vpp-maglev/internal/vpp/lbsync.go

// SPDX-License-Identifier: Apache-2.0

package vpp

import (
	"bytes"
	"errors"
	"fmt"
	"log/slog"
	"net"
	"regexp"
	"sort"
	"strconv"
	"strings"

	"go.fd.io/govpp/binapi/vlib"

	"git.ipng.ch/ipng/vpp-maglev/internal/config"
	"git.ipng.ch/ipng/vpp-maglev/internal/health"
	"git.ipng.ch/ipng/vpp-maglev/internal/metrics"
	ip_types "git.ipng.ch/ipng/vpp-maglev/internal/vpp/binapi/ip_types"
	lb "git.ipng.ch/ipng/vpp-maglev/internal/vpp/binapi/lb"
	lb_types "git.ipng.ch/ipng/vpp-maglev/internal/vpp/binapi/lb_types"
)

// sortedIPKeys returns the keys of m in ascending numeric IP order. Used
// by the AS iteration sites in reconcileVIP and recreateVIP to make the
// sequence of lb_as_add_del calls deterministic across maglevd instances
// on the same config.
//
// Why this matters for Maglev: VPP's LB plugin stores ASes in a vec in
// insertion order and rebuilds the new-flow lookup table by walking that
// vec. The per-AS permutation is a pure function of the AS address (so it
// matches across instances), but when two ASes want the same bucket on
// the same pass the tie is broken by whichever one comes first in the
// vec. If maglevd issues its add calls in Go map iteration order — which
// is randomised on every run — two independent maglevd instances with
// bit-identical configs can push the same ASes into VPP in different
// orders, and their lookup tables end up differing in tie-broken buckets.
// Sorting here is the first half of the fix; a matching sort inside VPP's
// lb_vip_update_new_flow_table closes the flap-history hole (where a
// remove+re-add after steady state reuses freed slots in the as_pool in
// locally-visited order) and is landing in a separate commit.
//
// The keys at every call site are IP literals from net.IP.String(). Sort
// order is numeric, not lexicographic: lexicographic would put 10.0.0.10
// before 10.0.0.2 (and 2001:db8::10 before 2001:db8::2), which is
// correct for determinism but confusing to operators reading the sync
// log. We parse each key back into a net.IP once, compare the canonical
// 16-byte form, and group IPv4 before IPv6 so mixed-family frontends
// read as "v4 block, v6 block" top-to-bottom. ParseIP always succeeds
// here because the caller built the map key via net.IP.String() in the
// first place.
func sortedIPKeys[V any](m map[string]V) []string {
	type kv struct {
		key string
		ip  net.IP
	}
	pairs := make([]kv, 0, len(m))
	for k := range m {
		pairs = append(pairs, kv{k, net.ParseIP(k)})
	}
	sort.Slice(pairs, func(i, j int) bool {
		return compareIPNumeric(pairs[i].ip, pairs[j].ip) < 0
	})
	out := make([]string, len(pairs))
	for i, p := range pairs {
		out[i] = p.key
	}
	return out
}

// compareIPNumeric returns <0, 0, >0 in the three-way convention. IPv4
// sorts before IPv6. Within each family the comparison runs against the
// canonical fixed-width byte form (4 bytes for v4, 16 bytes for v6),
// which makes byte ordering match numeric ordering. A nil input — which
// should not happen given sortedIPKeys's call contract — sorts before
// any non-nil address to keep the comparator total.
func compareIPNumeric(a, b net.IP) int {
	a4 := a.To4()
	b4 := b.To4()
	switch {
	case a == nil && b == nil:
		return 0
	case a == nil:
		return -1
	case b == nil:
		return 1
	case a4 != nil && b4 == nil:
		return -1
	case a4 == nil && b4 != nil:
		return 1
	case a4 != nil: // both v4
		return bytes.Compare(a4, b4)
	default: // both v6
		return bytes.Compare(a.To16(), b.To16())
	}
}

// ErrFrontendNotFound is returned by SyncLBStateVIP when the caller asks for
// a frontend name that does not exist in the config.
var ErrFrontendNotFound = errors.New("frontend not found in config")

// vipKey uniquely identifies a VPP LB VIP by its prefix, protocol, and port.
type vipKey struct {
	prefix   string // canonical CIDR form
	protocol uint8
	port     uint16
}

