Files
netbird/client/iface/bind/ice_bind.go
Maycon Santos a1415dbc05 [client] Fix the ICEBind races that wedge interface creation (#7377)
* [client] Add tests for the ICEBind open and close races

Running many embedded clients in one process intermittently wedges interface
creation. A goroutine dump taken from 50 clients shows ten of them parked for
seven minutes in Device.IpcSet, in closeBindLocked waiting on
device.net.stopping.Wait, holding device.net while every other device
goroutine queues behind it on Device.Up.

Open writes s.closed and Close reads it with no synchronisation, and Close
also closes s.closedChan without the mutex that Open swaps it under. Two
Closes can both pass the check and close the same channel, and a Close racing
an Open can mark the bind closed while a live channel and live receive
functions remain, after which every later Close takes its early return and
runs neither close(closedChan) nor StdNetBind.Close. The receive functions
never stop, so stopping.Wait never returns.

These tests do not fix that. The first pins the contract closeBindLocked
depends on and passes today. The other two fail under -race, reporting the
races at the three sites above, and pass again once closed and closedChan are
guarded consistently.

* [client] Release parked receivers so reopening a bind cannot stall

receiveRelayed held closedChanMu for the whole of its blocking select, so a
parked receiver kept the read lock indefinitely and Open could never take the
write lock it needs to install a fresh closedChan. wireguard-go reaches Open
from Device.IpcSet and Device.Up with device.net held, so the stall took the
device lock with it: interface creation never finished, every other device
goroutine queued behind Device.Up, and Engine.Start never returned.

Callers now copy the channel under a short read lock and select on the copy.
Copying alone would stand a new trap in the same place, because an Open that
follows an Open leaves the previous generation parked on a channel no later
Close can reach, so Open now closes the outgoing channel before swapping it.

closed and closedChan are also updated together under that mutex. Read and
written apart, Close could see a stale closed and skip both close(closedChan)
and StdNetBind.Close, leaving every receive function running and wedging
closeBindLocked on device.net.stopping.Wait, or two Close calls could pass the
check together and close the same channel twice.

TestICEBindOpenDoesNotBlockOnParkedReceiver fails without this change, without
needing the race detector. The other three cover the surrounding contract and
report the state races under -race.

* [client] Make the bind lifecycle transition atomic and tighten its tests

Review caught that the previous commit moved the torn transition rather than
removing it. Open published the new generation before calling StdNetBind.Open,
so an Open rejected because the bind was already open had already signalled the
outgoing generation, and a Close arriving in that window could mark the bind
closed while the same call went on to install live sockets. Every later Close
then returned early and never shut them down.

Open now calls StdNetBind.Open first, so a failure leaves the current
generation untouched, and both Open and Close hold the lock across the whole
transition. Ordering is safe: StdNetBind.Open reaches muUDPMux through
createReceiverFn, and no path takes muUDPMux before closedChanMu.

The tests were also weaker than they read. The stress test claimed to cover a
stale channel but only ever raced two Closes, and the concurrency test left
overlap to goroutine start order. Both now gate their goroutines on a common
start, the stress test races an Open against the Closes, and both assert the
surviving generation channel is actually closed. Waiting on receive functions
to be entered replaces part of the sleep in the reopen probe, and teardown
bounds its Close so a regression fails the assertion instead of hanging.

Two of the four now fail without the fix and no race detector, the stress test
by reproducing close of a closed channel at the Close early return.

* [client] Fail the reopen probe when its teardown does not complete

closeBounded swallowed its timeout and the cleanup discarded what
receiversStopped returned, so the bounds added in the previous commit only
stopped teardown hanging. A wedged Close or a parked receiver would have left
the test green with a leaked goroutine, which is the failure this test exists
to catch.

closeBounded now reports whether Close returned, and cleanup fails the test on
either bound.
2026-09-02 00:27:05 +02:00

