dae/control/control_plane.go
Ekko 6039b256de
chore: more detail of Choose DNS path log (#692)
Co-authored-by: mzz <2017@duck.com>
2024-11-25 20:25:01 +08:00

1014 lines
31 KiB
Go

/*
* SPDX-License-Identifier: AGPL-3.0-only
* Copyright (c) 2022-2024, daeuniverse Organization <dae@v2raya.org>
*/
package control
import (
"context"
"errors"
"fmt"
"net"
"net/netip"
"os"
"path/filepath"
"strconv"
"strings"
"sync"
"syscall"
"time"
"github.com/bits-and-blooms/bloom/v3"
"github.com/cilium/ebpf"
"github.com/cilium/ebpf/asm"
"github.com/cilium/ebpf/features"
"github.com/cilium/ebpf/rlimit"
"github.com/daeuniverse/dae/common"
"github.com/daeuniverse/dae/common/assets"
"github.com/daeuniverse/dae/common/consts"
"github.com/daeuniverse/dae/common/netutils"
"github.com/daeuniverse/dae/component/dns"
"github.com/daeuniverse/dae/component/outbound"
"github.com/daeuniverse/dae/component/outbound/dialer"
"github.com/daeuniverse/dae/component/routing"
"github.com/daeuniverse/dae/config"
"github.com/daeuniverse/dae/pkg/config_parser"
internal "github.com/daeuniverse/dae/pkg/ebpf_internal"
"github.com/daeuniverse/outbound/pool"
"github.com/daeuniverse/outbound/protocol/direct"
"github.com/daeuniverse/outbound/transport/grpc"
"github.com/daeuniverse/outbound/transport/meek"
dnsmessage "github.com/miekg/dns"
"github.com/mohae/deepcopy"
"github.com/sirupsen/logrus"
"golang.org/x/sys/unix"
)
type ControlPlane struct {
log *logrus.Logger
core *controlPlaneCore
deferFuncs []func() error
listenIp string
// TODO: add mutex?
outbounds []*outbound.DialerGroup
inConnections sync.Map
dnsController *DnsController
onceNetworkReady sync.Once
dialMode consts.DialMode
routingMatcher *RoutingMatcher
ctx context.Context
cancel context.CancelFunc
ready chan struct{}
muRealDomainSet sync.Mutex
realDomainSet *bloom.BloomFilter
wanInterface []string
lanInterface []string
sniffingTimeout time.Duration
tproxyPortProtect bool
soMarkFromDae uint32
mptcp bool
}
func NewControlPlane(
log *logrus.Logger,
_bpf interface{},
dnsCache map[string]*DnsCache,
tagToNodeList map[string][]string,
groups []config.Group,
routingA *config.Routing,
global *config.Global,
dnsConfig *config.Dns,
externGeoDataDirs []string,
) (*ControlPlane, error) {
// TODO: Some users reported that enabling GSO on the client would affect the performance of watching YouTube, so we disabled it by default.
if _, ok := os.LookupEnv("QUIC_GO_DISABLE_GSO"); !ok {
os.Setenv("QUIC_GO_DISABLE_GSO", "1")
}
var err error
kernelVersion, e := internal.KernelVersion()
if e != nil {
return nil, fmt.Errorf("failed to get kernel version: %w", e)
}
/// Check linux kernel requirements.
// Check version from high to low to reduce the number of user upgrading kernel.
if err := features.HaveProgramHelper(ebpf.SchedCLS, asm.FnLoop); err != nil {
return nil, fmt.Errorf("%w: your kernel version %v does not support bpf_loop (needed by routing); expect >=%v; upgrade your kernel and try again",
err,
kernelVersion.String(),
consts.BpfLoopFeatureVersion.String())
}
if requirement := consts.ChecksumFeatureVersion; kernelVersion.Less(requirement) {
return nil, fmt.Errorf("your kernel version %v does not support checksum related features; expect >=%v; upgrade your kernel and try again",
kernelVersion.String(),
requirement.String())
}
if requirement := consts.BpfTimerFeatureVersion; len(global.WanInterface) > 0 && kernelVersion.Less(requirement) {
return nil, fmt.Errorf("your kernel version %v does not support bind to WAN; expect >=%v; remove wan_interface in config file and try again",
kernelVersion.String(),
requirement.String())
}
if requirement := consts.SkAssignFeatureVersion; len(global.LanInterface) > 0 && kernelVersion.Less(requirement) {
return nil, fmt.Errorf("your kernel version %v does not support bind to LAN; expect >=%v; remove lan_interface in config file and try again",
kernelVersion.String(),
requirement.String())
}
if kernelVersion.Less(consts.BasicFeatureVersion) {
return nil, fmt.Errorf("your kernel version %v does not satisfy basic requirement; expect >=%v",
kernelVersion.String(),
consts.BasicFeatureVersion.String())
}
var deferFuncs []func() error
/// Allow the current process to lock memory for eBPF resources.
