.sh/docs/howto/charts_tips_and_tricks/#automatically-roll-deployments
-        checksum/config: 0814cc9f3baa94e76e83a108b04d05200485610c7f5950c584503af7151a9e86
+        checksum/config: 5204c8793a312441190994144e04cce286931362a4d92d3d692a945ad333fb65
       labels:
         app: antrea
         component: antrea-controller

diff --git a/docs/antrea-network-policy.md b/docs/antrea-network-policy.md
@@ -39,7 +39,7 @@
   - [Node Selector](#node-selector)
   - [toServices egress rules](#toservices-egress-rules)
   - [ServiceAccount based selection](#serviceaccount-based-selection)
-  - [ACNP appliedTo NodePort Service](#acnp-appliedto-nodeport-service)
+  - [Apply to NodePort Service](#apply-to-nodeport-service)
 - [ClusterGroup](#clustergroup)
   - [ClusterGroup CRD](#clustergroup-crd)
   - [kubectl commands for ClusterGroup](#kubectl-commands-for-clustergroup)
@@ -484,9 +484,9 @@ Specific Pods from specific Namespaces can be selected by providing both a
 The `appliedTo` field can also reference a ClusterGroup resource by setting
 the ClusterGroup's name in `group` field in place of the stand-alone selectors.
 The `appliedTo` field can also reference a Service by setting the Service's name
-and namespace in `service` field in place of the stand-alone selectors. Only a
+and Namespace in `service` field in place of the stand-alone selectors. Only a
 NodePort Service can be referred by this field. More details can be found in the
-[ACNPAppliedToNodePortService](#acnp-appliedto-nodeport-service) section.
+[ApplyToNodePortService](#apply-to-nodeport-service) section.
 IPBlock cannot be set in the `appliedTo` field.
 An IPBlock ClusterGroup referenced in an `appliedTo` field will be ignored,
 and the policy will have no effect.
@@ -1293,30 +1293,29 @@ spec:
 
 In this example, the policy will be applied to all Pods whose ServiceAccount is `sa-1` of `ns-1`.
 Let's call those Pods "appliedToPods".
-The egress `to` section will select all Pods whose ServiceAccount is in `ns-2` namespace and name as `sa-2`.
+The egress `to` section will select all Pods whose ServiceAccount is in `ns-2` Namespace and name as `sa-2`.
 Let's call those Pods "egressPods".
 After this policy is applied, traffic from "appliedToPods" to "egressPods" will be dropped.
 
 Note: Antrea will use a reserved label key for internal processing `serviceAccount`.
 The reserved label looks like: `internal.antrea.io/service-account:[ServiceAccountName]`. Users should avoid using
 this label key in any entities no matter if a policy with `serviceAccount` is applied in the cluster.
 
-### ACNP appliedTo NodePort Service
+### Apply to NodePort Service
 
-Antrea ClusterNetworkPolicy features a `service` field in `appliedTo` field to enable the ACNP could be enforced
-on the traffic from external client to a NodePort Service.
+Antrea ClusterNetworkPolicy features a `service` field in `appliedTo` field to enforce the ACNP rules on the
+traffic from external clients to a NodePort Service.
 
-`service` uses `namespace` and `name` to select the Service with a specific name under a specific namespace and
+`service` uses `namespace` and `name` to select the Service with a specific name under a specific Namespace;
 only a NodePort Service can be referred by `service` field.
 
-`service` field cannot be used with any other fields and a policy or a rule can't be applied to NodePort Service
-and other peers at the same time.
+There are a few **restrictions** on configuring a policy/rule that applies to NodePort Services:
 
-Since `service` field is used to control the external access of a NodePort Service, then
-
-1. If a `appliedTo` with `service` is used at policy level, then this policy can only contain ingress rules.
-2. If a `appliedTo` with `service` is used at rule level, then this rule can only be an ingress rule.
-3. If an ingress rule is applied to a NodePort Service, then this ingress can only use `ipBlock` in its `from` field.
+1. `service` field cannot be used with any other fields in `appliedTo`.
+2. a policy or a rule can't be applied to both a NodePort Service and other entities at the same time.
+3. If a `appliedTo` with `service` is used at policy level, then this policy can only contain ingress rules.
+4. If a `appliedTo` with `service` is used at rule level, then this rule can only be an ingress rule.
+5. If an ingress rule is applied to a NodePort Service, then this rule can only use `ipBlock` in its `from` field.
 
