mirror of
https://github.com/Control-D-Inc/ctrld.git
synced 2026-02-03 22:18:39 +00:00
Remove StopOnFirstMatch field that was defined but never used in the actual matching logic. The current implementation always evaluates all rule types and applies a fixed precedence (Domain > MAC > Network), making the StopOnFirstMatch field unnecessary. Changes: - Remove StopOnFirstMatch from MatchingConfig structs - Update DefaultMatchingConfig() function - Update all test cases and references - Simplify configuration to only include Order field This cleanup removes dead code and simplifies the configuration API without changing any functional behavior.
1013 lines
33 KiB
Go
1013 lines
33 KiB
Go
package ctrld
|
|
|
|
import (
|
|
"context"
|
|
crand "crypto/rand"
|
|
"crypto/tls"
|
|
"crypto/x509"
|
|
"encoding/hex"
|
|
"errors"
|
|
"fmt"
|
|
"io"
|
|
"math/rand"
|
|
"net"
|
|
"net/http"
|
|
"net/netip"
|
|
"net/url"
|
|
"os"
|
|
"runtime"
|
|
"sort"
|
|
"strconv"
|
|
"strings"
|
|
"sync"
|
|
"sync/atomic"
|
|
"time"
|
|
|
|
"github.com/ameshkov/dnsstamps"
|
|
"github.com/go-playground/validator/v10"
|
|
"github.com/miekg/dns"
|
|
"github.com/spf13/viper"
|
|
"golang.org/x/net/http2"
|
|
"golang.org/x/sync/singleflight"
|
|
"tailscale.com/logtail/backoff"
|
|
"tailscale.com/net/tsaddr"
|
|
|
|
"github.com/Control-D-Inc/ctrld/internal/dnsrcode"
|
|
ctrldnet "github.com/Control-D-Inc/ctrld/internal/net"
|
|
)
|
|
|
|
// IpStackBoth ...
|
|
const (
|
|
// IpStackBoth indicates that ctrld will use either ipv4 or ipv6 for connecting to upstream,
|
|
// depending on which stack is available when receiving the DNS query.
|
|
IpStackBoth = "both"
|
|
// IpStackV4 indicates that ctrld will use only ipv4 for connecting to upstream.
|
|
IpStackV4 = "v4"
|
|
// IpStackV6 indicates that ctrld will use only ipv6 for connecting to upstream.
|
|
IpStackV6 = "v6"
|
|
// IpStackSplit indicates that ctrld will use either ipv4 or ipv6 for connecting to upstream,
|
|
// depending on the record type of the DNS query.
|
|
IpStackSplit = "split"
|
|
|
|
// FreeDnsDomain is the domain name of free ControlD service.
|
|
FreeDnsDomain = "freedns.controld.com"
|
|
// FreeDNSBoostrapIP is the IP address of freedns.controld.com.
|
|
FreeDNSBoostrapIP = "76.76.2.11"
|
|
// FreeDNSBoostrapIPv6 is the IPv6 address of freedns.controld.com.
|
|
FreeDNSBoostrapIPv6 = "2606:1a40::11"
|
|
// PremiumDnsDomain is the domain name of premium ControlD service.
|
|
PremiumDnsDomain = "dns.controld.com"
|
|
// PremiumDNSBoostrapIP is the IP address of dns.controld.com.
|
|
PremiumDNSBoostrapIP = "76.76.2.22"
|
|
// PremiumDNSBoostrapIPv6 is the IPv6 address of dns.controld.com.
|
|
PremiumDNSBoostrapIPv6 = "2606:1a40::22"
|
|
|
|
// freeDnsDomainDev is the domain name of free ControlD service on dev env.
|
|
freeDnsDomainDev = "freedns.controld.dev"
|
|
// freeDNSBoostrapIP is the IP address of freedns.controld.dev.
|
|
freeDNSBoostrapIP = "176.125.239.11"
|
|
// freeDNSBoostrapIPv6 is the IPv6 address of freedns.controld.com.
|
|
freeDNSBoostrapIPv6 = "2606:1a40:f000::11"
|
|
// premiumDnsDomainDev is the domain name of premium ControlD service on dev env.
|
|
premiumDnsDomainDev = "dns.controld.dev"
|
|
// premiumDNSBoostrapIP is the IP address of dns.controld.dev.
|
|
premiumDNSBoostrapIP = "176.125.239.22"
|
|
// premiumDNSBoostrapIPv6 is the IPv6 address of dns.controld.dev.
|
|
premiumDNSBoostrapIPv6 = "2606:1a40:f000::22"
|
|
|
|
controlDComDomain = "controld.com"
|
|
controlDNetDomain = "controld.net"
|
|
controlDDevDomain = "controld.dev"
|
|
|
|
endpointPrefixHTTPS = "https://"
|
|
endpointPrefixQUIC = "quic://"
|
|
endpointPrefixH3 = "h3://"
|
|
endpointPrefixSdns = "sdns://"
|
|
)
|
|
|
|
var (
|
|
controldParentDomains = []string{controlDComDomain, controlDNetDomain, controlDDevDomain}
|
|
controldVerifiedDomain = map[string]string{
|
|
controlDComDomain: "verify.controld.com",
|
|
controlDDevDomain: "verify.controld.dev",
|
|
}
|
|
)
|
|
|
|
// SetConfigName set the config name that ctrld will look for.
|
|
// DEPRECATED: use SetConfigNameWithPath instead.
|
|
func SetConfigName(v *viper.Viper, name string) {
|
|
configPath := "$HOME"
|
|
// viper has its own way to get user home directory: https://github.com/spf13/viper/blob/v1.14.0/util.go#L134
|
|
// To be consistent, we prefer os.UserHomeDir instead.
|
|
if homeDir, err := os.UserHomeDir(); err == nil {
|
|
configPath = homeDir
|
|
}
|
|
SetConfigNameWithPath(v, name, configPath)
|
|
}
|
|
|
|
// SetConfigNameWithPath set the config path and name that ctrld will look for.
|
|
func SetConfigNameWithPath(v *viper.Viper, name, configPath string) {
|
|
v.SetConfigName(name)
|
|
v.AddConfigPath(configPath)
|
|
v.AddConfigPath(".")
