1 //===- Writer.cpp ---------------------------------------------------------===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 9 #include "Writer.h" 10 #include "Config.h" 11 #include "InputChunks.h" 12 #include "InputElement.h" 13 #include "MapFile.h" 14 #include "OutputSections.h" 15 #include "OutputSegment.h" 16 #include "Relocations.h" 17 #include "SymbolTable.h" 18 #include "SyntheticSections.h" 19 #include "WriterUtils.h" 20 #include "lld/Common/ErrorHandler.h" 21 #include "lld/Common/Memory.h" 22 #include "lld/Common/Strings.h" 23 #include "llvm/ADT/DenseSet.h" 24 #include "llvm/ADT/SmallSet.h" 25 #include "llvm/ADT/SmallVector.h" 26 #include "llvm/ADT/StringMap.h" 27 #include "llvm/BinaryFormat/Wasm.h" 28 #include "llvm/BinaryFormat/WasmTraits.h" 29 #include "llvm/Support/FileOutputBuffer.h" 30 #include "llvm/Support/Format.h" 31 #include "llvm/Support/FormatVariadic.h" 32 #include "llvm/Support/LEB128.h" 33 #include "llvm/Support/Parallel.h" 34 35 #include <cstdarg> 36 #include <map> 37 38 #define DEBUG_TYPE "lld" 39 40 using namespace llvm; 41 using namespace llvm::wasm; 42 43 namespace lld { 44 namespace wasm { 45 static constexpr int stackAlignment = 16; 46 47 namespace { 48 49 // The writer writes a SymbolTable result to a file. 50 class Writer { 51 public: 52 void run(); 53 54 private: 55 void openFile(); 56 57 bool needsPassiveInitialization(const OutputSegment *segment); 58 bool hasPassiveInitializedSegments(); 59 60 void createSyntheticInitFunctions(); 61 void createInitMemoryFunction(); 62 void createStartFunction(); 63 void createApplyDataRelocationsFunction(); 64 void createApplyGlobalRelocationsFunction(); 65 void createCallCtorsFunction(); 66 void createInitTLSFunction(); 67 void createCommandExportWrappers(); 68 void createCommandExportWrapper(uint32_t functionIndex, DefinedFunction *f); 69 70 void assignIndexes(); 71 void populateSymtab(); 72 void populateProducers(); 73 void populateTargetFeatures(); 74 void calculateInitFunctions(); 75 void calculateImports(); 76 void calculateExports(); 77 void calculateCustomSections(); 78 void calculateTypes(); 79 void createOutputSegments(); 80 OutputSegment *createOutputSegment(StringRef name); 81 void combineOutputSegments(); 82 void layoutMemory(); 83 void createHeader(); 84 85 void addSection(OutputSection *sec); 86 87 void addSections(); 88 89 void createCustomSections(); 90 void createSyntheticSections(); 91 void createSyntheticSectionsPostLayout(); 92 void finalizeSections(); 93 94 // Custom sections 95 void createRelocSections(); 96 97 void writeHeader(); 98 void writeSections(); 99 100 uint64_t fileSize = 0; 101 102 std::vector<WasmInitEntry> initFunctions; 103 llvm::StringMap<std::vector<InputSection *>> customSectionMapping; 104 105 // Stable storage for command export wrapper function name strings. 106 std::list<std::string> commandExportWrapperNames; 107 108 // Elements that are used to construct the final output 109 std::string header; 110 std::vector<OutputSection *> outputSections; 111 112 std::unique_ptr<FileOutputBuffer> buffer; 113 114 std::vector<OutputSegment *> segments; 115 llvm::SmallDenseMap<StringRef, OutputSegment *> segmentMap; 116 }; 117 118 } // anonymous namespace 119 120 void Writer::calculateCustomSections() { 121 log("calculateCustomSections"); 122 bool stripDebug = config->stripDebug || config->stripAll; 123 for (ObjFile *file : symtab->objectFiles) { 124 for (InputSection *section : file->customSections) { 125 // Exclude COMDAT sections that are not selected for inclusion 126 if (section->discarded) 127 continue; 128 StringRef name = section->getName(); 129 // These custom sections are known the linker and synthesized rather than 130 // blindly copied. 131 if (name == "linking" || name == "name" || name == "producers" || 132 name == "target_features" || name.startswith("reloc.")) 133 continue; 134 // These custom sections are generated by `clang -fembed-bitcode`. 135 // These are used by the rust toolchain to ship LTO data along with 136 // compiled object code, but they don't want this included in the linker 137 // output. 138 if (name == ".llvmbc" || name == ".llvmcmd") 139 continue; 140 // Strip debug section in that option was specified. 141 if (stripDebug && name.startswith(".debug_")) 142 continue; 143 // Otherwise include custom sections by default and concatenate their 144 // contents. 145 customSectionMapping[name].push_back(section); 146 } 147 } 148 } 149 150 void Writer::createCustomSections() { 151 log("createCustomSections"); 152 for (auto &pair : customSectionMapping) { 153 StringRef name = pair.first(); 154 LLVM_DEBUG(dbgs() << "createCustomSection: " << name << "\n"); 155 156 OutputSection *sec = make<CustomSection>(std::string(name), pair.second); 157 if (config->relocatable || config->emitRelocs) { 158 auto *sym = make<OutputSectionSymbol>(sec); 159 out.linkingSec->addToSymtab(sym); 160 sec->sectionSym = sym; 161 } 162 addSection(sec); 163 } 164 } 165 166 // Create relocations sections in the final output. 167 // These are only created when relocatable output is requested. 168 void Writer::createRelocSections() { 169 log("createRelocSections"); 170 // Don't use iterator here since we are adding to OutputSection 171 size_t origSize = outputSections.size(); 172 for (size_t i = 0; i < origSize; i++) { 173 LLVM_DEBUG(dbgs() << "check section " << i << "\n"); 174 OutputSection *sec = outputSections[i]; 175 176 // Count the number of needed sections. 177 uint32_t count = sec->getNumRelocations(); 178 if (!count) 179 continue; 180 181 StringRef name; 182 if (sec->type == WASM_SEC_DATA) 183 name = "reloc.DATA"; 184 else if (sec->type == WASM_SEC_CODE) 185 name = "reloc.CODE"; 186 else if (sec->type == WASM_SEC_CUSTOM) 187 name = saver.save("reloc." + sec->name); 188 else 189 llvm_unreachable( 190 "relocations only supported for code, data, or custom sections"); 191 192 addSection(make<RelocSection>(name, sec)); 193 } 194 } 195 196 void Writer::populateProducers() { 197 for (ObjFile *file : symtab->objectFiles) { 198 const WasmProducerInfo &info = file->getWasmObj()->getProducerInfo(); 199 out.producersSec->addInfo(info); 200 } 201 } 202 203 void Writer::writeHeader() { 204 memcpy(buffer->getBufferStart(), header.data(), header.size()); 205 } 206 207 void Writer::writeSections() { 208 uint8_t *buf = buffer->getBufferStart(); 209 parallelForEach(outputSections, [buf](OutputSection *s) { 210 assert(s->isNeeded()); 211 s->writeTo(buf); 212 }); 213 } 214 215 static void setGlobalPtr(DefinedGlobal *g, uint64_t memoryPtr) { 216 g->global->setPointerValue(memoryPtr); 217 } 218 219 // Fix the memory layout of the output binary. This assigns memory offsets 220 // to each of the input data sections as well as the explicit stack region. 