1 //===- llvm-link.cpp - Low-level LLVM linker ------------------------------===// 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 // This utility may be invoked in the following manner: 10 // llvm-link a.bc b.bc c.bc -o x.bc 11 // 12 //===----------------------------------------------------------------------===// 13 14 #include "llvm/ADT/STLExtras.h" 15 #include "llvm/BinaryFormat/Magic.h" 16 #include "llvm/Bitcode/BitcodeReader.h" 17 #include "llvm/Bitcode/BitcodeWriter.h" 18 #include "llvm/IR/AutoUpgrade.h" 19 #include "llvm/IR/DiagnosticInfo.h" 20 #include "llvm/IR/DiagnosticPrinter.h" 21 #include "llvm/IR/LLVMContext.h" 22 #include "llvm/IR/Module.h" 23 #include "llvm/IR/ModuleSummaryIndex.h" 24 #include "llvm/IR/Verifier.h" 25 #include "llvm/IRReader/IRReader.h" 26 #include "llvm/Linker/Linker.h" 27 #include "llvm/Object/Archive.h" 28 #include "llvm/Support/CommandLine.h" 29 #include "llvm/Support/FileSystem.h" 30 #include "llvm/Support/InitLLVM.h" 31 #include "llvm/Support/Path.h" 32 #include "llvm/Support/SourceMgr.h" 33 #include "llvm/Support/SystemUtils.h" 34 #include "llvm/Support/ToolOutputFile.h" 35 #include "llvm/Support/WithColor.h" 36 #include "llvm/Transforms/IPO/FunctionImport.h" 37 #include "llvm/Transforms/IPO/Internalize.h" 38 #include "llvm/Transforms/Utils/FunctionImportUtils.h" 39 40 #include <memory> 41 #include <utility> 42 using namespace llvm; 43 44 static cl::OptionCategory LinkCategory("Link Options"); 45 46 static cl::list<std::string> InputFilenames(cl::Positional, cl::OneOrMore, 47 cl::desc("<input bitcode files>"), 48 cl::cat(LinkCategory)); 49 50 static cl::list<std::string> OverridingInputs( 51 "override", cl::value_desc("filename"), 52 cl::desc( 53 "input bitcode file which can override previously defined symbol(s)"), 54 cl::cat(LinkCategory)); 55 56 // Option to simulate function importing for testing. This enables using 57 // llvm-link to simulate ThinLTO backend processes. 58 static cl::list<std::string> Imports( 59 "import", cl::value_desc("function:filename"), 60 cl::desc("Pair of function name and filename, where function should be " 61 "imported from bitcode in filename"), 62 cl::cat(LinkCategory)); 63 64 // Option to support testing of function importing. The module summary 65 // must be specified in the case were we request imports via the -import 66 // option, as well as when compiling any module with functions that may be 67 // exported (imported by a different llvm-link -import invocation), to ensure 68 // consistent promotion and renaming of locals. 69 static cl::opt<std::string> 70 SummaryIndex("summary-index", cl::desc("Module summary index filename"), 71 cl::init(""), cl::value_desc("filename"), 72 cl::cat(LinkCategory)); 73 74 static cl::opt<std::string> 75 OutputFilename("o", cl::desc("Override output filename"), cl::init("-"), 76 cl::value_desc("filename"), cl::cat(LinkCategory)); 77 78 static cl::opt<bool> Internalize("internalize", 79 cl::desc("Internalize linked symbols"), 80 cl::cat(LinkCategory)); 81 82 static cl::opt<bool> 83 DisableDITypeMap("disable-debug-info-type-map", 84 cl::desc("Don't use a uniquing type map for debug info"), 85 cl::cat(LinkCategory)); 86 87 static cl::opt<bool> OnlyNeeded("only-needed", 88 cl::desc("Link only needed symbols"), 89 cl::cat(LinkCategory)); 90 91 static cl::opt<bool> Force("f", cl::desc("Enable binary output on terminals"), 92 cl::cat(LinkCategory)); 93 94 static cl::opt<bool> DisableLazyLoad("disable-lazy-loading", 95 cl::desc("Disable lazy module loading"), 96 cl::cat(LinkCategory)); 97 98 static cl::opt<bool> OutputAssembly("S", 