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::ZeroOrMore, 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::ZeroOrMore, 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 ExitOnError ExitOnErr;
128 
129 // Read the specified bitcode file in and return it. This routine searches the
130 // link path for the specified file to try to find it...
131 //
132 static std::unique_ptr<Module> loadFile(const char *argv0,
133                                         std::unique_ptr<MemoryBuffer> Buffer,
134                                         LLVMContext &Context,
135                                         bool MaterializeMetadata = true) {
136   SMDiagnostic Err;
137   if (Verbose)
138     errs() << "Loading '" << Buffer->getBufferIdentifier() << "'\n";
139   std::unique_ptr<Module> Result;
140   if (DisableLazyLoad)
141     Result = parseIR(*Buffer, Err, Context);
142   else
143     Result =
144         getLazyIRModule(std::move(Buffer), Err, Context, !MaterializeMetadata);
145 
146   if (!Result) {
147     Err.print(argv0, errs());
148     return nullptr;
149   }
150 
151   if (MaterializeMetadata) {
152     ExitOnErr(Result->materializeMetadata());
153     UpgradeDebugInfo(*Result);
154   }
155 
156   return Result;
157 }
158 
159 static std::unique_ptr<Module> loadArFile(const char *Argv0,
160                                           std::unique_ptr<MemoryBuffer> Buffer,
161                                           LLVMContext &Context) {
162   std::unique_ptr<Module> Result(new Module("ArchiveModule", Context));
163   StringRef ArchiveName = Buffer->getBufferIdentifier();
164   if (Verbose)
165     errs() << "Reading library archive file '" << ArchiveName
166            << "' to memory\n";
167   Error Err = Error::success();
168   object::Archive Archive(*Buffer, Err);
169   ExitOnErr(std::move(Err));
170   Linker L(*Result);
171   for (const object::Archive::Child &C : Archive.children(Err)) {
172     Expected<StringRef> Ename = C.getName();
173     if (Error E = Ename.takeError()) {
174       errs() << Argv0 << ": ";
175       WithColor::error() << " failed to read name of archive member"
176                          << ArchiveName << "'\n";
177       return nullptr;
178     }
179     std::string ChildName = Ename.get().str();
180     if (Verbose)
181       errs() << "Parsing member '" << ChildName
182              << "' of archive library to module.\n";
183     SMDiagnostic ParseErr;
184     Expected<MemoryBufferRef> MemBuf = C.getMemoryBufferRef();
185     if (Error E = MemBuf.takeError()) {
186       errs() << Argv0 << ": ";
187       WithColor::error() << " loading memory for member '" << ChildName
188                          << "' of archive library failed'" << ArchiveName
189                          << "'\n";
190       return nullptr;
191     };
192 
193     if (!isBitcode(reinterpret_cast<const unsigned char *>(
194                        MemBuf.get().getBufferStart()),
195                    reinterpret_cast<const unsigned char *>(
196                        MemBuf.get().getBufferEnd()))) {
197       errs() << Argv0 << ": ";
198       WithColor::error() << "  member of archive is not a bitcode file: '"
199                          << ChildName << "'\n";
200       return nullptr;
201     }
202 
203     std::unique_ptr<Module> M;
204     if (DisableLazyLoad)
205       M = parseIR(MemBuf.get(), ParseErr, Context);
206     else
207       M = getLazyIRModule(MemoryBuffer::getMemBuffer(MemBuf.get(), false),
208                           ParseErr, Context);
209 
210     if (!M.get()) {
211       errs() << Argv0 << ": ";
212       WithColor::error() << " parsing member '" << ChildName
213                          << "' of archive library failed'" << ArchiveName
214                          << "'\n";
215       return nullptr;
216     }
217     if (Verbose)
218       errs() << "Linking member '" << ChildName << "' of archive library.\n";
219     if (L.linkInModule(std::move(M)))
220       return nullptr;
221   } // end for each child
222   ExitOnErr(std::move(Err));
223   return Result;
224 }
225 
226 namespace {
227 
228 /// Helper to load on demand a Module from file and cache it for subsequent
229 /// queries during function importing.
230 class ModuleLazyLoaderCache {
231   /// Cache of lazily loaded module for import.
232   StringMap<std::unique_ptr<Module>> ModuleMap;
233 
234   /// Retrieve a Module from the cache or lazily load it on demand.
235   std::function<std::unique_ptr<Module>(const char *argv0,
236                                         const std::string &FileName)>
237       createLazyModule;
238 
239 public:
240   /// Create the loader, Module will be initialized in \p Context.
241   ModuleLazyLoaderCache(std::function<std::unique_ptr<Module>(
242                             const char *argv0, const std::string &FileName)>
243                             createLazyModule)
244       : createLazyModule(std::move(createLazyModule)) {}
245 
246   /// Retrieve a Module from the cache or lazily load it on demand.
