Loading...
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 | /* * Copyright (c) 2021 Apple Inc. All rights reserved. * * @APPLE_LICENSE_HEADER_START@ * * This file contains Original Code and/or Modifications of Original Code * as defined in and that are subject to the Apple Public Source License * Version 2.0 (the 'License'). You may not use this file except in * compliance with the License. Please obtain a copy of the License at * http://www.opensource.apple.com/apsl/ and read it before using this * file. * * The Original Code and all software distributed under the License are * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES, * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY, * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT. * Please see the License for the specific language governing rights and * limitations under the License. * * @APPLE_LICENSE_HEADER_END@ */ #include <climits> #include <sys/types.h> #include <assert.h> #include <limits.h> #include <stdlib.h> #include <string.h> #include <mach-o/loader.h> // mach_o #include "Symbol.h" #include "Misc.h" #include "Algorithm.h" #include "Header.h" // mach_o_writer #include "ChunkBumpAllocator.h" #include "ExportsTrieWriter.h" using mach_o::Error; using mach_o::GenericTrieNode; namespace mach_o { // // MARK: --- Internal trie builder interface --- // // Expensive precondition checks are enabled in unit tests only. // FIXME: Actually implement these again for unit tests #define trieTraceAssert(...) struct VIS_HIDDEN GenericTrieWriterEntry { std::string_view name; std::span<uint8_t> payload; }; using WriterEntry = GenericTrieWriterEntry; struct VIS_HIDDEN Edge { std::string_view partialString; GenericTrieNode* child; Edge(const std::string_view& s, GenericTrieNode* n) : partialString(s), child(n) { } ~Edge() { } }; struct VIS_HIDDEN GenericTrieNode { std::string_view cummulativeString; std::vector<Edge> children; std::span<uint8_t> terminalPayload; uint32_t trieOffset = 0; uint32_t trieSize = 0; GenericTrieNode(const std::string_view& s) : cummulativeString(s) {} ~GenericTrieNode() = default; void updateOffset(uint32_t& curOffset); void writeToStream(std::span<uint8_t>& bytes) const; }; using Node = GenericTrieNode; struct VIS_HIDDEN SubtreeRoot { Node* parent=nullptr; std::span<const WriterEntry> entries; }; struct VIS_HIDDEN TrieBuilder { ChunkBumpAllocator allocator; std::vector<SubtreeRoot>* roots; TrieBuilder(ChunkBumpAllocatorZone& zone, std::vector<SubtreeRoot>* roots): allocator(zone), roots(roots) {} Error buildSubtree(Node& parentNode, uint32_t offset, std::span<const WriterEntry> entries); void addTerminalNode(Node& parentNode, const WriterEntry&); }; struct VIS_HIDDEN NodeWriter { // root nodes of subtrees that are written concurrently std::vector<const Node*> subtreeRoots; // list of nodes that are written written separately, without their children std::vector<const Node*> standaloneNodes; std::span<uint8_t> bytes; void collectRecursive(const Node* node, size_t depth); void write(const Node& node); void writeRecursive(const Node& node); }; // // MARK: --- GenericTrieWriter methods --- // // construct from an already built trie GenericTrieWriter::GenericTrieWriter() : GenericTrie(nullptr, 0), _trieSize(0), _allocatorZone(ChunkBumpAllocatorZone::make()) { } static void write_uleb128(uint64_t value, std::span<uint8_t>& out) { uint8_t* outByte = out.data(); uint8_t byte; do { byte = value & 0x7F; value &= ~0x7F; if ( value != 0 ) byte |= 0x80; *outByte = byte; ++outByte; value = value >> 7; } while ( byte >= 0x80 ); out = out.subspan(outByte - out.data()); } static void write_string(const std::string_view& str, std::span<uint8_t>& out) { std::copy(str.begin(), str.end(), out.begin()); *(out.data() + str.size()) = '\0'; out = out.subspan(str.size() + 1); } // find the number of leading elements in \a entries that have the same character \a ch at \a offset. // this assumes that the entries are sorted static uint32_t