Merge branch 'dns' into dns-geo

This commit is contained in:
秋のかえで 2021-05-01 08:48:29 +08:00
commit ee682b17d1
No known key found for this signature in database
GPG key ID: B208CDDD55BEF92C
12 changed files with 910 additions and 48 deletions

View file

@ -362,6 +362,9 @@ func TestChinaSites(t *testing.T) {
matcher, err := domain.NewDomainMatcher(domains)
common.Must(err)
acMatcher, err := domain.NewMphMatcherGroup(domains)
common.Must(err)
type TestCase struct {
Domain string
Output bool
@ -390,9 +393,96 @@ func TestChinaSites(t *testing.T) {
}
for _, testCase := range testCases {
r := matcher.ApplyDomain(testCase.Domain)
if r != testCase.Output {
t.Error("expected output ", testCase.Output, " for domain ", testCase.Domain, " but got ", r)
r1 := matcher.ApplyDomain(testCase.Domain)
r2 := acMatcher.ApplyDomain(testCase.Domain)
if r1 != testCase.Output {
t.Error("DomainMatcher expected output ", testCase.Output, " for domain ", testCase.Domain, " but got ", r1)
} else if r2 != testCase.Output {
t.Error("ACDomainMatcher expected output ", testCase.Output, " for domain ", testCase.Domain, " but got ", r2)
}
}
}
func BenchmarkMphDomainMatcher(b *testing.B) {
domains, err := loadGeoSite("CN")
common.Must(err)
matcher, err := domain.NewMphMatcherGroup(domains)
common.Must(err)
type TestCase struct {
Domain string
Output bool
}
testCases := []TestCase{
{
Domain: "163.com",
Output: true,
},
{
Domain: "163.com",
Output: true,
},
{
Domain: "164.com",
Output: false,
},
{
Domain: "164.com",
Output: false,
},
}
for i := 0; i < 1024; i++ {
testCases = append(testCases, TestCase{Domain: strconv.Itoa(i) + ".not-exists.com", Output: false})
}
b.ResetTimer()
for i := 0; i < b.N; i++ {
for _, testCase := range testCases {
_ = matcher.ApplyDomain(testCase.Domain)
}
}
}
func BenchmarkDomainMatcher(b *testing.B) {
domains, err := loadGeoSite("CN")
common.Must(err)
matcher, err := domain.NewDomainMatcher(domains)
common.Must(err)
type TestCase struct {
Domain string
Output bool
}
testCases := []TestCase{
{
Domain: "163.com",
Output: true,
},
{
Domain: "163.com",
Output: true,
},
{
Domain: "164.com",
Output: false,
},
{
Domain: "164.com",
Output: false,
},
}
for i := 0; i < 1024; i++ {
testCases = append(testCases, TestCase{Domain: strconv.Itoa(i) + ".not-exists.com", Output: false})
}
b.ResetTimer()
for i := 0; i < b.N; i++ {
for _, testCase := range testCases {
_ = matcher.ApplyDomain(testCase.Domain)
}
}
}

View file

@ -30,11 +30,24 @@ func (rr *RoutingRule) BuildCondition() (Condition, error) {
conds := NewConditionChan()
if len(rr.Domain) > 0 {
switch rr.DomainMatcher {
case "linear":
matcher, err := dm.NewDomainMatcher(rr.Domain)
if err != nil {
return nil, newError("failed to build domain condition").Base(err)
}
conds.Add(matcher)
case "mph", "hybrid":
fallthrough
default:
matcher, err := dm.NewMphMatcherGroup(rr.Domain)
if err != nil {
return nil, newError("failed to build domain condition with MphDomainMatcher").Base(err)
}
newError("MphDomainMatcher is enabled for ", len(rr.Domain), " domain rule(s)").AtDebug().WriteToLog()
conds.Add(matcher)
}
}
if len(rr.UserEmail) > 0 {

