StripSymbols.cpp revision 210299
1205407Srdivacky//===- StripSymbols.cpp - Strip symbols and debug info from a module ------===//
2205407Srdivacky//
3205407Srdivacky//                     The LLVM Compiler Infrastructure
4205407Srdivacky//
5205407Srdivacky// This file is distributed under the University of Illinois Open Source
6205407Srdivacky// License. See LICENSE.TXT for details.
7205407Srdivacky//
8205407Srdivacky//===----------------------------------------------------------------------===//
9205407Srdivacky//
10218893Sdim// The StripSymbols transformation implements code stripping. Specifically, it
11205407Srdivacky// can delete:
12205407Srdivacky//
13205407Srdivacky//   * names for virtual registers
14226633Sdim//   * symbols for internal globals and functions
15206083Srdivacky//   * debug information
16205407Srdivacky//
17234353Sdim// Note that this transformation makes code much less readable, so it should
18205407Srdivacky// only be used in situations where the 'strip' utility would be used, such as
19205407Srdivacky// reducing code size or making it harder to reverse engineer code.
20205407Srdivacky//
21208599Srdivacky//===----------------------------------------------------------------------===//
22205407Srdivacky
23218893Sdim#include "llvm/Transforms/IPO.h"
24239462Sdim#include "llvm/Constants.h"
25205407Srdivacky#include "llvm/DerivedTypes.h"
26205407Srdivacky#include "llvm/Instructions.h"
27205407Srdivacky#include "llvm/Module.h"
28205407Srdivacky#include "llvm/Pass.h"
29218893Sdim#include "llvm/Analysis/DebugInfo.h"
30205407Srdivacky#include "llvm/ValueSymbolTable.h"
31224145Sdim#include "llvm/TypeSymbolTable.h"
32224145Sdim#include "llvm/Transforms/Utils/Local.h"
33208599Srdivacky#include "llvm/ADT/SmallPtrSet.h"
34208599Srdivackyusing namespace llvm;
35208599Srdivacky
36208599Srdivackynamespace {
37208599Srdivacky  class StripSymbols : public ModulePass {
38208599Srdivacky    bool OnlyDebugInfo;
39208599Srdivacky  public:
40208599Srdivacky    static char ID; // Pass identification, replacement for typeid
41208599Srdivacky    explicit StripSymbols(bool ODI = false)
42208599Srdivacky      : ModulePass(&ID), OnlyDebugInfo(ODI) {}
43208599Srdivacky
44208599Srdivacky    virtual bool runOnModule(Module &M);
45208599Srdivacky
46224145Sdim    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
47224145Sdim      AU.setPreservesAll();
48224145Sdim    }
49224145Sdim  };
50224145Sdim
51205407Srdivacky  class StripNonDebugSymbols : public ModulePass {
52224145Sdim  public:
53224145Sdim    static char ID; // Pass identification, replacement for typeid
54224145Sdim    explicit StripNonDebugSymbols()
55205407Srdivacky      : ModulePass(&ID) {}
56224145Sdim
57224145Sdim    virtual bool runOnModule(Module &M);
58205407Srdivacky
59224145Sdim    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
60224145Sdim      AU.setPreservesAll();
61224145Sdim    }
62224145Sdim  };
63224145Sdim
64218893Sdim  class StripDebugDeclare : public ModulePass {
65224145Sdim  public:
66224145Sdim    static char ID; // Pass identification, replacement for typeid
67224145Sdim    explicit StripDebugDeclare()
68205407Srdivacky      : ModulePass(&ID) {}
69224145Sdim
70224145Sdim    virtual bool runOnModule(Module &M);
71243830Sdim
72243830Sdim    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
73243830Sdim      AU.setPreservesAll();
74243830Sdim    }
