1 /*
2  * Bittorrent Client using Qt and libtorrent.
3  * Copyright (C) 2015  Vladimir Golovnev <glassez@yandex.ru>
4  *
5  * This program is free software; you can redistribute it and/or
6  * modify it under the terms of the GNU General Public License
7  * as published by the Free Software Foundation; either version 2
8  * of the License, or (at your option) any later version.
9  *
10  * This program is distributed in the hope that it will be useful,
11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13  * GNU General Public License for more details.
14  *
15  * You should have received a copy of the GNU General Public License
16  * along with this program; if not, write to the Free Software
17  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
18  *
19  * In addition, as a special exception, the copyright holders give permission to
20  * link this program with the OpenSSL project's "OpenSSL" library (or with
21  * modified versions of it that use the same license as the "OpenSSL" library),
22  * and distribute the linked executables. You must obey the GNU General Public
23  * License in all respects for all of the code used other than "OpenSSL".  If you
24  * modify file(s), you may extend this exception to your version of the file(s),
25  * but you are not obligated to do so. If you do not wish to do so, delete this
26  * exception statement from your version.
27  */
28 
29 #include <QDateTime>
30 #include <QDebug>
31 #include <QFile>
32 #include <QHostAddress>
33 #include <QVariant>
34 
35 #include "geoipdatabase.h"
36 
37 namespace
38 {
39     const qint32 MAX_FILE_SIZE = 67108864; // 64MB
40     const quint32 MAX_METADATA_SIZE = 131072; // 128KB
41     const char METADATA_BEGIN_MARK[] = "\xab\xcd\xefMaxMind.com";
42     const char DATA_SECTION_SEPARATOR[16] = {0};
43 
44     enum class DataType
45     {
46         Unknown = 0,
47         Pointer = 1,
48         String = 2,
49         Double = 3,
50         Bytes = 4,
51         Integer16 = 5,
52         Integer32 = 6,
53         Map = 7,
54         SignedInteger32 = 8,
55         Integer64 = 9,
56         Integer128 = 10,
57         Array = 11,
58         DataCacheContainer = 12,
59         EndMarker = 13,
60         Boolean = 14,
61         Float = 15
62     };
63 }
64 
65 struct DataFieldDescriptor
66 {
67     DataType fieldType {DataType::Unknown};
68     union
69     {
70         quint32 fieldSize = 0;
71         quint32 offset; // Pointer
72     };
73 };
74 
GeoIPDatabase(const quint32 size)75 GeoIPDatabase::GeoIPDatabase(const quint32 size)
76     : m_ipVersion(0)
77     , m_recordSize(0)
78     , m_nodeCount(0)
79     , m_nodeSize(0)
80     , m_indexSize(0)
81     , m_recordBytes(0)
82     , m_size(size)
83     , m_data(new uchar[size])
84 {
85 }
86 
load(const QString & filename,QString & error)87 GeoIPDatabase *GeoIPDatabase::load(const QString &filename, QString &error)
88 {
89     GeoIPDatabase *db = nullptr;
90     QFile file(filename);
91     if (file.size() > MAX_FILE_SIZE)
92     {
93         error = tr("Unsupported database file size.");
94         return nullptr;
95     }
96 
97     if (!file.open(QFile::ReadOnly))
98     {
99         error = file.errorString();
100         return nullptr;
101     }
102 
103     db = new GeoIPDatabase(file.size());
104 
105     if (file.read(reinterpret_cast<char *>(db->m_data), db->m_size) != db->m_size)
106     {
107         error = file.errorString();
108         delete db;
109         return nullptr;
110     }
111 
112 
113     if (!db->parseMetadata(db->readMetadata(), error) || !db->loadDB(error))
114     {
115         delete db;
116         return nullptr;
117     }
118 
119     return db;
120 }
121 
load(const QByteArray & data,QString & error)122 GeoIPDatabase *GeoIPDatabase::load(const QByteArray &data, QString &error)
123 {
124     GeoIPDatabase *db = nullptr;
125     if (data.size() > MAX_FILE_SIZE)
126     {
127         error = tr("Unsupported database file size.");
128         return nullptr;
129     }
130 
131     db = new GeoIPDatabase(data.size());
132 
133     memcpy(reinterpret_cast<char *>(db->m_data), data.constData(), db->m_size);
134 
135     if (!db->parseMetadata(db->readMetadata(), error) || !db->loadDB(error))
136     {
137         delete db;
138         return nullptr;
139     }
140 
141     return db;
142 }
143 
~GeoIPDatabase()144 GeoIPDatabase::~GeoIPDatabase()