|
|
}
|
|
|
|
// InitConfig initializes default config values for given *viper.Viper instance.
|
|
func InitConfig(v *viper.Viper, name string) {
|
|
ctx := context.Background()
|
|
logger := LoggerFromCtx(ctx)
|
|
Log(ctx, logger.Debug(), "Config initialization started")
|
|
|
|
v.SetDefault("listener", map[string]*ListenerConfig{
|
|
"0": {
|
|
IP: "",
|
|
Port: 0,
|
|
Policy: &ListenerPolicyConfig{
|
|
Name: "Main Policy",
|
|
Networks: []Rule{
|
|
{"network.0": []string{"upstream.0"}},
|
|
},
|
|
Rules: []Rule{
|
|
{"example.com": []string{"upstream.0"}},
|
|
{"*.ads.com": []string{"upstream.1"}},
|
|
},
|
|
},
|
|
},
|
|
})
|
|
v.SetDefault("network", map[string]*NetworkConfig{
|
|
"0": {
|
|
Name: "Network 0",
|
|
Cidrs: []string{"0.0.0.0/0"},
|
|
},
|
|
})
|
|
v.SetDefault("upstream", map[string]*UpstreamConfig{
|
|
"0": {
|
|
BootstrapIP: FreeDNSBoostrapIP,
|
|
Name: "Control D - Anti-Malware",
|
|
Type: ResolverTypeDOH,
|
|
Endpoint: "https://freedns.controld.com/p1",
|
|
Timeout: 5000,
|
|
},
|
|
"1": {
|
|
BootstrapIP: FreeDNSBoostrapIP,
|
|
Name: "Control D - No Ads",
|
|
Type: ResolverTypeDOQ,
|
|
Endpoint: "p2.freedns.controld.com",
|
|
Timeout: 3000,
|
|
},
|
|
})
|
|
|
|
Log(ctx, logger.Debug(), "Config initialization completed")
|
|
}
|
|
|
|
// Config represents ctrld supported configuration.
|
|
type Config struct {
|
|
Service ServiceConfig `mapstructure:"service" toml:"service,omitempty"`
|
|
Listener map[string]*ListenerConfig `mapstructure:"listener" toml:"listener" validate:"min=1,dive"`
|
|
Network map[string]*NetworkConfig `mapstructure:"network" toml:"network" validate:"min=1,dive"`
|
|
Upstream map[string]*UpstreamConfig `mapstructure:"upstream" toml:"upstream" validate:"min=1,dive"`
|
|
}
|
|
|
|
// HasUpstreamSendClientInfo reports whether the config has any upstream
|
|
// is configured to send client info to Control D DNS server.
|
|
func (c *Config) HasUpstreamSendClientInfo() bool {
|
|
for _, uc := range c.Upstream {
|
|
if uc.UpstreamSendClientInfo() {
|
|
return true
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
// FirstListener returns the first listener config of current config. Listeners are sorted numerically.
|
|
//
|
|
// It panics if Config has no listeners configured.
|
|
func (c *Config) FirstListener() *ListenerConfig {
|
|
listeners := make([]int, 0, len(c.Listener))
|
|
for k := range c.Listener {
|
|
n, err := strconv.Atoi(k)
|
|
if err != nil {
|
|
continue
|
|
}
|
|
listeners = append(listeners, n)
|
|
}
|
|
if len(listeners) == 0 {
|
|
panic("missing listener config")
|
|
}
|
|
sort.Ints(listeners)
|
|
return c.Listener[strconv.Itoa(listeners[0])]
|
|
}
|
|
|
|
// FirstUpstream returns the first upstream of current config. Upstreams are sorted numerically.
|
|
//
|
|
// It panics if Config has no upstreams configured.
|
|
func (c *Config) FirstUpstream() *UpstreamConfig {
|
|
upstreams := make([]int, 0, len(c.Upstream))
|
|
for k := range c.Upstream {
|
|
n, err := strconv.Atoi(k)
|
|
if err != nil {
|
|
continue
|
|
}
|
|
upstreams = append(upstreams, n)
|
|
}
|
|
if len(upstreams) == 0 {
|
|
panic("missing listener config")
|
|
}
|
|
sort.Ints(upstreams)
|
|
return c.Upstream[strconv.Itoa(upstreams[0])]
|
|
}
|
|
|
|
// ServiceConfig specifies the general ctrld config.
|
|
type ServiceConfig struct {
|
|
LogLevel string `mapstructure:"log_level" toml:"log_level,omitempty"`
|
|
LogPath string `mapstructure:"log_path" toml:"log_path,omitempty"`
|
|
CacheEnable bool `mapstructure:"cache_enable" toml:"cache_enable,omitempty"`
|
|
CacheSize int `mapstructure:"cache_size" toml:"cache_size,omitempty"`
|
|
CacheTTLOverride int `mapstructure:"cache_ttl_override" toml:"cache_ttl_override,omitempty"`
|
|
CacheServeStale bool `mapstructure:"cache_serve_stale" toml:"cache_serve_stale,omitempty"`
|
|
CacheFlushDomains []string `mapstructure:"cache_flush_domains" toml:"cache_flush_domains" validate:"max=256"`
|
|
MaxConcurrentRequests *int `mapstructure:"max_concurrent_requests" toml:"max_concurrent_requests,omitempty" validate:"omitempty,gte=0"`
|
|
DHCPLeaseFile string `mapstructure:"dhcp_lease_file_path" toml:"dhcp_lease_file_path" validate:"omitempty,file"`
|
|
DHCPLeaseFileFormat string `mapstructure:"dhcp_lease_file_format" toml:"dhcp_lease_file_format" validate:"required_unless=DHCPLeaseFile '',omitempty,oneof=dnsmasq isc-dhcp kea-dhcp4"`
|
|
DiscoverMDNS *bool `mapstructure:"discover_mdns" toml:"discover_mdns,omitempty"`
|
|
DiscoverARP *bool `mapstructure:"discover_arp" toml:"discover_arp,omitempty"`
|
|
DiscoverDHCP *bool `mapstructure:"discover_dhcp" toml:"discover_dhcp,omitempty"`
|
|
DiscoverPtr *bool `mapstructure:"discover_ptr" toml:"discover_ptr,omitempty"`
|
|
DiscoverHosts *bool `mapstructure:"discover_hosts" toml:"discover_hosts,omitempty"`
|
|
DiscoverRefreshInterval int `mapstructure:"discover_refresh_interval" toml:"discover_refresh_interval,omitempty"`
|
|
ClientIDPref string `mapstructure:"client_id_preference" toml:"client_id_preference,omitempty" validate:"omitempty,oneof=host mac"`
|
|
MetricsQueryStats bool `mapstructure:"metrics_query_stats" toml:"metrics_query_stats,omitempty"`
|
|
MetricsListener string `mapstructure:"metrics_listener" toml:"metrics_listener,omitempty"`
|
|
DnsWatchdogEnabled *bool `mapstructure:"dns_watchdog_enabled" toml:"dns_watchdog_enabled,omitempty"`
|
|
DnsWatchdogInvterval *time.Duration `mapstructure:"dns_watchdog_interval" toml:"dns_watchdog_interval,omitempty"`
|
|
RefetchTime *int `mapstructure:"refetch_time" toml:"refetch_time,omitempty"`
|
|
ForceRefetchWaitTime *int `mapstructure:"force_refetch_wait_time" toml:"force_refetch_wait_time,omitempty"`
|
|
LeakOnUpstreamFailure *bool `mapstructure:"leak_on_upstream_failure" toml:"leak_on_upstream_failure,omitempty"`
|
|
Daemon bool `mapstructure:"-" toml:"-"`
|
|
AllocateIP bool `mapstructure:"-" toml:"-"`
|
|
}
|
|
|
|
// NetworkConfig specifies configuration for networks where ctrld will handle requests.