221 // The default memory layout is as follows, from low to high. 222 // 223 // - initialized data (starting at Config->globalBase) 224 // - BSS data (not currently implemented in llvm) 225 // - explicit stack (Config->ZStackSize) 226 // - heap start / unallocated 227 // 228 // The --stack-first option means that stack is placed before any static data. 229 // This can be useful since it means that stack overflow traps immediately 230 // rather than overwriting global data, but also increases code size since all 231 // static data loads and stores requires larger offsets. 232 void Writer::layoutMemory() { 233 uint64_t memoryPtr = 0; 234 235 auto placeStack = [&]() { 236 if (config->relocatable || config->isPic) 237 return; 238 memoryPtr = alignTo(memoryPtr, stackAlignment); 239 if (config->zStackSize != alignTo(config->zStackSize, stackAlignment)) 240 error("stack size must be " + Twine(stackAlignment) + "-byte aligned"); 241 log("mem: stack size = " + Twine(config->zStackSize)); 242 log("mem: stack base = " + Twine(memoryPtr)); 243 memoryPtr += config->zStackSize; 244 setGlobalPtr(cast<DefinedGlobal>(WasmSym::stackPointer), memoryPtr); 245 log("mem: stack top = " + Twine(memoryPtr)); 246 }; 247 248 if (config->stackFirst) { 249 placeStack(); 250 } else { 251 memoryPtr = config->globalBase; 252 log("mem: global base = " + Twine(config->globalBase)); 253 } 254 255 if (WasmSym::globalBase) 256 WasmSym::globalBase->setVA(memoryPtr); 257 258 uint64_t dataStart = memoryPtr; 259 260 // Arbitrarily set __dso_handle handle to point to the start of the data 261 // segments. 262 if (WasmSym::dsoHandle) 263 WasmSym::dsoHandle->setVA(dataStart); 264 265 out.dylinkSec->memAlign = 0; 266 for (OutputSegment *seg : segments) { 267 out.dylinkSec->memAlign = std::max(out.dylinkSec->memAlign, seg->alignment); 268 memoryPtr = alignTo(memoryPtr, 1ULL << seg->alignment); 269 seg->startVA = memoryPtr; 270 log(formatv("mem: {0,-15} offset={1,-8} size={2,-8} align={3}", seg->name, 271 memoryPtr, seg->size, seg->alignment)); 272 273 if (!config->relocatable && seg->isTLS()) { 274 if (config->sharedMemory) { 275 auto *tlsSize = cast<DefinedGlobal>(WasmSym::tlsSize); 276 setGlobalPtr(tlsSize, seg->size); 277 278 auto *tlsAlign = cast<DefinedGlobal>(WasmSym::tlsAlign); 279 setGlobalPtr(tlsAlign, int64_t{1} << seg->alignment); 280 } else { 281 auto *tlsBase = cast<DefinedGlobal>(WasmSym::tlsBase); 282 setGlobalPtr(tlsBase, memoryPtr); 283 } 284 } 285 286 memoryPtr += seg->size; 287 } 288 289 // Make space for the memory initialization flag 290 if (config->sharedMemory && hasPassiveInitializedSegments()) { 291 memoryPtr = alignTo(memoryPtr, 4); 292 WasmSym::initMemoryFlag = symtab->addSyntheticDataSymbol( 293 "__wasm_init_memory_flag", WASM_SYMBOL_VISIBILITY_HIDDEN); 294 WasmSym::initMemoryFlag->markLive(); 295 WasmSym::initMemoryFlag->setVA(memoryPtr); 296 log(formatv("mem: {0,-15} offset={1,-8} size={2,-8} align={3}", 297 "__wasm_init_memory_flag", memoryPtr, 4, 4)); 298 memoryPtr += 4; 299 } 300 301 if (WasmSym::dataEnd) 302 WasmSym::dataEnd->setVA(memoryPtr); 303 304 uint64_t staticDataSize = memoryPtr - dataStart; 305 log("mem: static data = " + Twine(staticDataSize)); 306 if (config->isPic) 307 out.dylinkSec->memSize = staticDataSize; 308 309 if (!config->stackFirst) 310 placeStack(); 311 312 if (WasmSym::heapBase) { 313 // Set `__heap_base` to directly follow the end of the stack or global data. 314 // The fact that this comes last means that a malloc/brk implementation 315 // can grow the heap at runtime. 316 log("mem: heap base = " + Twine(memoryPtr)); 317 WasmSym::heapBase->setVA(memoryPtr); 318 } 319 320 uint64_t maxMemorySetting = 1ULL 321 << (config->is64.getValueOr(false) ? 48 : 32); 322 323 if (config->initialMemory != 0) { 324 if (config->initialMemory != alignTo(config->initialMemory, WasmPageSize)) 325 error("initial memory must be " + Twine(WasmPageSize) + "-byte aligned"); 326 if (memoryPtr > config->initialMemory) 327 error("initial memory too small, " + Twine(memoryPtr) + " bytes needed"); 328 if (config->initialMemory > maxMemorySetting) 329 error("initial memory too large, cannot be greater than " + 330 Twine(maxMemorySetting)); 331 memoryPtr = config->initialMemory; 332 } 333 out.memorySec->numMemoryPages = 334 alignTo(memoryPtr, WasmPageSize) / WasmPageSize; 335 log("mem: total pages = " + Twine(out.memorySec->numMemoryPages)); 336 337 if (config->maxMemory != 0) { 338 if (config->maxMemory != alignTo(config->maxMemory, WasmPageSize)) 339 error("maximum memory must be " + Twine(WasmPageSize) + "-byte aligned"); 340 if (memoryPtr > config->maxMemory) 341 error("maximum memory too small, " + Twine(memoryPtr) + " bytes needed"); 342 if (config->maxMemory > maxMemorySetting) 343 error("maximum memory too large, cannot be greater than " + 344 Twine(maxMemorySetting)); 345 } 346 347 // Check max if explicitly supplied or required by shared memory 348 if (config->maxMemory != 0 || config->sharedMemory) { 349 uint64_t max = config->maxMemory; 350 if (max == 0) { 351 // If no maxMemory config was supplied but we are building with 352 // shared memory, we need to pick a sensible upper limit. 353 if (config->isPic) 354 max = maxMemorySetting; 355 else 356 max = alignTo(memoryPtr, WasmPageSize); 357 } 358 out.memorySec->maxMemoryPages = max / WasmPageSize; 359 log("mem: max pages = " + Twine(out.memorySec->maxMemoryPages)); 360 } 361 } 362 363 void Writer::addSection(OutputSection *sec) { 364 if (!sec->isNeeded()) 365 return; 366 log("addSection: " + toString(*sec)); 367 sec->sectionIndex = outputSections.size(); 368 outputSections.push_back(sec); 369 } 370 371 // If a section name is valid as a C identifier (which is rare because of 372 // the leading '.'), linkers are expected to define __start_<secname> and 373 // __stop_<secname> symbols. They are at beginning and end of the section, 374 // respectively. This is not requested by the ELF standard, but GNU ld and 375 // gold provide the feature, and used by many programs. 