433 lines
12 KiB
Go

//go:build !js
package bind
import (
"context"
"encoding/binary"
"fmt"
"net"
"net/netip"
"runtime"
"sync"
"github.com/pion/stun/v3"
"github.com/pion/transport/v3"
log "github.com/sirupsen/logrus"
"golang.org/x/net/ipv6"
wgConn "golang.zx2c4.com/wireguard/conn"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
nbnet "github.com/netbirdio/netbird/client/net"
)
const (
// wgMsgTypeHandshakeInitiation is the lowest WireGuard message type.
wgMsgTypeHandshakeInitiation uint32 = 1
// wgMsgTypeTransport is the highest WireGuard message type.
wgMsgTypeTransport uint32 = 4
// wgMinMsgSize is the smallest WireGuard message: transport data with an empty
// payload, which is what a keepalive is.
wgMinMsgSize = 32
)
type receiverCreator struct {
iceBind *ICEBind
}
func (rc receiverCreator) CreateReceiverFn(pc wgConn.BatchReader, conn *net.UDPConn, rxOffload bool, msgPool *sync.Pool) wgConn.ReceiveFunc {
return rc.iceBind.createReceiverFn(pc, conn, rxOffload, msgPool)
}
// ICEBind is a bind implementation with two main features:
// 1. filter out STUN messages and handle them
// 2. forward the received packets to the WireGuard interface from the relayed connection
//
// ICEBind.endpoints var is a map that stores the connection for each relayed peer. Fake address is just an IP address
// without port, in the format of 127.1.x.x where x.x is the last two octets of the peer address. We try to avoid to
// use the port because in the Send function the wgConn.Endpoint the port info is not exported.
type ICEBind struct {
*wgConn.StdNetBind
transportNet transport.Net
address wgaddr.Address
mtu uint16
endpoints map[netip.Addr]net.Conn
endpointsMu sync.Mutex
recvChan chan recvMessage
// Close() (i.e. BindUpdate()) closes closedChan to release receiveRelayed,
// and the following Open() installs a fresh one. closedChanMu guards both
// closedChan and closed: readers only ever hold it long enough to copy the
// channel, never across a blocking receive, so Open cannot be starved by a
// parked receiver.
closedChan chan struct{}
closedChanMu sync.RWMutex
closed bool
activityRecorder *ActivityRecorder
muUDPMux sync.Mutex
udpMux *udpmux.UniversalUDPMuxDefault
ipv4Conn *net.UDPConn
ipv6Conn *net.UDPConn
}
func NewICEBind(transportNet transport.Net, address wgaddr.Address, mtu uint16) *ICEBind {
b, _ := wgConn.NewStdNetBind().(*wgConn.StdNetBind)
ib := &ICEBind{
StdNetBind: b,
transportNet: transportNet,
address: address,
mtu: mtu,
endpoints: make(map[netip.Addr]net.Conn),
recvChan: make(chan recvMessage, 1),
closedChan: make(chan struct{}),
closed: true,
activityRecorder: NewActivityRecorder(),
}
rc := receiverCreator{
ib,
}
ib.StdNetBind = wgConn.NewStdNetBindWithReceiverCreator(rc)
return ib
}
func (s *ICEBind) Open(uport uint16) ([]wgConn.ReceiveFunc, uint16, error) {
s.closedChanMu.Lock()
defer s.closedChanMu.Unlock()
// Open the underlying bind before touching any state, so a failure leaves
// the current generation exactly as it was. Publishing the new generation
// first would strand it: StdNetBind rejects an Open while it is already
// open, and a Close arriving in that window would mark the bind closed
// while this call went on to install live sockets, after which every later
// Close returns early and never shuts them down.
fns, port, err := s.StdNetBind.Open(uport)
if err != nil {
return nil, 0, err
}
// Release whoever is parked on the outgoing generation before replacing it.
// An Open that follows an Open rather than a Close would otherwise leave
// them waiting on a channel no later Close can reach.
if !s.closed {
close(s.closedChan)
}
s.closed = false
s.closedChan = make(chan struct{})
fns = append(fns, s.receiveRelayed)
return fns, port, nil
}
func (s *ICEBind) Close() error {
s.closedChanMu.Lock()
defer s.closedChanMu.Unlock()
if s.closed {
return nil
}
s.closed = true
