Compare commits

..
Author SHA1 Message Date
JackDoan 0f59f895a4 cpupick: Keep single readers off tun.avoid_cpus without cpupick, warn and ignore bad lists 2026-10-15 12:21:59 -05:00
JackDoan df1c4d9364 cpupick: Make tun.avoid_cpus a pin selection rule instead of a process-wide mask 2026-10-15 12:13:21 -05:00
JackDoan 36579bae03 cpupick: Move AvoidCPUs from util into cpupick as Avoid 2026-10-15 12:00:28 -05:00
Ryan Huber 73d26dd87d tun: Allow excluding CPUs from process threads at startup
Set tun.avoid_cpus to a list of logical CPUs to remove from the affinity of
every thread of the process on Linux, before services start and reader CPUs
are chosen. Later threads inherit the restricted mask; enumeration repeats so
threads created meanwhile are covered. Inherited taskset and cgroup
restrictions are kept. Keeping Nebula off the underlay NIC's receive IRQ CPUs
and their SMT siblings is what this is for; the example configuration shows
how to find them.

Invalid IDs, a list that would leave a thread no CPU, and syscall failures
fail startup. Configuration tests only validate. Other platforms log and
ignore the setting. The exclusion is process-wide, including embedded use, and
not reloadable. It is independent of thread pinning.

New API: util.AvoidCPUs, next to the existing PinThreadToCPU and AllowedCPUs.

Tests run in a child process and check existing and new threads, preserved
inherited masks, validation without changes, and the unix.CPUSet bounds on
32- and 64-bit platforms. No isolated performance claim is made.
2026-10-15 14:40:45 +00:00
27 changed files with 401 additions and 1249 deletions

No files matched your search

+3 -12
View File
@@ -103,23 +103,14 @@ func (cs *ConnectionState) Decrypt(l *slog.Logger, messageCounter uint64, packet
if !result {
return nil, ErrAlreadySeen
}
out, err := cs.decryptPacket(messageCounter, packet, nb)
return cs.acceptDecrypted(l, messageCounter, out, err)
}
// decryptPacket authenticates and decrypts packet in place, leaving the replay window to the caller.
func (cs *ConnectionState) decryptPacket(messageCounter uint64, packet, nb []byte) ([]byte, error) {
return cs.dKey.DecryptDanger(packet[header.Len:header.Len], packet[:header.Len], packet[header.Len:], messageCounter, nb)
}
// acceptDecrypted finishes a decryptPacket result: an authenticated packet is recorded in the replay window, and
// rejected if its counter was recorded since it was checked.
func (cs *ConnectionState) acceptDecrypted(l *slog.Logger, messageCounter uint64, out []byte, err error) ([]byte, error) {
out, err := cs.dKey.DecryptDanger(packet[header.Len:header.Len], packet[:header.Len], packet[header.Len:], messageCounter, nb)
if err != nil {
return nil, err
}
cs.decryptLock.Lock()
result := cs.window.Update(l, messageCounter)
result = cs.window.Update(l, messageCounter)
cs.decryptLock.Unlock()
if !result {
return nil, ErrAlreadySeen
+3 -12
View File
@@ -23,20 +23,11 @@ import (
// runTestHandshake runs a complete IX handshake between two freshly-built
// peers and returns the initiator and responder Results. Used to produce
// real cipher states for tests that need to exercise post-handshake glue.
func runTestHandshake(t testing.TB) (initR, respR *handshake.Result) {
t.Helper()
return runTestHandshakeWith(t, noise.CipherChaChaPoly)
}
// runTestHandshakeWith is runTestHandshake with the data plane cipher picked by the caller.
func runTestHandshakeWith(t testing.TB, cipher noise.CipherFunc) (initR, respR *handshake.Result) {
func runTestHandshake(t *testing.T) (initR, respR *handshake.Result) {
t.Helper()
// Valid around now rather than the cert_test default, which is fixed at process start and lapses a minute
// later, inside a long -count run
now := time.Now()
ca, _, caKey, _ := ct.NewTestCaCert(
cert.Version2, cert.Curve_CURVE25519, now.Add(-time.Minute), now.Add(time.Hour), nil, nil, nil,
cert.Version2, cert.Curve_CURVE25519, time.Time{}, time.Time{}, nil, nil, nil,
)
caPool := ct.NewTestCAPool(ca)
@@ -49,7 +40,7 @@ func runTestHandshakeWith(t testing.TB, cipher noise.CipherFunc) (initR, respR *
require.NoError(t, err)
hsBytes, err := c.MarshalForHandshakes()
require.NoError(t, err)
ncs := noise.NewCipherSuite(noise.DH25519, cipher, noise.HashSHA256)
ncs := noise.NewCipherSuite(noise.DH25519, noise.CipherChaChaPoly, noise.HashSHA256)
cred := handshake.NewCredential(c, hsBytes, priv, ncs)
return func(v cert.Version) *handshake.Credential {
if v == cert.Version2 {
+4 -4
View File
@@ -144,10 +144,10 @@ type fakeConn struct {
rebinds int
}
func (c *fakeConn) Rebind() error { c.rebinds++; return nil }
func (c *fakeConn) LocalAddr() (netip.AddrPort, error) { return netip.AddrPort{}, nil }
func (c *fakeConn) ListenOut(_ udp.EncReader, _ func([][]byte) [][]byte) error { return nil }
func (c *fakeConn) WriteTo(_ []byte, _ netip.AddrPort) error { return nil }
func (c *fakeConn) Rebind() error { c.rebinds++; return nil }
func (c *fakeConn) LocalAddr() (netip.AddrPort, error) { return netip.AddrPort{}, nil }
func (c *fakeConn) ListenOut(_ udp.EncReader, _ func()) error { return nil }
func (c *fakeConn) WriteTo(_ []byte, _ netip.AddrPort) error { return nil }
func (c *fakeConn) WriteBatch(bufs [][]byte, _ []netip.AddrPort) (int, error) {
return len(bufs), nil
}
+75
View File
@@ -0,0 +1,75 @@
//go:build linux && !android && !e2e_testing
package cpupick
import (
"errors"
"fmt"
"os"
"strconv"
"unsafe"
"golang.org/x/sys/unix"
)
// Avoid removes cpus from the affinity of every thread of the process, for use at startup. New threads inherit
// their parent's mask, so it repeats until a pass finds nothing left to change; a thread that exits meanwhile is
// skipped. Every mask is checked before any is changed. With apply false it only validates, for configuration tests.
// This affects the entire process, including other users of an embedded Nebula. Not wired up yet: tun.avoid_cpus
// only steers Default for now.
func Avoid(cpus []int, apply bool) error {
const capacity = int(unsafe.Sizeof(unix.CPUSet{})) * 8
for _, cpu := range cpus {
if cpu < 0 || cpu >= capacity {
return fmt.Errorf("CPU %d is out of range", cpu)
}
}
if len(cpus) == 0 {
return nil
}
for {
masks, err := avoidingMasks(cpus)
if err != nil || !apply || len(masks) == 0 {
return err
}
for tid, mask := range masks {
if err := unix.SchedSetaffinity(tid, &mask); err != nil && !errors.Is(err, unix.ESRCH) {
return fmt.Errorf("set thread %d affinity: %w", tid, err)
}
}
}
}
// avoidingMasks returns the mask without cpus of every thread whose mask has any of them.
func avoidingMasks(cpus []int) (map[int]unix.CPUSet, error) {
entries, err := os.ReadDir("/proc/self/task")
if err != nil {
return nil, err
}
masks := make(map[int]unix.CPUSet, len(entries))
for _, entry := range entries {
tid, err := strconv.Atoi(entry.Name())
if err != nil {
return nil, err
}
var mask unix.CPUSet
err = unix.SchedGetaffinity(tid, &mask)
if errors.Is(err, unix.ESRCH) {
continue
}
if err != nil {
return nil, fmt.Errorf("read thread %d affinity: %w", tid, err)
}
next := mask
for _, cpu := range cpus {
next.Clear(cpu)
}
if next.Count() == 0 {
return nil, fmt.Errorf("avoid_cpus leaves thread %d with no allowed CPU", tid)
}
if next != mask {
masks[tid] = next
}
}
return masks, nil
}
+98
View File
@@ -0,0 +1,98 @@
//go:build linux && !android && !e2e_testing
package cpupick
import (
"fmt"
"os"
"os/exec"
"runtime"
"strconv"
"sync"
"testing"
"unsafe"
"github.com/slackhq/nebula/util"
"github.com/stretchr/testify/require"
"golang.org/x/sys/unix"
)
// TestAvoid changes the affinity of the whole process, so it runs in a child, leaving the test runner untouched.
func TestAvoid(t *testing.T) {
if os.Getenv("NEBULA_CPUPICK_AVOID_TEST") != "1" {
exe, err := os.Executable()
require.NoError(t, err)
cmd := exec.Command(exe, "-test.run=^TestAvoid$", "-test.v")
cmd.Env = append(os.Environ(), "NEBULA_CPUPICK_AVOID_TEST=1")
out, err := cmd.CombinedOutput()
require.NoError(t, err, "%s", out)
return
}
allowed, err := util.AllowedCPUs()
require.NoError(t, err)
if len(allowed) < 2 {
t.Skip("requires at least two allowed CPUs")
}
cpu := allowed[0]
release := make(chan struct{})
var wg sync.WaitGroup
// lockThreads occupies 8 OS threads until release, so the runtime has to start more
lockThreads := func() {
var ready sync.WaitGroup
for range 8 {
ready.Add(1)
wg.Go(func() {
runtime.LockOSThread()
ready.Done()
<-release
})
}
ready.Wait()
}
lockThreads()
require.NoError(t, Avoid([]int{cpu}, false))
after, err := util.AllowedCPUs()
require.NoError(t, err)
require.Equal(t, allowed, after, "configuration check must not mutate affinity")
require.Error(t, Avoid(allowed, true), "must reject an empty mask before applying")
require.NoError(t, Avoid([]int{cpu, cpu}, true))
requireAvoided(t, cpu, allowed)
// Existing threads stay locked while new ones inherit the restriction
lockThreads()
requireAvoided(t, cpu, allowed)
close(release)
wg.Wait()
}
// requireAvoided checks no thread may run on cpu, and every thread still may on some CPU of allowed only.
