1 //===- FunctionImport.cpp - ThinLTO Summary-based Function Import ---------===//
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 file implements Function import based on summaries.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "llvm/Transforms/IPO/FunctionImport.h"
15 
16 #include "llvm/ADT/StringSet.h"
17 #include "llvm/IR/AutoUpgrade.h"
18 #include "llvm/IR/DiagnosticPrinter.h"
19 #include "llvm/IR/IntrinsicInst.h"
20 #include "llvm/IR/Module.h"
21 #include "llvm/IRReader/IRReader.h"
22 #include "llvm/Linker/Linker.h"
23 #include "llvm/Object/FunctionIndexObjectFile.h"
24 #include "llvm/Support/CommandLine.h"
25 #include "llvm/Support/Debug.h"
26 #include "llvm/Support/SourceMgr.h"
27 using namespace llvm;
28 
29 #define DEBUG_TYPE "function-import"
30 
31 /// Limit on instruction count of imported functions.
32 static cl::opt<unsigned> ImportInstrLimit(
33     "import-instr-limit", cl::init(100), cl::Hidden, cl::value_desc("N"),
34     cl::desc("Only import functions with less than N instructions"));
35 
36 // Load lazily a module from \p FileName in \p Context.
37 static std::unique_ptr<Module> loadFile(const std::string &FileName,
38                                         LLVMContext &Context) {
39   SMDiagnostic Err;
40   DEBUG(dbgs() << "Loading '" << FileName << "'\n");
41   std::unique_ptr<Module> Result = getLazyIRFileModule(FileName, Err, Context);
42   if (!Result) {
43     Err.print("function-import", errs());
44     return nullptr;
45   }
46 
47   Result->materializeMetadata();
48   UpgradeDebugInfo(*Result);
49 
50   return Result;
51 }
52 
53 /// Walk through the instructions in \p F looking for external
54 /// calls not already in the \p CalledFunctions set. If any are
55 /// found they are added to the \p Worklist for importing.
56 static void findExternalCalls(const Function &F, StringSet<> &CalledFunctions,
57                               SmallVector<StringRef, 64> &Worklist) {
58   for (auto &BB : F) {
59     for (auto &I : BB) {
60       if (isa<CallInst>(I)) {
61         auto CalledFunction = cast<CallInst>(I).getCalledFunction();
62         // Insert any new external calls that have not already been
63         // added to set/worklist.
64         if (CalledFunction && CalledFunction->hasName() &&
65             CalledFunction->isDeclaration() &&
66             !CalledFunctions.count(CalledFunction->getName())) {
67           CalledFunctions.insert(CalledFunction->getName());
68           Worklist.push_back(CalledFunction->getName());
69         }
70       }
71     }
72   }
73 }
74 
75 // Helper function: given a worklist and an index, will process all the worklist
76 // and import them based on the summary information
77 static unsigned ProcessImportWorklist(
78     Module &DestModule, SmallVector<StringRef, 64> &Worklist,
79     StringSet<> &CalledFunctions, Linker &TheLinker,
80     const FunctionInfoIndex &Index,
81     std::function<std::unique_ptr<Module>(StringRef FileName)> &
82         LazyModuleLoader) {
83   unsigned ImportCount = 0;
84   while (!Worklist.empty()) {
85     auto CalledFunctionName = Worklist.pop_back_val();
86     DEBUG(dbgs() << DestModule.getModuleIdentifier() << "Process import for "
87                  << CalledFunctionName << "\n");
88 
89     // Try to get a summary for this function call.
90     auto InfoList = Index.findFunctionInfoList(CalledFunctionName);
91     if (InfoList == Index.end()) {
92       DEBUG(dbgs() << DestModule.getModuleIdentifier() << "No summary for "
93                    << CalledFunctionName << " Ignoring.\n");
94       continue;
95     }
96     assert(!InfoList->second.empty() && "No summary, error at import?");
97 
98     // Comdat can have multiple entries, FIXME: what do we do with them?
99     auto &Info = InfoList->second[0];
100     assert(Info && "Nullptr in list, error importing summaries?\n");
101 
102     auto *Summary = Info->functionSummary();
103     if (!Summary) {
104       // FIXME: in case we are lazyloading summaries, we can do it now.