376 static void addStartStopSymbols(const OutputSegment *seg) { 377 StringRef name = seg->name; 378 if (!isValidCIdentifier(name)) 379 return; 380 LLVM_DEBUG(dbgs() << "addStartStopSymbols: " << name << "\n"); 381 uint64_t start = seg->startVA; 382 uint64_t stop = start + seg->size; 383 symtab->addOptionalDataSymbol(saver.save("__start_" + name), start); 384 symtab->addOptionalDataSymbol(saver.save("__stop_" + name), stop); 385 } 386 387 void Writer::addSections() { 388 addSection(out.dylinkSec); 389 addSection(out.typeSec); 390 addSection(out.importSec); 391 addSection(out.functionSec); 392 addSection(out.tableSec); 393 addSection(out.memorySec); 394 addSection(out.eventSec); 395 addSection(out.globalSec); 396 addSection(out.exportSec); 397 addSection(out.startSec); 398 addSection(out.elemSec); 399 addSection(out.dataCountSec); 400 401 addSection(make<CodeSection>(out.functionSec->inputFunctions)); 402 addSection(make<DataSection>(segments)); 403 404 createCustomSections(); 405 406 addSection(out.linkingSec); 407 if (config->emitRelocs || config->relocatable) { 408 createRelocSections(); 409 } 410 411 addSection(out.nameSec); 412 addSection(out.producersSec); 413 addSection(out.targetFeaturesSec); 414 } 415 416 void Writer::finalizeSections() { 417 for (OutputSection *s : outputSections) { 418 s->setOffset(fileSize); 419 s->finalizeContents(); 420 fileSize += s->getSize(); 421 } 422 } 423 424 void Writer::populateTargetFeatures() { 425 StringMap<std::string> used; 426 StringMap<std::string> required; 427 StringMap<std::string> disallowed; 428 SmallSet<std::string, 8> &allowed = out.targetFeaturesSec->features; 429 bool tlsUsed = false; 430 431 // Only infer used features if user did not specify features 432 bool inferFeatures = !config->features.hasValue(); 433 434 if (!inferFeatures) { 435 auto &explicitFeatures = config->features.getValue(); 436 allowed.insert(explicitFeatures.begin(), explicitFeatures.end()); 437 if (!config->checkFeatures) 438 return; 439 } 440 441 // Find the sets of used, required, and disallowed features 442 for (ObjFile *file : symtab->objectFiles) { 443 StringRef fileName(file->getName()); 444 for (auto &feature : file->getWasmObj()->getTargetFeatures()) { 445 switch (feature.Prefix) { 446 case WASM_FEATURE_PREFIX_USED: 447 used.insert({feature.Name, std::string(fileName)}); 448 break; 449 case WASM_FEATURE_PREFIX_REQUIRED: 450 used.insert({feature.Name, std::string(fileName)}); 451 required.insert({feature.Name, std::string(fileName)}); 452 break; 453 case WASM_FEATURE_PREFIX_DISALLOWED: 454 disallowed.insert({feature.Name, std::string(fileName)}); 455 break; 456 default: 457 error("Unrecognized feature policy prefix " + 458 std::to_string(feature.Prefix)); 459 } 460 } 461 462 // Find TLS data segments 463 auto isTLS = [](InputSegment *segment) { 464 StringRef name = segment->getName(); 465 return segment->live && 466 (name.startswith(".tdata") || name.startswith(".tbss")); 467 }; 468 tlsUsed = tlsUsed || 469 std::any_of(file->segments.begin(), file->segments.end(), isTLS); 470 } 471 472 if (inferFeatures) 473 for (const auto &key : used.keys()) 474 allowed.insert(std::string(key)); 475 476 if (!config->checkFeatures) 477 return; 478 479 if (!config->relocatable && allowed.count("mutable-globals") == 0) { 480 for (const Symbol *sym : out.importSec->importedSymbols) { 481 if (auto *global = dyn_cast<GlobalSymbol>(sym)) { 482 if (global->getGlobalType()->Mutable) { 483 error(Twine("mutable global imported but 'mutable-globals' feature " 484 "not present in inputs: `") + 485 toString(*sym) + "`. Use --no-check-features to suppress."); 486 } 487 } 488 } 489 for (const Symbol *sym : out.exportSec->exportedSymbols) { 490 if (isa<GlobalSymbol>(sym)) { 491 error(Twine("mutable global exported but 'mutable-globals' feature " 492 "not present in inputs: `") + 493 toString(*sym) + "`. Use --no-check-features to suppress."); 494 } 495 } 496 } 497 498 if (config->sharedMemory) { 499 if (disallowed.count("shared-mem")) 500 error("--shared-memory is disallowed by " + disallowed["shared-mem"] + 501 " because it was not compiled with 'atomics' or 'bulk-memory' " 502 "features."); 503 504 for (auto feature : {"atomics", "bulk-memory"}) 505 if (!allowed.count(feature)) 506 error(StringRef("'") + feature + 507 "' feature must be used in order to use shared memory"); 508 } 509 510 if (tlsUsed) { 511 for (auto feature : {"atomics", "bulk-memory"}) 512 if (!allowed.count(feature)) 513 error(StringRef("'") + feature + 514 "' feature must be used in order to use thread-local storage"); 515 } 516 517 // Validate that used features are allowed in output 518 if (!inferFeatures) { 519 for (auto &feature : used.keys()) { 520 if (!allowed.count(std::string(feature))) 521 error(Twine("Target feature '") + feature + "' used by " + 522 used[feature] + " is not allowed."); 523 } 524 } 525 526 // Validate the required and disallowed constraints for each file 527 for (ObjFile *file : symtab->objectFiles) { 528 StringRef fileName(file->getName()); 529 SmallSet<std::string, 8> objectFeatures; 530 for (auto &feature : file->getWasmObj()->getTargetFeatures()) { 531 if (feature.Prefix == WASM_FEATURE_PREFIX_DISALLOWED) 532 continue; 533 objectFeatures.insert(feature.Name); 534 if (disallowed.count(feature.Name)) 535 error(Twine("Target feature '") + feature.Name + "' used in " + 536 fileName + " is disallowed by " + disallowed[feature.Name] + 537 ". Use --no-check-features to suppress."); 538 } 539 for (auto &feature : required.keys()) { 540 if (!objectFeatures.count(std::string(feature))) 541 error(Twine("Missing target feature '") + feature + "' in " + fileName + 542 ", required by " + required[feature] + 543 ". Use --no-check-features to suppress."); 544 } 545 } 546 } 547 548 static bool shouldImport(Symbol *sym) { 549 if (!sym->isUndefined()) 550 return false; 551 if (sym->isWeak() && !config->relocatable) 552 return false; 553 if (!sym->isLive()) 554 return false; 555 if (!sym->isUsedInRegularObj) 556 return false; 557 558 // We don't generate imports for data symbols. They however can be imported 559 // as GOT entries. 560 if (isa<DataSymbol>(sym)) 561 return false; 562 563 if ((config->isPic || config->relocatable) || 564 config->unresolvedSymbols == UnresolvedPolicy::ImportFuncs) 565 return true; 566 if (config->allowUndefinedSymbols.count(sym->getName()) != 0) 567 return true; 568 if (auto *g = dyn_cast<UndefinedGlobal>(sym)) 569 return g->importName.hasValue(); 570 if (auto *f = dyn_cast<UndefinedFunction>(sym)) 571 return f->importName.hasValue(); 572 if (auto *t = dyn_cast<UndefinedTable>(sym)) 573 return t->importName.hasValue(); 574 575 return false; 576 } 577 578 void Writer::calculateImports() { 579 // Some inputs require that the indirect function table be assigned to table 580 // number 0, so if it is present and is an import, allocate it before any 581 // other tables. 582 if (WasmSym::indirectFunctionTable && 583 shouldImport(WasmSym::indirectFunctionTable)) 584 out.importSec->addImport(WasmSym::indirectFunctionTable); 585 586 for (Symbol *sym : symtab->getSymbols()) { 587 if (!shouldImport(sym)) 588 continue; 589 if (sym == WasmSym::indirectFunctionTable) 590 continue; 591 LLVM_DEBUG(dbgs() << "import: " << sym->getName() << "\n"); 592 out.importSec->addImport(sym); 593 } 594 } 595 596 void Writer::calculateExports() { 597 if (config->relocatable) 598 return; 599 600 if (!config->relocatable && !config->importMemory) 601 out.exportSec->exports.push_back( 602 WasmExport{"memory", WASM_EXTERNAL_MEMORY, 0}); 603 604 unsigned globalIndex = 605 out.importSec->getNumImportedGlobals() + out.globalSec->numGlobals(); 606 607 for (Symbol *sym : symtab->getSymbols()) { 608 if (!sym->isExported()) 609 continue; 610 if (!sym->isLive()) 611 continue; 612 613 StringRef name = sym->getName(); 614 WasmExport export_; 615 if (auto *f = dyn_cast<DefinedFunction>(sym)) { 616 if (Optional<StringRef> exportName = f->function->getExportName()) { 617 name = *exportName; 618 } 619 export_ = {name, WASM_EXTERNAL_FUNCTION, f->getFunctionIndex()}; 620 } else if (auto *g = dyn_cast<DefinedGlobal>(sym)) { 621 if (g->getGlobalType()->Mutable && !g->getFile() && !g->forceExport) { 622 // Avoid exporting mutable globals are linker synthesized (e.g. 623 // __stack_pointer or __tls_base) unless they are explicitly exported 624 // from the command line. 625 // Without this check `--export-all` would cause any program using the 626 // stack pointer to export a mutable global even if none of the input 627 // files were built with the `mutable-globals` feature. 