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 cl::opt<bool> TryUseNewDbgInfoFormat(
133 "try-experimental-debuginfo-iterators",
134 cl::desc("Enable debuginfo iterator positions, if they're built in"),
135 cl::init(false));
136
137 extern cl::opt<bool> UseNewDbgInfoFormat;
138 extern cl::opt<cl::boolOrDefault> PreserveInputDbgFormat;
139 extern cl::opt<bool> WriteNewDbgInfoFormat;
140 extern bool WriteNewDbgInfoFormatToBitcode;
141
142 extern cl::opt<cl::boolOrDefault> LoadBitcodeIntoNewDbgInfoFormat;
143
144 static ExitOnError ExitOnErr;
145
146 // Read the specified bitcode file in and return it. This routine searches the
147 // link path for the specified file to try to find it...
148 //
loadFile(const char * argv0,std::unique_ptr<MemoryBuffer> Buffer,LLVMContext & Context,bool MaterializeMetadata=true)149 static std::unique_ptr<Module> loadFile(const char *argv0,
150 std::unique_ptr<MemoryBuffer> Buffer,
151 LLVMContext &Context,
152 bool MaterializeMetadata = true) {
153 SMDiagnostic Err;
154 if (Verbose)
155 errs() << "Loading '" << Buffer->getBufferIdentifier() << "'\n";
156 std::unique_ptr<Module> Result;
157 if (DisableLazyLoad)
158 Result = parseIR(*Buffer, Err, Context);
159 else
160 Result =
161 getLazyIRModule(std::move(Buffer), Err, Context, !MaterializeMetadata);
162
163 if (!Result) {
164 Err.print(argv0, errs());
165 return nullptr;
166 }
167
168 if (MaterializeMetadata) {
169 ExitOnErr(Result->materializeMetadata());
170 UpgradeDebugInfo(*Result);
171 }
172
173 return Result;
174 }
175
loadArFile(const char * Argv0,std::unique_ptr<MemoryBuffer> Buffer,LLVMContext & Context)176 static std::unique_ptr<Module> loadArFile(const char *Argv0,
177 std::unique_ptr<MemoryBuffer> Buffer,
178 LLVMContext &Context) {
179 std::unique_ptr<Module> Result(new Module("ArchiveModule", Context));
180 StringRef ArchiveName = Buffer->getBufferIdentifier();
181 if (Verbose)
182 errs() << "Reading library archive file '" << ArchiveName
183 << "' to memory\n";
184 Expected<std::unique_ptr<object::Archive>> ArchiveOrError =
185 object::Archive::create(Buffer->getMemBufferRef());
186 if (!ArchiveOrError)
187 ExitOnErr(ArchiveOrError.takeError());
188
189 std::unique_ptr<object::Archive> Archive = std::move(ArchiveOrError.get());
190
191 Linker L(*Result);
192 Error Err = Error::success();
193 for (const object::Archive::Child &C : Archive->children(Err)) {
194 Expected<StringRef> Ename = C.getName();
195 if (Error E = Ename.takeError()) {
196 errs() << Argv0 << ": ";
197 WithColor::error() << " failed to read name of archive member"
198 << ArchiveName << "'\n";
199 return nullptr;
200 }
201 std::string ChildName = Ename.get().str();
202 if (Verbose)
203 errs() << "Parsing member '" << ChildName
204 << "' of archive library to module.\n";
205 SMDiagnostic ParseErr;
206 Expected<MemoryBufferRef> MemBuf = C.getMemoryBufferRef();
207 if (Error E = MemBuf.takeError()) {
208 errs() << Argv0 << ": ";
209 WithColor::error() << " loading memory for member '" << ChildName
210 << "' of archive library failed'" << ArchiveName
211 << "'\n";
212 return nullptr;
213 };
214
215 if (!isBitcode(reinterpret_cast<const unsigned char *>(
216 MemBuf.get().getBufferStart()),
217 reinterpret_cast<const unsigned char *>(
218 MemBuf.get().getBufferEnd()))) {
219 if (IgnoreNonBitcode)
220 continue;
221 errs() << Argv0 << ": ";
222 WithColor::error() << " member of archive is not a bitcode file: '"
223 << ChildName << "'\n";
224 return nullptr;
225 }
226
227 std::unique_ptr<Module> M;
228 if (DisableLazyLoad)
229 M = parseIR(MemBuf.get(), ParseErr, Context);
230 else
231 M = getLazyIRModule(MemoryBuffer::getMemBuffer(MemBuf.get(), false),
232 ParseErr, Context);
233
234 if (!M) {
235 errs() << Argv0 << ": ";
236 WithColor::error() << " parsing member '" << ChildName
237 << "' of archive library failed'" << ArchiveName
238 << "'\n";
239 return nullptr;
240 }
241 if (Verbose)
242 errs() << "Linking member '" << ChildName << "' of archive library.\n";
243 if (L.linkInModule(std::move(M)))
244 return nullptr;
245 } // end for each child
246 ExitOnErr(std::move(Err));
247 return Result;
248 }
249
250 namespace {
251
252 /// Helper to load on demand a Module from file and cache it for subsequent
253 /// queries during function importing.