// desiredVIP is the sync's view of one VIP derived from the maglev config.
type desiredVIP struct {
	Prefix      *net.IPNet
	Protocol    uint8 // 6=TCP, 17=UDP, 255=any
	Port        uint16
	SrcIPSticky bool                 // lb_add_del_vip_v2.src_ip_sticky
	Encap       lb_types.LbEncapType // GRE4 / GRE6; matches the backend family, not the VIP's
	ASes        map[string]desiredAS // keyed by AS IP string
}

// desiredAS is one application server to be installed under a VIP.
type desiredAS struct {
	Address net.IP
	Weight  uint8 // 0-100
	Flush   bool  // if true, drop existing flows when transitioning to weight 0
}

// syncStats counts changes made to the dataplane during a sync run.
type syncStats struct {
	vipAdd   int
	vipDel   int
	asAdd    int
	asDel    int
	asWeight int
}

// recordSyncStats increments the Prometheus lbsync counters for one sync
// run. scope is "all" for SyncLBStateAll and "vip" for SyncLBStateVIP.
// Zero-valued kinds emit no increment (the counter stays at its previous
// value for that label set).
func recordSyncStats(scope string, st *syncStats) {
	if st.vipAdd > 0 {
		metrics.LBSyncTotal.WithLabelValues(scope, "vip_added").Add(float64(st.vipAdd))
	}
	if st.vipDel > 0 {
		metrics.LBSyncTotal.WithLabelValues(scope, "vip_removed").Add(float64(st.vipDel))
	}
	if st.asAdd > 0 {
		metrics.LBSyncTotal.WithLabelValues(scope, "as_added").Add(float64(st.asAdd))
	}
	if st.asDel > 0 {
		metrics.LBSyncTotal.WithLabelValues(scope, "as_removed").Add(float64(st.asDel))
	}
	if st.asWeight > 0 {
		metrics.LBSyncTotal.WithLabelValues(scope, "as_weight_updated").Add(float64(st.asWeight))
	}
}

// SyncLBStateAll reconciles the full VPP load-balancer state with the given
// config. For every frontend in cfg:
//   - if the VIP does not exist in VPP, create it;
//   - for every pool backend, add the application server if missing, or
//     update its weight if different.
//
// VIPs and ASes present in VPP but absent from the config are removed.
// Returns an error if any VPP API call fails.
func (c *Client) SyncLBStateAll(cfg *config.Config) error {
	if !c.IsConnected() {
		return errNotConnected
	}
	src := c.getStateSource()
	if src == nil {
		return fmt.Errorf("no state source configured")
	}

	cur, err := c.GetLBStateAll()
	if err != nil {
		return fmt.Errorf("read VPP LB state: %w", err)
	}
	desired := desiredFromConfig(cfg, src)

	ch, err := c.apiChannel()
	if err != nil {
		return err
	}
	defer ch.Close()

	slog.Info("vpp-lb-sync-start",
		"scope", "all",
		"vips-desired", len(desired),
		"vips-current", len(cur.VIPs))

	// Index both sides by (prefix, protocol, port).
	curByKey := make(map[vipKey]LBVIP, len(cur.VIPs))
	for _, v := range cur.VIPs {
		curByKey[makeVIPKey(v.Prefix, v.Protocol, v.Port)] = v
	}
	desByKey := make(map[vipKey]desiredVIP, len(desired))
	for _, d := range desired {
		desByKey[makeVIPKey(d.Prefix, d.Protocol, d.Port)] = d
	}

	var st syncStats

	// ---- pass 1: remove VIPs that are in VPP but not in config ----
	for k, v := range curByKey {
		if _, keep := desByKey[k]; keep {
			continue
		}
		if err := removeVIP(ch, v, &st); err != nil {
			return err
		}
	}

	// ---- pass 2: add/update VIPs that are in config ----
	for k, d := range desByKey {
		cur, existing := curByKey[k]
		var curPtr *LBVIP
		var curSticky bool
		if existing {
			curPtr = &cur
			curSticky = cur.SrcIPSticky
		}
		if err := reconcileVIP(ch, d, curPtr, curSticky, "", &st); err != nil {
			return err
		}
	}

	recordSyncStats("all", &st)
	slog.Info("vpp-lb-sync-done",
		"scope", "all",
		"vip-added", st.vipAdd,
		"vip-removed", st.vipDel,
		"as-added", st.asAdd,
		"as-removed", st.asDel,
		"as-weight-updated", st.asWeight)
	return nil
}

