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 "InputEvent.h" 13 #include "InputGlobal.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 "lld/Common/Threads.h" 24 #include "llvm/ADT/DenseSet.h" 25 #include "llvm/ADT/SmallSet.h" 26 #include "llvm/ADT/SmallVector.h" 27 #include "llvm/ADT/StringMap.h" 28 #include "llvm/BinaryFormat/Wasm.h" 29 #include "llvm/Object/WasmTraits.h" 30 #include "llvm/Support/FileOutputBuffer.h" 31 #include "llvm/Support/Format.h" 32 #include "llvm/Support/FormatVariadic.h" 33 #include "llvm/Support/LEB128.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 void createInitMemoryFunction(); 58 void createApplyRelocationsFunction(); 59 void createCallCtorsFunction(); 60 void createInitTLSFunction(); 61 62 void assignIndexes(); 63 void populateSymtab(); 64 void populateProducers(); 65 void populateTargetFeatures(); 66 void calculateInitFunctions(); 67 void calculateImports(); 68 void calculateExports(); 69 void calculateCustomSections(); 70 void calculateTypes(); 71 void createOutputSegments(); 72 void layoutMemory(); 73 void createHeader(); 74 75 void addSection(OutputSection *sec); 76 77 void addSections(); 78 79 void createCustomSections(); 80 void createSyntheticSections(); 81 void finalizeSections(); 82 83 // Custom sections 84 void createRelocSections(); 85 86 void writeHeader(); 87 void writeSections(); 88 89 uint64_t fileSize = 0; 90 91 std::vector<WasmInitEntry> initFunctions; 92 llvm::StringMap<std::vector<InputSection *>> customSectionMapping; 93 94 // Elements that are used to construct the final output 95 std::string header; 96 std::vector<OutputSection *> outputSections; 97 98 std::unique_ptr<FileOutputBuffer> buffer; 99 100 std::vector<OutputSegment *> segments; 101 llvm::SmallDenseMap<StringRef, OutputSegment *> segmentMap; 102 }; 103 104 } // anonymous namespace 105 106 void Writer::calculateCustomSections() { 107 log("calculateCustomSections"); 108 bool stripDebug = config->stripDebug || config->stripAll; 109 for (ObjFile *file : symtab->objectFiles) { 110 for (InputSection *section : file->customSections) { 111 StringRef name = section->getName(); 112 // These custom sections are known the linker and synthesized rather than 113 // blindly copied 114 if (name == "linking" || name == "name" || name == "producers" || 115 name == "target_features" || name.startswith("reloc.")) 116 continue; 117 // .. or it is a debug section 118 if (stripDebug && name.startswith(".debug_")) 119 continue; 120 customSectionMapping[name].push_back(section); 121 } 122 } 123 } 124 125 void Writer::createCustomSections() { 126 log("createCustomSections"); 127 for (auto &pair : customSectionMapping) { 128 StringRef name = pair.first(); 129 LLVM_DEBUG(dbgs() << "createCustomSection: " << name << "\n"); 130 131 OutputSection *sec = make<CustomSection>(std::string(name), pair.second); 132 if (config->relocatable || config->emitRelocs) { 133 auto *sym = make<OutputSectionSymbol>(sec); 134 out.linkingSec->addToSymtab(sym); 135 sec->sectionSym = sym; 136 } 137 addSection(sec); 138 } 139 } 140 141 // Create relocations sections in the final output. 142 // These are only created when relocatable output is requested. 143 void Writer::createRelocSections() { 144 log("createRelocSections"); 145 // Don't use iterator here since we are adding to OutputSection 146 size_t origSize = outputSections.size(); 147 for (size_t i = 0; i < origSize; i++) { 148 LLVM_DEBUG(dbgs() << "check section " << i << "\n"); 149 OutputSection *sec = outputSections[i]; 150 151 // Count the number of needed sections. 152 uint32_t count = sec->getNumRelocations(); 153 if (!count) 154 continue; 155 156 StringRef name; 157 if (sec->type == WASM_SEC_DATA) 158 name = "reloc.DATA"; 159 else if (sec->type == WASM_SEC_CODE) 160 name = "reloc.CODE"; 161 else if (sec->type == WASM_SEC_CUSTOM) 162 name = saver.save("reloc." + sec->name); 163 else 164 llvm_unreachable( 165 "relocations only supported for code, data, or custom sections"); 166 167 addSection(make<RelocSection>(name, sec)); 168 } 169 } 170 171 void Writer::populateProducers() { 172 for (ObjFile *file : symtab->objectFiles) { 173 const WasmProducerInfo &info = file->getWasmObj()->getProducerInfo(); 174 out.producersSec->addInfo(info); 175 } 176 } 177 178 void Writer::writeHeader() { 179 memcpy(buffer->getBufferStart(), header.data(), header.size()); 180 } 181 182 void Writer::writeSections() { 183 uint8_t *buf = buffer->getBufferStart(); 184 parallelForEach(outputSections, [buf](OutputSection *s) { 185 assert(s->isNeeded()); 186 s->writeTo(buf); 187 }); 188 } 189 190 // Fix the memory layout of the output binary. This assigns memory offsets 191 // to each of the input data sections as well as the explicit stack region. 