View file

@ -1,13 +1,12 @@
// Code generated by protoc-gen-go. DO NOT EDIT.
// versions:
// protoc-gen-go v1.25.0
// protoc v3.15.6
// protoc-gen-go v1.26.0
// protoc v3.15.8
// source: app/router/config.proto
package router
import (
proto "github.com/golang/protobuf/proto"
domain "github.com/xtls/xray-core/common/matcher/domain"
geoip "github.com/xtls/xray-core/common/matcher/geoip"
net "github.com/xtls/xray-core/common/net"
@ -24,10 +23,6 @@ const (
_ = protoimpl.EnforceVersion(protoimpl.MaxVersion - 20)
)
// This is a compile-time assertion that a sufficiently up-to-date version
// of the legacy proto package is being used.
const _ = proto.ProtoPackageIsVersion4
type Config_DomainStrategy int32
const (
@ -131,6 +126,7 @@ type RoutingRule struct {
InboundTag []string `protobuf:"bytes,8,rep,name=inbound_tag,json=inboundTag,proto3" json:"inbound_tag,omitempty"`
Protocol []string `protobuf:"bytes,9,rep,name=protocol,proto3" json:"protocol,omitempty"`
Attributes string `protobuf:"bytes,15,opt,name=attributes,proto3" json:"attributes,omitempty"`
DomainMatcher string `protobuf:"bytes,17,opt,name=domain_matcher,json=domainMatcher,proto3" json:"domain_matcher,omitempty"`
}
func (x *RoutingRule) Reset() {
@ -288,6 +284,13 @@ func (x *RoutingRule) GetAttributes() string {
return ""
}
func (x *RoutingRule) GetDomainMatcher() string {
if x != nil {
return x.DomainMatcher
}
return ""
}
type isRoutingRule_TargetTag interface {
isRoutingRule_TargetTag()
}
@ -437,7 +440,7 @@ var file_app_router_config_proto_rawDesc = []byte{
0x69, 0x6e, 0x2f, 0x64, 0x6f, 0x6d, 0x61, 0x69, 0x6e, 0x2e, 0x70, 0x72, 0x6f, 0x74, 0x6f, 0x1a,
0x20, 0x63, 0x6f, 0x6d, 0x6d, 0x6f, 0x6e, 0x2f, 0x6d, 0x61, 0x74, 0x63, 0x68, 0x65, 0x72, 0x2f,
0x67, 0x65, 0x6f, 0x69, 0x70, 0x2f, 0x67, 0x65, 0x6f, 0x69, 0x70, 0x2e, 0x70, 0x72, 0x6f, 0x74,
0x6f, 0x22, 0xc1, 0x06, 0x0a, 0x0b, 0x52, 0x6f, 0x75, 0x74, 0x69, 0x6e, 0x67, 0x52, 0x75, 0x6c,
0x6f, 0x22, 0xe8, 0x06, 0x0a, 0x0b, 0x52, 0x6f, 0x75, 0x74, 0x69, 0x6e, 0x67, 0x52, 0x75, 0x6c,
0x65, 0x12, 0x12, 0x0a, 0x03, 0x74, 0x61, 0x67, 0x18, 0x01, 0x20, 0x01, 0x28, 0x09, 0x48, 0x00,
0x52, 0x03, 0x74, 0x61, 0x67, 0x12, 0x25, 0x0a, 0x0d, 0x62, 0x61, 0x6c, 0x61, 0x6e, 0x63, 0x69,
0x6e, 0x67, 0x5f, 0x74, 0x61, 0x67, 0x18, 0x0c, 0x20, 0x01, 0x28, 0x09, 0x48, 0x00, 0x52, 0x0c,
@ -488,36 +491,39 @@ var file_app_router_config_proto_rawDesc = []byte{
0x72, 0x6f, 0x74, 0x6f, 0x63, 0x6f, 0x6c, 0x18, 0x09, 0x20, 0x03, 0x28, 0x09, 0x52, 0x08, 0x70,
0x72, 0x6f, 0x74, 0x6f, 0x63, 0x6f, 0x6c, 0x12, 0x1e, 0x0a, 0x0a, 0x61, 0x74, 0x74, 0x72, 0x69,
0x62, 0x75, 0x74, 0x65, 0x73, 0x18, 0x0f, 0x20, 0x01, 0x28, 0x09, 0x52, 0x0a, 0x61, 0x74, 0x74,
0x72, 0x69, 0x62, 0x75, 0x74, 0x65, 0x73, 0x42, 0x0c, 0x0a, 0x0a, 0x74, 0x61, 0x72, 0x67, 0x65,
0x74, 0x5f, 0x74, 0x61, 0x67, 0x22, 0x4e, 0x0a, 0x0d, 0x42, 0x61, 0x6c, 0x61, 0x6e, 0x63, 0x69,