binSearchNumEntriesWithChar(std::span<const WriterEntry> entries, uint32_t offset, char ch) { if ( entries.empty() ) assert(false && "cant divide empty list"); if ( entries.size() == 1 ) return 1; auto cmpOp = [&](const WriterEntry& entry) { return offset < entry.name.size() && entry.name[offset] == ch; }; // optimize case where last entry has the same character if ( cmpOp(entries.back()) ) { return (uint32_t)entries.size(); } auto low = entries.begin(); // optimize case where second entry has a different character if ( !cmpOp(*(low + 1)) ) return (uint32_t)1; auto high = entries.end() - 1; while ( true ) { auto middle = low + ((high - low) / 2); if ( cmpOp(*middle) ) { low = middle; } else { high = middle; } if ( (high - low) == 1 ) return (uint32_t)std::distance(entries.begin(), low) + 1; } } // find first offset that has a different character in any of the entries, starting at an input offset static Error findFirstDifferentChar(uint32_t& inOutOffset, std::span<const WriterEntry> entries) { if ( entries.size() < 2 ) return Error::none(); // entries are sorted, so we only need to compare first and last uint32_t diffOffset = inOutOffset; while ( entries.front().name.size() > diffOffset && entries.back().name.size() > diffOffset && entries.front().name[diffOffset] == entries.back().name[diffOffset] ) ++diffOffset; inOutOffset = diffOffset; // no duplicates, but there's more than one entry so there are duplicates if ( diffOffset == entries.back().name.size() ) { const WriterEntry& newEntry = entries.back(); char cstr[entries.back().name.size()+2]; memcpy(cstr, newEntry.name.data(), newEntry.name.size()); cstr[newEntry.name.size()] = '\0'; return Error("duplicate symbol '%s'", (const char*)cstr); // cast is to work around va_list aliasing issue } return Error::none(); } // Find the number of leading entries from \a entries span that all have the same character at \a offset. // If all entries in \a entries have the same character at \a offset then \a edgeBreak // will be set to the number of all entries and offset will point to the last common character. static Error nextEdgeBreak(uint32_t& offset, uint32_t& edgeBreak, std::span<const WriterEntry> entries) { assert(entries.size() >= 1); auto& entry = entries.front(); assert(offset < entry.name.size()); uint32_t diffOffset = offset; if ( Error err = findFirstDifferentChar(diffOffset, entries) ) return err; if ( diffOffset != offset ) { // common characters found in all entries, update offset to point to the last common character offset = diffOffset - 1; edgeBreak = (uint32_t)entries.size(); return Error::none(); } edgeBreak = binSearchNumEntriesWithChar(entries, offset, entry.name[offset]); return Error::none(); } void TrieBuilder::addTerminalNode(Node& parentNode, const WriterEntry& entry) { std::string_view name = entry.name; trieTraceAssert(name.starts_with(parentNode.cummulativeString)); assert(name.size() >= parentNode.cummulativeString.size()); std::string_view tail = name.substr(parentNode.cummulativeString.size()); if ( tail.empty() ) { assert(parentNode.terminalPayload.empty() && "duplicate node should have been handled before calling addTerminalNode"); parentNode.terminalPayload = entry.payload; } else { Node* newNode = allocator.allocate<Node>(); new (newNode) Node(name); newNode->terminalPayload = entry.payload; Edge newEdge(tail, newNode); parentNode.children.push_back(newEdge); } } // Trie building algorithm is based on the requirement that the input entries are sorted. // Sorted input allows to use binary search to quickly find the number of nodes in a subtree // and compare first/last entries from the subrange to determine the longest common partial string for an edge. // The problem of creating the trie is defined recursively - given a parent node, a list of // entries and a name offset create a subtree with optimal partial strings in edges. Thanks to the sorted input // we know that all characters are the same up to, but not