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
|
/*
* Copyright (C) 2005-2024 Apple Inc. All rights reserved.
* Copyright (C) 2008 David Levin <levin@chromium.org>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public License
* along with this library; see the file COPYING.LIB. If not, write to
* the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
* Boston, MA 02110-1301, USA.
*
*/
#pragma once
#include <atomic>
#include <iterator>
#include <mutex>
#include <string.h>
#include <type_traits>
#include <utility>
#include <wtf/AlignedStorage.h>
#include <wtf/Assertions.h>
#include <wtf/DebugHeap.h>
#include <wtf/FastMalloc.h>
#include <wtf/HashTraits.h>
#include <wtf/Lock.h>
#include <wtf/MathExtras.h>
#include <wtf/StdLibExtras.h>
#include <wtf/ValueCheck.h>
#include <wtf/WeakRandomNumber.h>
// Configuration of WTF::HashTable.
// - 75% load factor for small tables.
// - 50% load factor for large tables.
// - Use quadratic probing.
// - Always use power-of-two hashtable size, which is also important to make quadratic probing work.
#define DUMP_HASHTABLE_STATS 0
#define DUMP_HASHTABLE_STATS_PER_TABLE 0
#if DUMP_HASHTABLE_STATS_PER_TABLE
#include <wtf/DataLog.h>
#endif
WTF_ALLOW_UNSAFE_BUFFER_USAGE_BEGIN
namespace WTF {
DECLARE_ALLOCATOR_WITH_HEAP_IDENTIFIER(HashTable);
// Enables internal WTF consistency checks that are invoked automatically. Non-WTF callers can call checkTableConsistency() even if internal checks are disabled.
#define CHECK_HASHTABLE_CONSISTENCY 0
#ifdef NDEBUG
#define CHECK_HASHTABLE_ITERATORS 0
#define CHECK_HASHTABLE_USE_AFTER_DESTRUCTION 0
#else
#define CHECK_HASHTABLE_ITERATORS 1
#define CHECK_HASHTABLE_USE_AFTER_DESTRUCTION 1
#endif
#if DUMP_HASHTABLE_STATS
struct HashTableStats {
// The following variables are all atomically incremented when modified.
WTF_EXPORT_PRIVATE static std::atomic<unsigned> numAccesses;
WTF_EXPORT_PRIVATE static std::atomic<unsigned> numRehashes;
WTF_EXPORT_PRIVATE static std::atomic<unsigned> numRemoves;
WTF_EXPORT_PRIVATE static std::atomic<unsigned> numReinserts;
// The following variables are only modified in the recordCollisionAtCount method within a mutex.
WTF_EXPORT_PRIVATE static unsigned maxCollisions;
WTF_EXPORT_PRIVATE static unsigned numCollisions;
WTF_EXPORT_PRIVATE static unsigned collisionGraph[4096];
WTF_EXPORT_PRIVATE static void recordCollisionAtCount(unsigned count);
WTF_EXPORT_PRIVATE static void dumpStats();
};
#endif
template<typename HashTable, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
class HashTableIterator;
template<typename HashTable, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
class HashTableConstIterator;
template<typename HashTableType, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
void addIterator(const HashTableType*, HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits>*);
template<typename HashTableType, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
void removeIterator(HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits>*);
template<typename HashTableType>
void invalidateIterators(const HashTableType*);
#if !CHECK_HASHTABLE_ITERATORS
template<typename HashTableType, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
inline void addIterator(const HashTableType*, HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits>*) { }
template<typename HashTableType, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
inline void removeIterator(HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits>*) { }
template<typename HashTableType>
void invalidateIterators(const HashTableType*) { }
#endif
typedef enum { HashItemKnownGood } HashItemKnownGoodTag;
template<typename HashTable, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
class HashTableConstIterator {
WTF_MAKE_FAST_ALLOCATED;
public:
using iterator_category = std::forward_iterator_tag;
using value_type = Value;
using difference_type = ptrdiff_t;
using pointer = const value_type*;
using reference = const value_type&;
private:
using HashTableType = HashTable;
typedef HashTableIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits> iterator;
typedef HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits> const_iterator;
typedef Value ValueType;
typedef const ValueType& ReferenceType;
typedef const ValueType* PointerType;
friend HashTableType;
friend iterator;
void skipEmptyBuckets()
{
while (m_position != m_endPosition && HashTableType::isEmptyOrDeletedBucket(*m_position))
++m_position;
}
HashTableConstIterator(const HashTableType* table, PointerType position, PointerType endPosition)
: m_position(position), m_endPosition(endPosition)
{
addIterator(table, this);
skipEmptyBuckets();
}
HashTableConstIterator(const HashTableType* table, PointerType position, PointerType endPosition, HashItemKnownGoodTag)
: m_position(position), m_endPosition(endPosition)
{
addIterator(table, this);
}
public:
HashTableConstIterator()
{
addIterator(static_cast<const HashTableType*>(0), this);
}
// default copy, assignment and destructor are OK if CHECK_HASHTABLE_ITERATORS is 0
#if CHECK_HASHTABLE_ITERATORS
~HashTableConstIterator()
{
removeIterator(this);
}
HashTableConstIterator(const const_iterator& other)
: m_position(other.m_position), m_endPosition(other.m_endPosition)
{
addIterator(other.m_table, this);
}
const_iterator& operator=(const const_iterator& other)
{
m_position = other.m_position;
m_endPosition = other.m_endPosition;
removeIterator(this);
addIterator(other.m_table, this);
return *this;
}
#endif
PointerType get() const
{
checkValidity();
return m_position;
}
ReferenceType operator*() const { return *get(); }
PointerType operator->() const { return get(); }
const_iterator& operator++()
{
checkValidity();
ASSERT(m_position != m_endPosition);
++m_position;
skipEmptyBuckets();
return *this;
}
// postfix ++ intentionally omitted
// Comparison.
bool operator==(const const_iterator& other) const
{
checkValidity(other);
return m_position == other.m_position;
}
bool operator==(const iterator& other) const
{
return *this == static_cast<const_iterator>(other);
}
private:
void checkValidity() const
{
#if CHECK_HASHTABLE_ITERATORS
ASSERT(m_table);
#endif
}
#if CHECK_HASHTABLE_ITERATORS
void checkValidity(const const_iterator& other) const
{
ASSERT(m_table);
ASSERT_UNUSED(other, other.m_table);
ASSERT(m_table == other.m_table);
}
#else
void checkValidity(const const_iterator&) const { }
#endif
PointerType m_position { nullptr };
PointerType m_endPosition { nullptr };
#if CHECK_HASHTABLE_ITERATORS
public:
// Any modifications of the m_next or m_previous of an iterator that is in a linked list of a HashTable::m_iterator,
// should be guarded with m_table->m_mutex.
