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
|
/**
* Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
* SPDX-License-Identifier: Apache-2.0.
*/
#include <aws/io/event_loop.h>
#include <aws/io/host_resolver.h>
#include <aws/common/atomics.h>
#include <aws/common/clock.h>
#include <aws/common/condition_variable.h>
#include <aws/common/hash_table.h>
#include <aws/common/lru_cache.h>
#include <aws/common/mutex.h>
#include <aws/common/string.h>
#include <aws/common/task_scheduler.h>
#include <aws/common/thread.h>
#include <aws/io/logging.h>
#include <inttypes.h>
const uint64_t NS_PER_SEC = 1000000000;
const size_t AWS_DEFAULT_DNS_TTL = 30;
int aws_host_address_copy(const struct aws_host_address *from, struct aws_host_address *to) {
to->allocator = from->allocator;
to->address = aws_string_new_from_string(to->allocator, from->address);
to->host = aws_string_new_from_string(to->allocator, from->host);
to->record_type = from->record_type;
to->use_count = from->use_count;
to->connection_failure_count = from->connection_failure_count;
to->expiry = from->expiry;
to->weight = from->weight;
return AWS_OP_SUCCESS;
}
void aws_host_address_move(struct aws_host_address *from, struct aws_host_address *to) {
to->allocator = from->allocator;
to->address = from->address;
to->host = from->host;
to->record_type = from->record_type;
to->use_count = from->use_count;
to->connection_failure_count = from->connection_failure_count;
to->expiry = from->expiry;
to->weight = from->weight;
AWS_ZERO_STRUCT(*from);
}
void aws_host_address_clean_up(struct aws_host_address *address) {
if (address->address) {
aws_string_destroy((void *)address->address);
}
if (address->host) {
aws_string_destroy((void *)address->host);
}
AWS_ZERO_STRUCT(*address);
}
int aws_host_resolver_resolve_host(
struct aws_host_resolver *resolver,
const struct aws_string *host_name,
aws_on_host_resolved_result_fn *res,
const struct aws_host_resolution_config *config,
void *user_data) {
AWS_ASSERT(resolver->vtable && resolver->vtable->resolve_host);
return resolver->vtable->resolve_host(resolver, host_name, res, config, user_data);
}
int aws_host_resolver_purge_cache(struct aws_host_resolver *resolver) {
AWS_ASSERT(resolver->vtable && resolver->vtable->purge_cache);
return resolver->vtable->purge_cache(resolver);
}
int aws_host_resolver_purge_cache_with_callback(
struct aws_host_resolver *resolver,
aws_simple_completion_callback *on_purge_cache_complete_callback,
void *user_data) {
AWS_PRECONDITION(resolver);
AWS_PRECONDITION(resolver->vtable);
if (!resolver->vtable->purge_cache_with_callback) {
AWS_LOGF_ERROR(AWS_LS_IO_DNS, "purge_cache_with_callback function is not supported");
return aws_raise_error(AWS_ERROR_UNSUPPORTED_OPERATION);
}
return resolver->vtable->purge_cache_with_callback(resolver, on_purge_cache_complete_callback, user_data);
}
int aws_host_resolver_purge_host_cache(
struct aws_host_resolver *resolver,
const struct aws_host_resolver_purge_host_options *options) {
AWS_PRECONDITION(resolver);
AWS_PRECONDITION(resolver->vtable);
if (!resolver->vtable->purge_host_cache) {
AWS_LOGF_ERROR(AWS_LS_IO_DNS, "purge_host_cache function is not supported");
return aws_raise_error(AWS_ERROR_UNSUPPORTED_OPERATION);
}
return resolver->vtable->purge_host_cache(resolver, options);
}
int aws_host_resolver_record_connection_failure(
struct aws_host_resolver *resolver,
const struct aws_host_address *address) {
AWS_ASSERT(resolver->vtable && resolver->vtable->record_connection_failure);
return resolver->vtable->record_connection_failure(resolver, address);
}
/*
* Used by both the resolver for its lifetime state as well as individual host entries for theirs.
*/
enum default_resolver_state {
DRS_ACTIVE,
DRS_SHUTTING_DOWN,
};
struct default_host_resolver {
struct aws_allocator *allocator;
/*
* Mutually exclusion for the whole resolver, includes all member data and all host_entry_table operations. Once
* an entry is retrieved, this lock MAY be dropped but certain logic may hold both the resolver and the entry lock.
* The two locks must be taken in that order.
*/
struct aws_mutex resolver_lock;
/* host_name (aws_string*) -> host_entry* */
struct aws_hash_table host_entry_table;
/* Hash table of listener entries per host name. We keep this decoupled from the host entry table to allow for
* listeners to be added/removed regardless of whether or not a corresponding host entry exists.
*
* Any time the listener list in the listener entry becomes empty, we remove the entry from the table. This
* includes when a resolver thread moves all of the available listeners to its local list.
*/
/* host_name (aws_string*) -> host_listener_entry* */
struct aws_hash_table listener_entry_table;
enum default_resolver_state state;
/*
* Tracks the number of launched resolution threads that have not yet invoked their shutdown completion
* callback.
*/
uint32_t pending_host_entry_shutdown_completion_callbacks;
/*
* Function to use to query current time. Overridable in construction options.
*/
aws_io_clock_fn *system_clock_fn;
struct aws_event_loop_group *event_loop_group;
};
struct host_entry {
/* immutable post-creation */
struct aws_allocator *allocator;
struct aws_host_resolver *resolver;
struct aws_thread resolver_thread;
const struct aws_string *host_name;
int64_t resolve_frequency_ns;
struct aws_host_resolution_config resolution_config;
/* synchronized data and its lock */
struct aws_mutex entry_lock;
struct aws_condition_variable entry_signal;
struct aws_cache *aaaa_records;
struct aws_cache *a_records;
struct aws_cache *failed_connection_aaaa_records;
struct aws_cache *failed_connection_a_records;
struct aws_linked_list pending_resolution_callbacks;
uint32_t resolves_since_last_request;
uint64_t last_resolve_request_timestamp_ns;
enum default_resolver_state state;
struct aws_array_list new_addresses;
struct aws_array_list expired_addresses;
aws_simple_completion_callback *on_host_purge_complete;
void *on_host_purge_complete_user_data;
};
/*
* A host entry's caches hold things of this type. By using this and not the host_address directly, our
* on_remove callbacks for the cache have access to the host_entry. We wouldn't need to do this if those
* callbacks supported user data injection, but they don't and too many internal code bases already depend
* on the public API.
*/
struct aws_host_address_cache_entry {
struct aws_host_address address;
struct host_entry *entry;
};
int aws_host_address_cache_entry_copy(
const struct aws_host_address_cache_entry *from,
struct aws_host_address_cache_entry *to) {
if (aws_host_address_copy(&from->address, &to->address)) {
return AWS_OP_ERR;
}
to->entry = from->entry;
return AWS_OP_SUCCESS;
}
static void s_shutdown_host_entry(struct host_entry *entry) {
aws_mutex_lock(&entry->entry_lock);
entry->state = DRS_SHUTTING_DOWN;
/*
* intentionally signal under the lock; we can't guarantee the resolver
* is still around once the lock is released.
