1 //===- llvm-link.cpp - Low-level LLVM linker ------------------------------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // This utility may be invoked in the following manner:
11 //  llvm-link a.bc b.bc c.bc -o x.bc
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "llvm/ADT/STLExtras.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/Support/CommandLine.h"
28 #include "llvm/Support/FileSystem.h"
29 #include "llvm/Support/ManagedStatic.h"
30 #include "llvm/Support/Path.h"
31 #include "llvm/Support/PrettyStackTrace.h"
32 #include "llvm/Support/Signals.h"
33 #include "llvm/Support/SourceMgr.h"
34 #include "llvm/Support/SystemUtils.h"
35 #include "llvm/Support/ToolOutputFile.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>
45 InputFilenames(cl::Positional, cl::OneOrMore,
46                cl::desc("<input bitcode files>"));
47 
48 static cl::list<std::string> OverridingInputs(
49     "override", cl::ZeroOrMore, cl::value_desc("filename"),
50     cl::desc(
51         "input bitcode file which can override previously defined symbol(s)"));
52 
53 // Option to simulate function importing for testing. This enables using
54 // llvm-link to simulate ThinLTO backend processes.
55 static cl::list<std::string> Imports(
56     "import", cl::ZeroOrMore, cl::value_desc("function:filename"),
57     cl::desc("Pair of function name and filename, where function should be "
58              "imported from bitcode in filename"));
59 
60 // Option to support testing of function importing. The module summary
61 // must be specified in the case were we request imports via the -import
62 // option, as well as when compiling any module with functions that may be
63 // exported (imported by a different llvm-link -import invocation), to ensure
64 // consistent promotion and renaming of locals.
65 static cl::opt<std::string>
66     SummaryIndex("summary-index", cl::desc("Module summary index filename"),
67                  cl::init(""), cl::value_desc("filename"));
68 
69 static cl::opt<std::string>
70 OutputFilename("o", cl::desc("Override output filename"), cl::init("-"),
71                cl::value_desc("filename"));
72 
73 static cl::opt<bool>
74 Internalize("internalize", 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>
81 OnlyNeeded("only-needed", cl::desc("Link only needed symbols"));
82 
83 static cl::opt<bool>
84 Force("f", cl::desc("Enable binary output on terminals"));
85 
86 static cl::opt<bool>
87     DisableLazyLoad("disable-lazy-loading",
88                     cl::desc("Disable lazy module loading"));
89 
90 static cl::opt<bool>
91     OutputAssembly("S", cl::desc("Write output as LLVM assembly"), cl::Hidden);
92 
93 static cl::opt<bool>
94 Verbose("v", cl::desc("Print information about actions taken"));
95 
96 static cl::opt<bool>
97 DumpAsm("d", cl::desc("Print assembly as linked"), cl::Hidden);
98 
99 static cl::opt<bool>
100 SuppressWarnings("suppress-warnings", cl::desc("Suppress all linking warnings"),
101                  cl::init(false));
102 
103 static cl::opt<bool> PreserveBitcodeUseListOrder(
104     "preserve-bc-uselistorder",
105     cl::desc("Preserve use-list order when writing LLVM bitcode."),
106     cl::init(true), cl::Hidden);
107 
108 static cl::opt<bool> PreserveAssemblyUseListOrder(
109     "preserve-ll-uselistorder",
110     cl::desc("Preserve use-list order when writing LLVM assembly."),
111     cl::init(false), cl::Hidden);
112 
113 static ExitOnError ExitOnErr;
114 
115 // Read the specified bitcode file in and return it. This routine searches the
116 // link path for the specified file to try to find it...
117 //
118 static std::unique_ptr<Module> loadFile(const char *argv0,
119                                         const std::string &FN,
120                                         LLVMContext &Context,
121                                         bool MaterializeMetadata = true) {
122   SMDiagnostic Err;
123   if (Verbose) errs() << "Loading '" << FN << "'\n";
124   std::unique_ptr<Module> Result;
125   if (DisableLazyLoad)
126     Result = parseIRFile(FN, Err, Context);
127   else
128     Result = getLazyIRFileModule(FN, Err, Context, !MaterializeMetadata);
129 
130   if (!Result) {
131     Err.print(argv0, errs());
132     return nullptr;
133   }
134 
135   if (MaterializeMetadata) {
136     ExitOnErr(Result->materializeMetadata());
137     UpgradeDebugInfo(*Result);
138   }
139 
140   return Result;
141 }
142 
143 namespace {
144 
145 /// Helper to load on demand a Module from file and cache it for subsequent
146 /// queries during function importing.
147 class ModuleLazyLoaderCache {
148   /// Cache of lazily loaded module for import.
149   StringMap<std::unique_ptr<Module>> ModuleMap;
150 
151   /// Retrieve a Module from the cache or lazily load it on demand.
