StripSymbols.cpp revision 55d4c38074145bf9f594142b6b4cdca60699f4d1 
1//===- StripSymbols.cpp - Strip symbols and debug info from a module ------===//
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// The StripSymbols transformation implements code stripping. Specifically, it
11// can delete:
12//
13//   * names for virtual registers
14//   * symbols for internal globals and functions
15//   * debug information
16//
17// Note that this transformation makes code much less readable, so it should
18// only be used in situations where the 'strip' utility would be used, such as
19// reducing code size or making it harder to reverse engineer code.
20//
21//===----------------------------------------------------------------------===//
22
23#include "llvm/Transforms/IPO.h"
24#include "llvm/ADT/DenseMap.h"
25#include "llvm/ADT/SmallPtrSet.h"
26#include "llvm/DebugInfo.h"
27#include "llvm/IR/Constants.h"
28#include "llvm/IR/DerivedTypes.h"
29#include "llvm/IR/Instructions.h"
30#include "llvm/IR/Module.h"
31#include "llvm/IR/TypeFinder.h"
32#include "llvm/IR/ValueSymbolTable.h"
33#include "llvm/Pass.h"
34#include "llvm/Transforms/Utils/Local.h"
35using namespace llvm;
36
37namespace {
38  class StripSymbols : public ModulePass {
39    bool OnlyDebugInfo;
40  public:
41    static char ID; // Pass identification, replacement for typeid
42    explicit StripSymbols(bool ODI = false)
43      : ModulePass(ID), OnlyDebugInfo(ODI) {
44        initializeStripSymbolsPass(*PassRegistry::getPassRegistry());
45      }
46
47    virtual bool runOnModule(Module &M);
48
49    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
50      AU.setPreservesAll();
51    }
52  };
53
54  class StripNonDebugSymbols : public ModulePass {
55  public:
56    static char ID; // Pass identification, replacement for typeid
57    explicit StripNonDebugSymbols()
58      : ModulePass(ID) {
59        initializeStripNonDebugSymbolsPass(*PassRegistry::getPassRegistry());
60      }
61
62    virtual bool runOnModule(Module &M);
63
64    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
65      AU.setPreservesAll();
66    }
67  };
68
69  class StripDebugDeclare : public ModulePass {
70  public:
71    static char ID; // Pass identification, replacement for typeid
72    explicit StripDebugDeclare()
73      : ModulePass(ID) {
74        initializeStripDebugDeclarePass(*PassRegistry::getPassRegistry());
75      }
76
77    virtual bool runOnModule(Module &M);
78
79    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
80      AU.setPreservesAll();
81    }
82  };
83
84  class StripDeadDebugInfo : public ModulePass {
85  public:
86    static char ID; // Pass identification, replacement for typeid
87    explicit StripDeadDebugInfo()
88      : ModulePass(ID) {
89        initializeStripDeadDebugInfoPass(*PassRegistry::getPassRegistry());
90      }
91
92    virtual bool runOnModule(Module &M);
93
94    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
95      AU.setPreservesAll();
96    }
97  };
98}
99
100char StripSymbols::ID = 0;
101INITIALIZE_PASS(StripSymbols, "strip",
102                "Strip all symbols from a module", false, false)
103
104ModulePass *llvm::createStripSymbolsPass(bool OnlyDebugInfo) {
105  return new StripSymbols(OnlyDebugInfo);
106}
107
108char StripNonDebugSymbols::ID = 0;
109INITIALIZE_PASS(StripNonDebugSymbols, "strip-nondebug",
110                "Strip all symbols, except dbg symbols, from a module",
111                false, false)
112
113ModulePass *llvm::createStripNonDebugSymbolsPass() {
114  return new StripNonDebugSymbols();
115}
116
117char StripDebugDeclare::ID = 0;
118INITIALIZE_PASS(StripDebugDeclare, "strip-debug-declare",
119                "Strip all llvm.dbg.declare intrinsics", false, false)
120
121ModulePass *llvm::createStripDebugDeclarePass() {
122  return new StripDebugDeclare();
123}
124
125char StripDeadDebugInfo::ID = 0;
126INITIALIZE_PASS(StripDeadDebugInfo, "strip-dead-debug-info",
127                "Strip debug info for unused symbols", false, false)
128
129ModulePass *llvm::createStripDeadDebugInfoPass() {
130  return new StripDeadDebugInfo();
131}
132
133/// OnlyUsedBy - Return true if V is only used by Usr.