if err = rlimit.RemoveMemlock(); err != nil {
return nil, fmt.Errorf("rlimit.RemoveMemlock:%v", err)
}
InitDaeNetns(log)
if err = InitSysctlManager(log); err != nil {
return nil, err
}
if err = GetDaeNetns().Setup(); err != nil {
return nil, fmt.Errorf("failed to setup dae netns: %w", err)
}
pinPath := filepath.Join(consts.BpfPinRoot, consts.AppName)
if err = os.MkdirAll(pinPath, 0755); err != nil && !os.IsExist(err) {
if os.IsNotExist(err) {
log.Warnln("Perhaps you are in a container environment (such as lxc). If so, please use higher virtualization (kvm/qemu).")
}
return nil, err
}
/// Load pre-compiled programs and maps into the kernel.
if _bpf == nil {
log.Infof("Loading eBPF programs and maps into the kernel...")
log.Infof("The loading process takes about 120MB free memory, which will be released after loading. Insufficient memory will cause loading failure.")
}
//var bpf bpfObjects
var ProgramOptions = ebpf.ProgramOptions{
KernelTypes: nil,
LogSize: ebpf.DefaultVerifierLogSize * 10,
}
if log.Level == logrus.PanicLevel {
ProgramOptions.LogLevel = ebpf.LogLevelBranch | ebpf.LogLevelStats
// ProgramOptions.LogLevel = ebpf.LogLevelInstruction | ebpf.LogLevelStats
}
collectionOpts := &ebpf.CollectionOptions{
Maps: ebpf.MapOptions{
PinPath: pinPath,
},
Programs: ProgramOptions,
}
var bpf *bpfObjects
if _bpf != nil {
if _bpf, ok := _bpf.(*bpfObjects); ok {
bpf = _bpf
} else {
return nil, fmt.Errorf("unexpected bpf type: %T", _bpf)
}
} else {
bpf = new(bpfObjects)
if err = fullLoadBpfObjects(log, bpf, &loadBpfOptions{
PinPath: pinPath,
BigEndianTproxyPort: uint32(common.Htons(global.TproxyPort)),
CollectionOptions: collectionOpts,
}); err != nil {
if log.Level == logrus.PanicLevel {
log.Panicln(err)
}
return nil, fmt.Errorf("load eBPF objects: %w", err)
}
}
log.Infof("Loaded eBPF programs and maps")
// outboundId2Name can be modified later.
outboundId2Name := make(map[uint8]string)
core := newControlPlaneCore(
log,
bpf,
outboundId2Name,
&kernelVersion,
_bpf != nil,
)
defer func() {
if err != nil {
// Flip back.
core.Flip()
_ = core.Close()
}
}()
/// Bind to links. Binding should be advance of dialerGroups to avoid un-routable old connection.
// Bind to LAN
if len(global.LanInterface) > 0 {
if global.AutoConfigKernelParameter {
_ = SetIpv4forward("1")
_ = setForwarding("all", consts.IpVersionStr_6, "1")
}
global.LanInterface = common.Deduplicate(global.LanInterface)
for _, ifname := range global.LanInterface {
if err = core.bindLan(ifname, global.AutoConfigKernelParameter); err != nil {
return nil, fmt.Errorf("bindLan: %v: %w", ifname, err)
}
}
}
// Bind to WAN
if len(global.WanInterface) > 0 {
if err = core.setupSkPidMonitor(); err != nil {
log.WithError(err).Warnln("cgroup2 is not enabled; pname routing cannot be used")
}
if global.EnableLocalTcpFastRedirect {
if err = core.setupLocalTcpFastRedirect(); err != nil {
log.WithError(err).Warnln("failed to setup local tcp fast redirect")
}
}
for _, ifname := range global.WanInterface {
if len(global.LanInterface) > 0 {
// FIXME: Code is not elegant here.
// bindLan setting conf.ipv6.all.forwarding=1 suppresses accept_ra=1,
// thus we set it 2 as a workaround.