 An example policy using `service` in `appliedTo` could look like this:
 
@@ -1339,7 +1338,7 @@ spec:
               cidr: 1.1.1.0/24
 ```
 
-In this example, the policy will be applied to the NodePort Service `svc-1` in Namespace `ns-1`
+In this example, the policy will be applied to the NodePort Service `svc-1` in Namespace `ns-1`,
 and drop all packets from CIDR `1.1.1.0/24`.
 
 ## ClusterGroup

diff --git a/pkg/agent/controller/networkpolicy/networkpolicy_controller_test.go b/pkg/agent/controller/networkpolicy/networkpolicy_controller_test.go
@@ -674,7 +674,7 @@ func TestValidate(t *testing.T) {
 	groups := v1beta2.GroupMemberSet{}
 	groupAddress1, groupAddress2 := "225.1.2.3", "225.1.2.4"
 
-	groups["ns1/pod1"] = newAppliedToGroupMemberPod("pod1", "ns1")
+	groups["Pod:ns1/pod1"] = newAppliedToGroupMemberPod("pod1", "ns1")
 	controller.ruleCache.appliedToSetByGroup["appliedToGroup01"] = groups
 	controller.ruleCache.rules.Add(rule1)
 	controller.ruleCache.rules.Add(rule2)

diff --git a/pkg/agent/controller/networkpolicy/reconciler.go b/pkg/agent/controller/networkpolicy/reconciler.go
@@ -162,24 +162,24 @@ type lastRealized struct {
 	// the fqdn selector of this policy rule. It must be empty for policy rule
 	// that is not egress and does not have toFQDN field.
 	fqdnIPAddresses sets.String
-	// groupIDAddresses tracks the last realized set of groupIDs resolved for the
+	// serviceGroupIDs tracks the last realized set of groupIDs resolved for the
 	// toServices of this policy rule or services of TargetMember of this policy rule.
 	// It must be empty for policy rule that is neither an egress rule with toServices
 	// field nor an ingress rule that is applied to Services.
-	groupIDAddresses sets.Int64
+	serviceGroupIDs sets.Int64
 	// groupAddresses track the latest realized set of multicast groups for the multicast traffic
 	groupAddresses sets.String
 }
 
 func newLastRealized(rule *CompletedRule) *lastRealized {
 	return &lastRealized{
-		ofIDs:            map[servicesKey]uint32{},
-		CompletedRule:    rule,
-		podOFPorts:       map[servicesKey]sets.Int32{},
-		podIPs:           nil,
-		fqdnIPAddresses:  nil,
-		groupIDAddresses: nil,
-		groupAddresses:   nil,
+		ofIDs:           map[servicesKey]uint32{},
+		CompletedRule:   rule,
+		podOFPorts:      map[servicesKey]sets.Int32{},
+		podIPs:          nil,
+		fqdnIPAddresses: nil,
+		serviceGroupIDs: nil,
+		groupAddresses:  nil,
 	}
 }
 
@@ -504,7 +504,7 @@ func (r *reconciler) add(rule *CompletedRule, ofPriority *uint16, table uint8) e
 			if r.fqdnController != nil {
 				lastRealized.fqdnIPAddresses = nil
 			}
-			lastRealized.groupIDAddresses = nil
+			lastRealized.serviceGroupIDs = nil
 			return err
 		}
 		// Record ofID only if its Openflow is installed successfully.
@@ -551,13 +551,9 @@ func (r *reconciler) computeOFRulesForAdd(rule *CompletedRule, ofPriority *uint1
 		for svcKey, members := range membersByServicesMap {
 			toAddresses := make([]types.Address, 0)
 			if isRuleAppliedToService {
-				addressSet := sets.NewInt64()
-				for _, member := range members.Items() {
-					curAddresses, curAddressSet := r.svcRefToOFAddresses(*member.Service)
-					toAddresses = append(toAddresses, curAddresses...)
-					addressSet = addressSet.Union(curAddressSet)
-				}
-				lastRealized.groupIDAddresses = addressSet
+				svcGroupIDs := r.getSvcGroupIDs(members)
+				toAddresses = svcGroupIDsToOFAddresses(svcGroupIDs)
+				lastRealized.serviceGroupIDs = svcGroupIDs
 			} else {
