1 //===- Driver.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 "Driver.h" 10 #include "Config.h" 11 #include "ICF.h" 12 #include "InputFiles.h" 13 #include "LTO.h" 14 #include "MarkLive.h" 15 #include "ObjC.h" 16 #include "OutputSection.h" 17 #include "OutputSegment.h" 18 #include "SectionPriorities.h" 19 #include "SymbolTable.h" 20 #include "Symbols.h" 21 #include "SyntheticSections.h" 22 #include "Target.h" 23 #include "UnwindInfoSection.h" 24 #include "Writer.h" 25 26 #include "lld/Common/Args.h" 27 #include "lld/Common/Driver.h" 28 #include "lld/Common/ErrorHandler.h" 29 #include "lld/Common/LLVM.h" 30 #include "lld/Common/Memory.h" 31 #include "lld/Common/Reproduce.h" 32 #include "lld/Common/Version.h" 33 #include "llvm/ADT/DenseSet.h" 34 #include "llvm/ADT/StringExtras.h" 35 #include "llvm/ADT/StringRef.h" 36 #include "llvm/BinaryFormat/MachO.h" 37 #include "llvm/BinaryFormat/Magic.h" 38 #include "llvm/Config/llvm-config.h" 39 #include "llvm/LTO/LTO.h" 40 #include "llvm/Object/Archive.h" 41 #include "llvm/Option/ArgList.h" 42 #include "llvm/Support/CommandLine.h" 43 #include "llvm/Support/FileSystem.h" 44 #include "llvm/Support/Host.h" 45 #include "llvm/Support/MemoryBuffer.h" 46 #include "llvm/Support/Parallel.h" 47 #include "llvm/Support/Path.h" 48 #include "llvm/Support/TarWriter.h" 49 #include "llvm/Support/TargetSelect.h" 50 #include "llvm/Support/TimeProfiler.h" 51 #include "llvm/TextAPI/PackedVersion.h" 52 53 #include <algorithm> 54 55 using namespace llvm; 56 using namespace llvm::MachO; 57 using namespace llvm::object; 58 using namespace llvm::opt; 59 using namespace llvm::sys; 60 using namespace lld; 61 using namespace lld::macho; 62 63 std::unique_ptr<Configuration> macho::config; 64 std::unique_ptr<DependencyTracker> macho::depTracker; 65 66 static HeaderFileType getOutputType(const InputArgList &args) { 67 // TODO: -r, -dylinker, -preload... 68 Arg *outputArg = args.getLastArg(OPT_bundle, OPT_dylib, OPT_execute); 69 if (outputArg == nullptr) 70 return MH_EXECUTE; 71 72 switch (outputArg->getOption().getID()) { 73 case OPT_bundle: 74 return MH_BUNDLE; 75 case OPT_dylib: 76 return MH_DYLIB; 77 case OPT_execute: 78 return MH_EXECUTE; 79 default: 80 llvm_unreachable("internal error"); 81 } 82 } 83 84 static DenseMap<CachedHashStringRef, StringRef> resolvedLibraries; 85 static Optional<StringRef> findLibrary(StringRef name) { 86 CachedHashStringRef key(name); 87 auto entry = resolvedLibraries.find(key); 88 if (entry != resolvedLibraries.end()) 89 return entry->second; 90 91 auto doFind = [&] { 92 if (config->searchDylibsFirst) { 93 if (Optional<StringRef> path = findPathCombination( 94 "lib" + name, config->librarySearchPaths, {".tbd", ".dylib"})) 95 return path; 96 return findPathCombination("lib" + name, config->librarySearchPaths, 97 {".a"}); 98 } 99 return findPathCombination("lib" + name, config->librarySearchPaths, 100 {".tbd", ".dylib", ".a"}); 101 }; 102 103 Optional<StringRef> path = doFind(); 104 if (path) 105 resolvedLibraries[key] = *path; 106 107 return path; 108 } 109 110 static DenseMap<CachedHashStringRef, StringRef> resolvedFrameworks; 111 static Optional<StringRef> findFramework(StringRef name) { 112 CachedHashStringRef key(name); 113 auto entry = resolvedFrameworks.find(key); 114 if (entry != resolvedFrameworks.end()) 115 return entry->second; 116 117 SmallString<260> symlink; 118 StringRef suffix; 119 std::tie(name, suffix) = name.split(","); 120 for (StringRef dir : config->frameworkSearchPaths) { 121 symlink = dir; 122 path::append(symlink, name + ".framework", name); 123 124 if (!suffix.empty()) { 125 // NOTE: we must resolve the symlink before trying the suffixes, because 126 // there are no symlinks for the suffixed paths. 127 SmallString<260> location; 128 if (!fs::real_path(symlink, location)) { 129 // only append suffix if realpath() succeeds 130 Twine suffixed = location + suffix; 131 if (fs::exists(suffixed)) 132 return resolvedFrameworks[key] = saver().save(suffixed.str()); 133 } 134 // Suffix lookup failed, fall through to the no-suffix case. 135 } 136 137 if (Optional<StringRef> path = resolveDylibPath(symlink.str())) 138 return resolvedFrameworks[key] = *path; 139 } 140 return {}; 141 } 142 143 static bool warnIfNotDirectory(StringRef option, StringRef path) { 144 if (!fs::exists(path)) { 145 warn("directory not found for option -" + option + path); 146 return false; 147 } else if (!fs::is_directory(path)) { 148 warn("option -" + option + path + " references a non-directory path"); 149 return false; 150 } 151 return true; 152 } 153 154 static std::vector<StringRef> 155 getSearchPaths(unsigned optionCode, InputArgList &args, 156 const std::vector<StringRef> &roots, 157 const SmallVector<StringRef, 2> &systemPaths) { 158 std::vector<StringRef> paths; 159 StringRef optionLetter{optionCode == OPT_F ? "F" : "L"}; 160 for (StringRef path : args::getStrings(args, optionCode)) { 161 // NOTE: only absolute paths are re-rooted to syslibroot(s) 162 bool found = false; 163 if (path::is_absolute(path, path::Style::posix)) { 164 for (StringRef root : roots) { 165 SmallString<261> buffer(root); 166 path::append(buffer, path); 167 // Do not warn about paths that are computed via the syslib roots 168 if (fs::is_directory(buffer)) { 169 paths.push_back(saver().save(buffer.str())); 170 found = true; 171 } 172 } 173 } 174 if (!found && warnIfNotDirectory(optionLetter, path)) 175 paths.push_back(path); 176 } 177 178 // `-Z` suppresses the standard "system" search paths. 179 if (args.hasArg(OPT_Z)) 180 return paths; 181 182 for (const StringRef &path : systemPaths) { 183 for (const StringRef &root : roots) { 184 SmallString<261> buffer(root); 185 path::append(buffer, path); 186 if (fs::is_directory(buffer)) 187 paths.push_back(saver().save(buffer.str())); 188 } 189 } 190 return paths; 191 } 192 193 static std::vector<StringRef> getSystemLibraryRoots(InputArgList &args) { 194 std::vector<StringRef> roots; 195 for (const Arg *arg : args.filtered(OPT_syslibroot)) 196 roots.push_back(arg->getValue()); 197 // NOTE: the final `-syslibroot` being `/` will ignore all roots 198 if (!roots.empty() && roots.back() == "/") 199 roots.clear(); 200 // NOTE: roots can never be empty - add an empty root to simplify the library 201 // and framework search path computation. 202 if (roots.empty()) 203 roots.emplace_back(""); 204 return roots; 205 } 206 207 static std::vector<StringRef> 208 getLibrarySearchPaths(InputArgList &args, const std::vector<StringRef> &roots) { 209 return getSearchPaths(OPT_L, args, roots, {"/usr/lib", "/usr/local/lib"}); 210 } 211 212 static std::vector<StringRef> 213 getFrameworkSearchPaths(InputArgList &args, 214 const std::vector<StringRef> &roots) { 215 return getSearchPaths(OPT_F, args, roots, 216 {"/Library/Frameworks", "/System/Library/Frameworks"}); 217 } 218 219 static llvm::CachePruningPolicy getLTOCachePolicy(InputArgList &args) { 220 SmallString<128> ltoPolicy; 221 auto add = [<oPolicy](Twine val) { 222 if (!ltoPolicy.empty()) 223 ltoPolicy += ":"; 224 val.toVector(ltoPolicy); 225 }; 226 for (const Arg *arg : 227 args.filtered(OPT_thinlto_cache_policy, OPT_prune_interval_lto, 228 OPT_prune_after_lto, OPT_max_relative_cache_size_lto)) { 229 switch (arg->getOption().getID()) { 230 case OPT_thinlto_cache_policy: 231 add(arg->getValue()); 232 break; 233 case OPT_prune_interval_lto: 234 if (!strcmp("-1", arg->getValue())) 235 add("prune_interval=87600h"); // 10 years 236 else 237 add(Twine("prune_interval=") + arg->getValue() + "s"); 238 break; 239 case OPT_prune_after_lto: 240 add(Twine("prune_after=") + arg->getValue() + "s"); 241 break; 242 case OPT_max_relative_cache_size_lto: 243 add(Twine("cache_size=") + arg->getValue() + "%"); 244 break; 245 } 246 } 247 return CHECK(parseCachePruningPolicy(ltoPolicy), "invalid LTO cache policy"); 248 } 249 250 // What caused a given library to be loaded. Only relevant for archives. 