mutable const HashTableType* m_table;
mutable const_iterator* m_next;
mutable const_iterator* m_previous;
#endif
};
template<typename HashTable, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
class HashTableIterator {
WTF_MAKE_FAST_ALLOCATED;
public:
using iterator_category = std::forward_iterator_tag;
using value_type = Value;
using difference_type = ptrdiff_t;
using pointer = value_type*;
using reference = value_type&;
private:
using HashTableType = HashTable;
typedef HashTableIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits> iterator;
typedef HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits> const_iterator;
typedef Value ValueType;
typedef ValueType& ReferenceType;
typedef ValueType* PointerType;
friend HashTableType;
HashTableIterator(HashTableType* table, PointerType pos, PointerType end) : m_iterator(table, pos, end) { }
HashTableIterator(HashTableType* table, PointerType pos, PointerType end, HashItemKnownGoodTag tag) : m_iterator(table, pos, end, tag) { }
public:
HashTableIterator() { }
// default copy, assignment and destructor are OK
PointerType get() const { return const_cast<PointerType>(m_iterator.get()); }
ReferenceType operator*() const { return *get(); }
PointerType operator->() const { return get(); }
iterator& operator++() { ++m_iterator; return *this; }
// postfix ++ intentionally omitted
// Comparison.
bool operator==(const iterator& other) const { return m_iterator == other.m_iterator; }
bool operator==(const const_iterator& other) const { return m_iterator == other; }
operator const_iterator() const { return m_iterator; }
private:
const_iterator m_iterator;
};
template<typename ValueTraits, typename HashFunctions>
class IdentityHashTranslator {
public:
static unsigned hash(const auto& key) { return HashFunctions::hash(key); }
static bool equal(const auto& a, const auto& b) { return HashFunctions::equal(a, b); }
static void translate(auto& location, const auto&, NOESCAPE const Invocable<typename ValueTraits::TraitType()> auto& functor)
{
ValueTraits::assignToEmpty(location, functor());
}
};
template<typename IteratorType> struct HashTableAddResult {
HashTableAddResult() : isNewEntry(false) { }
HashTableAddResult(IteratorType iter, bool isNewEntry) : iterator(iter), isNewEntry(isNewEntry) { }
IteratorType iterator;
bool isNewEntry;
explicit operator bool() const { return isNewEntry; }
};
// HashTableCapacityForSize computes the upper power of two capacity to hold the size parameter.
// This is done at compile time to initialize the HashTraits.
template<unsigned size>
struct HashTableCapacityForSize {
// Load-factor for small table is 75%.
static constexpr unsigned smallMaxLoadNumerator = 3;
static constexpr unsigned smallMaxLoadDenominator = 4;
// Load-factor for large table is 50%.
static constexpr unsigned largeMaxLoadNumerator = 1;
static constexpr unsigned largeMaxLoadDenominator = 2;
static constexpr unsigned maxSmallTableCapacity = 1024;
static constexpr unsigned minLoad = 6;
static constexpr bool shouldExpand(uint64_t keyAndDeleteCount, uint64_t tableSize)
{
if (tableSize <= maxSmallTableCapacity)
return keyAndDeleteCount * smallMaxLoadDenominator >= tableSize * smallMaxLoadNumerator;
return keyAndDeleteCount * largeMaxLoadDenominator >= tableSize * largeMaxLoadNumerator;
}
static constexpr unsigned capacityForSize(uint32_t sizeArg)
{
if (!sizeArg)
return 0;
constexpr unsigned maxCapacity = 1U << 31;
UNUSED_PARAM(maxCapacity);
ASSERT_UNDER_CONSTEXPR_CONTEXT(sizeArg <= maxCapacity);
uint32_t capacity = roundUpToPowerOfTwo(sizeArg);
ASSERT_UNDER_CONSTEXPR_CONTEXT(capacity <= maxCapacity);
if (shouldExpand(sizeArg, capacity)) {
ASSERT_UNDER_CONSTEXPR_CONTEXT((static_cast<uint64_t>(capacity) * 2) <= maxCapacity);
return capacity * 2;
}
return capacity;
}
static constexpr unsigned value = capacityForSize(size);
static_assert(size > 0, "HashTableNonZeroMinimumCapacity");
static_assert(!static_cast<unsigned>(value >> 31), "HashTableNoCapacityOverflow");
};
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
class HashTable {
public:
using HashTableType = HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>;
typedef HashTableIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits> iterator;
typedef HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits> const_iterator;
typedef Traits ValueTraits;
typedef Key KeyType;
typedef Value ValueType;
using TakeType = typename ValueTraits::TakeType;
typedef IdentityHashTranslator<ValueTraits, HashFunctions> IdentityTranslatorType;
typedef HashTableAddResult<iterator> AddResult;
using HashTableSizePolicy = HashTableCapacityForSize<1>;
#if DUMP_HASHTABLE_STATS_PER_TABLE
struct Stats {
WTF_MAKE_STRUCT_FAST_ALLOCATED;
Stats()
: numAccesses(0)
, numRehashes(0)
, numRemoves(0)
, numReinserts(0)
, maxCollisions(0)
, numCollisions(0)
, collisionGraph()
{
}
unsigned numAccesses;
unsigned numRehashes;
unsigned numRemoves;
unsigned numReinserts;
unsigned maxCollisions;
unsigned numCollisions;
unsigned collisionGraph[4096];
void recordCollisionAtCount(unsigned count)
{
if (count > maxCollisions)
maxCollisions = count;
numCollisions++;
collisionGraph[count]++;
}
void dumpStats()
{
dataLogF("\nWTF::HashTable::Stats dump\n\n");
dataLogF("%d accesses\n", numAccesses);
dataLogF("%d total collisions, average %.2f probes per access\n", numCollisions, 1.0 * (numAccesses + numCollisions) / numAccesses);
dataLogF("longest collision chain: %d\n", maxCollisions);
for (unsigned i = 1; i <= maxCollisions; i++) {
dataLogF(" %d lookups with exactly %d collisions (%.2f%% , %.2f%% with this many or more)\n", collisionGraph[i], i, 100.0 * (collisionGraph[i] - collisionGraph[i+1]) / numAccesses, 100.0 * collisionGraph[i] / numAccesses);
}
dataLogF("%d rehashes\n", numRehashes);
dataLogF("%d reinserts\n", numReinserts);
}
};
#endif
HashTable();
~HashTable()
{
invalidateIterators(this);
if (m_table)
deallocateTable(m_table);
#if CHECK_HASHTABLE_USE_AFTER_DESTRUCTION
m_table = (ValueType*)(uintptr_t)0xbbadbeef;
#endif
}
HashTable(const HashTable&);
void swap(HashTable&);
HashTable& operator=(const HashTable&);
HashTable(HashTable&&);
HashTable& operator=(HashTable&&);
// When the hash table is empty, just return the same iterator for end as for begin.