105       DEBUG(dbgs() << DestModule.getModuleIdentifier()
106                    << " Missing summary for  " << CalledFunctionName
107                    << ", error at import?\n");
108       llvm_unreachable("Missing summary");
109     }
110 
111     if (Summary->instCount() > ImportInstrLimit) {
112       DEBUG(dbgs() << DestModule.getModuleIdentifier() << " Skip import of "
113                    << CalledFunctionName << " with " << Summary->instCount()
114                    << " instructions (limit " << ImportInstrLimit << ")\n");
115       continue;
116     }
117 
118     // Get the module path from the summary.
119     auto FileName = Summary->modulePath();
120     DEBUG(dbgs() << "Importing " << CalledFunctionName << " from " << FileName
121                  << "\n");
122 
123     // Get the module for the import
124     auto SrcModule = LazyModuleLoader(FileName);
125     assert(&SrcModule->getContext() == &DestModule.getContext());
126 
127     // The function that we will import!
128     GlobalValue *SGV = SrcModule->getNamedValue(CalledFunctionName);
129     StringRef ImportFunctionName = CalledFunctionName;
130     if (!SGV) {
131       // Might be local in source Module, promoted/renamed in DestModule.
132       std::pair<StringRef, StringRef> Split =
133           CalledFunctionName.split(".llvm.");
134       SGV = SrcModule->getNamedValue(Split.first);
135 #ifndef NDEBUG
136       // Assert that Split.second is module id
137       uint64_t ModuleId;
138       assert(!Split.second.getAsInteger(10, ModuleId));
139       assert(ModuleId == Index.getModuleId(FileName));
140 #endif
141     }
142     Function *F = dyn_cast<Function>(SGV);
143     if (!F && isa<GlobalAlias>(SGV)) {
144       auto *SGA = dyn_cast<GlobalAlias>(SGV);
145       F = dyn_cast<Function>(SGA->getBaseObject());
146       ImportFunctionName = F->getName();
147     }
148     if (!F) {
149       errs() << "Can't load function '" << CalledFunctionName << "' in Module '"
150              << FileName << "', error in the summary?\n";
151       llvm_unreachable("Can't load function in Module");
152     }
153 
154     // We cannot import weak_any functions/aliases without possibly affecting
155     // the order they are seen and selected by the linker, changing program
156     // semantics.
157     if (SGV->hasWeakAnyLinkage()) {
158       DEBUG(dbgs() << DestModule.getModuleIdentifier()
159                    << " Ignoring import request for weak-any "
160                    << (isa<Function>(SGV) ? "function " : "alias ")
161                    << CalledFunctionName << " from " << FileName << "\n");
162       continue;
163     }
164 
165     // Link in the specified function.
166     DenseSet<const GlobalValue *> FunctionsToImport;
167     FunctionsToImport.insert(F);
168     if (TheLinker.linkInModule(*SrcModule, Linker::Flags::None, &Index,
169                                &FunctionsToImport))
170       report_fatal_error("Function Import: link error");
171 
172     // Process the newly imported function and add callees to the worklist.
173     GlobalValue *NewGV = DestModule.getNamedValue(ImportFunctionName);
174     assert(NewGV);
175     Function *NewF = dyn_cast<Function>(NewGV);
176     assert(NewF);
177     findExternalCalls(*NewF, CalledFunctions, Worklist);
178     ++ImportCount;
179   }
180   return ImportCount;
181 }
182 
183 // Automatically import functions in Module \p DestModule based on the summaries
184 // index.
185 //
186 // The current implementation imports every called functions that exists in the
187 // summaries index.
188 bool FunctionImporter::importFunctions(Module &DestModule) {
189   DEBUG(dbgs() << "Starting import for Module "
190                << DestModule.getModuleIdentifier() << "\n");
191   unsigned ImportedCount = 0;
192 
193   /// First step is collecting the called external functions.
194   StringSet<> CalledFunctions;
195   SmallVector<StringRef, 64> Worklist;
196   for (auto &F : DestModule) {
197     if (F.isDeclaration() || F.hasFnAttribute(Attribute::OptimizeNone))
198       continue;
199     findExternalCalls(F, CalledFunctions, Worklist);
200   }
201   if (Worklist.empty())
202     return false;
203 
204   /// Second step: for every call to an external function, try to import it.