152   std::function<std::unique_ptr<Module>(const char *argv0,
153                                         const std::string &FileName)>
154       createLazyModule;
155 
156 public:
157   /// Create the loader, Module will be initialized in \p Context.
158   ModuleLazyLoaderCache(std::function<std::unique_ptr<Module>(
159                             const char *argv0, const std::string &FileName)>
160                             createLazyModule)
161       : createLazyModule(std::move(createLazyModule)) {}
162 
163   /// Retrieve a Module from the cache or lazily load it on demand.
164   Module &operator()(const char *argv0, const std::string &FileName);
165 
166   std::unique_ptr<Module> takeModule(const std::string &FileName) {
167     auto I = ModuleMap.find(FileName);
168     assert(I != ModuleMap.end());
169     std::unique_ptr<Module> Ret = std::move(I->second);
170     ModuleMap.erase(I);
171     return Ret;
172   }
173 };
174 
175 // Get a Module for \p FileName from the cache, or load it lazily.
176 Module &ModuleLazyLoaderCache::operator()(const char *argv0,
177                                           const std::string &Identifier) {
178   auto &Module = ModuleMap[Identifier];
179   if (!Module)
180     Module = createLazyModule(argv0, Identifier);
181   return *Module;
182 }
183 } // anonymous namespace
184 
185 static void diagnosticHandler(const DiagnosticInfo &DI, void *C) {
186   unsigned Severity = DI.getSeverity();
187   switch (Severity) {
188   case DS_Error:
189     errs() << "ERROR: ";
190     break;
191   case DS_Warning:
192     if (SuppressWarnings)
193       return;
194     errs() << "WARNING: ";
195     break;
196   case DS_Remark:
197   case DS_Note:
198     llvm_unreachable("Only expecting warnings and errors");
199   }
200 
201   DiagnosticPrinterRawOStream DP(errs());
202   DI.print(DP);
203   errs() << '\n';
204 }
205 
206 /// Import any functions requested via the -import option.
207 static bool importFunctions(const char *argv0, Module &DestModule) {
208   if (SummaryIndex.empty())
209     return true;
210   std::unique_ptr<ModuleSummaryIndex> Index =
211       ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex));
212 
213   // Map of Module -> List of globals to import from the Module
214   FunctionImporter::ImportMapTy ImportList;
215 
216   auto ModuleLoader = [&DestModule](const char *argv0,
217                                     const std::string &Identifier) {
218     return loadFile(argv0, Identifier, DestModule.getContext(), false);
219   };
220 
221   ModuleLazyLoaderCache ModuleLoaderCache(ModuleLoader);
222   for (const auto &Import : Imports) {
223     // Identify the requested function and its bitcode source file.
224     size_t Idx = Import.find(':');
225     if (Idx == std::string::npos) {
226       errs() << "Import parameter bad format: " << Import << "\n";
227       return false;
228     }
229     std::string FunctionName = Import.substr(0, Idx);
230     std::string FileName = Import.substr(Idx + 1, std::string::npos);
231 
232     // Load the specified source module.
233     auto &SrcModule = ModuleLoaderCache(argv0, FileName);
234 
235     if (verifyModule(SrcModule, &errs())) {
236       errs() << argv0 << ": " << FileName
237              << ": error: input module is broken!\n";
238       return false;
239     }
240 
241     Function *F = SrcModule.getFunction(FunctionName);
242     if (!F) {
243       errs() << "Ignoring import request for non-existent function "
244              << FunctionName << " from " << FileName << "\n";
245       continue;
246     }
247     // We cannot import weak_any functions without possibly affecting the
248     // order they are seen and selected by the linker, changing program
249     // semantics.
250     if (F->hasWeakAnyLinkage()) {
251       errs() << "Ignoring import request for weak-any function " << FunctionName
252              << " from " << FileName << "\n";
253       continue;
254     }
255 
256     if (Verbose)
257       errs() << "Importing " << FunctionName << " from " << FileName << "\n";
258 
259     auto &Entry = ImportList[FileName];
260     Entry.insert(std::make_pair(F->getGUID(), /* (Unused) threshold */ 1.0));
261   }
262   auto CachedModuleLoader = [&](StringRef Identifier) {
263     return ModuleLoaderCache.takeModule(Identifier);
264   };
265   FunctionImporter Importer(*Index, CachedModuleLoader);
266   ExitOnErr(Importer.importFunctions(DestModule, ImportList));
267 
268   return true;
269 }
270 
271 static bool linkFiles(const char *argv0, LLVMContext &Context, Linker &L,
272                       const cl::list<std::string> &Files,
273                       unsigned Flags) {
274   // Filter out flags that don't apply to the first file we load.