// This is more efficient because we don't have to skip all the empty and deleted
// buckets, and iterating an empty table is a common case that's worth optimizing.
iterator begin() { return isEmpty() ? end() : makeIterator(m_table); }
iterator end() { return makeKnownGoodIterator(m_table + tableSize()); }
const_iterator begin() const { return isEmpty() ? end() : makeConstIterator(m_table); }
const_iterator end() const { return makeKnownGoodConstIterator(m_table + tableSize()); }
iterator random()
{
if (isEmpty())
return end();
while (true) {
auto& bucket = m_table[weakRandomNumber<uint32_t>() & tableSizeMask()];
if (!isEmptyOrDeletedBucket(bucket))
return makeKnownGoodIterator(&bucket);
};
}
const_iterator random() const { return static_cast<const_iterator>(const_cast<HashTable*>(this)->random()); }
unsigned size() const { return keyCount(); }
unsigned capacity() const { return tableSize(); }
size_t byteSize() const { return metadataSize + tableSize() * sizeof(ValueType); }
bool isEmpty() const { return !keyCount(); }
void reserveInitialCapacity(unsigned keyCount)
{
ASSERT(!m_table);
ASSERT(!tableSize());
unsigned minimumTableSize = KeyTraits::minimumTableSize;
unsigned newTableSize = std::max(minimumTableSize, computeBestTableSize(keyCount));
m_table = allocateTable(newTableSize);
setTableSize(newTableSize);
setTableSizeMask(newTableSize - 1);
setDeletedCount(0);
setKeyCount(0);
}
AddResult add(const ValueType& value) { return add<IdentityTranslatorType>(Extractor::extract(value), [&]() ALWAYS_INLINE_LAMBDA { return value; }); }
AddResult add(ValueType&& value) { return add<IdentityTranslatorType>(Extractor::extract(value), [&]() ALWAYS_INLINE_LAMBDA { return WTFMove(value); }); }
// A special version of add() that finds the object by hashing and comparing
// with some other type, to avoid the cost of type conversion if the object is already
// in the table.
template<typename HashTranslator> AddResult add(auto&& key, NOESCAPE const std::invocable<> auto& functor);
template<typename HashTranslator> AddResult addPassingHashCode(auto&& key, NOESCAPE const std::invocable<> auto& functor);
iterator find(const KeyType& key) { return find<IdentityTranslatorType>(key); }
const_iterator find(const KeyType& key) const { return find<IdentityTranslatorType>(key); }
bool contains(const KeyType& key) const { return contains<IdentityTranslatorType>(key); }
template<typename HashTranslator, typename T> iterator find(const T&);
template<typename HashTranslator, typename T> const_iterator find(const T&) const;
template<typename HashTranslator, typename T> bool contains(const T&) const;
void remove(const KeyType&);
void remove(iterator);
void removeWithoutEntryConsistencyCheck(iterator);
void removeWithoutEntryConsistencyCheck(const_iterator);
// FIXME: This feels like it should be Invocable<bool(const ValueType&)> but that breaks many HashMap users.
bool removeIf(NOESCAPE const Invocable<bool(ValueType&)> auto&);
void clear();
template<size_t inlineCapacity>
Vector<TakeType, inlineCapacity> takeIf(NOESCAPE const Invocable<bool(const ValueType&)> auto&);
static bool isEmptyBucket(const ValueType& value) { return isHashTraitsEmptyValue<KeyTraits>(Extractor::extract(value)); }
static bool isReleasedWeakBucket(const ValueType& value) { return isHashTraitsReleasedWeakValue<KeyTraits>(Extractor::extract(value)); }
static bool isDeletedBucket(const ValueType& value) { return KeyTraits::isDeletedValue(Extractor::extract(value)); }
static bool isEmptyOrDeletedBucket(const ValueType& value) { return isEmptyBucket(value) || isDeletedBucket(value); }
ValueType* lookup(const Key& key) { return lookup<IdentityTranslatorType>(key); }
template<typename HashTranslator, typename T> ValueType* lookup(const T&);
template<typename HashTranslator, typename T> ValueType* inlineLookup(const T&);
ALWAYS_INLINE bool isNullStorage() const { return !m_table; }
#if ASSERT_ENABLED
void checkTableConsistency() const;
#else
static void checkTableConsistency() { }
#endif
#if CHECK_HASHTABLE_CONSISTENCY
void internalCheckTableConsistency() const { checkTableConsistency(); }
void internalCheckTableConsistencyExceptSize() const { checkTableConsistencyExceptSize(); }
#else
static void internalCheckTableConsistencyExceptSize() { }
static void internalCheckTableConsistency() { }
#endif
private:
static ValueType* allocateTable(unsigned size);
static void deallocateTable(ValueType* table);
typedef std::pair<ValueType*, bool> LookupType;
typedef std::pair<LookupType, unsigned> FullLookupType;
ValueType* lookupForReinsert(const Key& key) { return lookupForReinsert<IdentityTranslatorType>(key); };
template<typename HashTranslator, typename T> ValueType* lookupForReinsert(const T&);
template<typename HashTranslator, typename T> FullLookupType fullLookupForWriting(const T&);
template<typename HashTranslator> void addUniqueForInitialization(auto&& key, NOESCAPE const std::invocable<> auto& functor);
template<typename HashTranslator, typename T> void checkKey(const T&);
void removeAndInvalidateWithoutEntryConsistencyCheck(ValueType*);
void removeAndInvalidate(ValueType*);
void remove(ValueType*);
static constexpr unsigned computeBestTableSize(unsigned keyCount);
bool shouldExpand() const { return HashTableSizePolicy::shouldExpand(keyCount() + deletedCount(), tableSize()); }
bool mustRehashInPlace() const { return keyCount() * minLoad < tableSize() * 2; }
bool shouldShrink() const { return keyCount() * minLoad < tableSize() && tableSize() > KeyTraits::minimumTableSize; }
ValueType* expand(ValueType* entry = nullptr);
void shrink() { rehash(tableSize() / 2, nullptr); }
void shrinkToBestSize();
void deleteReleasedWeakBuckets();
ValueType* rehash(unsigned newTableSize, ValueType* entry);
ValueType* reinsert(ValueType&&);
static void initializeBucket(ValueType& bucket);
static void deleteBucket(ValueType& bucket) { hashTraitsDeleteBucket<Traits>(bucket); }
FullLookupType makeLookupResult(ValueType* position, bool found, unsigned hash)
{ return FullLookupType(LookupType(position, found), hash); }
iterator makeIterator(ValueType* pos) { return iterator(this, pos, m_table + tableSize()); }
const_iterator makeConstIterator(ValueType* pos) const { return const_iterator(this, pos, m_table + tableSize()); }
iterator makeKnownGoodIterator(ValueType* pos) { return iterator(this, pos, m_table + tableSize(), HashItemKnownGood); }
const_iterator makeKnownGoodConstIterator(ValueType* pos) const { return const_iterator(this, pos, m_table + tableSize(), HashItemKnownGood); }
#if ASSERT_ENABLED
void checkTableConsistencyExceptSize() const;
#else
static void checkTableConsistencyExceptSize() { }
#endif
// Load-factor for small table is 75%.