// See https://sysctl-explorer.net/net/ipv6/accept_ra/ for more information.
if global.AutoConfigKernelParameter {
acceptRa := sysctl.Keyf("net.ipv6.conf.%v.accept_ra", ifname)
val, _ := acceptRa.Get()
if val == "1" {
_ = acceptRa.Set("2", false)
}
}
}
if err = core.bindWan(ifname, global.AutoConfigKernelParameter); err != nil {
return nil, fmt.Errorf("bindWan: %v: %w", ifname, err)
}
}
}
// Bind to dae0 and dae0peer
if err = core.bindDaens(); err != nil {
return nil, fmt.Errorf("bindDaens: %w", err)
}
/// DialerGroups (outbounds).
if global.AllowInsecure {
log.Warnln("AllowInsecure is enabled, but it is not recommended. Please make sure you have to turn it on.")
}
option := dialer.NewGlobalOption(global, log)
// Dial mode.
dialMode, err := consts.ParseDialMode(global.DialMode)
if err != nil {
return nil, err
}
sniffingTimeout := global.SniffingTimeout
if dialMode == consts.DialMode_Ip {
sniffingTimeout = 0
}
disableKernelAliveCallback := dialMode != consts.DialMode_Ip
_direct, directProperty := dialer.NewDirectDialer(option, true)
direct := dialer.NewDialer(_direct, option, dialer.InstanceOption{DisableCheck: true}, directProperty)
_block, blockProperty := dialer.NewBlockDialer(option, func() { /*Dialer Outbound*/ })
block := dialer.NewDialer(_block, option, dialer.InstanceOption{DisableCheck: true}, blockProperty)
outbounds := []*outbound.DialerGroup{
outbound.NewDialerGroup(option, consts.OutboundDirect.String(),
[]*dialer.Dialer{direct}, []*dialer.Annotation{{}},
outbound.DialerSelectionPolicy{
Policy: consts.DialerSelectionPolicy_Fixed,
FixedIndex: 0,
}, core.outboundAliveChangeCallback(0, disableKernelAliveCallback)),
outbound.NewDialerGroup(option, consts.OutboundBlock.String(),
[]*dialer.Dialer{block}, []*dialer.Annotation{{}},
outbound.DialerSelectionPolicy{
Policy: consts.DialerSelectionPolicy_Fixed,
FixedIndex: 0,
}, core.outboundAliveChangeCallback(1, disableKernelAliveCallback)),
}
// Filter out groups.
// FIXME: Ugly code here: reset grpc and meek clients manually.
grpc.CleanGlobalClientConnectionCache()
meek.CleanGlobalRoundTripperCache()
dialerSet := outbound.NewDialerSetFromLinks(option, tagToNodeList)
deferFuncs = append(deferFuncs, dialerSet.Close)
for _, group := range groups {
// Parse policy.
policy, err := outbound.NewDialerSelectionPolicyFromGroupParam(&group)
if err != nil {
return nil, fmt.Errorf("failed to create group %v: %w", group.Name, err)
}
// Filter nodes with user given filters.
dialers, annos, err := dialerSet.FilterAndAnnotate(group.Filter, group.FilterAnnotation)
if err != nil {
return nil, fmt.Errorf(`failed to create group "%v": %w`, group.Name, err)
}
// Convert node links to dialers.
log.Infof(`Group "%v" node list:`, group.Name)
for _, d := range dialers {
log.Infoln("\t" + d.Property().Name)
}
if len(dialers) == 0 {
log.Infoln("\t<Empty>")
}
groupOption, err := ParseGroupOverrideOption(group, *global, log)
finalOption := option
if err == nil && groupOption != nil {
newDialers := make([]*dialer.Dialer, 0)
for _, d := range dialers {
newDialer := d.Clone()
deferFuncs = append(deferFuncs, newDialer.Close)
newDialer.GlobalOption = groupOption
newDialers = append(newDialers, newDialer)
}
log.Infof(`Group "%v"'s check option has been override.`, group.Name)
dialers = newDialers
finalOption = groupOption
}
// Create dialer group and append it to outbounds.
dialerGroup := outbound.NewDialerGroup(finalOption, group.Name, dialers, annos, *policy,
core.outboundAliveChangeCallback(uint8(len(outbounds)), disableKernelAliveCallback))
outbounds = append(outbounds, dialerGroup)
}
/// Routing.