including, the current offset. // For example, given symbols `foo`, `fop`, `read`, a root node and offset 0 - use binary search to find // the first symbol whose character at offset 0 is different from the character of the first symbol. // Character of the first entry at offset 0 is `f` and next different entry is `read`, so `foo` and `fop` entries // will form a subtree. Edge from a parent node to the subtree should cover the longest common prefix of all entries // in the subtree. Again, thanks to the input being sorted it's sufficient to compare character of the first and last // subtree entries, because if they're the same then so are all entries in between. // In this approach we know at all times how many more terminal nodes will be created in subtree, which makes it // easy to parallelize. Once a threshold of remaining entries in a subtree is reached the root of the subtree // is placed in a list of root nodes and after initial iteration all subtrees will be built concurrently. // // // Below is a visualization of the algorithm with a bit more complex example: // ``` // ┌───────────────────┐ ┌──────────────────┐ ┌──────────────────┐ ┌──────────────┐ // │ offset: 0 │ │ offset: 6 │ │ offset: 7 │ │ offset: 8 │ // └───────────────────┘ └──────────────────┘ └──────────────────┘ └──────────────┘ // // ┌───────────────────┐ // │ ┌──────────────┐ │ // │ │ aaaaaaaa │ │ // │ └──────────────┘ │ // │ ┌──────────────┐ │ // │ │ aaaaaaab │ │ ┌──────────────────┐ ┌──────────────────┐ ┌──────────────┐ // │ └──────────────┘ │ │ ┌──────────────┐ │ │ ┌──────────────┐ │ ┌───▶│ aaaaaaaa │ // │ ┌──────────────┐ │ │ │ aaaaaaaa │ │ │ │ aaaaaaaa │ │ a └──────────────┘ // │ │ aaaaaaba │ │ │ └──────────────┘ │ a │ └──────────────┘ │ │ // │ └──────────────┘ │ │ │ ┌────▶│ ┌──────────────┐ │────┤ // │ ┌─────┐ │ │ ┌──────────────┐ │ │ │ │ aaaaaaab │ │ │ ┌──────────────┐ // │ │ foo │ │────────────▶│ │ aaaaaaab │ │────┤ │ └──────────────┘ │ b───▶│ aaaaaaab │ // │ └─────┘ │ aaaaaa │ └──────────────┘ │ │ └──────────────────┘ └──────────────┘ // │ ┌─────┐ │ │ │ │ // │ │ fop │ │ │ ┌──────────────┐ │ │ ┌──────────────────┐ // │ └─────┘ │ │ │ aaaaaaba │ │ └────▶│ aaaaaaba │ // │ ┌─────┐ │ │ └──────────────┘ │ b └──────────────────┘ // │ │read │ │ └──────────────────┘ // │ └─────┘ │ // │ ┌─────┐ │ // │ │write│ │ // │ └─────┘ │ // └───────────────────┘ // // // ┌───────────────────┐ ┌───────────┐ ┌─────────┐ // │ offset: 0 │ │ offset: 2 │ │offset: 3│ // └───────────────────┘ └───────────┘ └─────────┘ // // ┌───────────────────┐ // │ ┌─────┐ │ // │ │ foo │ │ // │ └─────┘ │ ┌───────────┐ // │ ┌─────┐ │ │ ┌─────┐ │ o ┌─────┐ // │ │ fop │ │ │ │ foo │ │ ┌───────▶│ foo │ // │ └─────┘ │ fo │ └─────┘ │ │ └─────┘ // │ ┌─────┐ │────────────▶│ ┌─────┐ │────────┤ // │ │read │ │ │ │ fop │ │ │ ┌─────┐ // │ └─────┘ │ │ └─────┘ │ └───────▶│ fop │ // │ ┌─────┐ │ └───────────┘ p └─────┘ // │ │write│ │ // │ └─────┘ │ // └───────────────────┘ // // ┌───────────────────┐ ┌─────────┐ // │ offset: 0 │ │offset: 4│ // └───────────────────┘ └─────────┘ // ┌───────────────────┐ // │ ┌─────┐ │ // │ │read │ │ // │ └─────┘ │ read ┌─────┐ // │ ┌─────┐ │─────────────▶│read │ // │ │write│ │ └─────┘ // │ └─────┘ │ // └───────────────────┘ // // ┌───────────────────┐ // │ offset: 0 │ // └───────────────────┘ // // ┌───────────────────┐ // │ │ // │ ┌─────┐ │ // │ │write│ │ write ┌─────┐ // │ └─────┘ │─────────────▶│write│ // │ │ └─────┘ // │ │ // └───────────────────┘ // ``` Error TrieBuilder::buildSubtree(Node& parentNode, uint32_t offset, std::span<const WriterEntry> entries) { while ( !entries.empty() ) { // one entry left, add the terminal if ( entries.size() == 1 ) { addTerminalNode(parentNode, entries.front()); return Error::none(); } // offset equal to the current offset, there's no free characters to make // an edge, so this must the current root terminal payload if ( entries.front().name.size() == offset ) { addTerminalNode(parentNode, entries.front()); entries = entries.subspan(1); // another entry also with length equal to offset, it's a duplicate then // otherwise it wouldn't be in the same edge const WriterEntry& newEntry = entries.front(); if ( newEntry.name.size() == offset ) { char cstr[newEntry.name.size()+2]; memcpy(cstr, newEntry.name.data(), newEntry.name.size()); cstr[newEntry.name.size()] = '\0'; return Error("duplicate symbol '%s'", (const char*)cstr); // cast is to work around va_list aliasing issue } } uint32_t edgeBreak; if ( Error err = nextEdgeBreak(offset, edgeBreak, entries) ) return err; auto edgeNodes = std::span(entries).subspan(0, edgeBreak); entries = entries.subspan(edgeNodes.size()); if ( edgeNodes.size() == 1 ) { addTerminalNode(parentNode, edgeNodes.front()); continue; } // multiple entries with the same character at the current offset, // so make an edge with the common characters // first, find all common characters among the nodes in this edge this is to make the edge // partial string as long as possible uint32_t commonLen = offset; if ( Error err = findFirstDifferentChar(commonLen, edgeNodes) ) return err; if ( commonLen == offset ) assert(false && "edgeNodes come from the size determined by nextEdgeBreak, there must be at least one common character"); std::string_view firstEntryName = edgeNodes.front().name; // first entry will have size longer than the common length or equal, if it's // equal then it will become a terminal entry in the newly created node assert(firstEntryName.size() >= commonLen); std::string_view cummulativeStr = firstEntryName.substr(0, commonLen); trieTraceAssert(cummulativeStr.starts_with(parentNode.cummulativeString)); std::string_view edgePartialStr= cummulativeStr.substr(parentNode.cummulativeString.size()); assert(edgePartialStr.size() >= 1); Node* child = allocator.allocate<Node>(); new (child) Node(cummulativeStr); Edge edge(edgePartialStr, child); parentNode.children.push_back(edge); // place this node and its items in the roots vector if requested and match a threshold // they'll be processed concurrenctly later if ( roots && edgeNodes.size() < 0x4000 ) { roots->push_back({child, edgeNodes}); } else { if ( Error err = buildSubtree(*child, commonLen, edgeNodes) ) return err; } } return Error::none(); } #define DUMP_NODES 0 #if DUMP_NODES static void dumpNodes(const Node* node, size_t depth=0) { for ( size_t i = 0; i < depth; ++i ) { fprintf(stderr, " "); } fprintf(stderr, "%.*s (%d)\n", (int)node->cummulativeString.size(), node->cummulativeString.data(), !node->terminalEntry.payload.empty()); for ( auto& edge : node->children ) { dumpNodes(edge.child, depth + 1); } } #endif const uint8_t* GenericTrieWriter::bytes(size_t& size) { size = _trieEnd - _trieStart; return _trieStart; } static void updateOffsetPostorder(Node* node, uint32_t& curOffset) { for ( Edge& e : node->children ) { updateOffsetPostorder(e.child, curOffset); } node->updateOffset(curOffset); } void GenericTrieWriter::buildNodes(std::span<const WriterEntry> entries) { // build nodes _rootNode = _allocatorZone.makeAllocator().allocate<Node>(); new (_rootNode) Node(""); std::vector<SubtreeRoot> roots; roots.reserve(entries.size() / 0x4000); // build initial set of nodes, collecting some subtree roots along the way // subtries will be then build concurrently if ( !entries.empty() ) { TrieBuilder builder(_allocatorZone, &roots); if ( Error err = builder.buildSubtree(*_rootNode, 0, entries) ) { _buildError = std::move(err); return; } } // build subtrees in parallel mapReduce(std::span(roots), 1, ^(size_t, Error& chunkErr, std::span<SubtreeRoot> current) { // create a builder per a subtree root, without using a roots vector, so all the nodes will be built TrieBuilder builder(_allocatorZone, nullptr); for ( const SubtreeRoot& root : current ) { if ( Error err = builder.buildSubtree(*root.parent, (uint32_t)root.parent->cummulativeString.size(), root.entries) ) chunkErr = std::move(err); } }, ^(std::span<Error> errors) { for ( Error& err : errors ) { if ( err ) { _buildError = std::move(err); return; } } }); #if DUMP_NODES dumpNodes(&rootNode); #endif uint32_t curOffset = 0; { // set a dummy large trie offset for all chidren of the root node to ensure // enough space is reserved for their actual offset, so that the root // node size is stable for ( Edge& e : _rootNode->children ) { e.child->trieOffset = UINT_MAX; } _rootNode->updateOffset(curOffset); } // now that the size of the root node is known, offsets can // be computed recursively in a single iteration through a postorder traversal for ( Edge& e : _rootNode->children ) updateOffsetPostorder(e.child, curOffset); _trieSize = curOffset; if ( uint32_t pad = _trieSize % 8; pad != 0 ) _trieSize += 8 - pad; } void NodeWriter::collectRecursive(const Node* node, size_t depth) { standaloneNodes.push_back(node); if ( (depth + 1) > 4 ) { subtreeRoots.reserve(subtreeRoots.size() + node->children.size()); std::transform(node->children.begin(), node->children.end(), std::back_inserter(subtreeRoots), [](const Edge& e) { return e.child; }); } else { for ( const Edge& e : node->children ) { collectRecursive(e.child, depth + 1); } } } void NodeWriter::write(const Node& node) { std::span<uint8_t> nodeChunk = bytes.subspan(node.trieOffset); assert(nodeChunk.size() >= node.trieSize); nodeChunk = nodeChunk.subspan(0, node.trieSize); node.writeToStream(nodeChunk); } void NodeWriter::writeRecursive(const Node& node) { write(node); for ( const Edge& e : node.children ) { writeRecursive(*e.child); } } void GenericTrieWriter::writeTrieBytes(std::span<uint8_t> bytes) { // set up trie buffer _trieStart = bytes.data(); _trieEnd = bytes.end().base(); assert(_rootNode != nullptr); NodeWriter writer; // reserve some initial space for nodes writer.subtreeRoots.reserve(0x1000); writer.standaloneNodes.reserve(0x1000); writer.bytes = bytes; writer.collectRecursive(_rootNode, 0); // write subtrees dispatchForEach(std::span(writer.subtreeRoots), 1, [&writer](size_t, const Node* node) { writer.writeRecursive(*node); }); // write standalone nodes dispatchForEach(std::span(writer.standaloneNodes), 64, [&writer](size_t, const Node* node) { writer.write(*node); }); } // byte for terminal node size in bytes, or 0x00 if not terminal node // teminal node (uleb128 flags, uleb128 addr [uleb128 other]) // byte for child node count // each child: zero terminated substring, uleb128 node offset void Node::updateOffset(uint32_t& curOffset) { trieSize = 1; // length of node payload info when there is no payload (non-terminal) if ( !terminalPayload.empty() ) { // in terminal nodes, size is uleb128 encoded, so we include that in calculation trieSize = (uint32_t)terminalPayload.size(); trieSize += uleb128_size(trieSize); } // add children ++trieSize; // byte for count of chidren for ( Edge& edge : this->children ) { trieSize += edge.partialString.size() + 1 + uleb128_size(edge.child->trieOffset); } trieOffset = curOffset; curOffset += trieSize; } void Node::writeToStream(std::span<uint8_t>& bytes) const { if ( !terminalPayload.empty() ) { std::span<const uint8_t> payload = terminalPayload; write_uleb128(payload.size(), bytes); std::copy(payload.begin(), payload.end(), bytes.begin()); bytes = bytes.subspan(payload.size()); } else { // no terminal uleb128 of zero is one byte of zero *bytes.data() = 0; bytes = bytes.subspan(1); } // write number of children *bytes.data() = children.size(); bytes = bytes.subspan(1); // write each child for ( const Edge& e : children ) { write_string(e.partialString, bytes); write_uleb128(e.child->trieOffset, bytes); } } template<typename WriterEntryGetter> static inline std::vector<WriterEntry> buildWriterEntries(ChunkBumpAllocatorZone& allocatorZone, size_t entriesCount, bool needsSort, WriterEntryGetter get) { std::vector<WriterEntry> allEntries; allEntries.resize(entriesCount); // create generic trie's writer entries from higher-level entries used by one of the trie's subclasses // this