*/
aws_condition_variable_notify_all(&entry->entry_signal);
aws_mutex_unlock(&entry->entry_lock);
}
struct host_purge_callback_options {
struct aws_allocator *allocator;
struct aws_ref_count ref_count;
aws_simple_completion_callback *on_purge_cache_complete_callback;
void *user_data;
};
static void s_host_purge_callback_options_destroy(void *user_data) {
struct host_purge_callback_options *options = user_data;
options->on_purge_cache_complete_callback(options->user_data);
aws_mem_release(options->allocator, options);
}
static struct host_purge_callback_options *s_host_purge_callback_options_new(
struct aws_allocator *allocator,
aws_simple_completion_callback *on_purge_cache_complete_callback,
void *user_data) {
struct host_purge_callback_options *purge_callback_options =
aws_mem_calloc(allocator, 1, sizeof(struct host_purge_callback_options));
purge_callback_options->allocator = allocator;
aws_ref_count_init(
&purge_callback_options->ref_count, purge_callback_options, s_host_purge_callback_options_destroy);
purge_callback_options->on_purge_cache_complete_callback = on_purge_cache_complete_callback;
purge_callback_options->user_data = user_data;
return purge_callback_options;
}
static void s_purge_cache_callback(void *user_data) {
struct host_purge_callback_options *purge_callback_options = user_data;
aws_ref_count_release(&purge_callback_options->ref_count);
}
/*
* resolver lock must be held before calling this function
*/
static void s_clear_default_resolver_entry_table_synced(struct default_host_resolver *resolver) {
struct aws_hash_table *table = &resolver->host_entry_table;
for (struct aws_hash_iter iter = aws_hash_iter_begin(table); !aws_hash_iter_done(&iter);
aws_hash_iter_next(&iter)) {
struct host_entry *entry = iter.element.value;
s_shutdown_host_entry(entry);
}
aws_hash_table_clear(table);
}
static int s_resolver_purge_cache(struct aws_host_resolver *resolver) {
struct default_host_resolver *default_host_resolver = resolver->impl;
aws_mutex_lock(&default_host_resolver->resolver_lock);
s_clear_default_resolver_entry_table_synced(default_host_resolver);
aws_mutex_unlock(&default_host_resolver->resolver_lock);
return AWS_OP_SUCCESS;
}
static void s_purge_host_cache_callback_task(struct aws_task *task, void *arg, enum aws_task_status status) {
(void)status;
struct host_purge_callback_options *options = arg;
aws_mem_release(options->allocator, task);
aws_ref_count_release(&options->ref_count);
}
static void s_sechdule_purge_cache_callback_async(
struct default_host_resolver *default_host_resolver,
struct host_purge_callback_options *purge_callback_options) {
struct aws_task *task = aws_mem_calloc(default_host_resolver->allocator, 1, sizeof(struct aws_task));
aws_task_init(task, s_purge_host_cache_callback_task, purge_callback_options, "async_purge_host_callback_task");
struct aws_event_loop *loop = aws_event_loop_group_get_next_loop(default_host_resolver->event_loop_group);
AWS_FATAL_ASSERT(loop != NULL);
aws_event_loop_schedule_task_now(loop, task);
}
static int s_resolver_purge_cache_with_callback(
struct aws_host_resolver *resolver,
aws_simple_completion_callback *on_purge_cache_complete_callback,
void *user_data) {
if (!on_purge_cache_complete_callback) {
return s_resolver_purge_cache(resolver);
}
struct default_host_resolver *default_host_resolver = resolver->impl;
aws_mutex_lock(&default_host_resolver->resolver_lock);
struct aws_hash_table *table = &default_host_resolver->host_entry_table;
struct host_purge_callback_options *purge_callback_options = s_host_purge_callback_options_new(
default_host_resolver->allocator, on_purge_cache_complete_callback, user_data);
/* purge all cache */
for (struct aws_hash_iter iter = aws_hash_iter_begin(table); !aws_hash_iter_done(&iter);
aws_hash_iter_next(&iter)) {
struct host_entry *entry = iter.element.value;
/* acquire a refernce to wait for the callback to trigger */
aws_ref_count_acquire(&purge_callback_options->ref_count);
aws_mutex_lock(&entry->entry_lock);
entry->on_host_purge_complete = s_purge_cache_callback;
entry->on_host_purge_complete_user_data = purge_callback_options;
entry->state = DRS_SHUTTING_DOWN;
aws_mutex_unlock(&entry->entry_lock);
}
aws_hash_table_clear(table);
aws_mutex_unlock(&default_host_resolver->resolver_lock);
/* release the original reference async */
s_sechdule_purge_cache_callback_async(default_host_resolver, purge_callback_options);
return AWS_OP_SUCCESS;
}
static void s_cleanup_default_resolver(struct aws_host_resolver *resolver) {
struct default_host_resolver *default_host_resolver = resolver->impl;
aws_event_loop_group_release(default_host_resolver->event_loop_group);
aws_hash_table_clean_up(&default_host_resolver->host_entry_table);
aws_hash_table_clean_up(&default_host_resolver->listener_entry_table);
aws_mutex_clean_up(&default_host_resolver->resolver_lock);
aws_simple_completion_callback *shutdown_callback = resolver->shutdown_options.shutdown_callback_fn;
void *shutdown_completion_user_data = resolver->shutdown_options.shutdown_callback_user_data;
aws_mem_release(resolver->allocator, resolver);
/* invoke shutdown completion finally */
if (shutdown_callback != NULL) {
shutdown_callback(shutdown_completion_user_data);
}
}
static void resolver_destroy(struct aws_host_resolver *resolver) {
struct default_host_resolver *default_host_resolver = resolver->impl;
bool cleanup_resolver = false;
aws_mutex_lock(&default_host_resolver->resolver_lock);
AWS_FATAL_ASSERT(default_host_resolver->state == DRS_ACTIVE);
s_clear_default_resolver_entry_table_synced(default_host_resolver);
default_host_resolver->state = DRS_SHUTTING_DOWN;
if (default_host_resolver->pending_host_entry_shutdown_completion_callbacks == 0) {
cleanup_resolver = true;
}
aws_mutex_unlock(&default_host_resolver->resolver_lock);
if (cleanup_resolver) {
s_cleanup_default_resolver(resolver);
}
}
struct pending_callback {
aws_on_host_resolved_result_fn *callback;
void *user_data;
struct aws_linked_list_node node;
};
static void s_clear_address_list(struct aws_array_list *address_list) {
for (size_t i = 0; i < aws_array_list_length(address_list); ++i) {
struct aws_host_address *address = NULL;
aws_array_list_get_at_ptr(address_list, (void **)&address, i);
aws_host_address_clean_up(address);
}
aws_array_list_clear(address_list);
}
static void s_clean_up_host_entry(struct host_entry *entry) {
if (entry == NULL) {
return;
}
/*
* This can happen if the resolver's final reference drops while an unanswered query is pending on an entry.
*
* You could add an assertion that the resolver is in the shut down state if this condition hits but that
* requires additional locking just to make the assert.