// SyncLBStateVIP reconciles a single VIP (identified by frontend name) with
// the given config. Unlike SyncLBStateAll, it never removes VIPs: if the
// frontend is missing from cfg, SyncLBStateVIP returns ErrFrontendNotFound.
// This is the right tool for targeted updates on a busy load-balancer with
// many VIPs — only one VIP is read from VPP and only its ASes are modified.
//
// flushAddress, when non-empty, is the IP of an application server whose
// weight change (if any) should be pushed with IsFlush=true regardless of
// the usual "only flush on non-zero → zero" heuristic. This is how the
// SetFrontendPoolBackendWeight RPC exposes an explicit "drop flows now"
// knob: the server handler resolves the backend's config address and
// passes it here. Callers that don't need forced flushing pass "".
func (c *Client) SyncLBStateVIP(cfg *config.Config, feName, flushAddress string) error {
	if !c.IsConnected() {
		return errNotConnected
	}
	src := c.getStateSource()
	if src == nil {
		return fmt.Errorf("no state source configured")
	}
	fe, ok := cfg.Frontends[feName]
	if !ok {
		return fmt.Errorf("%q: %w", feName, ErrFrontendNotFound)
	}
	d := desiredFromFrontend(cfg, fe, src)

	cur, err := c.GetLBStateVIP(d.Prefix, d.Protocol, d.Port)
	if err != nil {
		return fmt.Errorf("read VPP VIP state: %w", err)
	}

	ch, err := c.apiChannel()
	if err != nil {
		return err
	}
	defer ch.Close()

	slog.Info("vpp-lb-sync-start",
		"scope", "vip",
		"frontend", feName,
		"vip", d.Prefix.IP.String(),
		"protocol", protocolName(d.Protocol),
		"port", d.Port)

	var curSticky bool
	if cur != nil {
		curSticky = cur.SrcIPSticky
	}

	var st syncStats
	if err := reconcileVIP(ch, d, cur, curSticky, flushAddress, &st); err != nil {
		return err
	}
	recordSyncStats("vip", &st)
	slog.Info("vpp-lb-sync-done",
		"scope", "vip",
		"frontend", feName,
		"vip-added", st.vipAdd,
		"as-added", st.asAdd,
		"as-removed", st.asDel,
		"as-weight-updated", st.asWeight)
	return nil
}

// reconcileVIP brings one VIP's state in VPP into alignment with the desired
// state. If cur is nil the VIP is added from scratch; otherwise ASes are
// added, removed, and reweighted individually. Stats are accumulated into st.
//
// curSticky is the src_ip_sticky flag VPP currently has programmed for this
// VIP, as scraped by queryLBSticky. Callers only consult it when cur != nil,
// and the scrape reads the same live lb_main.vips pool as lb_vip_dump, so a
// matching entry is always present. When the flag differs from the desired
// value, the VIP is torn down (ASes del+flushed, VIP deleted) and recreated
// — VPP has no API to mutate src_ip_sticky on an existing VIP.
func reconcileVIP(ch *loggedChannel, d desiredVIP, cur *LBVIP, curSticky bool, flushAddress string, st *syncStats) error {
	if cur == nil {
		if err := addVIP(ch, d); err != nil {
			return err
		}
		st.vipAdd++
		for _, addr := range sortedIPKeys(d.ASes) {
			if err := addAS(ch, d.Prefix, d.Protocol, d.Port, d.ASes[addr]); err != nil {
				return err
			}
			st.asAdd++
		}
		return nil
	}

	if curSticky != d.SrcIPSticky {
		return recreateVIP(ch, d, *cur, st, "src-ip-sticky-changed",
			"from", curSticky, "to", d.SrcIPSticky)
	}