close(s.closedChan)
s.muUDPMux.Lock()
s.ipv4Conn = nil
s.ipv6Conn = nil
s.udpMux = nil
s.muUDPMux.Unlock()
return s.StdNetBind.Close()
}
// currentClosedChan copies the channel that signals the current Open
// generation is closing. Callers select on the copy so the lock is never held
// across a blocking receive, which would otherwise stall the next Open.
func (s *ICEBind) currentClosedChan() chan struct{} {
s.closedChanMu.RLock()
defer s.closedChanMu.RUnlock()
return s.closedChan
}
func (s *ICEBind) ActivityRecorder() *ActivityRecorder {
return s.activityRecorder
}
// GetICEMux returns the ICE UDPMux that was created and used by ICEBind
func (s *ICEBind) GetICEMux() (*udpmux.UniversalUDPMuxDefault, error) {
s.muUDPMux.Lock()
defer s.muUDPMux.Unlock()
if s.udpMux == nil {
return nil, fmt.Errorf("ICEBind has not been initialized yet")
}
return s.udpMux, nil
}
func (b *ICEBind) SetEndpoint(fakeIP netip.Addr, conn net.Conn) {
b.endpointsMu.Lock()
b.endpoints[fakeIP] = conn
b.endpointsMu.Unlock()
}
func (b *ICEBind) RemoveEndpoint(fakeIP netip.Addr) {
b.endpointsMu.Lock()
defer b.endpointsMu.Unlock()
delete(b.endpoints, fakeIP)
}
func (b *ICEBind) ReceiveFromEndpoint(ctx context.Context, ep *Endpoint, buf []byte) {
closedChan := b.currentClosedChan()
select {
case <-closedChan:
return
case <-ctx.Done():
return
case b.recvChan <- recvMessage{ep, buf}:
}
}
func (b *ICEBind) Send(bufs [][]byte, ep wgConn.Endpoint) error {
b.endpointsMu.Lock()
conn, ok := b.endpoints[ep.DstIP()]
b.endpointsMu.Unlock()
if !ok {
return b.StdNetBind.Send(bufs, ep)
}
for _, buf := range bufs {
if _, err := conn.Write(buf); err != nil {
return err
}
}
return nil
}
func (s *ICEBind) createReceiverFn(pc wgConn.BatchReader, conn *net.UDPConn, rxOffload bool, msgsPool *sync.Pool) wgConn.ReceiveFunc {
s.muUDPMux.Lock()
defer s.muUDPMux.Unlock()
// Detect IPv4 vs IPv6 from connection's local address
if localAddr := conn.LocalAddr().(*net.UDPAddr); localAddr.IP.To4() != nil {
s.ipv4Conn = conn
} else {
s.ipv6Conn = conn
}
s.createOrUpdateMux()
return func(bufs [][]byte, sizes []int, eps []wgConn.Endpoint) (n int, err error) {
msgs := getMessages(msgsPool)
for i := range bufs {
(*msgs)[i].Buffers[0] = bufs[i]
(*msgs)[i].OOB = (*msgs)[i].OOB[:cap((*msgs)[i].OOB)]
}
defer putMessages(msgs, msgsPool)
var numMsgs int
if runtime.GOOS == "linux" || runtime.GOOS == "android" {
if rxOffload {
readAt := len(*msgs) - (wgConn.IdealBatchSize / wgConn.UdpSegmentMaxDatagrams)
//nolint:staticcheck
_, err = pc.ReadBatch((*msgs)[readAt:], 0)
if err != nil {
return 0, err
}
numMsgs, err = wgConn.SplitCoalescedMessages(*msgs, readAt, wgConn.GetGSOSize)
if err != nil {
return 0, err
}
} else {
numMsgs, err = pc.ReadBatch(*msgs, 0)
if err != nil {
return 0, err
}
}
} else {
msg := &(*msgs)[0]
msg.N, msg.NN, _, msg.Addr, err = conn.ReadMsgUDP(msg.Buffers[0], msg.OOB)
if err != nil {
return 0, err
}
numMsgs = 1
}
for i := 0; i < numMsgs; i++ {
msg := &(*msgs)[i]
if ok, err := s.filterOutStunMessages(msg.Buffers, msg.N, msg.Addr); ok {
if err != nil {
log.Debugf("failed to handle STUN packet from %s: %v", msg.Addr, err)
}
// WireGuard reuses sizes and eps across reads and only skips a slot
// whose size is below the minimum message size. Leaving a consumed
// slot untouched makes it process this buffer again under the
// previous packet's length and endpoint.
sizes[i] = 0
continue
}
sizes[i] = msg.N
if sizes[i] == 0 {
continue
}
addrPort := msg.Addr.(*net.UDPAddr).AddrPort()
if isTransportPkg(msg.Buffers, msg.N) {
s.activityRecorder.record(addrPort)
}
ep := &wgConn.StdNetEndpoint{AddrPort: addrPort} // TODO: remove allocation
wgConn.GetSrcFromControl(msg.OOB[:msg.NN], ep)
eps[i] = ep
}
return numMsgs, nil
}
}
// createOrUpdateMux creates or updates the UDP mux with the available connections.
// Must be called with muUDPMux held.