251 // Note that this does not tell us *how* we should load the library, i.e. 252 // whether we should do it lazily or eagerly (AKA force loading). The "how" is 253 // decided within addFile(). 254 enum class LoadType { 255 CommandLine, // Library was passed as a regular CLI argument 256 CommandLineForce, // Library was passed via `-force_load` 257 LCLinkerOption, // Library was passed via LC_LINKER_OPTIONS 258 }; 259 260 struct ArchiveFileInfo { 261 ArchiveFile *file; 262 bool isCommandLineLoad; 263 }; 264 265 static DenseMap<StringRef, ArchiveFileInfo> loadedArchives; 266 267 static InputFile *addFile(StringRef path, LoadType loadType, 268 bool isLazy = false, bool isExplicit = true, 269 bool isBundleLoader = false) { 270 Optional<MemoryBufferRef> buffer = readFile(path); 271 if (!buffer) 272 return nullptr; 273 MemoryBufferRef mbref = *buffer; 274 InputFile *newFile = nullptr; 275 276 file_magic magic = identify_magic(mbref.getBuffer()); 277 switch (magic) { 278 case file_magic::archive: { 279 bool isCommandLineLoad = loadType != LoadType::LCLinkerOption; 280 // Avoid loading archives twice. If the archives are being force-loaded, 281 // loading them twice would create duplicate symbol errors. In the 282 // non-force-loading case, this is just a minor performance optimization. 283 // We don't take a reference to cachedFile here because the 284 // loadArchiveMember() call below may recursively call addFile() and 285 // invalidate this reference. 286 auto entry = loadedArchives.find(path); 287 288 ArchiveFile *file; 289 if (entry == loadedArchives.end()) { 290 // No cached archive, we need to create a new one 291 std::unique_ptr<object::Archive> archive = CHECK( 292 object::Archive::create(mbref), path + ": failed to parse archive"); 293 294 if (!archive->isEmpty() && !archive->hasSymbolTable()) 295 error(path + ": archive has no index; run ranlib to add one"); 296 file = make<ArchiveFile>(std::move(archive)); 297 } else { 298 file = entry->second.file; 299 // Command-line loads take precedence. If file is previously loaded via 300 // command line, or is loaded via LC_LINKER_OPTION and being loaded via 301 // LC_LINKER_OPTION again, using the cached archive is enough. 302 if (entry->second.isCommandLineLoad || !isCommandLineLoad) 303 return file; 304 } 305 306 bool isLCLinkerForceLoad = loadType == LoadType::LCLinkerOption && 307 config->forceLoadSwift && 308 path::filename(path).startswith("libswift"); 309 if ((isCommandLineLoad && config->allLoad) || 310 loadType == LoadType::CommandLineForce || isLCLinkerForceLoad) { 311 if (Optional<MemoryBufferRef> buffer = readFile(path)) { 312 Error e = Error::success(); 313 for (const object::Archive::Child &c : file->getArchive().children(e)) { 314 StringRef reason; 315 switch (loadType) { 316 case LoadType::LCLinkerOption: 317 reason = "LC_LINKER_OPTION"; 318 break; 319 case LoadType::CommandLineForce: 320 reason = "-force_load"; 321 break; 322 case LoadType::CommandLine: 323 reason = "-all_load"; 324 break; 325 } 326 if (Error e = file->fetch(c, reason)) 327 error(toString(file) + ": " + reason + 328 " failed to load archive member: " + toString(std::move(e))); 329 } 330 if (e) 331 error(toString(file) + 332 ": Archive::children failed: " + toString(std::move(e))); 333 } 334 } else if (isCommandLineLoad && config->forceLoadObjC) { 335 for (const object::Archive::Symbol &sym : file->getArchive().symbols()) 336 if (sym.getName().startswith(objc::klass)) 337 file->fetch(sym); 338 339 // TODO: no need to look for ObjC sections for a given archive member if 340 // we already found that it contains an ObjC symbol. 341 if (Optional<MemoryBufferRef> buffer = readFile(path)) { 342 Error e = Error::success(); 343 for (const object::Archive::Child &c : file->getArchive().children(e)) { 344 Expected<MemoryBufferRef> mb = c.getMemoryBufferRef(); 345 if (!mb || !hasObjCSection(*mb)) 346 continue; 347 if (Error e = file->fetch(c, "-ObjC")) 348 error(toString(file) + ": -ObjC failed to load archive member: " + 349 toString(std::move(e))); 350 } 351 if (e) 352 error(toString(file) + 353 ": Archive::children failed: " + toString(std::move(e))); 354 } 355 } 356 357 file->addLazySymbols(); 358 loadedArchives[path] = ArchiveFileInfo{file, isCommandLineLoad}; 359 newFile = file; 360 break; 361 } 362 case file_magic::macho_object: 363 newFile = make<ObjFile>(mbref, getModTime(path), "", isLazy); 364 break; 365 case file_magic::macho_dynamically_linked_shared_lib: 366 case file_magic::macho_dynamically_linked_shared_lib_stub: 367 case file_magic::tapi_file: 368 if (DylibFile *dylibFile = 369 loadDylib(mbref, nullptr, /*isBundleLoader=*/false, isExplicit)) 370 newFile = dylibFile; 371 break; 372 case file_magic::bitcode: 373 newFile = make<BitcodeFile>(mbref, "", 0, isLazy); 374 break; 375 case file_magic::macho_executable: 376 case file_magic::macho_bundle: 377 // We only allow executable and bundle type here if it is used 378 // as a bundle loader. 379 if (!isBundleLoader) 380 error(path + ": unhandled file type"); 381 if (DylibFile *dylibFile = loadDylib(mbref, nullptr, isBundleLoader)) 382 newFile = dylibFile; 383 break; 384 default: 385 error(path + ": unhandled file type"); 386 } 387 if (newFile && !isa<DylibFile>(newFile)) { 388 if ((isa<ObjFile>(newFile) || isa<BitcodeFile>(newFile)) && newFile->lazy && 389 config->forceLoadObjC) { 390 for (Symbol *sym : newFile->symbols) 391 if (sym && sym->getName().startswith(objc::klass)) { 392 extract(*newFile, "-ObjC"); 393 break; 394 } 395 if (newFile->lazy && hasObjCSection(mbref)) 396 extract(*newFile, "-ObjC"); 397 } 398 399 // printArchiveMemberLoad() prints both .a and .o names, so no need to 400 // print the .a name here. Similarly skip lazy files. 401 if (config->printEachFile && magic != file_magic::archive && !isLazy) 402 message(toString(newFile)); 403 inputFiles.insert(newFile); 404 } 405 return newFile; 406 } 407 408 static void addLibrary(StringRef name, bool isNeeded, bool isWeak, 409 bool isReexport, bool isExplicit, LoadType loadType) { 410 if (Optional<StringRef> path = findLibrary(name)) { 411 if (auto *dylibFile = dyn_cast_or_null<DylibFile>( 412 addFile(*path, loadType, /*isLazy=*/false, isExplicit))) { 413 if (isNeeded) 414 dylibFile->forceNeeded = true; 415 if (isWeak) 416 dylibFile->forceWeakImport = true; 417 if (isReexport) { 418 config->hasReexports = true; 419 dylibFile->reexport = true; 420 } 421 } 422 return; 423 } 424 error("library not found for -l" + name); 425 } 426 427 static DenseSet<StringRef> loadedObjectFrameworks; 428 static void addFramework(StringRef name, bool isNeeded, bool isWeak, 429 bool isReexport, bool isExplicit, LoadType loadType) { 430 if (Optional<StringRef> path = findFramework(name)) { 431 if (loadedObjectFrameworks.contains(*path)) 432 return; 433 434 InputFile *file = 435 addFile(*path, loadType, /*isLazy=*/false, isExplicit, false); 436 if (auto *dylibFile = dyn_cast_or_null<DylibFile>(file)) { 437 if (isNeeded) 438 dylibFile->forceNeeded = true; 439 if (isWeak) 440 dylibFile->forceWeakImport = true; 441 if (isReexport) { 442 config->hasReexports = true; 443 dylibFile->reexport = true; 444 } 445 } else if (isa_and_nonnull<ObjFile>(file) || 446 isa_and_nonnull<BitcodeFile>(file)) { 447 // Cache frameworks containing object or bitcode files to avoid duplicate 448 // symbols. Frameworks containing static archives are cached separately 449 // in addFile() to share caching with libraries, and frameworks 450 // containing dylibs should allow overwriting of attributes such as 451 // forceNeeded by subsequent loads 452 loadedObjectFrameworks.insert(*path); 453 } 454 return; 455 } 456 error("framework not found for -framework " + name); 457 } 458 459 // Parses LC_LINKER_OPTION contents, which can add additional command line 460 // flags. This directly parses the flags instead of using the standard argument 461 // parser to improve performance. 