99 cl::desc("Write output as LLVM assembly"), 100 cl::Hidden, cl::cat(LinkCategory)); 101 102 static cl::opt<bool> Verbose("v", 103 cl::desc("Print information about actions taken"), 104 cl::cat(LinkCategory)); 105 106 static cl::opt<bool> DumpAsm("d", cl::desc("Print assembly as linked"), 107 cl::Hidden, cl::cat(LinkCategory)); 108 109 static cl::opt<bool> SuppressWarnings("suppress-warnings", 110 cl::desc("Suppress all linking warnings"), 111 cl::init(false), cl::cat(LinkCategory)); 112 113 static cl::opt<bool> PreserveBitcodeUseListOrder( 114 "preserve-bc-uselistorder", 115 cl::desc("Preserve use-list order when writing LLVM bitcode."), 116 cl::init(true), cl::Hidden, cl::cat(LinkCategory)); 117 118 static cl::opt<bool> PreserveAssemblyUseListOrder( 119 "preserve-ll-uselistorder", 120 cl::desc("Preserve use-list order when writing LLVM assembly."), 121 cl::init(false), cl::Hidden, cl::cat(LinkCategory)); 122 123 static cl::opt<bool> NoVerify("disable-verify", 124 cl::desc("Do not run the verifier"), cl::Hidden, 125 cl::cat(LinkCategory)); 126 127 static cl::opt<bool> IgnoreNonBitcode( 128 "ignore-non-bitcode", 129 cl::desc("Do not report an error for non-bitcode files in archives"), 130 cl::Hidden); 131 132 static ExitOnError ExitOnErr; 133 134 // Read the specified bitcode file in and return it. This routine searches the 135 // link path for the specified file to try to find it... 136 // 137 static std::unique_ptr<Module> loadFile(const char *argv0, 138 std::unique_ptr<MemoryBuffer> Buffer, 139 LLVMContext &Context, 140 bool MaterializeMetadata = true) { 141 SMDiagnostic Err; 142 if (Verbose) 143 errs() << "Loading '" << Buffer->getBufferIdentifier() << "'\n"; 144 std::unique_ptr<Module> Result; 145 if (DisableLazyLoad) 146 Result = parseIR(*Buffer, Err, Context); 147 else 148 Result = 149 getLazyIRModule(std::move(Buffer), Err, Context, !MaterializeMetadata); 150 151 if (!Result) { 152 Err.print(argv0, errs()); 153 return nullptr; 154 } 155 156 if (MaterializeMetadata) { 157 ExitOnErr(Result->materializeMetadata()); 158 UpgradeDebugInfo(*Result); 159 } 160 161 return Result; 162 } 163 164 static std::unique_ptr<Module> loadArFile(const char *Argv0, 165 std::unique_ptr<MemoryBuffer> Buffer, 166 LLVMContext &Context) { 167 std::unique_ptr<Module> Result(new Module("ArchiveModule", Context)); 168 StringRef ArchiveName = Buffer->getBufferIdentifier(); 169 if (Verbose) 170 errs() << "Reading library archive file '" << ArchiveName 171 << "' to memory\n"; 172 Expected<std::unique_ptr<object::Archive>> ArchiveOrError = 173 object::Archive::create(Buffer->getMemBufferRef()); 174 if (!ArchiveOrError) 175 ExitOnErr(ArchiveOrError.takeError()); 176 177 std::unique_ptr<object::Archive> Archive = std::move(ArchiveOrError.get()); 178 179 Linker L(*Result); 180 Error Err = Error::success(); 181 for (const object::Archive::Child &C : Archive->children(Err)) { 182 Expected<StringRef> Ename = C.getName(); 183 if (Error E = Ename.takeError()) { 184 errs() << Argv0 << ": "; 185 WithColor::error() << " failed to read name of archive member" 186 << ArchiveName << "'\n"; 187 return nullptr; 188 } 189 std::string ChildName = Ename.get().str(); 190 if (Verbose) 191 errs() << "Parsing member '" << ChildName 192 << "' of archive library to module.