1 //===- ExtractFunction.cpp - Extract a function from Program --------------===//
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 several methods that are used to extract functions,
11 // loops, or portions of a module from the rest of the module.
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "BugDriver.h"
16 #include "llvm/IR/Constants.h"
17 #include "llvm/IR/DataLayout.h"
18 #include "llvm/IR/DerivedTypes.h"
19 #include "llvm/IR/LLVMContext.h"
20 #include "llvm/IR/LegacyPassManager.h"
21 #include "llvm/IR/Module.h"
22 #include "llvm/IR/Verifier.h"
23 #include "llvm/Pass.h"
24 #include "llvm/Support/CommandLine.h"
25 #include "llvm/Support/Debug.h"
26 #include "llvm/Support/FileUtilities.h"
27 #include "llvm/Support/Path.h"
28 #include "llvm/Support/Signals.h"
29 #include "llvm/Support/ToolOutputFile.h"
30 #include "llvm/Transforms/IPO.h"
31 #include "llvm/Transforms/Scalar.h"
32 #include "llvm/Transforms/Utils/Cloning.h"
33 #include "llvm/Transforms/Utils/CodeExtractor.h"
34 #include <set>
35 using namespace llvm;
36 
37 #define DEBUG_TYPE "bugpoint"
38 
39 namespace llvm {
40 bool DisableSimplifyCFG = false;
41 extern cl::opt<std::string> OutputPrefix;
42 } // End llvm namespace
43 
44 namespace {
45 cl::opt<bool> NoDCE("disable-dce",
46                     cl::desc("Do not use the -dce pass to reduce testcases"));
47 cl::opt<bool, true>
48     NoSCFG("disable-simplifycfg", cl::location(DisableSimplifyCFG),
49            cl::desc("Do not use the -simplifycfg pass to reduce testcases"));
50 
51 Function *globalInitUsesExternalBA(GlobalVariable *GV) {
52   if (!GV->hasInitializer())
53     return nullptr;
54 
55   Constant *I = GV->getInitializer();
56 
57   // walk the values used by the initializer
58   // (and recurse into things like ConstantExpr)
59   std::vector<Constant *> Todo;
60   std::set<Constant *> Done;
61   Todo.push_back(I);
62 
63   while (!Todo.empty()) {
64     Constant *V = Todo.back();
65     Todo.pop_back();
66     Done.insert(V);
67 
68     if (BlockAddress *BA = dyn_cast<BlockAddress>(V)) {
69       Function *F = BA->getFunction();
70       if (F->isDeclaration())
71         return F;
72     }
73 
74     for (User::op_iterator i = V->op_begin(), e = V->op_end(); i != e; ++i) {
75       Constant *C = dyn_cast<Constant>(*i);
76       if (C && !isa<GlobalValue>(C) && !Done.count(C))
77         Todo.push_back(C);
78     }
79   }
80   return nullptr;
81 }
82 } // end anonymous namespace
83 
84 std::unique_ptr<Module>
85 BugDriver::deleteInstructionFromProgram(const Instruction *I,
86                                         unsigned Simplification) {
87   // FIXME, use vmap?
88   Module *Clone = CloneModule(Program).release();
89 
90   const BasicBlock *PBB = I->getParent();
91   const Function *PF = PBB->getParent();
92 
93   Module::iterator RFI = Clone->begin(); // Get iterator to corresponding fn
94   std::advance(
95       RFI, std::distance(PF->getParent()->begin(), Module::const_iterator(PF)));
96 
97   Function::iterator RBI = RFI->begin(); // Get iterator to corresponding BB
98   std::advance(RBI, std::distance(PF->begin(), Function::const_iterator(PBB)));
99 
100   BasicBlock::iterator RI = RBI->begin(); // Get iterator to corresponding inst
101   std::advance(RI, std::distance(PBB->begin(), BasicBlock::const_iterator(I)));
102   Instruction *TheInst = &*RI; // Got the corresponding instruction!
103 
104   // If this instruction produces a value, replace any users with null values
105   if (!TheInst->getType()->isVoidTy())
106     TheInst->replaceAllUsesWith(Constant::getNullValue(TheInst->getType()));
107 
108   // Remove the instruction from the program.
109   TheInst->getParent()->getInstList().erase(TheInst);
110 
111   // Spiff up the output a little bit.
112   std::vector<std::string> Passes;
113 
114   /// Can we get rid of the -disable-* options?