254 class ModuleLazyLoaderCache {
255 /// Cache of lazily loaded module for import.
256 StringMap<std::unique_ptr<Module>> ModuleMap;
257
258 /// Retrieve a Module from the cache or lazily load it on demand.
259 std::function<std::unique_ptr<Module>(const char *argv0,
260 const std::string &FileName)>
261 createLazyModule;
262
263 public:
264 /// 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)265 ModuleLazyLoaderCache(std::function<std::unique_ptr<Module>(
266 const char *argv0, const std::string &FileName)>
267 createLazyModule)
268 : createLazyModule(std::move(createLazyModule)) {}
269
270 /// Retrieve a Module from the cache or lazily load it on demand.
271 Module &operator()(const char *argv0, const std::string &FileName);
272
takeModule(const std::string & FileName)273 std::unique_ptr<Module> takeModule(const std::string &FileName) {
274 auto I = ModuleMap.find(FileName);
275 assert(I != ModuleMap.end());
276 std::unique_ptr<Module> Ret = std::move(I->second);
277 ModuleMap.erase(I);
278 return Ret;
279 }
280 };
281
282 // Get a Module for \p FileName from the cache, or load it lazily.
operator ()(const char * argv0,const std::string & Identifier)283 Module &ModuleLazyLoaderCache::operator()(const char *argv0,
284 const std::string &Identifier) {
285 auto &Module = ModuleMap[Identifier];
286 if (!Module) {
287 Module = createLazyModule(argv0, Identifier);
288 assert(Module && "Failed to create lazy module!");
289 }
290 return *Module;
291 }
292 } // anonymous namespace
293
294 namespace {
295 struct LLVMLinkDiagnosticHandler : public DiagnosticHandler {
handleDiagnostics__anon40c295050211::LLVMLinkDiagnosticHandler296 bool handleDiagnostics(const DiagnosticInfo &DI) override {
297 unsigned Severity = DI.getSeverity();
298 switch (Severity) {
299 case DS_Error:
300 WithColor::error();
301 break;
302 case DS_Warning:
303 if (SuppressWarnings)
304 return true;
305 WithColor::warning();
306 break;
307 case DS_Remark:
308 case DS_Note:
309 llvm_unreachable("Only expecting warnings and errors");
310 }
311
312 DiagnosticPrinterRawOStream DP(errs());
313 DI.print(DP);
314 errs() << '\n';
315 return true;
316 }
317 };
318 } // namespace
319
320 /// Import any functions requested via the -import option.
importFunctions(const char * argv0,Module & DestModule)321 static bool importFunctions(const char *argv0, Module &DestModule) {
322 if (SummaryIndex.empty())
323 return true;
324 std::unique_ptr<ModuleSummaryIndex> Index =
325 ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex));
326
327 // Map of Module -> List of globals to import from the Module
328 FunctionImporter::ImportMapTy ImportList;
329
330 auto ModuleLoader = [&DestModule](const char *argv0,
331 const std::string &Identifier) {
332 std::unique_ptr<MemoryBuffer> Buffer =
333 ExitOnErr(errorOrToExpected(MemoryBuffer::getFileOrSTDIN(Identifier)));
334 return loadFile(argv0, std::move(Buffer), DestModule.getContext(), false);
335 };
336
337 ModuleLazyLoaderCache ModuleLoaderCache(ModuleLoader);
338 // Owns the filename strings used to key into the ImportList. Normally this is
339 // constructed from the index and the strings are owned by the index, however,
340 // since we are synthesizing this data structure from options we need a cache
341 // to own those strings.
342 StringSet<> FileNameStringCache;
343 for (const auto &Import : Imports) {
344 // Identify the requested function and its bitcode source file.