192 // The default memory layout is as follows, from low to high. 193 // 194 // - initialized data (starting at Config->globalBase) 195 // - BSS data (not currently implemented in llvm) 196 // - explicit stack (Config->ZStackSize) 197 // - heap start / unallocated 198 // 199 // The --stack-first option means that stack is placed before any static data. 200 // This can be useful since it means that stack overflow traps immediately 201 // rather than overwriting global data, but also increases code size since all 202 // static data loads and stores requires larger offsets. 203 void Writer::layoutMemory() { 204 uint64_t memoryPtr = 0; 205 206 auto placeStack = [&]() { 207 if (config->relocatable || config->isPic) 208 return; 209 memoryPtr = alignTo(memoryPtr, stackAlignment); 210 if (config->zStackSize != alignTo(config->zStackSize, stackAlignment)) 211 error("stack size must be " + Twine(stackAlignment) + "-byte aligned"); 212 log("mem: stack size = " + Twine(config->zStackSize)); 213 log("mem: stack base = " + Twine(memoryPtr)); 214 memoryPtr += config->zStackSize; 215 auto *sp = cast<DefinedGlobal>(WasmSym::stackPointer); 216 sp->global->global.InitExpr.Value.Int32 = memoryPtr; 217 log("mem: stack top = " + Twine(memoryPtr)); 218 }; 219 220 if (config->stackFirst) { 221 placeStack(); 222 } else { 223 memoryPtr = config->globalBase; 224 log("mem: global base = " + Twine(config->globalBase)); 225 } 226 227 if (WasmSym::globalBase) 228 WasmSym::globalBase->setVirtualAddress(memoryPtr); 229 230 uint64_t dataStart = memoryPtr; 231 232 // Arbitrarily set __dso_handle handle to point to the start of the data 233 // segments. 234 if (WasmSym::dsoHandle) 235 WasmSym::dsoHandle->setVirtualAddress(dataStart); 236 237 out.dylinkSec->memAlign = 0; 238 for (OutputSegment *seg : segments) { 239 out.dylinkSec->memAlign = std::max(out.dylinkSec->memAlign, seg->alignment); 240 memoryPtr = alignTo(memoryPtr, 1ULL << seg->alignment); 241 seg->startVA = memoryPtr; 242 log(formatv("mem: {0,-15} offset={1,-8} size={2,-8} align={3}", seg->name, 243 memoryPtr, seg->size, seg->alignment)); 244 memoryPtr += seg->size; 245 246 if (WasmSym::tlsSize && seg->name == ".tdata") { 247 auto *tlsSize = cast<DefinedGlobal>(WasmSym::tlsSize); 248 tlsSize->global->global.InitExpr.Value.Int32 = seg->size; 249 250 auto *tlsAlign = cast<DefinedGlobal>(WasmSym::tlsAlign); 251 tlsAlign->global->global.InitExpr.Value.Int32 = 1U << seg->alignment; 252 } 253 } 254 255 // Make space for the memory initialization flag 256 if (WasmSym::initMemoryFlag) { 257 memoryPtr = alignTo(memoryPtr, 4); 258 WasmSym::initMemoryFlag->setVirtualAddress(memoryPtr); 259 log(formatv("mem: {0,-15} offset={1,-8} size={2,-8} align={3}", 260 "__wasm_init_memory_flag", memoryPtr, 4, 4)); 261 memoryPtr += 4; 262 } 263 264 if (WasmSym::dataEnd) 265 WasmSym::dataEnd->setVirtualAddress(memoryPtr); 266 267 log("mem: static data = " + Twine(memoryPtr - dataStart)); 268 269 if (config->shared) { 270 out.dylinkSec->memSize = memoryPtr; 271 return; 272 } 273 274 if (!config->stackFirst) 275 placeStack(); 276 277 // Set `__heap_base` to directly follow the end of the stack or global data. 278 // The fact that this comes last means that a malloc/brk implementation 279 // can grow the heap at runtime. 280 log("mem: heap base = " + Twine(memoryPtr)); 281 if (WasmSym::heapBase) 282 WasmSym::heapBase->setVirtualAddress(memoryPtr); 283 284 if (config->initialMemory != 0) { 285 if (config->initialMemory != alignTo(config->initialMemory, WasmPageSize)) 286 error("initial memory must be " + Twine(WasmPageSize) + "-byte aligned"); 287 if (memoryPtr > config->initialMemory) 288 error("initial memory too small, " + Twine(memoryPtr) + " bytes needed"); 289 if (config->initialMemory > (1ULL << 32)) 290 error("initial memory too large, cannot be greater than 4294967296"); 291 memoryPtr = config->initialMemory; 292 } 293 out.dylinkSec->memSize = memoryPtr; 294 out.memorySec->numMemoryPages = 295 alignTo(memoryPtr, WasmPageSize) / WasmPageSize; 296 log("mem: total pages = " + Twine(out.memorySec->numMemoryPages)); 297 298 // Check max if explicitly supplied or required by shared memory 299 if (config->maxMemory != 0 || config->sharedMemory) { 300 if (config->maxMemory != alignTo(config->maxMemory, WasmPageSize)) 301 error("maximum memory must be " + Twine(WasmPageSize) + "-byte aligned"); 302 if (memoryPtr > config->maxMemory) 303 error("maximum memory too