*/
if (!aws_linked_list_empty(&entry->pending_resolution_callbacks)) {
aws_raise_error(AWS_IO_DNS_HOST_REMOVED_FROM_CACHE);
}
while (!aws_linked_list_empty(&entry->pending_resolution_callbacks)) {
struct aws_linked_list_node *resolution_callback_node =
aws_linked_list_pop_front(&entry->pending_resolution_callbacks);
struct pending_callback *pending_callback =
AWS_CONTAINER_OF(resolution_callback_node, struct pending_callback, node);
pending_callback->callback(
entry->resolver, entry->host_name, AWS_IO_DNS_HOST_REMOVED_FROM_CACHE, NULL, pending_callback->user_data);
aws_mem_release(entry->allocator, pending_callback);
}
aws_cache_destroy(entry->aaaa_records);
aws_cache_destroy(entry->a_records);
aws_cache_destroy(entry->failed_connection_a_records);
aws_cache_destroy(entry->failed_connection_aaaa_records);
aws_string_destroy((void *)entry->host_name);
s_clear_address_list(&entry->new_addresses);
aws_array_list_clean_up(&entry->new_addresses);
s_clear_address_list(&entry->expired_addresses);
aws_array_list_clean_up(&entry->expired_addresses);
aws_mem_release(entry->allocator, entry);
}
static void s_on_host_entry_shutdown_completion(void *user_data) {
struct host_entry *entry = user_data;
struct aws_host_resolver *resolver = entry->resolver;
struct default_host_resolver *default_host_resolver = resolver->impl;
s_clean_up_host_entry(entry);
bool cleanup_resolver = false;
aws_mutex_lock(&default_host_resolver->resolver_lock);
--default_host_resolver->pending_host_entry_shutdown_completion_callbacks;
if (default_host_resolver->state == DRS_SHUTTING_DOWN &&
default_host_resolver->pending_host_entry_shutdown_completion_callbacks == 0) {
cleanup_resolver = true;
}
aws_mutex_unlock(&default_host_resolver->resolver_lock);
if (cleanup_resolver) {
s_cleanup_default_resolver(resolver);
}
}
static int s_copy_address_into_array_list(struct aws_host_address *address, struct aws_array_list *address_list) {
/*
* This is the worst.
*
* We have to copy the cache address while we still have a write lock. Otherwise, connection failures
* can sneak in and destroy our address by moving the address to/from the various lru caches.
*
* But there's no nice copy construction into an array list, so we get to
* (1) Push a zeroed dummy element onto the array list
* (2) Get its pointer
* (3) Call aws_host_address_copy onto it. If that fails, pop the dummy element.
*/
struct aws_host_address dummy;
AWS_ZERO_STRUCT(dummy);
if (aws_array_list_push_back(address_list, &dummy)) {
return AWS_OP_ERR;
}
struct aws_host_address *dest_copy = NULL;
aws_array_list_get_at_ptr(address_list, (void **)&dest_copy, aws_array_list_length(address_list) - 1);
AWS_FATAL_ASSERT(dest_copy != NULL);
if (aws_host_address_copy(address, dest_copy)) {
aws_array_list_pop_back(address_list);
return AWS_OP_ERR;
}
return AWS_OP_SUCCESS;
}
static uint64_t s_get_system_time_for_default_resolver(struct aws_host_resolver *resolver) {
struct default_host_resolver *default_resolver = resolver->impl;
uint64_t timestamp = 0;
(*default_resolver->system_clock_fn)(×tamp);
return timestamp;
}
/* this only ever gets called after resolution has already run. We expect that the entry's lock
has been acquired for writing before this function is called and released afterwards. */
static inline void process_records(
struct host_entry *host_entry,
struct aws_cache *records,
struct aws_cache *failed_records) {
struct aws_host_resolver *resolver = host_entry->resolver;
uint64_t timestamp = s_get_system_time_for_default_resolver(resolver);
size_t record_count = aws_cache_get_element_count(records);
size_t expired_records = 0;
/* since this only ever gets called after resolution has already run, we're in a dns outage
* if everything is expired. Leave an element so we can keep trying. */
for (size_t index = 0; index < record_count && expired_records < record_count - 1; ++index) {
struct aws_host_address_cache_entry *lru_element_entry = aws_lru_cache_use_lru_element(records);
if (lru_element_entry->address.expiry < timestamp) {
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"static: purging expired record %s for %s",
lru_element_entry->address.address->bytes,
lru_element_entry->address.host->bytes);
expired_records++;
aws_cache_remove(records, lru_element_entry->address.address);
}
}
record_count = aws_cache_get_element_count(records);
AWS_LOGF_TRACE(AWS_LS_IO_DNS, "static: remaining record count for host %d", (int)record_count);
/* if we don't have any known good addresses, take the least recently used, but not expired address with a history
* of spotty behavior and upgrade it for reuse. If it's expired, leave it and let the resolve fail. Better to fail
* than accidentally give a kids' app an IP address to somebody's adult website when the IP address gets rebound to
* a different endpoint. The moral of the story here is to not disable SSL verification! */
if (!record_count) {
size_t failed_count = aws_cache_get_element_count(failed_records);
for (size_t index = 0; index < failed_count; ++index) {
struct aws_host_address_cache_entry *lru_element_entry = aws_lru_cache_use_lru_element(failed_records);
if (timestamp >= lru_element_entry->address.expiry) {
continue;
}
struct aws_host_address_cache_entry *to_add =
aws_mem_calloc(host_entry->allocator, 1, sizeof(struct aws_host_address_cache_entry));
if (to_add == NULL) {
continue;
}
if (aws_host_address_cache_entry_copy(lru_element_entry, to_add) ||
aws_cache_put(records, to_add->address.address, to_add)) {
aws_host_address_clean_up(&to_add->address);
aws_mem_release(host_entry->allocator, to_add);
continue;
}
/*
* Promoting an address from failed to good should trigger the new address callback
*/
s_copy_address_into_array_list(&lru_element_entry->address, &host_entry->new_addresses);
AWS_LOGF_INFO(
AWS_LS_IO_DNS,
"static: promoting spotty record %s for %s back to good list",
lru_element_entry->address.address->bytes,
lru_element_entry->address.host->bytes);
aws_cache_remove(failed_records, lru_element_entry->address.address);
/* we only want to promote one per process run.*/
break;
}
}
}
static int s_resolver_purge_host_cache(
struct aws_host_resolver *resolver,
const struct aws_host_resolver_purge_host_options *options) {
AWS_PRECONDITION(resolver);
if (options == NULL) {