	// Encap mismatch: every backend flipped address family (e.g. all
	// IPv6 → all IPv4 after a config edit). VPP's encap is a VIP-level
	// attribute set at lb_add_del_vip time with no mutation API, so
	// adding new-family ASes under the old encap wedges the
	// reconciler: packets would be wrapped for the wrong family and
	// the new backends never see traffic. The only recovery is a VIP
	// recreate. The old-family ASes end up orphaned in VPP's pool and
	// are GC'd on the plugin's own ~40s "USED-flag=false" schedule;
	// the next regular sync tick confirms steady state. A recreate
	// does tear down existing flows to the VIP, which is why we gate
	// on an explicit encap difference rather than reconciling every
	// sync cycle.
	if desiredEncap := encapString(d.Encap); desiredEncap != cur.Encap {
		return recreateVIP(ch, d, *cur, st, "encap-changed",
			"from", cur.Encap, "to", desiredEncap)
	}

	// VIP exists in both — reconcile ASes.
	curASes := make(map[string]LBAS, len(cur.ASes))
	for _, a := range cur.ASes {
		curASes[a.Address.String()] = a
	}

	// Remove ASes that are in VPP but not desired.
	for _, addr := range sortedIPKeys(curASes) {
		a := curASes[addr]
		if _, keep := d.ASes[addr]; keep {
			continue
		}
		if err := delAS(ch, cur.Prefix, cur.Protocol, cur.Port, a.Address, a.Weight); err != nil {
			return err
		}
		st.asDel++
	}

	// Add new ASes, update weights on existing ones.
	for _, addr := range sortedIPKeys(d.ASes) {
		a := d.ASes[addr]
		c, hit := curASes[addr]
		if !hit {
			if err := addAS(ch, d.Prefix, d.Protocol, d.Port, a); err != nil {
				return err
			}
			st.asAdd++
			continue
		}
		// setASWeight is issued whenever the weight changes OR whenever
		// the state machine asks for a flush (a.Flush=true, currently
		// emitted only for StateDisabled). The a.Flush path has to
		// fire even on a no-op weight change, because a backend can
		// reach StateDisabled via a pool-failover that already drained
		// its VPP weight to 0 on an earlier tick — at that moment
		// c.Weight == a.Weight == 0, and a gate keyed solely on the
		// weight diff would silently drop the flush intent and leave
		// stale sticky-cache entries pointing at the now-disabled AS
		// (see the "disable nlams0 after fallback deactivation" trace
		// in the bug investigation).
		//
		// Firing unconditionally on a.Flush is idempotent at VPP's
		// side: lb_as_set_weight with an unchanged weight is a no-op
		// on the Maglev table (lb_vip_update_new_flow_table rebuilds
		// the same table), and a redundant lb_flush_vip_as is bounded
		// — it walks each per-worker sticky_ht once. The trade-off is
		// that disabled backends re-issue the flush on every periodic
		// SyncLBStateAll tick, and any sticky entries that happened to
		// land in the meantime get cleared; both are acceptable for
		// the "correctness over churn" semantics we want here.
		weightChanged := c.Weight != a.Weight
		flush := a.Flush
		// Caller-forced flush: used by SetFrontendPoolBackendWeight
		// with flush=true to explicitly drop live sessions for a
		// single backend. The address match is exact — no other
		// AS's weight change is affected, even if several happen
		// in the same reconcile pass.
		if flushAddress != "" && addr == flushAddress {
			flush = true
		}
		if weightChanged || flush {
			if err := setASWeight(ch, d.Prefix, d.Protocol, d.Port, a, c.Weight, flush); err != nil {
				return err
			}
			st.asWeight++
		}
	}
	return nil
}

// recreateVIP tears down an existing VIP and rebuilds it with the
// desired configuration and ASes. Used when a VIP attribute that VPP
// can't mutate in place has changed — today src_ip_sticky and the
// encap family. reason is logged as an operator-facing explanation;
// extra is appended to the slog call as additional fields (typically
// "from", <oldvalue>, "to", <newvalue>).
func recreateVIP(ch *loggedChannel, d desiredVIP, cur LBVIP, st *syncStats, reason string, extra ...any) error {
	logAttrs := []any{
		"vip", d.Prefix.IP.String(),
		"protocol", protocolName(d.Protocol),
		"port", d.Port,
		"reason", reason,
	}
	logAttrs = append(logAttrs, extra...)
	slog.Info("vpp-lb-sync-vip-recreate", logAttrs...)
	if err := removeVIP(ch, cur, st); err != nil {
		return err
	}
	if err := addVIP(ch, d); err != nil {
		return err
	}
	st.vipAdd++
	for _, addr := range sortedIPKeys(d.ASes) {
		if err := addAS(ch, d.Prefix, d.Protocol, d.Port, d.ASes[addr]); err != nil {
			return err
		}
		st.asAdd++
	}
	return nil
}