// Generate outboundName2Id from outbounds.
if len(outbounds) > int(consts.OutboundUserDefinedMax) {
return nil, fmt.Errorf("too many outbounds")
}
outboundName2Id := make(map[string]uint8)
for i, o := range outbounds {
if _, exist := outboundName2Id[o.Name]; exist {
return nil, fmt.Errorf("duplicated outbound name: %v", o.Name)
}
outboundName2Id[o.Name] = uint8(i)
outboundId2Name[uint8(i)] = o.Name
}
// Apply rules optimizers.
locationFinder := assets.NewLocationFinder(externGeoDataDirs)
var rules []*config_parser.RoutingRule
if rules, err = routing.ApplyRulesOptimizers(routingA.Rules,
&routing.AliasOptimizer{},
&routing.DatReaderOptimizer{Logger: log, LocationFinder: locationFinder},
&routing.MergeAndSortRulesOptimizer{},
&routing.DeduplicateParamsOptimizer{},
); err != nil {
return nil, fmt.Errorf("ApplyRulesOptimizers error:\n%w", err)
}
routingA.Rules = nil // Release.
if log.IsLevelEnabled(logrus.DebugLevel) {
var debugBuilder strings.Builder
for _, rule := range rules {
debugBuilder.WriteString(rule.String(true, false, false) + "\n")
}
log.Debugf("RoutingA:\n%vfallback: %v\n", debugBuilder.String(), routingA.Fallback)
}
// Parse rules and build.
builder, err := NewRoutingMatcherBuilder(log, rules, outboundName2Id, core.bpf, routingA.Fallback)
if err != nil {
return nil, fmt.Errorf("NewRoutingMatcherBuilder: %w", err)
}
if err = builder.BuildKernspace(log); err != nil {
return nil, fmt.Errorf("RoutingMatcherBuilder.BuildKernspace: %w", err)
}
routingMatcher, err := builder.BuildUserspace()
if err != nil {
return nil, fmt.Errorf("RoutingMatcherBuilder.BuildUserspace: %w", err)
}
// New control plane.
ctx, cancel := context.WithCancel(context.Background())
plane := &ControlPlane{
log: log,
core: core,
deferFuncs: deferFuncs,
listenIp: "0.0.0.0",
outbounds: outbounds,
dnsController: nil,
onceNetworkReady: sync.Once{},
dialMode: dialMode,
routingMatcher: routingMatcher,
ctx: ctx,
cancel: cancel,
ready: make(chan struct{}),
muRealDomainSet: sync.Mutex{},
realDomainSet: bloom.NewWithEstimates(2048, 0.001),
lanInterface: global.LanInterface,
wanInterface: global.WanInterface,
sniffingTimeout: sniffingTimeout,
tproxyPortProtect: global.TproxyPortProtect,
soMarkFromDae: global.SoMarkFromDae,
mptcp: global.Mptcp,
}
defer func() {
if err != nil {
cancel()
}
}()
/// DNS upstream.
dnsUpstream, err := dns.New(dnsConfig, &dns.NewOption{
Logger: log,
LocationFinder: locationFinder,
UpstreamReadyCallback: plane.dnsUpstreamReadyCallback,
UpstreamResolverNetwork: common.MagicNetwork("udp", global.SoMarkFromDae, global.Mptcp),
})
if err != nil {
return nil, err
}
/// Dns controller.
fixedDomainTtl, err := ParseFixedDomainTtl(dnsConfig.FixedDomainTtl)
if err != nil {
return nil, err
}
if plane.dnsController, err = NewDnsController(dnsUpstream, &DnsControllerOption{
Log: log,
CacheAccessCallback: func(cache *DnsCache) (err error) {
// Write mappings into eBPF map:
// IP record (from dns lookup) -> domain routing
if err = core.BatchUpdateDomainRouting(cache); err != nil {
return fmt.Errorf("BatchUpdateDomainRouting: %w", err)
}
return nil
},
CacheRemoveCallback: func(cache *DnsCache) (err error) {
// Write mappings into eBPF map:
// IP record (from dns lookup) -> domain routing
if err = core.BatchRemoveDomainRouting(cache); err != nil {
return fmt.Errorf("BatchUpdateDomainRouting: %w", err)
}
return nil
},
NewCache: func(fqdn string, answers []dnsmessage.RR, deadline time.Time, originalDeadline time.Time) (cache *DnsCache, err error) {
return &DnsCache{
DomainBitmap: plane.routingMatcher.domainMatcher.MatchDomainBitmap(fqdn),
Answer: answers,
Deadline: deadline,
OriginalDeadline: originalDeadline,
}, nil
},
BestDialerChooser: plane.chooseBestDnsDialer,
TimeoutExceedCallback: func(dialArgument *dialArgument, err error) {
dialArgument.bestDialer.ReportUnavailable(&dialer.NetworkType{
L4Proto: dialArgument.l4proto,
IpVersion: dialArgument.ipversion,
IsDns: true,
}, err)
},
IpVersionPrefer: dnsConfig.IpVersionPrefer,
FixedDomainTtl: fixedDomainTtl,
}); err != nil {
return nil, err
}
// Refresh domain routing cache with new routing.