func requireAvoided(t *testing.T, cpu int, allowed []int) {
t.Helper()
entries, err := os.ReadDir("/proc/self/task")
require.NoError(t, err)
for _, entry := range entries {
tid, err := strconv.Atoi(entry.Name())
require.NoError(t, err)
var mask unix.CPUSet
err = unix.SchedGetaffinity(tid, &mask)
if err == unix.ESRCH {
continue
}
require.NoError(t, err)
require.False(t, mask.IsSet(cpu), "thread %d", tid)
require.Positive(t, mask.Count())
for id := range len(mask) * 64 {
if mask.IsSet(id) {
require.Contains(t, allowed, id, "must not widen the inherited mask")
}
}
}
}
// TestAvoidRange checks the bounds of unix.CPUSet, whose words are 32 bits wide on 32-bit platforms.
func TestAvoidRange(t *testing.T) {
capacity := int(unsafe.Sizeof(unix.CPUSet{})) * 8
require.NoError(t, Avoid([]int{capacity - 1}, false))
require.EqualError(t, Avoid([]int{capacity}, false), fmt.Sprintf("CPU %d is out of range", capacity))
require.EqualError(t, Avoid([]int{-1}, false), "CPU -1 is out of range")
}
+8
View File
@@ -0,0 +1,8 @@
//go:build !linux || android || e2e_testing
package cpupick
import "errors"
// Avoid is unsupported outside Linux: there is no thread affinity API to restrict.
func Avoid(_ []int, _ bool) error { return errors.ErrUnsupported }
+17 -6
View File
@@ -16,10 +16,15 @@
// across distinct physical cores when the topology permits, CPU 0's physical
// core demoted to last resort, and the order rotated by a stable per-instance
// key so co-located instances spread instead of stacking.
//
// CPUs the operator asked to keep readers off (tun.avoid_cpus, typically the
// underlay NIC's receive IRQ CPUs) are removed from the allowed set before
// any of that, so no rule can bring one back.
package cpupick
import (
"log/slog"
"slices"
"github.com/slackhq/nebula/util"
)
@@ -57,12 +62,17 @@ func flatTopology(cpus []int) topology {
// Default computes the pin order for `routines` tun readers. key is any
// stable per-instance value; the bound UDP port is ideal — distinct across
// co-located instances, stable across restarts so benchmark runs stay
// comparable. Returns nil when there is nothing useful to say (no affinity
// support on this platform, lookup failure); callers keep their existing
// fallback spread.
func Default(routines int, key uint64, l *slog.Logger) []int {
// comparable. avoid lists CPUs never to return, matched literally. Returns
// nil when there is nothing useful to say (no affinity support on this
// platform, lookup failure, avoid covering every allowed CPU); callers keep
// their existing fallback spread.
func Default(routines int, key uint64, avoid []int, l *slog.Logger) []int {
allowed, err := util.AllowedCPUs()
if err != nil || len(allowed) == 0 {
if err != nil {
return nil
}
allowed = slices.DeleteFunc(allowed, func(c int) bool { return slices.Contains(avoid, c) })
if len(allowed) == 0 {
return nil
}
perf, signal := perfCPUs(allowed)
@@ -77,7 +87,8 @@ func Default(routines int, key uint64, l *slog.Logger) []int {
if l != nil {
l.Info("chose default pin CPUs for tun readers",
"cpus", cpus[:min(routines, len(cpus))],
"perfSignal", signal)
"perfSignal", signal,
"avoided", avoid)
}
return cpus
}
+35
View File
@@ -0,0 +1,35 @@
//go:build linux && !android && !e2e_testing
package cpupick
import (
"slices"
"testing"
"github.com/slackhq/nebula/util"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestDefaultSkipsAvoided(t *testing.T) {
allowed, err := util.AllowedCPUs()
require.NoError(t, err)
if len(allowed) < 2 {
t.Skip("requires at least two allowed CPUs")
}
avoid := allowed[:len(allowed)/2]
for key := range uint64(16) {
got := Default(len(allowed), key, avoid, nil)
require.NotEmpty(t, got)
for _, cpu := range got {
assert.NotContains(t, avoid, cpu, "key %d", key)
assert.Contains(t, allowed, cpu, "key %d", key)
}
}
assert.Nil(t, Default(2, 0, allowed, nil), "avoiding everything leaves nothing to suggest")
// Without avoid every allowed CPU stays in play
got := Default(len(allowed), 0, nil, nil)
slices.Sort(got)
assert.Equal(t, allowed, got)
}
-28
View File
@@ -5,11 +5,9 @@ package e2e
import (
"fmt"
"net/netip"
"runtime"
"testing"
"time"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/e2e/router"
@@ -572,29 +570,3 @@ func TestCloseTunnelAuthenticated(t *testing.T) {
myControl.Stop()
theirControl.Stop()
}
// TestStagedTunnel completes a handshake and traffic both ways with staged receive. Every received packet,
// handshakes included, uses the receive pipeline on Linux. Other platforms must do the same over the inline fallback.
func TestStagedTunnel(t *testing.T) {
t.Parallel()
received := metrics.GetOrRegisterCounter("flow.rx.inline", nil)
before := received.Count()
staged := m{"tun": m{"staged": m{"workers": 2}}}
ca, _, caKey, _ := cert_test.NewTestCaCert(cert.Version1, cert.Curve_CURVE25519, time.Now(), time.Now().Add(10*time.Minute), nil, nil, []string{})
myControl, myVpnIpNet, myUdpAddr, _ := newSimpleServer(cert.Version1, ca, caKey, "me", "10.128.0.1/24", staged)
theirControl, theirVpnIpNet, theirUdpAddr, _ := newSimpleServer(cert.Version1, ca, caKey, "them", "10.128.0.2/24", staged)
myControl.InjectLightHouseAddr(theirVpnIpNet[0].Addr(), theirUdpAddr)
theirControl.InjectLightHouseAddr(myVpnIpNet[0].Addr(), myUdpAddr)
myControl.Start()
theirControl.Start()
r := router.NewR(t, myControl, theirControl)
defer r.RenderFlow()
assertTunnel(t, myVpnIpNet[0].Addr(), theirVpnIpNet[0].Addr(), myControl, theirControl, r)
myControl.Stop()
theirControl.Stop()
if runtime.GOOS != "linux" {
assert.Zero(t, received.Count()-before)
return
}
assert.GreaterOrEqual(t, received.Count()-before, int64(2))
}
+9 -8
View File
@@ -271,14 +271,6 @@ tun:
# the use_offloads setting controls if Nebula will attempt to enable GSO and GRO for the tun device. Linux only, not reloadable.
#use_offloads: false
# Linux only. staged.workers enables staged receive: each routine hands its received batches to this many decrypt
# workers, then an ordered writer commits them in arrival order through the same replay, roaming and firewall
# handling as inline receive. 0, the default, keeps receive inline. Batches below 256 packets on an idle listener
# stay on the listener. Needs no support from peers. Costs about 24 MiB of buffers per routine at listen.batch 64
# with GRO. Not reloadable.
#staged:
# workers: 0
# Linux only. pin_threads pins each tun reader/encrypt OS thread to a single CPU. This keeps every goroutine's
# batched sends flowing through one XPS-selected NIC TX ring, so packets within a flow stay ordered on the wire
# instead of being sprayed across multiple TX rings and reordered. Not reloadable. Coerced to false if routines <= 1.
@@ -303,6 +295,15 @@ tun:
# - 2
# - 4
# Linux only. avoid_cpus lists CPU IDs the tun readers are never pinned to, exactly as listed (SMT siblings are not
# added for you). The default pin selection skips them, and a cpu_affinity entry naming one is rejected like any
# other invalid entry. Other threads are not restricted, and the setting does nothing with pin_threads: false. An
# invalid list, or one covering every allowed CPU, is ignored with a warning. Not reloadable. To keep Nebula off the
# underlay NIC's receive IRQ CPUs, find them in /proc/interrupts and /proc/irq/<IRQ>/effective_affinity_list, add
# their SMT siblings from /sys/devices/system/cpu/cpu<N>/topology/thread_siblings_list if you want those kept clear
# too, and revisit the list when IRQ placement changes.
#avoid_cpus: []
# Route based MTU overrides, you have known vpn ip paths that can support larger MTUs you can increase/decrease them here
routes:
#- mtu: 8800
-40
View File
@@ -1,40 +0,0 @@
package nebula
import (
"log/slog"
"runtime"
"github.com/slackhq/nebula/config"
)
// Staged receive (tun.staged.workers) moves decryption off each udp listener onto worker goroutines, behind which an
// ordered writer commits every batch in arrival order. It reuses the inline packet handlers: staging changes ownership
// and scheduling, not policy. Each routine has its own workers and writer, so order is per routine. Full queues block
// the listener. A read error drains the batches in flight, while shutdown waits for the workers and then drops what
// is left. Each routine keeps six batches; at listen.batch=64 with GRO their buffers take about 24 MiB per routine.
//
// Metrics are registered only for active pipelines. flow.rx.batches counts batches handed to the writer and
// flow.rx.inline those an idle listener handled itself; their rates show whether staging is used.