0x6e, 0x67, 0x52, 0x75, 0x6c, 0x65, 0x12, 0x10, 0x0a, 0x03, 0x74, 0x61, 0x67, 0x18, 0x01, 0x20,
0x01, 0x28, 0x09, 0x52, 0x03, 0x74, 0x61, 0x67, 0x12, 0x2b, 0x0a, 0x11, 0x6f, 0x75, 0x74, 0x62,
0x6f, 0x75, 0x6e, 0x64, 0x5f, 0x73, 0x65, 0x6c, 0x65, 0x63, 0x74, 0x6f, 0x72, 0x18, 0x02, 0x20,
0x03, 0x28, 0x09, 0x52, 0x10, 0x6f, 0x75, 0x74, 0x62, 0x6f, 0x75, 0x6e, 0x64, 0x53, 0x65, 0x6c,
0x65, 0x63, 0x74, 0x6f, 0x72, 0x22, 0x9b, 0x02, 0x0a, 0x06, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67,
0x12, 0x4f, 0x0a, 0x0f, 0x64, 0x6f, 0x6d, 0x61, 0x69, 0x6e, 0x5f, 0x73, 0x74, 0x72, 0x61, 0x74,
0x65, 0x67, 0x79, 0x18, 0x01, 0x20, 0x01, 0x28, 0x0e, 0x32, 0x26, 0x2e, 0x78, 0x72, 0x61, 0x79,
0x2e, 0x61, 0x70, 0x70, 0x2e, 0x72, 0x6f, 0x75, 0x74, 0x65, 0x72, 0x2e, 0x43, 0x6f, 0x6e, 0x66,
0x69, 0x67, 0x2e, 0x44, 0x6f, 0x6d, 0x61, 0x69, 0x6e, 0x53, 0x74, 0x72, 0x61, 0x74, 0x65, 0x67,
0x79, 0x52, 0x0e, 0x64, 0x6f, 0x6d, 0x61, 0x69, 0x6e, 0x53, 0x74, 0x72, 0x61, 0x74, 0x65, 0x67,
0x79, 0x12, 0x30, 0x0a, 0x04, 0x72, 0x75, 0x6c, 0x65, 0x18, 0x02, 0x20, 0x03, 0x28, 0x0b, 0x32,
0x1c, 0x2e, 0x78, 0x72, 0x61, 0x79, 0x2e, 0x61, 0x70, 0x70, 0x2e, 0x72, 0x6f, 0x75, 0x74, 0x65,
0x72, 0x2e, 0x52, 0x6f, 0x75, 0x74, 0x69, 0x6e, 0x67, 0x52, 0x75, 0x6c, 0x65, 0x52, 0x04, 0x72,
0x75, 0x6c, 0x65, 0x12, 0x45, 0x0a, 0x0e, 0x62, 0x61, 0x6c, 0x61, 0x6e, 0x63, 0x69, 0x6e, 0x67,
0x5f, 0x72, 0x75, 0x6c, 0x65, 0x18, 0x03, 0x20, 0x03, 0x28, 0x0b, 0x32, 0x1e, 0x2e, 0x78, 0x72,
0x61, 0x79, 0x2e, 0x61, 0x70, 0x70, 0x2e, 0x72, 0x6f, 0x75, 0x74, 0x65, 0x72, 0x2e, 0x42, 0x61,
0x6c, 0x61, 0x6e, 0x63, 0x69, 0x6e, 0x67, 0x52, 0x75, 0x6c, 0x65, 0x52, 0x0d, 0x62, 0x61, 0x6c,
0x61, 0x6e, 0x63, 0x69, 0x6e, 0x67, 0x52, 0x75, 0x6c, 0x65, 0x22, 0x47, 0x0a, 0x0e, 0x44, 0x6f,
0x6d, 0x61, 0x69, 0x6e, 0x53, 0x74, 0x72, 0x61, 0x74, 0x65, 0x67, 0x79, 0x12, 0x08, 0x0a, 0x04,
0x41, 0x73, 0x49, 0x73, 0x10, 0x00, 0x12, 0x09, 0x0a, 0x05, 0x55, 0x73, 0x65, 0x49, 0x70, 0x10,
0x01, 0x12, 0x10, 0x0a, 0x0c, 0x49, 0x70, 0x49, 0x66, 0x4e, 0x6f, 0x6e, 0x4d, 0x61, 0x74, 0x63,
0x68, 0x10, 0x02, 0x12, 0x0e, 0x0a, 0x0a, 0x49, 0x70, 0x4f, 0x6e, 0x44, 0x65, 0x6d, 0x61, 0x6e,
0x64, 0x10, 0x03, 0x42, 0x4f, 0x0a, 0x13, 0x63, 0x6f, 0x6d, 0x2e, 0x78, 0x72, 0x61, 0x79, 0x2e,
0x61, 0x70, 0x70, 0x2e, 0x72, 0x6f, 0x75, 0x74, 0x65, 0x72, 0x50, 0x01, 0x5a, 0x24, 0x67, 0x69,
0x74, 0x68, 0x75, 0x62, 0x2e, 0x63, 0x6f, 0x6d, 0x2f, 0x78, 0x74, 0x6c, 0x73, 0x2f, 0x78, 0x72,
0x61, 0x79, 0x2d, 0x63, 0x6f, 0x72, 0x65, 0x2f, 0x61, 0x70, 0x70, 0x2f, 0x72, 0x6f, 0x75, 0x74,
0x65, 0x72, 0xaa, 0x02, 0x0f, 0x58, 0x72, 0x61, 0x79, 0x2e, 0x41, 0x70, 0x70, 0x2e, 0x52, 0x6f,
0x75, 0x74, 0x65, 0x72, 0x62, 0x06, 0x70, 0x72, 0x6f, 0x74, 0x6f, 0x33,
0x72, 0x69, 0x62, 0x75, 0x74, 0x65, 0x73, 0x12, 0x25, 0x0a, 0x0e, 0x64, 0x6f, 0x6d, 0x61, 0x69,
0x6e, 0x5f, 0x6d, 0x61, 0x74, 0x63, 0x68, 0x65, 0x72, 0x18, 0x11, 0x20, 0x01, 0x28, 0x09, 0x52,