// FIXME: We temperarily disable it because we want to make change of DNS section take effects immediately.
// TODO: Add change detection.
if false && len(dnsCache) > 0 {
for cacheKey, cache := range dnsCache {
// Also refresh out-dated routing because kernel map items have no expiration.
lastDot := strings.LastIndex(cacheKey, ".")
if lastDot == -1 || lastDot == len(cacheKey)-1 {
// Not a valid key.
log.Warnln("Invalid cache key:", cacheKey)
continue
}
host := cacheKey[:lastDot]
_typ := cacheKey[lastDot+1:]
typ, err := strconv.ParseUint(_typ, 10, 16)
if err != nil {
// Unexpected.
return nil, err
}
_ = plane.dnsController.UpdateDnsCacheDeadline(host, uint16(typ), cache.Answer, cache.Deadline)
}
} else if _bpf != nil {
// Is reloading, and dnsCache == nil.
// Remove all map items.
// Normally, it is due to the change of ip version preference.
var key [4]uint32
var val bpfDomainRouting
iter := core.bpf.DomainRoutingMap.Iterate()
for iter.Next(&key, &val) {
_ = core.bpf.DomainRoutingMap.Delete(&key)
}
}
// Init immediately to avoid DNS leaking in the very beginning because param control_plane_dns_routing will
// be set in callback.
if err = dnsUpstream.CheckUpstreamsFormat(); err != nil {
return nil, err
}
go dnsUpstream.InitUpstreams()
close(plane.ready)
return plane, nil
}
func ParseFixedDomainTtl(ks []config.KeyableString) (map[string]int, error) {
m := make(map[string]int)
for _, k := range ks {
key, value, _ := strings.Cut(string(k), ":")
ttl, err := strconv.ParseInt(strings.TrimSpace(value), 0, strconv.IntSize)
if err != nil {
return nil, fmt.Errorf("failed to parse ttl: %v", err)
}
m[strings.TrimSpace(key)] = int(ttl)
}
return m, nil
}
func ParseGroupOverrideOption(group config.Group, global config.Global, log *logrus.Logger) (*dialer.GlobalOption, error) {
result := global
changed := false
if group.TcpCheckUrl != nil {
result.TcpCheckUrl = group.TcpCheckUrl
changed = true
}
if group.TcpCheckHttpMethod != "" {
result.TcpCheckHttpMethod = group.TcpCheckHttpMethod
changed = true
}
if group.UdpCheckDns != nil {
result.UdpCheckDns = group.UdpCheckDns
changed = true
}
if group.CheckInterval != 0 {
result.CheckInterval = group.CheckInterval
changed = true
}
if group.CheckTolerance != 0 {
result.CheckTolerance = group.CheckTolerance
changed = true
}
if changed {
option := dialer.NewGlobalOption(&result, log)
return option, nil
}
return nil, nil
}
// EjectBpf will resect bpf from destroying life-cycle of control plane.
func (c *ControlPlane) EjectBpf() *bpfObjects {
return c.core.EjectBpf()
}
func (c *ControlPlane) InjectBpf(bpf *bpfObjects) {
c.core.InjectBpf(bpf)
}
func (c *ControlPlane) CloneDnsCache() map[string]*DnsCache {
c.dnsController.dnsCacheMu.Lock()
defer c.dnsController.dnsCacheMu.Unlock()
return deepcopy.Copy(c.dnsController.dnsCache).(map[string]*DnsCache)
}
func (c *ControlPlane) dnsUpstreamReadyCallback(dnsUpstream *dns.Upstream) (err error) {
// Waiting for ready.
select {
case <-c.ctx.Done():
return nil
case <-c.ready:
}
/// Notify dialers to check.
c.onceNetworkReady.Do(func() {
for _, out := range c.outbounds {
for _, d := range out.Dialers {
d.NotifyCheck()
}
}
})
if dnsUpstream == nil {
return nil
}
/// Updates dns cache to support domain routing for hostname of dns_upstream.