// stagedConfig is tun.staged. workers is per routine; zero keeps receive inline.
type stagedConfig struct {
workers int
// inlineBelow is the packet count below which an idle listener skips the worker handoff
inlineBelow int
}
// parseStagedConfig enables staging on Linux only; conns that cannot hand off batches also stay inline.
func parseStagedConfig(c *config.C, l *slog.Logger) stagedConfig {
sc := stagedConfig{workers: c.GetInt("tun.staged.workers", 0)}
if sc.workers <= 0 {
return stagedConfig{}
}
if runtime.GOOS != "linux" {
l.Warn("tun.staged is only supported on Linux; packet processing stays inline")
return stagedConfig{}
}
// 256 is the measured crossover
sc.inlineBelow = 256
l.Info("Staged receive processing enabled", "workers", sc.workers)
return sc
}
-303
View File
@@ -1,303 +0,0 @@
package nebula
import (
"net/netip"
"sync"
"sync/atomic"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/udp"
)
// handoffConn is a udp conn that receives in batches of a fixed shape, so staged receive can have replacement buffers
// ready for every batch it keeps.
type handoffConn interface {
ListenOut(udp.EncReader, func([][]byte) [][]byte) error
HandoffShape() (n, size int)
}
const stagedRxSlots = 6
const rxChunkMin = 16
// rxItem is one received packet. A data packet for an established tunnel has ci set; out and err are written by
// whoever decrypts it.
type rxItem struct {
packet []byte
ci *ConnectionState
counter uint64
via netip.AddrPort
out []byte
err error
}
// rxChunk is a contiguous run of a batch's items for one worker to decrypt.
type rxChunk struct {
items []rxItem
done chan struct{}
}
// decryptRxItems decrypts the data packets among items in place, skipping the replays checkReplays marked. Recording
// the counters is left to readOutsidePackets on the writer, which also handles the packets without a tunnel.
func decryptRxItems(items []rxItem, nb []byte) {
for i := range items {
it := &items[i]
if it.ci == nil || it.err != nil {
continue
}
it.out, it.err = it.ci.decryptPacket(it.counter, it.packet, nb)
}
}
// rxBatch is one recvmmsg batch moving through the staged receive pipeline.
type rxBatch struct {
// items is every packet of the batch in arrival order. Data packets are decrypted by a worker; anything else is
// handled whole by the writer, in its place.
items []rxItem
// bufs is the set of recvmmsg buffers the items point into, owned by the batch from handoff until it is reused.
bufs [][]byte
// chunks are the decrypt jobs, nChunks of them in use.
chunks []rxChunk
nChunks int
}
func (b *rxBatch) reset() {
clear(b.items)
b.items = b.items[:0]
for i := range b.nChunks {
b.chunks[i].items = nil
}
b.nChunks = 0
}
// stagedRx is the staged receive pipeline of one udp listener. deliver and handoff run on the listener goroutine.
type stagedRx struct {
f *Interface
q int
conn handoffConn
// jobs feeds the decrypt workers. The listener is the only producer and closes it once the writer has drained.
jobs chan *rxChunk
workers sync.WaitGroup
// h and hostmapCache are the listener's own. wrxc is the commit context: the writer's, or the listener's while
// nothing is in flight.
h *header.H
hostmapCache map[uint32]*HostInfo
wrxc *rxContext
// cur is the batch the listener is filling. The others wait in free, or travel to the writer on ordered and come
// back through free; inFlight counts those, and the writer drops it only once it is done with a batch.
cur *rxBatch
free chan *rxBatch
ordered chan *rxBatch
inFlight atomic.Int32
writerDone chan struct{}
metricBatches metrics.Counter
metricInline metrics.Counter
}
func newStagedRx(f *Interface, q int, wrxc *rxContext, conn handoffConn) *stagedRx {
p := &stagedRx{
f: f,
q: q,
conn: conn,
h: &header.H{},
hostmapCache: map[uint32]*HostInfo{},
wrxc: wrxc,
free: make(chan *rxBatch, stagedRxSlots),
ordered: make(chan *rxBatch, stagedRxSlots),
writerDone: make(chan struct{}),
metricBatches: metrics.GetOrRegisterCounter("flow.rx.batches", nil),
metricInline: metrics.GetOrRegisterCounter("flow.rx.inline", nil),
}
n := f.staged.workers
p.jobs = make(chan *rxChunk, stagedRxSlots*2*n)
for range n {
p.workers.Go(p.decrypt)
}
bufs, size := conn.HandoffShape()
for i := range stagedRxSlots {
// Two chunks per worker, so the writer can start committing while the workers are still on the rest
b := &rxBatch{items: make([]rxItem, 0, 256), chunks: make([]rxChunk, 2*n)}
for j := range b.chunks {
b.chunks[j].done = make(chan struct{}, 1)
}
if i == 0 {
// The first batch reads into the conn's own buffers; every other one brings its own
p.cur = b
continue
}
b.bufs = make([][]byte, bufs)
for j := range b.bufs {
b.bufs[j] = make([]byte, size)
}
p.free <- b
}
go p.writer()
return p
}
// decrypt is a worker: it decrypts the chunks it is handed and signals each one done.
func (p *stagedRx) decrypt() {
nb := make([]byte, 12)
for c := range p.jobs {
decryptRxItems(c.items, nb)
c.done <- struct{}{}
}
}
// listen runs the conn's receive loop through the pipeline until it fails, then drains the pipeline. The conn hands
// its buffers over, so a batch moves between stages without a copy.
func (p *stagedRx) listen() error {
defer p.close()
return p.conn.ListenOut(p.deliver, p.handoff)
}
// deliver is the listener's per-packet work: classify the packet and add it to the current batch. Nothing here
// decrypts or touches the tun.
func (p *stagedRx) deliver(via netip.AddrPort, packet []byte) {
it := rxItem{packet: packet, via: via}
p.classify(&it)
p.cur.items = append(p.cur.items, it)
}
// classify fills in the tunnel for a plain data packet of an established, non-relayed tunnel, leaving anything else
// for the writer to handle whole. It is speculative: readOutsidePackets resolves the tunnel again on the writer and
// uses the worker's result only when it is still the same one.
func (p *stagedRx) classify(it *rxItem) {
f := p.f
h := p.h
packet := it.packet
if h.Parse(packet) != nil || h.Version != header.Version || h.Type != header.Message || h.Subtype != header.MessageNone {
return
}
if f.myVpnNetworksTable.Contains(it.via.Addr()) {
return
}
hostinfo := f.hostMap.QueryIndexCached(h.RemoteIndex, p.hostmapCache)
if hostinfo == nil || hostinfo.ConnectionState == nil {
return
}
ci := hostinfo.ConnectionState
if len(packet) < header.Len+ci.dKey.Overhead() {
return
}
it.ci, it.counter = ci, h.MessageCounter
}
// handoff ends a batch: the batch takes bufs, and the conn reads the next batch into the buffers of the next free
// batch.
func (p *stagedRx) handoff(bufs [][]byte) [][]byte {
p.cur.bufs = bufs
p.endBatch()
return p.cur.bufs
}
// endBatch hands the batch to the workers and the ordered writer, then takes a free batch.
func (p *stagedRx) endBatch() {
clear(p.hostmapCache)
b := p.cur
n := len(b.items)
if n == 0 {
b.reset()
return
}
p.checkReplays(b.items)
if p.inFlight.Load() == 0 && n < p.f.staged.inlineBelow {
// The writer is idle with nothing in flight, so the listener may commit with its context and flush the tun
p.metricInline.Inc(1)
p.processInline(b, p.wrxc)
b.reset()
return
}
parts := max(1, min(len(b.chunks), n/rxChunkMin))
size := (n + parts - 1) / parts
for off := 0; off < n; off += size {
c := &b.chunks[b.nChunks]
c.items = b.items[off:min(off+size, n)]
b.nChunks++
}
// The jobs go out before the batch does: once the writer has it, it may finish and reset it at any moment
for i := range b.nChunks {
p.jobs <- &b.chunks[i]
}
p.metricBatches.Inc(1)
p.inFlight.Add(1)
p.ordered <- b
// With every batch in flight this waits for the writer, the socket buffer absorbs the backlog meanwhile
p.cur = <-p.free
}
// checkReplays marks the replays among items before anything decrypts them, as Decrypt checks before it decrypts,
// taking each tunnel's lock once per run of its packets rather than once per packet. The window only moves forward
// until the writer records these counters, so a stale check can only let a replay through to acceptDecrypted, which
// rejects it.
func (p *stagedRx) checkReplays(items []rxItem) {
for i := 0; i < len(items); {
ci := items[i].ci
if ci == nil {
i++
continue
}
ci.decryptLock.Lock()
for ; i < len(items) && items[i].ci == ci; i++ {
if !ci.window.Check(p.f.l, items[i].counter) {
items[i].err = ErrAlreadySeen
}
}
ci.decryptLock.Unlock()
}
}
// processInline decrypts and commits a batch on the calling goroutine, then flushes the tun.
func (p *stagedRx) processInline(b *rxBatch, rxc *rxContext) {
decryptRxItems(b.items, rxc.nb)
p.commit(b.items, rxc)
p.f.flushOutside(rxc)
}
// commit runs every item through readOutsidePackets in arrival order, handing it the worker's result for a data
// packet, exactly as inline receive handles them.
func (p *stagedRx) commit(items []rxItem, rxc *rxContext) {
for i := range items {
it := &items[i]
if it.ci != nil {
rxc.decrypted = it
}
p.f.readOutsidePackets(ViaSender{UdpAddr: it.via}, it.packet, rxc)
rxc.decrypted = nil
}
}
// writer is the ordered receive stage: it commits each batch chunk by chunk as the workers finish them, so the commit
// overlaps the workers still on later chunks, then flushes the tun and returns the batch.
func (p *stagedRx) writer() {
defer close(p.writerDone)
for b := range p.ordered {
// Once the tun is going away a write would only log errors, so only wait out the workers
closed := p.f.closed.Load()
for i := range b.nChunks {
c := &b.chunks[i]
<-c.done
if !closed {
p.commit(c.items, p.wrxc)
}
}
if !closed {
p.f.flushOutside(p.wrxc)
}
b.reset()
p.inFlight.Add(-1)
p.free <- b
}
}
// close stops the writer once it has committed everything in flight, in order, then the workers.
func (p *stagedRx) close() {
close(p.ordered)
<-p.writerDone
close(p.jobs)
p.workers.Wait()
}
-219
View File
@@ -1,219 +0,0 @@
//go:build linux && !android && !e2e_testing
package nebula
import (
"fmt"
"net/netip"
"sync/atomic"
"testing"
"time"
"github.com/flynn/noise"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/noiseutil"
"github.com/slackhq/nebula/udp"
"github.com/stretchr/testify/require"
"go.uber.org/goleak"
)
// TestFlowStagedRXOrderAndReplay feeds every packet behind a forgery of it, from another address, and a replay of
// it, then a test packet that jumps the window and the whole stream again, with shuffled workers and a slow tun,
// through small and large batches: the tun sees each segment once, in order, the window marks them all, the forgeries
// neither consume a counter nor roam, and the jump does not get the earlier data packets of its batch rejected.
func TestFlowStagedRXOrderAndReplay(t *testing.T) {
defer goleak.VerifyNone(t, goleak.IgnoreCurrent())
for _, cipher := range []noise.CipherFunc{noiseutil.CipherAESGCM, noise.CipherChaChaPoly} {
for _, batchLen := range []int{8, 1000} {
t.Run(fmt.Sprintf("%s/batch=%d", cipher.CipherName(), batchLen), func(t *testing.T) {
sender, receiver := newFlowTestPair(t, cipher, stagedOn(4))
delayCrypto(receiver, func(n uint64) {
if n%32 == 0 {
time.Sleep(time.Millisecond)
}
})
receiver.tun.delay = 100 * time.Microsecond
batches, want := flowTestBatches(30, 4, 1, 10)
wire := wireFor(t, sender, batches)
var feed [][]byte
for i, p := range wire {
// Corrupt the tag or the payload
bad := append([]byte(nil), p...)