134static bool OnlyUsedBy(Value *V, Value *Usr) {
135  for(Value::use_iterator I = V->use_begin(), E = V->use_end(); I != E; ++I) {
136    User *U = *I;
137    if (U != Usr)
138      return false;
139  }
140  return true;
141}
142
143static void RemoveDeadConstant(Constant *C) {
144  assert(C->use_empty() && "Constant is not dead!");
145  SmallPtrSet<Constant*, 4> Operands;
146  for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i)
147    if (OnlyUsedBy(C->getOperand(i), C))
148      Operands.insert(cast<Constant>(C->getOperand(i)));
149  if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
150    if (!GV->hasLocalLinkage()) return;   // Don't delete non static globals.
151    GV->eraseFromParent();
152  }
153  else if (!isa<Function>(C))
154    if (isa<CompositeType>(C->getType()))
155      C->destroyConstant();
156
157  // If the constant referenced anything, see if we can delete it as well.
158  for (SmallPtrSet<Constant*, 4>::iterator OI = Operands.begin(),
159         OE = Operands.end(); OI != OE; ++OI)
160    RemoveDeadConstant(*OI);
161}
162
163// Strip the symbol table of its names.
164//
165static void StripSymtab(ValueSymbolTable &ST, bool PreserveDbgInfo) {
166  for (ValueSymbolTable::iterator VI = ST.begin(), VE = ST.end(); VI != VE; ) {
167    Value *V = VI->getValue();
168    ++VI;
169    if (!isa<GlobalValue>(V) || cast<GlobalValue>(V)->hasLocalLinkage()) {
170      if (!PreserveDbgInfo || !V->getName().startswith("llvm.dbg"))
171        // Set name to "", removing from symbol table!
172        V->setName("");
173    }
174  }
175}
176
177// Strip any named types of their names.
178static void StripTypeNames(Module &M, bool PreserveDbgInfo) {
179  TypeFinder StructTypes;
180  StructTypes.run(M, false);
181
182  for (unsigned i = 0, e = StructTypes.size(); i != e; ++i) {
183    StructType *STy = StructTypes[i];
184    if (STy->isLiteral() || STy->getName().empty()) continue;
185
186    if (PreserveDbgInfo && STy->getName().startswith("llvm.dbg"))
187      continue;
188
189    STy->setName("");
190  }
191}
192
193/// Find values that are marked as llvm.used.
194static void findUsedValues(GlobalVariable *LLVMUsed,
195                           SmallPtrSet<const GlobalValue*, 8> &UsedValues) {
196  if (LLVMUsed == 0) return;
197  UsedValues.insert(LLVMUsed);
198
199  ConstantArray *Inits = cast<ConstantArray>(LLVMUsed->getInitializer());
200
201  for (unsigned i = 0, e = Inits->getNumOperands(); i != e; ++i)
202    if (GlobalValue *GV =
203          dyn_cast<GlobalValue>(Inits->getOperand(i)->stripPointerCasts()))
204      UsedValues.insert(GV);
205}
206
207/// StripSymbolNames - Strip symbol names.