0x0d, 0x64, 0x6f, 0x6d, 0x61, 0x69, 0x6e, 0x4d, 0x61, 0x74, 0x63, 0x68, 0x65, 0x72, 0x42, 0x0c,
0x0a, 0x0a, 0x74, 0x61, 0x72, 0x67, 0x65, 0x74, 0x5f, 0x74, 0x61, 0x67, 0x22, 0x4e, 0x0a, 0x0d,
0x42, 0x61, 0x6c, 0x61, 0x6e, 0x63, 0x69, 0x6e, 0x67, 0x52, 0x75, 0x6c, 0x65, 0x12, 0x10, 0x0a,
0x03, 0x74, 0x61, 0x67, 0x18, 0x01, 0x20, 0x01, 0x28, 0x09, 0x52, 0x03, 0x74, 0x61, 0x67, 0x12,
0x2b, 0x0a, 0x11, 0x6f, 0x75, 0x74, 0x62, 0x6f, 0x75, 0x6e, 0x64, 0x5f, 0x73, 0x65, 0x6c, 0x65,
0x63, 0x74, 0x6f, 0x72, 0x18, 0x02, 0x20, 0x03, 0x28, 0x09, 0x52, 0x10, 0x6f, 0x75, 0x74, 0x62,
0x6f, 0x75, 0x6e, 0x64, 0x53, 0x65, 0x6c, 0x65, 0x63, 0x74, 0x6f, 0x72, 0x22, 0x9b, 0x02, 0x0a,
0x06, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67, 0x12, 0x4f, 0x0a, 0x0f, 0x64, 0x6f, 0x6d, 0x61, 0x69,
0x6e, 0x5f, 0x73, 0x74, 0x72, 0x61, 0x74, 0x65, 0x67, 0x79, 0x18, 0x01, 0x20, 0x01, 0x28, 0x0e,
0x32, 0x26, 0x2e, 0x78, 0x72, 0x61, 0x79, 0x2e, 0x61, 0x70, 0x70, 0x2e, 0x72, 0x6f, 0x75, 0x74,
0x65, 0x72, 0x2e, 0x43, 0x6f, 0x6e, 0x66, 0x69, 0x67, 0x2e, 0x44, 0x6f, 0x6d, 0x61, 0x69, 0x6e,
0x53, 0x74, 0x72, 0x61, 0x74, 0x65, 0x67, 0x79, 0x52, 0x0e, 0x64, 0x6f, 0x6d, 0x61, 0x69, 0x6e,
0x53, 0x74, 0x72, 0x61, 0x74, 0x65, 0x67, 0x79, 0x12, 0x30, 0x0a, 0x04, 0x72, 0x75, 0x6c, 0x65,
0x18, 0x02, 0x20, 0x03, 0x28, 0x0b, 0x32, 0x1c, 0x2e, 0x78, 0x72, 0x61, 0x79, 0x2e, 0x61, 0x70,
0x70, 0x2e, 0x72, 0x6f, 0x75, 0x74, 0x65, 0x72, 0x2e, 0x52, 0x6f, 0x75, 0x74, 0x69, 0x6e, 0x67,
0x52, 0x75, 0x6c, 0x65, 0x52, 0x04, 0x72, 0x75, 0x6c, 0x65, 0x12, 0x45, 0x0a, 0x0e, 0x62, 0x61,
0x6c, 0x61, 0x6e, 0x63, 0x69, 0x6e, 0x67, 0x5f, 0x72, 0x75, 0x6c, 0x65, 0x18, 0x03, 0x20, 0x03,
0x28, 0x0b, 0x32, 0x1e, 0x2e, 0x78, 0x72, 0x61, 0x79, 0x2e, 0x61, 0x70, 0x70, 0x2e, 0x72, 0x6f,
0x75, 0x74, 0x65, 0x72, 0x2e, 0x42, 0x61, 0x6c, 0x61, 0x6e, 0x63, 0x69, 0x6e, 0x67, 0x52, 0x75,
0x6c, 0x65, 0x52, 0x0d, 0x62, 0x61, 0x6c, 0x61, 0x6e, 0x63, 0x69, 0x6e, 0x67, 0x52, 0x75, 0x6c,
0x65, 0x22, 0x47, 0x0a, 0x0e, 0x44, 0x6f, 0x6d, 0x61, 0x69, 0x6e, 0x53, 0x74, 0x72, 0x61, 0x74,
0x65, 0x67, 0x79, 0x12, 0x08, 0x0a, 0x04, 0x41, 0x73, 0x49, 0x73, 0x10, 0x00, 0x12, 0x09, 0x0a,
0x05, 0x55, 0x73, 0x65, 0x49, 0x70, 0x10, 0x01, 0x12, 0x10, 0x0a, 0x0c, 0x49, 0x70, 0x49, 0x66,
0x4e, 0x6f, 0x6e, 0x4d, 0x61, 0x74, 0x63, 0x68, 0x10, 0x02, 0x12, 0x0e, 0x0a, 0x0a, 0x49, 0x70,
0x4f, 0x6e, 0x44, 0x65, 0x6d, 0x61, 0x6e, 0x64, 0x10, 0x03, 0x42, 0x4f, 0x0a, 0x13, 0x63, 0x6f,
0x6d, 0x2e, 0x78, 0x72, 0x61, 0x79, 0x2e, 0x61, 0x70, 0x70, 0x2e, 0x72, 0x6f, 0x75, 0x74, 0x65,
0x72, 0x50, 0x01, 0x5a, 0x24, 0x67, 0x69, 0x74, 0x68, 0x75, 0x62, 0x2e, 0x63, 0x6f, 0x6d, 0x2f,
0x78, 0x74, 0x6c, 0x73, 0x2f, 0x78, 0x72, 0x61, 0x79, 0x2d, 0x63, 0x6f, 0x72, 0x65, 0x2f, 0x61,
0x70, 0x70, 0x2f, 0x72, 0x6f, 0x75, 0x74, 0x65, 0x72, 0xaa, 0x02, 0x0f, 0x58, 0x72, 0x61, 0x79,
0x2e, 0x41, 0x70, 0x70, 0x2e, 0x52, 0x6f, 0x75, 0x74, 0x65, 0x72, 0x62, 0x06, 0x70, 0x72, 0x6f,
0x74, 0x6f, 0x33,
}
var (

