代码拉取完成,页面将自动刷新
// Package dns provides support for parsing DNS messages and reporting the
// results. This package supports the DNS protocol as defined by RFC 1034
// and RFC 1035. It does not have any special support for RFC 2671 (EDNS) or
// RFC 4035 (DNS Security Extensions), but since those specifications only
// add backwards compatible features there will be no issues handling the
// messages.
package dns
import (
"bytes"
"expvar"
"fmt"
"net"
"sort"
"strconv"
"strings"
"time"
"github.com/elastic/beats/libbeat/common"
"github.com/elastic/beats/libbeat/logp"
"github.com/elastic/beats/packetbeat/protos"
"github.com/elastic/beats/packetbeat/publish"
mkdns "github.com/miekg/dns"
"golang.org/x/net/publicsuffix"
)
type dnsPlugin struct {
// Configuration data.
ports []int
sendRequest bool
sendResponse bool
includeAuthorities bool
includeAdditionals bool
// Cache of active DNS transactions. The map key is the HashableDnsTuple
// associated with the request.
transactions *common.Cache
transactionTimeout time.Duration
results publish.Transactions // Channel where results are pushed.
}
var (
debugf = logp.MakeDebug("dns")
)
const maxDNSTupleRawSize = 16 + 16 + 2 + 2 + 4 + 1
// Constants used to associate the DNS QR flag with a meaningful value.
const (
query = false
response = true
)
// Transport protocol.
type transport uint8
var (
unmatchedRequests = expvar.NewInt("dns.unmatched_requests")
unmatchedResponses = expvar.NewInt("dns.unmatched_responses")
)
const (
transportTCP = iota
transportUDP
)
var transportNames = []string{
"tcp",
"udp",
}
func (t transport) String() string {
if int(t) >= len(transportNames) {
return "impossible"
}
return transportNames[t]
}
type hashableDNSTuple [maxDNSTupleRawSize]byte
// DnsMessage contains a single DNS message.
type dnsMessage struct {
ts time.Time // Time when the message was received.
tuple common.IPPortTuple // Source and destination addresses of packet.
cmdlineTuple *common.CmdlineTuple
data *mkdns.Msg // Parsed DNS packet data.
length int // Length of the DNS message in bytes (without DecodeOffset).
}
// DnsTuple contains source IP/port, destination IP/port, transport protocol,
// and DNS ID.
type dnsTuple struct {
ipLength int
srcIP, dstIP net.IP
srcPort, dstPort uint16
transport transport
id uint16
raw hashableDNSTuple // Src_ip:Src_port:Dst_ip:Dst_port:Transport:Id
revRaw hashableDNSTuple // Dst_ip:Dst_port:Src_ip:Src_port:Transport:Id
}
func dnsTupleFromIPPort(t *common.IPPortTuple, trans transport, id uint16) dnsTuple {
tuple := dnsTuple{
ipLength: t.IPLength,
srcIP: t.SrcIP,
dstIP: t.DstIP,
srcPort: t.SrcPort,
dstPort: t.DstPort,
transport: trans,
id: id,
}
tuple.computeHashebles()
return tuple
}
func (t dnsTuple) reverse() dnsTuple {
return dnsTuple{
ipLength: t.ipLength,
srcIP: t.dstIP,
dstIP: t.srcIP,
srcPort: t.dstPort,
dstPort: t.srcPort,
transport: t.transport,
id: t.id,
raw: t.revRaw,
revRaw: t.raw,
}
}
func (t *dnsTuple) computeHashebles() {
copy(t.raw[0:16], t.srcIP)
copy(t.raw[16:18], []byte{byte(t.srcPort >> 8), byte(t.srcPort)})
copy(t.raw[18:34], t.dstIP)
copy(t.raw[34:36], []byte{byte(t.dstPort >> 8), byte(t.dstPort)})
copy(t.raw[36:38], []byte{byte(t.id >> 8), byte(t.id)})
t.raw[39] = byte(t.transport)
copy(t.revRaw[0:16], t.dstIP)
copy(t.revRaw[16:18], []byte{byte(t.dstPort >> 8), byte(t.dstPort)})
copy(t.revRaw[18:34], t.srcIP)
copy(t.revRaw[34:36], []byte{byte(t.srcPort >> 8), byte(t.srcPort)})
copy(t.revRaw[36:38], []byte{byte(t.id >> 8), byte(t.id)})
t.revRaw[39] = byte(t.transport)
}
func (t *dnsTuple) String() string {
return fmt.Sprintf("DnsTuple src[%s:%d] dst[%s:%d] transport[%s] id[%d]",
t.srcIP.String(),
t.srcPort,
t.dstIP.String(),
t.dstPort,
t.transport,
t.id)
}
// Hashable returns a hashable value that uniquely identifies
// the DNS tuple.
