Update address.HoldExternal to allow holding IPv6 as well

This is done in order to reuse address holding logic in future mixed protocol NetLB IPv6 implementation. Similar thing has been already done in IPv4.

* Updates config with fields that are IPv4/IPv6 specific.
* Adds tests for HoldExternal IPv6 addresses.
* Updates forwarding_rules_ipv6.go to use address.HoldExternal.
* Removes tearDownResourcesWithWrongNetworkTier from l4netlb as it is already implemented inside address.HoldExternal.

pkg/address/hold.go

@@ -1,6 +1,8 @@
 package address
 
 import (
+	"fmt"
+
 	"github.com/GoogleCloudPlatform/k8s-cloud-provider/pkg/cloud"
 	api_v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/types"
@@ -9,6 +11,7 @@ import (
 	"k8s.io/ingress-gce/pkg/annotations"
 	"k8s.io/ingress-gce/pkg/composite"
 	"k8s.io/ingress-gce/pkg/utils"
+	"k8s.io/ingress-gce/pkg/utils/namer"
 	"k8s.io/klog/v2"
 )
 
@@ -19,6 +22,8 @@ type HoldConfig struct {
 	Service               *api_v1.Service
 	ExistingRules         []*composite.ForwardingRule
 	ForwardingRuleDeleter ForwardingRuleDeleter
+	IPVersion             IPVersion
+	SubnetworkURL         string
 }
 
 type ForwardingRuleDeleter interface {
@@ -31,28 +36,34 @@ type HoldResult struct {
 	Release func() error
 }
 
-// HoldExternalIPv4 will determine which IP to use for forwarding rules
+// HoldExternal will determine which IP to use for forwarding rules
 // and will hold it for future forwarding rules. After binding
 // IP to a forwarding rule call Release to prevent leaks.
-func HoldExternalIPv4(cfg HoldConfig) (HoldResult, error) {
+func HoldExternal(cfg HoldConfig) (HoldResult, error) {
 	var err error
 	res := HoldResult{
 		Release: func() error { return nil },
 	}
-	log := cfg.Logger.WithName("HoldIPv4")
-
-	// external specific
-	subnet := ""
-	name := utils.LegacyForwardingRuleName(cfg.Service)
+	log := cfg.Logger.WithName("HoldExternal")
 
 	// Determine IP which will be used for this LB. If no forwarding rule has been established
 	// or specified in the Service spec, then requestedIP = "".
 	rule := pickForwardingRuleToInferIP(cfg.ExistingRules)
-	res.IP, err = IPv4ToUse(cfg.Cloud, cfg.Recorder, cfg.Service, rule, subnet)
+
+	switch cfg.IPVersion {
+	case IPv4Version:
+		res.IP, err = IPv4ToUse(cfg.Cloud, cfg.Recorder, cfg.Service, rule, cfg.SubnetworkURL)
+	case IPv6Version:
+		res.IP, err = IPv6ToUse(cfg.Cloud, cfg.Service, rule, cfg.SubnetworkURL, cfg.Logger)
+	default:
+		return res, fmt.Errorf("unsupported IP version: '%s', only IPv4 and IPv6 are supported", cfg.IPVersion)
+	}
+
 	if err != nil {
-		log.Error(err, "IPv4ToUse for service returned error")
+		log.Error(err, "IPvXToUse for service returned error")
 		return res, err
 	}
+	log.V(2).Info("IP for service", "ip", res.IP)
 
 	// We can't use manager for legacy networks
 	if cfg.Cloud.IsLegacyNetwork() {
@@ -67,8 +78,8 @@ func HoldExternalIPv4(cfg HoldConfig) (HoldResult, error) {
 
 	addrMgr := NewManager(
 		cfg.Cloud, nm, cfg.Cloud.Region(),
-		subnet, name, res.IP,
-		cloud.SchemeExternal, netTier, IPv4Version, cfg.Logger,
+		cfg.SubnetworkURL, name(cfg), res.IP,
+		cloud.SchemeExternal, netTier, cfg.IPVersion, cfg.Logger,
 	)
 
 	// If network tier annotation in Service Spec is present
@@ -124,3 +135,11 @@ func tearDownRulesIfNetworkTierMismatch(deleter ForwardingRuleDeleter, existingR
 	}
 	return nil
 }
+
+func name(cfg HoldConfig) string {
+	name := utils.LegacyForwardingRuleName(cfg.Service)
+	if cfg.IPVersion == IPv6Version {
+		name = namer.GetSuffixedName(name, "-ipv6")
+	}
+	return name
+}