View file

@ -61,6 +61,8 @@ message RoutingRule {
repeated string protocol = 9;
string attributes = 15;
string domain_matcher = 17;
}
message BalancingRule {

View file

@ -32,6 +32,24 @@ type DomainMatcher struct {
matchers str.IndexMatcher
}
func NewMphMatcherGroup(domains []*Domain) (*DomainMatcher, error) {
g := str.NewMphMatcherGroup()
for _, d := range domains {
matcherType, f := matcherTypeMap[d.Type]
if !f {
return nil, newError("unsupported domain type", d.Type)
}
_, err := g.AddPattern(d.Value, matcherType)
if err != nil {
return nil, err
}
}
g.Build()
return &DomainMatcher{
matchers: g,
}, nil
}
func NewDomainMatcher(domains []*Domain) (*DomainMatcher, error) {
g := new(str.MatcherGroup)
for _, d := range domains {
@ -48,7 +66,7 @@ func NewDomainMatcher(domains []*Domain) (*DomainMatcher, error) {
}
func (m *DomainMatcher) ApplyDomain(domain string) bool {
return len(m.matchers.Match(domain)) > 0
return len(m.matchers.Match(strings.ToLower(domain))) > 0
}
// Apply implements Condition.
@ -57,5 +75,5 @@ func (m *DomainMatcher) Apply(ctx routing.Context) bool {
if len(domain) == 0 {
return false
}
return m.ApplyDomain(strings.ToLower(domain))
return m.ApplyDomain(domain)
}

View file

@ -0,0 +1,243 @@
package str
import (
"container/list"
)
const validCharCount = 53
type MatchType struct {
matchType Type
exist bool
}
const (
TrieEdge bool = true
FailEdge bool = false
)
type Edge struct {
edgeType bool
nextNode int
}
type ACAutomaton struct {
trie [][validCharCount]Edge
fail []int
exists []MatchType
count int
}
func newNode() [validCharCount]Edge {
var s [validCharCount]Edge
for i := range s {
s[i] = Edge{
edgeType: FailEdge,
nextNode: 0,
}
}
return s
}
var char2Index = []int{
'A': 0,
'a': 0,
'B': 1,
'b': 1,
'C': 2,
'c': 2,
'D': 3,
'd': 3,
'E': 4,
'e': 4,
'F': 5,
'f': 5,
'G': 6,
'g': 6,
'H': 7,
'h': 7,
'I': 8,
'i': 8,
'J': 9,
'j': 9,
'K': 10,
'k': 10,
'L': 11,
'l': 11,
'M': 12,
'm': 12,
'N': 13,
'n': 13,
'O': 14,
'o': 14,
'P': 15,
'p': 15,
'Q': 16,
'q': 16,
'R': 17,
'r': 17,
'S': 18,
's': 18,
'T': 19,
't': 19,
'U': 20,
'u': 20,
'V': 21,
'v': 21,
'W': 22,
'w': 22,
'X': 23,
'x': 23,
'Y': 24,
'y': 24,
'Z': 25,
'z': 25,
'!': 26,
'$': 27,
'&': 28,
'\'': 29,
'(': 30,
')': 31,
'*': 32,
'+': 33,
',': 34,
';': 35,
'=': 36,
':': 37,
'%': 38,
'-': 39,
'.': 40,
'_': 41,
'~': 42,
'0': 43,
'1': 44,
'2': 45,
'3': 46,
'4': 47,
'5': 48,
'6': 49,
'7': 50,
'8': 51,
'9': 52,
}
func NewACAutomaton() *ACAutomaton {
var ac = new(ACAutomaton)
ac.trie = append(ac.trie, newNode())
ac.fail = append(ac.fail, 0)
ac.exists = append(ac.exists, MatchType{
matchType: Full,
exist: false,
})
return ac
}
func (ac *ACAutomaton) Add(domain string, t Type) {
var node = 0
for i := len(domain) - 1; i >= 0; i-- {
var idx = char2Index[domain[i]]
if ac.trie[node][idx].nextNode == 0 {
ac.count++
if len(ac.trie) < ac.count+1 {
ac.trie = append(ac.trie, newNode())
ac.fail = append(ac.fail, 0)
ac.exists = append(ac.exists, MatchType{
matchType: Full,
exist: false,
})
}
ac.trie[node][idx] = Edge{
edgeType: TrieEdge,
nextNode: ac.count,
}
}
node = ac.trie[node][idx].nextNode
}
ac.exists[node] = MatchType{
matchType: t,
exist: true,
}
switch t {
case Domain:
ac.exists[node] = MatchType{
matchType: Full,
exist: true,
}
var idx = char2Index['.']
if ac.trie[node][idx].nextNode == 0 {
ac.count++
if len(ac.trie) < ac.count+1 {
ac.trie = append(ac.trie, newNode())
ac.fail = append(ac.fail, 0)
ac.exists = append(ac.exists, MatchType{
matchType: Full,
exist: false,
})
}
ac.trie[node][idx] = Edge{
edgeType: TrieEdge,
nextNode: ac.count,
}
}
node = ac.trie[node][idx].nextNode
ac.exists[node] = MatchType{
matchType: t,
exist: true,
}
default:
break
}
}
func (ac *ACAutomaton) Build() {
var queue = list.New()
for i := 0; i < validCharCount; i++ {
if ac.trie[0][i].nextNode != 0 {
queue.PushBack(ac.trie[0][i])
}
}
for {
var front = queue.Front()
if front == nil {
break
} else {
var node = front.Value.(Edge).nextNode
queue.Remove(front)
for i := 0; i < validCharCount; i++ {
if ac.trie[node][i].nextNode != 0 {
ac.fail[ac.trie[node][i].nextNode] = ac.trie[ac.fail[node]][i].nextNode
queue.PushBack(ac.trie[node][i])
} else {
ac.trie[node][i] = Edge{
edgeType: FailEdge,
nextNode: ac.trie[ac.fail[node]][i].nextNode,
}
}
}
}
}
}
func (ac *ACAutomaton) Match(s string) bool {
var node = 0
var fullMatch = true
// 1. the match string is all through trie edge. FULL MATCH or DOMAIN
// 2. the match string is through a fail edge. NOT FULL MATCH
// 2.1 Through a fail edge, but there exists a valid node. SUBSTR
for i := len(s) - 1; i >= 0; i-- {
var idx = char2Index[s[i]]
fullMatch = fullMatch && ac.trie[node][idx].edgeType
node = ac.trie[node][idx].nextNode
switch ac.exists[node].matchType {
case Substr:
return true
case Domain:
if fullMatch {
return true
}
default:
break
}
}
return fullMatch && ac.exists[node].exist
}