462 void macho::parseLCLinkerOption(InputFile *f, unsigned argc, StringRef data) { 463 SmallVector<StringRef, 4> argv; 464 size_t offset = 0; 465 for (unsigned i = 0; i < argc && offset < data.size(); ++i) { 466 argv.push_back(data.data() + offset); 467 offset += strlen(data.data() + offset) + 1; 468 } 469 if (argv.size() != argc || offset > data.size()) 470 fatal(toString(f) + ": invalid LC_LINKER_OPTION"); 471 472 unsigned i = 0; 473 StringRef arg = argv[i]; 474 if (arg.consume_front("-l")) { 475 addLibrary(arg, /*isNeeded=*/false, /*isWeak=*/false, 476 /*isReexport=*/false, /*isExplicit=*/false, 477 LoadType::LCLinkerOption); 478 } else if (arg == "-framework") { 479 StringRef name = argv[++i]; 480 addFramework(name, /*isNeeded=*/false, /*isWeak=*/false, 481 /*isReexport=*/false, /*isExplicit=*/false, 482 LoadType::LCLinkerOption); 483 } else { 484 error(arg + " is not allowed in LC_LINKER_OPTION"); 485 } 486 } 487 488 static void addFileList(StringRef path, bool isLazy) { 489 Optional<MemoryBufferRef> buffer = readFile(path); 490 if (!buffer) 491 return; 492 MemoryBufferRef mbref = *buffer; 493 for (StringRef path : args::getLines(mbref)) 494 addFile(rerootPath(path), LoadType::CommandLine, isLazy); 495 } 496 497 // We expect sub-library names of the form "libfoo", which will match a dylib 498 // with a path of .*/libfoo.{dylib, tbd}. 499 // XXX ld64 seems to ignore the extension entirely when matching sub-libraries; 500 // I'm not sure what the use case for that is. 501 static bool markReexport(StringRef searchName, ArrayRef<StringRef> extensions) { 502 for (InputFile *file : inputFiles) { 503 if (auto *dylibFile = dyn_cast<DylibFile>(file)) { 504 StringRef filename = path::filename(dylibFile->getName()); 505 if (filename.consume_front(searchName) && 506 (filename.empty() || llvm::is_contained(extensions, filename))) { 507 dylibFile->reexport = true; 508 return true; 509 } 510 } 511 } 512 return false; 513 } 514 515 // This function is called on startup. We need this for LTO since 516 // LTO calls LLVM functions to compile bitcode files to native code. 517 // Technically this can be delayed until we read bitcode files, but 518 // we don't bother to do lazily because the initialization is fast. 519 static void initLLVM() { 520 InitializeAllTargets(); 521 InitializeAllTargetMCs(); 522 InitializeAllAsmPrinters(); 523 InitializeAllAsmParsers(); 524 } 525 526 static void compileBitcodeFiles() { 527 TimeTraceScope timeScope("LTO"); 528 auto *lto = make<BitcodeCompiler>(); 529 for (InputFile *file : inputFiles) 530 if (auto *bitcodeFile = dyn_cast<BitcodeFile>(file)) 531 if (!file->lazy) 532 lto->add(*bitcodeFile); 533 534 for (ObjFile *file : lto->compile()) 535 inputFiles.insert(file); 536 } 537 538 // Replaces common symbols with defined symbols residing in __common sections. 539 // This function must be called after all symbol names are resolved (i.e. after 540 // all InputFiles have been loaded.) As a result, later operations won't see 541 // any CommonSymbols. 542 static void replaceCommonSymbols() { 543 TimeTraceScope timeScope("Replace common symbols"); 544 ConcatOutputSection *osec = nullptr; 545 for (Symbol *sym : symtab->getSymbols()) { 546 auto *common = dyn_cast<CommonSymbol>(sym); 547 if (common == nullptr) 548 continue; 549 550 // Casting to size_t will truncate large values on 32-bit architectures, 551 // but it's not really worth supporting the linking of 64-bit programs on 552 // 32-bit archs. 553 ArrayRef<uint8_t> data = {nullptr, static_cast<size_t>(common->size)}; 554 // FIXME avoid creating one Section per symbol? 555 auto *section = 556 make<Section>(common->getFile(), segment_names::data, 557 section_names::common, S_ZEROFILL, /*addr=*/0); 558 auto *isec = make<ConcatInputSection>(*section, data, common->align); 559 if (!osec) 560 osec = ConcatOutputSection::getOrCreateForInput(isec); 561 isec->parent = osec; 562 inputSections.push_back(isec); 563 564 // FIXME: CommonSymbol should store isReferencedDynamically, noDeadStrip 565 // and pass them on here. 566 replaceSymbol<Defined>( 567 sym, sym->getName(), common->getFile(), isec, /*value=*/0, /*size=*/0, 568 /*isWeakDef=*/false, /*isExternal=*/true, common->privateExtern, 569 /*includeInSymtab=*/true, /*isThumb=*/false, 570 /*isReferencedDynamically=*/false, /*noDeadStrip=*/false); 571 } 572 } 573 574 static void initializeSectionRenameMap() { 575 if (config->dataConst) { 576 SmallVector<StringRef> v{section_names::got, 577 section_names::authGot, 578 section_names::authPtr, 579 section_names::nonLazySymbolPtr, 580 section_names::const_, 581 section_names::cfString, 582 section_names::moduleInitFunc, 583 section_names::moduleTermFunc, 584 section_names::objcClassList, 585 section_names::objcNonLazyClassList, 586 section_names::objcCatList, 587 section_names::objcNonLazyCatList, 588 section_names::objcProtoList, 589 section_names::objcImageInfo}; 590 for (StringRef s : v) 591 config->sectionRenameMap[{segment_names::data, s}] = { 592 segment_names::dataConst, s}; 593 } 594 config->sectionRenameMap[{segment_names::text, section_names::staticInit}] = { 595 segment_names::text, section_names::text}; 596 config->sectionRenameMap[{segment_names::import, section_names::pointers}] = { 597 config->dataConst ? segment_names::dataConst : segment_names::data, 598 section_names::nonLazySymbolPtr}; 599 } 600 601 static inline char toLowerDash(char x) { 602 if (x >= 'A' && x <= 'Z') 603 return x - 'A' + 'a'; 604 else if (x == ' ') 605 return '-'; 606 return x; 607 } 608 609 static std::string lowerDash(StringRef s) { 610 return std::string(map_iterator(s.begin(), toLowerDash), 611 map_iterator(s.end(), toLowerDash)); 612 } 613 614 struct PlatformVersion { 615 PlatformType platform = PLATFORM_UNKNOWN; 616 llvm::VersionTuple minimum; 617 llvm::VersionTuple sdk; 618 }; 619 620 static PlatformVersion parsePlatformVersion(const Arg *arg) { 621 assert(arg->getOption().getID() == OPT_platform_version); 622 StringRef platformStr = arg->getValue(0); 623 StringRef minVersionStr = arg->getValue(1); 624 StringRef sdkVersionStr = arg->getValue(2); 625 626 PlatformVersion platformVersion; 627 628 // TODO(compnerd) see if we can generate this case list via XMACROS 629 platformVersion.platform = 630 StringSwitch<PlatformType>(lowerDash(platformStr)) 631 .Cases("macos", "1", PLATFORM_MACOS) 632 .Cases("ios", "2", PLATFORM_IOS) 633 .Cases("tvos", "3", PLATFORM_TVOS) 634 .Cases("watchos", "4", PLATFORM_WATCHOS) 635 .Cases("bridgeos", "5", PLATFORM_BRIDGEOS) 636 .Cases("mac-catalyst", "6", PLATFORM_MACCATALYST) 637 .Cases("ios-simulator", "7", PLATFORM_IOSSIMULATOR) 638 .Cases("tvos-simulator", "8", PLATFORM_TVOSSIMULATOR) 639 .Cases("watchos-simulator", "9", PLATFORM_WATCHOSSIMULATOR) 640 .Cases("driverkit", "10", PLATFORM_DRIVERKIT) 641 .Default(PLATFORM_UNKNOWN); 642 if (platformVersion.platform == PLATFORM_UNKNOWN) 643 error(Twine("malformed platform: ") + platformStr); 644 // TODO: check validity of version strings, which varies by platform 645 // NOTE: ld64 accepts version strings with 5 components 646 // llvm::VersionTuple accepts no more than 4 components 647 // Has Apple ever published version strings with 5 components? 648 if (platformVersion.minimum.tryParse(minVersionStr)) 649 error(Twine("malformed minimum version: ") + minVersionStr); 650 if (platformVersion.sdk.tryParse(sdkVersionStr)) 651 error(Twine("malformed sdk version: ") + sdkVersionStr); 652 return platformVersion; 653 } 654 655 // Has the side-effect of setting Config::platformInfo. 