247   Module &operator()(const char *argv0, const std::string &FileName);
248 
249   std::unique_ptr<Module> takeModule(const std::string &FileName) {
250     auto I = ModuleMap.find(FileName);
251     assert(I != ModuleMap.end());
252     std::unique_ptr<Module> Ret = std::move(I->second);
253     ModuleMap.erase(I);
254     return Ret;
255   }
256 };
257 
258 // Get a Module for \p FileName from the cache, or load it lazily.
259 Module &ModuleLazyLoaderCache::operator()(const char *argv0,
260                                           const std::string &Identifier) {
261   auto &Module = ModuleMap[Identifier];
262   if (!Module) {
263     Module = createLazyModule(argv0, Identifier);
264     assert(Module && "Failed to create lazy module!");
265   }
266   return *Module;
267 }
268 } // anonymous namespace
269 
270 namespace {
271 struct LLVMLinkDiagnosticHandler : public DiagnosticHandler {
272   bool handleDiagnostics(const DiagnosticInfo &DI) override {
273     unsigned Severity = DI.getSeverity();
274     switch (Severity) {
275     case DS_Error:
276       WithColor::error();
277       break;
278     case DS_Warning:
279       if (SuppressWarnings)
280         return true;
281       WithColor::warning();
282       break;
283     case DS_Remark:
284     case DS_Note:
285       llvm_unreachable("Only expecting warnings and errors");
286     }
287 
288     DiagnosticPrinterRawOStream DP(errs());
289     DI.print(DP);
290     errs() << '\n';
291     return true;
292   }
293 };
294 } // namespace
295 
296 /// Import any functions requested via the -import option.
297 static bool importFunctions(const char *argv0, Module &DestModule) {
298   if (SummaryIndex.empty())
299     return true;
300   std::unique_ptr<ModuleSummaryIndex> Index =
301       ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex));
302 
303   // Map of Module -> List of globals to import from the Module
304   FunctionImporter::ImportMapTy ImportList;
305 
306   auto ModuleLoader = [&DestModule](const char *argv0,
307                                     const std::string &Identifier) {
308     std::unique_ptr<MemoryBuffer> Buffer =
309         ExitOnErr(errorOrToExpected(MemoryBuffer::getFileOrSTDIN(Identifier)));
310     return loadFile(argv0, std::move(Buffer), DestModule.getContext(), false);
311   };
312 
313   ModuleLazyLoaderCache ModuleLoaderCache(ModuleLoader);
314   for (const auto &Import : Imports) {
315     // Identify the requested function and its bitcode source file.
316     size_t Idx = Import.find(':');
317     if (Idx == std::string::npos) {
318       errs() << "Import parameter bad format: " << Import << "\n";
319       return false;
320     }
321     std::string FunctionName = Import.substr(0, Idx);
322     std::string FileName = Import.substr(Idx + 1, std::string::npos);
323 
324     // Load the specified source module.
325     auto &SrcModule = ModuleLoaderCache(argv0, FileName);
326 
327     if (!NoVerify && verifyModule(SrcModule, &errs())) {
328       errs() << argv0 << ": " << FileName;
329       WithColor::error() << "input module is broken!\n";
330       return false;
331     }
332 
333     Function *F = SrcModule.getFunction(FunctionName);
334     if (!F) {
335       errs() << "Ignoring import request for non-existent function "
336              << FunctionName << " from " << FileName << "\n";
337       continue;
338     }
339     // We cannot import weak_any functions without possibly affecting the
340     // order they are seen and selected by the linker, changing program
341     // semantics.
342     if (F->hasWeakAnyLinkage()) {
343       errs() << "Ignoring import request for weak-any function " << FunctionName
344              << " from " << FileName << "\n";
345       continue;
346     }
347 
348     if (Verbose)
349       errs() << "Importing " << FunctionName << " from " << FileName << "\n";
350 
351     auto &Entry = ImportList[FileName];
352     Entry.insert(F->getGUID());
353   }
354   auto CachedModuleLoader = [&](StringRef Identifier) {
355     return ModuleLoaderCache.takeModule(std::string(Identifier));
356   };
357   FunctionImporter Importer(*Index, CachedModuleLoader,
358                             /*ClearDSOLocalOnDeclarations=*/false);
359   ExitOnErr(Importer.importFunctions(DestModule, ImportList));
360 
361   return true;
362 }
363 
364 static bool linkFiles(const char *argv0, LLVMContext &Context, Linker &L,
365                       const cl::list<std::string> &Files, unsigned Flags) {
366   // Filter out flags that don't apply to the first file we load.
367   unsigned ApplicableFlags = Flags & Linker::Flags::OverrideFromSrc;
368   // Similar to some flags, internalization doesn't apply to the first file.