|
|
type NetworkConfig struct {
|
|
Name string `mapstructure:"name" toml:"name,omitempty"`
|
|
Cidrs []string `mapstructure:"cidrs" toml:"cidrs,omitempty" validate:"dive,cidr"`
|
|
IPNets []*net.IPNet `mapstructure:"-" toml:"-"`
|
|
}
|
|
|
|
// UpstreamConfig specifies configuration for upstreams that ctrld will forward requests to.
|
|
type UpstreamConfig struct {
|
|
Name string `mapstructure:"name" toml:"name,omitempty"`
|
|
Type string `mapstructure:"type" toml:"type,omitempty" validate:"oneof=doh doh3 dot doq os legacy sdns ''"`
|
|
Endpoint string `mapstructure:"endpoint" toml:"endpoint,omitempty"`
|
|
BootstrapIP string `mapstructure:"bootstrap_ip" toml:"bootstrap_ip,omitempty"`
|
|
Domain string `mapstructure:"-" toml:"-"`
|
|
IPStack string `mapstructure:"ip_stack" toml:"ip_stack,omitempty" validate:"ipstack"`
|
|
Timeout int `mapstructure:"timeout" toml:"timeout,omitempty" validate:"gte=0"`
|
|
// The caller should not access this field directly.
|
|
// Use UpstreamSendClientInfo instead.
|
|
SendClientInfo *bool `mapstructure:"send_client_info" toml:"send_client_info,omitempty"`
|
|
// The caller should not access this field directly.
|
|
// Use IsDiscoverable instead.
|
|
Discoverable *bool `mapstructure:"discoverable" toml:"discoverable"`
|
|
|
|
g singleflight.Group
|
|
rebootstrap atomic.Bool
|
|
bootstrapIPs []string
|
|
bootstrapIPs4 []string
|
|
bootstrapIPs6 []string
|
|
transport *http.Transport
|
|
transportOnce sync.Once
|
|
transport4 *http.Transport
|
|
transport6 *http.Transport
|
|
http3RoundTripper http.RoundTripper
|
|
http3RoundTripper4 http.RoundTripper
|
|
http3RoundTripper6 http.RoundTripper
|
|
certPool *x509.CertPool
|
|
u *url.URL
|
|
fallbackOnce sync.Once
|
|
uid string
|
|
}
|
|
|
|
// ListenerConfig specifies the networks configuration that ctrld will run on.
|
|
type ListenerConfig struct {
|
|
IP string `mapstructure:"ip" toml:"ip,omitempty" validate:"iporempty"`
|
|
Port int `mapstructure:"port" toml:"port,omitempty" validate:"gte=0"`
|
|
Restricted bool `mapstructure:"restricted" toml:"restricted,omitempty"`
|
|
AllowWanClients bool `mapstructure:"allow_wan_clients" toml:"allow_wan_clients,omitempty"`
|
|
Policy *ListenerPolicyConfig `mapstructure:"policy" toml:"policy,omitempty"`
|
|
}
|
|
|
|
// IsDirectDnsListener reports whether ctrld can be a direct listener on port 53.
|
|
// It returns true only if ctrld can listen on port 53 for all interfaces. That means
|
|
// there's no other software listening on port 53.
|
|
//
|
|
// If someone listening on port 53, or ctrld could only listen on port 53 for a specific
|
|
// interface, ctrld could only be configured as a DNS forwarder.
|
|
func (lc *ListenerConfig) IsDirectDnsListener() bool {
|
|
if lc == nil || lc.Port != 53 {
|
|
return false
|
|
}
|
|
switch lc.IP {
|
|
case "", "::", "0.0.0.0":
|
|
return true
|
|
default:
|
|
return false
|
|
}
|
|
}
|
|
|
|
// MatchingConfig defines the configuration for rule matching behavior
|
|
type MatchingConfig struct {
|
|
Order []string `mapstructure:"order" toml:"order,omitempty" json:"order" yaml:"order"`
|
|
}
|
|
|
|
// ListenerPolicyConfig specifies the policy rules for ctrld to filter incoming requests.
|
|
type ListenerPolicyConfig struct {
|
|
Name string `mapstructure:"name" toml:"name,omitempty"`
|
|
Networks []Rule `mapstructure:"networks" toml:"networks,omitempty,inline,multiline" validate:"dive,len=1"`
|
|
Rules []Rule `mapstructure:"rules" toml:"rules,omitempty,inline,multiline" validate:"dive,len=1"`
|
|
Macs []Rule `mapstructure:"macs" toml:"macs,omitempty,inline,multiline" validate:"dive,len=1"`
|
|
FailoverRcodes []string `mapstructure:"failover_rcodes" toml:"failover_rcodes,omitempty" validate:"dive,dnsrcode"`
|
|
FailoverRcodeNumbers []int `mapstructure:"-" toml:"-"`
|
|
Matching *MatchingConfig `mapstructure:"-" toml:"-"`
|
|
}
|
|
|
|
// Rule is a map from source to list of upstreams.
