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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879
|
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
//
#pragma once
#include <algorithm>
#include <limits>
#ifdef ROCKSDB_MALLOC_USABLE_SIZE
#ifdef OS_FREEBSD
#include <malloc_np.h>
#else // OS_FREEBSD
#include <malloc.h>
#endif // OS_FREEBSD
#endif // ROCKSDB_MALLOC_USABLE_SIZE
#include <string>
#include "memory/memory_allocator_impl.h"
#include "rocksdb/options.h"
#include "rocksdb/table.h"
#include "table/block_based/block_type.h"
#include "test_util/sync_point.h"
#include "util/coding.h"
#include "util/compression_context_cache.h"
#include "util/string_util.h"
#ifdef SNAPPY
#include <snappy.h>
#endif
#ifdef ZLIB
#include <zlib.h>
#endif
#ifdef BZIP2
#include <bzlib.h>
#endif
#if defined(LZ4)
#include <lz4.h>
#include <lz4hc.h>
#endif
#if defined(ZSTD)
#include <zstd.h>
// v1.1.3+
#if ZSTD_VERSION_NUMBER >= 10103
#include <zdict.h>
#endif // ZSTD_VERSION_NUMBER >= 10103
// v1.4.0+
// ZSTD_Compress2(), ZSTD_compressStream2() and frame parameters all belong to
// advanced APIs and require v1.4.0+.
// https://github.com/facebook/zstd/blob/eb9f881eb810f2242f1ef36b3f3e7014eecb8fa6/lib/zstd.h#L297C40-L297C45
#if ZSTD_VERSION_NUMBER >= 10400
#define ZSTD_ADVANCED
#endif // ZSTD_VERSION_NUMBER >= 10400
namespace ROCKSDB_NAMESPACE {
// Need this for the context allocation override
// On windows we need to do this explicitly
#if (ZSTD_VERSION_NUMBER >= 500)
#if defined(ROCKSDB_JEMALLOC) && defined(OS_WIN) && \
defined(ZSTD_STATIC_LINKING_ONLY)
#define ROCKSDB_ZSTD_CUSTOM_MEM
namespace port {
ZSTD_customMem GetJeZstdAllocationOverrides();
} // namespace port
#endif // defined(ROCKSDB_JEMALLOC) && defined(OS_WIN) &&
// defined(ZSTD_STATIC_LINKING_ONLY)
// We require `ZSTD_sizeof_DDict` and `ZSTD_createDDict_byReference` to use
// `ZSTD_DDict`. The former was introduced in v1.0.0 and the latter was
// introduced in v1.1.3. But an important bug fix for `ZSTD_sizeof_DDict` came
// in v1.1.4, so that is the version we require. As of today's latest version
// (v1.3.8), they are both still in the experimental API, which means they are
// only exported when the compiler flag `ZSTD_STATIC_LINKING_ONLY` is set.
#if defined(ZSTD_STATIC_LINKING_ONLY) && ZSTD_VERSION_NUMBER >= 10104
#define ROCKSDB_ZSTD_DDICT
#endif // defined(ZSTD_STATIC_LINKING_ONLY) && ZSTD_VERSION_NUMBER >= 10104
// Cached data represents a portion that can be re-used
// If, in the future we have more than one native context to
// cache we can arrange this as a tuple
class ZSTDUncompressCachedData {
public:
using ZSTDNativeContext = ZSTD_DCtx*;
ZSTDUncompressCachedData() {}
// Init from cache
ZSTDUncompressCachedData(const ZSTDUncompressCachedData& o) = delete;
ZSTDUncompressCachedData& operator=(const ZSTDUncompressCachedData&) = delete;
ZSTDUncompressCachedData(ZSTDUncompressCachedData&& o) noexcept
: ZSTDUncompressCachedData() {
*this = std::move(o);
}
ZSTDUncompressCachedData& operator=(ZSTDUncompressCachedData&& o) noexcept {
assert(zstd_ctx_ == nullptr);
std::swap(zstd_ctx_, o.zstd_ctx_);
std::swap(cache_idx_, o.cache_idx_);
return *this;
}
ZSTDNativeContext Get() const { return zstd_ctx_; }
int64_t GetCacheIndex() const { return cache_idx_; }
void CreateIfNeeded() {
if (zstd_ctx_ == nullptr) {
#ifdef ROCKSDB_ZSTD_CUSTOM_MEM
zstd_ctx_ =
ZSTD_createDCtx_advanced(port::GetJeZstdAllocationOverrides());
#else // ROCKSDB_ZSTD_CUSTOM_MEM
zstd_ctx_ = ZSTD_createDCtx();
#endif // ROCKSDB_ZSTD_CUSTOM_MEM
cache_idx_ = -1;
}
}
void InitFromCache(const ZSTDUncompressCachedData& o, int64_t idx) {
zstd_ctx_ = o.zstd_ctx_;
cache_idx_ = idx;
}
~ZSTDUncompressCachedData() {
if (zstd_ctx_ != nullptr && cache_idx_ == -1) {
ZSTD_freeDCtx(zstd_ctx_);
}
}
private:
ZSTDNativeContext zstd_ctx_ = nullptr;
int64_t cache_idx_ = -1; // -1 means this instance owns the context
};
#endif // (ZSTD_VERSION_NUMBER >= 500)
} // namespace ROCKSDB_NAMESPACE
#endif // ZSTD
#if !(defined ZSTD) || !(ZSTD_VERSION_NUMBER >= 500)
namespace ROCKSDB_NAMESPACE {
class ZSTDUncompressCachedData {
void* padding; // unused
public:
using ZSTDNativeContext = void*;
ZSTDUncompressCachedData() {}
ZSTDUncompressCachedData(const ZSTDUncompressCachedData&) {}
ZSTDUncompressCachedData& operator=(const ZSTDUncompressCachedData&) = delete;
ZSTDUncompressCachedData(ZSTDUncompressCachedData&&) noexcept = default;
ZSTDUncompressCachedData& operator=(ZSTDUncompressCachedData&&) noexcept =
default;
ZSTDNativeContext Get() const { return nullptr; }
int64_t GetCacheIndex() const { return -1; }
void CreateIfNeeded() {}
void InitFromCache(const ZSTDUncompressCachedData&, int64_t) {}
private:
void ignore_padding__() { padding = nullptr; }
};
} // namespace ROCKSDB_NAMESPACE
#endif
#if defined(XPRESS)
#include "port/xpress.h"
#endif
namespace ROCKSDB_NAMESPACE {
// Holds dictionary and related data, like ZSTD's digested compression
// dictionary.
struct CompressionDict {
#if ZSTD_VERSION_NUMBER >= 700
ZSTD_CDict* zstd_cdict_ = nullptr;
#endif // ZSTD_VERSION_NUMBER >= 700
std::string dict_;
public:
#if ZSTD_VERSION_NUMBER >= 700
CompressionDict(std::string dict, CompressionType type, int level) {
#else // ZSTD_VERSION_NUMBER >= 700
CompressionDict(std::string dict, CompressionType /*type*/, int /*level*/) {
#endif // ZSTD_VERSION_NUMBER >= 700
dict_ = std::move(dict);
#if ZSTD_VERSION_NUMBER >= 700
zstd_cdict_ = nullptr;
if (!dict_.empty() && (type == kZSTD || type == kZSTDNotFinalCompression)) {
if (level == CompressionOptions::kDefaultCompressionLevel) {
// 3 is the value of ZSTD_CLEVEL_DEFAULT (not exposed publicly), see
// https://github.com/facebook/zstd/issues/1148
// TODO(cbi): ZSTD_CLEVEL_DEFAULT is exposed after
// https://github.com/facebook/zstd/pull/1174. Use ZSTD_CLEVEL_DEFAULT
// instead of hardcoding 3.