369   bool InternalizeLinkedSymbols = false;
370   for (const auto &File : Files) {
371     std::unique_ptr<MemoryBuffer> Buffer =
372         ExitOnErr(errorOrToExpected(MemoryBuffer::getFileOrSTDIN(File)));
373 
374     std::unique_ptr<Module> M =
375         identify_magic(Buffer->getBuffer()) == file_magic::archive
376             ? loadArFile(argv0, std::move(Buffer), Context)
377             : loadFile(argv0, std::move(Buffer), Context);
378     if (!M.get()) {
379       errs() << argv0 << ": ";
380       WithColor::error() << " loading file '" << File << "'\n";
381       return false;
382     }
383 
384     // Note that when ODR merging types cannot verify input files in here When
385     // doing that debug metadata in the src module might already be pointing to
386     // the destination.
387     if (DisableDITypeMap && !NoVerify && verifyModule(*M, &errs())) {
388       errs() << argv0 << ": " << File << ": ";
389       WithColor::error() << "input module is broken!\n";
390       return false;
391     }
392 
393     // If a module summary index is supplied, load it so linkInModule can treat
394     // local functions/variables as exported and promote if necessary.
395     if (!SummaryIndex.empty()) {
396       std::unique_ptr<ModuleSummaryIndex> Index =
397           ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex));
398 
399       // Conservatively mark all internal values as promoted, since this tool
400       // does not do the ThinLink that would normally determine what values to
401       // promote.
402       for (auto &I : *Index) {
403         for (auto &S : I.second.SummaryList) {
404           if (GlobalValue::isLocalLinkage(S->linkage()))
405             S->setLinkage(GlobalValue::ExternalLinkage);
406         }
407       }
408 
409       // Promotion
410       if (renameModuleForThinLTO(*M, *Index,
411                                  /*ClearDSOLocalOnDeclarations=*/false))
412         return true;
413     }
414 
415     if (Verbose)
416       errs() << "Linking in '" << File << "'\n";
417 
418     bool Err = false;
419     if (InternalizeLinkedSymbols) {
420       Err = L.linkInModule(
421           std::move(M), ApplicableFlags, [](Module &M, const StringSet<> &GVS) {
422             internalizeModule(M, [&GVS](const GlobalValue &GV) {
423               return !GV.hasName() || (GVS.count(GV.getName()) == 0);
424             });
425           });
426     } else {
427       Err = L.linkInModule(std::move(M), ApplicableFlags);
428     }
429 
430     if (Err)
431       return false;
432 
433     // Internalization applies to linking of subsequent files.
434     InternalizeLinkedSymbols = Internalize;
435 
436     // All linker flags apply to linking of subsequent files.
437     ApplicableFlags = Flags;
438   }
439 
440   return true;
441 }
442 
443 int main(int argc, char **argv) {
444   InitLLVM X(argc, argv);
445   ExitOnErr.setBanner(std::string(argv[0]) + ": ");
446 
447   LLVMContext Context;
448   Context.setDiagnosticHandler(std::make_unique<LLVMLinkDiagnosticHandler>(),
449                                true);
450   cl::HideUnrelatedOptions({&LinkCategory, &getColorCategory()});
451   cl::ParseCommandLineOptions(argc, argv, "llvm linker\n");
452 
453   if (!DisableDITypeMap)
454     Context.enableDebugTypeODRUniquing();
455 
456   auto Composite = std::make_unique<Module>("llvm-link", Context);
457   Linker L(*Composite);
458 
459   unsigned Flags = Linker::Flags::None;
460   if (OnlyNeeded)
461     Flags |= Linker::Flags::LinkOnlyNeeded;
462 
463   // First add all the regular input files
464   if (!linkFiles(argv[0], Context, L, InputFilenames, Flags))
465     return 1;
466 
467   // Next the -override ones.
468   if (!linkFiles(argv[0], Context, L, OverridingInputs,
469                  Flags | Linker::Flags::OverrideFromSrc))
470     return 1;
471 
472   // Import any functions requested via -import
473   if (!importFunctions(argv[0], *Composite))
474     return 1;
475 
476   if (DumpAsm)
477     errs() << "Here's the assembly:\n" << *Composite;
478 
479   std::error_code EC;
480   ToolOutputFile Out(OutputFilename, EC,
481                      OutputAssembly ? sys::fs::OF_TextWithCRLF
482                                     : sys::fs::OF_None);
483   if (EC) {
484     WithColor::error() << EC.message() << '\n';
485     return 1;
486   }
487 
488   if (!NoVerify && verifyModule(*Composite, &errs())) {
489     errs() << argv[0] << ": ";
490     WithColor::error() << "linked module is broken!\n";
491     return 1;
492   }
493 
494   if (Verbose)
495     errs() << "Writing bitcode...\n";
496   if (OutputAssembly) {
497     Composite->print(Out.os(), nullptr, PreserveAssemblyUseListOrder);
498   } else if (Force || !CheckBitcodeOutputToConsole(Out.os()))
499     WriteBitcodeToFile(*Composite, Out.os(), PreserveBitcodeUseListOrder);
500 
501   // Declare success.
502   Out.keep();
503 
504   return 0;
505 }
506