func (t *dnsTuple) hashable() hashableDNSTuple {
return t.raw
}
// Hashable returns a hashable value that uniquely identifies
// the DNS tuple after swapping the source and destination.
func (t *dnsTuple) revHashable() hashableDNSTuple {
return t.revRaw
}
// getTransaction returns the transaction associated with the given
// HashableDnsTuple. The lookup key should be the HashableDnsTuple associated
// with the request (src is the requestor). Nil is returned if the entry
// does not exist.
func (dns *dnsPlugin) getTransaction(k hashableDNSTuple) *dnsTransaction {
v := dns.transactions.Get(k)
if v != nil {
return v.(*dnsTransaction)
}
return nil
}
type dnsTransaction struct {
ts time.Time // Time when the request was received.
tuple dnsTuple // Key used to track this transaction in the transactionsMap.
responseTime int32 // Elapsed time in milliseconds between the request and response.
src common.Endpoint
dst common.Endpoint
transport transport
notes []string
request *dnsMessage
response *dnsMessage
}
func init() {
protos.Register("dns", New)
}
func New(
testMode bool,
results publish.Transactions,
cfg *common.Config,
) (protos.Plugin, error) {
p := &dnsPlugin{}
config := defaultConfig
if !testMode {
if err := cfg.Unpack(&config); err != nil {
return nil, err
}
}
if err := p.init(results, &config); err != nil {
return nil, err
}
return p, nil
}
func (dns *dnsPlugin) init(results publish.Transactions, config *dnsConfig) error {
dns.setFromConfig(config)
dns.transactions = common.NewCacheWithRemovalListener(
dns.transactionTimeout,
protos.DefaultTransactionHashSize,
func(k common.Key, v common.Value) {
trans, ok := v.(*dnsTransaction)
if !ok {
logp.Err("Expired value is not a *DnsTransaction.")
return
}
dns.expireTransaction(trans)
})
dns.transactions.StartJanitor(dns.transactionTimeout)
dns.results = results
return nil
}
func (dns *dnsPlugin) setFromConfig(config *dnsConfig) error {
dns.ports = config.Ports
dns.sendRequest = config.SendRequest
dns.sendResponse = config.SendResponse
dns.includeAuthorities = config.IncludeAuthorities
dns.includeAdditionals = config.IncludeAdditionals
dns.transactionTimeout = config.TransactionTimeout
return nil
}
func newTransaction(ts time.Time, tuple dnsTuple, cmd common.CmdlineTuple) *dnsTransaction {
trans := &dnsTransaction{
transport: tuple.transport,
ts: ts,
tuple: tuple,
}
trans.src = common.Endpoint{
IP: tuple.srcIP.String(),
Port: tuple.srcPort,
Proc: string(cmd.Src),
}
trans.dst = common.Endpoint{
IP: tuple.dstIP.String(),
Port: tuple.dstPort,
Proc: string(cmd.Dst),
}
return trans
}
// deleteTransaction deletes an entry from the transaction map and returns
// the deleted element. If the key does not exist then nil is returned.
func (dns *dnsPlugin) deleteTransaction(k hashableDNSTuple) *dnsTransaction {
v := dns.transactions.Delete(k)
if v != nil {
return v.(*dnsTransaction)
}
return nil
}
func (dns *dnsPlugin) GetPorts() []int {
return dns.ports
}
func (dns *dnsPlugin) ConnectionTimeout() time.Duration {
return dns.transactionTimeout
}
func (dns *dnsPlugin) receivedDNSRequest(tuple *dnsTuple, msg *dnsMessage) {
debugf("Processing query. %s", tuple.String())
trans := dns.deleteTransaction(tuple.hashable())
if trans != nil {
// This happens if a client puts multiple requests in flight
// with the same ID.