|
|
// ctrld uses rule to perform requests matching and forward
|
|
// the request to corresponding upstreams if it's matched.
|
|
type Rule map[string][]string
|
|
|
|
// Init initialized necessary values for an UpstreamConfig.
|
|
func (uc *UpstreamConfig) Init(ctx context.Context) {
|
|
logger := LoggerFromCtx(ctx)
|
|
if err := uc.initDnsStamps(); err != nil {
|
|
logger.Fatal().Err(err).Msg("Invalid dns stamps")
|
|
}
|
|
uc.initDoHScheme()
|
|
uc.uid = upstreamUID(ctx)
|
|
if u, err := url.Parse(uc.Endpoint); err == nil {
|
|
uc.Domain = u.Hostname()
|
|
switch uc.Type {
|
|
case ResolverTypeDOH, ResolverTypeDOH3:
|
|
uc.u = u
|
|
}
|
|
}
|
|
if uc.Domain == "" {
|
|
if !strings.Contains(uc.Endpoint, ":") {
|
|
uc.Domain = uc.Endpoint
|
|
uc.Endpoint = net.JoinHostPort(uc.Endpoint, defaultPortFor(uc.Type))
|
|
}
|
|
host, _, _ := net.SplitHostPort(uc.Endpoint)
|
|
uc.Domain = host
|
|
if net.ParseIP(uc.Domain) != nil {
|
|
uc.BootstrapIP = uc.Domain
|
|
}
|
|
}
|
|
if uc.IPStack == "" {
|
|
// Set default IP stack based on upstream type
|
|
// Control-D upstreams use split stack for better IPv4/IPv6 handling,
|
|
// while other upstreams use both stacks for maximum compatibility
|
|
if uc.IsControlD() {
|
|
uc.IPStack = IpStackSplit
|
|
} else {
|
|
uc.IPStack = IpStackBoth
|
|
}
|
|
}
|
|
}
|
|
|
|
// VerifyMsg creates and returns a new DNS message could be used for testing upstream health.
|
|
func (uc *UpstreamConfig) VerifyMsg() *dns.Msg {
|
|
msg := new(dns.Msg)
|
|
msg.RecursionDesired = true
|
|
msg.SetQuestion(".", dns.TypeNS)
|
|
msg.SetEdns0(4096, false) // ensure handling of large DNS response
|
|
return msg
|
|
}
|
|
|
|
// VerifyDomain returns the domain name that could be resolved by the upstream endpoint.
|
|
// It returns empty for non-ControlD upstream endpoint.
|
|
func (uc *UpstreamConfig) VerifyDomain() string {
|
|
domain := uc.Domain
|
|
if domain == "" {
|
|
if u, err := url.Parse(uc.Endpoint); err == nil {
|
|
domain = u.Hostname()
|
|
}
|
|
}
|
|
for _, parent := range controldParentDomains {
|
|
if dns.IsSubDomain(parent, domain) {
|
|
return controldVerifiedDomain[parent]
|
|
}
|
|
}
|
|
return ""
|
|
}
|
|
|
|
// UpstreamSendClientInfo reports whether the upstream is
|
|
// configured to send client info to Control D DNS server.
|
|
//
|
|
// Client info includes:
|
|
// - MAC
|
|
// - Lan IP
|
|
// - Hostname
|
|
func (uc *UpstreamConfig) UpstreamSendClientInfo() bool {
|
|
if uc.SendClientInfo != nil {
|
|
return *uc.SendClientInfo
|
|
}
|
|
switch uc.Type {
|
|
case ResolverTypeDOH, ResolverTypeDOH3:
|
|
if uc.IsControlD() || uc.isNextDNS() {
|
|
return true
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
// IsDiscoverable reports whether the upstream can be used for PTR discovery.
|
|
// The caller must ensure uc.Init() was called before calling this.
|
|
func (uc *UpstreamConfig) IsDiscoverable() bool {
|
|
if uc.Discoverable != nil {
|
|
return *uc.Discoverable
|
|
}
|
|
switch uc.Type {
|
|
case ResolverTypeOS, ResolverTypeLegacy, ResolverTypePrivate:
|
|
if ip, err := netip.ParseAddr(uc.Domain); err == nil {
|
|
return ip.IsLoopback() || ip.IsPrivate() || ip.IsLinkLocalUnicast() || tsaddr.CGNATRange().Contains(ip)
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
// BootstrapIPs returns the bootstrap IPs list of upstreams.
|
|
func (uc *UpstreamConfig) BootstrapIPs() []string {
|
|
return uc.bootstrapIPs
|
|
}
|
|
|
|
// SetCertPool sets the system cert pool used for TLS connections.
|
|
func (uc *UpstreamConfig) SetCertPool(cp *x509.CertPool) {
|
|
uc.certPool = cp
|
|
}
|
|
|
|
// UID returns the unique identifier of the upstream.
|
|
func (uc *UpstreamConfig) UID() string {
|
|
return uc.uid
|
|
}
|
|
|
|
// SetupBootstrapIP sets up bootstrap IPs for the upstream config.
|
|
// The setup process will block until there's usable IPs found.
|
|
func (uc *UpstreamConfig) SetupBootstrapIP(ctx context.Context) {
|
|
logger := LoggerFromCtx(ctx)
|
|
Log(ctx, logger.Debug(), "Setting up bootstrap IPs for upstream: %s", uc.Name)
|
|
|
|
b := backoff.NewBackoff("setupBootstrapIP", func(format string, args ...any) {}, 10*time.Second)
|
|
isControlD := uc.IsControlD()
|
|
nss := initDefaultOsResolver(ctx)
|
|
for {
|
|
Log(ctx, logger.Debug(), "Looking up bootstrap IPs for domain: %s", uc.Domain)
|
|
uc.bootstrapIPs = lookupIP(ctx, uc.Domain, uc.Timeout, nss)
|
|
// For ControlD upstream, the bootstrap IPs could not be RFC 1918 addresses,
|
|
// filtering them out here to prevent weird behavior.