static constexpr unsigned smallMaxLoadNumerator = HashTableSizePolicy::smallMaxLoadNumerator;
static constexpr unsigned smallMaxLoadDenominator = HashTableSizePolicy::smallMaxLoadDenominator;
// Load-factor for large table is 50%.
static constexpr unsigned largeMaxLoadNumerator = HashTableSizePolicy::largeMaxLoadNumerator;
static constexpr unsigned largeMaxLoadDenominator = HashTableSizePolicy::largeMaxLoadDenominator;
static constexpr unsigned maxSmallTableCapacity = HashTableSizePolicy::maxSmallTableCapacity;
static constexpr unsigned minLoad = HashTableSizePolicy::minLoad;
static constexpr int tableSizeOffset = -1;
static constexpr int tableSizeMaskOffset = -2;
static constexpr int keyCountOffset = -3;
static constexpr int deletedCountOffset = -4;
static constexpr unsigned metadataSize = std::max(4 * sizeof(unsigned), alignof(ValueType));
unsigned tableSize() const { return m_table ? reinterpret_cast_ptr<unsigned*>(m_table)[tableSizeOffset] : 0; }
void setTableSize(unsigned size) const { ASSERT(m_table); reinterpret_cast_ptr<unsigned*>(m_table)[tableSizeOffset] = size; }
unsigned tableSizeMask() const { ASSERT(m_table); return m_table ? reinterpret_cast_ptr<unsigned*>(m_table)[tableSizeMaskOffset] : 0; }
void setTableSizeMask(unsigned mask) { ASSERT(m_table); reinterpret_cast_ptr<unsigned*>(m_table)[tableSizeMaskOffset] = mask; }
unsigned keyCount() const { return m_table ? reinterpret_cast_ptr<unsigned*>(m_table)[keyCountOffset] : 0; }
void setKeyCount(unsigned count) const { ASSERT(m_table); reinterpret_cast_ptr<unsigned*>(m_table)[keyCountOffset] = count; }
unsigned deletedCount() const { ASSERT(m_table); return reinterpret_cast_ptr<unsigned*>(m_table)[deletedCountOffset]; }
void setDeletedCount(unsigned count) const { ASSERT(m_table); reinterpret_cast_ptr<unsigned*>(m_table)[deletedCountOffset] = count; }
union {
ValueType* m_table { nullptr };
unsigned* m_tableForLLDB;
};
#if CHECK_HASHTABLE_ITERATORS
public:
// All access to m_iterators should be guarded with m_mutex.
mutable const_iterator* m_iterators;
// Use std::unique_ptr so HashTable can still be memmove'd or memcpy'ed.