level = 3;
}
// Should be safe (but slower) if below call fails as we'll use the
// raw dictionary to compress.
zstd_cdict_ = ZSTD_createCDict(dict_.data(), dict_.size(), level);
assert(zstd_cdict_ != nullptr);
}
#endif // ZSTD_VERSION_NUMBER >= 700
}
~CompressionDict() {
#if ZSTD_VERSION_NUMBER >= 700
size_t res = 0;
if (zstd_cdict_ != nullptr) {
res = ZSTD_freeCDict(zstd_cdict_);
}
assert(res == 0); // Last I checked they can't fail
(void)res; // prevent unused var warning
#endif // ZSTD_VERSION_NUMBER >= 700
}
#if ZSTD_VERSION_NUMBER >= 700
const ZSTD_CDict* GetDigestedZstdCDict() const { return zstd_cdict_; }
#endif // ZSTD_VERSION_NUMBER >= 700
Slice GetRawDict() const { return dict_; }
static const CompressionDict& GetEmptyDict() {
static CompressionDict empty_dict{};
return empty_dict;
}
CompressionDict() = default;
// Disable copy/move
CompressionDict(const CompressionDict&) = delete;
CompressionDict& operator=(const CompressionDict&) = delete;
CompressionDict(CompressionDict&&) = delete;
CompressionDict& operator=(CompressionDict&&) = delete;
};
// Holds dictionary and related data, like ZSTD's digested uncompression
// dictionary.
struct UncompressionDict {
// Block containing the data for the compression dictionary in case the
// constructor that takes a string parameter is used.
std::string dict_;
// Block containing the data for the compression dictionary in case the
// constructor that takes a Slice parameter is used and the passed in
// CacheAllocationPtr is not nullptr.
CacheAllocationPtr allocation_;
// Slice pointing to the compression dictionary data. Can point to
// dict_, allocation_, or some other memory location, depending on how
// the object was constructed.
Slice slice_;
#ifdef ROCKSDB_ZSTD_DDICT
// Processed version of the contents of slice_ for ZSTD compression.
ZSTD_DDict* zstd_ddict_ = nullptr;
#endif // ROCKSDB_ZSTD_DDICT
#ifdef ROCKSDB_ZSTD_DDICT
UncompressionDict(std::string dict, bool using_zstd)
#else // ROCKSDB_ZSTD_DDICT
UncompressionDict(std::string dict, bool /* using_zstd */)
#endif // ROCKSDB_ZSTD_DDICT
: dict_(std::move(dict)), slice_(dict_) {
#ifdef ROCKSDB_ZSTD_DDICT
if (!slice_.empty() && using_zstd) {
zstd_ddict_ = ZSTD_createDDict_byReference(slice_.data(), slice_.size());
assert(zstd_ddict_ != nullptr);
}
#endif // ROCKSDB_ZSTD_DDICT
}
#ifdef ROCKSDB_ZSTD_DDICT
UncompressionDict(Slice slice, CacheAllocationPtr&& allocation,
bool using_zstd)
#else // ROCKSDB_ZSTD_DDICT
UncompressionDict(Slice slice, CacheAllocationPtr&& allocation,
bool /* using_zstd */)
#endif // ROCKSDB_ZSTD_DDICT
: allocation_(std::move(allocation)), slice_(std::move(slice)) {
#ifdef ROCKSDB_ZSTD_DDICT
if (!slice_.empty() && using_zstd) {
zstd_ddict_ = ZSTD_createDDict_byReference(slice_.data(), slice_.size());
assert(zstd_ddict_ != nullptr);
}
#endif // ROCKSDB_ZSTD_DDICT
}
UncompressionDict(UncompressionDict&& rhs)
: dict_(std::move(rhs.dict_)),
allocation_(std::move(rhs.allocation_)),
slice_(std::move(rhs.slice_))
#ifdef ROCKSDB_ZSTD_DDICT
,
zstd_ddict_(rhs.zstd_ddict_)
#endif
{
#ifdef ROCKSDB_ZSTD_DDICT
rhs.zstd_ddict_ = nullptr;
#endif
}
~UncompressionDict() {
#ifdef ROCKSDB_ZSTD_DDICT
size_t res = 0;
if (zstd_ddict_ != nullptr) {
res = ZSTD_freeDDict(zstd_ddict_);
}
assert(res == 0); // Last I checked they can't fail
(void)res; // prevent unused var warning
#endif // ROCKSDB_ZSTD_DDICT
}
UncompressionDict& operator=(UncompressionDict&& rhs) {
if (this == &rhs) {
return *this;
}
dict_ = std::move(rhs.dict_);
allocation_ = std::move(rhs.allocation_);
slice_ = std::move(rhs.slice_);
#ifdef ROCKSDB_ZSTD_DDICT
zstd_ddict_ = rhs.zstd_ddict_;
rhs.zstd_ddict_ = nullptr;
#endif
return *this;
}
// The object is self-contained if the string constructor is used, or the
// Slice constructor is invoked with a non-null allocation. Otherwise, it
// is the caller's responsibility to ensure that the underlying storage
// outlives this object.
bool own_bytes() const { return !dict_.empty() || allocation_; }
const Slice& GetRawDict() const { return slice_; }
// For TypedCacheInterface
const Slice& ContentSlice() const { return slice_; }
static constexpr CacheEntryRole kCacheEntryRole = CacheEntryRole::kOtherBlock;
static constexpr BlockType kBlockType = BlockType::kCompressionDictionary;
#ifdef ROCKSDB_ZSTD_DDICT
const ZSTD_DDict* GetDigestedZstdDDict() const { return zstd_ddict_; }
#endif // ROCKSDB_ZSTD_DDICT
static const UncompressionDict& GetEmptyDict() {
static UncompressionDict empty_dict{};
return empty_dict;
}
size_t ApproximateMemoryUsage() const {
size_t usage = sizeof(struct UncompressionDict);
usage += dict_.size();
if (allocation_) {
auto allocator = allocation_.get_deleter().allocator;
if (allocator) {
usage += allocator->UsableSize(allocation_.get(), slice_.size());
} else {
usage += slice_.size();
}
}
#ifdef ROCKSDB_ZSTD_DDICT
usage += ZSTD_sizeof_DDict(zstd_ddict_);
#endif // ROCKSDB_ZSTD_DDICT
return usage;
}
UncompressionDict() = default;
// Disable copy
UncompressionDict(const CompressionDict&) = delete;
UncompressionDict& operator=(const CompressionDict&) = delete;
};
class CompressionContext {
private:
#if defined(ZSTD) && (ZSTD_VERSION_NUMBER >= 500)
ZSTD_CCtx* zstd_ctx_ = nullptr;
ZSTD_CCtx* CreateZSTDContext() {
#ifdef ROCKSDB_ZSTD_CUSTOM_MEM
return ZSTD_createCCtx_advanced(port::GetJeZstdAllocationOverrides());
#else // ROCKSDB_ZSTD_CUSTOM_MEM
return ZSTD_createCCtx();
#endif // ROCKSDB_ZSTD_CUSTOM_MEM
}
void CreateNativeContext(CompressionType type, int level, bool checksum) {
if (type == kZSTD || type == kZSTDNotFinalCompression) {
zstd_ctx_ = CreateZSTDContext();
#ifdef ZSTD_ADVANCED
if (level == CompressionOptions::kDefaultCompressionLevel) {
// 3 is the value of ZSTD_CLEVEL_DEFAULT (not exposed publicly), see
// https://github.com/facebook/zstd/issues/1148
level = 3;
}
size_t err =
ZSTD_CCtx_setParameter(zstd_ctx_, ZSTD_c_compressionLevel, level);
if (ZSTD_isError(err)) {
assert(false);
ZSTD_freeCCtx(zstd_ctx_);
zstd_ctx_ = CreateZSTDContext();
}
if (checksum) {
err = ZSTD_CCtx_setParameter(zstd_ctx_, ZSTD_c_checksumFlag, 1);
if (ZSTD_isError(err)) {
assert(false);
ZSTD_freeCCtx(zstd_ctx_);
zstd_ctx_ = CreateZSTDContext();
}
}
#else
(void)level;
(void)checksum;
#endif
}
}
void DestroyNativeContext() {
if (zstd_ctx_ != nullptr) {
ZSTD_freeCCtx(zstd_ctx_);
}
}
public:
// callable inside ZSTD_Compress
ZSTD_CCtx* ZSTDPreallocCtx() const {
assert(zstd_ctx_ != nullptr);
return zstd_ctx_;
}
#else // ZSTD && (ZSTD_VERSION_NUMBER >= 500)
private:
void CreateNativeContext(CompressionType /* type */, int /* level */,