656 static PlatformType parsePlatformVersions(const ArgList &args) { 657 std::map<PlatformType, PlatformVersion> platformVersions; 658 const PlatformVersion *lastVersionInfo = nullptr; 659 for (const Arg *arg : args.filtered(OPT_platform_version)) { 660 PlatformVersion version = parsePlatformVersion(arg); 661 662 // For each platform, the last flag wins: 663 // `-platform_version macos 2 3 -platform_version macos 4 5` has the same 664 // effect as just passing `-platform_version macos 4 5`. 665 // FIXME: ld64 warns on multiple flags for one platform. Should we? 666 platformVersions[version.platform] = version; 667 lastVersionInfo = &platformVersions[version.platform]; 668 } 669 670 if (platformVersions.empty()) { 671 error("must specify -platform_version"); 672 return PLATFORM_UNKNOWN; 673 } 674 if (platformVersions.size() > 2) { 675 error("must specify -platform_version at most twice"); 676 return PLATFORM_UNKNOWN; 677 } 678 if (platformVersions.size() == 2) { 679 bool isZipperedCatalyst = platformVersions.count(PLATFORM_MACOS) && 680 platformVersions.count(PLATFORM_MACCATALYST); 681 682 if (!isZipperedCatalyst) { 683 error("lld supports writing zippered outputs only for " 684 "macos and mac-catalyst"); 685 } else if (config->outputType != MH_DYLIB && 686 config->outputType != MH_BUNDLE) { 687 error("writing zippered outputs only valid for -dylib and -bundle"); 688 } else { 689 config->platformInfo.minimum = platformVersions[PLATFORM_MACOS].minimum; 690 config->platformInfo.sdk = platformVersions[PLATFORM_MACOS].sdk; 691 config->secondaryPlatformInfo = PlatformInfo{}; 692 config->secondaryPlatformInfo->minimum = 693 platformVersions[PLATFORM_MACCATALYST].minimum; 694 config->secondaryPlatformInfo->sdk = 695 platformVersions[PLATFORM_MACCATALYST].sdk; 696 } 697 return PLATFORM_MACOS; 698 } 699 700 config->platformInfo.minimum = lastVersionInfo->minimum; 701 config->platformInfo.sdk = lastVersionInfo->sdk; 702 return lastVersionInfo->platform; 703 } 704 705 // Has the side-effect of setting Config::target. 706 static TargetInfo *createTargetInfo(InputArgList &args) { 707 StringRef archName = args.getLastArgValue(OPT_arch); 708 if (archName.empty()) { 709 error("must specify -arch"); 710 return nullptr; 711 } 712 713 PlatformType platform = parsePlatformVersions(args); 714 config->platformInfo.target = 715 MachO::Target(getArchitectureFromName(archName), platform); 716 if (config->secondaryPlatformInfo) { 717 config->secondaryPlatformInfo->target = 718 MachO::Target(getArchitectureFromName(archName), PLATFORM_MACCATALYST); 719 } 720 721 uint32_t cpuType; 722 uint32_t cpuSubtype; 723 std::tie(cpuType, cpuSubtype) = getCPUTypeFromArchitecture(config->arch()); 724 725 switch (cpuType) { 726 case CPU_TYPE_X86_64: 727 return createX86_64TargetInfo(); 728 case CPU_TYPE_ARM64: 729 return createARM64TargetInfo(); 730 case CPU_TYPE_ARM64_32: 731 return createARM64_32TargetInfo(); 732 case CPU_TYPE_ARM: 733 return createARMTargetInfo(cpuSubtype); 734 default: 735 error("missing or unsupported -arch " + archName); 736 return nullptr; 737 } 738 } 739 740 static UndefinedSymbolTreatment 741 getUndefinedSymbolTreatment(const ArgList &args) { 742 StringRef treatmentStr = args.getLastArgValue(OPT_undefined); 743 auto treatment = 744 StringSwitch<UndefinedSymbolTreatment>(treatmentStr) 745 .Cases("error", "", UndefinedSymbolTreatment::error) 746 .Case("warning", UndefinedSymbolTreatment::warning) 747 .Case("suppress", UndefinedSymbolTreatment::suppress) 748 .Case("dynamic_lookup", UndefinedSymbolTreatment::dynamic_lookup) 749 .Default(UndefinedSymbolTreatment::unknown); 750 if (treatment == UndefinedSymbolTreatment::unknown) { 751 warn(Twine("unknown -undefined TREATMENT '") + treatmentStr + 752 "', defaulting to 'error'"); 753 treatment = UndefinedSymbolTreatment::error; 754 } else if (config->namespaceKind == NamespaceKind::twolevel && 755 (treatment == UndefinedSymbolTreatment::warning || 756 treatment == UndefinedSymbolTreatment::suppress)) { 757 if (treatment == UndefinedSymbolTreatment::warning) 758 error("'-undefined warning' only valid with '-flat_namespace'"); 759 else 760 error("'-undefined suppress' only valid with '-flat_namespace'"); 761 treatment = UndefinedSymbolTreatment::error; 762 } 763 return treatment; 764 } 765 766 static ICFLevel getICFLevel(const ArgList &args) { 767 StringRef icfLevelStr = args.getLastArgValue(OPT_icf_eq); 768 auto icfLevel = StringSwitch<ICFLevel>(icfLevelStr) 769 .Cases("none", "", ICFLevel::none) 770 .Case("safe", ICFLevel::safe) 771 .Case("all", ICFLevel::all) 772 .Default(ICFLevel::unknown); 773 if (icfLevel == ICFLevel::unknown) { 774 warn(Twine("unknown --icf=OPTION `") + icfLevelStr + 775 "', defaulting to `none'"); 776 icfLevel = ICFLevel::none; 777 } 778 return icfLevel; 779 } 780 781 static void warnIfDeprecatedOption(const Option &opt) { 782 if (!opt.getGroup().isValid()) 783 return; 784 if (opt.getGroup().getID() == OPT_grp_deprecated) { 785 warn("Option `" + opt.getPrefixedName() + "' is deprecated in ld64:"); 786 warn(opt.getHelpText()); 787 } 788 } 789 790 static void warnIfUnimplementedOption(const Option &opt) { 791 if (!opt.getGroup().isValid() || !opt.hasFlag(DriverFlag::HelpHidden)) 792 return; 793 switch (opt.getGroup().getID()) { 794 case OPT_grp_deprecated: 795 // warn about deprecated options elsewhere 796 break; 797 case OPT_grp_undocumented: 798 warn("Option `" + opt.getPrefixedName() + 799 "' is undocumented. Should lld implement it?"); 800 break; 801 case OPT_grp_obsolete: 802 warn("Option `" + opt.getPrefixedName() + 803 "' is obsolete. Please modernize your usage."); 804 break; 805 case OPT_grp_ignored: 806 warn("Option `" + opt.getPrefixedName() + "' is ignored."); 807 break; 808 case OPT_grp_ignored_silently: 809 break; 810 default: 811 warn("Option `" + opt.getPrefixedName() + 812 "' is not yet implemented. Stay tuned..."); 813 break; 814 } 815 } 816 817 static const char *getReproduceOption(InputArgList &args) { 818 if (const Arg *arg = args.getLastArg(OPT_reproduce)) 819 return arg->getValue(); 820 return getenv("LLD_REPRODUCE"); 821 } 822 823 static void parseClangOption(StringRef opt, const Twine &msg) { 824 std::string err; 825 raw_string_ostream os(err); 826 827 const char *argv[] = {"lld", opt.data()}; 828 if (cl::ParseCommandLineOptions(2, argv, "", &os)) 829 return; 830 os.flush(); 831 error(msg + ": " + StringRef(err).trim()); 832 } 833 834 static uint32_t parseDylibVersion(const ArgList &args, unsigned id) { 835 const Arg *arg = args.getLastArg(id); 836 if (!arg) 837 return 0; 838 839 if (config->outputType != MH_DYLIB) { 840 error(arg->getAsString(args) + ": only valid with -dylib"); 841 return 0; 842 } 843 844 PackedVersion version; 845 if (!version.parse32(arg->getValue())) { 846 error(arg->getAsString(args) + ": malformed version"); 847 return 0; 848 } 849 850 return version.rawValue(); 851 } 852 853 static uint32_t parseProtection(StringRef protStr) { 854 uint32_t prot = 0; 855 for (char c : protStr) { 856 switch (c) { 857 case 'r': 858 prot |= VM_PROT_READ; 859 break; 860 case 'w': 861 prot |= VM_PROT_WRITE; 862 break; 863 case 'x': 864 prot |= VM_PROT_EXECUTE; 865 break; 866 case '-': 867 break; 868 default: 869 error("unknown -segprot letter '" + Twine(c) + "' in " + protStr); 870 return 0; 871 } 872 } 873 return prot; 874 } 875 876 static std::vector<SectionAlign> parseSectAlign(const opt::InputArgList &args) { 877 std::vector<SectionAlign> sectAligns; 878 for (const Arg *arg : args.filtered(OPT_sectalign)) { 879 StringRef segName = arg->getValue(0); 880 StringRef sectName = arg->getValue(1); 881 StringRef alignStr = arg->getValue(2); 882 if (alignStr.startswith("0x") || alignStr.startswith("0X")) 883 alignStr = alignStr.drop_front(2); 884 uint32_t align; 885 if (alignStr.getAsInteger(16, align)) { 