 				ofPorts := r.getOFPorts(members)
 				toAddresses = ofPortsToOFAddresses(ofPorts)
@@ -647,16 +643,11 @@ func (r *reconciler) computeOFRulesForAdd(rule *CompletedRule, ofPriority *uint1
 				lastRealized.fqdnIPAddresses = addressSet
 			}
 			if len(rule.To.ToServices) > 0 {
-				var addresses []types.Address
-				addressSet := sets.NewInt64()
-				for _, svcRef := range rule.To.ToServices {
-					curAddresses, curAddressSet := r.svcRefToOFAddresses(svcRef)
-					addresses = append(addresses, curAddresses...)
-					addressSet = addressSet.Union(curAddressSet)
-				}
+				svcGroupIDs := r.svcRefsToGroupIDs(rule.To.ToServices)
+				addresses := svcGroupIDsToOFAddresses(svcGroupIDs)
 				ofRule.To = append(ofRule.To, addresses...)
 				// If the rule installation fails, this will be reset.
-				lastRealized.groupIDAddresses = addressSet
+				lastRealized.serviceGroupIDs = svcGroupIDs
 			}
 		}
 	}
@@ -762,18 +753,14 @@ func (r *reconciler) update(lastRealized *lastRealized, newRule *CompletedRule,
 		membersByServicesMap, servicesMap := groupMembersByServices(newRule.Services, newRule.TargetMembers)
 		for svcKey, members := range membersByServicesMap {
 			newOFPorts := r.getOFPorts(members)
-			newGroupIDAddressSet := sets.NewInt64()
-			ofID, exists := lastRealized.ofIDs[svcKey]
-			toAddresses := make([]types.Address, 0)
+			newGroupIDSet := r.getSvcGroupIDs(members)
+			var toAddresses []types.Address
 			if isRuleAppliedToService {
-				for _, member := range members.Items() {
-					curAddresses, curAddressSet := r.svcRefToOFAddresses(*member.Service)
-					toAddresses = append(toAddresses, curAddresses...)
-					newGroupIDAddressSet = newGroupIDAddressSet.Union(curAddressSet)
-				}
+				toAddresses = svcGroupIDsToOFAddresses(newGroupIDSet)
 			} else {
 				toAddresses = ofPortsToOFAddresses(newOFPorts)
 			}
+			ofID, exists := lastRealized.ofIDs[svcKey]
 			// Install a new Openflow rule if this group doesn't exist, otherwise do incremental update.
 			if !exists {
 				ofRule := &types.PolicyRule{
@@ -797,22 +784,21 @@ func (r *reconciler) update(lastRealized *lastRealized, newRule *CompletedRule,
 				}
 				lastRealized.ofIDs[svcKey] = ofRule.FlowID
 			} else {
-				addedTo := ofPortsToOFAddresses(newOFPorts.Difference(lastRealized.podOFPorts[svcKey]))
-				deletedTo := ofPortsToOFAddresses(lastRealized.podOFPorts[svcKey].Difference(newOFPorts))
+				var addedTo, deletedTo []types.Address
 				if isRuleAppliedToService {
-					originalGroupIDAddressSet := sets.NewInt64()
-					if lastRealized.groupIDAddresses != nil {
-						originalGroupIDAddressSet = lastRealized.groupIDAddresses
-					}
-					addedGroupIDAddress := newGroupIDAddressSet.Difference(originalGroupIDAddressSet)
-					removedGroupIDAddress := originalGroupIDAddressSet.Difference(newGroupIDAddressSet)
-					for a := range addedGroupIDAddress {
-						addedTo = append(addedTo, openflow.NewServiceGroupIDAddress(binding.GroupIDType(a)))
+					originalGroupIDSet := sets.NewInt64()
+					if lastRealized.serviceGroupIDs != nil {
+						originalGroupIDSet = lastRealized.serviceGroupIDs
 					}
-					for r := range removedGroupIDAddress {
-						deletedTo = append(deletedTo, openflow.NewServiceGroupIDAddress(binding.GroupIDType(r)))
+					addedTo = svcGroupIDsToOFAddresses(newGroupIDSet.Difference(originalGroupIDSet))
+					deletedTo = svcGroupIDsToOFAddresses(originalGroupIDSet.Difference(newGroupIDSet))
+				} else {
+					originalOfPortsSet := sets.NewInt32()
+					if lastRealized.podOFPorts[svcKey] != nil {
