src/base/net/geoipdatabase.cpp

   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 "geoipdatabase.h"
  30
  31 #include <QDateTime>
  32 #include <QDebug>
  33 #include <QFile>
  34 #include <QHostAddress>
  35 #include <QVariant>
  36
  37 #include "base/global.h"
  38 #include "base/path.h"
  39
  40 namespace
  41 {
  42     const qint32 MAX_FILE_SIZE = 67108864; // 64MB
  43     const quint32 MAX_METADATA_SIZE = 131072; // 128KB
  44     const char METADATA_BEGIN_MARK[] = "\xab\xcd\xefMaxMind.com";
  45     const char DATA_SECTION_SEPARATOR[16] = {0};
  46
  47     enum class DataType
  48     {
  49         Unknown = 0,
  50         Pointer = 1,
  51         String = 2,
  52         Double = 3,
  53         Bytes = 4,
  54         Integer16 = 5,
  55         Integer32 = 6,
  56         Map = 7,
  57         SignedInteger32 = 8,
  58         Integer64 = 9,
  59         Integer128 = 10,
  60         Array = 11,
  61         DataCacheContainer = 12,
  62         EndMarker = 13,
  63         Boolean = 14,
  64         Float = 15
  65     };
  66 }
  67
  68 struct DataFieldDescriptor
  69 {
  70     DataType fieldType {DataType::Unknown};
  71     union
  72     {
  73         quint32 fieldSize = 0;
  74         quint32 offset; // Pointer
  75     };
  76 };
  77
  78 GeoIPDatabase::GeoIPDatabase(const quint32 size)
  79     : m_ipVersion(0)
  80     , m_recordSize(0)
  81     , m_nodeCount(0)
  82     , m_nodeSize(0)
  83     , m_indexSize(0)
  84     , m_recordBytes(0)
  85     , m_size(size)
  86     , m_data(new uchar[size])
  87 {
  88 }
  89
  90 GeoIPDatabase *GeoIPDatabase::load(const Path &filename, QString &error)
  91 {
  92     GeoIPDatabase *db = nullptr;
  93     QFile file {filename.data()};
  94     if (file.size() > MAX_FILE_SIZE)
  95     {
  96         error = tr("Unsupported database file size.");
  97         return nullptr;
  98     }
  99
 100     if (!file.open(QFile::ReadOnly))
 101     {
 102         error = file.errorString();
 103         return nullptr;
 104     }
 105
 106     db = new GeoIPDatabase(file.size());
 107
 108     if (file.read(reinterpret_cast<char *>(db->m_data), db->m_size) != db->m_size)
 109     {
 110         error = file.errorString();
 111         delete db;
 112         return nullptr;
 113     }
 114
 115
 116     if (!db->parseMetadata(db->readMetadata(), error) || !db->loadDB(error))
 117     {
 118         delete db;
 119         return nullptr;
 120     }
 121
 122     return db;
 123 }
 124
 125 GeoIPDatabase *GeoIPDatabase::load(const QByteArray &data, QString &error)
 126 {
 127     GeoIPDatabase *db = nullptr;
 128     if (data.size() > MAX_FILE_SIZE)
 129     {
 130         error = tr("Unsupported database file size.");
 131         return nullptr;
 132     }
 133
 134     db = new GeoIPDatabase(data.size());
 135
 136     memcpy(reinterpret_cast<char *>(db->m_data), data.constData(), db->m_size);
 137
 138     if (!db->parseMetadata(db->readMetadata(), error) || !db->loadDB(error))
 139     {
 140         delete db;
 141         return nullptr;
 142     }
 143
 144     return db;
 145 }
 146
 147 GeoIPDatabase::~GeoIPDatabase()
 148 {
 149     delete [] m_data;
 150 }
 151
 152 QString GeoIPDatabase::type() const
 153 {
 154     return m_dbType;
 155 }
 156
 157 quint16 GeoIPDatabase::ipVersion() const
 158 {
 159     return m_ipVersion;
 160 }
 161
 162 QDateTime GeoIPDatabase::buildEpoch() const
 163 {
 164     return m_buildEpoch;
 165 }
 166
 167 QString GeoIPDatabase::lookup(const QHostAddress &hostAddr) const
 168 {
 169     Q_IPV6ADDR addr = hostAddr.toIPv6Address();
 170
 171     const uchar *ptr = m_data;
 172
 173     for (int i = 0; i < 16; ++i)
 174     {
 175         for (int j = 0; j < 8; ++j)
 176         {
 177             const bool right = static_cast<bool>((addr[i] >> (7 - j)) & 1);
 178             // Interpret the left/right record as number
 179             if (right)
 180                 ptr += m_recordBytes;
 181
 182             quint32 id = 0;
 183             auto *idPtr = reinterpret_cast<uchar *>(&id);
