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
234bool StripSymbols::runOnModule(Module &M) {
235  bool Changed = false;
236  Changed |= StripDebugInfo(M);
237  if (!OnlyDebugInfo)
238    Changed |= StripSymbolNames(M, false);
239  return Changed;
240}
241
242bool StripNonDebugSymbols::runOnModule(Module &M) {
243  return StripSymbolNames(M, true);
244}
245
246bool StripDebugDeclare::runOnModule(Module &M) {
247
248  Function *Declare = M.getFunction("llvm.dbg.declare");
249  std::vector<Constant*> DeadConstants;
250
251  if (Declare) {
252    while (!Declare->use_empty()) {
253      CallInst *CI = cast<CallInst>(Declare->use_back());
254      Value *Arg1 = CI->getArgOperand(0);
255      Value *Arg2 = CI->getArgOperand(1);
256      assert(CI->use_empty() && "llvm.dbg intrinsic should have void result");
257      CI->eraseFromParent();
258      if (Arg1->use_empty()) {
259        if (Constant *C = dyn_cast<Constant>(Arg1))
260          DeadConstants.push_back(C);
261        else
262          RecursivelyDeleteTriviallyDeadInstructions(Arg1);
263      }
264      if (Arg2->use_empty())
265        if (Constant *C = dyn_cast<Constant>(Arg2))
266          DeadConstants.push_back(C);
267    }
268    Declare->eraseFromParent();
269  }
270
271  while (!DeadConstants.empty()) {
272    Constant *C = DeadConstants.back();
273    DeadConstants.pop_back();
274    if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
275      if (GV->hasLocalLinkage())
276        RemoveDeadConstant(GV);
277    } else
278      RemoveDeadConstant(C);
279  }
280
281  return true;
282}
283
284/// Remove any debug info for global variables/functions in the given module for
285/// which said global variable/function no longer exists (i.e. is null).
286///
287/// Debugging information is encoded in llvm IR using metadata. This is designed
288/// such a way that debug info for symbols preserved even if symbols are
289/// optimized away by the optimizer. This special pass removes debug info for
290/// such symbols.
291bool StripDeadDebugInfo::runOnModule(Module &M) {
292  bool Changed = false;
293
294  LLVMContext &C = M.getContext();
295
296  // Find all debug info in F. This is actually overkill in terms of what we
297  // want to do, but we want to try and be as resilient as possible in the face
298  // of potential debug info changes by using the formal interfaces given to us
299  // as much as possible.
300  DebugInfoFinder F;
301  F.processModule(M);
302
303  // For each compile unit, find the live set of global variables/functions and
304  // replace the current list of potentially dead global variables/functions
305  // with the live list.
306  SmallVector<Value *, 64> LiveGlobalVariables;
307  SmallVector<Value *, 64> LiveSubprograms;
308  DenseSet<const MDNode *> VisitedSet;
309
310  for (DebugInfoFinder::iterator CI = F.compile_unit_begin(),
311         CE = F.compile_unit_end(); CI != CE; ++CI) {
312    // Create our compile unit.
313    DICompileUnit DIC(*CI);
314    assert(DIC.Verify() && "DIC must verify as a DICompileUnit.");
315
316    // Create our live subprogram list.
317    DIArray SPs = DIC.getSubprograms();
318    bool SubprogramChange = false;
319    for (unsigned i = 0, e = SPs.getNumElements(); i != e; ++i) {
320      DISubprogram DISP(SPs.getElement(i));
321      assert(DISP.Verify() && "DISP must verify as a DISubprogram.");
322
323      // Make sure we visit each subprogram only once.
324      if (!VisitedSet.insert(DISP).second)
325        continue;
326
327      // If the function referenced by DISP is not null, the function is live.
328      if (DISP.getFunction())
329        LiveSubprograms.push_back(DISP);
330      else
331        SubprogramChange = true;
332    }
333
334    // Create our live global variable list.
335    DIArray GVs = DIC.getGlobalVariables();
336    bool GlobalVariableChange = false;
337    for (unsigned i = 0, e = GVs.getNumElements(); i != e; ++i) {
338      DIGlobalVariable DIG(GVs.getElement(i));
339      assert(DIG.Verify() && "DIG must verify as DIGlobalVariable.");
340
341      // Make sure we only visit each global variable only once.
342      if (!VisitedSet.insert(DIG).second)
343        continue;
344
345      // If the global variable referenced by DIG is not null, the global
346      // variable is live.
347      if (DIG.getGlobal())
348        LiveGlobalVariables.push_back(DIG);
349      else
350        GlobalVariableChange = true;
351    }
352
353    // If we found dead subprograms or global variables, replace the current
354    // subprogram list/global variable list with our new live subprogram/global
355    // variable list.
356    if (SubprogramChange) {
357      // Make sure that 9 is still the index of the subprograms. This is to make
358      // sure that an assert is hit if the location of the subprogram array
359      // changes. This is just to make sure that this is updated if such an
360      // event occurs.
361      assert(DIC->getNumOperands() >= 10 &&
362             SPs == DIC->getOperand(9) &&
363             "DICompileUnits is expected to store Subprograms in operand "
364             "9.");
365      DIC->replaceOperandWith(9, MDNode::get(C, LiveSubprograms));
366      Changed = true;
367    }
368
369    if (GlobalVariableChange) {
370      // Make sure that 10 is still the index of global variables. This is to
371      // make sure that an assert is hit if the location of the subprogram array
372      // changes. This is just to make sure that this index is updated if such
373      // an event occurs.
374      assert(DIC->getNumOperands() >= 11 &&
375             GVs == DIC->getOperand(10) &&
376             "DICompileUnits is expected to store Global Variables in operand "
377             "10.");
378      DIC->replaceOperandWith(10, MDNode::get(C, LiveGlobalVariables));
379      Changed = true;
380    }
381
382    // Reset lists for the next iteration.
383    LiveSubprograms.clear();
384    LiveGlobalVariables.clear();
385  }
386
387  return Changed;
388}
389