205 
206   // Linker that will be used for importing function
207   Linker TheLinker(DestModule, DiagnosticHandler);
208 
209   ImportedCount += ProcessImportWorklist(DestModule, Worklist, CalledFunctions,
210                                          TheLinker, Index, ModuleLoader);
211 
212   DEBUG(errs() << "Imported " << ImportedCount << " functions for Module "
213                << DestModule.getModuleIdentifier() << "\n");
214   return ImportedCount;
215 }
216 
217 /// Summary file to use for function importing when using -function-import from
218 /// the command line.
219 static cl::opt<std::string>
220     SummaryFile("summary-file",
221                 cl::desc("The summary file to use for function importing."));
222 
223 static void diagnosticHandler(const DiagnosticInfo &DI) {
224   raw_ostream &OS = errs();
225   DiagnosticPrinterRawOStream DP(OS);
226   DI.print(DP);
227   OS << '\n';
228 }
229 
230 /// Parse the function index out of an IR file and return the function
231 /// index object if found, or nullptr if not.
232 static std::unique_ptr<FunctionInfoIndex>
233 getFunctionIndexForFile(StringRef Path, std::string &Error,
234                         DiagnosticHandlerFunction DiagnosticHandler) {
235   std::unique_ptr<MemoryBuffer> Buffer;
236   ErrorOr<std::unique_ptr<MemoryBuffer>> BufferOrErr =
237       MemoryBuffer::getFile(Path);
238   if (std::error_code EC = BufferOrErr.getError()) {
239     Error = EC.message();
240     return nullptr;
241   }
242   Buffer = std::move(BufferOrErr.get());
243   ErrorOr<std::unique_ptr<object::FunctionIndexObjectFile>> ObjOrErr =
244       object::FunctionIndexObjectFile::create(Buffer->getMemBufferRef(),
245                                               DiagnosticHandler);
246   if (std::error_code EC = ObjOrErr.getError()) {
247     Error = EC.message();
248     return nullptr;
249   }
250   return (*ObjOrErr)->takeIndex();
251 }
252 
253 /// Pass that performs cross-module function import provided a summary file.
254 class FunctionImportPass : public ModulePass {
255   /// Optional function summary index to use for importing, otherwise
256   /// the summary-file option must be specified.
257   FunctionInfoIndex *Index;
258 
259 public:
260   /// Pass identification, replacement for typeid
261   static char ID;
262 
263   /// Specify pass name for debug output
264   const char *getPassName() const override {
265     return "Function Importing";
266   }
267 
268   explicit FunctionImportPass(FunctionInfoIndex *Index = nullptr)
269       : ModulePass(ID), Index(Index) {}
270 
271   bool runOnModule(Module &M) override {
272     if (SummaryFile.empty() && !Index)
273       report_fatal_error("error: -function-import requires -summary-file or "
274                          "file from frontend\n");
275     std::unique_ptr<FunctionInfoIndex> IndexPtr;
276     if (!SummaryFile.empty()) {
277       if (Index)
278         report_fatal_error("error: -summary-file and index from frontend\n");
279       std::string Error;
280       IndexPtr = getFunctionIndexForFile(SummaryFile, Error, diagnosticHandler);
281       if (!IndexPtr) {
282         errs() << "Error loading file '" << SummaryFile << "': " << Error
283                << "\n";
284         return false;
285       }
286       Index = IndexPtr.get();
287     }
288 
289     // Perform the import now.
290     auto ModuleLoader = [&M](StringRef Identifier) {
291       return loadFile(Identifier, M.getContext());
292     };
293     FunctionImporter Importer(*Index, diagnosticHandler, ModuleLoader);
294     return Importer.importFunctions(M);
295 
296     return false;
297   }
298 };
299 
300 char FunctionImportPass::ID = 0;
301 INITIALIZE_PASS_BEGIN(FunctionImportPass, "function-import",
302                       "Summary Based Function Import", false, false)
303 INITIALIZE_PASS_END(FunctionImportPass, "function-import",
304                     "Summary Based Function Import", false, false)
305 
306 namespace llvm {
307 Pass *createFunctionImportPass(FunctionInfoIndex *Index = nullptr) {
308   return new FunctionImportPass(Index);
309 }
310 }
311