886 error("-sectalign: failed to parse '" + StringRef(arg->getValue(2)) + 887 "' as number"); 888 continue; 889 } 890 if (!isPowerOf2_32(align)) { 891 error("-sectalign: '" + StringRef(arg->getValue(2)) + 892 "' (in base 16) not a power of two"); 893 continue; 894 } 895 sectAligns.push_back({segName, sectName, align}); 896 } 897 return sectAligns; 898 } 899 900 PlatformType macho::removeSimulator(PlatformType platform) { 901 switch (platform) { 902 case PLATFORM_IOSSIMULATOR: 903 return PLATFORM_IOS; 904 case PLATFORM_TVOSSIMULATOR: 905 return PLATFORM_TVOS; 906 case PLATFORM_WATCHOSSIMULATOR: 907 return PLATFORM_WATCHOS; 908 default: 909 return platform; 910 } 911 } 912 913 static bool dataConstDefault(const InputArgList &args) { 914 static const std::vector<std::pair<PlatformType, VersionTuple>> minVersion = { 915 {PLATFORM_MACOS, VersionTuple(10, 15)}, 916 {PLATFORM_IOS, VersionTuple(13, 0)}, 917 {PLATFORM_TVOS, VersionTuple(13, 0)}, 918 {PLATFORM_WATCHOS, VersionTuple(6, 0)}, 919 {PLATFORM_BRIDGEOS, VersionTuple(4, 0)}}; 920 PlatformType platform = removeSimulator(config->platformInfo.target.Platform); 921 auto it = llvm::find_if(minVersion, 922 [&](const auto &p) { return p.first == platform; }); 923 if (it != minVersion.end()) 924 if (config->platformInfo.minimum < it->second) 925 return false; 926 927 switch (config->outputType) { 928 case MH_EXECUTE: 929 return !args.hasArg(OPT_no_pie); 930 case MH_BUNDLE: 931 // FIXME: return false when -final_name ... 932 // has prefix "/System/Library/UserEventPlugins/" 933 // or matches "/usr/libexec/locationd" "/usr/libexec/terminusd" 934 return true; 935 case MH_DYLIB: 936 return true; 937 case MH_OBJECT: 938 return false; 939 default: 940 llvm_unreachable( 941 "unsupported output type for determining data-const default"); 942 } 943 return false; 944 } 945 946 void SymbolPatterns::clear() { 947 literals.clear(); 948 globs.clear(); 949 } 950 951 void SymbolPatterns::insert(StringRef symbolName) { 952 if (symbolName.find_first_of("*?[]") == StringRef::npos) 953 literals.insert(CachedHashStringRef(symbolName)); 954 else if (Expected<GlobPattern> pattern = GlobPattern::create(symbolName)) 955 globs.emplace_back(*pattern); 956 else 957 error("invalid symbol-name pattern: " + symbolName); 958 } 959 960 bool SymbolPatterns::matchLiteral(StringRef symbolName) const { 961 return literals.contains(CachedHashStringRef(symbolName)); 962 } 963 964 bool SymbolPatterns::matchGlob(StringRef symbolName) const { 965 for (const GlobPattern &glob : globs) 966 if (glob.match(symbolName)) 967 return true; 968 return false; 969 } 970 971 bool SymbolPatterns::match(StringRef symbolName) const { 972 return matchLiteral(symbolName) || matchGlob(symbolName); 973 } 974 975 static void parseSymbolPatternsFile(const Arg *arg, 976 SymbolPatterns &symbolPatterns) { 977 StringRef path = arg->getValue(); 978 Optional<MemoryBufferRef> buffer = readFile(path); 979 if (!buffer) { 980 error("Could not read symbol file: " + path); 981 return; 982 } 983 MemoryBufferRef mbref = *buffer; 984 for (StringRef line : args::getLines(mbref)) { 985 line = line.take_until([](char c) { return c == '#'; }).trim(); 986 if (!line.empty()) 987 symbolPatterns.insert(line); 988 } 989 } 990 991 static void handleSymbolPatterns(InputArgList &args, 992 SymbolPatterns &symbolPatterns, 993 unsigned singleOptionCode, 994 unsigned listFileOptionCode) { 995 for (const Arg *arg : args.filtered(singleOptionCode)) 996 symbolPatterns.insert(arg->getValue()); 997 for (const Arg *arg : args.filtered(listFileOptionCode)) 998 parseSymbolPatternsFile(arg, symbolPatterns); 999 } 1000 1001 static void createFiles(const InputArgList &args) { 1002 TimeTraceScope timeScope("Load input files"); 1003 // This loop should be reserved for options whose exact ordering matters. 1004 // Other options should be handled via filtered() and/or getLastArg(). 1005 bool isLazy = false; 1006 for (const Arg *arg : args) { 1007 const Option &opt = arg->getOption(); 1008 warnIfDeprecatedOption(opt); 1009 warnIfUnimplementedOption(opt); 1010 1011 switch (opt.getID()) { 1012 case OPT_INPUT: 1013 addFile(rerootPath(arg->getValue()), LoadType::CommandLine, isLazy); 1014 break; 1015 case OPT_needed_library: 1016 if (auto *dylibFile = dyn_cast_or_null<DylibFile>( 1017 addFile(rerootPath(arg->getValue()), LoadType::CommandLine))) 1018 dylibFile->forceNeeded = true; 1019 break; 1020 case OPT_reexport_library: 1021 if (auto *dylibFile = dyn_cast_or_null<DylibFile>( 1022 addFile(rerootPath(arg->getValue()), LoadType::CommandLine))) { 1023 config->hasReexports = true; 1024 dylibFile->reexport = true; 1025 } 1026 break; 1027 case OPT_weak_library: 1028 if (auto *dylibFile = dyn_cast_or_null<DylibFile>( 1029 addFile(rerootPath(arg->getValue()), LoadType::CommandLine))) 1030 dylibFile->forceWeakImport = true; 1031 break; 1032 case OPT_filelist: 1033 addFileList(arg->getValue(), isLazy); 1034 break; 1035 case OPT_force_load: 1036 addFile(rerootPath(arg->getValue()), LoadType::CommandLineForce); 1037 break; 1038 case OPT_l: 1039 case OPT_needed_l: 1040 case OPT_reexport_l: 1041 case OPT_weak_l: 1042 addLibrary(arg->getValue(), opt.getID() == OPT_needed_l, 1043 opt.getID() == OPT_weak_l, opt.getID() == OPT_reexport_l, 1044 /*isExplicit=*/true, LoadType::CommandLine); 1045 break; 1046 case OPT_framework: 1047 case OPT_needed_framework: 1048 case OPT_reexport_framework: 1049 case OPT_weak_framework: 1050 addFramework(arg->getValue(), opt.getID() == OPT_needed_framework, 1051 opt.getID() == OPT_weak_framework, 1052 opt.getID() == OPT_reexport_framework, /*isExplicit=*/true, 1053 LoadType::CommandLine); 1054 break; 1055 case OPT_start_lib: 1056 if (isLazy) 1057 error("nested --start-lib"); 1058 isLazy = true; 1059 break; 1060 case OPT_end_lib: 1061 if (!isLazy) 1062 error("stray --end-lib"); 1063 isLazy = false; 1064 break; 1065 default: 1066 break; 1067 } 1068 } 1069 } 1070 1071 static void gatherInputSections() { 1072 TimeTraceScope timeScope("Gathering input sections"); 1073 int inputOrder = 0; 1074 for (const InputFile *file : inputFiles) { 1075 for (const Section *section : file->sections) { 1076 // Compact unwind entries require special handling elsewhere. (In 1077 // contrast, EH frames are handled like regular ConcatInputSections.) 1078 if (section->name == section_names::compactUnwind) 1079 continue; 1080 ConcatOutputSection *osec = nullptr; 1081 for (const Subsection &subsection : section->subsections) { 1082 if (auto *isec = dyn_cast<ConcatInputSection>(subsection.isec)) { 1083 if (isec->isCoalescedWeak()) 1084 continue; 1085 isec->outSecOff = inputOrder++; 1086 if (!osec) 1087 osec = ConcatOutputSection::getOrCreateForInput(isec); 1088 isec->parent = osec; 1089 inputSections.push_back(isec); 1090 } else if (auto *isec = 1091 dyn_cast<CStringInputSection>(subsection.isec)) { 1092 if (in.cStringSection->inputOrder == UnspecifiedInputOrder) 1093 in.cStringSection->inputOrder = inputOrder++; 1094 in.cStringSection->addInput(isec); 1095 } else if (auto *isec = 1096 dyn_cast<WordLiteralInputSection>(subsection.isec)) { 1097 if (in.wordLiteralSection->inputOrder == UnspecifiedInputOrder) 1098 in.wordLiteralSection->inputOrder = inputOrder++; 1099 in.wordLiteralSection->addInput(isec); 1100 } else { 1101 llvm_unreachable("unexpected input section kind"); 1102 } 1103 } 1104 } 1105 } 1106 assert(inputOrder <= UnspecifiedInputOrder); 1107 } 1108 1109 static void foldIdenticalLiterals() { 1110 TimeTraceScope timeScope("Fold identical literals"); 1111 // We always create a cStringSection, regardless of whether dedupLiterals is 1112 // true. If it isn't, we simply create a non-deduplicating CStringSection. 