bad[[]int{len(p) - 1, header.Len + 3}[i%2]] ^= 1
feed = append(feed, bad, p, p)
}
// The jump lands in a batch with fresh data before it, at batch=1000
sender.hostinfo.ConnectionState.messageCounter.Add(2 * ReplayWindow)
feed = append(append(feed, sendControl(sender, header.Test, header.TestReply)), wire...)
forged := func(i int) netip.AddrPort {
if i < 3*len(wire) && i%3 == 0 {
return netip.MustParseAddrPort("198.51.100.9:4242")
}
return flowTestUnder
}
rx0 := rxBatches.Count()
conn := &rxConn{wire: feed, batchLen: batchLen, from: forged}
receiver.receive(conn)
require.Positive(t, rxBatches.Count()-rx0)
require.Equal(t, (len(feed)+batchLen-1)/batchLen, conn.swaps, "every batch is handed off")
require.Equal(t, want, receiver.tun.segments(t), "every packet exactly once, in order")
require.Equal(t, flowTestUnder, receiver.hostinfo.GetRemote())
for _, p := range wire {
h := &header.H{}
require.NoError(t, h.Parse(p))
require.False(t, receiver.hostinfo.ConnectionState.window.Check(receiver.f.l, h.MessageCounter), "counter %d must be marked seen", h.MessageCounter)
}
})
}
}
}
// TestFlowStagedRXClose: a CloseTunnel in the middle of a batch is handled in its place, by the writer. The data after
// it in the batch is still delivered from the batch's hostmap cache, as inline receive does.
func TestFlowStagedRXClose(t *testing.T) {
defer goleak.VerifyNone(t, goleak.IgnoreCurrent())
sender, receiver := newFlowTestPair(t, noise.CipherChaChaPoly, stagedOn(2))
batches, want := flowTestBatches(4, 4, 1, 8)
wire := wireFor(t, sender, batches)
half := len(wire) / 2
feed := append(append(append([][]byte{}, wire[:half]...), sendControl(sender, header.CloseTunnel, 0)), wire[half:]...)
receiver.receive(&rxConn{wire: feed, batchLen: len(feed)})
require.Equal(t, want, receiver.tun.segments(t))
require.Nil(t, receiver.f.hostMap.QueryIndex(receiver.hostinfo.localIndexId))
}
// sendControl returns the wire packet of a control message from sender.
func sendControl(sender *flowTestPeer, typ header.MessageType, st header.MessageSubType) []byte {
sender.f.sendNoMetrics(typ, st, sender.hostinfo.ConnectionState, sender.hostinfo, flowTestUnder, nil, make([]byte, 12), make([]byte, mtu), 0)
wire := sender.conn.captured()
sender.conn.pkts, sender.conn.addrs = nil, nil
return wire[len(wire)-1]
}
// receivePaths is inline, staged and, with the threshold, adaptive receive.
func receivePaths() []stagedConfig {
return []stagedConfig{{}, stagedOn(2), adaptive(2)}
}
// TestFlowStagedRXRecvError feeds a RecvError that deletes the tunnel, then a known replay and new data for it, in one
// batch. Every receive path must drop the data through its own view of the hostmap, without touching the replay window,
// and answer each packet with a RecvError, as inline receive does, without decrypting the replay again.
func TestFlowStagedRXRecvError(t *testing.T) {
for _, sc := range receivePaths() {
t.Run(stagedName(sc), func(t *testing.T) {
sender, receiver := newFlowTestPair(t, noise.CipherChaChaPoly, sc)
receiver.f.messageMetrics = newMessageMetrics()
receiver.f.acceptRecvErrorConfig = recvErrorAlways
receiver.f.sendRecvErrorConfig = recvErrorAlways
batches, want := flowTestBatches(1, 1, 1, 33)
wire := wireFor(t, sender, batches)
rxc := newRxContext(receiver.f, 0)
receiver.f.readOutsidePackets(ViaSender{UdpAddr: flowTestUnder}, append([]byte(nil), wire[0]...), rxc)
receiver.f.flushOutside(rxc)
h := &header.H{}
require.NoError(t, h.Parse(wire[0]))
replay := h.MessageCounter
var decrypts atomic.Int64
delayCrypto(receiver, func(n uint64) {
if n == replay {
decrypts.Add(1)
}
})
ctl := header.Encode(make([]byte, header.Len), header.Version, header.RecvError, 0, receiver.hostinfo.remoteIndexId, 0)
conn := &rxConn{wire: append([][]byte{ctl}, wire...), batchLen: len(wire) + 1}
receiver.receive(conn)
require.Nil(t, receiver.f.hostMap.QueryIndex(receiver.hostinfo.localIndexId))
require.Equal(t, want[:1], receiver.tun.segments(t))
require.Zero(t, decrypts.Load(), "known replays must not be decrypted again")
replies := conn.captured()
require.Len(t, replies, len(wire))
for i, reply := range replies {
require.NoError(t, h.Parse(reply))
require.Equal(t, header.RecvError, h.Type)
require.NoError(t, h.Parse(wire[i]))
require.True(t, i == 0 || receiver.hostinfo.ConnectionState.window.Check(nil, h.MessageCounter), "dropped data must not update replay state")
}
})
}
}
// TestFlowStagedRXReplacedTunnel classifies a data packet, then a close and a new handshake replace its tunnel at the
// same index before the writer commits it, as when the listener runs ahead: it must be dropped untouched.
func TestFlowStagedRXReplacedTunnel(t *testing.T) {
sender, receiver := newFlowTestPair(t, noise.CipherChaChaPoly, stagedOn(2))
receiver.f.messageMetrics = newMessageMetrics()
batches, _ := flowTestBatches(1, 1, 1, 1)
wire := wireFor(t, sender, batches)
closePacket := sendControl(sender, header.CloseTunnel, 0)
p := &stagedRx{f: receiver.f, h: &header.H{}, hostmapCache: map[uint32]*HostInfo{}}
item := rxItem{packet: wire[0], via: flowTestUnder}
p.classify(&item)
require.NotNil(t, item.ci)
rxc := newRxContext(receiver.f, 0)
p.processInline(&rxBatch{items: []rxItem{{packet: closePacket, via: flowTestUnder}}}, rxc)
require.Nil(t, receiver.f.hostMap.QueryIndex(receiver.hostinfo.localIndexId))
_, replacement := newFlowTestPair(t, noise.CipherChaChaPoly, stagedOn(2))
receiver.f.hostMap.Lock()
receiver.f.hostMap.unlockedAddHostInfo(replacement.hostinfo, receiver.f)
receiver.f.hostMap.Unlock()
p.processInline(&rxBatch{items: []rxItem{item}}, rxc)
require.Empty(t, receiver.tun.writes)
require.True(t, item.ci.window.Check(nil, item.counter))
require.True(t, replacement.hostinfo.ConnectionState.window.Check(nil, item.counter))
}
// TestFlowStagedRXRealSocket exercises buffer handoff through the real Linux UDP receive loop.
func TestFlowStagedRXRealSocket(t *testing.T) {
defer goleak.VerifyNone(t, goleak.IgnoreCurrent())
sender, receiver := newFlowTestPair(t, noise.CipherChaChaPoly, stagedOn(3))
batches, want := flowTestBatches(60, 4, 1, 8)
wire := wireFor(t, sender, batches)
l := receiver.f.l
conn, err := udp.NewListener(l, udp.Settings{Listen: netip.MustParseAddrPort("127.0.0.1:0"), Batch: 16, Offloads: true})
require.NoError(t, err)
addr, err := conn.LocalAddr()
require.NoError(t, err)
out, err := udp.NewListener(l, udp.Settings{Listen: netip.MustParseAddrPort("127.0.0.1:0"), Batch: 16, Offloads: true})
require.NoError(t, err)
defer out.Close()
// The sender's address is the tunnel's remote, so no roaming
from, err := out.LocalAddr()
require.NoError(t, err)
receiver.hostinfo.remote.Store(&from)
done := make(chan struct{})
go func() { receiver.receive(conn); close(done) }()
// Send in bursts small enough for the default socket buffer
for off := 0; off < len(wire); off += 32 {
end := min(off+32, len(wire))
addrs := make([]netip.AddrPort, end-off)
for i := range addrs {
addrs[i] = addr
}
_, err := out.WriteBatch(wire[off:end], addrs)
require.NoError(t, err)
require.Eventually(t, func() bool { return len(receiver.tun.segments(t)) >= end }, 5*time.Second, 100*time.Microsecond)
}
receiver.f.closed.Store(true)
require.NoError(t, conn.Close())
<-done
require.Equal(t, want, receiver.tun.segments(t))
}
// TestFlowStagedRXInlineBelow: small idle batches stay on the listener, larger batches still use the ordered writer.
func TestFlowStagedRXInlineBelow(t *testing.T) {
defer goleak.VerifyNone(t, goleak.IgnoreCurrent())
sc := adaptive(3)
sc.inlineBelow = 16
sender, receiver := newFlowTestPair(t, noise.CipherChaChaPoly, sc)
batches, want := flowTestBatches(40, 4, 1, 16)
wire := wireFor(t, sender, batches)
inline := metrics.GetOrRegisterCounter("flow.rx.inline", nil)
inline0, rx0 := inline.Count(), rxBatches.Count()
receiver.receive(&rxConn{wire: wire, sizes: []int{8, 8, 8, 128, 128, 8, 8, 128, 4, 128}})
require.Equal(t, want, receiver.tun.segments(t))
require.Positive(t, inline.Count()-inline0)
require.Positive(t, rxBatches.Count()-rx0)
}
-489
View File
@@ -1,489 +0,0 @@
//go:build linux && !android && !e2e_testing
package nebula
import (
"encoding/binary"
"io"
"math"
"net/netip"
"os"
"sync"
"sync/atomic"
"testing"
"time"
"github.com/flynn/noise"
"github.com/gaissmai/bart"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/firewall"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/iputil"
"github.com/slackhq/nebula/noiseutil"
"github.com/slackhq/nebula/overlay/batch"
"github.com/slackhq/nebula/overlay/tio"
"github.com/slackhq/nebula/test"
"github.com/slackhq/nebula/udp"
"github.com/stretchr/testify/require"
"go.uber.org/goleak"
)
// flowTestMSS is the segment payload size of the synthetic TSO superpackets.
