summaryrefslogtreecommitdiff
path: root/vendor/github.com/oschwald/maxminddb-golang/reader.go
blob: 97b96070fc62c948cafe3f39a19b836f86a1bf4b (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
package maxminddb

import (
	"bytes"
	"errors"
	"fmt"
	"net"
	"reflect"
)

const (
	// NotFound is returned by LookupOffset when a matched root record offset
	// cannot be found.
	NotFound = ^uintptr(0)

	dataSectionSeparatorSize = 16
)

var metadataStartMarker = []byte("\xAB\xCD\xEFMaxMind.com")

// Reader holds the data corresponding to the MaxMind DB file. Its only public
// field is Metadata, which contains the metadata from the MaxMind DB file.
type Reader struct {
	hasMappedFile bool
	buffer        []byte
	decoder       decoder
	Metadata      Metadata
	ipv4Start     uint
}

// Metadata holds the metadata decoded from the MaxMind DB file. In particular
// in has the format version, the build time as Unix epoch time, the database
// type and description, the IP version supported, and a slice of the natural
// languages included.
type Metadata struct {
	BinaryFormatMajorVersion uint              `maxminddb:"binary_format_major_version"`
	BinaryFormatMinorVersion uint              `maxminddb:"binary_format_minor_version"`
	BuildEpoch               uint              `maxminddb:"build_epoch"`
	DatabaseType             string            `maxminddb:"database_type"`
	Description              map[string]string `maxminddb:"description"`
	IPVersion                uint              `maxminddb:"ip_version"`
	Languages                []string          `maxminddb:"languages"`
	NodeCount                uint              `maxminddb:"node_count"`
	RecordSize               uint              `maxminddb:"record_size"`
}

// FromBytes takes a byte slice corresponding to a MaxMind DB file and returns
// a Reader structure or an error.
func FromBytes(buffer []byte) (*Reader, error) {
	metadataStart := bytes.LastIndex(buffer, metadataStartMarker)

	if metadataStart == -1 {
		return nil, newInvalidDatabaseError("error opening database: invalid MaxMind DB file")
	}

	metadataStart += len(metadataStartMarker)
	metadataDecoder := decoder{buffer[metadataStart:]}

	var metadata Metadata

	rvMetdata := reflect.ValueOf(&metadata)
	_, err := metadataDecoder.decode(0, rvMetdata, 0)
	if err != nil {
		return nil, err
	}

	searchTreeSize := metadata.NodeCount * metadata.RecordSize / 4
	dataSectionStart := searchTreeSize + dataSectionSeparatorSize
	dataSectionEnd := uint(metadataStart - len(metadataStartMarker))
	if dataSectionStart > dataSectionEnd {
		return nil, newInvalidDatabaseError("the MaxMind DB contains invalid metadata")
	}
	d := decoder{
		buffer[searchTreeSize+dataSectionSeparatorSize : metadataStart-len(metadataStartMarker)],
	}

	reader := &Reader{
		buffer:    buffer,
		decoder:   d,
		Metadata:  metadata,
		ipv4Start: 0,
	}

	reader.ipv4Start, err = reader.startNode()

	return reader, err
}

func (r *Reader) startNode() (uint, error) {
	if r.Metadata.IPVersion != 6 {
		return 0, nil
	}

	nodeCount := r.Metadata.NodeCount

	node := uint(0)
	var err error
	for i := 0; i < 96 && node < nodeCount; i++ {
		node, err = r.readNode(node, 0)
		if err != nil {
			return 0, err
		}
	}
	return node, err
}

// Lookup takes an IP address as a net.IP structure and a pointer to the
// result value to Decode into.
func (r *Reader) Lookup(ipAddress net.IP, result interface{}) error {
	if r.buffer == nil {
		return errors.New("cannot call Lookup on a closed database")
	}
	pointer, err := r.lookupPointer(ipAddress)
	if pointer == 0 || err != nil {
		return err
	}
	return r.retrieveData(pointer, result)
}

// LookupOffset maps an argument net.IP to a corresponding record offset in the
// database. NotFound is returned if no such record is found, and a record may
// otherwise be extracted by passing the returned offset to Decode. LookupOffset
// is an advanced API, which exists to provide clients with a means to cache
// previously-decoded records.
func (r *Reader) LookupOffset(ipAddress net.IP) (uintptr, error) {
	if r.buffer == nil {
		return 0, errors.New("cannot call LookupOffset on a closed database")
	}
	pointer, err := r.lookupPointer(ipAddress)
	if pointer == 0 || err != nil {
		return NotFound, err
	}
	return r.resolveDataPointer(pointer)
}