trans.notes = append(trans.notes, duplicateQueryMsg.Error())
debugf("%s %s", duplicateQueryMsg.Error(), tuple.String())
dns.publishTransaction(trans)
dns.deleteTransaction(trans.tuple.hashable())
}
trans = newTransaction(msg.ts, *tuple, *msg.cmdlineTuple)
if tuple.transport == transportUDP && (msg.data.IsEdns0() != nil) && msg.length > maxDNSPacketSize {
trans.notes = append(trans.notes, udpPacketTooLarge.Error())
debugf("%s", udpPacketTooLarge.Error())
}
dns.transactions.Put(tuple.hashable(), trans)
trans.request = msg
}
func (dns *dnsPlugin) receivedDNSResponse(tuple *dnsTuple, msg *dnsMessage) {
debugf("Processing response. %s", tuple.String())
trans := dns.getTransaction(tuple.revHashable())
if trans == nil {
trans = newTransaction(msg.ts, tuple.reverse(), common.CmdlineTuple{
Src: msg.cmdlineTuple.Dst, Dst: msg.cmdlineTuple.Src})
trans.notes = append(trans.notes, orphanedResponse.Error())
debugf("%s %s", orphanedResponse.Error(), tuple.String())
unmatchedResponses.Add(1)
}
trans.response = msg
if tuple.transport == transportUDP {
respIsEdns := msg.data.IsEdns0() != nil
if !respIsEdns && msg.length > maxDNSPacketSize {
trans.notes = append(trans.notes, udpPacketTooLarge.responseError())
debugf("%s", udpPacketTooLarge.responseError())
}
request := trans.request
if request != nil {
reqIsEdns := request.data.IsEdns0() != nil
switch {
case reqIsEdns && !respIsEdns:
trans.notes = append(trans.notes, respEdnsNoSupport.Error())
debugf("%s %s", respEdnsNoSupport.Error(), tuple.String())
case !reqIsEdns && respIsEdns:
trans.notes = append(trans.notes, respEdnsUnexpected.Error())
debugf("%s %s", respEdnsUnexpected.Error(), tuple.String())
}
}
}
dns.publishTransaction(trans)
dns.deleteTransaction(trans.tuple.hashable())
}
func (dns *dnsPlugin) publishTransaction(t *dnsTransaction) {
if dns.results == nil {
return
}
debugf("Publishing transaction. %s", t.tuple.String())
event := common.MapStr{}
event["@timestamp"] = common.Time(t.ts)
event["type"] = "dns"
event["transport"] = t.transport.String()
event["src"] = &t.src
event["dst"] = &t.dst
event["status"] = common.ERROR_STATUS
if len(t.notes) == 1 {
event["notes"] = t.notes[0]
} else if len(t.notes) > 1 {
event["notes"] = strings.Join(t.notes, " ")
}
dnsEvent := common.MapStr{}
event["dns"] = dnsEvent
if t.request != nil && t.response != nil {
event["bytes_in"] = t.request.length
event["bytes_out"] = t.response.length
event["responsetime"] = int32(t.response.ts.Sub(t.ts).Nanoseconds() / 1e6)
event["method"] = dnsOpCodeToString(t.request.data.Opcode)
if len(t.request.data.Question) > 0 {
event["query"] = dnsQuestionToString(t.request.data.Question[0])
event["resource"] = t.request.data.Question[0].Name
}
addDNSToMapStr(dnsEvent, t.response.data, dns.includeAuthorities,
dns.includeAdditionals)
if t.response.data.Rcode == 0 {
event["status"] = common.OK_STATUS
}
if dns.sendRequest {
event["request"] = dnsToString(t.request.data)
}
if dns.sendResponse {
event["response"] = dnsToString(t.response.data)
}
} else if t.request != nil {
event["bytes_in"] = t.request.length
event["method"] = dnsOpCodeToString(t.request.data.Opcode)
if len(t.request.data.Question) > 0 {
event["query"] = dnsQuestionToString(t.request.data.Question[0])
event["resource"] = t.request.data.Question[0].Name
}
addDNSToMapStr(dnsEvent, t.request.data, dns.includeAuthorities,
dns.includeAdditionals)
if dns.sendRequest {
event["request"] = dnsToString(t.request.data)
}
} else if t.response != nil {
event["bytes_out"] = t.response.length
event["method"] = dnsOpCodeToString(t.response.data.Opcode)
if len(t.response.data.Question) > 0 {
event["query"] = dnsQuestionToString(t.response.data.Question[0])
event["resource"] = t.response.data.Question[0].Name
}
addDNSToMapStr(dnsEvent, t.response.data, dns.includeAuthorities,
dns.includeAdditionals)
if dns.sendResponse {
event["response"] = dnsToString(t.response.data)
}
}
dns.results.PublishTransaction(event)
}
func (dns *dnsPlugin) expireTransaction(t *dnsTransaction) {
t.notes = append(t.notes, noResponse.Error())
debugf("%s %s", noResponse.Error(), t.tuple.String())
dns.publishTransaction(t)
unmatchedRequests.Add(1)
}
// Adds the DNS message data to the supplied MapStr.