mutable std::unique_ptr<Lock> m_mutex;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
public:
mutable std::unique_ptr<Stats> m_stats;
#endif
};
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::HashTable()
: m_table(nullptr)
#if CHECK_HASHTABLE_ITERATORS
, m_iterators(0)
, m_mutex(makeUnique<Lock>())
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
, m_stats(makeUnique<Stats>())
#endif
{
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
inline void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::checkKey(const T& key)
{
if constexpr (!ASSERT_ENABLED && shouldValidateKey == ShouldValidateKey::No)
return;
if constexpr (std::is_convertible_v<T, Key>) {
RELEASE_ASSERT(!isHashTraitsEmptyValue<KeyTraits>(key));
RELEASE_ASSERT(!KeyTraits::isDeletedValue(key));
} else if constexpr (HashFunctions::safeToCompareToEmptyOrDeleted) {
RELEASE_ASSERT(!HashTranslator::equal(KeyTraits::emptyValue(), key));
AlignedStorage<ValueType> deletedValueBuffer;
auto& deletedValue = *deletedValueBuffer;
Traits::constructDeletedValue(deletedValue);
RELEASE_ASSERT(!HashTranslator::equal(Extractor::extract(deletedValue), key));
}
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
inline auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::lookup(const T& key) -> ValueType*
{
return inlineLookup<HashTranslator>(key);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
ALWAYS_INLINE auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::inlineLookup(const T& key) -> ValueType*
{
static_assert(sizeof(Value) <= 150, "Your HashTable types are too big to efficiently move when rehashing. Consider using UniqueRef instead");
checkKey<HashTranslator>(key);
ValueType* table = m_table;
if (!table)
return nullptr;
unsigned sizeMask = tableSizeMask();
unsigned h = HashTranslator::hash(key);
unsigned i = h & sizeMask;
unsigned probeCount = 0;
#if DUMP_HASHTABLE_STATS
++HashTableStats::numAccesses;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
++m_stats->numAccesses;
#endif
while (true) {
ValueType* entry = table + i;
// we count on the compiler to optimize out this branch
if (HashFunctions::safeToCompareToEmptyOrDeleted) {
if (HashTranslator::equal(Extractor::extract(*entry), key))
return entry;
if (isEmptyBucket(*entry))
return nullptr;
} else {
if (isEmptyBucket(*entry))
return nullptr;
if (!isDeletedBucket(*entry) && HashTranslator::equal(Extractor::extract(*entry), key))
return entry;
}
++probeCount;
#if DUMP_HASHTABLE_STATS
HashTableStats::recordCollisionAtCount(probeCount);
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
m_stats->recordCollisionAtCount(probeCount);
#endif
i = (i + probeCount) & sizeMask;
}
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
inline auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::lookupForReinsert(const T& key) -> ValueType*
{
ASSERT(m_table);
checkKey<HashTranslator>(key);
ValueType* table = m_table;
unsigned sizeMask = tableSizeMask();
unsigned h = HashTranslator::hash(key);
unsigned i = h & sizeMask;
unsigned probeCount = 0;
#if DUMP_HASHTABLE_STATS
++HashTableStats::numAccesses;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
++m_stats->numAccesses;
#endif
while (true) {
ValueType* entry = table + i;
if (isEmptyBucket(*entry))
return entry;
++probeCount;
#if DUMP_HASHTABLE_STATS
HashTableStats::recordCollisionAtCount(probeCount);
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
m_stats->recordCollisionAtCount(probeCount);
#endif
i = (i + probeCount) & sizeMask;
}
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
inline auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::fullLookupForWriting(const T& key) -> FullLookupType
{
ASSERT(m_table);
checkKey<HashTranslator>(key);
ValueType* table = m_table;
unsigned sizeMask = tableSizeMask();
unsigned h = HashTranslator::hash(key);
unsigned i = h & sizeMask;
unsigned probeCount = 0;
#if DUMP_HASHTABLE_STATS
++HashTableStats::numAccesses;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
++m_stats->numAccesses;
#endif
ValueType* deletedEntry = nullptr;
while (true) {
ValueType* entry = table + i;
// we count on the compiler to optimize out this branch
if (HashFunctions::safeToCompareToEmptyOrDeleted) {
if (isEmptyBucket(*entry))
return makeLookupResult(deletedEntry ? deletedEntry : entry, false, h);
if (HashTranslator::equal(Extractor::extract(*entry), key))
return makeLookupResult(entry, true, h);
if (isDeletedBucket(*entry))
deletedEntry = entry;
} else {
if (isEmptyBucket(*entry))
return makeLookupResult(deletedEntry ? deletedEntry : entry, false, h);
if (isDeletedBucket(*entry))
deletedEntry = entry;
else if (HashTranslator::equal(Extractor::extract(*entry), key))
return makeLookupResult(entry, true, h);
}
++probeCount;
#if DUMP_HASHTABLE_STATS
HashTableStats::recordCollisionAtCount(probeCount);
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
m_stats->recordCollisionAtCount(probeCount);
#endif
i = (i + probeCount) & sizeMask;
}
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
ALWAYS_INLINE void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::addUniqueForInitialization(T&& key, NOESCAPE const std::invocable<> auto& functor)
{
ASSERT(m_table);
checkKey<HashTranslator>(key);
invalidateIterators(this);
internalCheckTableConsistency();
Value* entry = lookupForReinsert<HashTranslator>(key);
HashTranslator::translate(*entry, std::forward<T>(key), functor);
internalCheckTableConsistency();
}
template<bool emptyValueIsZero> struct HashTableBucketInitializer;
template<> struct HashTableBucketInitializer<false> {
template<typename Traits, typename Value> static void initialize(Value& bucket)
{
Traits::template constructEmptyValue<Traits>(bucket);
}
};
template<> struct HashTableBucketInitializer<true> {
template<typename Traits, typename Value> static void initialize(Value& bucket)
{
// This initializes the bucket without copying the empty value.
// That makes it possible to use this with types that don't support copying.
// The memset to 0 looks like a slow operation but is optimized by the compilers.
zeroBytes(bucket);
}
};
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::initializeBucket(ValueType& bucket)
{
HashTableBucketInitializer<Traits::emptyValueIsZero>::template initialize<Traits>(bucket);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
ALWAYS_INLINE auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::add(T&& key, NOESCAPE const std::invocable<> auto& functor) -> AddResult
{
checkKey<HashTranslator>(key);
invalidateIterators(this);
if (!m_table)
expand(nullptr);
internalCheckTableConsistency();
ASSERT(m_table);
ValueType* table = m_table;
unsigned sizeMask = tableSizeMask();
unsigned h = HashTranslator::hash(key);
unsigned i = h & sizeMask;
unsigned probeCount = 0;
#if DUMP_HASHTABLE_STATS
++HashTableStats::numAccesses;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
++m_stats->numAccesses;
#endif
ValueType* deletedEntry = nullptr;
ValueType* entry;
while (true) {
entry = table + i;
// we count on the compiler to optimize out this branch
if (HashFunctions::safeToCompareToEmptyOrDeleted) {
if (isEmptyBucket(*entry))
break;
if (HashTranslator::equal(Extractor::extract(*entry), key))
return AddResult(makeKnownGoodIterator(entry), false);
if (isDeletedBucket(*entry))
deletedEntry = entry;
} else {
if (isEmptyBucket(*entry))
break;
if (isDeletedBucket(*entry))
deletedEntry = entry;
else if (HashTranslator::equal(Extractor::extract(*entry), key))
return AddResult(makeKnownGoodIterator(entry), false);
}
++probeCount;
#if DUMP_HASHTABLE_STATS
HashTableStats::recordCollisionAtCount(probeCount);
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
m_stats->recordCollisionAtCount(probeCount);
#endif
i = (i + probeCount) & sizeMask;
}
if (deletedEntry) {
initializeBucket(*deletedEntry);
entry = deletedEntry;
setDeletedCount(deletedCount() - 1);