208static bool StripSymbolNames(Module &M, bool PreserveDbgInfo) {
209
210  SmallPtrSet<const GlobalValue*, 8> llvmUsedValues;
211  findUsedValues(M.getGlobalVariable("llvm.used"), llvmUsedValues);
212  findUsedValues(M.getGlobalVariable("llvm.compiler.used"), llvmUsedValues);
213
214  for (Module::global_iterator I = M.global_begin(), E = M.global_end();
215       I != E; ++I) {
216    if (I->hasLocalLinkage() && llvmUsedValues.count(I) == 0)
217      if (!PreserveDbgInfo || !I->getName().startswith("llvm.dbg"))
218        I->setName("");     // Internal symbols can't participate in linkage
219  }
220
221  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I) {
222    if (I->hasLocalLinkage() && llvmUsedValues.count(I) == 0)
223      if (!PreserveDbgInfo || !I->getName().startswith("llvm.dbg"))
224        I->setName("");     // Internal symbols can't participate in linkage
225    StripSymtab(I->getValueSymbolTable(), PreserveDbgInfo);
226  }
227
228  // Remove all names from types.
229  StripTypeNames(M, PreserveDbgInfo);
230
231  return true;
232}
233
234// StripDebugInfo - Strip debug info in the module if it exists.
235// To do this, we remove llvm.dbg.func.start, llvm.dbg.stoppoint, and
236// llvm.dbg.region.end calls, and any globals they point to if now dead.
237static bool StripDebugInfo(Module &M) {
238
239  bool Changed = false;
240
241  // Remove all of the calls to the debugger intrinsics, and remove them from
242  // the module.
243  if (Function *Declare = M.getFunction("llvm.dbg.declare")) {
244    while (!Declare->use_empty()) {
245      CallInst *CI = cast<CallInst>(Declare->use_back());
246      CI->eraseFromParent();
247    }
248    Declare->eraseFromParent();
249    Changed = true;
250  }
251
252  if (Function *DbgVal = M.getFunction("llvm.dbg.value")) {
253    while (!DbgVal->use_empty()) {
254      CallInst *CI = cast<CallInst>(DbgVal->use_back());
255      CI->eraseFromParent();
256    }
257    DbgVal->eraseFromParent();
258    Changed = true;
259  }
260
261  for (Module::named_metadata_iterator NMI = M.named_metadata_begin(),
262         NME = M.named_metadata_end(); NMI != NME;) {
263    NamedMDNode *NMD = NMI;
264    ++NMI;
265    if (NMD->getName().startswith("llvm.dbg.")) {
266      NMD->eraseFromParent();
267      Changed = true;
268    }
269  }
270
271  for (Module::iterator MI = M.begin(), ME = M.end(); MI != ME; ++MI)
272    for (Function::iterator FI = MI->begin(), FE = MI->end(); FI != FE;
273         ++FI)
274      for (BasicBlock::iterator BI = FI->begin(), BE = FI->end(); BI != BE;
275           ++BI) {
276        if (!BI->getDebugLoc().isUnknown()) {
277          Changed = true;
278          BI->setDebugLoc(DebugLoc());
279        }
280      }
281
282  return Changed;
283}
284
285bool StripSymbols::runOnModule(Module &M) {
286  bool Changed = false;
287  Changed |= StripDebugInfo(M);
288  if (!OnlyDebugInfo)
289    Changed |= StripSymbolNames(M, false);
290  return Changed;
291}
292
293bool StripNonDebugSymbols::runOnModule(Module &M) {
294  return StripSymbolNames(M, true);
295}
296
297bool StripDebugDeclare::runOnModule(Module &M) {
298
299  Function *Declare = M.getFunction("llvm.dbg.declare");
300  std::vector<Constant*> DeadConstants;
301
302  if (Declare) {
303    while (!Declare->use_empty()) {
304      CallInst *CI = cast<CallInst>(Declare->use_back());
305      Value *Arg1 = CI->getArgOperand(0);
306      Value *Arg2 = CI->getArgOperand(1);
307      assert(CI->use_empty() && "llvm.dbg intrinsic should have void result");
308      CI->eraseFromParent();
309      if (Arg1->use_empty()) {
310        if (Constant *C = dyn_cast<Constant>(Arg1))
311          DeadConstants.push_back(C);
312        else
313          RecursivelyDeleteTriviallyDeadInstructions(Arg1);
314      }
315      if (Arg2->use_empty())
316        if (Constant *C = dyn_cast<Constant>(Arg2))
317          DeadConstants.push_back(C);
318    }
319    Declare->eraseFromParent();
320  }
321
322  while (!DeadConstants.empty()) {
323    Constant *C = DeadConstants.back();
324    DeadConstants.pop_back();
325    if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
326      if (GV->hasLocalLinkage())
327        RemoveDeadConstant(GV);
328    } else
329      RemoveDeadConstant(C);
330  }
331
332  return true;
333}
334
335bool StripDeadDebugInfo::runOnModule(Module &M) {
336  bool Changed = false;
337
338  // Debugging information is encoded in llvm IR using metadata. This is designed
339  // such a way that debug info for symbols preserved even if symbols are
340  // optimized away by the optimizer. This special pass removes debug info for
341  // such symbols.
