StripSymbols.cpp revision 199481 
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/Constants.h"
25#include "llvm/DerivedTypes.h"
26#include "llvm/Instructions.h"
27#include "llvm/LLVMContext.h"
28#include "llvm/Module.h"
29#include "llvm/Pass.h"
30#include "llvm/Analysis/DebugInfo.h"
31#include "llvm/ValueSymbolTable.h"
32#include "llvm/TypeSymbolTable.h"
33#include "llvm/Transforms/Utils/Local.h"
34#include "llvm/ADT/SmallPtrSet.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
45    virtual bool runOnModule(Module &M);
46
47    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
48      AU.setPreservesAll();
49    }
50  };
51
52  class StripNonDebugSymbols : public ModulePass {
53  public:
54    static char ID; // Pass identification, replacement for typeid
55    explicit StripNonDebugSymbols()
56      : ModulePass(&ID) {}
57
58    virtual bool runOnModule(Module &M);
59
60    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
61      AU.setPreservesAll();
62    }
63  };
64
65  class StripDebugDeclare : public ModulePass {
66  public:
67    static char ID; // Pass identification, replacement for typeid
68    explicit StripDebugDeclare()
69      : ModulePass(&ID) {}
70
71    virtual bool runOnModule(Module &M);
72
73    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
74      AU.setPreservesAll();
75    }
76  };
77}
78
79char StripSymbols::ID = 0;
80static RegisterPass<StripSymbols>
81X("strip", "Strip all symbols from a module");
82
83ModulePass *llvm::createStripSymbolsPass(bool OnlyDebugInfo) {
84  return new StripSymbols(OnlyDebugInfo);
85}
86
87char StripNonDebugSymbols::ID = 0;
88static RegisterPass<StripNonDebugSymbols>
89Y("strip-nondebug", "Strip all symbols, except dbg symbols, from a module");
90
91ModulePass *llvm::createStripNonDebugSymbolsPass() {
92  return new StripNonDebugSymbols();
93}
94
95char StripDebugDeclare::ID = 0;
96static RegisterPass<StripDebugDeclare>
97Z("strip-debug-declare", "Strip all llvm.dbg.declare intrinsics");
98
99ModulePass *llvm::createStripDebugDeclarePass() {
100  return new StripDebugDeclare();
101}
102
103/// OnlyUsedBy - Return true if V is only used by Usr.
104static bool OnlyUsedBy(Value *V, Value *Usr) {
105  for(Value::use_iterator I = V->use_begin(), E = V->use_end(); I != E; ++I) {
106    User *U = *I;
107    if (U != Usr)
108      return false;
109  }
110  return true;
111}
112
113static void RemoveDeadConstant(Constant *C) {
114  assert(C->use_empty() && "Constant is not dead!");
115  SmallPtrSet<Constant*, 4> Operands;
116  for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i)
117    if (isa<DerivedType>(C->getOperand(i)->getType()) &&
118        OnlyUsedBy(C->getOperand(i), C))
119      Operands.insert(cast<Constant>(C->getOperand(i)));
120  if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
121    if (!GV->hasLocalLinkage()) return;   // Don't delete non static globals.
122    GV->eraseFromParent();
123  }
124  else if (!isa<Function>(C))
125    if (isa<CompositeType>(C->getType()))
126      C->destroyConstant();
127
128  // If the constant referenced anything, see if we can delete it as well.
129  for (SmallPtrSet<Constant*, 4>::iterator OI = Operands.begin(),
130         OE = Operands.end(); OI != OE; ++OI)
131    RemoveDeadConstant(*OI);
132}
133
134// Strip the symbol table of its names.
135//
136static void StripSymtab(ValueSymbolTable &ST, bool PreserveDbgInfo) {
137  for (ValueSymbolTable::iterator VI = ST.begin(), VE = ST.end(); VI != VE; ) {
138    Value *V = VI->getValue();
139    ++VI;
140    if (!isa<GlobalValue>(V) || cast<GlobalValue>(V)->hasLocalLinkage()) {
141      if (!PreserveDbgInfo || !V->getName().startswith("llvm.dbg"))
142        // Set name to "", removing from symbol table!
143        V->setName("");
144    }
145  }
146}
147
148// Strip the symbol table of its names.
149static void StripTypeSymtab(TypeSymbolTable &ST, bool PreserveDbgInfo) {
150  for (TypeSymbolTable::iterator TI = ST.begin(), E = ST.end(); TI != E; ) {
151    if (PreserveDbgInfo && strncmp(TI->first.c_str(), "llvm.dbg", 8) == 0)
152      ++TI;
153    else
154      ST.remove(TI++);
155  }
156}
157
158/// Find values that are marked as llvm.used.