}
HashTranslator::translate(*entry, std::forward<T>(key), functor);
setKeyCount(keyCount() + 1);
if (shouldExpand())
entry = expand(entry);
internalCheckTableConsistency();
return AddResult(makeKnownGoodIterator(entry), true);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<typename HashTranslator, typename T>
inline auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::addPassingHashCode(T&& key, NOESCAPE const std::invocable<> auto& functor) -> AddResult
{
checkKey<HashTranslator>(key);
invalidateIterators(this);
if (!m_table)
expand();
internalCheckTableConsistency();
FullLookupType lookupResult = fullLookupForWriting<HashTranslator>(key);
ValueType* entry = lookupResult.first.first;
bool found = lookupResult.first.second;
unsigned h = lookupResult.second;
if (found)
return AddResult(makeKnownGoodIterator(entry), false);
if (isDeletedBucket(*entry)) {
initializeBucket(*entry);
setDeletedCount(deletedCount() - 1);
}
HashTranslator::translate(*entry, std::forward<T>(key), functor, h);
setKeyCount(keyCount() + 1);
if (shouldExpand())
entry = expand(entry);
internalCheckTableConsistency();
return AddResult(makeKnownGoodIterator(entry), true);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::reinsert(ValueType&& entry) -> ValueType*
{
ASSERT(m_table);
ASSERT(!isDeletedBucket(*lookupForReinsert(Extractor::extract(entry))));
ASSERT(isEmptyBucket(*lookupForReinsert(Extractor::extract(entry))));
#if DUMP_HASHTABLE_STATS
++HashTableStats::numReinserts;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
++m_stats->numReinserts;
#endif
Value* newEntry = lookupForReinsert(Extractor::extract(entry));
newEntry->~Value();
new (NotNull, newEntry) ValueType(WTFMove(entry));
return newEntry;
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template <typename HashTranslator, typename T>
auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::find(const T& key) -> iterator
{
if (!m_table)
return end();
ValueType* entry = lookup<HashTranslator>(key);
if (!entry)
return end();
return makeKnownGoodIterator(entry);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template <typename HashTranslator, typename T>
auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::find(const T& key) const -> const_iterator
{
if (!m_table)
return end();
ValueType* entry = const_cast<HashTable*>(this)->lookup<HashTranslator>(key);
if (!entry)
return end();
return makeKnownGoodConstIterator(entry);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template <typename HashTranslator, typename T>
bool HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::contains(const T& key) const
{
if (!m_table)
return false;
return const_cast<HashTable*>(this)->lookup<HashTranslator>(key);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::removeAndInvalidateWithoutEntryConsistencyCheck(ValueType* pos)
{
invalidateIterators(this);
remove(pos);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::removeAndInvalidate(ValueType* pos)
{
invalidateIterators(this);
internalCheckTableConsistency();
remove(pos);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::remove(ValueType* pos)
{
#if DUMP_HASHTABLE_STATS
++HashTableStats::numRemoves;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
++m_stats->numRemoves;
#endif
deleteBucket(*pos);
setDeletedCount(deletedCount() + 1);
setKeyCount(keyCount() - 1);
if (shouldShrink())
shrink();
internalCheckTableConsistency();
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::remove(iterator it)
{
if (it == end())
return;
removeAndInvalidate(const_cast<ValueType*>(it.m_iterator.m_position));
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::removeWithoutEntryConsistencyCheck(iterator it)
{
if (it == end())
return;
removeAndInvalidateWithoutEntryConsistencyCheck(const_cast<ValueType*>(it.m_iterator.m_position));
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::removeWithoutEntryConsistencyCheck(const_iterator it)
{
if (it == end())
return;
removeAndInvalidateWithoutEntryConsistencyCheck(const_cast<ValueType*>(it.m_position));
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::remove(const KeyType& key)
{
remove(find(key));
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline bool HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::removeIf(NOESCAPE const Invocable<bool(ValueType&)> auto& functor)
{
// We must use local copies in case "functor" or "deleteBucket"
// make a function call, which prevents the compiler from keeping
// the values in register.
unsigned removedBucketCount = 0;
ValueType* table = m_table;
for (unsigned i = tableSize(); i--;) {
ValueType& bucket = table[i];
if (isEmptyOrDeletedBucket(bucket))
continue;
if (!functor(bucket))
continue;
deleteBucket(bucket);
++removedBucketCount;
}
if (removedBucketCount) {
setDeletedCount(deletedCount() + removedBucketCount);
setKeyCount(keyCount() - removedBucketCount);
}
if (shouldShrink())
shrinkToBestSize();
internalCheckTableConsistency();
return removedBucketCount;
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
template<size_t inlineCapacity>
inline auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::takeIf(NOESCAPE const Invocable<bool(const ValueType&)> auto& functor) -> Vector<TakeType, inlineCapacity>
{
// We must use local copies in case "functor" or "deleteBucket"
// make a function call, which prevents the compiler from keeping
// the values in register.
unsigned removedBucketCount = 0;
ValueType* table = m_table;
Vector<TakeType, inlineCapacity> result;
for (unsigned i = tableSize(); i--;) {
ValueType& bucket = table[i];
if (isEmptyOrDeletedBucket(bucket))
continue;
if (!functor(bucket))
continue;
result.append(ValueTraits::take(WTFMove(bucket)));
deleteBucket(bucket);
++removedBucketCount;
}
if (removedBucketCount) {
setDeletedCount(deletedCount() + removedBucketCount);
setKeyCount(keyCount() - removedBucketCount);
}
if (shouldShrink())
shrinkToBestSize();
internalCheckTableConsistency();
return result;
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::allocateTable(unsigned size) -> ValueType*
{
static_assert(!(metadataSize % alignof(ValueType)));
// would use a template member function with explicit specializations here, but
// gcc doesn't appear to support that
if (Traits::emptyValueIsZero)
return reinterpret_cast_ptr<ValueType*>(static_cast<char*>(HashTableMalloc::zeroedMalloc(metadataSize + size * sizeof(ValueType))) + metadataSize);
ValueType* result = reinterpret_cast_ptr<ValueType*>(static_cast<char*>(HashTableMalloc::malloc(metadataSize + size * sizeof(ValueType))) + metadataSize);
for (unsigned i = 0; i < size; i++)
initializeBucket(result[i]);
return result;
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::deallocateTable(ValueType* table)
{
unsigned size = reinterpret_cast_ptr<unsigned*>(table)[tableSizeOffset];
for (unsigned i = 0; i < size; ++i) {
if (!isDeletedBucket(table[i]))
table[i].~ValueType();
}
HashTableMalloc::free(reinterpret_cast<char*>(table) - metadataSize);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::expand(ValueType* entry) -> ValueType*
{
if (KeyTraits::hasIsReleasedWeakValueFunction)
deleteReleasedWeakBuckets();
unsigned newSize;
unsigned oldSize = tableSize();
if (!oldSize)
newSize = KeyTraits::minimumTableSize;
else if (mustRehashInPlace())
newSize = oldSize;
else
newSize = oldSize * 2;
return rehash(newSize, entry);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
constexpr unsigned HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::computeBestTableSize(unsigned keyCount)
{
unsigned bestTableSize = WTF::roundUpToPowerOfTwo(keyCount);
if (HashTableSizePolicy::shouldExpand(keyCount, bestTableSize))
bestTableSize *= 2;
auto aboveThresholdForEagerExpansion = [](double loadFactor, unsigned keyCount, unsigned tableSize)
{
// Here is the rationale behind this calculation, using 3/4 load-factor.