small, " + Twine(memoryPtr) + " bytes needed"); 304 if (config->maxMemory > (1ULL << 32)) 305 error("maximum memory too large, cannot be greater than 4294967296"); 306 out.memorySec->maxMemoryPages = config->maxMemory / WasmPageSize; 307 log("mem: max pages = " + Twine(out.memorySec->maxMemoryPages)); 308 } 309 } 310 311 void Writer::addSection(OutputSection *sec) { 312 if (!sec->isNeeded()) 313 return; 314 log("addSection: " + toString(*sec)); 315 sec->sectionIndex = outputSections.size(); 316 outputSections.push_back(sec); 317 } 318 319 // If a section name is valid as a C identifier (which is rare because of 320 // the leading '.'), linkers are expected to define __start_<secname> and 321 // __stop_<secname> symbols. They are at beginning and end of the section, 322 // respectively. This is not requested by the ELF standard, but GNU ld and 323 // gold provide the feature, and used by many programs. 324 static void addStartStopSymbols(const OutputSegment *seg) { 325 StringRef name = seg->name; 326 if (!isValidCIdentifier(name)) 327 return; 328 LLVM_DEBUG(dbgs() << "addStartStopSymbols: " << name << "\n"); 329 uint32_t start = seg->startVA; 330 uint32_t stop = start + seg->size; 331 symtab->addOptionalDataSymbol(saver.save("__start_" + name), start); 332 symtab->addOptionalDataSymbol(saver.save("__stop_" + name), stop); 333 } 334 335 void Writer::addSections() { 336 addSection(out.dylinkSec); 337 addSection(out.typeSec); 338 addSection(out.importSec); 339 addSection(out.functionSec); 340 addSection(out.tableSec); 341 addSection(out.memorySec); 342 addSection(out.eventSec); 343 addSection(out.globalSec); 344 addSection(out.exportSec); 345 addSection(out.startSec); 346 addSection(out.elemSec); 347 addSection(out.dataCountSec); 348 349 addSection(make<CodeSection>(out.functionSec->inputFunctions)); 350 addSection(make<DataSection>(segments)); 351 352 createCustomSections(); 353 354 addSection(out.linkingSec); 355 if (config->emitRelocs || config->relocatable) { 356 createRelocSections(); 357 } 358 359 addSection(out.nameSec); 360 addSection(out.producersSec); 361 addSection(out.targetFeaturesSec); 362 } 363 364 void Writer::finalizeSections() { 365 for (OutputSection *s : outputSections) { 366 s->setOffset(fileSize); 367 s->finalizeContents(); 368 fileSize += s->getSize(); 369 } 370 } 371 372 void Writer::populateTargetFeatures() { 373 StringMap<std::string> used; 374 StringMap<std::string> required; 375 StringMap<std::string> disallowed; 376 SmallSet<std::string, 8> &allowed = out.targetFeaturesSec->features; 377 bool tlsUsed = false; 378 379 // Only infer used features if user did not specify features 380 bool inferFeatures = !config->features.hasValue(); 381 382 if (!inferFeatures) { 383 auto &explicitFeatures = config->features.getValue(); 384 allowed.insert(explicitFeatures.begin(), explicitFeatures.end()); 385 if (!config->checkFeatures) 386 return; 387 } 388 389 // Find the sets of used, required, and disallowed features 390 for (ObjFile *file : symtab->objectFiles) { 391 StringRef fileName(file->getName()); 392 for (auto &feature : file->getWasmObj()->getTargetFeatures()) { 393 switch (feature.Prefix) { 394 case WASM_FEATURE_PREFIX_USED: 395 used.insert({feature.Name, std::string(fileName)}); 396 break; 397 case WASM_FEATURE_PREFIX_REQUIRED: 398 used.insert({feature.Name, std::string(fileName)}); 399 required.insert({feature.Name, std::string(fileName)}); 400 break; 401 case WASM_FEATURE_PREFIX_DISALLOWED: 402 disallowed.insert({feature.Name, std::string(fileName)}); 403 break; 404 default: 405 error("Unrecognized feature policy prefix " + 406 std::to_string(feature.Prefix)); 407 } 408 } 409 410 // Find TLS data segments 411 auto isTLS = [](InputSegment *segment) { 412 StringRef name = segment->getName(); 413 return segment->live && 414 (name.startswith(".tdata") || name.startswith(".tbss")); 415 }; 416 tlsUsed = tlsUsed || 417 std::any_of(file->segments.begin(), file->segments.end(), isTLS); 418 } 419 420 if (inferFeatures) 421 for (const auto &key : used.keys()) 422 allowed.insert(std::string(key)); 423 424 if (allowed.count("atomics") && !config->sharedMemory) { 425 if (inferFeatures) 426 error(Twine("'atomics' feature is used by ") + used["atomics"] + 427 ", so --shared-memory must be used"); 428 else 429 error("'atomics' feature is used, so --shared-memory must be used"); 430 } 431 432 if (!config->checkFeatures) 433 return; 434 435 if (disallowed.count("atomics") && config->sharedMemory) 436 error("'atomics' feature is disallowed by " + disallowed["atomics"] + 437 ", so --shared-memory must not be used"); 438 439 if (!allowed.count("atomics") && config->sharedMemory) 440 error("'atomics' feature must be used in order to use shared " 441 "memory"); 442 443 if (!allowed.count("bulk-memory") && config->sharedMemory) 444 error("'bulk-memory' feature must be used in order to use shared " 445 "memory"); 446 447 if (!allowed.count("bulk-memory") && tlsUsed) 448 error("'bulk-memory' feature must be used in order to use thread-local " 449 "storage"); 450 451 // Validate that used features are allowed in output 452 if (!inferFeatures) { 453 for (auto &feature : used.keys()) { 454 if (!allowed.count(std::string(feature))) 455 error(Twine("Target feature '") + feature + "' used by " + 456 used[feature] + " is not allowed."); 457 } 458 } 459 460 // Validate the required and disallowed constraints for each file 461 for (ObjFile *file : symtab->objectFiles) { 462 StringRef fileName(file->getName()); 463 SmallSet<std::string, 8> objectFeatures; 464 for (auto &feature : file->getWasmObj()->getTargetFeatures()) { 465 if (feature.Prefix == WASM_FEATURE_PREFIX_DISALLOWED) 466 continue; 467 objectFeatures.insert(feature.Name); 468 if (disallowed.count(feature.Name)) 469 error(Twine("Target feature '") + feature.Name + "' used in " + 470 fileName + " is disallowed by " + disallowed[feature.Name] + 471 ". Use --no-check-features to suppress."); 472 } 473 for (auto &feature : required.keys()) { 474 if (!objectFeatures.count(std::string(feature))) 475 error(Twine("Missing target feature '") + feature + "' in " + fileName + 476 ", required by " + required[feature] + 477 ". Use --no-check-features to suppress."); 478 } 479 } 480 } 481 482 void Writer::calculateImports() { 483 for (Symbol *sym : symtab->getSymbols()) { 484 if (!sym->isUndefined()) 485 continue; 486 if (sym->isWeak() && !config->relocatable) 487 continue; 488 if (!sym->isLive()) 489 continue; 490 if (!sym->isUsedInRegularObj) 491 continue; 492 // We don't generate imports for data symbols. They however can be imported 493 // as GOT entries. 494 if (isa<DataSymbol>(sym)) 495 continue; 496 497 LLVM_DEBUG(dbgs() << "import: " << sym->getName() << "\n"); 498 out.importSec->addImport(sym); 499 } 500 } 501 502 void Writer::calculateExports() { 503 if (config->relocatable) 504 return; 505 506 if (!config->relocatable && !config->importMemory) 507 out.exportSec->exports.push_back( 508 WasmExport{"memory", WASM_EXTERNAL_MEMORY, 0}); 509 510 if (!config->relocatable && config->exportTable) 511 out.exportSec->exports.push_back( 512 WasmExport{functionTableName, WASM_EXTERNAL_TABLE, 0}); 513 514 unsigned globalIndex = 515 out.importSec->getNumImportedGlobals() + out.globalSec->numGlobals(); 516 517 for (Symbol *sym : symtab->getSymbols()) { 518 if (!sym->isExported()) 519 continue; 520 if (!sym->isLive()) 521 continue; 522 523 StringRef name = sym->getName(); 524 WasmExport export_; 525 if (auto *f = dyn_cast<DefinedFunction>(sym)) { 526 if (Optional<StringRef> exportName = f->function->getExportName()) { 527 name = *exportName; 528 } 529 export_ = {name, WASM_EXTERNAL_FUNCTION, f->getFunctionIndex()}; 530 } else if (auto *g = dyn_cast<DefinedGlobal>(sym)) { 531 // TODO(sbc): Remove this check once to mutable global proposal is 532 // implement in all major browsers. 533 // See: https://github.com/WebAssembly/mutable-global 534 if (g->getGlobalType()->Mutable) { 535 // Only __stack_pointer and __tls_base should ever be create as mutable. 536 assert(g == WasmSym::stackPointer || g == WasmSym::tlsBase); 537 continue; 538 } 539 export_ = {name, WASM_EXTERNAL_GLOBAL, g->getGlobalIndex()}; 540 } else if (auto *e = dyn_cast<DefinedEvent>(sym)) { 541 export_ = {name, WASM_EXTERNAL_EVENT, e->getEventIndex()}; 542 } else { 543 auto *d = cast<DefinedData>(sym); 544 out.globalSec->dataAddressGlobals.push_back(d); 545 export_ = {name, WASM_EXTERNAL_GLOBAL, globalIndex++}; 546 } 547 548 LLVM_DEBUG(dbgs() << "Export: " << name << "\n"); 549 out.exportSec->exports.push_back(export_); 550 } 551 } 552 553 void Writer::populateSymtab() { 554 if (!config->relocatable && !config->emitRelocs) 555 return; 556 557 for (Symbol *sym : symtab->getSymbols()) 558 if (sym->isUsedInRegularObj && sym->isLive()) 559 out.linkingSec->addToSymtab(sym); 560 561 for (ObjFile *file : symtab->objectFiles) { 562 LLVM_DEBUG(dbgs() << "Local symtab entries: " << file->getName() << "\n"); 563 for (Symbol *sym : file->getSymbols()) 564 if (sym->isLocal() && !isa<SectionSymbol>(sym) && sym->isLive()) 565 out.linkingSec->addToSymtab(sym); 566 } 567 } 568 569 void Writer::calculateTypes() { 570 // The output type section is the union of the following