\n"; 193 SMDiagnostic ParseErr; 194 Expected<MemoryBufferRef> MemBuf = C.getMemoryBufferRef(); 195 if (Error E = MemBuf.takeError()) { 196 errs() << Argv0 << ": "; 197 WithColor::error() << " loading memory for member '" << ChildName 198 << "' of archive library failed'" << ArchiveName 199 << "'\n"; 200 return nullptr; 201 }; 202 203 if (!isBitcode(reinterpret_cast<const unsigned char *>( 204 MemBuf.get().getBufferStart()), 205 reinterpret_cast<const unsigned char *>( 206 MemBuf.get().getBufferEnd()))) { 207 if (IgnoreNonBitcode) 208 continue; 209 errs() << Argv0 << ": "; 210 WithColor::error() << " member of archive is not a bitcode file: '" 211 << ChildName << "'\n"; 212 return nullptr; 213 } 214 215 std::unique_ptr<Module> M; 216 if (DisableLazyLoad) 217 M = parseIR(MemBuf.get(), ParseErr, Context); 218 else 219 M = getLazyIRModule(MemoryBuffer::getMemBuffer(MemBuf.get(), false), 220 ParseErr, Context); 221 222 if (!M.get()) { 223 errs() << Argv0 << ": "; 224 WithColor::error() << " parsing member '" << ChildName 225 << "' of archive library failed'" << ArchiveName 226 << "'\n"; 227 return nullptr; 228 } 229 if (Verbose) 230 errs() << "Linking member '" << ChildName << "' of archive library.\n"; 231 if (L.linkInModule(std::move(M))) 232 return nullptr; 233 } // end for each child 234 ExitOnErr(std::move(Err)); 235 return Result; 236 } 237 238 namespace { 239 240 /// Helper to load on demand a Module from file and cache it for subsequent 241 /// queries during function importing. 242 class ModuleLazyLoaderCache { 243 /// Cache of lazily loaded module for import. 244 StringMap<std::unique_ptr<Module>> ModuleMap; 245 246 /// Retrieve a Module from the cache or lazily load it on demand. 247 std::function<std::unique_ptr<Module>(const char *argv0, 248 const std::string &FileName)> 249 createLazyModule; 250 251 public: 252 /// Create the loader, Module will be initialized in \p Context. 253 ModuleLazyLoaderCache(std::function<std::unique_ptr<Module>( 254 const char *argv0, const std::string &FileName)> 255 createLazyModule) 256 : createLazyModule(std::move(createLazyModule)) {} 257 258 /// Retrieve a Module from the cache or lazily load it on demand. 259 Module &operator()(const char *argv0, const std::string &FileName); 260 261 std::unique_ptr<Module> takeModule(const std::string &FileName) { 262 auto I = ModuleMap.find(FileName); 263 assert(I != ModuleMap.end()); 264 std::unique_ptr<Module> Ret = std::move(I->second); 265 ModuleMap.erase(I); 266 return Ret; 267 } 268 }; 269 270 // Get a Module for \p FileName from the cache, or load it lazily. 271 Module &ModuleLazyLoaderCache::operator()(const char *argv0, 272 const std::string &Identifier) { 273 auto &Module = ModuleMap[Identifier]; 274 if (!Module) { 275 Module = createLazyModule(argv0, Identifier); 276 assert(Module && "Failed to create lazy module!"); 277 } 278 return *Module; 279 } 280 } // anonymous namespace 281 282 namespace { 283 struct LLVMLinkDiagnosticHandler : public DiagnosticHandler { 284 bool handleDiagnostics(const DiagnosticInfo &DI) override { 285 unsigned Severity = DI.getSeverity(); 286 switch (Severity) { 287 case DS_Error: 288 WithColor::error(); 289 break; 290 case DS_Warning: 291 if (SuppressWarnings) 292 return true; 293 WithColor::warning(); 294 break; 295 case DS_Remark: 296 case DS_Note: 297 llvm_unreachable("Only expecting warnings and errors"); 298 } 299 300 DiagnosticPrinterRawOStream DP(errs()); 301 DI.print(DP); 302 errs() << '\n'; 303 return true; 304 } 305 }; 306 } // namespace 307 308 /// Import any functions requested via the -import option. 