bool /* checksum */) {}
void DestroyNativeContext() {}
#endif // ZSTD && (ZSTD_VERSION_NUMBER >= 500)
public:
explicit CompressionContext(CompressionType type,
const CompressionOptions& options) {
CreateNativeContext(type, options.level, options.checksum);
}
~CompressionContext() { DestroyNativeContext(); }
CompressionContext(const CompressionContext&) = delete;
CompressionContext& operator=(const CompressionContext&) = delete;
};
class CompressionInfo {
const CompressionOptions& opts_;
const CompressionContext& context_;
const CompressionDict& dict_;
const CompressionType type_;
const uint64_t sample_for_compression_;
public:
CompressionInfo(const CompressionOptions& _opts,
const CompressionContext& _context,
const CompressionDict& _dict, CompressionType _type,
uint64_t _sample_for_compression)
: opts_(_opts),
context_(_context),
dict_(_dict),
type_(_type),
sample_for_compression_(_sample_for_compression) {}
const CompressionOptions& options() const { return opts_; }
const CompressionContext& context() const { return context_; }
const CompressionDict& dict() const { return dict_; }
CompressionType type() const { return type_; }
uint64_t SampleForCompression() const { return sample_for_compression_; }
};
class UncompressionContext {
private:
CompressionContextCache* ctx_cache_ = nullptr;
ZSTDUncompressCachedData uncomp_cached_data_;
public:
explicit UncompressionContext(CompressionType type) {
if (type == kZSTD || type == kZSTDNotFinalCompression) {
ctx_cache_ = CompressionContextCache::Instance();
uncomp_cached_data_ = ctx_cache_->GetCachedZSTDUncompressData();
}
}
~UncompressionContext() {
if (uncomp_cached_data_.GetCacheIndex() != -1) {
assert(ctx_cache_ != nullptr);
ctx_cache_->ReturnCachedZSTDUncompressData(
uncomp_cached_data_.GetCacheIndex());
}
}
UncompressionContext(const UncompressionContext&) = delete;
UncompressionContext& operator=(const UncompressionContext&) = delete;
ZSTDUncompressCachedData::ZSTDNativeContext GetZSTDContext() const {
return uncomp_cached_data_.Get();
}
};
class UncompressionInfo {
const UncompressionContext& context_;
const UncompressionDict& dict_;
const CompressionType type_;
public:
UncompressionInfo(const UncompressionContext& _context,
const UncompressionDict& _dict, CompressionType _type)
: context_(_context), dict_(_dict), type_(_type) {}
const UncompressionContext& context() const { return context_; }
const UncompressionDict& dict() const { return dict_; }
CompressionType type() const { return type_; }
};
inline bool Snappy_Supported() {
#ifdef SNAPPY
return true;
#else
return false;
#endif
}
inline bool Zlib_Supported() {
#ifdef ZLIB
return true;
#else
return false;
#endif
}
inline bool BZip2_Supported() {
#ifdef BZIP2
return true;
#else
return false;
#endif
}
inline bool LZ4_Supported() {
#ifdef LZ4
return true;
#else
return false;
#endif
}
inline bool XPRESS_Supported() {
#ifdef XPRESS
return true;
#else
return false;
#endif
}
inline bool ZSTD_Supported() {
#ifdef ZSTD
// ZSTD format is finalized since version 0.8.0.
return (ZSTD_versionNumber() >= 800);
#else
return false;
#endif
}
inline bool ZSTDNotFinal_Supported() {
#ifdef ZSTD
return true;
#else
return false;
#endif
}
inline bool ZSTD_Streaming_Supported() {
#if defined(ZSTD_ADVANCED)
return true;
#else
return false;
#endif
}
inline bool StreamingCompressionTypeSupported(
CompressionType compression_type) {
switch (compression_type) {
case kNoCompression:
return true;
case kZSTD:
return ZSTD_Streaming_Supported();
default:
return false;
}
}
inline bool CompressionTypeSupported(CompressionType compression_type) {
switch (compression_type) {
case kNoCompression:
return true;
case kSnappyCompression:
return Snappy_Supported();
case kZlibCompression:
return Zlib_Supported();
case kBZip2Compression:
return BZip2_Supported();
case kLZ4Compression:
return LZ4_Supported();
case kLZ4HCCompression:
return LZ4_Supported();
case kXpressCompression:
return XPRESS_Supported();
case kZSTDNotFinalCompression:
return ZSTDNotFinal_Supported();
case kZSTD:
return ZSTD_Supported();
default:
assert(false);
return false;
}
}
inline bool DictCompressionTypeSupported(CompressionType compression_type) {
switch (compression_type) {
case kNoCompression:
return false;
case kSnappyCompression:
return false;
case kZlibCompression:
return Zlib_Supported();
case kBZip2Compression:
return false;
case kLZ4Compression:
case kLZ4HCCompression:
#if LZ4_VERSION_NUMBER >= 10400 // r124+
return LZ4_Supported();
#else
return false;
#endif
case kXpressCompression:
return false;
case kZSTDNotFinalCompression:
#if ZSTD_VERSION_NUMBER >= 500 // v0.5.0+
return ZSTDNotFinal_Supported();
#else
return false;
#endif
case kZSTD:
#if ZSTD_VERSION_NUMBER >= 500 // v0.5.0+
return ZSTD_Supported();
#else
return false;
#endif
default:
assert(false);
return false;
}
}
inline std::string CompressionTypeToString(CompressionType compression_type) {
switch (compression_type) {
case kNoCompression:
return "NoCompression";
case kSnappyCompression:
return "Snappy";
case kZlibCompression:
return "Zlib";
case kBZip2Compression:
return "BZip2";
case kLZ4Compression:
return "LZ4";
case kLZ4HCCompression:
return "LZ4HC";
case kXpressCompression:
return "Xpress";
case kZSTD:
return "ZSTD";
case kZSTDNotFinalCompression:
return "ZSTDNotFinal";
case kDisableCompressionOption:
return "DisableOption";
default:
assert(false);
return "";
}
}
inline std::string CompressionOptionsToString(
CompressionOptions& compression_options) {
std::string result;
result.reserve(512);
result.append("window_bits=")
.append(std::to_string(compression_options.window_bits))
.append("; ");
result.append("level=")
.append(std::to_string(compression_options.level))
.append("; ");
result.append("strategy=")
.append(std::to_string(compression_options.strategy))
.append("; ");
result.append("max_dict_bytes=")
.append(std::to_string(compression_options.max_dict_bytes))
.append("; ");
result.append("zstd_max_train_bytes=")
.append(std::to_string(compression_options.zstd_max_train_bytes))
.append("; ");
result.append("enabled=")
.append(std::to_string(compression_options.enabled))
.append("; ");
result.append("max_dict_buffer_bytes=")
.append(std::to_string(compression_options.max_dict_buffer_bytes))
.append("; ");
result.append("use_zstd_dict_trainer=")
.append(std::to_string(compression_options.use_zstd_dict_trainer))
.append("; ");
return result;
}
// compress_format_version can have two values:
// 1 -- decompressed sizes for BZip2 and Zlib are not included in the compressed
// block. Also, decompressed sizes for LZ4 are encoded in platform-dependent
// way.