628 continue; 629 } 630 export_ = {name, WASM_EXTERNAL_GLOBAL, g->getGlobalIndex()}; 631 } else if (auto *e = dyn_cast<DefinedEvent>(sym)) { 632 export_ = {name, WASM_EXTERNAL_EVENT, e->getEventIndex()}; 633 } else if (auto *d = dyn_cast<DefinedData>(sym)) { 634 if (d->segment && d->segment->isTLS()) { 635 // We can't currenly export TLS data symbols. 636 if (sym->isExportedExplicit()) 637 error("TLS symbols cannot yet be exported: `" + toString(*sym) + "`"); 638 continue; 639 } 640 out.globalSec->dataAddressGlobals.push_back(d); 641 export_ = {name, WASM_EXTERNAL_GLOBAL, globalIndex++}; 642 } else { 643 auto *t = cast<DefinedTable>(sym); 644 export_ = {name, WASM_EXTERNAL_TABLE, t->getTableNumber()}; 645 } 646 647 LLVM_DEBUG(dbgs() << "Export: " << name << "\n"); 648 out.exportSec->exports.push_back(export_); 649 out.exportSec->exportedSymbols.push_back(sym); 650 } 651 } 652 653 void Writer::populateSymtab() { 654 if (!config->relocatable && !config->emitRelocs) 655 return; 656 657 for (Symbol *sym : symtab->getSymbols()) 658 if (sym->isUsedInRegularObj && sym->isLive()) 659 out.linkingSec->addToSymtab(sym); 660 661 for (ObjFile *file : symtab->objectFiles) { 662 LLVM_DEBUG(dbgs() << "Local symtab entries: " << file->getName() << "\n"); 663 for (Symbol *sym : file->getSymbols()) 664 if (sym->isLocal() && !isa<SectionSymbol>(sym) && sym->isLive()) 665 out.linkingSec->addToSymtab(sym); 666 } 667 } 668 669 void Writer::calculateTypes() { 670 // The output type section is the union of the following sets: 671 // 1. Any signature used in the TYPE relocation 672 // 2. The signatures of all imported functions 673 // 3. The signatures of all defined functions 674 // 4. The signatures of all imported events 675 // 5. The signatures of all defined events 676 677 for (ObjFile *file : symtab->objectFiles) { 678 ArrayRef<WasmSignature> types = file->getWasmObj()->types(); 679 for (uint32_t i = 0; i < types.size(); i++) 680 if (file->typeIsUsed[i]) 681 file->typeMap[i] = out.typeSec->registerType(types[i]); 682 } 683 684 for (const Symbol *sym : out.importSec->importedSymbols) { 685 if (auto *f = dyn_cast<FunctionSymbol>(sym)) 686 out.typeSec->registerType(*f->signature); 687 else if (auto *e = dyn_cast<EventSymbol>(sym)) 688 out.typeSec->registerType(*e->signature); 689 } 690 691 for (const InputFunction *f : out.functionSec->inputFunctions) 692 out.typeSec->registerType(f->signature); 693 694 for (const InputEvent *e : out.eventSec->inputEvents) 695 out.typeSec->registerType(e->signature); 696 } 697 698 // In a command-style link, create a wrapper for each exported symbol 699 // which calls the constructors and destructors. 700 void Writer::createCommandExportWrappers() { 701 // This logic doesn't currently support Emscripten-style PIC mode. 702 assert(!config->isPic); 703 704 // If there are no ctors and there's no libc `__wasm_call_dtors` to 705 // call, don't wrap the exports. 706 if (initFunctions.empty() && WasmSym::callDtors == NULL) 707 return; 708 709 std::vector<DefinedFunction *> toWrap; 710 711 for (Symbol *sym : symtab->getSymbols()) 712 if (sym->isExported()) 713 if (auto *f = dyn_cast<DefinedFunction>(sym)) 714 toWrap.push_back(f); 715 716 for (auto *f : toWrap) { 717 auto funcNameStr = (f->getName() + ".command_export").str(); 718 commandExportWrapperNames.push_back(funcNameStr); 719 const std::string &funcName = commandExportWrapperNames.back(); 720 721 auto func = make<SyntheticFunction>(*f->getSignature(), funcName); 722 if (f->function->getExportName().hasValue()) 723 func->setExportName(f->function->getExportName()->str()); 724 else 725 func->setExportName(f->getName().str()); 726 727 DefinedFunction *def = 728 symtab->addSyntheticFunction(funcName, f->flags, func); 729 def->markLive(); 730 731 def->flags |= WASM_SYMBOL_EXPORTED; 732 def->flags &= ~WASM_SYMBOL_VISIBILITY_HIDDEN; 733 def->forceExport = f->forceExport; 734 735 f->flags |= WASM_SYMBOL_VISIBILITY_HIDDEN; 736 f->flags &= ~WASM_SYMBOL_EXPORTED; 737 f->forceExport = false; 738 739 out.functionSec->addFunction(func); 740 741 createCommandExportWrapper(f->getFunctionIndex(), def); 742 } 743 } 744 745 static void finalizeIndirectFunctionTable() { 746 if (!WasmSym::indirectFunctionTable) 747 return; 748 749 if (shouldImport(WasmSym::indirectFunctionTable) && 750 !WasmSym::indirectFunctionTable->hasTableNumber()) { 751 // Processing -Bsymbolic relocations resulted in a late requirement that the 752 // indirect function table be present, and we are running in --import-table 753 // mode. Add the table now to the imports section. Otherwise it will be 754 // added to the tables section later in assignIndexes. 755 out.importSec->addImport(WasmSym::indirectFunctionTable); 756 } 757 758 uint32_t tableSize = config->tableBase + out.elemSec->numEntries(); 759 WasmLimits limits = {0, tableSize, 0}; 760 if (WasmSym::indirectFunctionTable->isDefined() && !config->growableTable) { 761 limits.Flags |= WASM_LIMITS_FLAG_HAS_MAX; 762 limits.Maximum = limits.Minimum; 763 } 764 WasmSym::indirectFunctionTable->setLimits(limits); 765 } 766 767 static void scanRelocations() { 768 for (ObjFile *file : symtab->objectFiles) { 769 LLVM_DEBUG(dbgs() << "scanRelocations: " << file->getName() << "\n"); 770 for (InputChunk *chunk : file->functions) 771 scanRelocations(chunk); 772 for (InputChunk *chunk : file->segments) 773 scanRelocations(chunk); 774 for (auto &p : file->customSections) 775 scanRelocations(p); 776 } 777 } 778 779 void Writer::assignIndexes() { 780 // Seal the import section, since other index spaces such as function and 781 // global are effected by the number of imports. 