sets: 571 // 1. Any signature used in the TYPE relocation 572 // 2. The signatures of all imported functions 573 // 3. The signatures of all defined functions 574 // 4. The signatures of all imported events 575 // 5. The signatures of all defined events 576 577 for (ObjFile *file : symtab->objectFiles) { 578 ArrayRef<WasmSignature> types = file->getWasmObj()->types(); 579 for (uint32_t i = 0; i < types.size(); i++) 580 if (file->typeIsUsed[i]) 581 file->typeMap[i] = out.typeSec->registerType(types[i]); 582 } 583 584 for (const Symbol *sym : out.importSec->importedSymbols) { 585 if (auto *f = dyn_cast<FunctionSymbol>(sym)) 586 out.typeSec->registerType(*f->signature); 587 else if (auto *e = dyn_cast<EventSymbol>(sym)) 588 out.typeSec->registerType(*e->signature); 589 } 590 591 for (const InputFunction *f : out.functionSec->inputFunctions) 592 out.typeSec->registerType(f->signature); 593 594 for (const InputEvent *e : out.eventSec->inputEvents) 595 out.typeSec->registerType(e->signature); 596 } 597 598 static void scanRelocations() { 599 for (ObjFile *file : symtab->objectFiles) { 600 LLVM_DEBUG(dbgs() << "scanRelocations: " << file->getName() << "\n"); 601 for (InputChunk *chunk : file->functions) 602 scanRelocations(chunk); 603 for (InputChunk *chunk : file->segments) 604 scanRelocations(chunk); 605 for (auto &p : file->customSections) 606 scanRelocations(p); 607 } 608 } 609 610 void Writer::assignIndexes() { 611 // Seal the import section, since other index spaces such as function and 612 // global are effected by the number of imports. 613 out.importSec->seal(); 614 615 for (InputFunction *func : symtab->syntheticFunctions) 616 out.functionSec->addFunction(func); 617 618 for (ObjFile *file : symtab->objectFiles) { 619 LLVM_DEBUG(dbgs() << "Functions: " << file->getName() << "\n"); 620 for (InputFunction *func : file->functions) 621 out.functionSec->addFunction(func); 622 } 623 624 for (InputGlobal *global : symtab->syntheticGlobals) 625 out.globalSec->addGlobal(global); 626 627 for (ObjFile *file : symtab->objectFiles) { 628 LLVM_DEBUG(dbgs() << "Globals: " << file->getName() << "\n"); 629 for (InputGlobal *global : file->globals) 630 out.globalSec->addGlobal(global); 631 } 632 633 for (ObjFile *file : symtab->objectFiles) { 634 LLVM_DEBUG(dbgs() << "Events: " << file->getName() << "\n"); 635 for (InputEvent *event : file->events) 636 out.eventSec->addEvent(event); 637 } 638 639 out.globalSec->assignIndexes(); 640 } 641 642 static StringRef getOutputDataSegmentName(StringRef name) { 643 // With PIC code we currently only support a single data segment since 644 // we only have a single __memory_base to use as our base address. 645 if (config->isPic) 646 return ".data"; 647 // We only support one thread-local segment, so we must merge the segments 648 // despite --no-merge-data-segments. 649 // We also need to merge .tbss into .tdata so they share the same offsets. 650 if (name.startswith(".tdata") || name.startswith(".tbss")) 651 return ".tdata"; 652 if (!config->mergeDataSegments) 653 return name; 654 if (name.startswith(".text.")) 655 return ".text"; 656 if (name.startswith(".data.")) 657 return ".data"; 658 if (name.startswith(".bss.")) 659 return ".bss"; 660 if (name.startswith(".rodata.")) 661 return ".rodata"; 662 return name; 663 } 664 665 void Writer::createOutputSegments() { 666 for (ObjFile *file : symtab->objectFiles) { 667 for (InputSegment *segment : file->segments) { 668 if (!segment->live) 669 continue; 670 StringRef name = getOutputDataSegmentName(segment->getName()); 671 OutputSegment *&s = segmentMap[name]; 672 if (s == nullptr) { 673 LLVM_DEBUG(dbgs() << "new segment: " << name << "\n"); 674 s = make<OutputSegment>(name); 675 if (config->sharedMemory || name == ".tdata") 676 s->initFlags = WASM_SEGMENT_IS_PASSIVE; 677 // Exported memories are guaranteed to be zero-initialized, so no need 678 // to emit data segments for bss sections. 679 // TODO: consider initializing bss sections with memory.fill 680 // instructions when memory is imported and bulk-memory is available. 