1113 // Either way, we must unconditionally finalize it here. 1114 in.cStringSection->finalizeContents(); 1115 if (in.wordLiteralSection) 1116 in.wordLiteralSection->finalizeContents(); 1117 } 1118 1119 static void referenceStubBinder() { 1120 bool needsStubHelper = config->outputType == MH_DYLIB || 1121 config->outputType == MH_EXECUTE || 1122 config->outputType == MH_BUNDLE; 1123 if (!needsStubHelper || !symtab->find("dyld_stub_binder")) 1124 return; 1125 1126 // dyld_stub_binder is used by dyld to resolve lazy bindings. This code here 1127 // adds a opportunistic reference to dyld_stub_binder if it happens to exist. 1128 // dyld_stub_binder is in libSystem.dylib, which is usually linked in. This 1129 // isn't needed for correctness, but the presence of that symbol suppresses 1130 // "no symbols" diagnostics from `nm`. 1131 // StubHelperSection::setup() adds a reference and errors out if 1132 // dyld_stub_binder doesn't exist in case it is actually needed. 1133 symtab->addUndefined("dyld_stub_binder", /*file=*/nullptr, /*isWeak=*/false); 1134 } 1135 1136 bool macho::link(ArrayRef<const char *> argsArr, llvm::raw_ostream &stdoutOS, 1137 llvm::raw_ostream &stderrOS, bool exitEarly, 1138 bool disableOutput) { 1139 // This driver-specific context will be freed later by lldMain(). 1140 auto *ctx = new CommonLinkerContext; 1141 1142 ctx->e.initialize(stdoutOS, stderrOS, exitEarly, disableOutput); 1143 ctx->e.cleanupCallback = []() { 1144 resolvedFrameworks.clear(); 1145 resolvedLibraries.clear(); 1146 cachedReads.clear(); 1147 concatOutputSections.clear(); 1148 inputFiles.clear(); 1149 inputSections.clear(); 1150 loadedArchives.clear(); 1151 loadedObjectFrameworks.clear(); 1152 syntheticSections.clear(); 1153 thunkMap.clear(); 1154 1155 firstTLVDataSection = nullptr; 1156 tar = nullptr; 1157 memset(&in, 0, sizeof(in)); 1158 1159 resetLoadedDylibs(); 1160 resetOutputSegments(); 1161 resetWriter(); 1162 InputFile::resetIdCount(); 1163 }; 1164 1165 ctx->e.logName = args::getFilenameWithoutExe(argsArr[0]); 1166 1167 MachOOptTable parser; 1168 InputArgList args = parser.parse(argsArr.slice(1)); 1169 1170 ctx->e.errorLimitExceededMsg = "too many errors emitted, stopping now " 1171 "(use --error-limit=0 to see all errors)"; 1172 ctx->e.errorLimit = args::getInteger(args, OPT_error_limit_eq, 20); 1173 ctx->e.verbose = args.hasArg(OPT_verbose); 1174 1175 if (args.hasArg(OPT_help_hidden)) { 1176 parser.printHelp(argsArr[0], /*showHidden=*/true); 1177 return true; 1178 } 1179 if (args.hasArg(OPT_help)) { 1180 parser.printHelp(argsArr[0], /*showHidden=*/false); 1181 return true; 1182 } 1183 if (args.hasArg(OPT_version)) { 1184 message(getLLDVersion()); 1185 return true; 1186 } 1187 1188 config = std::make_unique<Configuration>(); 1189 symtab = std::make_unique<SymbolTable>(); 1190 config->outputType = getOutputType(args); 1191 target = createTargetInfo(args); 1192 depTracker = std::make_unique<DependencyTracker>( 1193 args.getLastArgValue(OPT_dependency_info)); 1194 if (errorCount()) 1195 return false; 1196 1197 if (args.hasArg(OPT_pagezero_size)) { 1198 uint64_t pagezeroSize = args::getHex(args, OPT_pagezero_size, 0); 1199 1200 // ld64 does something really weird. It attempts to realign the value to the 1201 // page size, but assumes the the page size is 4K. This doesn't work with 1202 // most of Apple's ARM64 devices, which use a page size of 16K. This means 1203 // that it will first 4K align it by rounding down, then round up to 16K. 1204 // This probably only happened because no one using this arg with anything 1205 // other then 0, so no one checked if it did what is what it says it does. 1206 1207 // So we are not copying this weird behavior and doing the it in a logical 1208 // way, by always rounding down to page size. 1209 if (!isAligned(Align(target->getPageSize()), pagezeroSize)) { 1210 pagezeroSize -= pagezeroSize % target->getPageSize(); 1211 warn("__PAGEZERO size is not page aligned, rounding down to 0x" + 1212 Twine::utohexstr(pagezeroSize)); 1213 } 1214 1215 target->pageZeroSize = pagezeroSize; 1216 } 1217 1218 config->osoPrefix = args.getLastArgValue(OPT_oso_prefix); 1219 if (!config->osoPrefix.empty()) { 1220 // Expand special characters, such as ".", "..", or "~", if present. 1221 // Note: LD64 only expands "." and not other special characters. 1222 // That seems silly to imitate so we will not try to follow it, but rather 1223 // just use real_path() to do it. 1224 1225 // The max path length is 4096, in theory. However that seems quite long 1226 // and seems unlikely that any one would want to strip everything from the 1227 // path. Hence we've picked a reasonably large number here. 1228 SmallString<1024> expanded; 1229 if (!fs::real_path(config->osoPrefix, expanded, 1230 /*expand_tilde=*/true)) { 1231 // Note: LD64 expands "." to be `<current_dir>/` 1232 // (ie., it has a slash suffix) whereas real_path() doesn't. 1233 // So we have to append '/' to be consistent. 1234 StringRef sep = sys::path::get_separator(); 1235 // real_path removes trailing slashes as part of the normalization, but 1236 // these are meaningful for our text based stripping 1237 if (config->osoPrefix.equals(".") || config->osoPrefix.endswith(sep)) 1238 expanded += sep; 1239 config->osoPrefix = saver().save(expanded.str()); 1240 } 1241 } 1242 1243 // Must be set before any InputSections and Symbols are created. 1244 config->deadStrip = args.hasArg(OPT_dead_strip); 1245 1246 config->systemLibraryRoots = getSystemLibraryRoots(args); 1247 if (const char *path = getReproduceOption(args)) { 1248 // Note that --reproduce is a debug option so you can ignore it 1249 // if you are trying to understand the whole picture of the code. 1250 Expected<std::unique_ptr<TarWriter>> errOrWriter = 1251 TarWriter::create(path, path::stem(path)); 1252 if (errOrWriter) { 1253 tar = std::move(*errOrWriter); 1254 tar->append("response.txt", createResponseFile(args)); 1255 tar->append("version.txt", getLLDVersion() + "\n"); 1256 } else { 1257 error("--reproduce: " + toString(errOrWriter.takeError())); 1258 } 1259 } 1260 1261 if (auto *arg = args.getLastArg(OPT_threads_eq)) { 1262 StringRef v(arg->getValue()); 1263 unsigned threads = 0; 1264 if (!llvm::to_integer(v, threads, 0) || threads == 0) 1265 error(arg->getSpelling() + ": expected a positive integer, but got '" + 1266 arg->getValue() + "'"); 1267 parallel::strategy = hardware_concurrency(threads); 1268 config->thinLTOJobs = v; 1269 } 1270 if (auto *arg = args.getLastArg(OPT_thinlto_jobs_eq)) 1271 config->thinLTOJobs = arg->getValue(); 1272 if (!get_threadpool_strategy(config->thinLTOJobs)) 1273 error("--thinlto-jobs: invalid job count: " + config->thinLTOJobs); 1274 1275 for (const Arg *arg : args.filtered(OPT_u)) { 1276 config->explicitUndefineds.push_back(symtab->addUndefined( 1277 arg->getValue(), /*file=*/nullptr, /*isWeakRef=*/false)); 1278 } 1279 1280 for (const Arg *arg : args.filtered(OPT_U)) 1281 config->explicitDynamicLookups.insert(arg->getValue()); 1282 1283 config->mapFile = args.getLastArgValue(OPT_map); 1284 config->optimize = args::getInteger(args, OPT_O, 1); 1285 config->outputFile = args.getLastArgValue(OPT_o, "a.out"); 1286 config->finalOutput = 1287 args.getLastArgValue(OPT_final_output, config->outputFile); 1288 config->astPaths = args.getAllArgValues(OPT_add_ast_path); 1289 config->headerPad = args::getHex(args, OPT_headerpad, /*Default=*/32); 1290 config->headerPadMaxInstallNames = 1291 args.hasArg(OPT_headerpad_max_install_names); 1292 config->printDylibSearch = 1293 args.hasArg(OPT_print_dylib_search) || getenv("RC_TRACE_DYLIB_SEARCHING"); 1294 config->printEachFile = args.hasArg(OPT_t); 1295 config->printWhyLoad = args.hasArg(OPT_why_load); 1296 config->omitDebugInfo = args.hasArg(OPT_S); 1297 config->errorForArchMismatch = args.hasArg(OPT_arch_errors_fatal); 1298 if (const Arg *arg = args.getLastArg(OPT_bundle_loader)) { 