782 out.importSec->seal(); 783 784 for (InputFunction *func : symtab->syntheticFunctions) 785 out.functionSec->addFunction(func); 786 787 for (ObjFile *file : symtab->objectFiles) { 788 LLVM_DEBUG(dbgs() << "Functions: " << file->getName() << "\n"); 789 for (InputFunction *func : file->functions) 790 out.functionSec->addFunction(func); 791 } 792 793 for (InputGlobal *global : symtab->syntheticGlobals) 794 out.globalSec->addGlobal(global); 795 796 for (ObjFile *file : symtab->objectFiles) { 797 LLVM_DEBUG(dbgs() << "Globals: " << file->getName() << "\n"); 798 for (InputGlobal *global : file->globals) 799 out.globalSec->addGlobal(global); 800 } 801 802 for (ObjFile *file : symtab->objectFiles) { 803 LLVM_DEBUG(dbgs() << "Events: " << file->getName() << "\n"); 804 for (InputEvent *event : file->events) 805 out.eventSec->addEvent(event); 806 } 807 808 for (ObjFile *file : symtab->objectFiles) { 809 LLVM_DEBUG(dbgs() << "Tables: " << file->getName() << "\n"); 810 for (InputTable *table : file->tables) 811 out.tableSec->addTable(table); 812 } 813 814 for (InputTable *table : symtab->syntheticTables) 815 out.tableSec->addTable(table); 816 817 out.globalSec->assignIndexes(); 818 out.tableSec->assignIndexes(); 819 } 820 821 static StringRef getOutputDataSegmentName(StringRef name) { 822 // We only support one thread-local segment, so we must merge the segments 823 // despite --no-merge-data-segments. 824 // We also need to merge .tbss into .tdata so they share the same offsets. 825 if (name.startswith(".tdata") || name.startswith(".tbss")) 826 return ".tdata"; 827 if (!config->mergeDataSegments) 828 return name; 829 if (name.startswith(".text.")) 830 return ".text"; 831 if (name.startswith(".data.")) 832 return ".data"; 833 if (name.startswith(".bss.")) 834 return ".bss"; 835 if (name.startswith(".rodata.")) 836 return ".rodata"; 837 return name; 838 } 839 840 OutputSegment *Writer::createOutputSegment(StringRef name) { 841 LLVM_DEBUG(dbgs() << "new segment: " << name << "\n"); 842 OutputSegment *s = make<OutputSegment>(name); 843 if (config->sharedMemory) 844 s->initFlags = WASM_DATA_SEGMENT_IS_PASSIVE; 845 // Exported memories are guaranteed to be zero-initialized, so no need 846 // to emit data segments for bss sections. 847 // TODO: consider initializing bss sections with memory.fill 848 // instructions when memory is imported and bulk-memory is available. 849 if (!config->importMemory && !config->relocatable && name.startswith(".bss")) 850 s->isBss = true; 851 segments.push_back(s); 852 return s; 853 } 854 855 void Writer::createOutputSegments() { 856 for (ObjFile *file : symtab->objectFiles) { 857 for (InputSegment *segment : file->segments) { 858 if (!segment->live) 859 continue; 860 StringRef name = getOutputDataSegmentName(segment->getName()); 861 OutputSegment *s = nullptr; 862 // When running in relocatable mode we can't merge segments that are part 863 // of comdat groups since the ultimate linker needs to be able exclude or 864 // include them individually. 865 if (config->relocatable && !segment->getComdatName().empty()) { 866 s = createOutputSegment(name); 867 } else { 868 if (segmentMap.count(name) == 0) 869 segmentMap[name] = createOutputSegment(name); 870 s = segmentMap[name]; 871 } 872 s->addInputSegment(segment); 873 } 874 } 875 876 // Sort segments by type, placing .bss last 877 std::stable_sort(segments.begin(), segments.end(), 878 [](const OutputSegment *a, const OutputSegment *b) { 879 auto order = [](StringRef name) { 880 return StringSwitch<int>(name) 881 .StartsWith(".tdata", 0) 882 .StartsWith(".rodata", 1) 883 .StartsWith(".data", 2) 884 .StartsWith(".bss", 4) 885 .Default(3); 886 }; 887 return order(a->name) < order(b->name); 888 }); 889 890 for (size_t i = 0; i < segments.size(); ++i) 891 segments[i]->index = i; 892 893 // Merge MergeInputSections into a single MergeSyntheticSection. 894 LLVM_DEBUG(dbgs() << "-- finalize input semgments\n"); 895 for (OutputSegment *seg : segments) 896 seg->finalizeInputSegments(); 897 } 898 899 void Writer::combineOutputSegments() { 900 // With PIC code we currently only support a single data segment since 901 // we only have a single __memory_base to use as our base address. 902 // This pass combines all non-TLS data segments into a single .data 903 // segment. 904 // This restructions can be relaxed once we have extended constant 905 // expressions available: 906 // https://github.com/WebAssembly/extended-const 907 assert(config->isPic); 908 if (segments.size() <= 1) 909 return; 910 OutputSegment *combined = nullptr; 911 std::vector<OutputSegment *> new_segments; 912 for (OutputSegment *s : segments) { 913 if (s->isTLS()) { 914 new_segments.push_back(s); 915 } else { 916 if (!combined) { 917 LLVM_DEBUG(dbgs() << "created combined output segment: .data\n"); 918 combined = make<OutputSegment>(".data"); 919 combined->startVA = s->startVA; 920 if (config->sharedMemory) 921 combined->initFlags = WASM_DATA_SEGMENT_IS_PASSIVE; 922 } 923 bool first = true; 924 for (InputSegment *inSeg : s->inputSegments) { 925 if (first) 926 inSeg->alignment = std::max(inSeg->alignment, s->alignment); 927 first = false; 928 #ifndef NDEBUG 929 uint64_t oldVA = inSeg->getVA(); 930 #endif 931 combined->addInputSegment(inSeg); 932 #ifndef NDEBUG 933 uint64_t newVA = inSeg->getVA(); 934 LLVM_DEBUG(dbgs() << "added input segment. name=" << inSeg->getName() 935 << " oldVA=" << oldVA << " newVA=" << newVA << "\n"); 936 assert(oldVA == newVA); 937 #endif 938 } 939 } 940 } 941 if (combined) { 942 new_segments.push_back(combined); 943 segments = new_segments; 944 for (size_t i = 0; i < segments.size(); ++i) 945 segments[i]->index = i; 946 } 947 } 948 949 static void createFunction(DefinedFunction *func, StringRef bodyContent) { 950 std::string functionBody; 951 { 952 raw_string_ostream os(functionBody); 953 writeUleb128(os, bodyContent.size(), "function size"); 954 os << bodyContent; 955 } 956 ArrayRef<uint8_t> body = arrayRefFromStringRef(saver.save(functionBody)); 957 cast<SyntheticFunction>(func->function)->setBody(body); 958 } 959 960 bool Writer::needsPassiveInitialization(const OutputSegment *segment) { 961 return segment->initFlags & WASM_DATA_SEGMENT_IS_PASSIVE && 962 !segment->isTLS() && !segment->isBss; 963 } 964 965 bool Writer::hasPassiveInitializedSegments() { 966 return std::find_if(segments.begin(), segments.end(), 967 [this](const OutputSegment *s) { 968 return this->needsPassiveInitialization(s); 969 }) != segments.end(); 970 } 971 972 void Writer::createSyntheticInitFunctions() { 973 if (config->relocatable) 974 return; 975 976 static WasmSignature nullSignature = {{}, {}}; 977 978 // Passive segments are used to avoid memory being reinitialized on each 979 // thread's instantiation. These passive segments are initialized and 980 // dropped in __wasm_init_memory, which is registered as the start function 981 if (config->sharedMemory && hasPassiveInitializedSegments()) { 982 WasmSym::initMemory = symtab->addSyntheticFunction( 983 "__wasm_init_memory", WASM_SYMBOL_VISIBILITY_HIDDEN, 984 make<SyntheticFunction>(nullSignature, "__wasm_init_memory")); 985 WasmSym::initMemory->markLive(); 986 } 987 988 if (config->isPic) { 989 // For PIC code we create synthetic functions that apply relocations. 990 // These get called from __wasm_call_ctors before the user-level 991 // constructors. 