309 static bool importFunctions(const char *argv0, Module &DestModule) { 310 if (SummaryIndex.empty()) 311 return true; 312 std::unique_ptr<ModuleSummaryIndex> Index = 313 ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex)); 314 315 // Map of Module -> List of globals to import from the Module 316 FunctionImporter::ImportMapTy ImportList; 317 318 auto ModuleLoader = [&DestModule](const char *argv0, 319 const std::string &Identifier) { 320 std::unique_ptr<MemoryBuffer> Buffer = 321 ExitOnErr(errorOrToExpected(MemoryBuffer::getFileOrSTDIN(Identifier))); 322 return loadFile(argv0, std::move(Buffer), DestModule.getContext(), false); 323 }; 324 325 ModuleLazyLoaderCache ModuleLoaderCache(ModuleLoader); 326 // Owns the filename strings used to key into the ImportList. Normally this is 327 // constructed from the index and the strings are owned by the index, however, 328 // since we are synthesizing this data structure from options we need a cache 329 // to own those strings. 330 StringSet<> FileNameStringCache; 331 for (const auto &Import : Imports) { 332 // Identify the requested function and its bitcode source file. 333 size_t Idx = Import.find(':'); 334 if (Idx == std::string::npos) { 335 errs() << "Import parameter bad format: " << Import << "\n"; 336 return false; 337 } 338 std::string FunctionName = Import.substr(0, Idx); 339 std::string FileName = Import.substr(Idx + 1, std::string::npos); 340 341 // Load the specified source module. 342 auto &SrcModule = ModuleLoaderCache(argv0, FileName); 343 344 if (!NoVerify && verifyModule(SrcModule, &errs())) { 345 errs() << argv0 << ": " << FileName; 346 WithColor::error() << "input module is broken!\n"; 347 return false; 348 } 349 350 Function *F = SrcModule.getFunction(FunctionName); 351 if (!F) { 352 errs() << "Ignoring import request for non-existent function " 353 << FunctionName << " from " << FileName << "\n"; 354 continue; 355 } 356 // We cannot import weak_any functions without possibly affecting the 357 // order they are seen and selected by the linker, changing program 358 // semantics. 359 if (F->hasWeakAnyLinkage()) { 360 errs() << "Ignoring import request for weak-any function " << FunctionName 361 << " from " << FileName << "\n"; 362 continue; 363 } 364 365 if (Verbose) 366 errs() << "Importing " << FunctionName << " from " << FileName << "\n"; 367 368 auto &Entry = 369 ImportList[FileNameStringCache.insert(FileName).first->getKey()]; 370 Entry.insert(F->getGUID()); 371 } 372 auto CachedModuleLoader = [&](StringRef Identifier) { 373 return ModuleLoaderCache.takeModule(std::string(Identifier)); 374 }; 375 FunctionImporter Importer(*Index, CachedModuleLoader, 376 /*ClearDSOLocalOnDeclarations=*/false); 377 ExitOnErr(Importer.importFunctions(DestModule, ImportList)); 378 379 return true; 380 } 381 382 static bool linkFiles(const char *argv0, LLVMContext &Context, Linker &L, 383 const cl::list<std::string> &Files, unsigned Flags) { 384 // Filter out flags that don't apply to the first file we load. 385 unsigned ApplicableFlags = Flags & Linker::Flags::OverrideFromSrc; 386 // Similar to some flags, internalization doesn't apply to the first file. 387 bool InternalizeLinkedSymbols = false; 388 for (const auto &File : Files) { 389 std::unique_ptr<MemoryBuffer> Buffer = 390 ExitOnErr(errorOrToExpected(MemoryBuffer::getFileOrSTDIN(File))); 391 392 std::unique_ptr<Module> M = 393 identify_magic(Buffer->getBuffer()) == file_magic::archive 394 ? loadArFile(argv0, std::move(Buffer), Context) 395 : loadFile(argv0, std::move(Buffer), Context); 396 if (!M.get()) { 397 errs() << argv0 << ": "; 398 WithColor::error() << " loading file '" << File << "'\n"; 399 return false; 400 } 401 402 // Note that when ODR merging types cannot verify input files in here When 403 // doing that debug metadata in the src module might already be pointing to 404 // the destination. 