|
|
if isControlD {
|
|
n := 0
|
|
for _, ip := range uc.bootstrapIPs {
|
|
netIP := net.ParseIP(ip)
|
|
if netIP != nil && !netIP.IsPrivate() {
|
|
uc.bootstrapIPs[n] = ip
|
|
n++
|
|
}
|
|
}
|
|
uc.bootstrapIPs = uc.bootstrapIPs[:n]
|
|
if len(uc.bootstrapIPs) == 0 {
|
|
uc.bootstrapIPs = bootstrapIPsFromControlDDomain(uc.Domain)
|
|
logger.Warn().Msgf("No record found for %q, lookup from direct ip table", uc.Domain)
|
|
}
|
|
}
|
|
if len(uc.bootstrapIPs) == 0 {
|
|
logger.Warn().Msgf("No record found for %q, using bootstrap server: %s", uc.Domain, PremiumDNSBoostrapIP)
|
|
uc.bootstrapIPs = lookupIP(ctx, uc.Domain, uc.Timeout, []string{net.JoinHostPort(PremiumDNSBoostrapIP, "53")})
|
|
|
|
}
|
|
if len(uc.bootstrapIPs) > 0 {
|
|
break
|
|
}
|
|
logger.Warn().Msg("Could not resolve bootstrap ips, retrying...")
|
|
b.BackOff(context.Background(), errors.New("no bootstrap IPs"))
|
|
}
|
|
for _, ip := range uc.bootstrapIPs {
|
|
if ctrldnet.IsIPv6(ip) {
|
|
uc.bootstrapIPs6 = append(uc.bootstrapIPs6, ip)
|
|
} else {
|
|
uc.bootstrapIPs4 = append(uc.bootstrapIPs4, ip)
|
|
}
|
|
}
|
|
logger.Debug().Msgf("Bootstrap ips: %v", uc.bootstrapIPs)
|
|
Log(ctx, logger.Debug(), "Bootstrap IP setup completed for upstream: %s", uc.Name)
|
|
}
|
|
|
|
// ReBootstrap re-setup the bootstrap IP and the transport.
|
|
func (uc *UpstreamConfig) ReBootstrap(ctx context.Context) {
|
|
switch uc.Type {
|
|
case ResolverTypeDOH, ResolverTypeDOH3:
|
|
default:
|
|
return
|
|
}
|
|
_, _, _ = uc.g.Do("ReBootstrap", func() (any, error) {
|
|
if uc.rebootstrap.CompareAndSwap(false, true) {
|
|
logger := LoggerFromCtx(ctx)
|
|
Log(ctx, logger.Debug(), "Re-bootstrapping upstream: %s", uc.Name)
|
|
}
|
|
return true, nil
|
|
})
|
|
}
|
|
|
|
// SetupTransport initializes the network transport used to connect to upstream server.
|
|
// For now, only DoH upstream is supported.
|
|
func (uc *UpstreamConfig) SetupTransport(ctx context.Context) {
|
|
switch uc.Type {
|
|
case ResolverTypeDOH:
|
|
uc.setupDOHTransport(ctx)
|
|
case ResolverTypeDOH3:
|
|
uc.setupDOH3Transport(ctx)
|
|
}
|
|
}
|
|
|
|
func (uc *UpstreamConfig) setupDOHTransport(ctx context.Context) {
|
|
switch uc.IPStack {
|
|
case IpStackBoth, "":
|
|
uc.transport = uc.newDOHTransport(ctx, uc.bootstrapIPs)
|
|
case IpStackV4:
|
|
uc.transport = uc.newDOHTransport(ctx, uc.bootstrapIPs4)
|
|
case IpStackV6:
|
|
uc.transport = uc.newDOHTransport(ctx, uc.bootstrapIPs6)
|
|
case IpStackSplit:
|
|
uc.transport4 = uc.newDOHTransport(ctx, uc.bootstrapIPs4)
|
|
if HasIPv6(ctx) {
|
|
uc.transport6 = uc.newDOHTransport(ctx, uc.bootstrapIPs6)
|
|
} else {
|
|
uc.transport6 = uc.transport4
|
|
}
|
|
uc.transport = uc.newDOHTransport(ctx, uc.bootstrapIPs)
|
|
}
|
|
}
|
|
|
|
func (uc *UpstreamConfig) newDOHTransport(ctx context.Context, addrs []string) *http.Transport {
|
|
transport := http.DefaultTransport.(*http.Transport).Clone()
|
|
transport.MaxIdleConnsPerHost = 100
|
|
transport.TLSClientConfig = &tls.Config{
|
|
RootCAs: uc.certPool,
|
|
ClientSessionCache: tls.NewLRUClientSessionCache(0),
|
|
}
|
|
|
|
// Prevent bad tcp connection hanging the requests for too long.
|
|
// See: https://github.com/golang/go/issues/36026
|
|
if t2, err := http2.ConfigureTransports(transport); err == nil {
|
|
t2.ReadIdleTimeout = 10 * time.Second
|
|
t2.PingTimeout = 5 * time.Second
|
|
}
|
|
|
|
dialerTimeoutMs := 2000
|
|
if uc.Timeout > 0 && uc.Timeout < dialerTimeoutMs {
|
|
dialerTimeoutMs = uc.Timeout
|
|
}
|
|
dialerTimeout := time.Duration(dialerTimeoutMs) * time.Millisecond
|
|
logger := LoggerFromCtx(ctx)
|
|
transport.DialContext = func(ctx context.Context, network, addr string) (net.Conn, error) {
|
|
_, port, _ := net.SplitHostPort(addr)
|
|
if uc.BootstrapIP != "" {
|
|
dialer := net.Dialer{Timeout: dialerTimeout, KeepAlive: dialerTimeout}
|
|
addr := net.JoinHostPort(uc.BootstrapIP, port)
|
|
Log(ctx, logger.Debug(), "Sending doh request to: %s", addr)
|
|
return dialer.DialContext(ctx, network, addr)
|
|
}
|
|
pd := &ctrldnet.ParallelDialer{}
|
|
pd.Timeout = dialerTimeout
|
|
pd.KeepAlive = dialerTimeout
|
|
dialAddrs := make([]string, len(addrs))
|
|
for i := range addrs {
|
|
dialAddrs[i] = net.JoinHostPort(addrs[i], port)
|
|
}
|
|
conn, err := pd.DialContext(ctx, network, dialAddrs, logger.Logger)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
Log(ctx, logger.Debug(), "Sending doh request to: %s", conn.RemoteAddr())
|
|
return conn, nil
|
|
}
|
|
runtime.SetFinalizer(transport, func(transport *http.Transport) {
|
|
transport.CloseIdleConnections()
|
|
})
|
|
return transport
|
|
}
|
|
|
|
// Ping warms up the connection to DoH/DoH3 upstream.