992 WasmSym::applyDataRelocs = symtab->addSyntheticFunction( 993 "__wasm_apply_data_relocs", WASM_SYMBOL_VISIBILITY_HIDDEN, 994 make<SyntheticFunction>(nullSignature, "__wasm_apply_data_relocs")); 995 WasmSym::applyDataRelocs->markLive(); 996 997 if (out.globalSec->needsRelocations()) { 998 WasmSym::applyGlobalRelocs = symtab->addSyntheticFunction( 999 "__wasm_apply_global_relocs", WASM_SYMBOL_VISIBILITY_HIDDEN, 1000 make<SyntheticFunction>(nullSignature, "__wasm_apply_global_relocs")); 1001 WasmSym::applyGlobalRelocs->markLive(); 1002 } 1003 } 1004 1005 if (WasmSym::applyGlobalRelocs && WasmSym::initMemory) { 1006 WasmSym::startFunction = symtab->addSyntheticFunction( 1007 "__wasm_start", WASM_SYMBOL_VISIBILITY_HIDDEN, 1008 make<SyntheticFunction>(nullSignature, "__wasm_start")); 1009 WasmSym::startFunction->markLive(); 1010 } 1011 } 1012 1013 void Writer::createInitMemoryFunction() { 1014 LLVM_DEBUG(dbgs() << "createInitMemoryFunction\n"); 1015 assert(WasmSym::initMemory); 1016 assert(WasmSym::initMemoryFlag); 1017 assert(hasPassiveInitializedSegments()); 1018 uint64_t flagAddress = WasmSym::initMemoryFlag->getVA(); 1019 bool is64 = config->is64.getValueOr(false); 1020 std::string bodyContent; 1021 { 1022 raw_string_ostream os(bodyContent); 1023 // Initialize memory in a thread-safe manner. The thread that successfully 1024 // increments the flag from 0 to 1 is is responsible for performing the 1025 // memory initialization. Other threads go sleep on the flag until the 1026 // first thread finishing initializing memory, increments the flag to 2, 1027 // and wakes all the other threads. Once the flag has been set to 2, 1028 // subsequently started threads will skip the sleep. All threads 1029 // unconditionally drop their passive data segments once memory has been 1030 // initialized. The generated code is as follows: 1031 // 1032 // (func $__wasm_init_memory 1033 // (if 1034 // (i32.atomic.rmw.cmpxchg align=2 offset=0 1035 // (i32.const $__init_memory_flag) 1036 // (i32.const 0) 1037 // (i32.const 1) 1038 // ) 1039 // (then 1040 // (drop 1041 // (i32.atomic.wait align=2 offset=0 1042 // (i32.const $__init_memory_flag) 1043 // (i32.const 1) 1044 // (i32.const -1) 1045 // ) 1046 // ) 1047 // ) 1048 // (else 1049 // ( ... initialize data segments ... ) 1050 // (i32.atomic.store align=2 offset=0 1051 // (i32.const $__init_memory_flag) 1052 // (i32.const 2) 1053 // ) 1054 // (drop 1055 // (i32.atomic.notify align=2 offset=0 1056 // (i32.const $__init_memory_flag) 1057 // (i32.const -1u) 1058 // ) 1059 // ) 1060 // ) 1061 // ) 1062 // ( ... drop data segments ... ) 1063 // ) 1064 // 1065 // When we are building with PIC, calculate the flag location using: 1066 // 1067 // (global.get $__memory_base) 1068 // (i32.const $__init_memory_flag) 1069 // (i32.const 1) 1070 1071 // With PIC code we cache the flag address in local 0 1072 if (config->isPic) { 1073 writeUleb128(os, 1, "num local decls"); 1074 writeUleb128(os, 1, "local count"); 1075 writeU8(os, is64 ? WASM_TYPE_I64 : WASM_TYPE_I32, "address type"); 1076 writeU8(os, WASM_OPCODE_GLOBAL_GET, "GLOBAL_GET"); 1077 writeUleb128(os, WasmSym::memoryBase->getGlobalIndex(), "memory_base"); 1078 writePtrConst(os, flagAddress, is64, "flag address"); 1079 writeU8(os, WASM_OPCODE_I32_ADD, "add"); 1080 writeU8(os, WASM_OPCODE_LOCAL_SET, "local.set"); 1081 writeUleb128(os, 0, "local 0"); 1082 } else { 1083 writeUleb128(os, 0, "num locals"); 1084 } 1085 1086 auto writeGetFlagAddress = [&]() { 1087 if (config->isPic) { 1088 writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get"); 1089 writeUleb128(os, 0, "local 0"); 1090 } else { 1091 writePtrConst(os, flagAddress, is64, "flag address"); 1092 } 1093 }; 1094 1095 // Atomically check whether this is the main thread. 1096 writeGetFlagAddress(); 1097 writeI32Const(os, 0, "expected flag value"); 1098 writeI32Const(os, 1, "flag value"); 1099 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 1100 writeUleb128(os, WASM_OPCODE_I32_RMW_CMPXCHG, "i32.atomic.rmw.cmpxchg"); 1101 writeMemArg(os, 2, 0); 1102 writeU8(os, WASM_OPCODE_IF, "IF"); 1103 writeU8(os, WASM_TYPE_NORESULT, "blocktype"); 1104 1105 // Did not increment 0, so wait for main thread to initialize memory 1106 writeGetFlagAddress(); 1107 writeI32Const(os, 1, "expected flag value"); 1108 writeI64Const(os, -1, "timeout"); 1109 1110 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 1111 writeUleb128(os, WASM_OPCODE_I32_ATOMIC_WAIT, "i32.atomic.wait"); 1112 writeMemArg(os, 2, 0); 1113 writeU8(os, WASM_OPCODE_DROP, "drop"); 1114 1115 writeU8(os, WASM_OPCODE_ELSE, "ELSE"); 1116 1117 // Did increment 0, so conditionally initialize passive data segments 1118 for (const OutputSegment *s : segments) { 1119 if (needsPassiveInitialization(s)) { 1120 // destination address 1121 writePtrConst(os, s->startVA, is64, "destination address"); 1122 if (config->isPic) { 1123 writeU8(os, WASM_OPCODE_GLOBAL_GET, "GLOBAL_GET"); 1124 writeUleb128(os, WasmSym::memoryBase->getGlobalIndex(), 1125 "memory_base"); 1126 writeU8(os, WASM_OPCODE_I32_ADD, "i32.add"); 1127 } 1128 // source segment offset 1129 writeI32Const(os, 0, "segment offset"); 1130 // memory region size 1131 writeI32Const(os, s->size, "memory region size"); 1132 // memory.init instruction 1133 writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix"); 1134 writeUleb128(os, WASM_OPCODE_MEMORY_INIT, "memory.init"); 1135 writeUleb128(os, s->index, "segment index immediate"); 1136 writeU8(os, 0, "memory index immediate"); 1137 } 1138 } 1139 1140 // Set flag to 2 to mark end of initialization 1141 writeGetFlagAddress(); 1142 writeI32Const(os, 2, "flag value"); 1143 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 1144 writeUleb128(os, WASM_OPCODE_I32_ATOMIC_STORE, "i32.atomic.store"); 1145 writeMemArg(os, 2, 0); 1146 1147 // Notify any waiters that memory initialization is complete 1148 writeGetFlagAddress(); 1149 writeI32Const(os, -1, "number of waiters"); 1150 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 1151 writeUleb128(os, WASM_OPCODE_ATOMIC_NOTIFY, "atomic.notify"); 1152 writeMemArg(os, 2, 0); 1153 writeU8(os, WASM_OPCODE_DROP, "drop"); 1154 1155 writeU8(os, WASM_OPCODE_END, "END"); 1156 1157 // Unconditionally drop passive data segments 1158 for (const OutputSegment *s : segments) { 1159 if (needsPassiveInitialization(s)) { 1160 // data.drop instruction 1161 writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix"); 1162 writeUleb128(os, WASM_OPCODE_DATA_DROP, "data.drop"); 1163 writeUleb128(os, s->index, "segment index immediate"); 1164 } 1165 } 1166 writeU8(os, WASM_OPCODE_END, "END"); 1167 } 1168 1169 createFunction(WasmSym::initMemory, bodyContent); 1170 } 1171 1172 void Writer::createStartFunction() { 1173 if (WasmSym::startFunction) { 1174 std::string bodyContent; 1175 { 1176 raw_string_ostream os(bodyContent); 1177 writeUleb128(os, 0, "num locals"); 1178 writeU8(os, WASM_OPCODE_CALL, "CALL"); 1179 writeUleb128(os, WasmSym::initMemory->getFunctionIndex(), 1180 "function index"); 1181 writeU8(os, WASM_OPCODE_CALL, "CALL"); 1182 writeUleb128(os, WasmSym::applyGlobalRelocs->getFunctionIndex(), 1183 "function index"); 1184 writeU8(os, WASM_OPCODE_END, "END"); 1185 } 1186 createFunction(WasmSym::startFunction, bodyContent); 1187 } else if (WasmSym::initMemory) { 1188 WasmSym::startFunction = WasmSym::initMemory; 1189 } else if (WasmSym::applyGlobalRelocs) { 1190 WasmSym::startFunction = WasmSym::applyGlobalRelocs; 1191 } 1192 } 1193 1194 // For -shared (PIC) output, we create create a synthetic function which will 1195 // apply any relocations to the data segments on startup. This function is 1196 // called `__wasm_apply_data_relocs` and is added at the beginning of 1197 // `__wasm_call_ctors` before any of the constructors run. 1198 void Writer::createApplyDataRelocationsFunction() { 1199 LLVM_DEBUG(dbgs() << "createApplyDataRelocationsFunction\n"); 1200 // First write the body's contents to a string. 