// Decode the record at |offset| into |result|. The result value pointed to
// must be a data value that corresponds to a record in the database. This may
// include a struct representation of the data, a map capable of holding the
// data or an empty interface{} value.
//
// If result is a pointer to a struct, the struct need not include a field
// for every value that may be in the database. If a field is not present in
// the structure, the decoder will not decode that field, reducing the time
// required to decode the record.
//
// As a special case, a struct field of type uintptr will be used to capture
// the offset of the value. Decode may later be used to extract the stored
// value from the offset. MaxMind DBs are highly normalized: for example in
// the City database, all records of the same country will reference a
// single representative record for that country. This uintptr behavior allows
// clients to leverage this normalization in their own sub-record caching.
func (r *Reader) Decode(offset uintptr, result interface{}) error {
	if r.buffer == nil {
		return errors.New("cannot call Decode on a closed database")
	}
	return r.decode(offset, result)
}

func (r *Reader) decode(offset uintptr, result interface{}) error {
	rv := reflect.ValueOf(result)
	if rv.Kind() != reflect.Ptr || rv.IsNil() {
		return errors.New("result param must be a pointer")
	}

	_, err := r.decoder.decode(uint(offset), reflect.ValueOf(result), 0)
	return err
}

func (r *Reader) lookupPointer(ipAddress net.IP) (uint, error) {
	if ipAddress == nil {
		return 0, errors.New("ipAddress passed to Lookup cannot be nil")
	}

	ipV4Address := ipAddress.To4()
	if ipV4Address != nil {
		ipAddress = ipV4Address
	}
	if len(ipAddress) == 16 && r.Metadata.IPVersion == 4 {
		return 0, fmt.Errorf("error looking up '%s': you attempted to look up an IPv6 address in an IPv4-only database", ipAddress.String())
	}

	return r.findAddressInTree(ipAddress)
}

func (r *Reader) findAddressInTree(ipAddress net.IP) (uint, error) {

	bitCount := uint(len(ipAddress) * 8)

	var node uint
	if bitCount == 32 {
		node = r.ipv4Start
	}

	nodeCount := r.Metadata.NodeCount

	for i := uint(0); i < bitCount && node < nodeCount; i++ {
		bit := uint(1) & (uint(ipAddress[i>>3]) >> (7 - (i % 8)))

		var err error
		node, err = r.readNode(node, bit)
		if err != nil {
			return 0, err
		}
	}
	if node == nodeCount {
		// Record is empty
		return 0, nil
	} else if node > nodeCount {
		return node, nil
	}

	return 0, newInvalidDatabaseError("invalid node in search tree")
}

func (r *Reader) readNode(nodeNumber uint, index uint) (uint, error) {
	RecordSize := r.Metadata.RecordSize

	baseOffset := nodeNumber * RecordSize / 4

	var nodeBytes []byte
	var prefix uint
	switch RecordSize {
	case 24:
		offset := baseOffset + index*3
		nodeBytes = r.buffer[offset : offset+3]
	case 28:
		prefix = uint(r.buffer[baseOffset+3])
		if index != 0 {
			prefix &= 0x0F
		} else {
			prefix = (0xF0 & prefix) >> 4
		}
		offset := baseOffset + index*4
		nodeBytes = r.buffer[offset : offset+3]
	case 32:
		offset := baseOffset + index*4
		nodeBytes = r.buffer[offset : offset+4]
	default:
		return 0, newInvalidDatabaseError("unknown record size: %d", RecordSize)
	}
	return uintFromBytes(prefix, nodeBytes), nil
}

func (r *Reader) retrieveData(pointer uint, result interface{}) error {
	offset, err := r.resolveDataPointer(pointer)
	if err != nil {
		return err
	}
	return r.decode(offset, result)
}

func (r *Reader) resolveDataPointer(pointer uint) (uintptr, error) {
	var resolved = uintptr(pointer - r.Metadata.NodeCount - dataSectionSeparatorSize)

	if resolved > uintptr(len(r.buffer)) {
		return 0, newInvalidDatabaseError("the MaxMind DB file's search tree is corrupt")
	}
	return resolved, nil
}