+						originalOfPortsSet = lastRealized.podOFPorts[svcKey]
 					}
-					lastRealized.groupIDAddresses = newGroupIDAddressSet
+					addedTo = ofPortsToOFAddresses(newOFPorts.Difference(originalOfPortsSet))
+					deletedTo = ofPortsToOFAddresses(originalOfPortsSet.Difference(newOFPorts))
 				}
 				if err := r.updateOFRule(ofID, addedFrom, addedTo, deletedFrom, deletedTo, ofPriority); err != nil {
 					return err
@@ -821,6 +807,7 @@ func (r *reconciler) update(lastRealized *lastRealized, newRule *CompletedRule,
 				delete(staleOFIDs, svcKey)
 			}
 			lastRealized.podOFPorts[svcKey] = newOFPorts
+			lastRealized.serviceGroupIDs = newGroupIDSet
 		}
 	} else {
 		if r.fqdnController != nil && len(newRule.To.FQDNs) > 0 {
@@ -894,23 +881,14 @@ func (r *reconciler) update(lastRealized *lastRealized, newRule *CompletedRule,
 						}
 					}
 				}
-				originalGroupIDAddressSet, newGroupIDAddressSet := sets.NewInt64(), sets.NewInt64()
-				if lastRealized.groupIDAddresses != nil {
-					originalGroupIDAddressSet = lastRealized.groupIDAddresses
-				}
+				originalGroupIDSet, newGroupIDSet := sets.NewInt64(), sets.NewInt64()
 				if len(newRule.To.ToServices) > 0 {
-					for _, svcRef := range newRule.To.ToServices {
-						_, groupIDSets := r.svcRefToOFAddresses(svcRef)
-						newGroupIDAddressSet = newGroupIDAddressSet.Union(groupIDSets)
-					}
-					addedGroupIDAddress := newGroupIDAddressSet.Difference(originalGroupIDAddressSet)
-					removedGroupIDAddress := originalGroupIDAddressSet.Difference(newGroupIDAddressSet)
-					for a := range addedGroupIDAddress {
-						addedTo = append(addedTo, openflow.NewServiceGroupIDAddress(binding.GroupIDType(a)))
-					}
-					for r := range removedGroupIDAddress {
-						deletedTo = append(deletedTo, openflow.NewServiceGroupIDAddress(binding.GroupIDType(r)))
+					if lastRealized.serviceGroupIDs != nil {
+						originalGroupIDSet = lastRealized.serviceGroupIDs
 					}
+					newGroupIDSet = r.svcRefsToGroupIDs(newRule.To.ToServices)
+					addedTo = svcGroupIDsToOFAddresses(newGroupIDSet.Difference(originalGroupIDSet))
+					deletedTo = svcGroupIDsToOFAddresses(originalGroupIDSet.Difference(newGroupIDSet))
 				}
 				if err := r.updateOFRule(ofID, addedFrom, addedTo, deletedFrom, deletedTo, ofPriority); err != nil {
 					return err
@@ -920,7 +898,7 @@ func (r *reconciler) update(lastRealized *lastRealized, newRule *CompletedRule,
 					lastRealized.fqdnIPAddresses = newFQDNAddressSet
 				}
 				// Update the groupID address set if rule installation succeeds.
-				lastRealized.groupIDAddresses = newGroupIDAddressSet
+				lastRealized.serviceGroupIDs = newGroupIDSet
 				// Delete valid servicesKey from staleOFIDs.
 				delete(staleOFIDs, svcKey)
 			}
@@ -1094,6 +1072,19 @@ func (r *reconciler) getIPs(members v1beta2.GroupMemberSet) sets.String {
 	return ips
 }
 
+func (r *reconciler) getSvcGroupIDs(members v1beta2.GroupMemberSet) sets.Int64 {
+	var svcRefs []v1beta2.ServiceReference
+	for _, m := range members {
+		if m.Service != nil {
+			svcRefs = append(svcRefs, v1beta2.ServiceReference{
+				Name:      m.Service.Name,
+				Namespace: m.Service.Namespace,
+			})
+		}
+	}
+	return r.svcRefsToGroupIDs(svcRefs)
+}
+
 // groupMembersByServices groups the provided groupMembers based on their services resolving result.
 // A map of servicesHash to the grouped members and a map of servicesHash to the services resolving result will be returned.