345 size_t Idx = Import.find(':');
346 if (Idx == std::string::npos) {
347 errs() << "Import parameter bad format: " << Import << "\n";
348 return false;
349 }
350 std::string FunctionName = Import.substr(0, Idx);
351 std::string FileName = Import.substr(Idx + 1, std::string::npos);
352
353 // Load the specified source module.
354 auto &SrcModule = ModuleLoaderCache(argv0, FileName);
355
356 if (!NoVerify && verifyModule(SrcModule, &errs())) {
357 errs() << argv0 << ": " << FileName;
358 WithColor::error() << "input module is broken!\n";
359 return false;
360 }
361
362 Function *F = SrcModule.getFunction(FunctionName);
363 if (!F) {
364 errs() << "Ignoring import request for non-existent function "
365 << FunctionName << " from " << FileName << "\n";
366 continue;
367 }
368 // We cannot import weak_any functions without possibly affecting the
369 // order they are seen and selected by the linker, changing program
370 // semantics.
371 if (F->hasWeakAnyLinkage()) {
372 errs() << "Ignoring import request for weak-any function " << FunctionName
373 << " from " << FileName << "\n";
374 continue;
375 }
376
377 if (Verbose)
378 errs() << "Importing " << FunctionName << " from " << FileName << "\n";
379
380 // `-import` specifies the `<filename,function-name>` pairs to import as
381 // definition, so make the import type definition directly.
382 // FIXME: A follow-up patch should add test coverage for import declaration
383 // in `llvm-link` CLI (e.g., by introducing a new command line option).
384 auto &Entry =
385 ImportList[FileNameStringCache.insert(FileName).first->getKey()];
386 Entry[F->getGUID()] = GlobalValueSummary::Definition;
387 }
388 auto CachedModuleLoader = [&](StringRef Identifier) {
389 return ModuleLoaderCache.takeModule(std::string(Identifier));
390 };
391 FunctionImporter Importer(*Index, CachedModuleLoader,
392 /*ClearDSOLocalOnDeclarations=*/false);
393 ExitOnErr(Importer.importFunctions(DestModule, ImportList));
394
395 return true;
396 }
397
linkFiles(const char * argv0,LLVMContext & Context,Linker & L,const cl::list<std::string> & Files,unsigned Flags)398 static bool linkFiles(const char *argv0, LLVMContext &Context, Linker &L,
399 const cl::list<std::string> &Files, unsigned Flags) {
400 // Filter out flags that don't apply to the first file we load.
401 unsigned ApplicableFlags = Flags & Linker::Flags::OverrideFromSrc;
402 // Similar to some flags, internalization doesn't apply to the first file.
403 bool InternalizeLinkedSymbols = false;
404 for (const auto &File : Files) {
405 auto BufferOrErr = MemoryBuffer::getFileOrSTDIN(File);
406
407 // When we encounter a missing file, make sure we expose its name.
408 if (auto EC = BufferOrErr.getError())
409 if (EC == std::errc::no_such_file_or_directory)
410 ExitOnErr(createStringError(EC, "No such file or directory: '%s'",
411 File.c_str()));
412
413 std::unique_ptr<MemoryBuffer> Buffer =
414 ExitOnErr(errorOrToExpected(std::move(BufferOrErr)));
415
416 std::unique_ptr<Module> M =
417 identify_magic(Buffer->getBuffer()) == file_magic::archive
418 ? loadArFile(argv0, std::move(Buffer), Context)
419 : loadFile(argv0, std::move(Buffer), Context);
420 if (!M) {
421 errs() << argv0 << ": ";
422 WithColor::error() << " loading file '" << File << "'\n";
423 return false;
424 }
425
426 // Note that when ODR merging types cannot verify input files in here When
427 // doing that debug metadata in the src module might already be pointing to
428 // the destination.
429 if (DisableDITypeMap && !NoVerify && verifyModule(*M, &errs())) {
430 errs() << argv0 << ": " << File << ": ";
431 WithColor::error() << "input module is broken!\n";
432 return false;
433 }
434
435 // If a module summary index is supplied, load it so linkInModule can treat
436 // local functions/variables as exported and promote if necessary.
437 if (!SummaryIndex.empty()) {
438 std::unique_ptr<ModuleSummaryIndex> Index =
439 ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex));
440
441 // Conservatively mark all internal values as promoted, since this tool
442 // does not do the ThinLink that would normally determine what values to
443 // promote.