275   unsigned ApplicableFlags = Flags & Linker::Flags::OverrideFromSrc;
276   // Similar to some flags, internalization doesn't apply to the first file.
277   bool InternalizeLinkedSymbols = false;
278   for (const auto &File : Files) {
279     std::unique_ptr<Module> M = loadFile(argv0, File, Context);
280     if (!M.get()) {
281       errs() << argv0 << ": error loading file '" << File << "'\n";
282       return false;
283     }
284 
285     // Note that when ODR merging types cannot verify input files in here When
286     // doing that debug metadata in the src module might already be pointing to
287     // the destination.
288     if (DisableDITypeMap && verifyModule(*M, &errs())) {
289       errs() << argv0 << ": " << File << ": error: input module is broken!\n";
290       return false;
291     }
292 
293     // If a module summary index is supplied, load it so linkInModule can treat
294     // local functions/variables as exported and promote if necessary.
295     if (!SummaryIndex.empty()) {
296       std::unique_ptr<ModuleSummaryIndex> Index =
297           ExitOnErr(llvm::getModuleSummaryIndexForFile(SummaryIndex));
298 
299       // Conservatively mark all internal values as promoted, since this tool
300       // does not do the ThinLink that would normally determine what values to
301       // promote.
302       for (auto &I : *Index) {
303         for (auto &S : I.second.SummaryList) {
304           if (GlobalValue::isLocalLinkage(S->linkage()))
305             S->setLinkage(GlobalValue::ExternalLinkage);
306         }
307       }
308 
309       // Promotion
310       if (renameModuleForThinLTO(*M, *Index))
311         return true;
312     }
313 
314     if (Verbose)
315       errs() << "Linking in '" << File << "'\n";
316 
317     bool Err = false;
318     if (InternalizeLinkedSymbols) {
319       Err = L.linkInModule(
320           std::move(M), ApplicableFlags, [](Module &M, const StringSet<> &GVS) {
321             internalizeModule(M, [&GVS](const GlobalValue &GV) {
322               return !GV.hasName() || (GVS.count(GV.getName()) == 0);
323             });
324           });
325     } else {
326       Err = L.linkInModule(std::move(M), ApplicableFlags);
327     }
328 
329     if (Err)
330       return false;
331 
332     // Internalization applies to linking of subsequent files.
333     InternalizeLinkedSymbols = Internalize;
334 
335     // All linker flags apply to linking of subsequent files.
336     ApplicableFlags = Flags;
337   }
338 
339   return true;
340 }
341 
342 int main(int argc, char **argv) {
343   // Print a stack trace if we signal out.
344   sys::PrintStackTraceOnErrorSignal(argv[0]);
345   PrettyStackTraceProgram X(argc, argv);
346 
347   ExitOnErr.setBanner(std::string(argv[0]) + ": ");
348 
349   LLVMContext Context;
350   Context.setDiagnosticHandler(diagnosticHandler, nullptr, true);
351 
352   llvm_shutdown_obj Y;  // Call llvm_shutdown() on exit.
353   cl::ParseCommandLineOptions(argc, argv, "llvm linker\n");
354 
355   if (!DisableDITypeMap)
356     Context.enableDebugTypeODRUniquing();
357 
358   auto Composite = make_unique<Module>("llvm-link", Context);
359   Linker L(*Composite);
360 
361   unsigned Flags = Linker::Flags::None;
362   if (OnlyNeeded)
363     Flags |= Linker::Flags::LinkOnlyNeeded;
364 
365   // First add all the regular input files
366   if (!linkFiles(argv[0], Context, L, InputFilenames, Flags))
367     return 1;
368 
369   // Next the -override ones.
370   if (!linkFiles(argv[0], Context, L, OverridingInputs,
371                  Flags | Linker::Flags::OverrideFromSrc))
372     return 1;
373 
374   // Import any functions requested via -import
375   if (!importFunctions(argv[0], *Composite))
376     return 1;
377 
378   if (DumpAsm) errs() << "Here's the assembly:\n" << *Composite;
379 
380   std::error_code EC;
381   tool_output_file Out(OutputFilename, EC, sys::fs::F_None);
382   if (EC) {
383     errs() << EC.message() << '\n';
384     return 1;
385   }
386 
387   if (verifyModule(*Composite, &errs())) {
388     errs() << argv[0] << ": error: linked module is broken!\n";
389     return 1;
390   }
391 
392   if (Verbose) errs() << "Writing bitcode...\n";
393   if (OutputAssembly) {
394     Composite->print(Out.os(), nullptr, PreserveAssemblyUseListOrder);
395   } else if (Force || !CheckBitcodeOutputToConsole(Out.os(), true))
396     WriteBitcodeToFile(Composite.get(), Out.os(), PreserveBitcodeUseListOrder);
397 
398   // Declare success.
399   Out.keep();
400 
401   return 0;
402 }
403