// 2 -- Zlib, BZip2 and LZ4 encode decompressed size as Varint32 just before the
// start of compressed block. Snappy format is the same as version 1.
inline bool Snappy_Compress(const CompressionInfo& /*info*/, const char* input,
size_t length, ::std::string* output) {
#ifdef SNAPPY
output->resize(snappy::MaxCompressedLength(length));
size_t outlen;
snappy::RawCompress(input, length, &(*output)[0], &outlen);
output->resize(outlen);
return true;
#else
(void)input;
(void)length;
(void)output;
return false;
#endif
}
inline CacheAllocationPtr Snappy_Uncompress(
const char* input, size_t length, size_t* uncompressed_size,
MemoryAllocator* allocator = nullptr) {
#ifdef SNAPPY
size_t uncompressed_length = 0;
if (!snappy::GetUncompressedLength(input, length, &uncompressed_length)) {
return nullptr;
}
CacheAllocationPtr output = AllocateBlock(uncompressed_length, allocator);
if (!snappy::RawUncompress(input, length, output.get())) {
return nullptr;
}
*uncompressed_size = uncompressed_length;
return output;
#else
(void)input;
(void)length;
(void)uncompressed_size;
(void)allocator;
return nullptr;
#endif
}
namespace compression {
// returns size
inline size_t PutDecompressedSizeInfo(std::string* output, uint32_t length) {
PutVarint32(output, length);
return output->size();
}
inline bool GetDecompressedSizeInfo(const char** input_data,
size_t* input_length,
uint32_t* output_len) {
auto new_input_data =
GetVarint32Ptr(*input_data, *input_data + *input_length, output_len);
if (new_input_data == nullptr) {
return false;
}
*input_length -= (new_input_data - *input_data);
*input_data = new_input_data;
return true;
}
} // namespace compression
// compress_format_version == 1 -- decompressed size is not included in the
// block header
// compress_format_version == 2 -- decompressed size is included in the block
// header in varint32 format
// @param compression_dict Data for presetting the compression library's
// dictionary.
inline bool Zlib_Compress(const CompressionInfo& info,
uint32_t compress_format_version, const char* input,
size_t length, ::std::string* output) {
#ifdef ZLIB
if (length > std::numeric_limits<uint32_t>::max()) {
// Can't compress more than 4GB
return false;
}
size_t output_header_len = 0;
if (compress_format_version == 2) {
output_header_len = compression::PutDecompressedSizeInfo(
output, static_cast<uint32_t>(length));
}
// The memLevel parameter specifies how much memory should be allocated for
// the internal compression state.
// memLevel=1 uses minimum memory but is slow and reduces compression ratio.
// memLevel=9 uses maximum memory for optimal speed.
// The default value is 8. See zconf.h for more details.
static const int memLevel = 8;
int level;
if (info.options().level == CompressionOptions::kDefaultCompressionLevel) {
level = Z_DEFAULT_COMPRESSION;
} else {
level = info.options().level;
}
z_stream _stream;
memset(&_stream, 0, sizeof(z_stream));
int st = deflateInit2(&_stream, level, Z_DEFLATED, info.options().window_bits,
memLevel, info.options().strategy);
if (st != Z_OK) {
return false;
}
Slice compression_dict = info.dict().GetRawDict();
if (compression_dict.size()) {
// Initialize the compression library's dictionary
st = deflateSetDictionary(
&_stream, reinterpret_cast<const Bytef*>(compression_dict.data()),
static_cast<unsigned int>(compression_dict.size()));
if (st != Z_OK) {
deflateEnd(&_stream);
return false;
}
}
// Get an upper bound on the compressed size.
size_t upper_bound =
deflateBound(&_stream, static_cast<unsigned long>(length));
output->resize(output_header_len + upper_bound);
// Compress the input, and put compressed data in output.
_stream.next_in = (Bytef*)input;
_stream.avail_in = static_cast<unsigned int>(length);
// Initialize the output size.
_stream.avail_out = static_cast<unsigned int>(upper_bound);
_stream.next_out = reinterpret_cast<Bytef*>(&(*output)[output_header_len]);
bool compressed = false;
st = deflate(&_stream, Z_FINISH);
if (st == Z_STREAM_END) {
compressed = true;
output->resize(output->size() - _stream.avail_out);
}
// The only return value we really care about is Z_STREAM_END.
// Z_OK means insufficient output space. This means the compression is
// bigger than decompressed size. Just fail the compression in that case.
deflateEnd(&_stream);
return compressed;
#else
(void)info;
(void)compress_format_version;
(void)input;
(void)length;
(void)output;
return false;
#endif
}
// compress_format_version == 1 -- decompressed size is not included in the
// block header
// compress_format_version == 2 -- decompressed size is included in the block
// header in varint32 format
// @param compression_dict Data for presetting the compression library's
// dictionary.
inline CacheAllocationPtr Zlib_Uncompress(
const UncompressionInfo& info, const char* input_data, size_t input_length,
size_t* uncompressed_size, uint32_t compress_format_version,
MemoryAllocator* allocator = nullptr, int windowBits = -14) {
#ifdef ZLIB
uint32_t output_len = 0;
if (compress_format_version == 2) {
if (!compression::GetDecompressedSizeInfo(&input_data, &input_length,
&output_len)) {
return nullptr;
}
} else {
// Assume the decompressed data size will 5x of compressed size, but round
// to the page size
size_t proposed_output_len = ((input_length * 5) & (~(4096 - 1))) + 4096;
output_len = static_cast<uint32_t>(
std::min(proposed_output_len,
static_cast<size_t>(std::numeric_limits<uint32_t>::max())));
}
z_stream _stream;
memset(&_stream, 0, sizeof(z_stream));
// For raw inflate, the windowBits should be -8..-15.
// If windowBits is bigger than zero, it will use either zlib
// header or gzip header. Adding 32 to it will do automatic detection.
int st =
inflateInit2(&_stream, windowBits > 0 ? windowBits + 32 : windowBits);
if (st != Z_OK) {
return nullptr;
}
const Slice& compression_dict = info.dict().GetRawDict();
if (compression_dict.size()) {
// Initialize the compression library's dictionary
st = inflateSetDictionary(
&_stream, reinterpret_cast<const Bytef*>(compression_dict.data()),
static_cast<unsigned int>(compression_dict.size()));
if (st != Z_OK) {
return nullptr;
}
}
_stream.next_in = (Bytef*)input_data;
_stream.avail_in = static_cast<unsigned int>(input_length);
auto output = AllocateBlock(output_len, allocator);
_stream.next_out = (Bytef*)output.get();
_stream.avail_out = static_cast<unsigned int>(output_len);
bool done = false;
while (!done) {
st = inflate(&_stream, Z_SYNC_FLUSH);
switch (st) {
case Z_STREAM_END:
done = true;
break;
case Z_OK: {
// No output space. Increase the output space by 20%.
// We should never run out of output space if
// compress_format_version == 2
assert(compress_format_version != 2);
size_t old_sz = output_len;
uint32_t output_len_delta = output_len / 5;
output_len += output_len_delta < 10 ? 10 : output_len_delta;
auto tmp = AllocateBlock(output_len, allocator);
memcpy(tmp.get(), output.get(), old_sz);
output = std::move(tmp);
// Set more output.