115   if (Simplification > 1 && !NoDCE)
116     Passes.push_back("dce");
117   if (Simplification && !DisableSimplifyCFG)
118     Passes.push_back("simplifycfg"); // Delete dead control flow
119 
120   Passes.push_back("verify");
121   std::unique_ptr<Module> New = runPassesOn(Clone, Passes);
122   delete Clone;
123   if (!New) {
124     errs() << "Instruction removal failed.  Sorry. :(  Please report a bug!\n";
125     exit(1);
126   }
127   return New;
128 }
129 
130 std::unique_ptr<Module>
131 BugDriver::performFinalCleanups(Module *M, bool MayModifySemantics) {
132   // Make all functions external, so GlobalDCE doesn't delete them...
133   for (Module::iterator I = M->begin(), E = M->end(); I != E; ++I)
134     I->setLinkage(GlobalValue::ExternalLinkage);
135 
136   std::vector<std::string> CleanupPasses;
137   CleanupPasses.push_back("globaldce");
138 
139   if (MayModifySemantics)
140     CleanupPasses.push_back("deadarghaX0r");
141   else
142     CleanupPasses.push_back("deadargelim");
143 
144   std::unique_ptr<Module> New = runPassesOn(M, CleanupPasses);
145   if (!New) {
146     errs() << "Final cleanups failed.  Sorry. :(  Please report a bug!\n";
147     return nullptr;
148   }
149   delete M;
150   return New;
151 }
152 
153 std::unique_ptr<Module> BugDriver::extractLoop(Module *M) {
154   std::vector<std::string> LoopExtractPasses;
155   LoopExtractPasses.push_back("loop-extract-single");
156 
157   std::unique_ptr<Module> NewM = runPassesOn(M, LoopExtractPasses);
158   if (!NewM) {
159     outs() << "*** Loop extraction failed: ";
160     EmitProgressBitcode(M, "loopextraction", true);
161     outs() << "*** Sorry. :(  Please report a bug!\n";
162     return nullptr;
163   }
164 
165   // Check to see if we created any new functions.  If not, no loops were
166   // extracted and we should return null.  Limit the number of loops we extract
167   // to avoid taking forever.
168   static unsigned NumExtracted = 32;
169   if (M->size() == NewM->size() || --NumExtracted == 0) {
170     return nullptr;
171   } else {
172     assert(M->size() < NewM->size() && "Loop extract removed functions?");
173     Module::iterator MI = NewM->begin();
174     for (unsigned i = 0, e = M->size(); i != e; ++i)
175       ++MI;
176   }
177 
178   return NewM;
179 }
180 
181 static void eliminateAliases(GlobalValue *GV) {
182   // First, check whether a GlobalAlias references this definition.
183   // GlobalAlias MAY NOT reference declarations.
184   for (;;) {
185     // 1. Find aliases
186     SmallVector<GlobalAlias *, 1> aliases;
187     Module *M = GV->getParent();
188     for (Module::alias_iterator I = M->alias_begin(), E = M->alias_end();
189          I != E; ++I)
190       if (I->getAliasee()->stripPointerCasts() == GV)
191         aliases.push_back(&*I);
192     if (aliases.empty())
193       break;
194     // 2. Resolve aliases
195     for (unsigned i = 0, e = aliases.size(); i < e; ++i) {
196       aliases[i]->replaceAllUsesWith(aliases[i]->getAliasee());
197       aliases[i]->eraseFromParent();
198     }
199     // 3. Repeat until no more aliases found; there might
200     // be an alias to an alias...
201   }
202 }
203 
204 //
205 // DeleteGlobalInitializer - "Remove" the global variable by deleting its
206 // initializer,
207 // making it external.
208 //
209 void llvm::DeleteGlobalInitializer(GlobalVariable *GV) {
210   eliminateAliases(GV);
211   GV->setInitializer(nullptr);
212 }
213 
214 // DeleteFunctionBody - "Remove" the function by deleting all of its basic
215 // blocks, making it external.
216 //
217 void llvm::DeleteFunctionBody(Function *F) {
218   eliminateAliases(F);
219   // Function declarations can't have comdats.
220   F->setComdat(nullptr);
221 
222   // delete the body of the function...
223   F->deleteBody();
224   assert(F->isDeclaration() && "This didn't make the function external!");
225 }
226 
227 /// GetTorInit - Given a list of entries for static ctors/dtors, return them
228 /// as a constant array.