444 for (auto &I : *Index) {
445 for (auto &S : I.second.SummaryList) {
446 if (GlobalValue::isLocalLinkage(S->linkage()))
447 S->setLinkage(GlobalValue::ExternalLinkage);
448 }
449 }
450
451 // Promotion
452 if (renameModuleForThinLTO(*M, *Index,
453 /*ClearDSOLocalOnDeclarations=*/false))
454 return true;
455 }
456
457 if (Verbose)
458 errs() << "Linking in '" << File << "'\n";
459
460 bool Err = false;
461 if (InternalizeLinkedSymbols) {
462 Err = L.linkInModule(
463 std::move(M), ApplicableFlags, [](Module &M, const StringSet<> &GVS) {
464 internalizeModule(M, [&GVS](const GlobalValue &GV) {
465 return !GV.hasName() || (GVS.count(GV.getName()) == 0);
466 });
467 });
468 } else {
469 Err = L.linkInModule(std::move(M), ApplicableFlags);
470 }
471
472 if (Err)
473 return false;
474
475 // Internalization applies to linking of subsequent files.
476 InternalizeLinkedSymbols = Internalize;
477
478 // All linker flags apply to linking of subsequent files.
479 ApplicableFlags = Flags;
480 }
481
482 return true;
483 }
484
main(int argc,char ** argv)485 int main(int argc, char **argv) {
486 InitLLVM X(argc, argv);
487 ExitOnErr.setBanner(std::string(argv[0]) + ": ");
488
489 cl::HideUnrelatedOptions({&LinkCategory, &getColorCategory()});
490 cl::ParseCommandLineOptions(argc, argv, "llvm linker\n");
491
492 // Load bitcode into the new debug info format by default.
493 if (LoadBitcodeIntoNewDbgInfoFormat == cl::boolOrDefault::BOU_UNSET)
494 LoadBitcodeIntoNewDbgInfoFormat = cl::boolOrDefault::BOU_TRUE;
495
496 // Since llvm-link collects multiple IR modules together, for simplicity's
497 // sake we disable the "PreserveInputDbgFormat" flag to enforce a single
498 // debug info format.
499 PreserveInputDbgFormat = cl::boolOrDefault::BOU_FALSE;
500
501 LLVMContext Context;
502 Context.setDiagnosticHandler(std::make_unique<LLVMLinkDiagnosticHandler>(),
503 true);
504
505 if (!DisableDITypeMap)
506 Context.enableDebugTypeODRUniquing();
507
508 auto Composite = std::make_unique<Module>("llvm-link", Context);
509 Linker L(*Composite);
510
511 unsigned Flags = Linker::Flags::None;
512 if (OnlyNeeded)
513 Flags |= Linker::Flags::LinkOnlyNeeded;
514
515 // First add all the regular input files
516 if (!linkFiles(argv[0], Context, L, InputFilenames, Flags))
517 return 1;
518
519 // Next the -override ones.
520 if (!linkFiles(argv[0], Context, L, OverridingInputs,
521 Flags | Linker::Flags::OverrideFromSrc))
522 return 1;
523
524 // Import any functions requested via -import
525 if (!importFunctions(argv[0], *Composite))
526 return 1;
527
528 if (DumpAsm)
529 errs() << "Here's the assembly:\n" << *Composite;
530
531 std::error_code EC;
532 ToolOutputFile Out(OutputFilename, EC,
533 OutputAssembly ? sys::fs::OF_TextWithCRLF
534 : sys::fs::OF_None);
535 if (EC) {
536 WithColor::error() << EC.message() << '\n';
537 return 1;
538 }
539
540 if (!NoVerify && verifyModule(*Composite, &errs())) {
541 errs() << argv[0] << ": ";
542 WithColor::error() << "linked module is broken!\n";
543 return 1;
544 }
545
546 if (Verbose)
547 errs() << "Writing bitcode...\n";
548 auto SetFormat = [&](bool NewFormat) {
549 Composite->setIsNewDbgInfoFormat(NewFormat);
550 if (NewFormat)
551 Composite->removeDebugIntrinsicDeclarations();
552 };
553 if (OutputAssembly) {
554 SetFormat(WriteNewDbgInfoFormat);
555 Composite->print(Out.os(), nullptr, PreserveAssemblyUseListOrder);
556 } else if (Force || !CheckBitcodeOutputToConsole(Out.os())) {
557 SetFormat(UseNewDbgInfoFormat && WriteNewDbgInfoFormatToBitcode);
558 WriteBitcodeToFile(*Composite, Out.os(), PreserveBitcodeUseListOrder);
559 }
560
561 // Declare success.
562 Out.keep();
563
564 return 0;
565 }
566