145 {
146     delete [] m_data;
147 }
148 
type() const149 QString GeoIPDatabase::type() const
150 {
151     return m_dbType;
152 }
153 
ipVersion() const154 quint16 GeoIPDatabase::ipVersion() const
155 {
156     return m_ipVersion;
157 }
158 
buildEpoch() const159 QDateTime GeoIPDatabase::buildEpoch() const
160 {
161     return m_buildEpoch;
162 }
163 
lookup(const QHostAddress & hostAddr) const164 QString GeoIPDatabase::lookup(const QHostAddress &hostAddr) const
165 {
166     Q_IPV6ADDR addr = hostAddr.toIPv6Address();
167 
168     const uchar *ptr = m_data;
169 
170     for (int i = 0; i < 16; ++i)
171     {
172         for (int j = 0; j < 8; ++j)
173         {
174             const bool right = static_cast<bool>((addr[i] >> (7 - j)) & 1);
175             // Interpret the left/right record as number
176             if (right)
177                 ptr += m_recordBytes;
178 
179             quint32 id = 0;
180             auto *idPtr = reinterpret_cast<uchar *>(&id);
181             memcpy(&idPtr[4 - m_recordBytes], ptr, m_recordBytes);
182             fromBigEndian(idPtr, 4);
183 
184             if (id == m_nodeCount)
185             {
186                 return {};
187             }
188             if (id > m_nodeCount)
189             {
190                 QString country = m_countries.value(id);
191                 if (country.isEmpty())
192                 {
193                     const quint32 offset = id - m_nodeCount - sizeof(DATA_SECTION_SEPARATOR);
194                     quint32 tmp = offset + m_indexSize + sizeof(DATA_SECTION_SEPARATOR);
195                     const QVariant val = readDataField(tmp);
196                     if (val.userType() == QMetaType::QVariantHash)
197                     {
198                         country = val.toHash()["country"].toHash()["iso_code"].toString();
199                         m_countries[id] = country;
200                     }
201                 }
202                 return country;
203             }
204 
205             ptr = m_data + (id * m_nodeSize);
206         }
207     }
208 
209     return {};
210 }
211 
212 #define CHECK_METADATA_REQ(key, type) \
213 if (!metadata.contains(#key)) \
214 { \
215     error = errMsgNotFound.arg(#key); \
216     return false; \
217 } \
218 if (metadata.value(#key).userType() != QMetaType::type) \
219 { \
220     error = errMsgInvalid.arg(#key);  \
221     return false; \
222 }
223 
224 #define CHECK_METADATA_OPT(key, type) \
225 if (metadata.contains(#key)) \
226 { \
227     if (metadata.value(#key).userType() != QMetaType::type) \
228     { \
229         error = errMsgInvalid.arg(#key);  \
230         return false; \
231     } \
232 }
233 
parseMetadata(const QVariantHash & metadata,QString & error)234 bool GeoIPDatabase::parseMetadata(const QVariantHash &metadata, QString &error)
235 {
236     const QString errMsgNotFound = tr("Metadata error: '%1' entry not found.");
237     const QString errMsgInvalid = tr("Metadata error: '%1' entry has invalid type.");
238 
239     qDebug() << "Parsing MaxMindDB metadata...";
240 
241     CHECK_METADATA_REQ(binary_format_major_version, UShort);
242     CHECK_METADATA_REQ(binary_format_minor_version, UShort);
243     const uint versionMajor = metadata.value("binary_format_major_version").toUInt();
244     const uint versionMinor = metadata.value("binary_format_minor_version").toUInt();
245     if (versionMajor != 2)
246     {
247         error = tr("Unsupported database version: %1.%2").arg(versionMajor).arg(versionMinor);
248         return false;
249     }
250 
251     CHECK_METADATA_REQ(ip_version, UShort);
252     m_ipVersion = metadata.value("ip_version").value<quint16>();
253     if (m_ipVersion != 6)
254     {
255         error = tr("Unsupported IP version: %1").arg(m_ipVersion);
256         return false;
257     }
258 
259     CHECK_METADATA_REQ(record_size, UShort);
260     m_recordSize = metadata.value("record_size").value<quint16>();
261     if (m_recordSize != 24)
262     {
263         error = tr("Unsupported record size: %1").arg(m_recordSize);
264         return false;
265     }
266     m_nodeSize = m_recordSize / 4;
267     m_recordBytes = m_nodeSize / 2;
268 
269     CHECK_METADATA_REQ(node_count, UInt);
270     m_nodeCount = metadata.value("node_count").value<quint32>();
271     m_indexSize = m_nodeCount * m_nodeSize;
272 