// removeVIP flushes all ASes from a VIP and then deletes the VIP itself.
func removeVIP(ch *loggedChannel, v LBVIP, st *syncStats) error {
	for _, as := range v.ASes {
		if err := delAS(ch, v.Prefix, v.Protocol, v.Port, as.Address, as.Weight); err != nil {
			return err
		}
		st.asDel++
	}
	if err := delVIP(ch, v.Prefix, v.Protocol, v.Port); err != nil {
		return err
	}
	st.vipDel++
	return nil
}

// desiredFromConfig flattens every frontend in cfg into a desired VIP set.
// src provides the per-backend health state so weights and flush hints
// reflect the current runtime state, not just the static config.
func desiredFromConfig(cfg *config.Config, src StateSource) []desiredVIP {
	out := make([]desiredVIP, 0, len(cfg.Frontends))
	for _, fe := range cfg.Frontends {
		out = append(out, desiredFromFrontend(cfg, fe, src))
	}
	return out
}

// desiredFromFrontend builds the desired VIP for a single frontend.
//
// All backends across all pools of a frontend are merged into a single
// application-server list so VPP knows about every backend that could ever
// receive traffic. The per-AS weight and flush hint are computed by
// asFromBackend from three inputs: (pool index, backend health state,
// configured pool weight).
//
// When the same backend appears in multiple pools, the first pool it
// appears in wins.
func desiredFromFrontend(cfg *config.Config, fe config.Frontend, src StateSource) desiredVIP {
	bits := 32
	if fe.Address.To4() == nil {
		bits = 128
	}
	// Start with an encap derived from the VIP's own family as a
	// fallback. This only applies when the frontend has zero valid
	// backends (e.g. every referenced backend is missing from
	// cfg.Backends); any real backend below overrides it to the
	// backend family, which is the correct choice because the GRE
	// encap carries backend traffic, not VIP traffic. Config
	// validation already guarantees every backend in a frontend
	// shares the same family, so the first valid backend we see is
	// authoritative.
	d := desiredVIP{
		Prefix:      &net.IPNet{IP: fe.Address, Mask: net.CIDRMask(bits, bits)},
		Protocol:    protocolFromConfig(fe.Protocol),
		Port:        fe.Port,
		SrcIPSticky: fe.SrcIPSticky,
		Encap:       encapForIP(fe.Address),
		ASes:        make(map[string]desiredAS),
	}
	encapSet := false

	states := snapshotStates(fe, src)
	activePool := health.ActivePoolIndex(fe, states)

	for poolIdx, pool := range fe.Pools {
		for bName, pb := range pool.Backends {
			b, ok := cfg.Backends[bName]
			if !ok || b.Address == nil {
				continue
			}
			if !encapSet {
				d.Encap = encapForIP(b.Address)
				encapSet = true
			}
			// Disabled backends (either via operator action or config) are
			// kept in the desired set so they stay installed in VPP with
			// weight=0 — they must not be deleted, otherwise a subsequent
			// enable has to re-add them and existing flow-table state (if
			// any) is lost. The state machine drives what weight to set
			// via health.BackendEffectiveWeight; we never filter on
			// b.Enabled here.
			addr := b.Address.String()
			if _, already := d.ASes[addr]; already {
				continue
			}
			w, flush := health.BackendEffectiveWeight(poolIdx, activePool, states[bName], pb.Weight, fe.FlushOnDown)
			d.ASes[addr] = desiredAS{
				Address: b.Address,
				Weight:  w,
				Flush:   flush,
			}
		}
	}
	return d
}

// EffectiveWeights returns the current effective VPP weight for every backend
// in every pool of fe, keyed by poolIdx and backend name. Intended for
// observability (e.g. the GetFrontend gRPC handler) — the sync path and the
// checker's frontend-state logic use health.EffectiveWeights directly.
func EffectiveWeights(fe config.Frontend, src StateSource) map[int]map[string]uint8 {
	return health.EffectiveWeights(fe, snapshotStates(fe, src))
}