405 if (DisableDITypeMap && !NoVerify && verifyModule(*M, &errs())) { 406 errs() << argv0 << ": " << File << ": "; 407 WithColor::error() << "input module is broken!\n"; 408 return false; 409 } 410 411 // If a module summary index is supplied, load it so linkInModule can treat 412 // local functions/variables as exported and promote if necessary. 413 if (!SummaryIndex.empty()) { 414 std::unique_ptr<ModuleSummaryIndex> Index = 415 ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex)); 416 417 // Conservatively mark all internal values as promoted, since this tool 418 // does not do the ThinLink that would normally determine what values to 419 // promote. 420 for (auto &I : *Index) { 421 for (auto &S : I.second.SummaryList) { 422 if (GlobalValue::isLocalLinkage(S->linkage())) 423 S->setLinkage(GlobalValue::ExternalLinkage); 424 } 425 } 426 427 // Promotion 428 if (renameModuleForThinLTO(*M, *Index, 429 /*ClearDSOLocalOnDeclarations=*/false)) 430 return true; 431 } 432 433 if (Verbose) 434 errs() << "Linking in '" << File << "'\n"; 435 436 bool Err = false; 437 if (InternalizeLinkedSymbols) { 438 Err = L.linkInModule( 439 std::move(M), ApplicableFlags, [](Module &M, const StringSet<> &GVS) { 440 internalizeModule(M, [&GVS](const GlobalValue &GV) { 441 return !GV.hasName() || (GVS.count(GV.getName()) == 0); 442 }); 443 }); 444 } else { 445 Err = L.linkInModule(std::move(M), ApplicableFlags); 446 } 447 448 if (Err) 449 return false; 450 451 // Internalization applies to linking of subsequent files. 452 InternalizeLinkedSymbols = Internalize; 453 454 // All linker flags apply to linking of subsequent files. 455 ApplicableFlags = Flags; 456 } 457 458 return true; 459 } 460 461 int main(int argc, char **argv) { 462 InitLLVM X(argc, argv); 463 ExitOnErr.setBanner(std::string(argv[0]) + ": "); 464 465 cl::HideUnrelatedOptions({&LinkCategory, &getColorCategory()}); 466 cl::ParseCommandLineOptions(argc, argv, "llvm linker\n"); 467 468 LLVMContext Context; 469 Context.setDiagnosticHandler(std::make_unique<LLVMLinkDiagnosticHandler>(), 470 true); 471 472 if (!DisableDITypeMap) 473 Context.enableDebugTypeODRUniquing(); 474 475 auto Composite = std::make_unique<Module>("llvm-link", Context); 476 Linker L(*Composite); 477 478 unsigned Flags = Linker::Flags::None; 479 if (OnlyNeeded) 480 Flags |= Linker::Flags::LinkOnlyNeeded; 481 482 // First add all the regular input files 483 if (!linkFiles(argv[0], Context, L, InputFilenames, Flags)) 484 return 1; 485 486 // Next the -override ones. 487 if (!linkFiles(argv[0], Context, L, OverridingInputs, 488 Flags | Linker::Flags::OverrideFromSrc)) 489 return 1; 490 491 // Import any functions requested via -import 492 if (!importFunctions(argv[0], *Composite)) 493 return 1; 494 495 if (DumpAsm) 496 errs() << "Here's the assembly:\n" << *Composite; 497 498 std::error_code EC; 499 ToolOutputFile Out(OutputFilename, EC, 500 OutputAssembly ? sys::fs::OF_TextWithCRLF 501 : sys::fs::OF_None); 502 if (EC) { 503 WithColor::error() << EC.message() << '\n'; 504 return 1; 505 } 506 507 if (!NoVerify && verifyModule(*Composite, &errs())) { 508 errs() << argv[0] << ": "; 509 WithColor::error() << "linked module is broken!\n"; 510 return 1; 511 } 512 513 if (Verbose) 514 errs() << "Writing bitcode...\n"; 515 if (OutputAssembly) { 516 Composite->print(Out.os(), nullptr, PreserveAssemblyUseListOrder); 517 } else if (Force || !CheckBitcodeOutputToConsole(Out.os())) 518 WriteBitcodeToFile(*Composite, Out.os(), PreserveBitcodeUseListOrder); 519 520 // Declare success. 521 Out.keep(); 522 523 return 0; 524 } 525