273     CHECK_METADATA_REQ(database_type, QString);
274     m_dbType = metadata.value("database_type").toString();
275 
276     CHECK_METADATA_REQ(build_epoch, ULongLong);
277     m_buildEpoch = QDateTime::fromSecsSinceEpoch(metadata.value("build_epoch").toULongLong());
278 
279     CHECK_METADATA_OPT(languages, QVariantList);
280     CHECK_METADATA_OPT(description, QVariantHash);
281 
282     return true;
283 }
284 
loadDB(QString & error) const285 bool GeoIPDatabase::loadDB(QString &error) const
286 {
287     qDebug() << "Parsing IP geolocation database index tree...";
288 
289     const int nodeSize = m_recordSize / 4; // in bytes
290     const int indexSize = m_nodeCount * nodeSize;
291     if ((m_size < (indexSize + sizeof(DATA_SECTION_SEPARATOR)))
292         || (memcmp(m_data + indexSize, DATA_SECTION_SEPARATOR, sizeof(DATA_SECTION_SEPARATOR)) != 0))
293         {
294         error = tr("Database corrupted: no data section found.");
295         return false;
296     }
297 
298     return true;
299 }
300 
readMetadata() const301 QVariantHash GeoIPDatabase::readMetadata() const
302 {
303     const char *ptr = reinterpret_cast<const char *>(m_data);
304     quint32 size = m_size;
305     if (m_size > MAX_METADATA_SIZE)
306     {
307         ptr += m_size - MAX_METADATA_SIZE;
308         size = MAX_METADATA_SIZE;
309     }
310 
311     const QByteArray data = QByteArray::fromRawData(ptr, size);
312     int index = data.lastIndexOf(METADATA_BEGIN_MARK);
313     if (index >= 0)
314     {
315         if (m_size > MAX_METADATA_SIZE)
316             index += (m_size - MAX_METADATA_SIZE); // from begin of all data
317         auto offset = static_cast<quint32>(index + strlen(METADATA_BEGIN_MARK));
318         const QVariant metadata = readDataField(offset);
319         if (metadata.userType() == QMetaType::QVariantHash)
320             return metadata.toHash();
321     }
322 
323     return {};
324 }
325 
readDataField(quint32 & offset) const326 QVariant GeoIPDatabase::readDataField(quint32 &offset) const
327 {
328     DataFieldDescriptor descr;
329     if (!readDataFieldDescriptor(offset, descr))
330         return {};
331 
332     quint32 locOffset = offset;
333     bool usePointer = false;
334     if (descr.fieldType == DataType::Pointer)
335     {
336         usePointer = true;
337         // convert offset from data section to global
338         locOffset = descr.offset + (m_nodeCount * m_recordSize / 4) + sizeof(DATA_SECTION_SEPARATOR);
339         if (!readDataFieldDescriptor(locOffset, descr))
340             return {};
341     }
342 
343     QVariant fieldValue;
344     switch (descr.fieldType)
345     {
346     case DataType::Pointer:
347         qDebug() << "* Illegal Pointer using";
348         break;
349     case DataType::String:
350         fieldValue = QString::fromUtf8(reinterpret_cast<const char *>(m_data + locOffset), descr.fieldSize);
351         locOffset += descr.fieldSize;
352         break;
353     case DataType::Double:
354         if (descr.fieldSize == 8)
355             fieldValue = readPlainValue<double>(locOffset, descr.fieldSize);
356         else
357             qDebug() << "* Invalid field size for type: Double";
358         break;
359     case DataType::Bytes:
360         fieldValue = QByteArray(reinterpret_cast<const char *>(m_data + locOffset), descr.fieldSize);
361         locOffset += descr.fieldSize;
362         break;
363     case DataType::Integer16:
364         fieldValue = readPlainValue<quint16>(locOffset, descr.fieldSize);
365         break;
366     case DataType::Integer32:
367         fieldValue = readPlainValue<quint32>(locOffset, descr.fieldSize);
368         break;
369     case DataType::Map:
370         fieldValue = readMapValue(locOffset, descr.fieldSize);
371         break;
372     case DataType::SignedInteger32:
373         fieldValue = readPlainValue<qint32>(locOffset, descr.fieldSize);
374         break;
375     case DataType::Integer64:
376         fieldValue = readPlainValue<quint64>(locOffset, descr.fieldSize);
377         break;
378     case DataType::Integer128:
379         qDebug() << "* Unsupported data type: Integer128";
380         break;
381     case DataType::Array:
382         fieldValue = readArrayValue(locOffset, descr.fieldSize);
383         break;
384     case DataType::DataCacheContainer:
385         qDebug() << "* Unsupported data type: DataCacheContainer";