229 static Constant *GetTorInit(std::vector<std::pair<Function *, int>> &TorList) {
230   assert(!TorList.empty() && "Don't create empty tor list!");
231   std::vector<Constant *> ArrayElts;
232   Type *Int32Ty = Type::getInt32Ty(TorList[0].first->getContext());
233 
234   StructType *STy =
235       StructType::get(Int32Ty, TorList[0].first->getType(), nullptr);
236   for (unsigned i = 0, e = TorList.size(); i != e; ++i) {
237     Constant *Elts[] = {ConstantInt::get(Int32Ty, TorList[i].second),
238                         TorList[i].first};
239     ArrayElts.push_back(ConstantStruct::get(STy, Elts));
240   }
241   return ConstantArray::get(
242       ArrayType::get(ArrayElts[0]->getType(), ArrayElts.size()), ArrayElts);
243 }
244 
245 /// SplitStaticCtorDtor - A module was recently split into two parts, M1/M2, and
246 /// M1 has all of the global variables.  If M2 contains any functions that are
247 /// static ctors/dtors, we need to add an llvm.global_[cd]tors global to M2, and
248 /// prune appropriate entries out of M1s list.
249 static void SplitStaticCtorDtor(const char *GlobalName, Module *M1, Module *M2,
250                                 ValueToValueMapTy &VMap) {
251   GlobalVariable *GV = M1->getNamedGlobal(GlobalName);
252   if (!GV || GV->isDeclaration() || GV->hasLocalLinkage() || !GV->use_empty())
253     return;
254 
255   std::vector<std::pair<Function *, int>> M1Tors, M2Tors;
256   ConstantArray *InitList = dyn_cast<ConstantArray>(GV->getInitializer());
257   if (!InitList)
258     return;
259 
260   for (unsigned i = 0, e = InitList->getNumOperands(); i != e; ++i) {
261     if (ConstantStruct *CS =
262             dyn_cast<ConstantStruct>(InitList->getOperand(i))) {
263       if (CS->getNumOperands() != 2)
264         return; // Not array of 2-element structs.
265 
266       if (CS->getOperand(1)->isNullValue())
267         break; // Found a null terminator, stop here.
268 
269       ConstantInt *CI = dyn_cast<ConstantInt>(CS->getOperand(0));
270       int Priority = CI ? CI->getSExtValue() : 0;
271 
272       Constant *FP = CS->getOperand(1);
273       if (ConstantExpr *CE = dyn_cast<ConstantExpr>(FP))
274         if (CE->isCast())
275           FP = CE->getOperand(0);
276       if (Function *F = dyn_cast<Function>(FP)) {
277         if (!F->isDeclaration())
278           M1Tors.push_back(std::make_pair(F, Priority));
279         else {
280           // Map to M2's version of the function.
281           F = cast<Function>(VMap[F]);
282           M2Tors.push_back(std::make_pair(F, Priority));
283         }
284       }
285     }
286   }
287 
288   GV->eraseFromParent();
289   if (!M1Tors.empty()) {
290     Constant *M1Init = GetTorInit(M1Tors);
291     new GlobalVariable(*M1, M1Init->getType(), false,
292                        GlobalValue::AppendingLinkage, M1Init, GlobalName);
293   }
294 
295   GV = M2->getNamedGlobal(GlobalName);
296   assert(GV && "Not a clone of M1?");
297   assert(GV->use_empty() && "llvm.ctors shouldn't have uses!");
298 
299   GV->eraseFromParent();
300   if (!M2Tors.empty()) {
301     Constant *M2Init = GetTorInit(M2Tors);
302     new GlobalVariable(*M2, M2Init->getType(), false,
303                        GlobalValue::AppendingLinkage, M2Init, GlobalName);
304   }
305 }
306 
307 std::unique_ptr<Module>
308 llvm::SplitFunctionsOutOfModule(Module *M, const std::vector<Function *> &F,
309                                 ValueToValueMapTy &VMap) {
310   // Make sure functions & globals are all external so that linkage
311   // between the two modules will work.