|
|
func (uc *UpstreamConfig) Ping(ctx context.Context) {
|
|
if err := uc.ping(ctx); err != nil {
|
|
logger := LoggerFromCtx(ctx)
|
|
logger.Debug().Err(err).Msgf("Upstream ping failed: %s", uc.Endpoint)
|
|
_ = uc.FallbackToDirectIP(ctx)
|
|
}
|
|
}
|
|
|
|
// ErrorPing is like Ping, but return an error if any.
|
|
func (uc *UpstreamConfig) ErrorPing(ctx context.Context) error {
|
|
return uc.ping(ctx)
|
|
}
|
|
|
|
func (uc *UpstreamConfig) ping(ctx context.Context) error {
|
|
switch uc.Type {
|
|
case ResolverTypeDOH, ResolverTypeDOH3:
|
|
default:
|
|
return nil
|
|
}
|
|
|
|
ping := func(t http.RoundTripper) error {
|
|
if t == nil {
|
|
return nil
|
|
}
|
|
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
|
|
defer cancel()
|
|
req, err := http.NewRequestWithContext(ctx, "HEAD", uc.Endpoint, nil)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
resp, err := t.RoundTrip(req)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
defer resp.Body.Close()
|
|
_, _ = io.Copy(io.Discard, resp.Body)
|
|
return nil
|
|
}
|
|
|
|
for _, typ := range []uint16{dns.TypeA, dns.TypeAAAA} {
|
|
switch uc.Type {
|
|
case ResolverTypeDOH:
|
|
if err := ping(uc.dohTransport(ctx, typ)); err != nil {
|
|
return err
|
|
}
|
|
case ResolverTypeDOH3:
|
|
if err := ping(uc.doh3Transport(ctx, typ)); err != nil {
|
|
return err
|
|
}
|
|
}
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// IsControlD reports whether this is a ControlD upstream.
|
|
func (uc *UpstreamConfig) IsControlD() bool {
|
|
domain := uc.Domain
|
|
if domain == "" {
|
|
if u, err := url.Parse(uc.Endpoint); err == nil {
|
|
domain = u.Hostname()
|
|
}
|
|
}
|
|
for _, parent := range controldParentDomains {
|
|
if dns.IsSubDomain(parent, domain) {
|
|
return true
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
func (uc *UpstreamConfig) isNextDNS() bool {
|
|
domain := uc.Domain
|
|
if domain == "" {
|
|
if u, err := url.Parse(uc.Endpoint); err == nil {
|
|
domain = u.Hostname()
|
|
}
|
|
}
|
|
return domain == "dns.nextdns.io"
|
|
}
|
|
|
|
func (uc *UpstreamConfig) dohTransport(ctx context.Context, dnsType uint16) http.RoundTripper {
|
|
uc.transportOnce.Do(func() {
|
|
uc.SetupTransport(ctx)
|
|
})
|
|
if uc.rebootstrap.CompareAndSwap(true, false) {
|
|
uc.SetupTransport(ctx)
|
|
}
|
|
switch uc.IPStack {
|
|
case IpStackBoth, IpStackV4, IpStackV6:
|
|
return uc.transport
|
|
case IpStackSplit:
|
|
switch dnsType {
|
|
case dns.TypeA:
|
|
return uc.transport4
|
|
default:
|
|
return uc.transport6
|
|
}
|
|
}
|
|
return uc.transport
|
|
}
|
|
|
|
func (uc *UpstreamConfig) bootstrapIPForDNSType(ctx context.Context, dnsType uint16) string {
|
|
switch uc.IPStack {
|
|
case IpStackBoth:
|
|
return pick(uc.bootstrapIPs)
|
|
case IpStackV4:
|
|
return pick(uc.bootstrapIPs4)
|
|
case IpStackV6:
|
|
return pick(uc.bootstrapIPs6)
|
|
case IpStackSplit:
|
|
switch dnsType {
|
|
case dns.TypeA:
|
|
return pick(uc.bootstrapIPs4)
|
|
default:
|
|
if HasIPv6(ctx) {
|
|
return pick(uc.bootstrapIPs6)
|
|
}
|
|
return pick(uc.bootstrapIPs4)
|
|
}
|
|
}
|
|
return pick(uc.bootstrapIPs)
|
|
}
|
|
|
|
func (uc *UpstreamConfig) netForDNSType(ctx context.Context, dnsType uint16) (string, string) {
|
|
switch uc.IPStack {
|
|
case IpStackBoth:
|
|
return "tcp-tls", "udp"
|
|
case IpStackV4:
|
|
return "tcp4-tls", "udp4"
|
|
case IpStackV6:
|
|
return "tcp6-tls", "udp6"
|
|
case IpStackSplit:
|
|
switch dnsType {
|
|
case dns.TypeA:
|
|
return "tcp4-tls", "udp4"
|
|
default:
|
|
if HasIPv6(ctx) {
|
|
return "tcp6-tls", "udp6"
|
|
}
|
|
return "tcp4-tls", "udp4"
|
|
}
|
|
}
|
|
return "tcp-tls", "udp"
|
|
}
|
|
|
|
// initDoHScheme initializes the endpoint scheme for DoH/DoH3 upstream if not present.