1299 if (config->outputType != MH_BUNDLE) 1300 error("-bundle_loader can only be used with MachO bundle output"); 1301 addFile(arg->getValue(), LoadType::CommandLine, /*isLazy=*/false, 1302 /*isExplicit=*/false, /*isBundleLoader=*/true); 1303 } 1304 if (const Arg *arg = args.getLastArg(OPT_umbrella)) { 1305 if (config->outputType != MH_DYLIB) 1306 warn("-umbrella used, but not creating dylib"); 1307 config->umbrella = arg->getValue(); 1308 } 1309 config->ltoObjPath = args.getLastArgValue(OPT_object_path_lto); 1310 config->ltoo = args::getInteger(args, OPT_lto_O, 2); 1311 if (config->ltoo > 3) 1312 error("--lto-O: invalid optimization level: " + Twine(config->ltoo)); 1313 config->thinLTOCacheDir = args.getLastArgValue(OPT_cache_path_lto); 1314 config->thinLTOCachePolicy = getLTOCachePolicy(args); 1315 config->runtimePaths = args::getStrings(args, OPT_rpath); 1316 config->allLoad = args.hasFlag(OPT_all_load, OPT_noall_load, false); 1317 config->archMultiple = args.hasArg(OPT_arch_multiple); 1318 config->applicationExtension = args.hasFlag( 1319 OPT_application_extension, OPT_no_application_extension, false); 1320 config->exportDynamic = args.hasArg(OPT_export_dynamic); 1321 config->forceLoadObjC = args.hasArg(OPT_ObjC); 1322 config->forceLoadSwift = args.hasArg(OPT_force_load_swift_libs); 1323 config->deadStripDylibs = args.hasArg(OPT_dead_strip_dylibs); 1324 config->demangle = args.hasArg(OPT_demangle); 1325 config->implicitDylibs = !args.hasArg(OPT_no_implicit_dylibs); 1326 config->emitFunctionStarts = 1327 args.hasFlag(OPT_function_starts, OPT_no_function_starts, true); 1328 config->emitBitcodeBundle = args.hasArg(OPT_bitcode_bundle); 1329 config->emitDataInCodeInfo = 1330 args.hasFlag(OPT_data_in_code_info, OPT_no_data_in_code_info, true); 1331 config->icfLevel = getICFLevel(args); 1332 config->dedupLiterals = 1333 args.hasFlag(OPT_deduplicate_literals, OPT_icf_eq, false) || 1334 config->icfLevel != ICFLevel::none; 1335 config->warnDylibInstallName = args.hasFlag( 1336 OPT_warn_dylib_install_name, OPT_no_warn_dylib_install_name, false); 1337 config->ignoreOptimizationHints = args.hasArg(OPT_ignore_optimization_hints); 1338 config->callGraphProfileSort = args.hasFlag( 1339 OPT_call_graph_profile_sort, OPT_no_call_graph_profile_sort, true); 1340 config->printSymbolOrder = args.getLastArgValue(OPT_print_symbol_order); 1341 1342 for (const Arg *arg : args.filtered(OPT_alias)) { 1343 config->aliasedSymbols.push_back( 1344 std::make_pair(arg->getValue(0), arg->getValue(1))); 1345 } 1346 1347 // FIXME: Add a commandline flag for this too. 1348 config->zeroModTime = getenv("ZERO_AR_DATE"); 1349 1350 std::array<PlatformType, 3> encryptablePlatforms{ 1351 PLATFORM_IOS, PLATFORM_WATCHOS, PLATFORM_TVOS}; 1352 config->emitEncryptionInfo = 1353 args.hasFlag(OPT_encryptable, OPT_no_encryption, 1354 is_contained(encryptablePlatforms, config->platform())); 1355 1356 #ifndef LLVM_HAVE_LIBXAR 1357 if (config->emitBitcodeBundle) 1358 error("-bitcode_bundle unsupported because LLD wasn't built with libxar"); 1359 #endif 1360 1361 if (const Arg *arg = args.getLastArg(OPT_install_name)) { 1362 if (config->warnDylibInstallName && config->outputType != MH_DYLIB) 1363 warn( 1364 arg->getAsString(args) + 1365 ": ignored, only has effect with -dylib [--warn-dylib-install-name]"); 1366 else 1367 config->installName = arg->getValue(); 1368 } else if (config->outputType == MH_DYLIB) { 1369 config->installName = config->finalOutput; 1370 } 1371 1372 if (args.hasArg(OPT_mark_dead_strippable_dylib)) { 1373 if (config->outputType != MH_DYLIB) 1374 warn("-mark_dead_strippable_dylib: ignored, only has effect with -dylib"); 1375 else 1376 config->markDeadStrippableDylib = true; 1377 } 1378 1379 if (const Arg *arg = args.getLastArg(OPT_static, OPT_dynamic)) 1380 config->staticLink = (arg->getOption().getID() == OPT_static); 1381 1382 if (const Arg *arg = 1383 args.getLastArg(OPT_flat_namespace, OPT_twolevel_namespace)) 1384 config->namespaceKind = arg->getOption().getID() == OPT_twolevel_namespace 1385 ? NamespaceKind::twolevel 1386 : NamespaceKind::flat; 1387 1388 config->undefinedSymbolTreatment = getUndefinedSymbolTreatment(args); 1389 1390 if (config->outputType == MH_EXECUTE) 1391 config->entry = symtab->addUndefined(args.getLastArgValue(OPT_e, "_main"), 1392 /*file=*/nullptr, 1393 /*isWeakRef=*/false); 1394 1395 config->librarySearchPaths = 1396 getLibrarySearchPaths(args, config->systemLibraryRoots); 1397 config->frameworkSearchPaths = 1398 getFrameworkSearchPaths(args, config->systemLibraryRoots); 1399 if (const Arg *arg = 1400 args.getLastArg(OPT_search_paths_first, OPT_search_dylibs_first)) 1401 config->searchDylibsFirst = 1402 arg->getOption().getID() == OPT_search_dylibs_first; 1403 1404 config->dylibCompatibilityVersion = 1405 parseDylibVersion(args, OPT_compatibility_version); 1406 config->dylibCurrentVersion = parseDylibVersion(args, OPT_current_version); 1407 1408 config->dataConst = 1409 args.hasFlag(OPT_data_const, OPT_no_data_const, dataConstDefault(args)); 1410 // Populate config->sectionRenameMap with builtin default renames. 1411 // Options -rename_section and -rename_segment are able to override. 1412 initializeSectionRenameMap(); 1413 // Reject every special character except '.' and '$' 1414 // TODO(gkm): verify that this is the proper set of invalid chars 1415 StringRef invalidNameChars("!\"#%&'()*+,-/:;<=>?@[\\]^`{|}~"); 1416 auto validName = [invalidNameChars](StringRef s) { 1417 if (s.find_first_of(invalidNameChars) != StringRef::npos) 1418 error("invalid name for segment or section: " + s); 1419 return s; 1420 }; 1421 for (const Arg *arg : args.filtered(OPT_rename_section)) { 1422 config->sectionRenameMap[{validName(arg->getValue(0)), 1423 validName(arg->getValue(1))}] = { 1424 validName(arg->getValue(2)), validName(arg->getValue(3))}; 1425 } 1426 for (const Arg *arg : args.filtered(OPT_rename_segment)) { 1427 config->segmentRenameMap[validName(arg->getValue(0))] = 1428 validName(arg->getValue(1)); 1429 } 1430 1431 config->sectionAlignments = parseSectAlign(args); 1432 1433 for (const Arg *arg : args.filtered(OPT_segprot)) { 1434 StringRef segName = arg->getValue(0); 1435 uint32_t maxProt = parseProtection(arg->getValue(1)); 1436 uint32_t initProt = parseProtection(arg->getValue(2)); 1437 if (maxProt != initProt && config->arch() != AK_i386) 1438 error("invalid argument '" + arg->getAsString(args) + 1439 "': max and init must be the same for non-i386 archs"); 1440 if (segName == segment_names::linkEdit) 1441 error("-segprot cannot be used to change __LINKEDIT's protections"); 1442 config->segmentProtections.push_back({segName, maxProt, initProt}); 1443 } 1444 1445 config->hasExplicitExports = 1446 args.hasArg(OPT_no_exported_symbols) || 1447 args.hasArgNoClaim(OPT_exported_symbol, OPT_exported_symbols_list); 1448 handleSymbolPatterns(args, config->exportedSymbols, OPT_exported_symbol, 1449 OPT_exported_symbols_list); 1450 handleSymbolPatterns(args, config->unexportedSymbols, OPT_unexported_symbol, 1451 OPT_unexported_symbols_list); 1452 if (config->hasExplicitExports && !config->unexportedSymbols.empty()) 1453 error("cannot use both -exported_symbol* and -unexported_symbol* options"); 1454 1455 if (args.hasArg(OPT_no_exported_symbols) && !config->exportedSymbols.empty()) 1456 error("cannot use both -exported_symbol* and -no_exported_symbols options"); 1457 1458 // Imitating LD64's: 1459 // -non_global_symbols_no_strip_list and -non_global_symbols_strip_list can't 1460 // both be present. 1461 // But -x can be used with either of these two, in which case, the last arg 1462 // takes effect. 