View file

@ -8,6 +8,19 @@ import (
. "github.com/xtls/xray-core/common/matcher/str"
)
func BenchmarkACAutomaton(b *testing.B) {
ac := NewACAutomaton()
for i := 1; i <= 1024; i++ {
ac.Add(strconv.Itoa(i)+".v2ray.com", Domain)
}
ac.Build()
b.ResetTimer()
for i := 0; i < b.N; i++ {
_ = ac.Match("0.v2ray.com")
}
}
func BenchmarkDomainMatcherGroup(b *testing.B) {
g := new(DomainMatcherGroup)

View file

@ -0,0 +1,301 @@
package str
import (
"math/bits"
"regexp"
"sort"
"strings"
"unsafe"
)
// PrimeRK is the prime base used in Rabin-Karp algorithm.
const PrimeRK = 16777619
// calculate the rolling murmurHash of given string
func RollingHash(s string) uint32 {
h := uint32(0)
for i := len(s) - 1; i >= 0; i-- {
h = h*PrimeRK + uint32(s[i])
}
return h
}
// A MphMatcherGroup is divided into three parts:
// 1. `full` and `domain` patterns are matched by Rabin-Karp algorithm and minimal perfect hash table;
// 2. `substr` patterns are matched by ac automaton;
// 3. `regex` patterns are matched with the regex library.
type MphMatcherGroup struct {
ac *ACAutomaton
otherMatchers []matcherEntry
rules []string
level0 []uint32
level0Mask int
level1 []uint32
level1Mask int
count uint32
ruleMap *map[string]uint32
}
func (g *MphMatcherGroup) AddFullOrDomainPattern(pattern string, t Type) {
h := RollingHash(pattern)
switch t {
case Domain:
(*g.ruleMap)["."+pattern] = h*PrimeRK + uint32('.')
fallthrough
case Full:
(*g.ruleMap)[pattern] = h
default:
}
}
func NewMphMatcherGroup() *MphMatcherGroup {
return &MphMatcherGroup{
ac: nil,
otherMatchers: nil,
rules: nil,
level0: nil,
level0Mask: 0,
level1: nil,
level1Mask: 0,
count: 1,
ruleMap: &map[string]uint32{},
}
}
// AddPattern adds a pattern to MphMatcherGroup
func (g *MphMatcherGroup) AddPattern(pattern string, t Type) (uint32, error) {
switch t {
case Substr:
if g.ac == nil {
g.ac = NewACAutomaton()
}
g.ac.Add(pattern, t)
case Full, Domain:
pattern = strings.ToLower(pattern)
g.AddFullOrDomainPattern(pattern, t)
case Regex:
r, err := regexp.Compile(pattern)
if err != nil {
return 0, err
}
g.otherMatchers = append(g.otherMatchers, matcherEntry{
m: &regexMatcher{pattern: r},
id: g.count,
})
default:
panic("Unknown type")
}
return g.count, nil
}
// Build builds a minimal perfect hash table and ac automaton from insert rules
func (g *MphMatcherGroup) Build() {
if g.ac != nil {
g.ac.Build()
}
keyLen := len(*g.ruleMap)
if keyLen == 0 {
keyLen = 1
(*g.ruleMap)["empty___"] = RollingHash("empty___")
}
g.level0 = make([]uint32, nextPow2(keyLen/4))
g.level0Mask = len(g.level0) - 1
g.level1 = make([]uint32, nextPow2(keyLen))
g.level1Mask = len(g.level1) - 1
var sparseBuckets = make([][]int, len(g.level0))
var ruleIdx int
for rule, hash := range *g.ruleMap {
n := int(hash) & g.level0Mask
g.rules = append(g.rules, rule)
sparseBuckets[n] = append(sparseBuckets[n], ruleIdx)
ruleIdx++
}
g.ruleMap = nil
var buckets []indexBucket
for n, vals := range sparseBuckets {
if len(vals) > 0 {
buckets = append(buckets, indexBucket{n, vals})
}
}
sort.Sort(bySize(buckets))
occ := make([]bool, len(g.level1))
var tmpOcc []int
for _, bucket := range buckets {
var seed = uint32(0)
for {
findSeed := true
tmpOcc = tmpOcc[:0]
for _, i := range bucket.vals {
n := int(strhashFallback(unsafe.Pointer(&g.rules[i]), uintptr(seed))) & g.level1Mask
if occ[n] {
for _, n := range tmpOcc {
occ[n] = false
}
seed++
findSeed = false
break
}
occ[n] = true
tmpOcc = append(tmpOcc, n)
g.level1[n] = uint32(i)
}
if findSeed {