_stream.next_out = (Bytef*)(output.get() + old_sz);
_stream.avail_out = static_cast<unsigned int>(output_len - old_sz);
break;
}
case Z_BUF_ERROR:
default:
inflateEnd(&_stream);
return nullptr;
}
}
// If we encoded decompressed block size, we should have no bytes left
assert(compress_format_version != 2 || _stream.avail_out == 0);
assert(output_len >= _stream.avail_out);
*uncompressed_size = output_len - _stream.avail_out;
inflateEnd(&_stream);
return output;
#else
(void)info;
(void)input_data;
(void)input_length;
(void)uncompressed_size;
(void)compress_format_version;
(void)allocator;
(void)windowBits;
return nullptr;
#endif
}
// compress_format_version == 1 -- decompressed size is not included in the
// block header
// compress_format_version == 2 -- decompressed size is included in the block
// header in varint32 format
inline bool BZip2_Compress(const CompressionInfo& /*info*/,
uint32_t compress_format_version, const char* input,
size_t length, ::std::string* output) {
#ifdef BZIP2
if (length > std::numeric_limits<uint32_t>::max()) {
// Can't compress more than 4GB
return false;
}
size_t output_header_len = 0;
if (compress_format_version == 2) {
output_header_len = compression::PutDecompressedSizeInfo(
output, static_cast<uint32_t>(length));
}
// Resize output to be the plain data length.
// This may not be big enough if the compression actually expands data.
output->resize(output_header_len + length);
bz_stream _stream;
memset(&_stream, 0, sizeof(bz_stream));
// Block size 1 is 100K.
// 0 is for silent.
// 30 is the default workFactor
int st = BZ2_bzCompressInit(&_stream, 1, 0, 30);
if (st != BZ_OK) {
return false;
}
// Compress the input, and put compressed data in output.
_stream.next_in = (char*)input;
_stream.avail_in = static_cast<unsigned int>(length);
// Initialize the output size.
_stream.avail_out = static_cast<unsigned int>(length);
_stream.next_out = reinterpret_cast<char*>(&(*output)[output_header_len]);
bool compressed = false;
st = BZ2_bzCompress(&_stream, BZ_FINISH);
if (st == BZ_STREAM_END) {
compressed = true;
output->resize(output->size() - _stream.avail_out);
}
// The only return value we really care about is BZ_STREAM_END.
// BZ_FINISH_OK means insufficient output space. This means the compression
// is bigger than decompressed size. Just fail the compression in that case.
BZ2_bzCompressEnd(&_stream);
return compressed;
#else
(void)compress_format_version;
(void)input;
(void)length;
(void)output;
return false;
#endif
}
// compress_format_version == 1 -- decompressed size is not included in the
// block header
// compress_format_version == 2 -- decompressed size is included in the block
// header in varint32 format
inline CacheAllocationPtr BZip2_Uncompress(
const char* input_data, size_t input_length, size_t* uncompressed_size,
uint32_t compress_format_version, MemoryAllocator* allocator = nullptr) {
#ifdef BZIP2
uint32_t output_len = 0;
if (compress_format_version == 2) {
if (!compression::GetDecompressedSizeInfo(&input_data, &input_length,
&output_len)) {
return nullptr;
}
} else {
// Assume the decompressed data size will 5x of compressed size, but round
// to the next page size
size_t proposed_output_len = ((input_length * 5) & (~(4096 - 1))) + 4096;
output_len = static_cast<uint32_t>(
std::min(proposed_output_len,
static_cast<size_t>(std::numeric_limits<uint32_t>::max())));
}
bz_stream _stream;
memset(&_stream, 0, sizeof(bz_stream));
int st = BZ2_bzDecompressInit(&_stream, 0, 0);
if (st != BZ_OK) {
return nullptr;
}
_stream.next_in = (char*)input_data;
_stream.avail_in = static_cast<unsigned int>(input_length);
auto output = AllocateBlock(output_len, allocator);
_stream.next_out = (char*)output.get();
_stream.avail_out = static_cast<unsigned int>(output_len);
bool done = false;
while (!done) {
st = BZ2_bzDecompress(&_stream);
switch (st) {
case BZ_STREAM_END:
done = true;
break;
case BZ_OK: {
// No output space. Increase the output space by 20%.
// We should never run out of output space if
// compress_format_version == 2
assert(compress_format_version != 2);
uint32_t old_sz = output_len;
output_len = output_len * 1.2;
auto tmp = AllocateBlock(output_len, allocator);
memcpy(tmp.get(), output.get(), old_sz);
output = std::move(tmp);
// Set more output.
_stream.next_out = (char*)(output.get() + old_sz);
_stream.avail_out = static_cast<unsigned int>(output_len - old_sz);
break;
}
default:
BZ2_bzDecompressEnd(&_stream);
return nullptr;
}
}
// If we encoded decompressed block size, we should have no bytes left
assert(compress_format_version != 2 || _stream.avail_out == 0);
assert(output_len >= _stream.avail_out);
*uncompressed_size = output_len - _stream.avail_out;
BZ2_bzDecompressEnd(&_stream);
return output;
#else
(void)input_data;
(void)input_length;
(void)uncompressed_size;
(void)compress_format_version;
(void)allocator;
return nullptr;
#endif
}
// compress_format_version == 1 -- decompressed size is included in the
// block header using memcpy, which makes database non-portable)
// compress_format_version == 2 -- decompressed size is included in the block
// header in varint32 format
// @param compression_dict Data for presetting the compression library's
// dictionary.
inline bool LZ4_Compress(const CompressionInfo& info,
uint32_t compress_format_version, const char* input,
size_t length, ::std::string* output) {
#ifdef LZ4
if (length > std::numeric_limits<uint32_t>::max()) {
// Can't compress more than 4GB
return false;
}
size_t output_header_len = 0;
if (compress_format_version == 2) {
// new encoding, using varint32 to store size information
output_header_len = compression::PutDecompressedSizeInfo(
output, static_cast<uint32_t>(length));
} else {
// legacy encoding, which is not really portable (depends on big/little
// endianness)
output_header_len = 8;
output->resize(output_header_len);
char* p = const_cast<char*>(output->c_str());
memcpy(p, &length, sizeof(length));
}
int compress_bound = LZ4_compressBound(static_cast<int>(length));
output->resize(static_cast<size_t>(output_header_len + compress_bound));
int outlen;
#if LZ4_VERSION_NUMBER >= 10400 // r124+
LZ4_stream_t* stream = LZ4_createStream();
Slice compression_dict = info.dict().GetRawDict();
if (compression_dict.size()) {
LZ4_loadDict(stream, compression_dict.data(),
static_cast<int>(compression_dict.size()));
}
#if LZ4_VERSION_NUMBER >= 10700 // r129+
int acceleration;
if (info.options().level < 0) {
acceleration = -info.options().level;
} else {
acceleration = 1;
}
outlen = LZ4_compress_fast_continue(
stream, input, &(*output)[output_header_len], static_cast<int>(length),
compress_bound, acceleration);
#else // up to r128
outlen = LZ4_compress_limitedOutput_continue(
stream, input, &(*output)[output_header_len], static_cast<int>(length),
compress_bound);
#endif
LZ4_freeStream(stream);
#else // up to r123
outlen = LZ4_compress_limitedOutput(input, &(*output)[output_header_len],
static_cast<int>(length), compress_bound);
#endif // LZ4_VERSION_NUMBER >= 10400
if (outlen == 0) {
return false;
}
output->resize(static_cast<size_t>(output_header_len + outlen));
return true;
#else // LZ4
(void)info;
(void)compress_format_version;
(void)input;
(void)length;
(void)output;
return false;
#endif
}
// compress_format_version == 1 -- decompressed size is included in the
// block header using memcpy, which makes database non-portable)
// compress_format_version == 2 -- decompressed size is included in the block
// header in varint32 format
// @param compression_dict Data for presetting the compression library's
// dictionary.