1201 std::string bodyContent; 1202 { 1203 raw_string_ostream os(bodyContent); 1204 writeUleb128(os, 0, "num locals"); 1205 for (const OutputSegment *seg : segments) 1206 for (const InputSegment *inSeg : seg->inputSegments) 1207 inSeg->generateRelocationCode(os); 1208 1209 writeU8(os, WASM_OPCODE_END, "END"); 1210 } 1211 1212 createFunction(WasmSym::applyDataRelocs, bodyContent); 1213 } 1214 1215 // Similar to createApplyDataRelocationsFunction but generates relocation code 1216 // fro WebAssembly globals. Because these globals are not shared between threads 1217 // these relocation need to run on every thread. 1218 void Writer::createApplyGlobalRelocationsFunction() { 1219 // First write the body's contents to a string. 1220 std::string bodyContent; 1221 { 1222 raw_string_ostream os(bodyContent); 1223 writeUleb128(os, 0, "num locals"); 1224 out.globalSec->generateRelocationCode(os); 1225 writeU8(os, WASM_OPCODE_END, "END"); 1226 } 1227 1228 createFunction(WasmSym::applyGlobalRelocs, bodyContent); 1229 } 1230 1231 // Create synthetic "__wasm_call_ctors" function based on ctor functions 1232 // in input object. 1233 void Writer::createCallCtorsFunction() { 1234 // If __wasm_call_ctors isn't referenced, there aren't any ctors, and we 1235 // aren't calling `__wasm_apply_data_relocs` for Emscripten-style PIC, don't 1236 // define the `__wasm_call_ctors` function. 1237 if (!WasmSym::callCtors->isLive() && !WasmSym::applyDataRelocs && 1238 initFunctions.empty()) 1239 return; 1240 1241 // First write the body's contents to a string. 1242 std::string bodyContent; 1243 { 1244 raw_string_ostream os(bodyContent); 1245 writeUleb128(os, 0, "num locals"); 1246 1247 if (WasmSym::applyDataRelocs) { 1248 writeU8(os, WASM_OPCODE_CALL, "CALL"); 1249 writeUleb128(os, WasmSym::applyDataRelocs->getFunctionIndex(), 1250 "function index"); 1251 } 1252 1253 // Call constructors 1254 for (const WasmInitEntry &f : initFunctions) { 1255 writeU8(os, WASM_OPCODE_CALL, "CALL"); 1256 writeUleb128(os, f.sym->getFunctionIndex(), "function index"); 1257 for (size_t i = 0; i < f.sym->signature->Returns.size(); i++) { 1258 writeU8(os, WASM_OPCODE_DROP, "DROP"); 1259 } 1260 } 1261 1262 writeU8(os, WASM_OPCODE_END, "END"); 1263 } 1264 1265 createFunction(WasmSym::callCtors, bodyContent); 1266 } 1267 1268 // Create a wrapper around a function export which calls the 1269 // static constructors and destructors. 1270 void Writer::createCommandExportWrapper(uint32_t functionIndex, 1271 DefinedFunction *f) { 1272 // First write the body's contents to a string. 1273 std::string bodyContent; 1274 { 1275 raw_string_ostream os(bodyContent); 1276 writeUleb128(os, 0, "num locals"); 1277 1278 // Call `__wasm_call_ctors` which call static constructors (and 1279 // applies any runtime relocations in Emscripten-style PIC mode) 1280 if (WasmSym::callCtors->isLive()) { 1281 writeU8(os, WASM_OPCODE_CALL, "CALL"); 1282 writeUleb128(os, WasmSym::callCtors->getFunctionIndex(), 1283 "function index"); 1284 } 1285 1286 // Call the user's code, leaving any return values on the operand stack. 1287 for (size_t i = 0; i < f->signature->Params.size(); ++i) { 1288 writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get"); 1289 writeUleb128(os, i, "local index"); 1290 } 1291 writeU8(os, WASM_OPCODE_CALL, "CALL"); 1292 writeUleb128(os, functionIndex, "function index"); 1293 1294 // Call the function that calls the destructors. 1295 if (DefinedFunction *callDtors = WasmSym::callDtors) { 1296 writeU8(os, WASM_OPCODE_CALL, "CALL"); 1297 writeUleb128(os, callDtors->getFunctionIndex(), "function index"); 1298 } 1299 1300 // End the function, returning the return values from the user's code. 1301 writeU8(os, WASM_OPCODE_END, "END"); 1302 } 1303 1304 createFunction(f, bodyContent); 1305 } 1306 1307 void Writer::createInitTLSFunction() { 1308 std::string bodyContent; 1309 { 1310 raw_string_ostream os(bodyContent); 1311 1312 OutputSegment *tlsSeg = nullptr; 1313 for (auto *seg : segments) { 1314 if (seg->name == ".tdata") { 1315 tlsSeg = seg; 1316 break; 1317 } 1318 } 1319 1320 writeUleb128(os, 0, "num locals"); 1321 if (tlsSeg) { 1322 writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get"); 1323 writeUleb128(os, 0, "local index"); 1324 1325 writeU8(os, WASM_OPCODE_GLOBAL_SET, "global.set"); 1326 writeUleb128(os, WasmSym::tlsBase->getGlobalIndex(), "global index"); 1327 1328 // FIXME(wvo): this local needs to be I64 in wasm64, or we need an extend op. 1329 writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get"); 1330 writeUleb128(os, 0, "local index"); 1331 1332 writeI32Const(os, 0, "segment offset"); 1333 1334 writeI32Const(os, tlsSeg->size, "memory region size"); 1335 1336 writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix"); 1337 writeUleb128(os, WASM_OPCODE_MEMORY_INIT, "MEMORY.INIT"); 1338 writeUleb128(os, tlsSeg->index, "segment index immediate"); 1339 writeU8(os, 0, "memory index immediate"); 1340 } 1341 writeU8(os, WASM_OPCODE_END, "end function"); 1342 } 1343 1344 createFunction(WasmSym::initTLS, bodyContent); 1345 } 1346 1347 // Populate InitFunctions vector with init functions from all input objects. 1348 // This is then used either when creating the output linking section or to 1349 // synthesize the "__wasm_call_ctors" function. 1350 void Writer::calculateInitFunctions() { 1351 if (!config->relocatable && !WasmSym::callCtors->isLive()) 1352 return; 1353 1354 for (ObjFile *file : symtab->objectFiles) { 1355 const WasmLinkingData &l = file->getWasmObj()->linkingData(); 1356 for (const WasmInitFunc &f : l.InitFunctions) { 1357 FunctionSymbol *sym = file->getFunctionSymbol(f.Symbol); 1358 // comdat exclusions can cause init functions be discarded. 1359 if (sym->isDiscarded() || !sym->isLive()) 1360 continue; 1361 if (sym->signature->Params.size() != 0) 1362 error("constructor functions cannot take arguments: " + toString(*sym)); 1363 LLVM_DEBUG(dbgs() << "initFunctions: " << toString(*sym) << "\n"); 1364 initFunctions.emplace_back(WasmInitEntry{sym, f.Priority}); 1365 } 1366 } 1367 1368 // Sort in order of priority (lowest first) so that they are called 1369 // in the correct order. 