const flowTestMSS = 1200
var (
flowTestLocal = netip.MustParseAddr("10.0.0.1")
flowTestRemote = netip.MustParseAddr("10.0.0.2")
flowTestUnder = netip.MustParseAddrPort("192.0.2.2:4242")
)
// flowTCPSuperpacket builds an IPv4/TCP TSO superpacket from local to remote carrying segs segments. Segment i's
// payload starts with the flow id, the superpacket seq and i, so a receiver can check what it got and in which order.
func flowTCPSuperpacket(flow uint16, seq uint32, segs int) tio.Packet {
const hdr = 40
pay := segs * flowTestMSS
b := make([]byte, hdr+pay)
b[0] = 0x45
binary.BigEndian.PutUint16(b[2:4], uint16(min(hdr+pay, 65535)))
b[8] = 64
b[9] = iputil.IPProtocolTCP
copy(b[12:16], flowTestLocal.AsSlice())
copy(b[16:20], flowTestRemote.AsSlice())
binary.BigEndian.PutUint16(b[20:22], 10000+flow)
binary.BigEndian.PutUint16(b[22:24], 80)
binary.BigEndian.PutUint32(b[24:28], seq*uint32(pay))
b[32] = 0x50
b[33] = 0x10 // ACK
binary.BigEndian.PutUint16(b[34:36], 65535)
for i := range segs {
p := b[hdr+i*flowTestMSS:]
binary.BigEndian.PutUint16(p[0:2], flow)
binary.BigEndian.PutUint32(p[2:6], seq)
binary.BigEndian.PutUint16(p[6:8], uint16(i))
}
if segs == 1 {
return tio.Packet{Bytes: b}
}
return tio.Packet{Bytes: b, GSO: tio.GSOInfo{Size: flowTestMSS, HdrLen: hdr, CsumStart: 20, Proto: tio.GSOProtoTCP}}
}
// decodedSeg identifies one segment by the marker flowTCPSuperpacket put in its payload.
type decodedSeg struct {
flow uint16
seq uint32
idx uint16
}
func decodeSeg(p []byte) decodedSeg {
return decodedSeg{flow: binary.BigEndian.Uint16(p[0:2]), seq: binary.BigEndian.Uint32(p[2:6]), idx: binary.BigEndian.Uint16(p[6:8])}
}
// flowTestBatches builds n reads of per superpackets of segs segments, across flows interleaved flows, and the
// segments they carry in order.
func flowTestBatches(n, per, flows, segs int) ([][]tio.Packet, []decodedSeg) {
var batches [][]tio.Packet
var want []decodedSeg
seq := make([]uint32, flows)
for range n {
var b []tio.Packet
for j := range per {
flow := uint16(j % flows)
b = append(b, flowTCPSuperpacket(flow, seq[flow], segs))
for s := range segs {
want = append(want, decodedSeg{flow: flow, seq: seq[flow], idx: uint16(s)})
}
seq[flow]++
}
batches = append(batches, b)
}
return batches, want
}
// flowTestPeer is one end of a tunnel: the interface, the hostinfo for the other end, and its capture points.
type flowTestPeer struct {
f *Interface
hostinfo *HostInfo
conn *flowCaptureConn
tun *flowTunCapture
}
// newFlowTestPair builds two interfaces joined by one real tunnel (keys from an actual handshake). sender sends from
// flowTestLocal to flowTestRemote, receiver is the far side.
func newFlowTestPair(t testing.TB, cipher noise.CipherFunc, sc stagedConfig) (sender, receiver *flowTestPeer) {
t.Helper()
initR, respR := runTestHandshakeWith(t, cipher)
sci, err := newConnectionStateFromResult(initR)
require.NoError(t, err)
rci, err := newConnectionStateFromResult(respR)
require.NoError(t, err)
sender = newFlowTestPeer(t, flowTestLocal, flowTestRemote, sci, 100, 200)
receiver = newFlowTestPeer(t, flowTestRemote, flowTestLocal, rci, 200, 100)
sender.f.staged, receiver.f.staged = sc, sc
return sender, receiver
}
func newFlowTestPeer(t testing.TB, me, them netip.Addr, ci *ConnectionState, localIdx, remoteIdx uint32) *flowTestPeer {
l := test.NewLogger()
hm := newHostMap(l)
hm.preferredRanges.Store(&[]netip.Prefix{})
myNets, myAddrs := &bart.Lite{}, &bart.Lite{}
myNets.Insert(netip.PrefixFrom(me, 24))
myAddrs.Insert(netip.PrefixFrom(me, 32))
myCert := &dummyCert{version: cert.Version1, networks: []netip.Prefix{netip.PrefixFrom(me, 24)}}
theirCert := &dummyCert{version: cert.Version1, networks: []netip.Prefix{netip.PrefixFrom(them, 24)}, name: "peer"}
ci.peerCert = &cert.CachedCertificate{Certificate: theirCert, InvertedGroups: map[string]struct{}{}}
fw := NewFirewall(l, time.Minute, time.Minute, time.Minute, myCert)
require.NoError(t, fw.AddRule(false, firewall.ProtoAny, 0, 0, []string{"any"}, "", "", "", "", ""))
require.NoError(t, fw.AddRule(true, firewall.ProtoAny, 0, 0, []string{"any"}, "", "", "", "", ""))
conn, tun := &flowCaptureConn{}, &flowTunCapture{}
lh := newTestLighthouse()
lh.remoteAllowList.Store(&RemoteAllowList{})
f := &Interface{
hostMap: hm,
outside: conn,
firewall: fw,
lightHouse: lh,
pki: &PKI{},
myVpnNetworksTable: myNets,
myVpnAddrsTable: myAddrs,
routines: 1,
writers: []udp.Conn{conn},
messageMetrics: &MessageMetrics{txExhausted: metrics.NewCounter()},
metricTxDropped: metrics.NilCounter{},
l: l,
}
f.ctx = t.Context()
f.pki.caPool.Store(cert.NewCAPool())
// Never promote or requery, those would put extra test packets on the wire
f.tryPromoteEvery.Store(math.MaxUint32)
f.reQueryEvery.Store(math.MaxUint32)
f.batchers = []*batch.MultiCoalescer{batch.NewMultiCoalescer(tun, l)}
f.connectionManager = &connectionManager{
intf: f,
relayUsed: map[uint32]struct{}{},
relayUsedLock: &sync.RWMutex{},
trafficTimer: NewLockingTimerWheel[uint32](time.Second, time.Minute),
}
f.handshakeManager = NewHandshakeManager(l, hm, lh, conn, defaultHandshakeConfig)
f.handshakeManager.f = f
hostinfo := &HostInfo{
ConnectionState: ci,
vpnAddrs: []netip.Addr{them},
localIndexId: localIdx,
remoteIndexId: remoteIdx,
relayState: RelayState{relayForByAddr: map[netip.Addr]*Relay{}, relayForByIdx: map[uint32]*Relay{}},
}
hostinfo.remote.Store(&flowTestUnder)
hostinfo.buildNetworks(myNets, theirCert)
hm.unlockedAddHostInfo(hostinfo, f)
return &flowTestPeer{f: f, hostinfo: hostinfo, conn: conn, tun: tun}
}
// flowCaptureConn records every packet sent, in order, as copies.
type flowCaptureConn struct {
udp.NoopConn
mu sync.Mutex
pkts [][]byte
addrs []netip.AddrPort
// delay, when set, slows every WriteBatch down so batches pile up in flight
delay time.Duration
}
func (c *flowCaptureConn) WriteBatch(bufs [][]byte, addrs []netip.AddrPort) (int, error) {
time.Sleep(c.delay)
c.mu.Lock()
defer c.mu.Unlock()
for i, b := range bufs {
c.pkts = append(c.pkts, append([]byte(nil), b...))
c.addrs = append(c.addrs, addrs[i])
}
return len(bufs), nil
}
func (c *flowCaptureConn) WriteTo(b []byte, addr netip.AddrPort) error {
_, err := c.WriteBatch([][]byte{b}, []netip.AddrPort{addr})
return err
}
func (c *flowCaptureConn) captured() [][]byte {
c.mu.Lock()
defer c.mu.Unlock()
return append([][]byte(nil), c.pkts...)
}
// flowTunCapture is a tun queue that records every plaintext write.
type flowTunCapture struct {
mu sync.Mutex
writes [][]byte
// delay, when set, slows every write down
delay time.Duration
}
func (c *flowTunCapture) Read() ([]tio.Packet, error) { return nil, io.EOF }
func (c *flowTunCapture) Close() error { return nil }
func (c *flowTunCapture) Write(b []byte) (int, error) {
time.Sleep(c.delay)
c.mu.Lock()
defer c.mu.Unlock()
c.writes = append(c.writes, append([]byte(nil), b...))
return len(b), nil
}
// segments decodes what reached the tun. The coalescer may have merged segments back into superpackets, so this
// walks each payload in MSS steps.
func (c *flowTunCapture) segments(t testing.TB) []decodedSeg {
t.Helper()
c.mu.Lock()
defer c.mu.Unlock()
var out []decodedSeg
for _, w := range c.writes {
pay := w[40:]
require.Zero(t, len(pay)%flowTestMSS, "tun write payload %d bytes", len(pay))
for off := 0; off < len(pay); off += flowTestMSS {
out = append(out, decodeSeg(pay[off:]))
}
}
return out
}
// decryptWire checks every wire packet decrypts on the receiver, in order, with strictly increasing counters, and
// returns what each one carried. It uses its own replay window and copies, so the wire can still be fed to the
// receiver afterwards.