386         break;
387     case DataType::EndMarker:
388         qDebug() << "* Unsupported data type: EndMarker";
389         break;
390     case DataType::Boolean:
391         fieldValue = QVariant::fromValue(static_cast<bool>(descr.fieldSize));
392         break;
393     case DataType::Float:
394         if (descr.fieldSize == 4)
395             fieldValue = readPlainValue<float>(locOffset, descr.fieldSize);
396         else
397             qDebug() << "* Invalid field size for type: Float";
398         break;
399     default:
400         qDebug() << "* Unsupported data type: Unknown";
401     }
402 
403     if (!usePointer)
404         offset = locOffset;
405     return fieldValue;
406 }
407 
readDataFieldDescriptor(quint32 & offset,DataFieldDescriptor & out) const408 bool GeoIPDatabase::readDataFieldDescriptor(quint32 &offset, DataFieldDescriptor &out) const
409 {
410     const uchar *dataPtr = m_data + offset;
411     const int availSize = m_size - offset;
412     if (availSize < 1) return false;
413 
414     out.fieldType = static_cast<DataType>((dataPtr[0] & 0xE0) >> 5);
415     if (out.fieldType == DataType::Pointer)
416     {
417         const int size = ((dataPtr[0] & 0x18) >> 3);
418         if (availSize < (size + 2)) return false;
419 
420         if (size == 0)
421             out.offset = ((dataPtr[0] & 0x07) << 8) + dataPtr[1];
422         else if (size == 1)
423             out.offset = ((dataPtr[0] & 0x07) << 16) + (dataPtr[1] << 8) + dataPtr[2] + 2048;
424         else if (size == 2)
425             out.offset = ((dataPtr[0] & 0x07) << 24) + (dataPtr[1] << 16) + (dataPtr[2] << 8) + dataPtr[3] + 526336;
426         else if (size == 3)
427             out.offset = (dataPtr[1] << 24) + (dataPtr[2] << 16) + (dataPtr[3] << 8) + dataPtr[4];
428 
429         offset += size + 2;
430         return true;
431     }
432 
433     out.fieldSize = dataPtr[0] & 0x1F;
434     if (out.fieldSize <= 28)
435     {
436         if (out.fieldType == DataType::Unknown)
437         {
438             out.fieldType = static_cast<DataType>(dataPtr[1] + 7);
439             if ((out.fieldType <= DataType::Map) || (out.fieldType > DataType::Float) || (availSize < 3))
440                 return false;
441             offset += 2;
442         }
443         else
444         {
445             offset += 1;
446         }
447     }
448     else if (out.fieldSize == 29)
449     {
450         if (availSize < 2) return false;
451         out.fieldSize = dataPtr[1] + 29;
452         offset += 2;
453     }
454     else if (out.fieldSize == 30)
455     {
456         if (availSize < 3) return false;
457         out.fieldSize = (dataPtr[1] << 8) + dataPtr[2] + 285;
458         offset += 3;
459     }
460     else if (out.fieldSize == 31)
461     {
462         if (availSize < 4) return false;
463         out.fieldSize = (dataPtr[1] << 16) + (dataPtr[2] << 8) + dataPtr[3] + 65821;
464         offset += 4;
465     }
466 
467     return true;
468 }
469 
fromBigEndian(uchar * buf,const quint32 len) const470 void GeoIPDatabase::fromBigEndian(uchar *buf, const quint32 len) const
471 {
472 #if (Q_BYTE_ORDER == Q_LITTLE_ENDIAN)
473     std::reverse(buf, buf + len);
474 #else
475     Q_UNUSED(buf);
476     Q_UNUSED(len);
477 #endif
478 }
479 
readMapValue(quint32 & offset,const quint32 count) const480 QVariant GeoIPDatabase::readMapValue(quint32 &offset, const quint32 count) const
481 {
482     QVariantHash map;
483 
484     for (quint32 i = 0; i < count; ++i)
485     {
486         QVariant field = readDataField(offset);
487         if (field.userType() != QMetaType::QString)
488             return {};
489 
490         const QString key = field.toString();
491         field = readDataField(offset);
492         if (field.userType() == QVariant::Invalid)
493             return {};
494 
495         map[key] = field;
496     }
497 
498     return map;
499 }
500 
readArrayValue(quint32 & offset,const quint32 count) const501 QVariant GeoIPDatabase::readArrayValue(quint32 &offset, const quint32 count) const
502 {
503     QVariantList array;
504 
505     for (quint32 i = 0; i < count; ++i)
506     {
507         const QVariant field = readDataField(offset);
508         if (field.userType() == QVariant::Invalid)
509             return {};
510 
511         array.append(field);
512     }
513 
514     return array;
515 }
516