func addDNSToMapStr(m common.MapStr, dns *mkdns.Msg, authority bool, additional bool) {
m["id"] = dns.Id
m["op_code"] = dnsOpCodeToString(dns.Opcode)
m["flags"] = common.MapStr{
"authoritative": dns.Authoritative,
"truncated_response": dns.Truncated,
"recursion_desired": dns.RecursionDesired,
"recursion_available": dns.RecursionAvailable,
"authentic_data": dns.AuthenticatedData, // [RFC4035]
"checking_disabled": dns.CheckingDisabled, // [RFC4035]
}
m["response_code"] = dnsResponseCodeToString(dns.Rcode)
if len(dns.Question) > 0 {
q := dns.Question[0]
qMapStr := common.MapStr{
"name": q.Name,
"type": dnsTypeToString(q.Qtype),
"class": dnsClassToString(q.Qclass),
}
m["question"] = qMapStr
eTLDPlusOne, err := publicsuffix.EffectiveTLDPlusOne(strings.TrimRight(q.Name, "."))
if err == nil {
qMapStr["etld_plus_one"] = eTLDPlusOne + "."
}
}
rrOPT := dns.IsEdns0()
if rrOPT != nil {
m["opt"] = optToMapStr(rrOPT)
}
m["answers_count"] = len(dns.Answer)
if len(dns.Answer) > 0 {
m["answers"] = rrsToMapStrs(dns.Answer)
}
m["authorities_count"] = len(dns.Ns)
if authority && len(dns.Ns) > 0 {
m["authorities"] = rrsToMapStrs(dns.Ns)
}
if rrOPT != nil {
m["additionals_count"] = len(dns.Extra) - 1
} else {
m["additionals_count"] = len(dns.Extra)
}
if additional && len(dns.Extra) > 0 {
rrsMapStrs := rrsToMapStrs(dns.Extra)
// We do not want OPT RR to appear in the 'additional' section,
// that's why rrsMapStrs could be empty even though len(dns.Extra) > 0
if len(rrsMapStrs) > 0 {
m["additionals"] = rrsMapStrs
}
}
}
func optToMapStr(rrOPT *mkdns.OPT) common.MapStr {
optMapStr := common.MapStr{
"do": rrOPT.Do(), // true if DNSSEC
"version": strconv.FormatUint(uint64(rrOPT.Version()), 10),
"udp_size": rrOPT.UDPSize(),
"ext_rcode": dnsResponseCodeToString(rrOPT.ExtendedRcode()),
}
for _, o := range rrOPT.Option {
switch o.(type) {
case *mkdns.EDNS0_DAU:
optMapStr["dau"] = o.String()
case *mkdns.EDNS0_DHU:
optMapStr["dhu"] = o.String()
case *mkdns.EDNS0_EXPIRE:
optMapStr["local"] = o.String()
case *mkdns.EDNS0_LLQ:
optMapStr["llq"] = o.String()
case *mkdns.EDNS0_LOCAL:
optMapStr["local"] = o.String()
case *mkdns.EDNS0_N3U:
optMapStr["n3u"] = o.String()
case *mkdns.EDNS0_NSID:
optMapStr["nsid"] = o.String()
case *mkdns.EDNS0_SUBNET:
var draft string
if o.(*mkdns.EDNS0_SUBNET).DraftOption {
draft = " draft"
}
optMapStr["subnet"] = o.String() + draft
case *mkdns.EDNS0_UL:
optMapStr["ul"] = o.String()
}
}
return optMapStr
}
// rrsToMapStr converts an slice of RR's to an slice of MapStr's.