func decryptWire(t testing.TB, receiver *flowTestPeer, wire [][]byte) []decodedSeg {
t.Helper()
peer := receiver.hostinfo.ConnectionState
ci := &ConnectionState{dKey: peer.dKey, window: NewBits(peer.window.length)}
nb := make([]byte, 12)
var last uint64
out := make([]decodedSeg, 0, len(wire))
for i, pkt := range wire {
h := &header.H{}
require.NoError(t, h.Parse(pkt))
require.Equal(t, header.Message, h.Type)
require.Equal(t, receiver.hostinfo.localIndexId, h.RemoteIndex)
require.Greater(t, h.MessageCounter, last, "wire packet %d: counter %d after %d", i, h.MessageCounter, last)
last = h.MessageCounter
plain, err := ci.Decrypt(receiver.f.l, h.MessageCounter, append([]byte(nil), pkt...), nb)
require.NoError(t, err, "wire packet %d", i)
out = append(out, decodeSeg(plain[40:]))
}
return out
}
// flowBatchQueue is a tun queue that hands out scripted Read batches, then fails with io.EOF. Every batch is copied
// into the current receive buffer, so a test sees the buffer ownership the Offload queue has.
type flowBatchQueue struct {
batches [][]tio.Packet
rxBuf []byte
out []tio.Packet
closed chan struct{}
// reads counts the batches handed out
reads atomic.Int64
}
func newFlowBatchQueue(batches [][]tio.Packet) *flowBatchQueue {
return &flowBatchQueue{batches: batches, rxBuf: make([]byte, 8*65536), closed: make(chan struct{})}
}
func (q *flowBatchQueue) Read() ([]tio.Packet, error) {
select {
case <-q.closed:
return nil, os.ErrClosed
default:
}
n := int(q.reads.Load())
if n == len(q.batches) {
return nil, io.EOF
}
q.out = q.out[:0]
off := 0
for _, p := range q.batches[n] {
l := copy(q.rxBuf[off:], p.Bytes)
q.out = append(q.out, tio.Packet{Bytes: q.rxBuf[off : off+l : off+l], GSO: p.GSO})
off += l
}
q.reads.Add(1)
return q.out, nil
}
func (q *flowBatchQueue) Write(b []byte) (int, error) { return len(b), nil }
func (q *flowBatchQueue) Close() error { close(q.closed); return nil }
// delayedCipher changes scheduling through the real cipher interface, only in tests.
type delayedCipher struct {
noiseutil.CipherState
before func(uint64)
}
func (c delayedCipher) EncryptDanger(out, ad, plain []byte, n uint64, nb []byte) ([]byte, error) {
c.before(n)
return c.CipherState.EncryptDanger(out, ad, plain, n, nb)
}
func (c delayedCipher) DecryptDanger(out, ad, ciphertext []byte, n uint64, nb []byte) ([]byte, error) {
c.before(n)
return c.CipherState.DecryptDanger(out, ad, ciphertext, n, nb)
}
func delayCrypto(peer *flowTestPeer, before func(uint64)) {
ci := peer.hostinfo.ConnectionState
ci.eKey = delayedCipher{ci.eKey, before}
ci.dKey = delayedCipher{ci.dKey, before}
}
// stagedOn enables staging with n workers.
func stagedOn(n int) stagedConfig {
return stagedConfig{workers: n}
}
// adaptive is staging with the inline threshold parseStagedConfig sets.
func adaptive(n int) stagedConfig {
return stagedConfig{workers: n, inlineBelow: 256}
}
func stagedName(sc stagedConfig) string {
switch {
case sc.workers == 0:
return "inline"
case sc.inlineBelow > 0:
return "adaptive"
default:
return "staged"
}
}
// TestFlowStagedFallback: a conn that cannot hand off its batches keeps receive on the inline path.
func TestFlowStagedFallback(t *testing.T) {
defer goleak.VerifyNone(t, goleak.IgnoreCurrent())
sender, receiver := newFlowTestPair(t, noise.CipherChaChaPoly, stagedOn(3))
batches, want := flowTestBatches(10, 2, 1, 4)
wire := wireFor(t, sender, batches)
require.Equal(t, want, decryptWire(t, receiver, wire))
rx0 := rxBatches.Count()
receiver.receive(plainConn{&rxConn{wire: wire, batchLen: 16}})
require.Equal(t, want, receiver.tun.segments(t))
require.Equal(t, rx0, rxBatches.Count())
}
// wireFor produces the wire packets the sender emits for batches.
func wireFor(t testing.TB, sender *flowTestPeer, batches [][]tio.Packet) [][]byte {
t.Helper()
sender.f.listenIn(newFlowBatchQueue(batches), 0)
wire := sender.conn.captured()
sender.conn.pkts, sender.conn.addrs = nil, nil
return wire
}
// rxConn plays wire back to listenOut through the buffer handoff, in batches of batchLen or of the scripted sizes
// (the last one repeating), and captures what is sent back. Every buffer handed back is scribbled over before it is
// read into, so a stage still using one shows corruption. from, when set, picks each packet's source address by its
// index in wire.
type rxConn struct {
flowCaptureConn
wire [][]byte
batchLen int
sizes []int
from func(i int) netip.AddrPort
swaps int
}
func (c *rxConn) HandoffShape() (n, size int) {
n = c.batchLen
for _, s := range c.sizes {
n = max(n, s)
}
return n, udp.MTU
}
func (c *rxConn) ListenOut(r udp.EncReader, handoff func(bufs [][]byte) [][]byte) error {
n, _ := c.HandoffShape()
bufs := make([][]byte, n)
for i := range bufs {
bufs[i] = make([]byte, udp.MTU)
}
for off, k := 0, 0; off < len(c.wire); k++ {
size := c.batchLen
if len(c.sizes) > 0 {
size = c.sizes[min(k, len(c.sizes)-1)]
}
end := min(off+size, len(c.wire))
for i, p := range c.wire[off:end] {
n := copy(bufs[i], p)
from := flowTestUnder
if c.from != nil {
from = c.from(off + i)
}
r(from, bufs[i][:n:n])
}
next := handoff(bufs)
if len(next) != len(bufs) {
panic("handoff returned the wrong number of buffers")
}
bufs = next
for _, b := range bufs {
for j := range b[:udp.MTU] {
b[j] = 0x5a
}
}
c.swaps++
off = end
}
return os.ErrClosed
}
// plainConn hides HandoffShape, so receive stays inline.
type plainConn struct{ udp.Conn }
var rxBatches = metrics.GetOrRegisterCounter("flow.rx.batches", nil)
// receive runs the peer's listenOut over conn until conn runs out.
func (p *flowTestPeer) receive(conn udp.Conn) {
p.f.writers = []udp.Conn{conn}
p.f.outside = conn
p.f.listenOut(0)
}
// TestFlowStagedFirewallParity: both receive paths apply the same inbound policy. An authenticated inbound denial
// still consumes its replay counter, exactly as inline Decrypt does.
func TestFlowStagedFirewallParity(t *testing.T) {
for _, sc := range []stagedConfig{{}, stagedOn(2)} {
t.Run(stagedName(sc), func(t *testing.T) {
batches, _ := flowTestBatches(4, 3, 1, 8)
sender, receiver := newFlowTestPair(t, noise.CipherChaChaPoly, sc)
wire := wireFor(t, sender, batches)
receiver.f.firewall.InRules = newFirewallTable()
receiver.receive(&rxConn{wire: wire, batchLen: 64})
require.Empty(t, receiver.tun.writes)
for _, pkt := range wire {
h := &header.H{}
require.NoError(t, h.Parse(pkt))
require.False(t, receiver.hostinfo.ConnectionState.window.Check(receiver.f.l, h.MessageCounter))
}
})
}
}
// TestFlowStagedAESGCM runs the parsed configuration with the real AES cipher. Receive decryption has no nonce
// ordering requirement, so it stays staged in BoringCrypto and FIPS builds too.
func TestFlowStagedAESGCM(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
c.Settings = map[string]any{"tun": map[string]any{"staged": map[string]any{"workers": 2}}}
sender, receiver := newFlowTestPair(t, noiseutil.CipherAESGCM, parseStagedConfig(c, l))
require.Equal(t, 2, receiver.f.staged.workers)
batches, want := flowTestBatches(4, 8, 1, 45)
wire := wireFor(t, sender, batches)
rx0 := rxBatches.Count()
receiver.receive(&rxConn{wire: wire, batchLen: 512})
require.Greater(t, rxBatches.Count(), rx0)
require.Equal(t, want, receiver.tun.segments(t))
}
+15 -19
View File
@@ -60,12 +60,14 @@ type InterfaceConfig struct {
// shorter lists than `routines` cycle. Empty list keeps the default
// pin-to-(i % NumCPU) behavior. Only consulted when PinThreads is true.
CpuAffinity []int
// AvoidCPUs (tun.avoid_cpus) lists CPUs the default pin spread skips when
// CpuAffinity is empty. Only consulted when PinThreads is true.
AvoidCPUs []int
// PinThreads controls whether each TUN reader OS thread is pinned to a
// single CPU (via tun.pin_threads, default true). Pinning keeps each
// goroutine's sendmmsg on one XPS-selected NIC TX ring so per-flow
// packets stay ordered on the wire.
PinThreads bool
staged stagedConfig
l *slog.Logger
}
@@ -96,11 +98,12 @@ type Interface struct {
// Empty falls back to the default pin-to-(allowed CPU) behavior.
// Only consulted when pinThreads is true.
cpuAffinity []int
// avoidCPUs are removed from that fallback spread, unless they would empty it.
avoidCPUs []int
// pinThreads controls whether listenIn pins each TUN reader OS thread to
// a CPU at all (tun.pin_threads, default true). When false, threads are
// left free to migrate as on stock nebula.
pinThreads bool
staged stagedConfig
relayManager *relayManager
tryPromoteEvery atomic.Uint32
@@ -249,8 +252,8 @@ func NewInterface(ctx context.Context, c *InterfaceConfig) (*Interface, error) {
connectionManager: c.connectionManager,
conntrackCacheTimeout: c.ConntrackCacheTimeout,
cpuAffinity: c.CpuAffinity,
avoidCPUs: c.AvoidCPUs,
pinThreads: c.PinThreads,
staged: c.staged,
udpTxErrLimiter: udpTxErrLimiter,
udpRelayTxErrLimiter: udpRelayTxErrLimiter,
txOversizeLimiter: txOversizeLimiter,
@@ -387,8 +390,6 @@ type rxContext struct {
hostmapCache map[uint32]*HostInfo
lhh *LightHouseHandler
ctCache *firewall.ConntrackCacheTicker
// decrypted is staged receive's result for the packet being handled, decrypted ahead of time on a worker
decrypted *rxItem
}
func newRxContext(f *Interface, q int) *rxContext {
@@ -418,21 +419,14 @@ func (f *Interface) listenOut(i int) {
f.readOutsidePackets(ViaSender{UdpAddr: fromUdpAddr}, payload, rxc)
}
flusher := func(bufs [][]byte) [][]byte {
f.flushOutside(rxc)
return bufs
flusher := func() {
if err := f.batchers[i].Flush(); err != nil {
f.l.Error("Failed to flush tun coalescer", "error", err)
}
clear(rxc.hostmapCache)
}
var err error
if hc, ok := li.(handoffConn); ok && f.staged.workers > 0 {
// Every stage on its own goroutine, see flow_rx.go. rxc becomes the ordered writer's context.
err = newStagedRx(f, i, rxc, hc).listen()
} else {
if f.staged.workers > 0 {
f.l.Info("udp conn cannot hand off receive batches, receive stays inline", "reader", i)
}
err = li.ListenOut(listener, flusher)
}
err := li.ListenOut(listener, flusher)
// An error after teardown began is shutdown noise, the closed flag covers resources
// Close releases itself and the cancelled ctx covers ones torn down by their owners
@@ -449,12 +443,14 @@ func (f *Interface) pinThisThread(i int) {
var cpu int
if n := len(f.cpuAffinity); n > 0 {
// Explicit tun.cpu_affinity list wins; parseCpuAffinity already
// validated the entries against the allowed CPU set.