// snapshotStates builds a state map for every backend referenced by fe. It
// takes one read-lock per backend via the StateSource interface, so the
// caller gets a consistent view to feed into the pure health helpers.
func snapshotStates(fe config.Frontend, src StateSource) map[string]health.State {
	states := make(map[string]health.State)
	for _, pool := range fe.Pools {
		for bName := range pool.Backends {
			if s, ok := src.BackendState(bName); ok {
				states[bName] = s
			} else {
				states[bName] = health.StateUnknown
			}
		}
	}
	return states
}

// ---- API call helpers ------------------------------------------------------

// defaultFlowsTableLength is sent as NewFlowsTableLength in lb_add_del_vip_v2.
// The .api file declares default=1024 but that default is only applied by VAT/
// the CLI parser, not when a raw message is marshalled over the socket. If we
// send 0, the plugin's vec_validate explodes (OOM / panic). Must be a power of
// two — 1024 matches the default that would have been applied via CLI.
const defaultFlowsTableLength = 1024

func addVIP(ch *loggedChannel, d desiredVIP) error {
	req := &lb.LbAddDelVipV2{
		Pfx:                 ip_types.NewAddressWithPrefix(*d.Prefix),
		Protocol:            d.Protocol,
		Port:                d.Port,
		Encap:               d.Encap,
		Type:                lb_types.LB_API_SRV_TYPE_CLUSTERIP,
		NewFlowsTableLength: defaultFlowsTableLength,
		SrcIPSticky:         d.SrcIPSticky,
		IsDel:               false,
	}
	reply := &lb.LbAddDelVipV2Reply{}
	if err := ch.SendRequest(req).ReceiveReply(reply); err != nil {
		return fmt.Errorf("lb_add_del_vip_v2 add %s: %w", d.Prefix, err)
	}
	if reply.Retval != 0 {
		return fmt.Errorf("lb_add_del_vip_v2 add %s: retval=%d", d.Prefix, reply.Retval)
	}
	slog.Info("vpp-lb-sync-vip-added",
		"vip", d.Prefix.IP.String(),
		"protocol", protocolName(d.Protocol),
		"port", d.Port,
		"encap", encapName(d.Encap),
		"src-ip-sticky", d.SrcIPSticky)
	return nil
}

func delVIP(ch *loggedChannel, prefix *net.IPNet, protocol uint8, port uint16) error {
	req := &lb.LbAddDelVipV2{
		Pfx:      ip_types.NewAddressWithPrefix(*prefix),
		Protocol: protocol,
		Port:     port,
		IsDel:    true,
	}
	reply := &lb.LbAddDelVipV2Reply{}
	if err := ch.SendRequest(req).ReceiveReply(reply); err != nil {
		return fmt.Errorf("lb_add_del_vip_v2 del %s: %w", prefix, err)
	}
	if reply.Retval != 0 {
		return fmt.Errorf("lb_add_del_vip_v2 del %s: retval=%d", prefix, reply.Retval)
	}
	slog.Info("vpp-lb-sync-vip-removed",
		"vip", prefix.IP.String(),
		"protocol", protocolName(protocol),
		"port", port)
	return nil
}

func addAS(ch *loggedChannel, prefix *net.IPNet, protocol uint8, port uint16, a desiredAS) error {
	req := &lb.LbAddDelAsV2{
		Pfx:       ip_types.NewAddressWithPrefix(*prefix),
		Protocol:  protocol,
		Port:      port,
		AsAddress: ip_types.NewAddress(a.Address),
		Weight:    a.Weight,
		IsDel:     false,
	}
	reply := &lb.LbAddDelAsV2Reply{}
	if err := ch.SendRequest(req).ReceiveReply(reply); err != nil {
		return fmt.Errorf("lb_add_del_as_v2 add %s@%s: %w", a.Address, prefix, err)
	}
	if reply.Retval != 0 {
		return fmt.Errorf("lb_add_del_as_v2 add %s@%s: retval=%d", a.Address, prefix, reply.Retval)
	}
	slog.Info("vpp-lb-sync-as-added",
		"vip", prefix.IP.String(),
		"protocol", protocolName(protocol),
		"port", port,
		"address", a.Address.String(),
		"weight", a.Weight)
	return nil
}