AWS_LOGF_ERROR(AWS_LS_IO_DNS, "Cannot purge host cache; options structure is NULL.");
return aws_raise_error(AWS_ERROR_INVALID_ARGUMENT);
}
struct default_host_resolver *default_host_resolver = resolver->impl;
AWS_LOGF_INFO(AWS_LS_IO_DNS, "id=%p: purging record for %s", (void *)resolver, options->host->bytes);
aws_mutex_lock(&default_host_resolver->resolver_lock);
struct aws_hash_element *element = NULL;
aws_hash_table_find(&default_host_resolver->host_entry_table, options->host, &element);
/* Success if entry doesn't exist in cache. */
if (element == NULL) {
aws_mutex_unlock(&default_host_resolver->resolver_lock);
if (options->on_host_purge_complete_callback != NULL) {
/* Schedule completion callback asynchronouly */
struct host_purge_callback_options *purge_callback_options = s_host_purge_callback_options_new(
default_host_resolver->allocator, options->on_host_purge_complete_callback, options->user_data);
s_sechdule_purge_cache_callback_async(default_host_resolver, purge_callback_options);
}
return AWS_OP_SUCCESS;
}
struct host_entry *host_entry = element->value;
AWS_FATAL_ASSERT(host_entry);
/* Setup the on_host_purge_complete callback. */
aws_mutex_lock(&host_entry->entry_lock);
AWS_FATAL_ASSERT(!host_entry->on_host_purge_complete);
AWS_FATAL_ASSERT(!host_entry->on_host_purge_complete_user_data);
host_entry->on_host_purge_complete = options->on_host_purge_complete_callback;
host_entry->on_host_purge_complete_user_data = options->user_data;
aws_mutex_unlock(&host_entry->entry_lock);
s_shutdown_host_entry(host_entry);
aws_hash_table_remove_element(&default_host_resolver->host_entry_table, element);
aws_mutex_unlock(&default_host_resolver->resolver_lock);
return AWS_OP_SUCCESS;
}
static int resolver_record_connection_failure(
struct aws_host_resolver *resolver,
const struct aws_host_address *address) {
struct default_host_resolver *default_host_resolver = resolver->impl;
AWS_LOGF_INFO(
AWS_LS_IO_DNS,
"id=%p: recording failure for record %s for %s, moving to bad list",
(void *)resolver,
address->address->bytes,
address->host->bytes);
aws_mutex_lock(&default_host_resolver->resolver_lock);
struct aws_hash_element *element = NULL;
if (aws_hash_table_find(&default_host_resolver->host_entry_table, address->host, &element)) {
aws_mutex_unlock(&default_host_resolver->resolver_lock);
return AWS_OP_ERR;
}
struct host_entry *host_entry = NULL;
if (element != NULL) {
host_entry = element->value;
AWS_FATAL_ASSERT(host_entry);
}
if (host_entry) {
struct aws_host_address_cache_entry *cached_address_entry = NULL;
aws_mutex_lock(&host_entry->entry_lock);
aws_mutex_unlock(&default_host_resolver->resolver_lock);
struct aws_cache *address_table =
address->record_type == AWS_ADDRESS_RECORD_TYPE_AAAA ? host_entry->aaaa_records : host_entry->a_records;
struct aws_cache *failed_table = address->record_type == AWS_ADDRESS_RECORD_TYPE_AAAA
? host_entry->failed_connection_aaaa_records
: host_entry->failed_connection_a_records;
aws_cache_find(address_table, address->address, (void **)&cached_address_entry);
struct aws_host_address_cache_entry *address_entry_copy = NULL;
if (cached_address_entry) {
address_entry_copy = aws_mem_calloc(resolver->allocator, 1, sizeof(struct aws_host_address_cache_entry));
if (!address_entry_copy || aws_host_address_cache_entry_copy(cached_address_entry, address_entry_copy)) {
goto error_host_entry_cleanup;
}
/*
* This will trigger an expiration callback since the good caches add the removed address to the
* host_entry's expired list, via the cache's on_delete callback
*/
if (aws_cache_remove(address_table, cached_address_entry->address.address)) {
goto error_host_entry_cleanup;
}
address_entry_copy->address.connection_failure_count += 1;
if (aws_cache_put(failed_table, address_entry_copy->address.address, address_entry_copy)) {
goto error_host_entry_cleanup;
}
} else {
if (aws_cache_find(failed_table, address->address, (void **)&cached_address_entry)) {
goto error_host_entry_cleanup;
}
if (cached_address_entry) {
cached_address_entry->address.connection_failure_count += 1;
}
}
aws_mutex_unlock(&host_entry->entry_lock);
return AWS_OP_SUCCESS;
error_host_entry_cleanup:
if (address_entry_copy) {
aws_host_address_clean_up(&address_entry_copy->address);
aws_mem_release(resolver->allocator, address_entry_copy);
}
aws_mutex_unlock(&host_entry->entry_lock);
return AWS_OP_ERR;
}
aws_mutex_unlock(&default_host_resolver->resolver_lock);
return AWS_OP_SUCCESS;
}
/*
* A bunch of convenience functions for the host resolver background thread function
*/
static struct aws_host_address_cache_entry *s_find_cached_address_entry_aux(
struct aws_cache *primary_records,
struct aws_cache *fallback_records,
const struct aws_string *address) {
struct aws_host_address_cache_entry *found = NULL;
aws_cache_find(primary_records, address, (void **)&found);
if (found == NULL) {
aws_cache_find(fallback_records, address, (void **)&found);
}
return found;
}
/*
* Looks in both the good and failed connection record sets for a given host record
*/
static struct aws_host_address_cache_entry *s_find_cached_address_entry(
struct host_entry *entry,
const struct aws_string *address,
enum aws_address_record_type record_type) {
switch (record_type) {
case AWS_ADDRESS_RECORD_TYPE_AAAA:
return s_find_cached_address_entry_aux(entry->aaaa_records, entry->failed_connection_aaaa_records, address);
case AWS_ADDRESS_RECORD_TYPE_A:
return s_find_cached_address_entry_aux(entry->a_records, entry->failed_connection_a_records, address);
default:
return NULL;
}
}
static struct aws_host_address_cache_entry *s_get_lru_address_entry_aux(
struct aws_cache *primary_records,
struct aws_cache *fallback_records) {
struct aws_host_address_cache_entry *address_entry = aws_lru_cache_use_lru_element(primary_records);
if (address_entry == NULL) {
aws_lru_cache_use_lru_element(fallback_records);
}
return address_entry;
}
/*
* Looks in both the good and failed connection record sets for the LRU host record
*/
static struct aws_host_address_cache_entry *s_get_lru_address(
struct host_entry *entry,
enum aws_address_record_type record_type) {
switch (record_type) {
case AWS_ADDRESS_RECORD_TYPE_AAAA:
return s_get_lru_address_entry_aux(entry->aaaa_records, entry->failed_connection_aaaa_records);
case AWS_ADDRESS_RECORD_TYPE_A:
return s_get_lru_address_entry_aux(entry->a_records, entry->failed_connection_a_records);