inline CacheAllocationPtr LZ4_Uncompress(const UncompressionInfo& info,
const char* input_data,
size_t input_length,
size_t* uncompressed_size,
uint32_t compress_format_version,
MemoryAllocator* allocator = nullptr) {
#ifdef LZ4
uint32_t output_len = 0;
if (compress_format_version == 2) {
// new encoding, using varint32 to store size information
if (!compression::GetDecompressedSizeInfo(&input_data, &input_length,
&output_len)) {
return nullptr;
}
} else {
// legacy encoding, which is not really portable (depends on big/little
// endianness)
if (input_length < 8) {
return nullptr;
}
if (port::kLittleEndian) {
memcpy(&output_len, input_data, sizeof(output_len));
} else {
memcpy(&output_len, input_data + 4, sizeof(output_len));
}
input_length -= 8;
input_data += 8;
}
auto output = AllocateBlock(output_len, allocator);
int decompress_bytes = 0;
#if LZ4_VERSION_NUMBER >= 10400 // r124+
LZ4_streamDecode_t* stream = LZ4_createStreamDecode();
const Slice& compression_dict = info.dict().GetRawDict();
if (compression_dict.size()) {
LZ4_setStreamDecode(stream, compression_dict.data(),
static_cast<int>(compression_dict.size()));
}
decompress_bytes = LZ4_decompress_safe_continue(
stream, input_data, output.get(), static_cast<int>(input_length),
static_cast<int>(output_len));
LZ4_freeStreamDecode(stream);
#else // up to r123
decompress_bytes = LZ4_decompress_safe(input_data, output.get(),
static_cast<int>(input_length),
static_cast<int>(output_len));
#endif // LZ4_VERSION_NUMBER >= 10400
if (decompress_bytes < 0) {
return nullptr;
}
assert(decompress_bytes == static_cast<int>(output_len));
*uncompressed_size = decompress_bytes;
return output;
#else // LZ4
(void)info;
(void)input_data;
(void)input_length;
(void)uncompressed_size;
(void)compress_format_version;
(void)allocator;
return nullptr;
#endif
}
// compress_format_version == 1 -- decompressed size is included in the
// block header using memcpy, which makes database non-portable)
// compress_format_version == 2 -- decompressed size is included in the block
// header in varint32 format
// @param compression_dict Data for presetting the compression library's
// dictionary.
inline bool LZ4HC_Compress(const CompressionInfo& info,
uint32_t compress_format_version, const char* input,
size_t length, ::std::string* output) {
#ifdef LZ4
if (length > std::numeric_limits<uint32_t>::max()) {
// Can't compress more than 4GB
return false;
}
size_t output_header_len = 0;
if (compress_format_version == 2) {
// new encoding, using varint32 to store size information
output_header_len = compression::PutDecompressedSizeInfo(
output, static_cast<uint32_t>(length));
} else {
// legacy encoding, which is not really portable (depends on big/little
// endianness)
output_header_len = 8;
output->resize(output_header_len);
char* p = const_cast<char*>(output->c_str());
memcpy(p, &length, sizeof(length));
}
int compress_bound = LZ4_compressBound(static_cast<int>(length));
output->resize(static_cast<size_t>(output_header_len + compress_bound));
int outlen;
int level;
if (info.options().level == CompressionOptions::kDefaultCompressionLevel) {
level = 0; // lz4hc.h says any value < 1 will be sanitized to default
} else {
level = info.options().level;
}
#if LZ4_VERSION_NUMBER >= 10400 // r124+
LZ4_streamHC_t* stream = LZ4_createStreamHC();
LZ4_resetStreamHC(stream, level);
Slice compression_dict = info.dict().GetRawDict();
const char* compression_dict_data =
compression_dict.size() > 0 ? compression_dict.data() : nullptr;
size_t compression_dict_size = compression_dict.size();
if (compression_dict_data != nullptr) {
LZ4_loadDictHC(stream, compression_dict_data,
static_cast<int>(compression_dict_size));
}
#if LZ4_VERSION_NUMBER >= 10700 // r129+
outlen =
LZ4_compress_HC_continue(stream, input, &(*output)[output_header_len],
static_cast<int>(length), compress_bound);
#else // r124-r128
outlen = LZ4_compressHC_limitedOutput_continue(
stream, input, &(*output)[output_header_len], static_cast<int>(length),
compress_bound);
#endif // LZ4_VERSION_NUMBER >= 10700
LZ4_freeStreamHC(stream);
#elif LZ4_VERSION_MAJOR // r113-r123
outlen = LZ4_compressHC2_limitedOutput(input, &(*output)[output_header_len],
static_cast<int>(length),
compress_bound, level);
#else // up to r112
outlen =
LZ4_compressHC_limitedOutput(input, &(*output)[output_header_len],
static_cast<int>(length), compress_bound);
#endif // LZ4_VERSION_NUMBER >= 10400
if (outlen == 0) {
return false;
}
output->resize(static_cast<size_t>(output_header_len + outlen));
return true;
#else // LZ4
(void)info;
(void)compress_format_version;
(void)input;
(void)length;
(void)output;
return false;
#endif
}
#ifdef XPRESS
inline bool XPRESS_Compress(const char* input, size_t length,
std::string* output) {
return port::xpress::Compress(input, length, output);
}
#else
inline bool XPRESS_Compress(const char* /*input*/, size_t /*length*/,
std::string* /*output*/) {
return false;
}
#endif
#ifdef XPRESS
inline char* XPRESS_Uncompress(const char* input_data, size_t input_length,
size_t* uncompressed_size) {
return port::xpress::Decompress(input_data, input_length, uncompressed_size);
}
#else
inline char* XPRESS_Uncompress(const char* /*input_data*/,
size_t /*input_length*/,
size_t* /*uncompressed_size*/) {
return nullptr;
}
#endif
inline bool ZSTD_Compress(const CompressionInfo& info, const char* input,
size_t length, ::std::string* output) {
#ifdef ZSTD
if (length > std::numeric_limits<uint32_t>::max()) {
// Can't compress more than 4GB
return false;
}
size_t output_header_len = compression::PutDecompressedSizeInfo(
output, static_cast<uint32_t>(length));
size_t compressBound = ZSTD_compressBound(length);
output->resize(static_cast<size_t>(output_header_len + compressBound));
size_t outlen = 0;
#if ZSTD_VERSION_NUMBER >= 500 // v0.5.0+
ZSTD_CCtx* context = info.context().ZSTDPreallocCtx();
assert(context != nullptr);
#ifdef ZSTD_ADVANCED
if (info.dict().GetDigestedZstdCDict() != nullptr) {
ZSTD_CCtx_refCDict(context, info.dict().GetDigestedZstdCDict());
} else {
ZSTD_CCtx_loadDictionary(context, info.dict().GetRawDict().data(),
info.dict().GetRawDict().size());
}
// Compression level is set in `contex` during CreateNativeContext()
outlen = ZSTD_compress2(context, &(*output)[output_header_len], compressBound,
input, length);
#else // ZSTD_ADVANCED
#if ZSTD_VERSION_NUMBER >= 700 // v0.7.0+
if (info.dict().GetDigestedZstdCDict() != nullptr) {
outlen = ZSTD_compress_usingCDict(context, &(*output)[output_header_len],
compressBound, input, length,
info.dict().GetDigestedZstdCDict());
}
#endif // ZSTD_VERSION_NUMBER >= 700
// TODO (cbi): error handling for compression.
if (outlen == 0) {
int level;
if (info.options().level == CompressionOptions::kDefaultCompressionLevel) {
// 3 is the value of ZSTD_CLEVEL_DEFAULT (not exposed publicly), see
// https://github.com/facebook/zstd/issues/1148
level = 3;
} else {
level = info.options().level;
}
outlen = ZSTD_compress_usingDict(context, &(*output)[output_header_len],
compressBound, input, length,
info.dict().GetRawDict().data(),
info.dict().GetRawDict().size(), level);
}
#endif // ZSTD_ADVANCED
#else // up to v0.4.x
outlen = ZSTD_compress(&(*output)[output_header_len], compressBound, input,
length, level);
#endif // ZSTD_VERSION_NUMBER >= 500
if (outlen == 0) {
return false;
}
output->resize(output_header_len + outlen);
return true;
#else // ZSTD
(void)info;
(void)input;
(void)length;
(void)output;
return false;
#endif
}
// @param compression_dict Data for presetting the compression library's
// dictionary.