342
343  // llvm.dbg.gv keeps track of debug info for global variables.
344  if (NamedMDNode *NMD = M.getNamedMetadata("llvm.dbg.gv")) {
345    SmallVector<MDNode *, 8> MDs;
346    for (unsigned i = 0, e = NMD->getNumOperands(); i != e; ++i)
347      if (NMD->getOperand(i)) {
348        assert(DIGlobalVariable(NMD->getOperand(i)).isGlobalVariable() &&
349          "A MDNode in llvm.dbg.gv should be a DIGlobalVariable.");
350        MDs.push_back(NMD->getOperand(i));
351      }
352      else
353        Changed = true;
354    NMD->eraseFromParent();
355    NMD = NULL;
356
357    for (SmallVectorImpl<MDNode *>::iterator I = MDs.begin(),
358           E = MDs.end(); I != E; ++I) {
359      GlobalVariable *GV = DIGlobalVariable(*I).getGlobal();
360      if (GV && M.getGlobalVariable(GV->getName(), true)) {
361        if (!NMD)
362          NMD = M.getOrInsertNamedMetadata("llvm.dbg.gv");
363        NMD->addOperand(*I);
364      }
365      else
366        Changed = true;
367    }
368  }
369
370  // llvm.dbg.sp keeps track of debug info for subprograms.
371  if (NamedMDNode *NMD = M.getNamedMetadata("llvm.dbg.sp")) {
372    SmallVector<MDNode *, 8> MDs;
373    for (unsigned i = 0, e = NMD->getNumOperands(); i != e; ++i)
374      if (NMD->getOperand(i)) {
375        assert(DISubprogram(NMD->getOperand(i)).isSubprogram() &&
376          "A MDNode in llvm.dbg.sp should be a DISubprogram.");
377        MDs.push_back(NMD->getOperand(i));
378      }
379      else
380        Changed = true;
381    NMD->eraseFromParent();
382    NMD = NULL;
383
384    for (SmallVectorImpl<MDNode *>::iterator I = MDs.begin(),
385           E = MDs.end(); I != E; ++I) {
386      bool FnIsLive = false;
387      if (Function *F = DISubprogram(*I).getFunction())
388        if (M.getFunction(F->getName()))
389          FnIsLive = true;
390      if (FnIsLive) {
391          if (!NMD)
392            NMD = M.getOrInsertNamedMetadata("llvm.dbg.sp");
393          NMD->addOperand(*I);
394      } else {
395        // Remove llvm.dbg.lv.fnname named mdnode which may have been used
396        // to hold debug info for dead function's local variables.
397        StringRef FName = DISubprogram(*I).getLinkageName();
398        if (FName.empty())
399          FName = DISubprogram(*I).getName();
400        if (NamedMDNode *LVNMD = M.getNamedMetadata(
401                "llvm.dbg.lv." + Function::getRealLinkageName(FName)))
402          LVNMD->eraseFromParent();
403      }
404    }
405  }
406
407  return Changed;
408}
409