func (s *ICEBind) createOrUpdateMux() {
var muxConn net.PacketConn
switch {
case s.ipv4Conn != nil && s.ipv6Conn != nil:
muxConn = NewDualStackPacketConn(
nbnet.WrapPacketConn(s.ipv4Conn),
nbnet.WrapPacketConn(s.ipv6Conn),
)
case s.ipv4Conn != nil:
muxConn = nbnet.WrapPacketConn(s.ipv4Conn)
case s.ipv6Conn != nil:
muxConn = nbnet.WrapPacketConn(s.ipv6Conn)
default:
return
}
// Don't close the old mux - it doesn't own the underlying connections.
// The sockets are managed by WireGuard's StdNetBind, not by us.
s.udpMux = udpmux.NewUniversalUDPMuxDefault(
udpmux.UniversalUDPMuxParams{
UDPConn: muxConn,
Net: s.transportNet,
WGAddress: s.address,
MTU: s.mtu,
},
)
}
func (s *ICEBind) filterOutStunMessages(buffers [][]byte, n int, addr net.Addr) (bool, error) {
for i := range buffers {
if n > len(buffers[i]) {
continue
}
pkt := buffers[i][:n]
if isWireGuardMsg(pkt) || !stun.IsMessage(pkt) {
continue
}
msg, err := s.parseSTUNMessage(pkt)
if err != nil {
buffers[i] = []byte{}
return true, err
}
s.muUDPMux.Lock()
mux := s.udpMux
s.muUDPMux.Unlock()
if mux != nil {
if muxErr := mux.HandleSTUNMessage(msg, addr); muxErr != nil {
log.Warnf("failed to handle STUN packet: %v", muxErr)
}
}
buffers[i] = []byte{}
return true, nil
}
return false, nil
}
func (s *ICEBind) parseSTUNMessage(raw []byte) (*stun.Message, error) {
msg := &stun.Message{
Raw: raw,
}
if err := msg.Decode(); err != nil {
return nil, err
}
return msg, nil
}
// receiveRelayed is a receive function that is used to receive packets from the relayed connection and forward to the
// WireGuard. Critical part is do not block if the Closed() has been called.
func (c *ICEBind) receiveRelayed(buffs [][]byte, sizes []int, eps []wgConn.Endpoint) (int, error) {
closedChan := c.currentClosedChan()
select {
case <-closedChan:
return 0, net.ErrClosed
case msg, ok := <-c.recvChan:
if !ok {
return 0, net.ErrClosed
}
copy(buffs[0], msg.Buffer)
sizes[0] = len(msg.Buffer)
eps[0] = wgConn.Endpoint(msg.Endpoint)
if isTransportPkg(buffs, sizes[0]) {
if ep, ok := eps[0].(*Endpoint); ok {
c.activityRecorder.record(ep.AddrPort)
}
}
return 1, nil
}
}
func getMessages(msgsPool *sync.Pool) *[]ipv6.Message {
return msgsPool.Get().(*[]ipv6.Message)
}
func putMessages(msgs *[]ipv6.Message, msgsPool *sync.Pool) {
for i := range *msgs {
(*msgs)[i].OOB = (*msgs)[i].OOB[:0]
(*msgs)[i] = ipv6.Message{Buffers: (*msgs)[i].Buffers, OOB: (*msgs)[i].OOB}
}
msgsPool.Put(msgs)
}
// isWireGuardMsg reports whether the packet carries a WireGuard message header: a
// little-endian uint32 message type in the range 1..4, which leaves the three bytes
// after the type byte zero, in a packet long enough to hold any WireGuard message.
//
// A well formed STUN message cannot take that shape. Its length field sits in the two
// bytes the type must leave zero, and for a message of at least wgMinMsgSize bytes that
// field holds at least 12, so the two framings do not overlap. The test has to be this
// tight because stun.IsMessage only looks at the magic cookie, which in a WireGuard
// message overlaps the receiver index: a session whose index happens to equal the cookie
// would otherwise have all of its inbound data misrouted to the STUN handler until the
// next rekey.
func isWireGuardMsg(pkt []byte) bool {
if len(pkt) < wgMinMsgSize {
return false
}
msgType := binary.LittleEndian.Uint32(pkt[:4])
return msgType >= wgMsgTypeHandshakeInitiation && msgType <= wgMsgTypeTransport
}
// isTransportPkg reports whether the packet is WireGuard transport data carrying a
// payload, which is what counts as peer activity. A keepalive holds no payload and is
// exactly wgMinMsgSize bytes.
func isTransportPkg(buffers [][]byte, n int) bool {
if n < 4 || n > len(buffers[0]) {
return false
}
msgType := binary.LittleEndian.Uint32(buffers[0][:4])
return msgType == wgMsgTypeTransport && n > wgMinMsgSize
}