g.level0[bucket.n] = seed
break
}
}
}
}
func nextPow2(v int) int {
if v <= 1 {
return 1
}
const MaxUInt = ^uint(0)
n := (MaxUInt >> bits.LeadingZeros(uint(v))) + 1
return int(n)
}
// Lookup searches for s in t and returns its index and whether it was found.
func (g *MphMatcherGroup) Lookup(h uint32, s string) bool {
i0 := int(h) & g.level0Mask
seed := g.level0[i0]
i1 := int(strhashFallback(unsafe.Pointer(&s), uintptr(seed))) & g.level1Mask
n := g.level1[i1]
return s == g.rules[int(n)]
}
// Match implements IndexMatcher.Match.
func (g *MphMatcherGroup) Match(pattern string) []uint32 {
result := []uint32{}
hash := uint32(0)
for i := len(pattern) - 1; i >= 0; i-- {
hash = hash*PrimeRK + uint32(pattern[i])
if pattern[i] == '.' {
if g.Lookup(hash, pattern[i:]) {
result = append(result, 1)
return result
}
}
}
if g.Lookup(hash, pattern) {
result = append(result, 1)
return result
}
if g.ac != nil && g.ac.Match(pattern) {
result = append(result, 1)
return result
}
for _, e := range g.otherMatchers {
if e.m.Match(pattern) {
result = append(result, e.id)
return result
}
}
return nil
}
type indexBucket struct {
n int
vals []int
}
type bySize []indexBucket
func (s bySize) Len() int { return len(s) }
func (s bySize) Less(i, j int) bool { return len(s[i].vals) > len(s[j].vals) }
func (s bySize) Swap(i, j int) { s[i], s[j] = s[j], s[i] }
type stringStruct struct {
str unsafe.Pointer
len int
}
func strhashFallback(a unsafe.Pointer, h uintptr) uintptr {
x := (*stringStruct)(a)
return memhashFallback(x.str, h, uintptr(x.len))
}
const (
// Constants for multiplication: four random odd 64-bit numbers.
m1 = 16877499708836156737
m2 = 2820277070424839065
m3 = 9497967016996688599
m4 = 15839092249703872147
)
var hashkey = [4]uintptr{1, 1, 1, 1}
func memhashFallback(p unsafe.Pointer, seed, s uintptr) uintptr {
h := uint64(seed + s*hashkey[0])
tail:
switch {
case s == 0:
case s < 4:
h ^= uint64(*(*byte)(p))
h ^= uint64(*(*byte)(add(p, s>>1))) << 8
h ^= uint64(*(*byte)(add(p, s-1))) << 16
h = rotl31(h*m1) * m2
case s <= 8:
h ^= uint64(readUnaligned32(p))
h ^= uint64(readUnaligned32(add(p, s-4))) << 32
h = rotl31(h*m1) * m2
case s <= 16:
h ^= readUnaligned64(p)
h = rotl31(h*m1) * m2
h ^= readUnaligned64(add(p, s-8))
h = rotl31(h*m1) * m2
case s <= 32:
h ^= readUnaligned64(p)
h = rotl31(h*m1) * m2
h ^= readUnaligned64(add(p, 8))
h = rotl31(h*m1) * m2
h ^= readUnaligned64(add(p, s-16))
h = rotl31(h*m1) * m2
h ^= readUnaligned64(add(p, s-8))
h = rotl31(h*m1) * m2
default:
v1 := h
v2 := uint64(seed * hashkey[1])
v3 := uint64(seed * hashkey[2])
v4 := uint64(seed * hashkey[3])
for s >= 32 {
v1 ^= readUnaligned64(p)
v1 = rotl31(v1*m1) * m2
p = add(p, 8)
v2 ^= readUnaligned64(p)
v2 = rotl31(v2*m2) * m3
p = add(p, 8)
v3 ^= readUnaligned64(p)
v3 = rotl31(v3*m3) * m4
p = add(p, 8)
v4 ^= readUnaligned64(p)
v4 = rotl31(v4*m4) * m1
p = add(p, 8)
s -= 32
}
h = v1 ^ v2 ^ v3 ^ v4
goto tail
}
h ^= h >> 29
h *= m3
h ^= h >> 32
return uintptr(h)
}
func add(p unsafe.Pointer, x uintptr) unsafe.Pointer {
return unsafe.Pointer(uintptr(p) + x)
}
func readUnaligned32(p unsafe.Pointer) uint32 {
q := (*[4]byte)(p)
return uint32(q[0]) | uint32(q[1])<<8 | uint32(q[2])<<16 | uint32(q[3])<<24
}
func rotl31(x uint64) uint64 {
return (x << 31) | (x >> (64 - 31))
}
func readUnaligned64(p unsafe.Pointer) uint64 {
q := (*[8]byte)(p)
return uint64(q[0]) | uint64(q[1])<<8 | uint64(q[2])<<16 | uint64(q[3])<<24 | uint64(q[4])<<32 | uint64(q[5])<<40 | uint64(q[6])<<48 | uint64(q[7])<<56
}