1370 llvm::stable_sort(initFunctions, 1371 [](const WasmInitEntry &l, const WasmInitEntry &r) { 1372 return l.priority < r.priority; 1373 }); 1374 } 1375 1376 void Writer::createSyntheticSections() { 1377 out.dylinkSec = make<DylinkSection>(); 1378 out.typeSec = make<TypeSection>(); 1379 out.importSec = make<ImportSection>(); 1380 out.functionSec = make<FunctionSection>(); 1381 out.tableSec = make<TableSection>(); 1382 out.memorySec = make<MemorySection>(); 1383 out.eventSec = make<EventSection>(); 1384 out.globalSec = make<GlobalSection>(); 1385 out.exportSec = make<ExportSection>(); 1386 out.startSec = make<StartSection>(); 1387 out.elemSec = make<ElemSection>(); 1388 out.producersSec = make<ProducersSection>(); 1389 out.targetFeaturesSec = make<TargetFeaturesSection>(); 1390 } 1391 1392 void Writer::createSyntheticSectionsPostLayout() { 1393 out.dataCountSec = make<DataCountSection>(segments); 1394 out.linkingSec = make<LinkingSection>(initFunctions, segments); 1395 out.nameSec = make<NameSection>(segments); 1396 } 1397 1398 void Writer::run() { 1399 if (config->relocatable || config->isPic) 1400 config->globalBase = 0; 1401 1402 // For PIC code the table base is assigned dynamically by the loader. 1403 // For non-PIC, we start at 1 so that accessing table index 0 always traps. 1404 if (!config->isPic) { 1405 config->tableBase = 1; 1406 if (WasmSym::definedTableBase) 1407 WasmSym::definedTableBase->setVA(config->tableBase); 1408 } 1409 1410 log("-- createOutputSegments"); 1411 createOutputSegments(); 1412 log("-- createSyntheticSections"); 1413 createSyntheticSections(); 1414 log("-- layoutMemory"); 1415 layoutMemory(); 1416 1417 if (!config->relocatable) { 1418 // Create linker synthesized __start_SECNAME/__stop_SECNAME symbols 1419 // This has to be done after memory layout is performed. 1420 for (const OutputSegment *seg : segments) { 1421 addStartStopSymbols(seg); 1422 } 1423 } 1424 1425 for (auto &pair : config->exportedSymbols) { 1426 Symbol *sym = symtab->find(pair.first()); 1427 if (sym && sym->isDefined()) 1428 sym->forceExport = true; 1429 } 1430 1431 // Delay reporting error about explict exports until after addStartStopSymbols 1432 // which can create optional symbols. 1433 for (auto &name : config->requiredExports) { 1434 Symbol *sym = symtab->find(name); 1435 if (!sym || !sym->isDefined()) { 1436 if (config->unresolvedSymbols == UnresolvedPolicy::ReportError) 1437 error(Twine("symbol exported via --export not found: ") + name); 1438 if (config->unresolvedSymbols == UnresolvedPolicy::Warn) 1439 warn(Twine("symbol exported via --export not found: ") + name); 1440 } 1441 } 1442 1443 if (config->isPic) { 1444 log("-- combineOutputSegments"); 1445 combineOutputSegments(); 1446 } 1447 1448 log("-- createSyntheticSectionsPostLayout"); 1449 createSyntheticSectionsPostLayout(); 1450 log("-- populateProducers"); 1451 populateProducers(); 1452 log("-- calculateImports"); 1453 calculateImports(); 1454 log("-- scanRelocations"); 1455 scanRelocations(); 1456 log("-- finalizeIndirectFunctionTable"); 1457 finalizeIndirectFunctionTable(); 1458 log("-- createSyntheticInitFunctions"); 1459 createSyntheticInitFunctions(); 1460 log("-- assignIndexes"); 1461 assignIndexes(); 1462 log("-- calculateInitFunctions"); 1463 calculateInitFunctions(); 1464 1465 if (!config->relocatable) { 1466 // Create linker synthesized functions 1467 if (WasmSym::applyDataRelocs) 1468 createApplyDataRelocationsFunction(); 1469 if (WasmSym::applyGlobalRelocs) 1470 createApplyGlobalRelocationsFunction(); 1471 if (WasmSym::initMemory) 1472 createInitMemoryFunction(); 1473 createStartFunction(); 1474 1475 createCallCtorsFunction(); 1476 1477 // Create export wrappers for commands if needed. 1478 // 1479 // If the input contains a call to `__wasm_call_ctors`, either in one of 1480 // the input objects or an explicit export from the command-line, we 1481 // assume ctors and dtors are taken care of already. 1482 if (!config->relocatable && !config->isPic && 1483 !WasmSym::callCtors->isUsedInRegularObj && 1484 !WasmSym::callCtors->isExported()) { 1485 log("-- createCommandExportWrappers"); 1486 createCommandExportWrappers(); 1487 } 1488 } 1489 1490 if (WasmSym::initTLS && WasmSym::initTLS->isLive()) 1491 createInitTLSFunction(); 1492 1493 if (errorCount()) 1494 return; 1495 1496 log("-- calculateTypes"); 1497 calculateTypes(); 1498 log("-- calculateExports"); 1499 calculateExports(); 1500 log("-- calculateCustomSections"); 1501 calculateCustomSections(); 1502 log("-- populateSymtab"); 1503 populateSymtab(); 1504 log("-- populateTargetFeatures"); 1505 populateTargetFeatures(); 1506 log("-- addSections"); 1507 addSections(); 1508 1509 if (errorHandler().verbose) { 1510 log("Defined Functions: " + Twine(out.functionSec->inputFunctions.size())); 1511 log("Defined Globals : " + Twine(out.globalSec->numGlobals())); 1512 log("Defined Events : " + Twine(out.eventSec->inputEvents.size())); 1513 log("Defined Tables : " + Twine(out.tableSec->inputTables.size())); 1514 log("Function Imports : " + 1515 Twine(out.importSec->getNumImportedFunctions())); 1516 log("Global Imports : " + Twine(out.importSec->getNumImportedGlobals())); 1517 log("Event Imports : " + Twine(out.importSec->getNumImportedEvents())); 1518 log("Table Imports : " + Twine(out.importSec->getNumImportedTables())); 1519 for (ObjFile *file : symtab->objectFiles) 1520 file->dumpInfo(); 1521 } 1522 1523 createHeader(); 1524 log("-- finalizeSections"); 1525 finalizeSections(); 1526 1527 log("-- writeMapFile"); 1528 writeMapFile(outputSections); 1529 1530 log("-- openFile"); 1531 openFile(); 1532 if (errorCount()) 1533 return; 1534 1535 writeHeader(); 1536 1537 log("-- writeSections"); 1538 writeSections(); 1539 if (errorCount()) 1540 return; 1541 1542 if (Error e = buffer->commit()) 1543 fatal("failed to write the output file: " + toString(std::move(e))); 1544 } 1545 1546 // Open a result file. 1547 void Writer::openFile() { 1548 log("writing: " + config->outputFile); 1549 1550 Expected<std::unique_ptr<FileOutputBuffer>> bufferOrErr = 1551 FileOutputBuffer::create(config->outputFile, fileSize, 1552 FileOutputBuffer::F_executable); 1553 1554 if (!bufferOrErr) 1555 error("failed to open " + config->outputFile + ": " + 1556 toString(bufferOrErr.takeError())); 1557 else 1558 buffer = std::move(*bufferOrErr); 1559 } 1560 1561 void Writer::createHeader() { 1562 raw_string_ostream os(header); 1563 writeBytes(os, WasmMagic, sizeof(WasmMagic), "wasm magic"); 1564 writeU32(os, WasmVersion, "wasm version"); 1565 os.flush(); 1566 fileSize += header.size(); 1567 } 1568 1569 void writeResult() { Writer().run(); } 1570 1571 } // namespace wasm 1572 } // namespace lld 1573