default:
return NULL;
}
}
static void s_update_address_cache(
struct host_entry *host_entry,
struct aws_array_list *address_list,
uint64_t new_expiration) {
AWS_PRECONDITION(host_entry);
AWS_PRECONDITION(address_list);
for (size_t i = 0; i < aws_array_list_length(address_list); ++i) {
struct aws_host_address *fresh_resolved_address = NULL;
aws_array_list_get_at_ptr(address_list, (void **)&fresh_resolved_address, i);
struct aws_host_address_cache_entry *address_to_cache_entry = s_find_cached_address_entry(
host_entry, fresh_resolved_address->address, fresh_resolved_address->record_type);
if (address_to_cache_entry) {
address_to_cache_entry->address.expiry = new_expiration;
AWS_LOGF_TRACE(
AWS_LS_IO_DNS,
"static: updating expiry for %s for host %s to %llu",
address_to_cache_entry->address.address->bytes,
host_entry->host_name->bytes,
(unsigned long long)new_expiration);
} else {
address_to_cache_entry =
aws_mem_calloc(host_entry->allocator, 1, sizeof(struct aws_host_address_cache_entry));
aws_host_address_move(fresh_resolved_address, &address_to_cache_entry->address);
address_to_cache_entry->address.expiry = new_expiration;
address_to_cache_entry->entry = host_entry;
struct aws_cache *address_table =
address_to_cache_entry->address.record_type == AWS_ADDRESS_RECORD_TYPE_AAAA ? host_entry->aaaa_records
: host_entry->a_records;
if (aws_cache_put(address_table, address_to_cache_entry->address.address, address_to_cache_entry)) {
AWS_LOGF_ERROR(
AWS_LS_IO_DNS,
"static: could not add new address to host entry cache for host '%s' in "
"s_update_address_cache.",
host_entry->host_name->bytes);
continue;
}
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"static: new address resolved %s for host %s caching",
address_to_cache_entry->address.address->bytes,
host_entry->host_name->bytes);
struct aws_host_address new_address_copy;
if (aws_host_address_copy(&address_to_cache_entry->address, &new_address_copy)) {
AWS_LOGF_ERROR(
AWS_LS_IO_DNS,
"static: could not copy address for new-address list for host '%s' in s_update_address_cache.",
host_entry->host_name->bytes);
continue;
}
if (aws_array_list_push_back(&host_entry->new_addresses, &new_address_copy)) {
aws_host_address_clean_up(&new_address_copy);
AWS_LOGF_ERROR(
AWS_LS_IO_DNS,
"static: could not push address to new-address list for host '%s' in s_update_address_cache.",
host_entry->host_name->bytes);
continue;
}
}
}
}
static void s_copy_address_into_callback_set(
struct aws_host_address_cache_entry *entry,
struct aws_array_list *callback_addresses,
const struct aws_string *host_name) {
if (entry != NULL) {
if (s_copy_address_into_array_list(&entry->address, callback_addresses)) {
AWS_LOGF_ERROR(
AWS_LS_IO_DNS,
"static: failed to vend address %s for host %s to caller",
entry->address.address->bytes,
host_name->bytes);
return;
}
entry->address.use_count += 1;
AWS_LOGF_TRACE(
AWS_LS_IO_DNS,
"static: vending address %s for host %s to caller",
entry->address.address->bytes,
host_name->bytes);
}
}
static bool s_host_entry_finished_pred(void *user_data) {
struct host_entry *entry = user_data;
return entry->state == DRS_SHUTTING_DOWN;
}
static bool s_host_entry_finished_or_pending_request_pred(void *user_data) {
struct host_entry *entry = user_data;
return entry->state == DRS_SHUTTING_DOWN || !aws_linked_list_empty(&entry->pending_resolution_callbacks);
}
static const uint64_t AWS_MINIMUM_WAIT_BETWEEN_DNS_QUERIES_NS = 100000000; /* 100 ms */
static void aws_host_resolver_thread(void *arg) {
struct host_entry *host_entry = arg;
uint64_t max_no_solicitation_interval = aws_timestamp_convert(
aws_max_u64(1, host_entry->resolution_config.max_ttl), AWS_TIMESTAMP_SECS, AWS_TIMESTAMP_NANOS, NULL);
uint64_t wait_between_resolves_interval =
aws_min_u64(max_no_solicitation_interval, host_entry->resolve_frequency_ns);
uint64_t shutdown_only_wait_time = AWS_MINIMUM_WAIT_BETWEEN_DNS_QUERIES_NS;
uint64_t request_interruptible_wait_time = 0;
if (wait_between_resolves_interval > shutdown_only_wait_time) {
request_interruptible_wait_time = wait_between_resolves_interval - shutdown_only_wait_time;
}
struct aws_linked_list listener_list;
aws_linked_list_init(&listener_list);
struct aws_linked_list listener_destroy_list;
aws_linked_list_init(&listener_destroy_list);
bool keep_going = true;
struct aws_array_list address_list;
AWS_ZERO_STRUCT(address_list);
struct aws_array_list new_address_list;
AWS_ZERO_STRUCT(new_address_list);
struct aws_array_list expired_address_list;
AWS_ZERO_STRUCT(expired_address_list);
if (aws_array_list_init_dynamic(&address_list, host_entry->allocator, 4, sizeof(struct aws_host_address))) {
goto done;
}
if (aws_array_list_init_dynamic(&new_address_list, host_entry->allocator, 4, sizeof(struct aws_host_address))) {
goto done;
}
if (aws_array_list_init_dynamic(&expired_address_list, host_entry->allocator, 4, sizeof(struct aws_host_address))) {
goto done;
}
while (keep_going) {
/* resolve and then process each record */
int err_code = AWS_ERROR_SUCCESS;
if (host_entry->resolution_config.impl(
host_entry->allocator, host_entry->host_name, &address_list, host_entry->resolution_config.impl_data)) {
err_code = aws_last_error();
}
if (err_code == AWS_ERROR_SUCCESS) {
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"static, resolving host %s successful, returned %d addresses",
aws_string_c_str(host_entry->host_name),
(int)aws_array_list_length(&address_list));
} else {
AWS_LOGF_WARN(
AWS_LS_IO_DNS,
"static, resolving host %s failed, ec %d (%s)",
aws_string_c_str(host_entry->host_name),
err_code,
aws_error_debug_str(err_code));
}
uint64_t timestamp = s_get_system_time_for_default_resolver(host_entry->resolver);
uint64_t new_expiry = timestamp + (host_entry->resolution_config.max_ttl * NS_PER_SEC);
struct aws_linked_list pending_resolve_copy;
aws_linked_list_init(&pending_resolve_copy);
/*
* Within the lock we
* (1) Update the cache with the newly resolved addresses
* (2) Process all held addresses looking for expired or promotable ones
* (3) Prep for callback invocations
*/
aws_mutex_lock(&host_entry->entry_lock);
if (!err_code) {
s_update_address_cache(host_entry, &address_list, new_expiry);
}
/*
* process and clean_up records in the entry. occasionally, failed connect records will be upgraded
* for retry.