requires that the getter implementation is thread-safe const size_t elementsPerChunk = 0x4000; mapReduce(std::span(allEntries), elementsPerChunk, ^(size_t chunkIndex, int&, std::span<WriterEntry> entries) { ChunkBumpAllocator allocator = allocatorZone.makeAllocator(); size_t startIndex = chunkIndex * elementsPerChunk; for ( size_t i = 0; i < entries.size(); ++i ) entries[i] = get(startIndex + i, allocator); }); if ( needsSort ) { std::sort(allEntries.begin(), allEntries.end(), [](const WriterEntry& lhs, const WriterEntry& rhs) { return lhs.name < rhs.name; }); } else { trieTraceAssert(std::is_sorted(allEntries.begin(), allEntries.end(), [](const WriterEntry& lhs, const WriterEntry& rhs) { return lhs.name < rhs.name; })); } return allEntries; } // // MARK: --- ExportsTrieWriter methods --- // ExportsTrieWriter::operator ExportsTrie() const { return { _trieStart, _trieSize }; } bool ExportsTrieWriter::hasExportedSymbol(const char* symbolName, Symbol& symbol) const { ExportsTrie trie = *this; return trie.hasExportedSymbol(symbolName, symbol); } void ExportsTrieWriter::forEachExportedSymbol(void (^callback)(const Symbol& symbol, bool& stop)) const { ExportsTrie trie = *this; return trie.forEachExportedSymbol(callback); } Error ExportsTrieWriter::valid(uint64_t baseAddr, uint64_t maxVmOffset) const { if ( _buildError.hasError() ) return Error("%s", _buildError.message()); ExportsTrie trie = *this; return trie.valid(baseAddr, maxVmOffset); } // generic trie builder struct VIS_HIDDEN Export { std::string_view name; uint64_t offset=0; uint64_t flags=0; uint64_t other=0; std::string_view importName; }; static WriterEntry exportToEntry(const Export& exportInfo, ChunkBumpAllocator& allocator) { // encode exportInfo as uleb128s std::span<uint8_t> payload; assert(!exportInfo.name.empty() && "empty export info"); if ( exportInfo.flags & EXPORT_SYMBOL_FLAGS_REEXPORT ) { std::string_view importName = exportInfo.importName; // optimize case where re-export does not change name to just have a trailing empty string if ( importName == exportInfo.name ) importName = ""; // nodes with re-export info: size, flags, ordinal, string size_t size = uleb128_size(exportInfo.flags) + uleb128_size(exportInfo.other) + importName.size() + 1; payload = allocator.allocate(size); std::span<uint8_t> temp = payload; write_uleb128(exportInfo.flags, temp); write_uleb128(exportInfo.other, temp); write_string(importName, temp); assert(temp.size() == 0); } else if ( exportInfo.flags & EXPORT_SYMBOL_FLAGS_STUB_AND_RESOLVER ) { // nodes with stub and resolver: size, flags, address, other size_t size = uleb128_size(exportInfo.flags) + uleb128_size(exportInfo.offset) + uleb128_size(exportInfo.other); payload = allocator.allocate(size); std::span<uint8_t> temp = payload; write_uleb128(exportInfo.flags, temp); write_uleb128(exportInfo.offset, temp); write_uleb128(exportInfo.other, temp); assert(temp.size() == 0); } else if ( exportInfo.flags & EXPORT_SYMBOL_FLAGS_FUNCTION_VARIANT ) { // nodes with default and tableIndex: size, flags, address, other size_t size = uleb128_size(exportInfo.flags) + uleb128_size(exportInfo.offset) + uleb128_size(exportInfo.other); payload = allocator.allocate(size); std::span<uint8_t> temp = payload; write_uleb128(exportInfo.flags, temp); write_uleb128(exportInfo.offset, temp); write_uleb128(exportInfo.other, temp); assert(temp.size() == 0); } else { // nodes with export info: size, flags, address size_t size = uleb128_size(exportInfo.flags) + uleb128_size(exportInfo.offset); payload = allocator.allocate(size); std::span<uint8_t> temp = payload; write_uleb128(exportInfo.flags, temp); write_uleb128(exportInfo.offset, temp); assert(temp.size() == 0); } WriterEntry entry; entry.name = exportInfo.name; entry.payload = payload; return entry; } ExportsTrieWriter::ExportsTrieWriter(std::span<const Symbol> exports, bool writeBytes, bool needsSort) : GenericTrieWriter() { std::vector<WriterEntry> entries = buildWriterEntries(_allocatorZone, exports.size(), needsSort, [exports](size_t index, ChunkBumpAllocator& allocator) { const Symbol& sym = exports[index]; Export exp; const char* importName = nullptr; int libOrdinal = 0; bool weakImport = false; uint64_t resolverStub = 0; uint64_t absAddr = 0; uint32_t fvtIndex = 0; assert((sym.scope() == Symbol::Scope::global) || (sym.scope() == Symbol::Scope::globalNeverStrip)); assert(!sym.isUndefined(libOrdinal, weakImport)); exp.name = sym.name(); if ( sym.isThreadLocal() ) { exp.offset = sym.implOffset(); exp.flags = EXPORT_SYMBOL_FLAGS_KIND_THREAD_LOCAL; if ( sym.isWeakDef() ) exp.flags |= EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION; } else if ( sym.isAbsolute(absAddr) ) { exp.offset = absAddr; exp.flags = EXPORT_SYMBOL_FLAGS_KIND_ABSOLUTE; } else if ( sym.isReExport(libOrdinal, importName) ) { exp.flags = EXPORT_SYMBOL_FLAGS_REEXPORT; exp.other = libOrdinal; exp.importName = importName; if ( sym.isWeakDef() ) exp.flags |= EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION; } else if ( sym.isDynamicResolver(resolverStub) ) { exp.offset = resolverStub; exp.flags = EXPORT_SYMBOL_FLAGS_STUB_AND_RESOLVER; exp.other = sym.implOffset(); } else if ( sym.isFunctionVariant(fvtIndex) ) { exp.offset = sym.implOffset(); exp.flags = EXPORT_SYMBOL_FLAGS_FUNCTION_VARIANT; exp.other = fvtIndex; } else if ( sym.isWeakDef() ) { exp.offset = sym.implOffset(); exp.flags = EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION; } else { exp.offset = sym.implOffset(); exp.flags = 0; } return exportToEntry(exp, allocator); }); buildNodes(entries); if ( _buildError.hasError() || !writeBytes ) return; _trieBytes.resize(_trieSize); writeTrieBytes(_trieBytes); } // build a trie from an existing trie, but filter out some entries ExportsTrieWriter::ExportsTrieWriter(const ExportsTrie& inputExportsTrie, bool (^remove)(const Symbol& sym)) : GenericTrieWriter() { ChunkBumpAllocatorZone zone = ChunkBumpAllocatorZone::make(); __block ChunkBumpAllocator bumpAllocator(zone); __block std::vector<Symbol> keptSymbols; inputExportsTrie.forEachExportedSymbol(^(const Symbol& symbol, bool& stop) { if ( remove(symbol) ) return; // The CString in symbol is ephemeral. // Make a copy with a long term string Symbol newSymbol = symbol; const char* ephemString = symbol.name().c_str(); std::span<uint8_t> buffer = bumpAllocator.allocate(symbol.name().size()+1); strcpy((char*)buffer.data(), ephemString); newSymbol.setName((char*)buffer.data()); keptSymbols.push_back(newSymbol); }); // call constructor that takes span of symbols *this = ExportsTrieWriter(keptSymbols); } // // MARK: --- DylibsPathTrie methods --- // DylibsPathTrieWriter::DylibsPathTrieWriter(std::span<const mach_o::DylibsPathTrie::DylibAndIndex> dylibs, bool needsSort) : GenericTrieWriter() { std::vector<WriterEntry> entries = buildWriterEntries(_allocatorZone, dylibs.size(), /* needsSort */ needsSort, [dylibs](size_t index, ChunkBumpAllocator& allocator) { const mach_o::DylibsPathTrie::DylibAndIndex& info = dylibs[index]; // payload for DylibsPathTrie is just uleb128 encoded dylib index size_t size = uleb128_size(info.index); std::span<uint8_t> payload = allocator.allocate(size); std::span<uint8_t> temp = payload; write_uleb128(info.index, temp); assert(temp.size() == 0); WriterEntry entry; entry.name = info.path; entry.payload = payload; return entry; }); buildNodes(entries); if ( _buildError.hasError() ) return; _trieBytes.resize(_trieSize); writeTrieBytes(_trieBytes); } DylibsPathTrieWriter::operator DylibsPathTrie() const { return { _trieStart, _trieSize }; } bool DylibsPathTrieWriter::hasPath(const char* path, uint32_t& dylibIndex) const { DylibsPathTrie trie = *this; return trie.hasPath(path, dylibIndex); } void DylibsPathTrieWriter::forEachDylibPath(void (^callback)(const mach_o::DylibsPathTrie::DylibAndIndex& info, bool& stop)) const { DylibsPathTrie trie = *this; return trie.forEachDylibPath(callback); } } // namespace mach_o |