// With maxLoad at 3/4 and minLoad at 1/6, our average load is 11/24.
// If we are getting half-way between 11/24 and 3/4, we double the size
// to avoid being too close to loadMax and bring the ratio close to 11/24. This
// give us a load in the bounds [9/24, 15/24).
double maxLoadRatio = loadFactor;
double minLoadRatio = 1.0 / minLoad;
double averageLoadRatio = (maxLoadRatio + minLoadRatio) / 2;
double halfWayBetweenAverageAndMaxLoadRatio = (averageLoadRatio + maxLoadRatio) / 2;
return keyCount >= tableSize * halfWayBetweenAverageAndMaxLoadRatio;
};
if (bestTableSize <= maxSmallTableCapacity) {
constexpr double smallLoadFactor = static_cast<double>(smallMaxLoadNumerator) / smallMaxLoadDenominator;
if (aboveThresholdForEagerExpansion(smallLoadFactor, keyCount, bestTableSize))
bestTableSize *= 2;
} else {
constexpr double largeLoadFactor = static_cast<double>(largeMaxLoadNumerator) / largeMaxLoadDenominator;
if (aboveThresholdForEagerExpansion(largeLoadFactor, keyCount, bestTableSize))
bestTableSize *= 2;
}
unsigned minimumTableSize = KeyTraits::minimumTableSize;
return std::max(bestTableSize, minimumTableSize);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::shrinkToBestSize()
{
unsigned minimumTableSize = KeyTraits::minimumTableSize;
rehash(std::max(minimumTableSize, computeBestTableSize(keyCount())), nullptr);
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::deleteReleasedWeakBuckets()
{
unsigned tableSize = this->tableSize();
for (unsigned i = 0; i < tableSize; ++i) {
auto& entry = m_table[i];
if (isReleasedWeakBucket(entry)) {
deleteBucket(entry);
setDeletedCount(deletedCount() + 1);
setKeyCount(keyCount() - 1);
}
}
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::rehash(unsigned newTableSize, ValueType* entry) -> ValueType*
{
internalCheckTableConsistencyExceptSize();
unsigned oldTableSize = tableSize();
ValueType* oldTable = m_table;
#if DUMP_HASHTABLE_STATS
if (oldTableSize != 0)
++HashTableStats::numRehashes;
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
if (oldTableSize != 0)
++m_stats->numRehashes;
#endif
unsigned oldKeyCount = keyCount();
m_table = allocateTable(newTableSize);
setTableSize(newTableSize);
setTableSizeMask(newTableSize - 1);
setDeletedCount(0);
setKeyCount(oldKeyCount);
Value* newEntry = nullptr;
for (unsigned i = 0; i != oldTableSize; ++i) {
auto& oldEntry = oldTable[i];
if (isDeletedBucket(oldEntry)) {
ASSERT(std::addressof(oldEntry) != entry);
continue;
}
if (isEmptyBucket(oldEntry)) {
ASSERT(std::addressof(oldEntry) != entry);
oldTable[i].~ValueType();
continue;
}
if (isReleasedWeakBucket(oldEntry)) {
ASSERT(std::addressof(oldEntry) != entry);
oldEntry.~ValueType();
setKeyCount(keyCount() - 1);
continue;
}
Value* reinsertedEntry = reinsert(WTFMove(oldEntry));
oldEntry.~ValueType();
if (std::addressof(oldEntry) == entry) {
ASSERT(!newEntry);
newEntry = reinsertedEntry;
}
}
if (oldTable)
HashTableMalloc::free(reinterpret_cast<char*>(oldTable) - metadataSize);
internalCheckTableConsistency();
return newEntry;
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::clear()
{
invalidateIterators(this);
if (!m_table)
return;
deallocateTable(std::exchange(m_table, nullptr));
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::HashTable(const HashTable& other)
: m_table(nullptr)
#if CHECK_HASHTABLE_ITERATORS
, m_iterators(nullptr)
, m_mutex(makeUnique<Lock>())
#endif
#if DUMP_HASHTABLE_STATS_PER_TABLE
, m_stats(makeUnique<Stats>(*other.m_stats))
#endif
{
unsigned otherKeyCount = other.size();
if (!otherKeyCount)
return;
unsigned bestTableSize = computeBestTableSize(otherKeyCount);
m_table = allocateTable(bestTableSize);
setTableSize(bestTableSize);
setTableSizeMask(bestTableSize - 1);
setKeyCount(otherKeyCount);
setDeletedCount(0);
for (const auto& otherValue : other)
addUniqueForInitialization<IdentityTranslatorType>(Extractor::extract(otherValue), [&]() ALWAYS_INLINE_LAMBDA { return otherValue; });
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::swap(HashTable& other)
{
invalidateIterators(this);
invalidateIterators(&other);
std::swap(m_table, other.m_table);
#if DUMP_HASHTABLE_STATS_PER_TABLE
m_stats.swap(other.m_stats);
#endif
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::operator=(const HashTable& other) -> HashTable&
{
HashTable tmp(other);
swap(tmp);
return *this;
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::HashTable(HashTable&& other)
#if CHECK_HASHTABLE_ITERATORS
: m_iterators(nullptr)
, m_mutex(makeUnique<Lock>())
#endif
{
invalidateIterators(&other);
m_table = std::exchange(other.m_table, nullptr);
#if DUMP_HASHTABLE_STATS_PER_TABLE
m_stats = WTFMove(other.m_stats);
other.m_stats = nullptr;
#endif
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
inline auto HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::operator=(HashTable&& other) -> HashTable&
{
HashTable temp = WTFMove(other);
swap(temp);
return *this;
}
#if ASSERT_ENABLED
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::checkTableConsistency() const
{
checkTableConsistencyExceptSize();
ASSERT(!m_table || !shouldExpand());
ASSERT(!shouldShrink());
}
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
void HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>::checkTableConsistencyExceptSize() const
