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