// Ten years later.
deadline := time.Now().Add(time.Hour * 24 * 365 * 10)
fqdn := dnsmessage.CanonicalName(dnsUpstream.Hostname)
if dnsUpstream.Ip4.IsValid() {
typ := dnsmessage.TypeA
answers := []dnsmessage.RR{&dnsmessage.A{
Hdr: dnsmessage.RR_Header{
Name: dnsmessage.CanonicalName(fqdn),
Rrtype: typ,
Class: dnsmessage.ClassINET,
Ttl: 0, // Must be zero.
},
A: dnsUpstream.Ip4.AsSlice(),
}}
if err = c.dnsController.UpdateDnsCacheDeadline(dnsUpstream.Hostname, typ, answers, deadline); err != nil {
return err
}
}
if dnsUpstream.Ip6.IsValid() {
typ := dnsmessage.TypeAAAA
answers := []dnsmessage.RR{&dnsmessage.AAAA{
Hdr: dnsmessage.RR_Header{
Name: dnsmessage.CanonicalName(fqdn),
Rrtype: typ,
Class: dnsmessage.ClassINET,
Ttl: 0, // Must be zero.
},
AAAA: dnsUpstream.Ip6.AsSlice(),
}}
if err = c.dnsController.UpdateDnsCacheDeadline(dnsUpstream.Hostname, typ, answers, deadline); err != nil {
return err
}
}
return nil
}
func (c *ControlPlane) ActivateCheck() {
for _, g := range c.outbounds {
for _, d := range g.Dialers {
// We only activate check of nodes that have a group.
d.ActivateCheck()
}
}
}
func (c *ControlPlane) ChooseDialTarget(outbound consts.OutboundIndex, dst netip.AddrPort, domain string) (dialTarget string, shouldReroute bool, dialIp bool) {
dialMode := consts.DialMode_Ip
if !outbound.IsReserved() && domain != "" {
switch c.dialMode {
case consts.DialMode_Domain:
if cache := c.dnsController.LookupDnsRespCache(c.dnsController.cacheKey(domain, common.AddrToDnsType(dst.Addr())), true); cache != nil {
// Has A/AAAA records. It is a real domain.
dialMode = consts.DialMode_Domain
} else {
// Check if the domain is in real-domain set (bloom filter).
c.muRealDomainSet.Lock()
if c.realDomainSet.TestString(domain) {
c.muRealDomainSet.Unlock()
dialMode = consts.DialMode_Domain
// Should use this domain to reroute
shouldReroute = true
} else {
c.muRealDomainSet.Unlock()
// Lookup A/AAAA to make sure it is a real domain.
ctx, cancel := context.WithTimeout(context.TODO(), 5*time.Second)
defer cancel()
// TODO: use DNS controller and re-route by control plane.
systemDns, err := netutils.SystemDns()
if err == nil {
if ip46, err := netutils.ResolveIp46(ctx, direct.SymmetricDirect, systemDns, domain, common.MagicNetwork("udp", c.soMarkFromDae, c.mptcp), true); err == nil && (ip46.Ip4.IsValid() || ip46.Ip6.IsValid()) {
// Has A/AAAA records. It is a real domain.
dialMode = consts.DialMode_Domain
// Add it to real-domain set.
c.muRealDomainSet.Lock()
c.realDomainSet.AddString(domain)
c.muRealDomainSet.Unlock()
// Should use this domain to reroute
shouldReroute = true
}
}
}
}
case consts.DialMode_DomainCao:
shouldReroute = true
fallthrough
case consts.DialMode_DomainPlus:
dialMode = consts.DialMode_Domain
}
}
switch dialMode {
case consts.DialMode_Ip:
dialTarget = dst.String()
dialIp = true
case consts.DialMode_Domain:
if strings.HasPrefix(domain, "[") && strings.HasSuffix(domain, "]") {
// Sniffed domain may be like `[2606:4700:20::681a:d1f]`. We should remove the brackets.
domain = domain[1 : len(domain)-1]
}
if _, err := netip.ParseAddr(domain); err == nil {
// domain is IPv4 or IPv6 (has colon)
dialTarget = net.JoinHostPort(domain, strconv.Itoa(int(dst.Port())))
dialIp = true
} else if _, _, err := net.SplitHostPort(domain); err == nil {
// domain is already domain:port
dialTarget = domain
} else {
dialTarget = net.JoinHostPort(domain, strconv.Itoa(int(dst.Port())))
}
c.log.WithFields(logrus.Fields{
"from": dst.String(),
"to": dialTarget,
}).Debugln("Rewrite dial target to domain")
}
return dialTarget, shouldReroute, dialIp
}
type Listener struct {
tcpListener net.Listener
packetConn net.PacketConn
port uint16
}
func (l *Listener) Close() error {
var (
err error
err2 error
)
if err, err2 = l.tcpListener.Close(), l.packetConn.Close(); err2 != nil {
if err == nil {
err = err2
} else {
err = fmt.Errorf("%w: %v", err, err2)
}
}
return err
}
func (c *ControlPlane) Serve(readyChan chan<- bool, listener *Listener) (err error) {
sentReady := false
defer func() {
if !sentReady {
readyChan <- false
}
}()
udpConn := listener.packetConn.(*net.UDPConn)
/// Serve.