*/
process_records(host_entry, host_entry->aaaa_records, host_entry->failed_connection_aaaa_records);
process_records(host_entry, host_entry->a_records, host_entry->failed_connection_a_records);
aws_linked_list_swap_contents(&pending_resolve_copy, &host_entry->pending_resolution_callbacks);
aws_mutex_unlock(&host_entry->entry_lock);
/*
* Clean up resolved addressed outside of the lock
*/
s_clear_address_list(&address_list);
struct aws_host_address address_array[2];
AWS_ZERO_ARRAY(address_array);
/*
* Perform the actual subscriber notifications
*/
while (!aws_linked_list_empty(&pending_resolve_copy)) {
struct aws_linked_list_node *resolution_callback_node = aws_linked_list_pop_front(&pending_resolve_copy);
struct pending_callback *pending_callback =
AWS_CONTAINER_OF(resolution_callback_node, struct pending_callback, node);
struct aws_array_list callback_address_list;
aws_array_list_init_static(&callback_address_list, address_array, 2, sizeof(struct aws_host_address));
aws_mutex_lock(&host_entry->entry_lock);
s_copy_address_into_callback_set(
s_get_lru_address(host_entry, AWS_ADDRESS_RECORD_TYPE_AAAA),
&callback_address_list,
host_entry->host_name);
s_copy_address_into_callback_set(
s_get_lru_address(host_entry, AWS_ADDRESS_RECORD_TYPE_A),
&callback_address_list,
host_entry->host_name);
aws_mutex_unlock(&host_entry->entry_lock);
size_t callback_address_list_size = aws_array_list_length(&callback_address_list);
if (callback_address_list_size > 0) {
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"static, invoking resolution callback for host %s with %d addresses",
aws_string_c_str(host_entry->host_name),
(int)callback_address_list_size);
} else {
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"static, invoking resolution callback for host %s with failure",
aws_string_c_str(host_entry->host_name));
}
if (callback_address_list_size > 0) {
pending_callback->callback(
host_entry->resolver,
host_entry->host_name,
AWS_OP_SUCCESS,
&callback_address_list,
pending_callback->user_data);
} else {
int error_code = (err_code != AWS_ERROR_SUCCESS) ? err_code : AWS_IO_DNS_QUERY_FAILED;
pending_callback->callback(
host_entry->resolver, host_entry->host_name, error_code, NULL, pending_callback->user_data);
}
s_clear_address_list(&callback_address_list);
aws_mem_release(host_entry->allocator, pending_callback);
}
aws_mutex_lock(&host_entry->entry_lock);
++host_entry->resolves_since_last_request;
/*
* A long resolve frequency matched with a connection failure can induce a state of DNS starvation, where
* additional resolution requests go into the queue but since there's no good records and the thread is sleeping
* for a long time, nothing happens.
*
* While we could make the wait predicate also check the queue of requests, there is a worry that a
* host that can't be resolved (user error, dns record removal, etc...) could lead to a "spammy" scenario
* where the thread generates DNS requests extremely quickly, ie, the sleep becomes almost instant.
*
* We'd like to be able to express the wait here as something a bit more complex:
*
* "Wait until either (1) shutdown notice, or (2) a small amount of time has passed and there are pending
* requests, or (3) the resolution interval has passed"
*
* While seemingly complicated, we can do this actually just by chaining two waits:
*
* (1) The first wait is for a short amount of time and only predicates on the shutdown notice
* (2) The second wait is for the remaining frequency interval and predicates on either the shutdown notice
* or a pending resolve request
*
* This leaves us with wait behavior where:
* (1) Shutdown always fully interrupts and immediately causes the thread function to complete
* (2) Absent shutdown, there is always a controllable, non-trivial sleep between resolves
* (3) Starvation is avoided as pending requests can wake the resolver thread independent of resolution
* frequency
*/
aws_condition_variable_wait_for_pred(
&host_entry->entry_signal,
&host_entry->entry_lock,
shutdown_only_wait_time,
s_host_entry_finished_pred,
host_entry);
if (request_interruptible_wait_time > 0) {
aws_condition_variable_wait_for_pred(
&host_entry->entry_signal,
&host_entry->entry_lock,
request_interruptible_wait_time,
s_host_entry_finished_or_pending_request_pred,
host_entry);
}
aws_mutex_unlock(&host_entry->entry_lock);
/*
* This is a bit awkward that we unlock the entry and then relock both the resolver and the entry, but it
* is mandatory that -- in order to maintain the consistent view of the resolver table (entry exist => entry
* is alive and can be queried) -- we have the resolver lock as well before making the decision to remove
* the entry from the table and terminate the thread.
*/
struct default_host_resolver *resolver = host_entry->resolver->impl;
aws_mutex_lock(&resolver->resolver_lock);
aws_mutex_lock(&host_entry->entry_lock);
uint64_t now = s_get_system_time_for_default_resolver(host_entry->resolver);
/*
* The only way we terminate the loop with pending queries is if the resolver itself has no more references
* to it and is going away. In that case, the pending queries will be completed (with failure) by the
* final clean up of this entry.
*/
if (aws_linked_list_empty(&host_entry->pending_resolution_callbacks) &&
host_entry->last_resolve_request_timestamp_ns + max_no_solicitation_interval < now) {
host_entry->state = DRS_SHUTTING_DOWN;
}
keep_going = host_entry->state == DRS_ACTIVE;
if (!keep_going) {
aws_hash_table_remove(&resolver->host_entry_table, host_entry->host_name, NULL, NULL);
}
aws_array_list_swap_contents(&host_entry->new_addresses, &new_address_list);
aws_array_list_swap_contents(&host_entry->expired_addresses, &expired_address_list);
aws_mutex_unlock(&host_entry->entry_lock);
aws_mutex_unlock(&resolver->resolver_lock);
s_clear_address_list(&new_address_list);
s_clear_address_list(&expired_address_list);
}
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"static: Either no requests have been made for an address for %s for the duration "
"of the ttl, or this thread is being forcibly shutdown. Killing thread.",
host_entry->host_name->bytes);
done:
AWS_FATAL_ASSERT(aws_array_list_length(&address_list) == 0);
AWS_FATAL_ASSERT(aws_array_list_length(&new_address_list) == 0);
AWS_FATAL_ASSERT(aws_array_list_length(&expired_address_list) == 0);
aws_array_list_clean_up(&address_list);
aws_array_list_clean_up(&new_address_list);
aws_array_list_clean_up(&expired_address_list);
/* trigger the purge complete callback */
if (host_entry->on_host_purge_complete != NULL) {
host_entry->on_host_purge_complete(host_entry->on_host_purge_complete_user_data);
}
/* please don't fail */
aws_thread_current_at_exit(s_on_host_entry_shutdown_completion, host_entry);