func delAS(ch *loggedChannel, prefix *net.IPNet, protocol uint8, port uint16, addr net.IP, fromWeight uint8) error {
	req := &lb.LbAddDelAsV2{
		Pfx:       ip_types.NewAddressWithPrefix(*prefix),
		Protocol:  protocol,
		Port:      port,
		AsAddress: ip_types.NewAddress(addr),
		IsDel:     true,
		IsFlush:   true,
	}
	reply := &lb.LbAddDelAsV2Reply{}
	if err := ch.SendRequest(req).ReceiveReply(reply); err != nil {
		return fmt.Errorf("lb_add_del_as_v2 del %s@%s: %w", addr, prefix, err)
	}
	if reply.Retval != 0 {
		return fmt.Errorf("lb_add_del_as_v2 del %s@%s: retval=%d", addr, prefix, reply.Retval)
	}
	slog.Info("vpp-lb-sync-as-removed",
		"vip", prefix.IP.String(),
		"protocol", protocolName(protocol),
		"port", port,
		"address", addr.String(),
		"weight", fromWeight)
	return nil
}

func setASWeight(ch *loggedChannel, prefix *net.IPNet, protocol uint8, port uint16, a desiredAS, fromWeight uint8, flush bool) error {
	req := &lb.LbAsSetWeight{
		Pfx:       ip_types.NewAddressWithPrefix(*prefix),
		Protocol:  protocol,
		Port:      port,
		AsAddress: ip_types.NewAddress(a.Address),
		Weight:    a.Weight,
		IsFlush:   flush,
	}
	reply := &lb.LbAsSetWeightReply{}
	if err := ch.SendRequest(req).ReceiveReply(reply); err != nil {
		return fmt.Errorf("lb_as_set_weight %s@%s: %w", a.Address, prefix, err)
	}
	if reply.Retval != 0 {
		return fmt.Errorf("lb_as_set_weight %s@%s: retval=%d", a.Address, prefix, reply.Retval)
	}
	slog.Info("vpp-lb-sync-as-weight-updated",
		"vip", prefix.IP.String(),
		"protocol", protocolName(protocol),
		"port", port,
		"address", a.Address.String(),
		"from", fromWeight,
		"to", a.Weight,
		"flush", flush)
	return nil
}

// ---- VIP snapshot scrape ---------------------------------------------------

// TEMPORARY WORKAROUND: VPP's lb_vip_dump / lb_vip_details message does not
// return the src_ip_sticky flag, and lb_vip_details also doesn't expose the
// LB plugin's pool index — which is what the stats segment uses to key
// per-VIP counters. We work around both by running `show lb vips verbose`
// via the cli_inband API and parsing the human-readable output. This is
// slow and fragile (the format is not a stable API) and must be replaced
// once VPP ships an lb_vip_v2_dump that includes these fields.

// lbVIPSnapshot holds the per-VIP facts that the scrape recovers. `index`
// is the LB pool index (`vip - lbm->vips` in lb.c) used as the second
// dimension of the vip_counters SimpleCounterStat in the stats segment.
type lbVIPSnapshot struct {
	index  uint32
	sticky bool
}

// queryLBVIPSnapshot runs `show lb vips verbose` and parses it into a map
// keyed by vipKey. The returned error is non-nil only when the cli_inband
// RPC itself fails.
func queryLBVIPSnapshot(ch *loggedChannel) (map[vipKey]lbVIPSnapshot, error) {
	req := &vlib.CliInband{Cmd: "show lb vips verbose"}
	reply := &vlib.CliInbandReply{}
	if err := ch.SendRequest(req).ReceiveReply(reply); err != nil {
		return nil, fmt.Errorf("cli_inband %q: %w", req.Cmd, err)
	}
	if reply.Retval != 0 {
		return nil, fmt.Errorf("cli_inband %q: retval=%d", req.Cmd, reply.Retval)
	}
	return parseLBVIPSnapshot(reply.Reply), nil
}

// queryLBSticky is a thin projection of queryLBVIPSnapshot for callers that
// only care about the src_ip_sticky flag (the sync path).
func queryLBSticky(ch *loggedChannel) (map[vipKey]bool, error) {
	snap, err := queryLBVIPSnapshot(ch)
	if err != nil {
		return nil, err
	}
	out := make(map[vipKey]bool, len(snap))
	for k, v := range snap {
		out[k] = v.sticky
	}
	return out, nil
}