View file

@ -27,6 +27,7 @@ const (
// New creates a new Matcher based on the given pattern.
func (t Type) New(pattern string) (Matcher, error) {
// 1. regex matching is case-sensitive
switch t {
case Full:
return fullMatcher(pattern), nil

View file

@ -91,3 +91,172 @@ func TestMatcherGroup(t *testing.T) {
}
}
}
func TestACAutomaton(t *testing.T) {
cases1 := []struct {
pattern string
mType Type
input string
output bool
}{
{
pattern: "xtls.github.io",
mType: Domain,
input: "www.xtls.github.io",
output: true,
},
{
pattern: "xtls.github.io",
mType: Domain,
input: "xtls.github.io",
output: true,
},
{
pattern: "xtls.github.io",
mType: Domain,
input: "www.xtis.github.io",
output: false,
},
{
pattern: "xtls.github.io",
mType: Domain,
input: "tls.github.io",
output: false,
},
{
pattern: "xtls.github.io",
mType: Domain,
input: "xxtls.github.io",
output: false,
},
{
pattern: "xtls.github.io",
mType: Full,
input: "xtls.github.io",
output: true,
},
{
pattern: "xtls.github.io",
mType: Full,
input: "xxtls.github.io",
output: false,
},
}
for _, test := range cases1 {
var ac = NewACAutomaton()
ac.Add(test.pattern, test.mType)
ac.Build()
if m := ac.Match(test.input); m != test.output {
t.Error("unexpected output: ", m, " for test case ", test)
}
}
{
cases2Input := []struct {
pattern string
mType Type
}{
{
pattern: "163.com",
mType: Domain,
},
{
pattern: "m.126.com",
mType: Full,
},
{
pattern: "3.com",
mType: Full,
},
{
pattern: "google.com",
mType: Substr,
},
{
pattern: "vgoogle.com",
mType: Substr,
},
}
var ac = NewACAutomaton()
for _, test := range cases2Input {
ac.Add(test.pattern, test.mType)
}
ac.Build()
cases2Output := []struct {
pattern string
res bool
}{
{
pattern: "126.com",
res: false,
},
{
pattern: "m.163.com",
res: true,
},
{
pattern: "mm163.com",
res: false,
},
{
pattern: "m.126.com",
res: true,
},
{
pattern: "163.com",
res: true,
},
{
pattern: "63.com",
res: false,
},
{
pattern: "oogle.com",
res: false,
},
{
pattern: "vvgoogle.com",
res: true,
},
}
for _, test := range cases2Output {
if m := ac.Match(test.pattern); m != test.res {
t.Error("unexpected output: ", m, " for test case ", test)
}
}
}
{
cases3Input := []struct {
pattern string
mType Type
}{
{
pattern: "video.google.com",
mType: Domain,
},
{
pattern: "gle.com",
mType: Domain,
},
}
var ac = NewACAutomaton()
for _, test := range cases3Input {
ac.Add(test.pattern, test.mType)
}
ac.Build()
cases3Output := []struct {
pattern string
res bool
}{
{
pattern: "google.com",
res: false,
},
}
for _, test := range cases3Output {
if m := ac.Match(test.pattern); m != test.res {
t.Error("unexpected output: ", m, " for test case ", test)
}
}
}
}

View file

@ -102,7 +102,7 @@ func ReadClientHello(data []byte, h *SniffHeader) error {
return errNotClientHello
}
if nameType == 0 {
serverName := strings.ToLower(string(d[:nameLen]))
serverName := string(d[:nameLen])
// An SNI value may not include a
// trailing dot. See
// https://tools.ietf.org/html/rfc6066#section-3.

View file

@ -97,6 +97,8 @@ type RouterRule struct {
Type string `json:"type"`
OutboundTag string `json:"outboundTag"`
BalancerTag string `json:"balancerTag"`
DomainMatcher string `json:"domainMatcher"`
}
func ParseIP(s string) (*geoip.CIDR, error) {
@ -182,6 +184,10 @@ func parseFieldRule(msg json.RawMessage) (*router.RoutingRule, error) {
return nil, newError("neither outboundTag nor balancerTag is specified in routing rule")
}
if rawFieldRule.DomainMatcher != "" {
rule.DomainMatcher = rawFieldRule.DomainMatcher
}
if rawFieldRule.Domain != nil {
for _, domain := range *rawFieldRule.Domain {
rules, err := conf.ParseDomainRule(domain)