|
|
func (uc *UpstreamConfig) initDoHScheme() {
|
|
if strings.HasPrefix(uc.Endpoint, endpointPrefixH3) && uc.Type == "" {
|
|
uc.Type = ResolverTypeDOH3
|
|
}
|
|
switch uc.Type {
|
|
case ResolverTypeDOH:
|
|
case ResolverTypeDOH3:
|
|
if after, found := strings.CutPrefix(uc.Endpoint, endpointPrefixH3); found {
|
|
uc.Endpoint = endpointPrefixHTTPS + after
|
|
}
|
|
default:
|
|
return
|
|
}
|
|
if !strings.HasPrefix(uc.Endpoint, endpointPrefixHTTPS) {
|
|
uc.Endpoint = endpointPrefixHTTPS + uc.Endpoint
|
|
}
|
|
}
|
|
|
|
// initDnsStamps initializes upstream config based on encoded DNS Stamps Endpoint.
|
|
func (uc *UpstreamConfig) initDnsStamps() error {
|
|
if strings.HasPrefix(uc.Endpoint, endpointPrefixSdns) && uc.Type == "" {
|
|
uc.Type = ResolverTypeSDNS
|
|
}
|
|
if uc.Type != ResolverTypeSDNS {
|
|
return nil
|
|
}
|
|
sdns, err := dnsstamps.NewServerStampFromString(uc.Endpoint)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
ip, port, _ := net.SplitHostPort(sdns.ServerAddrStr)
|
|
providerName, port2, _ := net.SplitHostPort(sdns.ProviderName)
|
|
if port2 != "" {
|
|
port = port2
|
|
}
|
|
if providerName == "" {
|
|
providerName = sdns.ProviderName
|
|
}
|
|
switch sdns.Proto {
|
|
case dnsstamps.StampProtoTypeDoH:
|
|
uc.Type = ResolverTypeDOH
|
|
host := sdns.ProviderName
|
|
if port != "" && port != defaultPortFor(uc.Type) {
|
|
host = net.JoinHostPort(providerName, port)
|
|
}
|
|
uc.Endpoint = "https://" + host + sdns.Path
|
|
case dnsstamps.StampProtoTypeTLS:
|
|
uc.Type = ResolverTypeDOT
|
|
uc.Endpoint = net.JoinHostPort(providerName, port)
|
|
case dnsstamps.StampProtoTypeDoQ:
|
|
uc.Type = ResolverTypeDOQ
|
|
uc.Endpoint = net.JoinHostPort(providerName, port)
|
|
case dnsstamps.StampProtoTypePlain:
|
|
uc.Type = ResolverTypeLegacy
|
|
uc.Endpoint = sdns.ServerAddrStr
|
|
default:
|
|
return fmt.Errorf("unsupported stamp protocol %q", sdns.Proto)
|
|
}
|
|
uc.BootstrapIP = ip
|
|
return nil
|
|
}
|
|
|
|
// Context returns a new context with timeout set from upstream config.
|
|
func (uc *UpstreamConfig) Context(ctx context.Context) (context.Context, context.CancelFunc) {
|
|
if uc.Timeout > 0 {
|
|
return context.WithTimeout(ctx, time.Millisecond*time.Duration(uc.Timeout))
|
|
}
|
|
return context.WithCancel(ctx)
|
|
}
|
|
|
|
// FallbackToDirectIP changes ControlD upstream endpoint to use direct IP instead of domain.
|
|
func (uc *UpstreamConfig) FallbackToDirectIP(ctx context.Context) bool {
|
|
if !uc.IsControlD() {
|
|
return false
|
|
}
|
|
if uc.u == nil || uc.Domain == "" {
|
|
return false
|
|
}
|
|
|
|
done := false
|
|
uc.fallbackOnce.Do(func() {
|
|
var ip string
|
|
switch {
|
|
case dns.IsSubDomain(PremiumDnsDomain, uc.Domain):
|
|
ip = PremiumDNSBoostrapIP
|
|
case dns.IsSubDomain(FreeDnsDomain, uc.Domain):
|
|
ip = FreeDNSBoostrapIP
|
|
default:
|
|
return
|
|
}
|
|
logger := LoggerFromCtx(ctx)
|
|
Log(ctx, logger.Warn(), "Using direct IP for %q: %s", uc.Endpoint, ip)
|
|
uc.u.Host = ip
|
|
done = true
|
|
})
|
|
return done
|
|
}
|
|
|
|
// Init initialized necessary values for an ListenerConfig.
|
|
func (lc *ListenerConfig) Init() {
|
|
logger := LoggerFromCtx(context.Background())
|
|
Log(context.Background(), logger.Debug(), "Initializing listener config")
|
|
|
|
if lc.Policy != nil {
|
|
lc.Policy.FailoverRcodeNumbers = make([]int, len(lc.Policy.FailoverRcodes))
|
|
for i, rcode := range lc.Policy.FailoverRcodes {
|
|
lc.Policy.FailoverRcodeNumbers[i] = dnsrcode.FromString(rcode)
|
|
}
|
|
Log(context.Background(), logger.Debug(), "Listener policy initialized with %d failover rcodes", len(lc.Policy.FailoverRcodes))
|
|
}
|
|
|
|
Log(context.Background(), logger.Debug(), "Listener config initialization completed")
|
|
}
|
|
|
|
// ValidateConfig validates the given config.
|
|
func ValidateConfig(validate *validator.Validate, cfg *Config) error {
|
|
_ = validate.RegisterValidation("dnsrcode", validateDnsRcode)
|
|
_ = validate.RegisterValidation("ipstack", validateIpStack)
|
|
_ = validate.RegisterValidation("iporempty", validateIpOrEmpty)
|
|
validate.RegisterStructValidation(upstreamConfigStructLevelValidation, UpstreamConfig{})
|
|
return validate.Struct(cfg)
|
|
}
|
|
|
|
func validateDnsRcode(fl validator.FieldLevel) bool {
|
|
return dnsrcode.FromString(fl.Field().String()) != -1
|
|
}
|
|
|
|
func validateIpStack(fl validator.FieldLevel) bool {
|
|
switch fl.Field().String() {
|
|
case IpStackBoth, IpStackV4, IpStackV6, IpStackSplit, "":
|
|
return true
|
|
default:
|
|
return false
|
|
}
|
|
}
|
|
|
|
func validateIpOrEmpty(fl validator.FieldLevel) bool {
|
|
val := fl.Field().String()
|
|
if val == "" {
|
|
return true
|
|
}
|
|
return net.ParseIP(val) != nil
|
|
}
|
|
|
|
func upstreamConfigStructLevelValidation(sl validator.StructLevel) {
|
|
uc := sl.Current().Addr().Interface().(*UpstreamConfig)
|
|
if uc.Type == ResolverTypeOS {
|
|
return
|
|
}
|
|
|
|
// Endpoint is required for non os resolver.