// @param error_message If not null, will be set if decompression fails.
//
// Returns nullptr if decompression fails.
inline CacheAllocationPtr ZSTD_Uncompress(
const UncompressionInfo& info, const char* input_data, size_t input_length,
size_t* uncompressed_size, MemoryAllocator* allocator = nullptr,
const char** error_message = nullptr) {
#ifdef ZSTD
static const char* const kErrorDecodeOutputSize =
"Cannot decode output size.";
static const char* const kErrorOutputLenMismatch =
"Decompressed size does not match header.";
uint32_t output_len = 0;
if (!compression::GetDecompressedSizeInfo(&input_data, &input_length,
&output_len)) {
if (error_message) {
*error_message = kErrorDecodeOutputSize;
}
return nullptr;
}
CacheAllocationPtr output = AllocateBlock(output_len, allocator);
size_t actual_output_length = 0;
#if ZSTD_VERSION_NUMBER >= 500 // v0.5.0+
ZSTD_DCtx* context = info.context().GetZSTDContext();
assert(context != nullptr);
#ifdef ROCKSDB_ZSTD_DDICT
if (info.dict().GetDigestedZstdDDict() != nullptr) {
actual_output_length = ZSTD_decompress_usingDDict(
context, output.get(), output_len, input_data, input_length,
info.dict().GetDigestedZstdDDict());
} else {
#endif // ROCKSDB_ZSTD_DDICT
actual_output_length = ZSTD_decompress_usingDict(
context, output.get(), output_len, input_data, input_length,
info.dict().GetRawDict().data(), info.dict().GetRawDict().size());
#ifdef ROCKSDB_ZSTD_DDICT
}
#endif // ROCKSDB_ZSTD_DDICT
#else // up to v0.4.x
(void)info;
actual_output_length =
ZSTD_decompress(output.get(), output_len, input_data, input_length);
#endif // ZSTD_VERSION_NUMBER >= 500
if (ZSTD_isError(actual_output_length)) {
if (error_message) {
*error_message = ZSTD_getErrorName(actual_output_length);
}
return nullptr;
} else if (actual_output_length != output_len) {
if (error_message) {
*error_message = kErrorOutputLenMismatch;
}
return nullptr;
}
*uncompressed_size = actual_output_length;
return output;
#else // ZSTD
(void)info;
(void)input_data;
(void)input_length;
(void)uncompressed_size;
(void)allocator;
(void)error_message;
return nullptr;
#endif
}
inline bool ZSTD_TrainDictionarySupported() {
#ifdef ZSTD
// Dictionary trainer is available since v0.6.1 for static linking, but not
// available for dynamic linking until v1.1.3. For now we enable the feature
// in v1.1.3+ only.
return (ZSTD_versionNumber() >= 10103);
#else
return false;
#endif
}
inline std::string ZSTD_TrainDictionary(const std::string& samples,
const std::vector<size_t>& sample_lens,
size_t max_dict_bytes) {
// Dictionary trainer is available since v0.6.1 for static linking, but not
// available for dynamic linking until v1.1.3. For now we enable the feature
// in v1.1.3+ only.
#if ZSTD_VERSION_NUMBER >= 10103 // v1.1.3+
assert(samples.empty() == sample_lens.empty());
if (samples.empty()) {
return "";
}
std::string dict_data(max_dict_bytes, '\0');
size_t dict_len = ZDICT_trainFromBuffer(
&dict_data[0], max_dict_bytes, &samples[0], &sample_lens[0],
static_cast<unsigned>(sample_lens.size()));
if (ZDICT_isError(dict_len)) {
return "";
}
assert(dict_len <= max_dict_bytes);
dict_data.resize(dict_len);
return dict_data;
#else // up to v1.1.2
assert(false);
(void)samples;
(void)sample_lens;
(void)max_dict_bytes;
return "";
#endif // ZSTD_VERSION_NUMBER >= 10103
}
inline std::string ZSTD_TrainDictionary(const std::string& samples,
size_t sample_len_shift,
size_t max_dict_bytes) {
// Dictionary trainer is available since v0.6.1, but ZSTD was marked stable
// only since v0.8.0. For now we enable the feature in stable versions only.
#if ZSTD_VERSION_NUMBER >= 10103 // v1.1.3+
// skips potential partial sample at the end of "samples"
size_t num_samples = samples.size() >> sample_len_shift;
std::vector<size_t> sample_lens(num_samples, size_t(1) << sample_len_shift);
return ZSTD_TrainDictionary(samples, sample_lens, max_dict_bytes);
#else // up to v1.1.2
assert(false);
(void)samples;
(void)sample_len_shift;
(void)max_dict_bytes;
return "";
#endif // ZSTD_VERSION_NUMBER >= 10103
}
inline bool ZSTD_FinalizeDictionarySupported() {
#ifdef ZSTD
// ZDICT_finalizeDictionary API is stable since v1.4.5
return (ZSTD_versionNumber() >= 10405);
#else
return false;
#endif
}
inline std::string ZSTD_FinalizeDictionary(
const std::string& samples, const std::vector<size_t>& sample_lens,
size_t max_dict_bytes, int level) {
// ZDICT_finalizeDictionary is stable since version v1.4.5
#if ZSTD_VERSION_NUMBER >= 10405 // v1.4.5+
assert(samples.empty() == sample_lens.empty());
if (samples.empty()) {
return "";
}
if (level == CompressionOptions::kDefaultCompressionLevel) {
// 3 is the value of ZSTD_CLEVEL_DEFAULT (not exposed publicly), see
// https://github.com/facebook/zstd/issues/1148
level = 3;
}
std::string dict_data(max_dict_bytes, '\0');
size_t dict_len = ZDICT_finalizeDictionary(
dict_data.data(), max_dict_bytes, samples.data(),
std::min(static_cast<size_t>(samples.size()), max_dict_bytes),
samples.data(), sample_lens.data(),
static_cast<unsigned>(sample_lens.size()),
{level, 0 /* notificationLevel */, 0 /* dictID */});
if (ZDICT_isError(dict_len)) {
return "";
} else {
assert(dict_len <= max_dict_bytes);
dict_data.resize(dict_len);
return dict_data;
}
#else // up to v1.4.4
assert(false);
(void)samples;
(void)sample_lens;
(void)max_dict_bytes;
(void)level;
return "";
#endif // ZSTD_VERSION_NUMBER >= 10405
}
inline bool CompressData(const Slice& raw,
const CompressionInfo& compression_info,
uint32_t compress_format_version,
std::string* compressed_output) {
bool ret = false;
// Will return compressed block contents if (1) the compression method is
// supported in this platform and (2) the compression rate is "good enough".