681 if (!config->importMemory && !config->relocatable && 682 name.startswith(".bss")) 683 s->isBss = true; 684 segments.push_back(s); 685 } 686 s->addInputSegment(segment); 687 LLVM_DEBUG(dbgs() << "added data: " << name << ": " << s->size << "\n"); 688 } 689 } 690 691 // Sort segments by type, placing .bss last 692 std::stable_sort(segments.begin(), segments.end(), 693 [](const OutputSegment *a, const OutputSegment *b) { 694 auto order = [](StringRef name) { 695 return StringSwitch<int>(name) 696 .StartsWith(".rodata", 0) 697 .StartsWith(".data", 1) 698 .StartsWith(".tdata", 2) 699 .StartsWith(".bss", 4) 700 .Default(3); 701 }; 702 return order(a->name) < order(b->name); 703 }); 704 705 for (size_t i = 0; i < segments.size(); ++i) 706 segments[i]->index = i; 707 } 708 709 static void createFunction(DefinedFunction *func, StringRef bodyContent) { 710 std::string functionBody; 711 { 712 raw_string_ostream os(functionBody); 713 writeUleb128(os, bodyContent.size(), "function size"); 714 os << bodyContent; 715 } 716 ArrayRef<uint8_t> body = arrayRefFromStringRef(saver.save(functionBody)); 717 cast<SyntheticFunction>(func->function)->setBody(body); 718 } 719 720 void Writer::createInitMemoryFunction() { 721 LLVM_DEBUG(dbgs() << "createInitMemoryFunction\n"); 722 assert(WasmSym::initMemoryFlag); 723 uint32_t flagAddress = WasmSym::initMemoryFlag->getVirtualAddress(); 724 std::string bodyContent; 725 { 726 raw_string_ostream os(bodyContent); 727 writeUleb128(os, 0, "num locals"); 728 729 if (segments.size()) { 730 // Initialize memory in a thread-safe manner. The thread that successfully 731 // increments the flag from 0 to 1 is is responsible for performing the 732 // memory initialization. Other threads go sleep on the flag until the 733 // first thread finishing initializing memory, increments the flag to 2, 734 // and wakes all the other threads. Once the flag has been set to 2, 735 // subsequently started threads will skip the sleep. All threads 736 // unconditionally drop their passive data segments once memory has been 737 // initialized. The generated code is as follows: 738 // 739 // (func $__wasm_init_memory 740 // (if 741 // (i32.atomic.rmw.cmpxchg align=2 offset=0 742 // (i32.const $__init_memory_flag) 743 // (i32.const 0) 744 // (i32.const 1) 745 // ) 746 // (then 747 // (drop 748 // (i32.atomic.wait align=2 offset=0 749 // (i32.const $__init_memory_flag) 750 // (i32.const 1) 751 // (i32.const -1) 752 // ) 753 // ) 754 // ) 755 // (else 756 // ( ... initialize data segments ... ) 757 // (i32.atomic.store align=2 offset=0 758 // (i32.const $__init_memory_flag) 759 // (i32.const 2) 760 // ) 761 // (drop 762 // (i32.atomic.notify align=2 offset=0 763 // (i32.const $__init_memory_flag) 764 // (i32.const -1u) 765 // ) 766 // ) 767 // ) 768 // ) 769 // ( ... drop data segments ... ) 770 // ) 771 772 // Atomically check whether this is the main thread. 773 writeI32Const(os, flagAddress, "flag address"); 774 writeI32Const(os, 0, "expected flag value"); 775 writeI32Const(os, 1, "flag value"); 776 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 777 writeUleb128(os, WASM_OPCODE_I32_RMW_CMPXCHG, "i32.atomic.rmw.cmpxchg"); 778 writeMemArg(os, 2, 0); 779 writeU8(os, WASM_OPCODE_IF, "IF"); 780 writeU8(os, WASM_TYPE_NORESULT, "blocktype"); 781 782 // Did not increment 0, so wait for main thread to initialize memory 783 writeI32Const(os, flagAddress, "flag address"); 784 writeI32Const(os, 1, "expected flag value"); 785 writeI64Const(os, -1, "timeout"); 786 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 787 writeUleb128(os, WASM_OPCODE_I32_ATOMIC_WAIT, "i32.atomic.wait"); 788 writeMemArg(os, 2, 0); 789 writeU8(os, WASM_OPCODE_DROP, "drop"); 790 791 writeU8(os, WASM_OPCODE_ELSE, "ELSE"); 792 793 // Did increment 0, so conditionally initialize passive data segments 794 for (const OutputSegment *s : segments) { 795 if (s->initFlags & WASM_SEGMENT_IS_PASSIVE && s->name != ".tdata") { 796 // destination address 797 writeI32Const(os, s->startVA, "destination address"); 798 // source segment offset 799 writeI32Const(os, 0, "segment offset"); 800 // memory region size 801 writeI32Const(os, s->size, "memory region size"); 802 // memory.init instruction 803 writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix"); 804 writeUleb128(os, WASM_OPCODE_MEMORY_INIT, "memory.init"); 805 writeUleb128(os, s->index, "segment index immediate"); 806 writeU8(os, 0, "memory index immediate"); 807 } 808 } 809 810 // Set flag to 2 to mark end of initialization 811 writeI32Const(os, flagAddress, "flag address"); 812 writeI32Const(os, 2, "flag value"); 813 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 814 writeUleb128(os, WASM_OPCODE_I32_ATOMIC_STORE, "i32.atomic.store"); 815 writeMemArg(os, 2, 0); 816 817 // Notify any waiters that memory initialization is complete 818 writeI32Const(os, flagAddress, "flag address"); 819 writeI32Const(os, -1, "number of waiters"); 820 writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix"); 821 writeUleb128(os, WASM_OPCODE_ATOMIC_NOTIFY, "atomic.notify"); 822 writeMemArg(os, 2, 0); 823 writeU8(os, WASM_OPCODE_DROP, "drop"); 