// TCP socket.
tcpFile, err := listener.tcpListener.(*net.TCPListener).File()
if err != nil {
return fmt.Errorf("failed to retrieve copy of the underlying TCP connection file")
}
c.deferFuncs = append(c.deferFuncs, func() error {
return tcpFile.Close()
})
if err := c.core.bpf.ListenSocketMap.Update(consts.ZeroKey, uint64(tcpFile.Fd()), ebpf.UpdateAny); err != nil {
return err
}
// UDP socket.
udpFile, err := udpConn.File()
if err != nil {
return fmt.Errorf("failed to retrieve copy of the underlying UDP connection file")
}
c.deferFuncs = append(c.deferFuncs, func() error {
return udpFile.Close()
})
if err := c.core.bpf.ListenSocketMap.Update(consts.OneKey, uint64(udpFile.Fd()), ebpf.UpdateAny); err != nil {
return err
}
sentReady = true
readyChan <- true
go func() {
for {
select {
case <-c.ctx.Done():
return
default:
}
lconn, err := listener.tcpListener.Accept()
if err != nil {
if !strings.Contains(err.Error(), "use of closed network connection") {
c.log.Errorf("Error when accept: %v", err)
}
break
}
go func(lconn net.Conn) {
c.inConnections.Store(lconn, struct{}{})
defer c.inConnections.Delete(lconn)
if err := c.handleConn(lconn); err != nil {
c.log.Warnln("handleConn:", err)
}
}(lconn)
}
}()
go func() {
buf := pool.GetFullCap(consts.EthernetMtu)
var oob [120]byte // Size for original dest
defer buf.Put()
for {
select {
case <-c.ctx.Done():
return
default:
}
n, oobn, _, src, err := udpConn.ReadMsgUDPAddrPort(buf, oob[:])
if err != nil {
if !strings.Contains(err.Error(), "use of closed network connection") {
c.log.Errorf("ReadFromUDPAddrPort: %v, %v", src.String(), err)
}
break
}
newBuf := pool.Get(n)
copy(newBuf, buf[:n])
newOob := pool.Get(oobn)
copy(newOob, oob[:oobn])
newSrc := src
convergeSrc := common.ConvergeAddrPort(src)
// Debug:
// t := time.Now()
DefaultUdpTaskPool.EmitTask(convergeSrc.String(), func() {
data := newBuf
oob := newOob
src := newSrc
defer data.Put()
defer oob.Put()
var realDst netip.AddrPort
var routingResult *bpfRoutingResult
pktDst := RetrieveOriginalDest(oob)
routingResult, err := c.core.RetrieveRoutingResult(src, pktDst, unix.IPPROTO_UDP)
if err != nil {
c.log.Warnf("No AddrPort presented: %v", err)
return
} else {
realDst = pktDst
}
if e := c.handlePkt(udpConn, data, convergeSrc, common.ConvergeAddrPort(pktDst), common.ConvergeAddrPort(realDst), routingResult, false); e != nil {
c.log.Warnln("handlePkt:", e)
}
})
// if d := time.Since(t); d > 100*time.Millisecond {
// logrus.Println(d)
// }
}
}()
c.ActivateCheck()
<-c.ctx.Done()
return nil
}
func (c *ControlPlane) ListenAndServe(readyChan chan<- bool, port uint16) (listener *Listener, err error) {
// Listen.
var listenConfig = net.ListenConfig{
Control: func(network, address string, c syscall.RawConn) error {
return dialer.TproxyControl(c)
},
}
listenAddr := net.JoinHostPort(c.listenIp, strconv.Itoa(int(port)))
tcpListener, err := listenConfig.Listen(context.TODO(), "tcp", listenAddr)
if err != nil {
return nil, fmt.Errorf("listenTCP: %w", err)
}
packetConn, err := listenConfig.ListenPacket(context.TODO(), "udp", listenAddr)
if err != nil {
_ = tcpListener.Close()
return nil, fmt.Errorf("listenUDP: %w", err)
}
listener = &Listener{
tcpListener: tcpListener,
packetConn: packetConn,
port: port,
}
defer func() {
if err != nil {
_ = listener.Close()
}
}()
// Serve
if err = c.Serve(readyChan, listener); err != nil {
return nil, fmt.Errorf("failed to serve: %w", err)
}
return listener, nil
}
func (c *ControlPlane) chooseBestDnsDialer(
req *udpRequest,
dnsUpstream *dns.Upstream,
) (*dialArgument, error) {
/// Choose the best l4proto+ipversion dialer, and change taregt DNS to the best ipversion DNS upstream for DNS request.