// lbVIPHeaderRe matches the first line of each VIP block in the output of
// `show lb vips verbose`. VPP produces lines like:
//
//	ip4-gre4 [1] 192.0.2.1/32 src_ip_sticky
//	ip6-gre6 [2] 2001:db8::1/128
//
// Capture groups: 1 = pool index, 2 = prefix (CIDR), 3 = " src_ip_sticky" when present.
var lbVIPHeaderRe = regexp.MustCompile(
	`\b(?:ip4-gre4|ip6-gre6|ip4-gre6|ip6-gre4|ip4-l3dsr|ip4-nat4|ip6-nat6)\s+\[(\d+)\]\s+(\S+)(\s+src_ip_sticky)?`,
)

// lbVIPProtoRe matches the `protocol:<n> port:<n>` sub-line that appears
// under each non-all-port VIP block.
var lbVIPProtoRe = regexp.MustCompile(`protocol:(\d+)\s+port:(\d+)`)

// parseLBVIPSnapshot turns the reply text of `show lb vips verbose` into a
// map of vipKey → lbVIPSnapshot. It walks lines sequentially: each header
// line starts a new VIP, and the following `protocol:<p> port:<port>` line
// (if any) refines the key. All-port VIPs have no protocol/port sub-line
// and are recorded with protocol=255 and port=0 — the same key format used
// by the rest of the sync path (see protocolFromConfig).
func parseLBVIPSnapshot(text string) map[vipKey]lbVIPSnapshot {
	out := make(map[vipKey]lbVIPSnapshot)
	var havePending bool
	var curPrefix string
	var curIndex uint32
	var curSticky bool
	var curProto uint8 = 255
	var curPort uint16

	flush := func() {
		if !havePending || curPrefix == "" {
			return
		}
		out[vipKey{prefix: curPrefix, protocol: curProto, port: curPort}] = lbVIPSnapshot{
			index:  curIndex,
			sticky: curSticky,
		}
	}

	for _, line := range strings.Split(text, "\n") {
		if m := lbVIPHeaderRe.FindStringSubmatch(line); m != nil {
			flush()
			havePending = true
			if idx, err := strconv.ParseUint(m[1], 10, 32); err == nil {
				curIndex = uint32(idx)
			} else {
				curIndex = 0
			}
			curPrefix = canonicalCIDR(m[2])
			curSticky = m[3] != ""
			curProto = 255
			curPort = 0
			continue
		}
		if !havePending {
			continue
		}
		if m := lbVIPProtoRe.FindStringSubmatch(line); m != nil {
			if p, err := strconv.Atoi(m[1]); err == nil {
				curProto = uint8(p)
			}
			if p, err := strconv.Atoi(m[2]); err == nil {
				curPort = uint16(p)
			}
		}
	}
	flush()
	return out
}

// canonicalCIDR parses a CIDR string and returns it in Go's canonical
// net.IPNet.String() form so it matches the keys produced by makeVIPKey.
// Returns the input unchanged if parsing fails — the header regex should
// have validated it, but a mismatch shouldn't panic the sync path.
func canonicalCIDR(s string) string {
	_, ipnet, err := net.ParseCIDR(s)
	if err != nil {
		return s
	}
	return ipnet.String()
}

// ---- utility ---------------------------------------------------------------

func makeVIPKey(prefix *net.IPNet, protocol uint8, port uint16) vipKey {
	return vipKey{prefix: prefix.String(), protocol: protocol, port: port}
}

func protocolFromConfig(s string) uint8 {
	switch s {
	case "tcp":
		return 6
	case "udp":
		return 17
	}
	return 255 // any
}

func protocolName(p uint8) string {
	switch p {
	case 6:
		return "tcp"
	case 17:
		return "udp"
	case 255:
		return "any"
	}
	return fmt.Sprintf("%d", p)
}

func encapForIP(ip net.IP) lb_types.LbEncapType {
	if ip.To4() != nil {
		return lb_types.LB_API_ENCAP_TYPE_GRE4
	}
	return lb_types.LB_API_ENCAP_TYPE_GRE6
}

func encapName(e lb_types.LbEncapType) string {
	switch e {
	case lb_types.LB_API_ENCAP_TYPE_GRE4:
		return "gre4"
	case lb_types.LB_API_ENCAP_TYPE_GRE6:
		return "gre6"
	}
	return fmt.Sprintf("%d", e)
}