824 825 writeU8(os, WASM_OPCODE_END, "END"); 826 827 // Unconditionally drop passive data segments 828 for (const OutputSegment *s : segments) { 829 if (s->initFlags & WASM_SEGMENT_IS_PASSIVE && s->name != ".tdata") { 830 // data.drop instruction 831 writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix"); 832 writeUleb128(os, WASM_OPCODE_DATA_DROP, "data.drop"); 833 writeUleb128(os, s->index, "segment index immediate"); 834 } 835 } 836 } 837 writeU8(os, WASM_OPCODE_END, "END"); 838 } 839 840 createFunction(WasmSym::initMemory, bodyContent); 841 } 842 843 // For -shared (PIC) output, we create create a synthetic function which will 844 // apply any relocations to the data segments on startup. This function is 845 // called __wasm_apply_relocs and is added at the beginning of __wasm_call_ctors 846 // before any of the constructors run. 847 void Writer::createApplyRelocationsFunction() { 848 LLVM_DEBUG(dbgs() << "createApplyRelocationsFunction\n"); 849 // First write the body's contents to a string. 850 std::string bodyContent; 851 { 852 raw_string_ostream os(bodyContent); 853 writeUleb128(os, 0, "num locals"); 854 for (const OutputSegment *seg : segments) 855 for (const InputSegment *inSeg : seg->inputSegments) 856 inSeg->generateRelocationCode(os); 857 writeU8(os, WASM_OPCODE_END, "END"); 858 } 859 860 createFunction(WasmSym::applyRelocs, bodyContent); 861 } 862 863 // Create synthetic "__wasm_call_ctors" function based on ctor functions 864 // in input object. 865 void Writer::createCallCtorsFunction() { 866 if (!WasmSym::callCtors->isLive()) 867 return; 868 869 // First write the body's contents to a string. 870 std::string bodyContent; 871 { 872 raw_string_ostream os(bodyContent); 873 writeUleb128(os, 0, "num locals"); 874 875 if (config->isPic) { 876 writeU8(os, WASM_OPCODE_CALL, "CALL"); 877 writeUleb128(os, WasmSym::applyRelocs->getFunctionIndex(), 878 "function index"); 879 } 880 881 // Call constructors 882 for (const WasmInitEntry &f : initFunctions) { 883 writeU8(os, WASM_OPCODE_CALL, "CALL"); 884 writeUleb128(os, f.sym->getFunctionIndex(), "function index"); 885 } 886 writeU8(os, WASM_OPCODE_END, "END"); 887 } 888 889 createFunction(WasmSym::callCtors, bodyContent); 890 } 891 892 void Writer::createInitTLSFunction() { 893 if (!WasmSym::initTLS->isLive()) 894 return; 895 896 std::string bodyContent; 897 { 898 raw_string_ostream os(bodyContent); 899 900 OutputSegment *tlsSeg = nullptr; 901 for (auto *seg : segments) { 902 if (seg->name == ".tdata") { 903 tlsSeg = seg; 904 break; 905 } 906 } 907 908 writeUleb128(os, 0, "num locals"); 909 if (tlsSeg) { 910 writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get"); 911 writeUleb128(os, 0, "local index"); 912 913 writeU8(os, WASM_OPCODE_GLOBAL_SET, "global.set"); 914 writeUleb128(os, WasmSym::tlsBase->getGlobalIndex(), "global index"); 915 916 writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get"); 917 writeUleb128(os, 0, "local index"); 918 919 writeI32Const(os, 0, "segment offset"); 920 921 writeI32Const(os, tlsSeg->size, "memory region size"); 922 923 writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix"); 924 writeUleb128(os, WASM_OPCODE_MEMORY_INIT, "MEMORY.INIT"); 925 writeUleb128(os, tlsSeg->index, "segment index immediate"); 926 writeU8(os, 0, "memory index immediate"); 927 } 928 writeU8(os, WASM_OPCODE_END, "end function"); 929 } 930 931 createFunction(WasmSym::initTLS, bodyContent); 932 } 933 934 // Populate InitFunctions vector with init functions from all input objects. 935 // This is then used either when creating the output linking section or to 936 // synthesize the "__wasm_call_ctors" function. 937 void Writer::calculateInitFunctions() { 938 if (!config->relocatable && !WasmSym::callCtors->isLive()) 939 return; 940 941 for (ObjFile *file : symtab->objectFiles) { 942 const WasmLinkingData &l = file->getWasmObj()->linkingData(); 943 for (const WasmInitFunc &f : l.InitFunctions) { 944 FunctionSymbol *sym = file->getFunctionSymbol(f.Symbol); 945 // comdat exclusions can cause init functions be discarded. 946 if (sym->isDiscarded()) 947 continue; 948 assert(sym->isLive()); 949 if (*sym->signature != WasmSignature{{}, {}}) 950 error("invalid signature for init func: " + toString(*sym)); 951 LLVM_DEBUG(dbgs() << "initFunctions: " << toString(*sym) << "\n"); 952 initFunctions.emplace_back(WasmInitEntry{sym, f.Priority}); 953 } 954 } 955 956 // Sort in order of priority (lowest first) so that they are called 957 // in the correct order. 