// validated the entries against the allowed CPU set and tun.avoid_cpus.
cpu = f.cpuAffinity[i%n]
} else if allowed, err := util.AllowedCPUs(); err == nil && len(allowed) > 0 {
// Default: spread queues across the CPUs we're actually allowed to
// run on. Under a cpuset/taskset mask these aren't 0..NumCPU-1, so
// i % NumCPU would pick unrunnable IDs and every pin would fail.
// parseAvoidCPUs already dropped an avoid list that covers them all.
allowed = slices.DeleteFunc(allowed, func(c int) bool { return slices.Contains(f.avoidCPUs, c) })
cpu = allowed[i%len(allowed)]
} else {
cpu = i % runtime.NumCPU()
+20
View File
@@ -7,6 +7,7 @@ import (
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/test"
"github.com/slackhq/nebula/util"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
@@ -118,3 +119,22 @@ func TestReloadFirewall_NoChange(t *testing.T) {
assert.Same(t, fw, f.firewall, "firewall should not have been replaced")
}
// TestPinThisThreadFallbackSkipsAvoided covers the single reader, which never goes through cpupick: its fallback
// spread over the allowed set must still skip tun.avoid_cpus.
func TestPinThisThreadFallbackSkipsAvoided(t *testing.T) {
allowed, err := util.AllowedCPUs()
require.NoError(t, err)
if len(allowed) < 2 {
t.Skip("requires at least two allowed CPUs")
}
f := &Interface{l: test.NewLogger(), pinThreads: true, avoidCPUs: []int{allowed[0]}}
got := make(chan []int)
go func() {
// The pin leaves this goroutine locked to its thread, so the runtime discards the thread when it returns.
f.pinThisThread(0)
cpus, _ := util.AllowedCPUs()
got <- cpus
}()
assert.Equal(t, []int{allowed[1]}, <-got)
}
+52 -8
View File
@@ -243,12 +243,19 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
}
pinThreads := c.GetBool("tun.pin_threads", true)
cpuAffinity := parseCpuAffinity(c, l, routines)
staged := parseStagedConfig(c, l)
avoidCPUs := parseAvoidCPUs(c, l)
if len(avoidCPUs) > 0 && !pinThreads {
l.Warn("tun.avoid_cpus has no effect with tun.pin_threads disabled")
}
// Keeping every thread of the process off avoidCPUs, not just the pinned readers, is parked for now:
//if err := cpupick.Avoid(avoidCPUs, !configTest); err != nil && !errors.Is(err, errors.ErrUnsupported) {
// return nil, fmt.Errorf("tun.avoid_cpus: %w", err)
//}
cpuAffinity := parseCpuAffinity(c, l, routines, avoidCPUs)
if pinThreads && routines > 1 && len(cpuAffinity) == 0 && !configTest {
// The operator didn't choose pin CPUs, so pick a default set that
// prefers performance cores and doesn't stack co-located instances
// onto allowed[0].
// prefers performance cores, skips tun.avoid_cpus and doesn't stack
// co-located instances onto allowed[0].
// key is used to seed the spreading of routines->cores.
// use PID if you want to ensure many different Nebulas in VMs or containers land on different cores
@@ -271,7 +278,7 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
l.Warn("tun.pin_threads_key is invalid, using PID")
}
cpuAffinity = cpupick.Default(routines, key, l)
cpuAffinity = cpupick.Default(routines, key, avoidCPUs, l)
}
ifConfig := &InterfaceConfig{
@@ -296,8 +303,8 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
punchy: punchy,
ConntrackCacheTimeout: conntrackCacheTimeout,
CpuAffinity: cpuAffinity,
AvoidCPUs: avoidCPUs,
PinThreads: pinThreads,
staged: staged,
l: l,
}
@@ -365,8 +372,10 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
// mask (util.AllowedCPUs) rather than 0..NumCPU-1: under a cgroup cpuset or
// taskset the runnable IDs are frequently not that contiguous range, and
// pinning to an unrunnable ID always fails. If the allowed set can't be
// determined we fall back to a plain non-negative check.
func parseCpuAffinity(c *config.C, l *slog.Logger, routines int) []int {
// determined we fall back to a plain non-negative check. An entry listed in
// avoid (tun.avoid_cpus) contradicts the operator's own config and is treated
// the same as any other invalid entry.
func parseCpuAffinity(c *config.C, l *slog.Logger, routines int, avoid []int) []int {
raw := c.Get("tun.cpu_affinity")
if raw == nil {
return nil
@@ -409,6 +418,11 @@ func parseCpuAffinity(c *config.C, l *slog.Logger, routines int) []int {
"index", i, "cpu", cpu, "allowed", allowed)
return nil
}
if slices.Contains(avoid, cpu) {
l.Warn("tun.cpu_affinity entry is listed in tun.avoid_cpus; ignoring affinity",
"index", i, "cpu", cpu)
return nil
}
cpus = append(cpus, cpu)
}
if len(cpus) != routines {
@@ -418,6 +432,36 @@ func parseCpuAffinity(c *config.C, l *slog.Logger, routines int) []int {
return cpus
}
// parseAvoidCPUs reads tun.avoid_cpus, a list of CPU IDs the tun readers must not be pinned to. IDs outside the
// allowed set are accepted, since one config may serve differently shaped hosts. Like tun.cpu_affinity, an invalid
// list is a warning and is ignored entirely, as is one covering every allowed CPU, which would leave nowhere to pin.
func parseAvoidCPUs(c *config.C, l *slog.Logger) []int {
raw := c.Get("tun.avoid_cpus")
if raw == nil {
return nil
}
values, ok := raw.([]any)
if !ok {
l.Warn("tun.avoid_cpus must be a list of integers; ignoring", "value", raw)
return nil
}
cpus := make([]int, 0, len(values))
for i, value := range values {
cpu, ok := value.(int)
if !ok || cpu < 0 {
l.Warn("tun.avoid_cpus entry not a non-negative integer; ignoring avoid_cpus", "index", i, "value", value)
return nil
}
cpus = append(cpus, cpu)
}
if allowed, err := util.AllowedCPUs(); err == nil && len(allowed) > 0 &&
!slices.ContainsFunc(allowed, func(cpu int) bool { return !slices.Contains(cpus, cpu) }) {
l.Warn("tun.avoid_cpus covers every allowed CPU; ignoring avoid_cpus", "cpus", cpus, "allowed", allowed)
return nil
}
return cpus
}
func moduleVersion() string {
info, ok := debug.ReadBuildInfo()
if !ok {
+36 -26
View File
@@ -5,7 +5,6 @@ import (
"net/netip"
"os"
"path/filepath"
"runtime"
"testing"
"time"
@@ -33,7 +32,7 @@ func TestParseCpuAffinity(t *testing.T) {
}
// unset -> nil (listenIn falls back to spreading across the allowed set)
assert.Nil(t, parseCpuAffinity(newConfig(nil), l, 1))
assert.Nil(t, parseCpuAffinity(newConfig(nil), l, 1, nil))
// Pick a CPU we're actually allowed to run on so a valid list survives
// validation regardless of the host's affinity mask.
@@ -44,19 +43,50 @@ func TestParseCpuAffinity(t *testing.T) {
}
// valid list -> parsed through unchanged
assert.Equal(t, []int{validCPU, validCPU}, parseCpuAffinity(newConfig([]any{validCPU, validCPU}), l, 2))
assert.Equal(t, []int{validCPU, validCPU}, parseCpuAffinity(newConfig([]any{validCPU, validCPU}), l, 2, nil))
// a negative entry is out of range on every platform -> disables the override
assert.Nil(t, parseCpuAffinity(newConfig([]any{validCPU, -1}), l, 2))
assert.Nil(t, parseCpuAffinity(newConfig([]any{validCPU, -1}), l, 2, nil))
// a non-integer entry -> disables the override
assert.Nil(t, parseCpuAffinity(newConfig([]any{validCPU, "not-a-cpu"}), l, 2))
assert.Nil(t, parseCpuAffinity(newConfig([]any{validCPU, "not-a-cpu"}), l, 2, nil))
// a CPU id outside the allowed set -> disables the override. Only assertable
// where we can enumerate the allowed set (e.g. linux); 1<<20 is far beyond
// any representable CPU id so it can never be in the mask.