{
if (!m_table)
return;
unsigned count = 0;
unsigned deletedCount = 0;
unsigned tableSize = this->tableSize();
for (unsigned j = 0; j < tableSize; ++j) {
ValueType* entry = m_table + j;
if (isEmptyBucket(*entry))
continue;
if (isDeletedBucket(*entry)) {
++deletedCount;
continue;
}
auto& key = Extractor::extract(*entry);
const_iterator it = find(key);
ASSERT(entry == it.m_position);
++count;
ValueCheck<Key>::checkConsistency(key);
}
ASSERT(count == keyCount());
ASSERT(deletedCount == this->deletedCount());
ASSERT(this->tableSize() >= KeyTraits::minimumTableSize);
ASSERT(tableSizeMask());
ASSERT(this->tableSize() == tableSizeMask() + 1);
}
#endif // ASSERT_ENABLED
#if CHECK_HASHTABLE_ITERATORS
template<typename HashTableType>
void invalidateIterators(const HashTableType* table)
{
Locker locker { *table->m_mutex };
typename HashTableType::const_iterator* next;
for (typename HashTableType::const_iterator* p = table->m_iterators; p; p = next) {
next = p->m_next;
p->m_table = nullptr;
p->m_next = nullptr;
p->m_previous = nullptr;
}
table->m_iterators = nullptr;
}
template<typename HashTableType, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
void addIterator(const HashTableType* table, HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits>* it)
{
it->m_table = table;
it->m_previous = nullptr;
// Insert iterator at head of doubly-linked list of iterators.
if (!table) {
it->m_next = nullptr;
} else {
Locker locker { *table->m_mutex };
ASSERT(table->m_iterators != it);
it->m_next = table->m_iterators;
table->m_iterators = it;
if (it->m_next) {
ASSERT(!it->m_next->m_previous);
it->m_next->m_previous = it;
}
}
}
template<typename HashTableType, typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits>
void removeIterator(HashTableConstIterator<HashTableType, Key, Value, Extractor, HashFunctions, Traits, KeyTraits>* it)
{
// Delete iterator from doubly-linked list of iterators.
if (!it->m_table) {
ASSERT(!it->m_next);
ASSERT(!it->m_previous);
} else {
Locker locker { *it->m_table->m_mutex };
if (it->m_next) {
ASSERT(it->m_next->m_previous == it);
it->m_next->m_previous = it->m_previous;
}
if (it->m_previous) {
ASSERT(it->m_table->m_iterators != it);
ASSERT(it->m_previous->m_next == it);
it->m_previous->m_next = it->m_next;
} else {
ASSERT(it->m_table->m_iterators == it);
it->m_table->m_iterators = it->m_next;
}
}
it->m_table = nullptr;
it->m_next = nullptr;
it->m_previous = nullptr;
}
#endif // CHECK_HASHTABLE_ITERATORS
struct HashTableTraits {
template<typename Key, typename Value, typename Extractor, typename HashFunctions, typename Traits, typename KeyTraits, ShouldValidateKey shouldValidateKey>
using TableType = HashTable<Key, Value, Extractor, HashFunctions, Traits, KeyTraits, shouldValidateKey>;
};
// iterator adapters
template<typename HashTableType, typename ValueType> struct HashTableConstIteratorAdapter {
using iterator_category = std::forward_iterator_tag;
using value_type = ValueType;
using difference_type = ptrdiff_t;
using pointer = const value_type*;
using reference = const value_type&;
HashTableConstIteratorAdapter() {}
HashTableConstIteratorAdapter(const typename HashTableType::const_iterator& impl) : m_impl(impl) {}
const ValueType* get() const { return (const ValueType*)m_impl.get(); }
const ValueType& operator*() const { return *get(); }
const ValueType* operator->() const { return get(); }
HashTableConstIteratorAdapter& operator++() { ++m_impl; return *this; }
HashTableConstIteratorAdapter& operator++(int) { auto result = *this; ++m_impl; return result; }
typename HashTableType::const_iterator m_impl;
};
template<typename HashTableType, typename ValueType> struct HashTableIteratorAdapter {
using iterator_category = std::forward_iterator_tag;
using value_type = ValueType;
using difference_type = ptrdiff_t;
using pointer = value_type*;
using reference = value_type&;
HashTableIteratorAdapter() {}
HashTableIteratorAdapter(const typename HashTableType::iterator& impl) : m_impl(impl) {}
ValueType* get() const { return (ValueType*)m_impl.get(); }
ValueType& operator*() const { return *get(); }
ValueType* operator->() const { return get(); }
HashTableIteratorAdapter& operator++() { ++m_impl; return *this; }
HashTableIteratorAdapter& operator++(int) { auto result = *this; ++m_impl; return result; }
operator HashTableConstIteratorAdapter<HashTableType, ValueType>() {
typename HashTableType::const_iterator i = m_impl;
return i;
}
typename HashTableType::iterator m_impl;
};
template<typename T, typename U>
inline bool operator==(const HashTableConstIteratorAdapter<T, U>& a, const HashTableConstIteratorAdapter<T, U>& b)
{
return a.m_impl == b.m_impl;
}
template<typename T, typename U>
inline bool operator==(const HashTableIteratorAdapter<T, U>& a, const HashTableIteratorAdapter<T, U>& b)
{
return a.m_impl == b.m_impl;
}
template<typename T, typename U>
inline bool operator==(const HashTableConstIteratorAdapter<T, U>& a, const HashTableIteratorAdapter<T, U>& b)
{
return a.m_impl == b.m_impl;
}
template<typename T, typename U>
inline bool operator==(const HashTableIteratorAdapter<T, U>& a, const HashTableConstIteratorAdapter<T, U>& b)
{
return a.m_impl == b.m_impl;
}
} // namespace WTF
WTF_ALLOW_UNSAFE_BUFFER_USAGE_END
#include <wtf/HashIterators.h>
|