958 llvm::stable_sort(initFunctions, 959 [](const WasmInitEntry &l, const WasmInitEntry &r) { 960 return l.priority < r.priority; 961 }); 962 } 963 964 void Writer::createSyntheticSections() { 965 out.dylinkSec = make<DylinkSection>(); 966 out.typeSec = make<TypeSection>(); 967 out.importSec = make<ImportSection>(); 968 out.functionSec = make<FunctionSection>(); 969 out.tableSec = make<TableSection>(); 970 out.memorySec = make<MemorySection>(); 971 out.eventSec = make<EventSection>(); 972 out.globalSec = make<GlobalSection>(); 973 out.exportSec = make<ExportSection>(); 974 out.startSec = make<StartSection>(segments.size()); 975 out.elemSec = make<ElemSection>(); 976 out.dataCountSec = make<DataCountSection>(segments); 977 out.linkingSec = make<LinkingSection>(initFunctions, segments); 978 out.nameSec = make<NameSection>(); 979 out.producersSec = make<ProducersSection>(); 980 out.targetFeaturesSec = make<TargetFeaturesSection>(); 981 } 982 983 void Writer::run() { 984 if (config->relocatable || config->isPic) 985 config->globalBase = 0; 986 987 // For PIC code the table base is assigned dynamically by the loader. 988 // For non-PIC, we start at 1 so that accessing table index 0 always traps. 989 if (!config->isPic) { 990 config->tableBase = 1; 991 if (WasmSym::definedTableBase) 992 WasmSym::definedTableBase->setVirtualAddress(config->tableBase); 993 } 994 995 log("-- createOutputSegments"); 996 createOutputSegments(); 997 log("-- createSyntheticSections"); 998 createSyntheticSections(); 999 log("-- populateProducers"); 1000 populateProducers(); 1001 log("-- populateTargetFeatures"); 1002 populateTargetFeatures(); 1003 log("-- calculateImports"); 1004 calculateImports(); 1005 log("-- layoutMemory"); 1006 layoutMemory(); 1007 1008 if (!config->relocatable) { 1009 // Create linker synthesized __start_SECNAME/__stop_SECNAME symbols 1010 // This has to be done after memory layout is performed. 1011 for (const OutputSegment *seg : segments) 1012 addStartStopSymbols(seg); 1013 } 1014 1015 log("-- scanRelocations"); 1016 scanRelocations(); 1017 log("-- assignIndexes"); 1018 assignIndexes(); 1019 log("-- calculateInitFunctions"); 1020 calculateInitFunctions(); 1021 1022 if (!config->relocatable) { 1023 // Create linker synthesized functions 1024 if (config->sharedMemory) 1025 createInitMemoryFunction(); 1026 if (config->isPic) 1027 createApplyRelocationsFunction(); 1028 createCallCtorsFunction(); 1029 } 1030 1031 if (!config->relocatable && config->sharedMemory && !config->shared) 1032 createInitTLSFunction(); 1033 1034 if (errorCount()) 1035 return; 1036 1037 log("-- calculateTypes"); 1038 calculateTypes(); 1039 log("-- calculateExports"); 1040 calculateExports(); 1041 log("-- calculateCustomSections"); 1042 calculateCustomSections(); 1043 log("-- populateSymtab"); 1044 populateSymtab(); 1045 log("-- addSections"); 1046 addSections(); 1047 1048 if (errorHandler().verbose) { 1049 log("Defined Functions: " + Twine(out.functionSec->inputFunctions.size())); 1050 log("Defined Globals : " + Twine(out.globalSec->numGlobals())); 1051 log("Defined Events : " + Twine(out.eventSec->inputEvents.size())); 1052 log("Function Imports : " + 1053 Twine(out.importSec->getNumImportedFunctions())); 1054 log("Global Imports : " + Twine(out.importSec->getNumImportedGlobals())); 1055 log("Event Imports : " + Twine(out.importSec->getNumImportedEvents())); 1056 for (ObjFile *file : symtab->objectFiles) 1057 file->dumpInfo(); 1058 } 1059 1060 createHeader(); 1061 log("-- finalizeSections"); 1062 finalizeSections(); 1063 1064 log("-- openFile"); 1065 openFile(); 1066 if (errorCount()) 1067 return; 1068 1069 writeHeader(); 1070 1071 log("-- writeSections"); 1072 writeSections(); 1073 if (errorCount()) 1074 return; 1075 1076 if (Error e = buffer->commit()) 1077 fatal("failed to write the output file: " + toString(std::move(e))); 1078 } 1079 1080 // Open a result file. 1081 void Writer::openFile() { 1082 log("writing: " + config->outputFile); 1083 1084 Expected<std::unique_ptr<FileOutputBuffer>> bufferOrErr = 1085 FileOutputBuffer::create(config->outputFile, fileSize, 1086 FileOutputBuffer::F_executable); 1087 1088 if (!bufferOrErr) 1089 error("failed to open " + config->outputFile + ": " + 1090 toString(bufferOrErr.takeError())); 1091 else 1092 buffer = std::move(*bufferOrErr); 1093 } 1094 1095 void Writer::createHeader() { 1096 raw_string_ostream os(header); 1097 writeBytes(os, WasmMagic, sizeof(WasmMagic), "wasm magic"); 1098 writeU32(os, WasmVersion, "wasm version"); 1099 os.flush(); 1100 fileSize += header.size(); 1101 } 1102 1103 void writeResult() { Writer().run(); } 1104 1105 } // namespace wasm 1106 } // namespace lld 1107