}
static void on_cache_entry_removed_helper(struct aws_host_address_cache_entry *entry) {
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"static: purging address %s for host %s from "
"the cache due to cache eviction or shutdown",
entry->address.address->bytes,
entry->address.host->bytes);
struct aws_allocator *allocator = entry->address.allocator;
aws_host_address_clean_up(&entry->address);
aws_mem_release(allocator, entry);
}
static void on_good_address_entry_removed(void *value) {
struct aws_host_address_cache_entry *entry = value;
if (entry == NULL) {
return;
}
s_copy_address_into_array_list(&entry->address, &entry->entry->expired_addresses);
on_cache_entry_removed_helper(entry);
}
static void on_failed_address_entry_removed(void *value) {
struct aws_host_address_cache_entry *entry = value;
on_cache_entry_removed_helper(entry);
}
/*
* The resolver lock must be held before calling this function
*/
static inline int create_and_init_host_entry(
struct aws_host_resolver *resolver,
const struct aws_string *host_name,
aws_on_host_resolved_result_fn *res,
const struct aws_host_resolution_config *config,
uint64_t timestamp,
void *user_data) {
struct host_entry *new_host_entry = aws_mem_calloc(resolver->allocator, 1, sizeof(struct host_entry));
if (!new_host_entry) {
return AWS_OP_ERR;
}
new_host_entry->resolver = resolver;
new_host_entry->allocator = resolver->allocator;
new_host_entry->last_resolve_request_timestamp_ns = timestamp;
new_host_entry->resolves_since_last_request = 0;
new_host_entry->resolve_frequency_ns =
(config->resolve_frequency_ns != 0) ? config->resolve_frequency_ns : NS_PER_SEC;
new_host_entry->state = DRS_ACTIVE;
bool thread_init = false;
struct pending_callback *pending_callback = NULL;
const struct aws_string *host_string_copy = aws_string_new_from_string(resolver->allocator, host_name);
if (AWS_UNLIKELY(!host_string_copy)) {
goto setup_host_entry_error;
}
new_host_entry->host_name = host_string_copy;
new_host_entry->a_records = aws_cache_new_lru(
new_host_entry->allocator,
aws_hash_string,
aws_hash_callback_string_eq,
NULL,
on_good_address_entry_removed,
config->max_ttl);
if (AWS_UNLIKELY(!new_host_entry->a_records)) {
goto setup_host_entry_error;
}
new_host_entry->aaaa_records = aws_cache_new_lru(
new_host_entry->allocator,
aws_hash_string,
aws_hash_callback_string_eq,
NULL,
on_good_address_entry_removed,
config->max_ttl);
if (AWS_UNLIKELY(!new_host_entry->aaaa_records)) {
goto setup_host_entry_error;
}
new_host_entry->failed_connection_a_records = aws_cache_new_lru(
new_host_entry->allocator,
aws_hash_string,
aws_hash_callback_string_eq,
NULL,
on_failed_address_entry_removed,
config->max_ttl);
if (AWS_UNLIKELY(!new_host_entry->failed_connection_a_records)) {
goto setup_host_entry_error;
}
new_host_entry->failed_connection_aaaa_records = aws_cache_new_lru(
new_host_entry->allocator,
aws_hash_string,
aws_hash_callback_string_eq,
NULL,
on_failed_address_entry_removed,
config->max_ttl);
if (AWS_UNLIKELY(!new_host_entry->failed_connection_aaaa_records)) {
goto setup_host_entry_error;
}
if (aws_array_list_init_dynamic(
&new_host_entry->new_addresses, new_host_entry->allocator, 4, sizeof(struct aws_host_address))) {
goto setup_host_entry_error;
}
if (aws_array_list_init_dynamic(
&new_host_entry->expired_addresses, new_host_entry->allocator, 4, sizeof(struct aws_host_address))) {
goto setup_host_entry_error;
}
aws_linked_list_init(&new_host_entry->pending_resolution_callbacks);
pending_callback = aws_mem_acquire(resolver->allocator, sizeof(struct pending_callback));
if (AWS_UNLIKELY(!pending_callback)) {
goto setup_host_entry_error;
}
/*add the current callback here */
pending_callback->user_data = user_data;
pending_callback->callback = res;
aws_linked_list_push_back(&new_host_entry->pending_resolution_callbacks, &pending_callback->node);
aws_mutex_init(&new_host_entry->entry_lock);
new_host_entry->resolution_config = *config;
aws_condition_variable_init(&new_host_entry->entry_signal);
aws_thread_init(&new_host_entry->resolver_thread, resolver->allocator);
thread_init = true;
struct default_host_resolver *default_host_resolver = resolver->impl;
if (AWS_UNLIKELY(
aws_hash_table_put(&default_host_resolver->host_entry_table, host_string_copy, new_host_entry, NULL))) {
goto setup_host_entry_error;
}
struct aws_thread_options thread_options = *aws_default_thread_options();
thread_options.join_strategy = AWS_TJS_MANAGED;
thread_options.name = aws_byte_cursor_from_c_str("AwsHostResolver"); /* 15 characters is max for Linux */
if (aws_thread_launch(
&new_host_entry->resolver_thread, aws_host_resolver_thread, new_host_entry, &thread_options)) {
goto setup_host_entry_error;
}
++default_host_resolver->pending_host_entry_shutdown_completion_callbacks;
return AWS_OP_SUCCESS;
setup_host_entry_error:
if (thread_init) {
aws_thread_clean_up(&new_host_entry->resolver_thread);
}
// If we registered a callback, clear it. So that we don’t trigger callback and return an error.
if (!aws_linked_list_empty(&new_host_entry->pending_resolution_callbacks)) {
aws_linked_list_remove(&pending_callback->node);
}
s_clean_up_host_entry(new_host_entry);
return AWS_OP_ERR;
}
static int default_resolve_host(
struct aws_host_resolver *resolver,
const struct aws_string *host_name,
aws_on_host_resolved_result_fn *res,
const struct aws_host_resolution_config *config,
void *user_data) {
int result = AWS_OP_SUCCESS;
AWS_LOGF_DEBUG(AWS_LS_IO_DNS, "id=%p: Host resolution requested for %s", (void *)resolver, host_name->bytes);
uint64_t timestamp = s_get_system_time_for_default_resolver(resolver);
struct default_host_resolver *default_host_resolver = resolver->impl;
aws_mutex_lock(&default_host_resolver->resolver_lock);
struct aws_hash_element *element = NULL;
/* we don't care about the error code here, only that the host_entry was found or not. */
aws_hash_table_find(&default_host_resolver->host_entry_table, host_name, &element);
struct host_entry *host_entry = NULL;
if (element != NULL) {
host_entry = element->value;
AWS_FATAL_ASSERT(host_entry != NULL);
}
if (!host_entry) {
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"id=%p: No cached entries found for %s starting new resolver thread.",
(void *)resolver,
host_name->bytes);
result = create_and_init_host_entry(resolver, host_name, res, config, timestamp, user_data);
aws_mutex_unlock(&default_host_resolver->resolver_lock);
return result;
}
aws_mutex_lock(&host_entry->entry_lock);
/*
* We don't need to make any resolver side-affects in the remaining logic and it's impossible for the entry
* to disappear underneath us while holding its lock, so its safe to release the resolver lock and let other
* things query other entries.