switch (compression_info.type()) {
case kSnappyCompression:
ret = Snappy_Compress(compression_info, raw.data(), raw.size(),
compressed_output);
break;
case kZlibCompression:
ret = Zlib_Compress(compression_info, compress_format_version, raw.data(),
raw.size(), compressed_output);
break;
case kBZip2Compression:
ret = BZip2_Compress(compression_info, compress_format_version,
raw.data(), raw.size(), compressed_output);
break;
case kLZ4Compression:
ret = LZ4_Compress(compression_info, compress_format_version, raw.data(),
raw.size(), compressed_output);
break;
case kLZ4HCCompression:
ret = LZ4HC_Compress(compression_info, compress_format_version,
raw.data(), raw.size(), compressed_output);
break;
case kXpressCompression:
ret = XPRESS_Compress(raw.data(), raw.size(), compressed_output);
break;
case kZSTD:
case kZSTDNotFinalCompression:
ret = ZSTD_Compress(compression_info, raw.data(), raw.size(),
compressed_output);
break;
default:
// Do not recognize this compression type
break;
}
TEST_SYNC_POINT_CALLBACK("CompressData:TamperWithReturnValue",
static_cast<void*>(&ret));
return ret;
}
inline CacheAllocationPtr UncompressData(
const UncompressionInfo& uncompression_info, const char* data, size_t n,
size_t* uncompressed_size, uint32_t compress_format_version,
MemoryAllocator* allocator = nullptr,
const char** error_message = nullptr) {
switch (uncompression_info.type()) {
case kSnappyCompression:
return Snappy_Uncompress(data, n, uncompressed_size, allocator);
case kZlibCompression:
return Zlib_Uncompress(uncompression_info, data, n, uncompressed_size,
compress_format_version, allocator);
case kBZip2Compression:
return BZip2_Uncompress(data, n, uncompressed_size,
compress_format_version, allocator);
case kLZ4Compression:
case kLZ4HCCompression:
return LZ4_Uncompress(uncompression_info, data, n, uncompressed_size,
compress_format_version, allocator);
case kXpressCompression:
// XPRESS allocates memory internally, thus no support for custom
// allocator.
return CacheAllocationPtr(XPRESS_Uncompress(data, n, uncompressed_size));
case kZSTD:
case kZSTDNotFinalCompression:
// TODO(cbi): error message handling for other compression algorithms.
return ZSTD_Uncompress(uncompression_info, data, n, uncompressed_size,
allocator, error_message);
default:
return CacheAllocationPtr();
}
}
// Records the compression type for subsequent WAL records.
class CompressionTypeRecord {
public:
explicit CompressionTypeRecord(CompressionType compression_type)
: compression_type_(compression_type) {}
CompressionType GetCompressionType() const { return compression_type_; }
inline void EncodeTo(std::string* dst) const {
assert(dst != nullptr);
PutFixed32(dst, compression_type_);
}
inline Status DecodeFrom(Slice* src) {
constexpr char class_name[] = "CompressionTypeRecord";
uint32_t val;
if (!GetFixed32(src, &val)) {
return Status::Corruption(class_name,
"Error decoding WAL compression type");
}
CompressionType compression_type = static_cast<CompressionType>(val);
if (!StreamingCompressionTypeSupported(compression_type)) {
return Status::Corruption(class_name,
"WAL compression type not supported");
}
compression_type_ = compression_type;
return Status::OK();
}
inline std::string DebugString() const {
return "compression_type: " + CompressionTypeToString(compression_type_);
}
private:
CompressionType compression_type_;
};
// Base class to implement compression for a stream of buffers.
// Instantiate an implementation of the class using Create() with the
// compression type and use Compress() repeatedly.
// The output buffer needs to be at least max_output_len.
// Call Reset() in between frame boundaries or in case of an error.
// NOTE: This class is not thread safe.
class StreamingCompress {
public:
StreamingCompress(CompressionType compression_type,
const CompressionOptions& opts,
uint32_t compress_format_version, size_t max_output_len)
: compression_type_(compression_type),
opts_(opts),
compress_format_version_(compress_format_version),
max_output_len_(max_output_len) {}
virtual ~StreamingCompress() = default;
// compress should be called repeatedly with the same input till the method
// returns 0
// Parameters:
// input - buffer to compress
// input_size - size of input buffer
// output - compressed buffer allocated by caller, should be at least
// max_output_len
// output_size - size of the output buffer
// Returns -1 for errors, the remaining size of the input buffer that needs to
// be compressed
virtual int Compress(const char* input, size_t input_size, char* output,
size_t* output_pos) = 0;
// static method to create object of a class inherited from StreamingCompress
// based on the actual compression type.
static StreamingCompress* Create(CompressionType compression_type,
const CompressionOptions& opts,
uint32_t compress_format_version,
size_t max_output_len);
virtual void Reset() = 0;
protected:
const CompressionType compression_type_;
const CompressionOptions opts_;
const uint32_t compress_format_version_;
const size_t max_output_len_;
};
// Base class to uncompress a stream of compressed buffers.
// Instantiate an implementation of the class using Create() with the
// compression type and use Uncompress() repeatedly.
// The output buffer needs to be at least max_output_len.
// Call Reset() in between frame boundaries or in case of an error.
// NOTE: This class is not thread safe.
class StreamingUncompress {
public:
StreamingUncompress(CompressionType compression_type,
uint32_t compress_format_version, size_t max_output_len)
: compression_type_(compression_type),
compress_format_version_(compress_format_version),
max_output_len_(max_output_len) {}
virtual ~StreamingUncompress() = default;
// Uncompress can be called repeatedly to progressively process the same
// input buffer, or can be called with a new input buffer. When the input
// buffer is not fully consumed, the return value is > 0 or output_size
// == max_output_len. When calling uncompress to continue processing the
// same input buffer, the input argument should be nullptr.
// Parameters:
// input - buffer to uncompress
// input_size - size of input buffer
// output - uncompressed buffer allocated by caller, should be at least
// max_output_len
// output_size - size of the output buffer
// Returns -1 for errors, remaining input to be processed otherwise.
virtual int Uncompress(const char* input, size_t input_size, char* output,
size_t* output_pos) = 0;
static StreamingUncompress* Create(CompressionType compression_type,
uint32_t compress_format_version,
size_t max_output_len);
virtual void Reset() = 0;
protected:
CompressionType compression_type_;
uint32_t compress_format_version_;
size_t max_output_len_;
};
class ZSTDStreamingCompress final : public StreamingCompress {
public:
explicit ZSTDStreamingCompress(const CompressionOptions& opts,
uint32_t compress_format_version,
size_t max_output_len)
: StreamingCompress(kZSTD, opts, compress_format_version,
max_output_len) {
#ifdef ZSTD_ADVANCED
cctx_ = ZSTD_createCCtx();
// Each compressed frame will have a checksum
ZSTD_CCtx_setParameter(cctx_, ZSTD_c_checksumFlag, 1);
assert(cctx_ != nullptr);
input_buffer_ = {/*src=*/nullptr, /*size=*/0, /*pos=*/0};
#endif
}
~ZSTDStreamingCompress() override {
#ifdef ZSTD_ADVANCED
ZSTD_freeCCtx(cctx_);
#endif
}
int Compress(const char* input, size_t input_size, char* output,
size_t* output_pos) override;
void Reset() override;
#ifdef ZSTD_ADVANCED
ZSTD_CCtx* cctx_;
ZSTD_inBuffer input_buffer_;
#endif
};
class ZSTDStreamingUncompress final : public StreamingUncompress {
public:
explicit ZSTDStreamingUncompress(uint32_t compress_format_version,
size_t max_output_len)
: StreamingUncompress(kZSTD, compress_format_version, max_output_len) {
#ifdef ZSTD_ADVANCED
dctx_ = ZSTD_createDCtx();
assert(dctx_ != nullptr);
input_buffer_ = {/*src=*/nullptr, /*size=*/0, /*pos=*/0};
#endif
}
~ZSTDStreamingUncompress() override {
#ifdef ZSTD_ADVANCED
ZSTD_freeDCtx(dctx_);
#endif
}
int Uncompress(const char* input, size_t input_size, char* output,
size_t* output_size) override;
void Reset() override;
private:
#ifdef ZSTD_ADVANCED
ZSTD_DCtx* dctx_;
ZSTD_inBuffer input_buffer_;
#endif
};
} // namespace ROCKSDB_NAMESPACE
|