if len(allowed) > 0 {
assert.Nil(t, parseCpuAffinity(newConfig([]any{1 << 20}), l, 1))
assert.Nil(t, parseCpuAffinity(newConfig([]any{1 << 20}), l, 1, nil))
}
// an entry also listed in tun.avoid_cpus -> disables the override
assert.Nil(t, parseCpuAffinity(newConfig([]any{validCPU}), l, 1, []int{validCPU}))
}
func TestParseAvoidCPUs(t *testing.T) {
l := test.NewLogger()
for _, tc := range []struct {
value any
want []int
}{
{nil, nil}, {[]any{}, []int{}}, {[]any{1, 1 << 20}, []int{1, 1 << 20}},
{"auto", nil}, {"1,2", nil}, {[]any{-1}, nil}, {[]any{1.5}, nil}, {[]any{1, "2"}, nil},
} {
c := config.NewC(l)
if tc.value != nil {
c.Settings["tun"] = map[string]any{"avoid_cpus": tc.value}
}
assert.Equal(t, tc.want, parseAvoidCPUs(c, l), "%v", tc.value)
}
// avoiding every allowed CPU leaves nowhere to pin; only checkable where the allowed set is known
if allowed, _ := util.AllowedCPUs(); len(allowed) > 0 {
all := make([]any, len(allowed))
for i, cpu := range allowed {
all[i] = cpu
}
c := config.NewC(l)
c.Settings["tun"] = map[string]any{"avoid_cpus": all}
assert.Nil(t, parseAvoidCPUs(c, l))
}
}
@@ -124,23 +154,3 @@ firewall:
require.NoError(t, err)
require.Nil(t, ctrl, "a config test hands back nothing to stop, so it must stop itself")
}
func TestParseStagedConfig(t *testing.T) {
l := test.NewLogger()
staged := func(workers any) stagedConfig {
c := config.NewC(l)
if workers != nil {
c.Settings["tun"] = map[string]any{"staged": map[string]any{"workers": workers}}
}
return parseStagedConfig(c, l)
}
for _, workers := range []any{nil, -1, 0} {
require.Equal(t, stagedConfig{}, staged(workers))
}
if runtime.GOOS != "linux" {
require.Equal(t, stagedConfig{}, staged(2))
return
}
want := stagedConfig{workers: 2, inlineBelow: 256}
require.Equal(t, want, staged(2))
}
+1 -15
View File
@@ -129,13 +129,7 @@ func (f *Interface) readOutsidePackets(via ViaSender, packet []byte, rxc *rxCont
return
}
var out []byte
if pre := rxc.decrypted; pre != nil && pre.ci == hostinfo.ConnectionState {
// Staged receive already checked and decrypted this packet on a worker, only the replay window is left
out, err = pre.ci.acceptDecrypted(f.l, h.MessageCounter, pre.out, pre.err)
} else {
out, err = hostinfo.ConnectionState.Decrypt(f.l, h.MessageCounter, packet, rxc.nb)
}
out, err := hostinfo.ConnectionState.Decrypt(f.l, h.MessageCounter, packet, rxc.nb)
if err != nil {
if f.l.Enabled(context.Background(), slog.LevelDebug) {
hostinfo.logger(f.l).Debug("Failed to decrypt packet", "error", err, "from", via, "header", h)
@@ -479,14 +473,6 @@ func parseV4(data []byte, incoming bool, fp *firewall.ParsedPacket) error {
return nil
}
// flushOutside ends a receive batch: it writes out the tun coalescer and drops the batch's hostmap cache.
func (f *Interface) flushOutside(rxc *rxContext) {
if err := f.batchers[rxc.q].Flush(); err != nil {
f.l.Error("Failed to flush tun coalescer", "error", err)
}
clear(rxc.hostmapCache)
}
func (f *Interface) handleOutsideMessagePacket(hostinfo *HostInfo, messageCounter uint64, out []byte, rxc *rxContext) {
err := newPacket(out, true, rxc.fwPacket)
if err != nil {
+6 -7
View File
@@ -29,12 +29,11 @@ type Settings struct {
type Conn interface {
Rebind() error
LocalAddr() (netip.AddrPort, error)
// ListenOut invokes r for each received packet. The packets alias the receive buffers.
// After each batch is fully delivered (after each packet on single-packet backends), handoff takes the batch's
// buffers and returns the ones to receive the next batch into: the same ones once the caller is done with the
// packets, after flushing per-batch accumulators such as TUN write coalescers, or replacements of the same count
// and size to keep them. handoff must not be nil.
ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) error
// ListenOut invokes r for each received packet.
// On batch-capable backends (recvmmsg), flush is called after each batch is fully delivered.
// Callers use it to flush per-batch accumulators such as TUN write coalescers.
// Single-packet backends call flush after each packet. flush must not be nil.
ListenOut(r EncReader, flush func()) error
WriteTo(b []byte, addr netip.AddrPort) error
// WriteBatch sends a contiguous batch of packets, each with its own
// destination. bufs and addrs must have the same length. Linux uses
@@ -57,7 +56,7 @@ func (NoopConn) Rebind() error {
func (NoopConn) LocalAddr() (netip.AddrPort, error) {
return netip.AddrPort{}, nil
}
func (NoopConn) ListenOut(_ EncReader, _ func([][]byte) [][]byte) error {
func (NoopConn) ListenOut(_ EncReader, _ func()) error {
return nil
}
func (NoopConn) SupportsMultipleReaders() bool {
+3 -5
View File
@@ -180,9 +180,8 @@ func NewUDPStatsEmitter(udpConns []Conn) func() {
return func() {}
}
func (u *StdConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) error {
bufs := [][]byte{make([]byte, MTU)}
buffer := bufs[0]
func (u *StdConn) ListenOut(r EncReader, flush func()) error {
buffer := make([]byte, MTU)
for {
// Just read one packet at a time
@@ -196,8 +195,7 @@ func (u *StdConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) e
}
r(netip.AddrPortFrom(rua.Addr().Unmap(), rua.Port()), buffer[:n:n])
bufs = handoff(bufs)
buffer = bufs[0]
flush()
}
}
+3 -5
View File
@@ -86,9 +86,8 @@ type rawMessage struct {
Len uint32
}
func (u *GenericConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) error {
bufs := [][]byte{make([]byte, MTU)}
buffer := bufs[0]
func (u *GenericConn) ListenOut(r EncReader, flush func()) error {
buffer := make([]byte, MTU)
var lastRecvErr time.Time
@@ -108,8 +107,7 @@ func (u *GenericConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byt
}
r(netip.AddrPortFrom(rua.Addr().Unmap(), rua.Port()), buffer[:n:n])
bufs = handoff(bufs)
buffer = bufs[0]
flush()
}
}
+4 -24
View File
@@ -245,18 +245,11 @@ func getFrom(names [][]byte, i int, isV4 bool) netip.AddrPort {
return netip.AddrPortFrom(ip.Unmap(), binary.BigEndian.Uint16(names[i][2:4]))
}
// HandoffShape is the count and size of the buffers ListenOut hands over after each batch.
func (u *StdConn) HandoffShape() (n, size int) {
if u.groSupported {
return u.batch, udpGROBufferSize
}
return u.batch, MTU
}
func (u *StdConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) error {
_, bufSize := u.HandoffShape()
func (u *StdConn) ListenOut(r EncReader, flush func()) error {
bufSize := MTU
cmsgSpace := 0
if u.groSupported {
bufSize = udpGROBufferSize
cmsgSpace = unix.CmsgSpace(udpGROCmsgPayload)
}
msgs, buffers, names, _ := prepareRawMessages(u.batch, bufSize, cmsgSpace)
@@ -290,20 +283,7 @@ func (u *StdConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) e
deliverSegments(r, from, payload, segSize)
}
next := handoff(buffers)
if len(next) != len(buffers) {
panic(fmt.Sprintf("udp: handoff returned %d buffers, need %d", len(next), len(buffers)))
}
if &next[0] != &buffers[0] {
// Replacements, point the recvmmsg iovecs at them
for i := range next {
if len(next[i]) < bufSize {
panic(fmt.Sprintf("udp: handoff buffer is %d bytes, need %d", len(next[i]), bufSize))
}
msgs[i].Hdr.Iov.Base = &next[i][0]
}
buffers = next
}
flush()
}
}
+1 -1
View File
@@ -226,7 +226,7 @@ func TestOffloadsDisabledRxDelivers(t *testing.T) {
_ = sc.ListenOut(func(from netip.AddrPort, payload []byte) {
// payload aliases the shared recv buffer row; copy before handing off.
rxCh <- rxPkt{from, slices.Clone(payload)}
}, func(bufs [][]byte) [][]byte { return bufs })
}, func() {})
}()
tx, err := net.DialUDP("udp4", nil, net.UDPAddrFromAddrPort(addr))
+1 -1
View File
@@ -142,7 +142,7 @@ func runTeardownCase(t *testing.T, batch int, name string, traffic func(send net
go func() {
loopDone <- sc.ListenOut(func(netip.AddrPort, []byte) {
received.Add(1)
}, func(bufs [][]byte) [][]byte { return bufs })
}, func() {})
}()
send, err := net.Dial("udp", addr.String())
+3 -5
View File
@@ -142,9 +142,8 @@ func (u *RIOConn) bind(l *slog.Logger, sa windows.Sockaddr) error {
return nil
}
func (u *RIOConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) error {
bufs := [][]byte{make([]byte, MTU)}
buffer := bufs[0]
func (u *RIOConn) ListenOut(r EncReader, flush func()) error {
buffer := make([]byte, MTU)
var lastRecvErr time.Time
@@ -165,8 +164,7 @@ func (u *RIOConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) e
}
r(netip.AddrPortFrom(netip.AddrFrom16(rua.Addr).Unmap(), (rua.Port>>8)|((rua.Port&0xff)<<8)), buffer[:n:n])
bufs = handoff(bufs)
buffer = bufs[0]
flush()
}
}
+4 -12
View File
@@ -182,24 +182,16 @@ func (u *TesterConn) WriteBatch(bufs [][]byte, addrs []netip.AddrPort) (int, err
return written, nil
}
// HandoffShape is one buffer per handoff; the pool resizes the replacements it gets back as it needs.
func (u *TesterConn) HandoffShape() (n, size int) { return 1, MTU }
// ListenOut hands each packet's buffer over as a batch of one, the buffer it gets back goes to the pool. Like StdConn
// it carries on with the slice handoff returns, since a reader that keeps the packets keeps the one it was given.
// The batcher borrows plaintext decrypted in place inside p.Data until it flushes, inside handoff, so the packet
// must stay alive until handoff returns.
func (u *TesterConn) ListenOut(r EncReader, handoff func(bufs [][]byte) [][]byte) error {
bufs := make([][]byte, 1)
func (u *TesterConn) ListenOut(r EncReader, flush func()) error {
for {
select {
case <-u.done:
return os.ErrClosed
case p := <-u.RxPackets:
r(p.From, p.Data[:len(p.Data):len(p.Data)])
bufs[0] = p.Data
bufs = handoff(bufs)
p.Data = bufs[0]
// The batcher borrows plaintext decrypted in place inside p.Data
// until Flush, so the packet must stay alive across flush()
flush()
p.Release()
}
}