 184             memcpy(&idPtr[4 - m_recordBytes], ptr, m_recordBytes);
 185             fromBigEndian(idPtr, 4);
 186
 187             if (id == m_nodeCount)
 188             {
 189                 return {};
 190             }
 191             if (id > m_nodeCount)
 192             {
 193                 QString country = m_countries.value(id);
 194                 if (country.isEmpty())
 195                 {
 196                     const quint32 offset = id - m_nodeCount - sizeof(DATA_SECTION_SEPARATOR);
 197                     quint32 tmp = offset + m_indexSize + sizeof(DATA_SECTION_SEPARATOR);
 198                     const QVariant val = readDataField(tmp);
 199                     if (val.userType() == QMetaType::QVariantHash)
 200                     {
 201                         country = val.toHash()[u"country"_qs].toHash()[u"iso_code"_qs].toString();
 202                         m_countries[id] = country;
 203                     }
 204                 }
 205                 return country;
 206             }
 207
 208             ptr = m_data + (id * m_nodeSize);
 209         }
 210     }
 211
 212     return {};
 213 }
 214
 215 #define CHECK_METADATA_REQ(key, type) \
 216 if (!metadata.contains(key)) \
 217 { \
 218     error = errMsgNotFound.arg(key); \
 219     return false; \
 220 } \
 221 if (metadata.value(key).userType() != QMetaType::type) \
 222 { \
 223     error = errMsgInvalid.arg(key);  \
 224     return false; \
 225 }
 226
 227 #define CHECK_METADATA_OPT(key, type) \
 228 if (metadata.contains(key)) \
 229 { \
 230     if (metadata.value(key).userType() != QMetaType::type) \
 231     { \
 232         error = errMsgInvalid.arg(key);  \
 233         return false; \
 234     } \
 235 }
 236
 237 bool GeoIPDatabase::parseMetadata(const QVariantHash &metadata, QString &error)
 238 {
 239     const QString errMsgNotFound = tr("Metadata error: '%1' entry not found.");
 240     const QString errMsgInvalid = tr("Metadata error: '%1' entry has invalid type.");
 241
 242     qDebug() << "Parsing MaxMindDB metadata...";
 243
 244     CHECK_METADATA_REQ(u"binary_format_major_version"_qs, UShort);
 245     CHECK_METADATA_REQ(u"binary_format_minor_version"_qs, UShort);
 246     const uint versionMajor = metadata.value(u"binary_format_major_version"_qs).toUInt();
 247     const uint versionMinor = metadata.value(u"binary_format_minor_version"_qs).toUInt();
 248     if (versionMajor != 2)
 249     {
 250         error = tr("Unsupported database version: %1.%2").arg(versionMajor).arg(versionMinor);
 251         return false;
 252     }
 253
 254     CHECK_METADATA_REQ(u"ip_version"_qs, UShort);
 255     m_ipVersion = metadata.value(u"ip_version"_qs).value<quint16>();
 256     if (m_ipVersion != 6)
 257     {
 258         error = tr("Unsupported IP version: %1").arg(m_ipVersion);
 259         return false;
 260     }
 261
 262     CHECK_METADATA_REQ(u"record_size"_qs, UShort);
 263     m_recordSize = metadata.value(u"record_size"_qs).value<quint16>();
 264     if (m_recordSize != 24)
 265     {
 266         error = tr("Unsupported record size: %1").arg(m_recordSize);
 267         return false;
 268     }
 269     m_nodeSize = m_recordSize / 4;
 270     m_recordBytes = m_nodeSize / 2;
 271
 272     CHECK_METADATA_REQ(u"node_count"_qs, UInt);
 273     m_nodeCount = metadata.value(u"node_count"_qs).value<quint32>();
 274     m_indexSize = m_nodeCount * m_nodeSize;
 275
 276     CHECK_METADATA_REQ(u"database_type"_qs, QString);
 277     m_dbType = metadata.value(u"database_type"_qs).toString();
 278
 279     CHECK_METADATA_REQ(u"build_epoch"_qs, ULongLong);
 280     m_buildEpoch = QDateTime::fromSecsSinceEpoch(metadata.value(u"build_epoch"_qs).toULongLong());
 281
 282     CHECK_METADATA_OPT(u"languages"_qs, QVariantList);
 283     CHECK_METADATA_OPT(u"description"_qs, QVariantHash);
 284
 285     return true;
 286 }
 287
 288 bool GeoIPDatabase::loadDB(QString &error) const
 289 {
 290     qDebug() << "Parsing IP geolocation database index tree...";