|
|
if uc.Endpoint == "" {
|
|
sl.ReportError(uc.Endpoint, "endpoint", "Endpoint", "required_unless", "")
|
|
return
|
|
}
|
|
|
|
// Empty type is ok only for endpoints starts with "h3://" and "sdns://".
|
|
if uc.Type == "" && !strings.HasPrefix(uc.Endpoint, endpointPrefixH3) && !strings.HasPrefix(uc.Endpoint, endpointPrefixSdns) {
|
|
sl.ReportError(uc.Endpoint, "type", "type", "oneof", "doh doh3 dot doq os legacy sdns")
|
|
return
|
|
}
|
|
|
|
// initDoHScheme/initDnsStamps may change upstreams information,
|
|
// so restoring changed values after validation to keep original one.
|
|
defer func(ep, typ string) {
|
|
uc.Endpoint = ep
|
|
uc.Type = typ
|
|
}(uc.Endpoint, uc.Type)
|
|
|
|
if err := uc.initDnsStamps(); err != nil {
|
|
sl.ReportError(uc.Endpoint, "endpoint", "Endpoint", "http_url", "")
|
|
return
|
|
}
|
|
uc.initDoHScheme()
|
|
// DoH/DoH3 requires endpoint is an HTTP url.
|
|
if uc.Type == ResolverTypeDOH || uc.Type == ResolverTypeDOH3 {
|
|
u, err := url.Parse(uc.Endpoint)
|
|
if err != nil || u.Host == "" {
|
|
sl.ReportError(uc.Endpoint, "endpoint", "Endpoint", "http_url", "")
|
|
return
|
|
}
|
|
}
|
|
}
|
|
|
|
func defaultPortFor(typ string) string {
|
|
switch typ {
|
|
case ResolverTypeDOH, ResolverTypeDOH3:
|
|
return "443"
|
|
case ResolverTypeDOQ, ResolverTypeDOT:
|
|
return "853"
|
|
case ResolverTypeLegacy:
|
|
return "53"
|
|
}
|
|
return "53"
|
|
}
|
|
|
|
// ResolverTypeFromEndpoint tries guessing the resolver type with a given endpoint
|
|
// using following rules:
|
|
//
|
|
// - If endpoint is an IP address -> ResolverTypeLegacy
|
|
// - If endpoint starts with "https://" -> ResolverTypeDOH
|
|
// - If endpoint starts with "quic://" -> ResolverTypeDOQ
|
|
// - If endpoint starts with "h3://" -> ResolverTypeDOH3
|
|
// - If endpoint starts with "sdns://" -> ResolverTypeSDNS
|
|
// - For anything else -> ResolverTypeDOT
|
|
func ResolverTypeFromEndpoint(endpoint string) string {
|
|
switch {
|
|
case strings.HasPrefix(endpoint, endpointPrefixHTTPS):
|
|
return ResolverTypeDOH
|
|
case strings.HasPrefix(endpoint, endpointPrefixQUIC):
|
|
return ResolverTypeDOQ
|
|
case strings.HasPrefix(endpoint, endpointPrefixH3):
|
|
return ResolverTypeDOH3
|
|
case strings.HasPrefix(endpoint, endpointPrefixSdns):
|
|
return ResolverTypeSDNS
|
|
}
|
|
host := endpoint
|
|
if strings.Contains(endpoint, ":") {
|
|
host, _, _ = net.SplitHostPort(host)
|
|
}
|
|
if ip := net.ParseIP(host); ip != nil {
|
|
return ResolverTypeLegacy
|
|
}
|
|
return ResolverTypeDOT
|
|
}
|
|
|
|
func pick(s []string) string {
|
|
return s[rand.Intn(len(s))]
|
|
}
|
|
|
|
// upstreamUID generates an unique identifier for an upstream.
|
|
func upstreamUID(ctx context.Context) string {
|
|
logger := LoggerFromCtx(ctx)
|
|
b := make([]byte, 4)
|
|
for {
|
|
if _, err := crand.Read(b); err != nil {
|
|
logger.Warn().Err(err).Msg("Could not generate uid for upstream, retrying...")
|
|
continue
|
|
}
|
|
return hex.EncodeToString(b)
|
|
}
|
|
}
|
|
|
|
// String returns a string representation of the UpstreamConfig for logging.
|
|
func (uc *UpstreamConfig) String() string {
|
|
if uc == nil {
|
|
return "<nil>"
|
|
}
|
|
return fmt.Sprintf("{name: %q, type: %q, endpoint: %q, bootstrap_ip: %q, domain: %q, ip_stack: %q}",
|
|
uc.Name, uc.Type, uc.Endpoint, uc.BootstrapIP, uc.Domain, uc.IPStack)
|
|
}
|
|
|
|
// bootstrapIPsFromControlDDomain returns bootstrap IPs for ControlD domain.
|
|
func bootstrapIPsFromControlDDomain(domain string) []string {
|
|
switch {
|
|
case dns.IsSubDomain(PremiumDnsDomain, domain):
|
|
return []string{PremiumDNSBoostrapIP, PremiumDNSBoostrapIPv6}
|
|
case dns.IsSubDomain(FreeDnsDomain, domain):
|
|
return []string{FreeDNSBoostrapIP, FreeDNSBoostrapIPv6}
|
|
case dns.IsSubDomain(premiumDnsDomainDev, domain):
|
|
return []string{premiumDNSBoostrapIP, premiumDNSBoostrapIPv6}
|
|
case dns.IsSubDomain(freeDnsDomainDev, domain):
|
|
return []string{freeDNSBoostrapIP, freeDNSBoostrapIPv6}
|
|
}
|
|
return nil
|
|
}
|