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 //
loadFile(const char * argv0,std::unique_ptr<MemoryBuffer> Buffer,LLVMContext & Context,bool MaterializeMetadata=true)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
loadArFile(const char * Argv0,std::unique_ptr<MemoryBuffer> Buffer,LLVMContext & Context)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) {
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.
ModuleLazyLoaderCache(std::function<std::unique_ptr<Module> (const char * argv0,const std::string & FileName)> createLazyModule)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
takeModule(const std::string & FileName)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.
operator ()(const char * argv0,const std::string & Identifier)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 {
handleDiagnostics__anon40c295050211::LLVMLinkDiagnosticHandler284 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.
importFunctions(const char * argv0,Module & DestModule)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::ImportIDTable ImportIDs;
317 FunctionImporter::ImportMapTy ImportList(ImportIDs);
318
319 auto ModuleLoader = [&DestModule](const char *argv0,
320 const std::string &Identifier) {
321 std::unique_ptr<MemoryBuffer> Buffer = ExitOnErr(errorOrToExpected(
322 MemoryBuffer::getFileOrSTDIN(Identifier, /*IsText=*/true)));
323 return loadFile(argv0, std::move(Buffer), DestModule.getContext(), false);
324 };
325
326 ModuleLazyLoaderCache ModuleLoaderCache(ModuleLoader);
327 // Owns the filename strings used to key into the ImportList. Normally this is
328 // constructed from the index and the strings are owned by the index, however,
329 // since we are synthesizing this data structure from options we need a cache
330 // to own those strings.
331 StringSet<> FileNameStringCache;
332 for (const auto &Import : Imports) {
333 // Identify the requested function and its bitcode source file.
334 size_t Idx = Import.find(':');
335 if (Idx == std::string::npos) {
336 errs() << "Import parameter bad format: " << Import << "\n";
337 return false;
338 }
339 std::string FunctionName = Import.substr(0, Idx);
340 std::string FileName = Import.substr(Idx + 1, std::string::npos);
341
342 // Load the specified source module.
343 auto &SrcModule = ModuleLoaderCache(argv0, FileName);
344
345 if (!NoVerify && verifyModule(SrcModule, &errs())) {
346 errs() << argv0 << ": " << FileName;
347 WithColor::error() << "input module is broken!\n";
348 return false;
349 }
350
351 Function *F = SrcModule.getFunction(FunctionName);
352 if (!F) {
353 errs() << "Ignoring import request for non-existent function "
354 << FunctionName << " from " << FileName << "\n";
355 continue;
356 }
357 // We cannot import weak_any functions without possibly affecting the
358 // order they are seen and selected by the linker, changing program
359 // semantics.
360 if (F->hasWeakAnyLinkage()) {
361 errs() << "Ignoring import request for weak-any function " << FunctionName
362 << " from " << FileName << "\n";
363 continue;
364 }
365
366 if (Verbose)
367 errs() << "Importing " << FunctionName << " from " << FileName << "\n";
368
369 // `-import` specifies the `<filename,function-name>` pairs to import as
370 // definition, so make the import type definition directly.
371 // FIXME: A follow-up patch should add test coverage for import declaration
372 // in `llvm-link` CLI (e.g., by introducing a new command line option).
373 ImportList.addDefinition(
374 FileNameStringCache.insert(FileName).first->getKey(), F->getGUID());
375 }
376 auto CachedModuleLoader = [&](StringRef Identifier) {
377 return ModuleLoaderCache.takeModule(std::string(Identifier));
378 };
379 FunctionImporter Importer(*Index, CachedModuleLoader,
380 /*ClearDSOLocalOnDeclarations=*/false);
381 ExitOnErr(Importer.importFunctions(DestModule, ImportList));
382
383 return true;
384 }
385
linkFiles(const char * argv0,LLVMContext & Context,Linker & L,const cl::list<std::string> & Files,unsigned Flags)386 static bool linkFiles(const char *argv0, LLVMContext &Context, Linker &L,
387 const cl::list<std::string> &Files, unsigned Flags) {
388 // Filter out flags that don't apply to the first file we load.
389 unsigned ApplicableFlags = Flags & Linker::Flags::OverrideFromSrc;
390 // Similar to some flags, internalization doesn't apply to the first file.
391 bool InternalizeLinkedSymbols = false;
392 for (const auto &File : Files) {
393 auto BufferOrErr = MemoryBuffer::getFileOrSTDIN(File, /*IsText=*/true);
394
395 // When we encounter a missing file, make sure we expose its name.