1463 // (TODO: This is kind of confusing - considering disallowing using them 1464 // together for a more straightforward behaviour) 1465 { 1466 bool includeLocal = false; 1467 bool excludeLocal = false; 1468 for (const Arg *arg : 1469 args.filtered(OPT_x, OPT_non_global_symbols_no_strip_list, 1470 OPT_non_global_symbols_strip_list)) { 1471 switch (arg->getOption().getID()) { 1472 case OPT_x: 1473 config->localSymbolsPresence = SymtabPresence::None; 1474 break; 1475 case OPT_non_global_symbols_no_strip_list: 1476 if (excludeLocal) { 1477 error("cannot use both -non_global_symbols_no_strip_list and " 1478 "-non_global_symbols_strip_list"); 1479 } else { 1480 includeLocal = true; 1481 config->localSymbolsPresence = SymtabPresence::SelectivelyIncluded; 1482 parseSymbolPatternsFile(arg, config->localSymbolPatterns); 1483 } 1484 break; 1485 case OPT_non_global_symbols_strip_list: 1486 if (includeLocal) { 1487 error("cannot use both -non_global_symbols_no_strip_list and " 1488 "-non_global_symbols_strip_list"); 1489 } else { 1490 excludeLocal = true; 1491 config->localSymbolsPresence = SymtabPresence::SelectivelyExcluded; 1492 parseSymbolPatternsFile(arg, config->localSymbolPatterns); 1493 } 1494 break; 1495 default: 1496 llvm_unreachable("unexpected option"); 1497 } 1498 } 1499 } 1500 // Explicitly-exported literal symbols must be defined, but might 1501 // languish in an archive if unreferenced elsewhere or if they are in the 1502 // non-global strip list. Light a fire under those lazy symbols! 1503 for (const CachedHashStringRef &cachedName : config->exportedSymbols.literals) 1504 symtab->addUndefined(cachedName.val(), /*file=*/nullptr, 1505 /*isWeakRef=*/false); 1506 1507 for (const Arg *arg : args.filtered(OPT_why_live)) 1508 config->whyLive.insert(arg->getValue()); 1509 if (!config->whyLive.empty() && !config->deadStrip) 1510 warn("-why_live has no effect without -dead_strip, ignoring"); 1511 1512 config->saveTemps = args.hasArg(OPT_save_temps); 1513 1514 config->adhocCodesign = args.hasFlag( 1515 OPT_adhoc_codesign, OPT_no_adhoc_codesign, 1516 (config->arch() == AK_arm64 || config->arch() == AK_arm64e) && 1517 config->platform() == PLATFORM_MACOS); 1518 1519 if (args.hasArg(OPT_v)) { 1520 message(getLLDVersion(), lld::errs()); 1521 message(StringRef("Library search paths:") + 1522 (config->librarySearchPaths.empty() 1523 ? "" 1524 : "\n\t" + join(config->librarySearchPaths, "\n\t")), 1525 lld::errs()); 1526 message(StringRef("Framework search paths:") + 1527 (config->frameworkSearchPaths.empty() 1528 ? "" 1529 : "\n\t" + join(config->frameworkSearchPaths, "\n\t")), 1530 lld::errs()); 1531 } 1532 1533 config->progName = argsArr[0]; 1534 1535 config->timeTraceEnabled = args.hasArg(OPT_time_trace_eq); 1536 config->timeTraceGranularity = 1537 args::getInteger(args, OPT_time_trace_granularity_eq, 500); 1538 1539 // Initialize time trace profiler. 1540 if (config->timeTraceEnabled) 1541 timeTraceProfilerInitialize(config->timeTraceGranularity, config->progName); 1542 1543 { 1544 TimeTraceScope timeScope("ExecuteLinker"); 1545 1546 initLLVM(); // must be run before any call to addFile() 1547 createFiles(args); 1548 1549 config->isPic = config->outputType == MH_DYLIB || 1550 config->outputType == MH_BUNDLE || 1551 (config->outputType == MH_EXECUTE && 1552 args.hasFlag(OPT_pie, OPT_no_pie, true)); 1553 1554 // Now that all dylibs have been loaded, search for those that should be 1555 // re-exported. 1556 { 1557 auto reexportHandler = [](const Arg *arg, 1558 const std::vector<StringRef> &extensions) { 1559 config->hasReexports = true; 1560 StringRef searchName = arg->getValue(); 1561 if (!markReexport(searchName, extensions)) 1562 error(arg->getSpelling() + " " + searchName + 1563 " does not match a supplied dylib"); 1564 }; 1565 std::vector<StringRef> extensions = {".tbd"}; 1566 for (const Arg *arg : args.filtered(OPT_sub_umbrella)) 1567 reexportHandler(arg, extensions); 1568 1569 extensions.push_back(".dylib"); 1570 for (const Arg *arg : args.filtered(OPT_sub_library)) 1571 reexportHandler(arg, extensions); 1572 } 1573 1574 cl::ResetAllOptionOccurrences(); 1575 1576 // Parse LTO options. 1577 if (const Arg *arg = args.getLastArg(OPT_mcpu)) 1578 parseClangOption(saver().save("-mcpu=" + StringRef(arg->getValue())), 1579 arg->getSpelling()); 1580 1581 for (const Arg *arg : args.filtered(OPT_mllvm)) 1582 parseClangOption(arg->getValue(), arg->getSpelling()); 1583 1584 compileBitcodeFiles(); 1585 replaceCommonSymbols(); 1586 1587 StringRef orderFile = args.getLastArgValue(OPT_order_file); 1588 if (!orderFile.empty()) 1589 priorityBuilder.parseOrderFile(orderFile); 1590 1591 referenceStubBinder(); 1592 1593 // FIXME: should terminate the link early based on errors encountered so 1594 // far? 1595 1596 createSyntheticSections(); 1597 createSyntheticSymbols(); 1598 1599 for (const auto &pair : config->aliasedSymbols) { 1600 if (const auto &sym = symtab->find(pair.first)) { 1601 if (const auto &defined = dyn_cast<Defined>(sym)) { 1602 symtab->aliasDefined(defined, pair.second); 1603 continue; 1604 } 1605 } 1606 1607 warn("undefined base symbol '" + pair.first + "' for alias '" + 1608 pair.second + "'\n"); 1609 } 1610 1611 if (config->hasExplicitExports) { 1612 parallelForEach(symtab->getSymbols(), [](Symbol *sym) { 1613 if (auto *defined = dyn_cast<Defined>(sym)) { 1614 StringRef symbolName = defined->getName(); 1615 if (config->exportedSymbols.match(symbolName)) { 1616 if (defined->privateExtern) { 1617 if (defined->weakDefCanBeHidden) { 1618 // weak_def_can_be_hidden symbols behave similarly to 1619 // private_extern symbols in most cases, except for when 1620 // it is explicitly exported. 1621 // The former can be exported but the latter cannot. 1622 defined->privateExtern = false; 1623 } else { 1624 warn("cannot export hidden symbol " + toString(*defined) + 1625 "\n>>> defined in " + toString(defined->getFile())); 1626 } 1627 } 1628 } else { 1629 defined->privateExtern = true; 1630 } 1631 } 1632 }); 1633 } else if (!config->unexportedSymbols.empty()) { 1634 parallelForEach(symtab->getSymbols(), [](Symbol *sym) { 1635 if (auto *defined = dyn_cast<Defined>(sym)) 1636 if (config->unexportedSymbols.match(defined->getName())) 1637 defined->privateExtern = true; 1638 }); 1639 } 1640 1641 for (const Arg *arg : args.filtered(OPT_sectcreate)) { 1642 StringRef segName = arg->getValue(0); 1643 StringRef sectName = arg->getValue(1); 1644 StringRef fileName = arg->getValue(2); 1645 Optional<MemoryBufferRef> buffer = readFile(fileName); 1646 if (buffer) 1647 inputFiles.insert(make<OpaqueFile>(*buffer, segName, sectName)); 1648 } 1649 1650 for (const Arg *arg : args.filtered(OPT_add_empty_section)) { 1651 StringRef segName = arg->getValue(0); 1652 StringRef sectName = arg->getValue(1); 1653 inputFiles.insert(make<OpaqueFile>(MemoryBufferRef(), segName, sectName)); 1654 } 1655 1656 gatherInputSections(); 1657 if (config->callGraphProfileSort) 1658 priorityBuilder.extractCallGraphProfile(); 1659 1660 if (config->deadStrip) 1661 markLive(); 1662 1663 // ICF assumes that all literals have been folded already, so we must run 1664 // foldIdenticalLiterals before foldIdenticalSections. 1665 foldIdenticalLiterals(); 1666 if (config->icfLevel != ICFLevel::none) { 1667 if (config->icfLevel == ICFLevel::safe) 1668 markAddrSigSymbols(); 1669 foldIdenticalSections(/*onlyCfStrings=*/false); 1670 } else if (config->dedupLiterals) { 1671 foldIdenticalSections(/*onlyCfStrings=*/true); 1672 } 1673 1674 // Write to an output file. 1675 if (target->wordSize == 8) 1676 writeResult<LP64>(); 1677 else 1678 writeResult<ILP32>(); 1679 1680 depTracker->write(getLLDVersion(), inputFiles, config->outputFile); 1681 } 1682 1683 if (config->timeTraceEnabled) { 1684 checkError(timeTraceProfilerWrite( 1685 args.getLastArgValue(OPT_time_trace_eq).str(), config->outputFile)); 1686 1687 timeTraceProfilerCleanup(); 1688 } 1689 return errorCount() == 0; 1690 } 1691