312   for (Module::iterator I = M->begin(), E = M->end(); I != E; ++I)
313     I->setLinkage(GlobalValue::ExternalLinkage);
314   for (Module::global_iterator I = M->global_begin(), E = M->global_end();
315        I != E; ++I) {
316     if (I->hasName() && I->getName()[0] == '\01')
317       I->setName(I->getName().substr(1));
318     I->setLinkage(GlobalValue::ExternalLinkage);
319   }
320 
321   ValueToValueMapTy NewVMap;
322   std::unique_ptr<Module> New = CloneModule(M, NewVMap);
323 
324   // Remove the Test functions from the Safe module
325   std::set<Function *> TestFunctions;
326   for (unsigned i = 0, e = F.size(); i != e; ++i) {
327     Function *TNOF = cast<Function>(VMap[F[i]]);
328     DEBUG(errs() << "Removing function ");
329     DEBUG(TNOF->printAsOperand(errs(), false));
330     DEBUG(errs() << "\n");
331     TestFunctions.insert(cast<Function>(NewVMap[TNOF]));
332     DeleteFunctionBody(TNOF); // Function is now external in this module!
333   }
334 
335   // Remove the Safe functions from the Test module
336   for (Function &I : *New)
337     if (!TestFunctions.count(&I))
338       DeleteFunctionBody(&I);
339 
340   // Try to split the global initializers evenly
341   for (GlobalVariable &I : M->globals()) {
342     GlobalVariable *GV = cast<GlobalVariable>(NewVMap[&I]);
343     if (Function *TestFn = globalInitUsesExternalBA(&I)) {
344       if (Function *SafeFn = globalInitUsesExternalBA(GV)) {
345         errs() << "*** Error: when reducing functions, encountered "
346                   "the global '";
347         GV->printAsOperand(errs(), false);
348         errs() << "' with an initializer that references blockaddresses "
349                   "from safe function '"
350                << SafeFn->getName() << "' and from test function '"
351                << TestFn->getName() << "'.\n";
352         exit(1);
353       }
354       DeleteGlobalInitializer(&I); // Delete the initializer to make it external
355     } else {
356       // If we keep it in the safe module, then delete it in the test module
357       DeleteGlobalInitializer(GV);
358     }
359   }
360 
361   // Make sure that there is a global ctor/dtor array in both halves of the
362   // module if they both have static ctor/dtor functions.
363   SplitStaticCtorDtor("llvm.global_ctors", M, New.get(), NewVMap);
364   SplitStaticCtorDtor("llvm.global_dtors", M, New.get(), NewVMap);
365 
366   return New;
367 }
368 
369 //===----------------------------------------------------------------------===//
370 // Basic Block Extraction Code
371 //===----------------------------------------------------------------------===//
372 
373 std::unique_ptr<Module>
374 BugDriver::extractMappedBlocksFromModule(const std::vector<BasicBlock *> &BBs,
375                                          Module *M) {
376   SmallString<128> Filename;
377   int FD;
378   std::error_code EC = sys::fs::createUniqueFile(
379       OutputPrefix + "-extractblocks%%%%%%%", FD, Filename);
380   if (EC) {
381     outs() << "*** Basic Block extraction failed!\n";
382     errs() << "Error creating temporary file: " << EC.message() << "\n";
383     EmitProgressBitcode(M, "basicblockextractfail", true);
384     return nullptr;
385   }
386   sys::RemoveFileOnSignal(Filename);
387 
388   tool_output_file BlocksToNotExtractFile(Filename.c_str(), FD);
389   for (std::vector<BasicBlock *>::const_iterator I = BBs.begin(), E = BBs.end();
390        I != E; ++I) {
391     BasicBlock *BB = *I;
392     // If the BB doesn't have a name, give it one so we have something to key
393     // off of.
394     if (!BB->hasName())
395       BB->setName("tmpbb");
396     BlocksToNotExtractFile.os() << BB->getParent()->getName() << " "
397                                 << BB->getName() << "\n";
398   }
399   BlocksToNotExtractFile.os().close();
400   if (BlocksToNotExtractFile.os().has_error()) {
401     errs() << "Error writing list of blocks to not extract\n";
402     EmitProgressBitcode(M, "basicblockextractfail", true);
403     BlocksToNotExtractFile.os().clear_error();
404     return nullptr;
405   }
406   BlocksToNotExtractFile.keep();
407 
408   std::string uniqueFN = "--extract-blocks-file=";
409   uniqueFN += Filename.str();
410   const char *ExtraArg = uniqueFN.c_str();
411 
412   std::vector<std::string> PI;
413   PI.push_back("extract-blocks");
414   std::unique_ptr<Module> Ret = runPassesOn(M, PI, 1, &ExtraArg);
415 
416   sys::fs::remove(Filename.c_str());
417 
418   if (!Ret) {
419     outs() << "*** Basic Block extraction failed, please report a bug!\n";
420     EmitProgressBitcode(M, "basicblockextractfail", true);
421   }
422   return Ret;
423 }
424