159static void findUsedValues(GlobalVariable *LLVMUsed,
160                           SmallPtrSet<const GlobalValue*, 8> &UsedValues) {
161  if (LLVMUsed == 0) return;
162  UsedValues.insert(LLVMUsed);
163
164  ConstantArray *Inits = dyn_cast<ConstantArray>(LLVMUsed->getInitializer());
165  if (Inits == 0) return;
166
167  for (unsigned i = 0, e = Inits->getNumOperands(); i != e; ++i)
168    if (GlobalValue *GV =
169          dyn_cast<GlobalValue>(Inits->getOperand(i)->stripPointerCasts()))
170      UsedValues.insert(GV);
171}
172
173/// StripSymbolNames - Strip symbol names.
174static bool StripSymbolNames(Module &M, bool PreserveDbgInfo) {
175
176  SmallPtrSet<const GlobalValue*, 8> llvmUsedValues;
177  findUsedValues(M.getGlobalVariable("llvm.used"), llvmUsedValues);
178  findUsedValues(M.getGlobalVariable("llvm.compiler.used"), llvmUsedValues);
179
180  for (Module::global_iterator I = M.global_begin(), E = M.global_end();
181       I != E; ++I) {
182    if (I->hasLocalLinkage() && llvmUsedValues.count(I) == 0)
183      if (!PreserveDbgInfo || !I->getName().startswith("llvm.dbg"))
184        I->setName("");     // Internal symbols can't participate in linkage
185  }
186
187  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I) {
188    if (I->hasLocalLinkage() && llvmUsedValues.count(I) == 0)
189      if (!PreserveDbgInfo || !I->getName().startswith("llvm.dbg"))
190        I->setName("");     // Internal symbols can't participate in linkage
191    StripSymtab(I->getValueSymbolTable(), PreserveDbgInfo);
192  }
193
194  // Remove all names from types.
195  StripTypeSymtab(M.getTypeSymbolTable(), PreserveDbgInfo);
196
197  return true;
198}
199
200// StripDebugInfo - Strip debug info in the module if it exists.
201// To do this, we remove llvm.dbg.func.start, llvm.dbg.stoppoint, and
202// llvm.dbg.region.end calls, and any globals they point to if now dead.
203static bool StripDebugInfo(Module &M) {
204
205  bool Changed = false;
206
207  // Remove all of the calls to the debugger intrinsics, and remove them from
208  // the module.
209  if (Function *Declare = M.getFunction("llvm.dbg.declare")) {
210    while (!Declare->use_empty()) {
211      CallInst *CI = cast<CallInst>(Declare->use_back());
212      CI->eraseFromParent();
213    }
214    Declare->eraseFromParent();
215    Changed = true;
216  }
217
218  NamedMDNode *NMD = M.getNamedMetadata("llvm.dbg.gv");
219  if (NMD) {
220    Changed = true;
221    NMD->eraseFromParent();
222  }
223  MetadataContext &TheMetadata = M.getContext().getMetadata();
224  unsigned MDDbgKind = TheMetadata.getMDKind("dbg");
225  if (!MDDbgKind)
226    return Changed;
227
228  for (Module::iterator MI = M.begin(), ME = M.end(); MI != ME; ++MI)
229    for (Function::iterator FI = MI->begin(), FE = MI->end(); FI != FE;
230         ++FI)
231      for (BasicBlock::iterator BI = FI->begin(), BE = FI->end(); BI != BE;
232           ++BI)
233        TheMetadata.removeMD(MDDbgKind, BI);
234
235  return true;
236}
237
238bool StripSymbols::runOnModule(Module &M) {
239  bool Changed = false;
240  Changed |= StripDebugInfo(M);
241  if (!OnlyDebugInfo)
242    Changed |= StripSymbolNames(M, false);
243  return Changed;
244}
245
246bool StripNonDebugSymbols::runOnModule(Module &M) {
247  return StripSymbolNames(M, true);
248}
249
250bool StripDebugDeclare::runOnModule(Module &M) {
251
252  Function *Declare = M.getFunction("llvm.dbg.declare");
253  std::vector<Constant*> DeadConstants;
254
255  if (Declare) {
256    while (!Declare->use_empty()) {
257      CallInst *CI = cast<CallInst>(Declare->use_back());
258      Value *Arg1 = CI->getOperand(1);
259      Value *Arg2 = CI->getOperand(2);
260      assert(CI->use_empty() && "llvm.dbg intrinsic should have void result");
261      CI->eraseFromParent();
262      if (Arg1->use_empty()) {
263        if (Constant *C = dyn_cast<Constant>(Arg1))
264          DeadConstants.push_back(C);
265        else
266          RecursivelyDeleteTriviallyDeadInstructions(Arg1);
267      }
268      if (Arg2->use_empty())
269        if (Constant *C = dyn_cast<Constant>(Arg2))
270          DeadConstants.push_back(C);
271    }
272    Declare->eraseFromParent();
273  }
274
275  while (!DeadConstants.empty()) {
276    Constant *C = DeadConstants.back();
277    DeadConstants.pop_back();
278    if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
279      if (GV->hasLocalLinkage())
280        RemoveDeadConstant(GV);
281    } else
282      RemoveDeadConstant(C);
283  }
284
285  return true;
286}
287