396 if (auto EC = BufferOrErr.getError())
397 if (EC == std::errc::no_such_file_or_directory)
398 ExitOnErr(createStringError(EC, "No such file or directory: '%s'",
399 File.c_str()));
400
401 std::unique_ptr<MemoryBuffer> Buffer =
402 ExitOnErr(errorOrToExpected(std::move(BufferOrErr)));
403
404 std::unique_ptr<Module> M =
405 identify_magic(Buffer->getBuffer()) == file_magic::archive
406 ? loadArFile(argv0, std::move(Buffer), Context)
407 : loadFile(argv0, std::move(Buffer), Context);
408 if (!M) {
409 errs() << argv0 << ": ";
410 WithColor::error() << " loading file '" << File << "'\n";
411 return false;
412 }
413
414 // Note that when ODR merging types cannot verify input files in here When
415 // doing that debug metadata in the src module might already be pointing to
416 // the destination.
417 if (DisableDITypeMap && !NoVerify && verifyModule(*M, &errs())) {
418 errs() << argv0 << ": " << File << ": ";
419 WithColor::error() << "input module is broken!\n";
420 return false;
421 }
422
423 // If a module summary index is supplied, load it so linkInModule can treat
424 // local functions/variables as exported and promote if necessary.
425 if (!SummaryIndex.empty()) {
426 std::unique_ptr<ModuleSummaryIndex> Index =
427 ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex));
428
429 // Conservatively mark all internal values as promoted, since this tool
430 // does not do the ThinLink that would normally determine what values to
431 // promote.
432 for (auto &I : *Index) {
433 for (auto &S : I.second.SummaryList) {
434 if (GlobalValue::isLocalLinkage(S->linkage()))
435 S->setLinkage(GlobalValue::ExternalLinkage);
436 }
437 }
438
439 // Promotion
440 renameModuleForThinLTO(*M, *Index,
441 /*ClearDSOLocalOnDeclarations=*/false);
442 }
443
444 if (Verbose)
445 errs() << "Linking in '" << File << "'\n";
446
447 bool Err = false;
448 if (InternalizeLinkedSymbols) {
449 Err = L.linkInModule(
450 std::move(M), ApplicableFlags, [](Module &M, const StringSet<> &GVS) {
451 internalizeModule(M, [&GVS](const GlobalValue &GV) {
452 return !GV.hasName() || (GVS.count(GV.getName()) == 0);
453 });
454 });
455 } else {
456 Err = L.linkInModule(std::move(M), ApplicableFlags);
457 }
458
459 if (Err)
460 return false;
461
462 // Internalization applies to linking of subsequent files.
463 InternalizeLinkedSymbols = Internalize;
464
465 // All linker flags apply to linking of subsequent files.
466 ApplicableFlags = Flags;
467 }
468
469 return true;
470 }
471
main(int argc,char ** argv)472 int main(int argc, char **argv) {
473 InitLLVM X(argc, argv);
474 ExitOnErr.setBanner(std::string(argv[0]) + ": ");
475
476 cl::HideUnrelatedOptions({&LinkCategory, &getColorCategory()});
477 cl::ParseCommandLineOptions(argc, argv, "llvm linker\n");
478
479 LLVMContext Context;
480 Context.setDiagnosticHandler(std::make_unique<LLVMLinkDiagnosticHandler>(),
481 true);
482
483 if (!DisableDITypeMap)
484 Context.enableDebugTypeODRUniquing();
485
486 auto Composite = std::make_unique<Module>("llvm-link", Context);
487 Linker L(*Composite);
488
489 unsigned Flags = Linker::Flags::None;
490 if (OnlyNeeded)
491 Flags |= Linker::Flags::LinkOnlyNeeded;
492
493 // First add all the regular input files
494 if (!linkFiles(argv[0], Context, L, InputFilenames, Flags))
495 return 1;
496
497 // Next the -override ones.
498 if (!linkFiles(argv[0], Context, L, OverridingInputs,
499 Flags | Linker::Flags::OverrideFromSrc))
500 return 1;
501
502 // Import any functions requested via -import
503 if (!importFunctions(argv[0], *Composite))
504 return 1;
505
506 if (DumpAsm)
507 errs() << "Here's the assembly:\n" << *Composite;
508
509 std::error_code EC;
510 ToolOutputFile Out(OutputFilename, EC,
511 OutputAssembly ? sys::fs::OF_TextWithCRLF
512 : sys::fs::OF_None);
513 if (EC) {
514 WithColor::error() << EC.message() << '\n';
515 return 1;
516 }
517
518 if (!NoVerify && verifyModule(*Composite, &errs())) {
519 errs() << argv[0] << ": ";
520 WithColor::error() << "linked module is broken!\n";
521 return 1;
522 }
523
524 if (Verbose)
525 errs() << "Writing bitcode...\n";
526 Composite->removeDebugIntrinsicDeclarations();
527 if (OutputAssembly) {
528 Composite->print(Out.os(), nullptr, PreserveAssemblyUseListOrder);
529 } else if (Force || !CheckBitcodeOutputToConsole(Out.os())) {
530 WriteBitcodeToFile(*Composite, Out.os(), PreserveBitcodeUseListOrder);
531 }
532
533 // Declare success.
534 Out.keep();
535
536 return 0;
537 }
538