func rrsToMapStrs(records []mkdns.RR) []common.MapStr {
mapStrSlice := make([]common.MapStr, 0, len(records))
for _, rr := range records {
rrHeader := rr.Header()
mapStr := rrToMapStr(rr)
if len(mapStr) == 0 { // OPT pseudo-RR returns an empty MapStr
continue
}
mapStr["name"] = rrHeader.Name
mapStr["type"] = dnsTypeToString(rrHeader.Rrtype)
mapStr["class"] = dnsClassToString(rrHeader.Class)
mapStr["ttl"] = strconv.FormatInt(int64(rrHeader.Ttl), 10)
mapStrSlice = append(mapStrSlice, mapStr)
}
return mapStrSlice
}
// Convert all RDATA fields of a RR to a single string
// fields are ordered alphabetically with 'data' as the last element
//
// TODO An improvement would be to replace 'data' by the real field name
// It would require some changes in unit tests
func rrToString(rr mkdns.RR) string {
var st string
var keys []string
mapStr := rrToMapStr(rr)
data, ok := mapStr["data"]
delete(mapStr, "data")
for k := range mapStr {
keys = append(keys, k)
}
sort.Strings(keys)
var b bytes.Buffer
for _, k := range keys {
v := mapStr[k]
switch x := v.(type) {
case int:
fmt.Fprintf(&b, "%s %d, ", k, x)
case string:
fmt.Fprintf(&b, "%s %s, ", k, x)
}
}
if !ok {
st = strings.TrimSuffix(b.String(), ", ")
return st
}
switch x := data.(type) {
case int:
fmt.Fprintf(&b, "%d", x)
case string:
fmt.Fprintf(&b, "%s", x)
}
return b.String()
}
func rrToMapStr(rr mkdns.RR) common.MapStr {
mapStr := common.MapStr{}
rrType := rr.Header().Rrtype
switch x := rr.(type) {
default:
// We don't have special handling for this type
debugf("No special handling for RR type %s", dnsTypeToString(rrType))
unsupportedRR := new(mkdns.RFC3597)
err := unsupportedRR.ToRFC3597(x)
if err == nil {
rData, err := hexStringToString(unsupportedRR.Rdata)
mapStr["data"] = rData
if err != nil {
debugf("%s", err.Error())
}
} else {
debugf("Rdata for the unhandled RR type %s could not be fetched", dnsTypeToString(rrType))
}
case *mkdns.A:
mapStr["data"] = x.A.String()
case *mkdns.AAAA:
mapStr["data"] = x.AAAA.String()
case *mkdns.CNAME:
mapStr["data"] = x.Target
case *mkdns.DNSKEY:
mapStr["flags"] = strconv.Itoa(int(x.Flags))
mapStr["protocol"] = strconv.Itoa(int(x.Protocol))
mapStr["algorithm"] = dnsAlgorithmToString(x.Algorithm)
mapStr["data"] = x.PublicKey
case *mkdns.DS:
mapStr["key_tag"] = strconv.Itoa(int(x.KeyTag))
mapStr["algorithm"] = dnsAlgorithmToString(x.Algorithm)
mapStr["digest_type"] = dnsHashToString(x.DigestType)
mapStr["data"] = strings.ToUpper(x.Digest)
case *mkdns.MX:
mapStr["preference"] = x.Preference
mapStr["data"] = x.Mx
case *mkdns.NS:
mapStr["data"] = x.Ns
case *mkdns.NSEC:
mapStr["type_bits"] = dnsTypeBitsMapToString(x.TypeBitMap)
mapStr["data"] = x.NextDomain
case *mkdns.NSEC3:
mapStr["hash"] = dnsHashToString(x.Hash)
mapStr["flags"] = strconv.Itoa(int(x.Flags))
mapStr["iterations"] = strconv.Itoa(int(x.Iterations))
mapStr["salt"] = dnsSaltToString(x.Salt)
mapStr["type_bits"] = dnsTypeBitsMapToString(x.TypeBitMap)
mapStr["data"] = x.NextDomain
case *mkdns.NSEC3PARAM:
mapStr["hash"] = dnsHashToString(x.Hash)
mapStr["flags"] = strconv.Itoa(int(x.Flags))
mapStr["iterations"] = strconv.Itoa(int(x.Iterations))
mapStr["data"] = dnsSaltToString(x.Salt)
case *mkdns.OPT: // EDNS [RFC6891]
// OPT pseudo-RR is managed in addDnsToMapStr function
return nil
case *mkdns.PTR:
mapStr["data"] = x.Ptr
case *mkdns.RFC3597:
// Miekg/dns lib doesn't handle this type
debugf("Unknown RR type %s", dnsTypeToString(rrType))
rData, err := hexStringToString(x.Rdata)
mapStr["data"] = rData
if err != nil {
debugf("%s", err.Error())
}
case *mkdns.RRSIG:
mapStr["type_covered"] = dnsTypeToString(x.TypeCovered)
mapStr["algorithm"] = dnsAlgorithmToString(x.Algorithm)
mapStr["labels"] = strconv.Itoa(int(x.Labels))
mapStr["original_ttl"] = strconv.FormatInt(int64(x.OrigTtl), 10)
mapStr["expiration"] = mkdns.TimeToString(x.Expiration)
mapStr["inception"] = mkdns.TimeToString(x.Inception)
mapStr["key_tag"] = strconv.Itoa(int(x.KeyTag))
mapStr["signer_name"] = x.SignerName
mapStr["data"] = x.Signature
case *mkdns.SOA:
mapStr["rname"] = x.Mbox
mapStr["serial"] = x.Serial
mapStr["refresh"] = x.Refresh
mapStr["retry"] = x.Retry
mapStr["expire"] = x.Expire
mapStr["minimum"] = x.Minttl
mapStr["data"] = x.Ns
case *mkdns.SRV:
mapStr["priority"] = x.Priority
mapStr["weight"] = x.Weight
mapStr["port"] = x.Port
mapStr["data"] = x.Target
case *mkdns.TXT:
mapStr["data"] = strings.Join(x.Txt, " ")
}
return mapStr
}
// dnsQuestionToString converts a Question to a string.