 func groupMembersByServices(services []v1beta2.Service, memberSet v1beta2.GroupMemberSet) (map[servicesKey]v1beta2.GroupMemberSet, map[servicesKey][]v1beta2.Service) {
@@ -1143,16 +1134,25 @@ func ofPortsToOFAddresses(ofPorts sets.Int32) []types.Address {
 	return addresses
 }
 
-func (r *reconciler) svcRefToOFAddresses(svcRef v1beta2.ServiceReference) ([]types.Address, sets.Int64) {
-	var addresses []types.Address
-	addressSet := sets.NewInt64()
-	for _, groupCounter := range r.groupCounters {
-		for _, groupID := range groupCounter.GetAllGroupIDs(k8s.NamespacedName(svcRef.Namespace, svcRef.Name)) {
-			addresses = append(addresses, openflow.NewServiceGroupIDAddress(groupID))
-			addressSet.Insert(int64(groupID))
+func (r *reconciler) svcRefsToGroupIDs(svcRefs []v1beta2.ServiceReference) sets.Int64 {
+	groupIDs := sets.NewInt64()
+	for _, svcRef := range svcRefs {
+		for _, groupCounter := range r.groupCounters {
+			for _, groupID := range groupCounter.GetAllGroupIDs(k8s.NamespacedName(svcRef.Namespace, svcRef.Name)) {
+				groupIDs.Insert(int64(groupID))
+			}
 		}
 	}
-	return addresses, addressSet
+	return groupIDs
+}
+
+func svcGroupIDsToOFAddresses(groupIDs sets.Int64) []types.Address {
+	// Must not return nil as it means not restricted by addresses in Openflow implementation.
+	addresses := make([]types.Address, 0, len(groupIDs))
+	for _, groupID := range groupIDs.List() {
+		addresses = append(addresses, openflow.NewServiceGroupIDAddress(binding.GroupIDType(groupID)))
+	}
+	return addresses
 }
 
 func groupMembersToOFAddresses(groupMemberSet v1beta2.GroupMemberSet) []types.Address {

diff --git a/pkg/agent/controller/networkpolicy/reject.go b/pkg/agent/controller/networkpolicy/reject.go
@@ -145,7 +145,7 @@ func (c *Controller) rejectRequest(pktIn *ofctrl.PacketIn) error {
 		return fmt.Errorf("error when generating reject response for the packet from: %s to %s: neither source nor destination are on this Node", dstIP, srcIP)
 	}
 	if packetOutType == RejectServiceRemoteToExternal {
-		dstMAC = "aa:bb:cc:dd:ee:ff"
+		dstMAC = openflow.GlobalVirtualMAC.String()
 	}
 	// When in AntreaIPAM mode, even though srcPod and dstPod are on the same Node, MAC
 	// will still be re-written in L3ForwardingTable. During rejection, the reject

diff --git a/pkg/apis/controlplane/sets.go b/pkg/apis/controlplane/sets.go
@@ -30,21 +30,28 @@ type groupMemberKey string
 type GroupMemberSet map[groupMemberKey]*GroupMember
 
 // normalizeGroupMember calculates the groupMemberKey of the provided
-// GroupMember based on the Pod/ExternalEntity's namespaced name and IPs.
+// GroupMember based on the Pod/ExternalEntity/Service's namespaced name and IPs.
 // For GroupMembers in appliedToGroups, the IPs are not set, so the
 // generated key does not contain IP information.
 func normalizeGroupMember(member *GroupMember) groupMemberKey {
 	// "/" is illegal in Namespace and name so is safe as the delimiter.
 	const delimiter = "/"
 	var b strings.Builder
 	if member.Pod != nil {
+		b.WriteString("Pod:")
 		b.WriteString(member.Pod.Namespace)
 		b.WriteString(delimiter)
 		b.WriteString(member.Pod.Name)
 	} else if member.ExternalEntity != nil {
+		b.WriteString("ExternalEntity:")
 		b.WriteString(member.ExternalEntity.Namespace)
 		b.WriteString(delimiter)
 		b.WriteString(member.ExternalEntity.Name)
+	} else if member.Service != nil {
+		b.WriteString("Service:")
+		b.WriteString(member.Service.Namespace)
+		b.WriteString(delimiter)
+		b.WriteString(member.Service.Name)
 	}
 	for _, ip := range member.IPs {
 		b.Write(ip)