*/
aws_mutex_unlock(&default_host_resolver->resolver_lock);
host_entry->last_resolve_request_timestamp_ns = timestamp;
host_entry->resolves_since_last_request = 0;
struct aws_host_address_cache_entry *aaaa_entry = aws_lru_cache_use_lru_element(host_entry->aaaa_records);
struct aws_host_address *aaaa_record = (aaaa_entry != NULL) ? &aaaa_entry->address : NULL;
struct aws_host_address_cache_entry *a_entry = aws_lru_cache_use_lru_element(host_entry->a_records);
struct aws_host_address *a_record = (a_entry != NULL) ? &a_entry->address : NULL;
struct aws_host_address address_array[2];
AWS_ZERO_ARRAY(address_array);
struct aws_array_list callback_address_list;
aws_array_list_init_static(&callback_address_list, address_array, 2, sizeof(struct aws_host_address));
if ((aaaa_record || a_record)) {
AWS_LOGF_DEBUG(
AWS_LS_IO_DNS,
"id=%p: cached entries found for %s returning to caller.",
(void *)resolver,
host_name->bytes);
/* these will all need to be copied so that we don't hold the lock during the callback. */
if (aaaa_record) {
struct aws_host_address aaaa_record_cpy;
aws_host_address_copy(aaaa_record, &aaaa_record_cpy);
aws_array_list_push_back(&callback_address_list, &aaaa_record_cpy);
AWS_LOGF_TRACE(
AWS_LS_IO_DNS,
"id=%p: vending address %s for host %s to caller",
(void *)resolver,
aaaa_record->address->bytes,
host_entry->host_name->bytes);
}
if (a_record) {
struct aws_host_address a_record_cpy;
aws_host_address_copy(a_record, &a_record_cpy);
aws_array_list_push_back(&callback_address_list, &a_record_cpy);
AWS_LOGF_TRACE(
AWS_LS_IO_DNS,
"id=%p: vending address %s for host %s to caller",
(void *)resolver,
a_record->address->bytes,
host_entry->host_name->bytes);
}
aws_mutex_unlock(&host_entry->entry_lock);
/* we don't want to do the callback WHILE we hold the lock someone may reentrantly call us. */
// TODO: Fire the callback asynchronously
res(resolver, host_name, AWS_OP_SUCCESS, &callback_address_list, user_data);
for (size_t i = 0; i < aws_array_list_length(&callback_address_list); ++i) {
struct aws_host_address *address_ptr = NULL;
aws_array_list_get_at_ptr(&callback_address_list, (void **)&address_ptr, i);
aws_host_address_clean_up(address_ptr);
}
aws_array_list_clean_up(&callback_address_list);
return result;
}
struct pending_callback *pending_callback =
aws_mem_acquire(default_host_resolver->allocator, sizeof(struct pending_callback));
if (pending_callback != NULL) {
pending_callback->user_data = user_data;
pending_callback->callback = res;
aws_linked_list_push_back(&host_entry->pending_resolution_callbacks, &pending_callback->node);
/*
* intentionally signal under the lock; similar to the shutdown case, we can't guarantee the resolver
* is still around once the lock is released.
*/
aws_condition_variable_notify_all(&host_entry->entry_signal);
} else {
result = AWS_OP_ERR;
}
aws_mutex_unlock(&host_entry->entry_lock);
return result;
}
static size_t default_get_host_address_count(
struct aws_host_resolver *host_resolver,
const struct aws_string *host_name,
uint32_t flags) {
struct default_host_resolver *default_host_resolver = host_resolver->impl;
size_t address_count = 0;
aws_mutex_lock(&default_host_resolver->resolver_lock);
struct aws_hash_element *element = NULL;
aws_hash_table_find(&default_host_resolver->host_entry_table, host_name, &element);
if (element != NULL) {
struct host_entry *host_entry = element->value;
if (host_entry != NULL) {
aws_mutex_lock(&host_entry->entry_lock);
if ((flags & AWS_GET_HOST_ADDRESS_COUNT_RECORD_TYPE_A) != 0) {
address_count += aws_cache_get_element_count(host_entry->a_records);
}
if ((flags & AWS_GET_HOST_ADDRESS_COUNT_RECORD_TYPE_AAAA) != 0) {
address_count += aws_cache_get_element_count(host_entry->aaaa_records);
}
aws_mutex_unlock(&host_entry->entry_lock);
}
}
aws_mutex_unlock(&default_host_resolver->resolver_lock);
return address_count;
}
static struct aws_host_resolver_vtable s_vtable = {
.purge_cache = s_resolver_purge_cache,
.purge_cache_with_callback = s_resolver_purge_cache_with_callback,
.resolve_host = default_resolve_host,
.record_connection_failure = resolver_record_connection_failure,
.get_host_address_count = default_get_host_address_count,
.destroy = resolver_destroy,
.purge_host_cache = s_resolver_purge_host_cache,
};
static void s_aws_host_resolver_destroy(struct aws_host_resolver *resolver) {
AWS_ASSERT(resolver->vtable && resolver->vtable->destroy);
resolver->vtable->destroy(resolver);
}
struct aws_host_resolver *aws_host_resolver_new_default(
struct aws_allocator *allocator,
const struct aws_host_resolver_default_options *options) {
AWS_FATAL_ASSERT(options != NULL);
AWS_ASSERT(options->el_group);
struct aws_host_resolver *resolver = NULL;
struct default_host_resolver *default_host_resolver = NULL;
if (!aws_mem_acquire_many(
allocator,
2,
&resolver,
sizeof(struct aws_host_resolver),
&default_host_resolver,
sizeof(struct default_host_resolver))) {
return NULL;
}
AWS_ZERO_STRUCT(*resolver);
AWS_ZERO_STRUCT(*default_host_resolver);
AWS_LOGF_INFO(
AWS_LS_IO_DNS,
"id=%p: Initializing default host resolver with %llu max host entries.",
(void *)resolver,
(unsigned long long)options->max_entries);
resolver->vtable = &s_vtable;
resolver->allocator = allocator;
resolver->impl = default_host_resolver;
default_host_resolver->event_loop_group = aws_event_loop_group_acquire(options->el_group);
default_host_resolver->allocator = allocator;
default_host_resolver->pending_host_entry_shutdown_completion_callbacks = 0;
default_host_resolver->state = DRS_ACTIVE;
aws_mutex_init(&default_host_resolver->resolver_lock);
if (aws_hash_table_init(
&default_host_resolver->host_entry_table,
allocator,
options->max_entries,
aws_hash_string,
aws_hash_callback_string_eq,
NULL,
NULL)) {
goto on_error;
}
aws_ref_count_init(&resolver->ref_count, resolver, (aws_simple_completion_callback *)s_aws_host_resolver_destroy);
if (options->shutdown_options != NULL) {
resolver->shutdown_options = *options->shutdown_options;
}
if (options->system_clock_override_fn != NULL) {
default_host_resolver->system_clock_fn = options->system_clock_override_fn;
} else {
default_host_resolver->system_clock_fn = aws_high_res_clock_get_ticks;
}
return resolver;
on_error:
s_cleanup_default_resolver(resolver);
return NULL;
}
struct aws_host_resolver *aws_host_resolver_acquire(struct aws_host_resolver *resolver) {
if (resolver != NULL) {
aws_ref_count_acquire(&resolver->ref_count);
}
return resolver;
}
void aws_host_resolver_release(struct aws_host_resolver *resolver) {
if (resolver != NULL) {
aws_ref_count_release(&resolver->ref_count);
}
}
size_t aws_host_resolver_get_host_address_count(
struct aws_host_resolver *resolver,
const struct aws_string *host_name,
uint32_t flags) {
return resolver->vtable->get_host_address_count(resolver, host_name, flags);
}
struct aws_host_resolution_config aws_host_resolver_init_default_resolution_config(void) {
struct aws_host_resolution_config config = {
.impl = aws_default_dns_resolve,
.max_ttl = AWS_DEFAULT_DNS_TTL,
.impl_data = NULL,
.resolve_frequency_ns = NS_PER_SEC,
};
return config;
}
|