func dnsQuestionToString(q mkdns.Question) string {
name := q.Name
return fmt.Sprintf("class %s, type %s, %s", dnsClassToString(q.Qclass),
dnsTypeToString(q.Qtype), name)
}
// rrsToString converts an array of RR's to a
// string.
func rrsToString(r []mkdns.RR) string {
var rrStrs []string
for _, rr := range r {
rrStrs = append(rrStrs, rrToString(rr))
}
return strings.Join(rrStrs, "; ")
}
// dnsToString converts a DNS message to a string.
func dnsToString(dns *mkdns.Msg) string {
var msgType string
if dns.Response {
msgType = "response"
} else {
msgType = "query"
}
var t []string
if dns.Authoritative {
t = append(t, "aa")
}
if dns.Truncated {
t = append(t, "tc")
}
if dns.RecursionDesired {
t = append(t, "rd")
}
if dns.RecursionAvailable {
t = append(t, "ra")
}
if dns.AuthenticatedData {
t = append(t, "ad")
}
if dns.CheckingDisabled {
t = append(t, "cd")
}
flags := strings.Join(t, " ")
var a []string
a = append(a, fmt.Sprintf("ID %d; QR %s; OPCODE %s; FLAGS %s; RCODE %s",
dns.Id, msgType, dnsOpCodeToString(dns.Opcode), flags,
dnsResponseCodeToString(dns.Rcode)))
if len(dns.Question) > 0 {
t = []string{}
for _, question := range dns.Question {
t = append(t, dnsQuestionToString(question))
}
a = append(a, fmt.Sprintf("QUESTION %s", strings.Join(t, "; ")))
}
if len(dns.Answer) > 0 {
a = append(a, fmt.Sprintf("ANSWER %s",
rrsToString(dns.Answer)))
}
if len(dns.Ns) > 0 {
a = append(a, fmt.Sprintf("AUTHORITY %s",
rrsToString(dns.Ns)))
}
if len(dns.Extra) > 0 {
a = append(a, fmt.Sprintf("ADDITIONAL %s",
rrsToString(dns.Extra)))
}
return strings.Join(a, "; ")
}
// decodeDnsData decodes a byte array into a DNS struct. If an error occurs
// then the returned dns pointer will be nil. This method recovers from panics
// and is concurrency-safe.
// We do not handle Unpack ErrTruncated for now. See https://github.com/miekg/dns/pull/281
func decodeDNSData(transp transport, rawData []byte) (dns *mkdns.Msg, err error) {
var offset int
if transp == transportTCP {
offset = decodeOffset
}
// Recover from any panics that occur while parsing a packet.
defer func() {
if r := recover(); r != nil {
err = fmt.Errorf("panic: %v", r)
}
}()
msg := &mkdns.Msg{}
err = msg.Unpack(rawData[offset:])
// Message should be more than 12 bytes.
// The 12 bytes value corresponds to a message header length.
// We use this check because Unpack does not return an error for some unvalid messages.
// TODO: can a better solution be found?
if msg.Len() <= 12 || err != nil {
return nil, nonDNSMsg
}
return msg, nil
}
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。