// Get available ipversions and l4protos for DNS upstream.
ipversions, l4protos := dnsUpstream.SupportedNetworks()
var (
bestLatency time.Duration
l4proto consts.L4ProtoStr
ipversion consts.IpVersionStr
bestDialer *dialer.Dialer
bestOutbound *outbound.DialerGroup
bestTarget netip.AddrPort
dialMark uint32
)
// Get the min latency path.
networkType := dialer.NetworkType{
IsDns: true,
}
for _, ver := range ipversions {
for _, proto := range l4protos {
networkType.L4Proto = proto
networkType.IpVersion = ver
var dAddr netip.Addr
switch ver {
case consts.IpVersionStr_4:
dAddr = dnsUpstream.Ip4
case consts.IpVersionStr_6:
dAddr = dnsUpstream.Ip6
default:
return nil, fmt.Errorf("unexpected ipversion: %v", ver)
}
outboundIndex, mark, _, err := c.Route(req.realSrc, netip.AddrPortFrom(dAddr, dnsUpstream.Port), dnsUpstream.Hostname, proto.ToL4ProtoType(), req.routingResult)
if err != nil {
return nil, err
}
if mark == 0 {
mark = c.soMarkFromDae
}
if int(outboundIndex) >= len(c.outbounds) {
return nil, fmt.Errorf("bad outbound index: %v", outboundIndex)
}
dialerGroup := c.outbounds[outboundIndex]
// DNS always dial IP.
d, latency, err := dialerGroup.Select(&networkType, true)
if err != nil {
continue
}
//if c.log.IsLevelEnabled(logrus.TraceLevel) {
// c.log.WithFields(logrus.Fields{
// "name": d.Name(),
// "latency": latency,
// "network": networkType.String(),
// "outbound": dialerGroup.Name,
// }).Traceln("Choice")
//}
if bestDialer == nil || latency < bestLatency {
bestDialer = d
bestOutbound = dialerGroup
bestLatency = latency
l4proto = proto
ipversion = ver
dialMark = mark
if bestLatency == 0 {
break
}
}
}
}
if bestDialer == nil {
return nil, fmt.Errorf("no proper dialer for DNS upstream: %v", dnsUpstream.String())
}
switch ipversion {
case consts.IpVersionStr_4:
bestTarget = netip.AddrPortFrom(dnsUpstream.Ip4, dnsUpstream.Port)
case consts.IpVersionStr_6:
bestTarget = netip.AddrPortFrom(dnsUpstream.Ip6, dnsUpstream.Port)
}
if c.log.IsLevelEnabled(logrus.TraceLevel) {
c.log.WithFields(logrus.Fields{
"ipversions": ipversions,
"l4protos": l4protos,
"upstream": dnsUpstream.String(),
"choose": string(l4proto) + "+" + string(ipversion),
"use": bestTarget.String(),
"outbound": bestOutbound.Name,
"dialer": bestDialer.Property().Name,
}).Traceln("Choose DNS path")
}
return &dialArgument{
l4proto: l4proto,
ipversion: ipversion,
bestDialer: bestDialer,
bestOutbound: bestOutbound,
bestTarget: bestTarget,
mark: dialMark,
mptcp: c.mptcp,
}, nil
}
func (c *ControlPlane) AbortConnections() (err error) {
var errs []error
c.inConnections.Range(func(key, value any) bool {
if err = key.(net.Conn).Close(); err != nil {
errs = append(errs, err)
}
return true
})
return errors.Join(errs...)
}
func (c *ControlPlane) Close() (err error) {
// Invoke defer funcs in reverse order.
for i := len(c.deferFuncs) - 1; i >= 0; i-- {
if e := c.deferFuncs[i](); e != nil {
// Combine errors.
if err != nil {
err = fmt.Errorf("%w; %v", err, e)
} else {
err = e
}
}
}
c.cancel()
return c.core.Close()
}