 291
 292     const int nodeSize = m_recordSize / 4; // in bytes
 293     const int indexSize = m_nodeCount * nodeSize;
 294     if ((m_size < (indexSize + sizeof(DATA_SECTION_SEPARATOR)))
 295         || (memcmp(m_data + indexSize, DATA_SECTION_SEPARATOR, sizeof(DATA_SECTION_SEPARATOR)) != 0))
 296         {
 297         error = tr("Database corrupted: no data section found.");
 298         return false;
 299     }
 300
 301     return true;
 302 }
 303
 304 QVariantHash GeoIPDatabase::readMetadata() const
 305 {
 306     const char *ptr = reinterpret_cast<const char *>(m_data);
 307     quint32 size = m_size;
 308     if (m_size > MAX_METADATA_SIZE)
 309     {
 310         ptr += m_size - MAX_METADATA_SIZE;
 311         size = MAX_METADATA_SIZE;
 312     }
 313
 314     const QByteArray data = QByteArray::fromRawData(ptr, size);
 315     int index = data.lastIndexOf(METADATA_BEGIN_MARK);
 316     if (index >= 0)
 317     {
 318         if (m_size > MAX_METADATA_SIZE)
 319             index += (m_size - MAX_METADATA_SIZE); // from begin of all data
 320         auto offset = static_cast<quint32>(index + strlen(METADATA_BEGIN_MARK));
 321         const QVariant metadata = readDataField(offset);
 322         if (metadata.userType() == QMetaType::QVariantHash)
 323             return metadata.toHash();
 324     }
 325
 326     return {};
 327 }
 328
 329 QVariant GeoIPDatabase::readDataField(quint32 &offset) const
 330 {
 331     DataFieldDescriptor descr;
 332     if (!readDataFieldDescriptor(offset, descr))
 333         return {};
 334
 335     quint32 locOffset = offset;
 336     bool usePointer = false;
 337     if (descr.fieldType == DataType::Pointer)
 338     {
 339         usePointer = true;
 340         // convert offset from data section to global
 341         locOffset = descr.offset + (m_nodeCount * m_recordSize / 4) + sizeof(DATA_SECTION_SEPARATOR);
 342         if (!readDataFieldDescriptor(locOffset, descr))
 343             return {};
 344     }
 345
 346     QVariant fieldValue;
 347     switch (descr.fieldType)
 348     {
 349     case DataType::Pointer:
 350         qDebug() << "* Illegal Pointer using";
 351         break;
 352     case DataType::String:
 353         fieldValue = QString::fromUtf8(reinterpret_cast<const char *>(m_data + locOffset), descr.fieldSize);
 354         locOffset += descr.fieldSize;
 355         break;
 356     case DataType::Double:
 357         if (descr.fieldSize == 8)
 358             fieldValue = readPlainValue<double>(locOffset, descr.fieldSize);
 359         else
 360             qDebug() << "* Invalid field size for type: Double";
 361         break;
 362     case DataType::Bytes:
 363         fieldValue = QByteArray(reinterpret_cast<const char *>(m_data + locOffset), descr.fieldSize);
 364         locOffset += descr.fieldSize;
 365         break;
 366     case DataType::Integer16:
 367         fieldValue = readPlainValue<quint16>(locOffset, descr.fieldSize);
 368         break;
 369     case DataType::Integer32:
 370         fieldValue = readPlainValue<quint32>(locOffset, descr.fieldSize);
 371         break;
 372     case DataType::Map:
 373         fieldValue = readMapValue(locOffset, descr.fieldSize);
 374         break;
 375     case DataType::SignedInteger32:
 376         fieldValue = readPlainValue<qint32>(locOffset, descr.fieldSize);
 377         break;
 378     case DataType::Integer64:
 379         fieldValue = readPlainValue<quint64>(locOffset, descr.fieldSize);
 380         break;
 381     case DataType::Integer128:
 382         qDebug() << "* Unsupported data type: Integer128";
 383         break;
 384     case DataType::Array:
 385         fieldValue = readArrayValue(locOffset, descr.fieldSize);
 386         break;
 387     case DataType::DataCacheContainer:
 388         qDebug() << "* Unsupported data type: DataCacheContainer";
 389         break;
 390     case DataType::EndMarker:
 391         qDebug() << "* Unsupported data type: EndMarker";
 392         break;
 393     case DataType::Boolean:
 394         fieldValue = QVariant::fromValue(static_cast<bool>(descr.fieldSize));
 395         break;
 396     case DataType::Float:
 397         if (descr.fieldSize == 4)
 398             fieldValue = readPlainValue<float>(locOffset, descr.fieldSize);
 399         else
 400             qDebug() << "* Invalid field size for type: Float";
 401         break;
 402     default:
 403         qDebug() << "* Unsupported data type: Unknown";
 404     }
 405
 406     if (!usePointer)
 407         offset = locOffset;
 408     return fieldValue;
 409 }
 410
 411 bool GeoIPDatabase::readDataFieldDescriptor(quint32 &offset, DataFieldDescriptor &out) const
 412 {
 413     const uchar *dataPtr = m_data + offset;
 414     const int availSize = m_size - offset;
 415     if (availSize < 1) return false;
 416
 417     out.fieldType = static_cast<DataType>((dataPtr[0] & 0xE0) >> 5);
 418     if (out.fieldType == DataType::Pointer)
 419     {
 420         const int size = ((dataPtr[0] & 0x18) >> 3);
 421         if (availSize < (size + 2)) return false;
 422
 423         if (size == 0)
 424             out.offset = ((dataPtr[0] & 0x07) << 8) + dataPtr[1];
 425         else if (size == 1)
 426             out.offset = ((dataPtr[0] & 0x07) << 16) + (dataPtr[1] << 8) + dataPtr[2] + 2048;
 427         else if (size == 2)
 428             out.offset = ((dataPtr[0] & 0x07) << 24) + (dataPtr[1] << 16) + (dataPtr[2] << 8) + dataPtr[3] + 526336;
 429         else if (size == 3)
 430             out.offset = (dataPtr[1] << 24) + (dataPtr[2] << 16) + (dataPtr[3] << 8) + dataPtr[4];
 431
 432         offset += size + 2;
 433         return true;
 434     }
 435
 436     out.fieldSize = dataPtr[0] & 0x1F;
 437     if (out.fieldSize <= 28)
 438     {
 439         if (out.fieldType == DataType::Unknown)
 440         {
 441             out.fieldType = static_cast<DataType>(dataPtr[1] + 7);
 442             if ((out.fieldType <= DataType::Map) || (out.fieldType > DataType::Float) || (availSize < 3))
 443                 return false;
 444             offset += 2;
 445         }
 446         else
 447         {
 448             offset += 1;
 449         }
 450     }
 451     else if (out.fieldSize == 29)
 452     {
 453         if (availSize < 2) return false;
 454         out.fieldSize = dataPtr[1] + 29;
 455         offset += 2;
 456     }
 457     else if (out.fieldSize == 30)
 458     {
 459         if (availSize < 3) return false;
 460         out.fieldSize = (dataPtr[1] << 8) + dataPtr[2] + 285;
 461         offset += 3;
 462     }
 463     else if (out.fieldSize == 31)
 464     {
 465         if (availSize < 4) return false;
 466         out.fieldSize = (dataPtr[1] << 16) + (dataPtr[2] << 8) + dataPtr[3] + 65821;
 467         offset += 4;
 468     }
 469
 470     return true;
 471 }
 472
 473 void GeoIPDatabase::fromBigEndian(uchar *buf, const quint32 len) const
 474 {
 475 #if (Q_BYTE_ORDER == Q_LITTLE_ENDIAN)
 476     std::reverse(buf, buf + len);
 477 #else
 478     Q_UNUSED(buf);
 479     Q_UNUSED(len);
 480 #endif
 481 }
 482
 483 QVariant GeoIPDatabase::readMapValue(quint32 &offset, const quint32 count) const
 484 {
 485     QVariantHash map;
 486
 487     for (quint32 i = 0; i < count; ++i)
 488     {
 489         QVariant field = readDataField(offset);
 490         if (field.userType() != QMetaType::QString)
 491             return {};
 492
 493         const QString key = field.toString();
 494         field = readDataField(offset);
 495         if (field.userType() == QVariant::Invalid)
 496             return {};
 497
 498         map[key] = field;
 499     }
 500
 501     return map;
 502 }
 503
 504 QVariant GeoIPDatabase::readArrayValue(quint32 &offset, const quint32 count) const
 505 {
 506     QVariantList array;
 507
 508     for (quint32 i = 0; i < count; ++i)
 509     {
 510         const QVariant field = readDataField(offset);
 511         if (field.userType() == QVariant::Invalid)
 512             return {};
 513
 514         array.append(field);
 515     }
 516
 517     return array;
 518 }