75243830Sdim  };
76224145Sdim
77224145Sdim  class StripDeadDebugInfo : public ModulePass {
78224145Sdim  public:
79224145Sdim    static char ID; // Pass identification, replacement for typeid
80205407Srdivacky    explicit StripDeadDebugInfo()
81224145Sdim      : ModulePass(&ID) {}
82224145Sdim
83224145Sdim    virtual bool runOnModule(Module &M);
84224145Sdim
85224145Sdim    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
86224145Sdim      AU.setPreservesAll();
87224145Sdim    }
88205407Srdivacky  };
89224145Sdim}
90224145Sdim
91224145Sdimchar StripSymbols::ID = 0;
92224145Sdimstatic RegisterPass<StripSymbols>
93224145SdimX("strip", "Strip all symbols from a module");
94224145Sdim
95224145SdimModulePass *llvm::createStripSymbolsPass(bool OnlyDebugInfo) {
96224145Sdim  return new StripSymbols(OnlyDebugInfo);
97218893Sdim}
98205407Srdivacky
99224145Sdimchar StripNonDebugSymbols::ID = 0;
100224145Sdimstatic RegisterPass<StripNonDebugSymbols>
101224145SdimY("strip-nondebug", "Strip all symbols, except dbg symbols, from a module");
102218893Sdim
103224145SdimModulePass *llvm::createStripNonDebugSymbolsPass() {
104224145Sdim  return new StripNonDebugSymbols();
105224145Sdim}
106224145Sdim
107224145Sdimchar StripDebugDeclare::ID = 0;
108224145Sdimstatic RegisterPass<StripDebugDeclare>
109221345SdimZ("strip-debug-declare", "Strip all llvm.dbg.declare intrinsics");
110224145Sdim
111224145SdimModulePass *llvm::createStripDebugDeclarePass() {
112224145Sdim  return new StripDebugDeclare();
113224145Sdim}
114224145Sdim
115221345Sdimchar StripDeadDebugInfo::ID = 0;
116224145Sdimstatic RegisterPass<StripDeadDebugInfo>
117224145SdimA("strip-dead-debug-info", "Strip debug info for unused symbols");
118224145Sdim
119224145SdimModulePass *llvm::createStripDeadDebugInfoPass() {
120221345Sdim  return new StripDeadDebugInfo();
121224145Sdim}
122224145Sdim
123221345Sdim/// OnlyUsedBy - Return true if V is only used by Usr.
124224145Sdimstatic bool OnlyUsedBy(Value *V, Value *Usr) {
125224145Sdim  for(Value::use_iterator I = V->use_begin(), E = V->use_end(); I != E; ++I) {
126218893Sdim    User *U = *I;
127224145Sdim    if (U != Usr)
128224145Sdim      return false;
129218893Sdim  }
130224145Sdim  return true;
131218893Sdim}
132224145Sdim
133224145Sdimstatic void RemoveDeadConstant(Constant *C) {
134205407Srdivacky  assert(C->use_empty() && "Constant is not dead!");
135224145Sdim  SmallPtrSet<Constant*, 4> Operands;
136224145Sdim  for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i)
137224145Sdim    if (isa<DerivedType>(C->getOperand(i)->getType()) &&
138224145Sdim        OnlyUsedBy(C->getOperand(i), C))
139224145Sdim      Operands.insert(cast<Constant>(C->getOperand(i)));
140224145Sdim  if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
141218893Sdim    if (!GV->hasLocalLinkage()) return;   // Don't delete non static globals.
142224145Sdim    GV->eraseFromParent();
143224145Sdim  }
144224145Sdim  else if (!isa<Function>(C))
145205407Srdivacky    if (isa<CompositeType>(C->getType()))
146224145Sdim      C->destroyConstant();
147224145Sdim
148243830Sdim  // If the constant referenced anything, see if we can delete it as well.
149243830Sdim  for (SmallPtrSet<Constant*, 4>::iterator OI = Operands.begin(),
150224145Sdim         OE = Operands.end(); OI != OE; ++OI)
151224145Sdim    RemoveDeadConstant(*OI);
152224145Sdim}
153224145Sdim
154224145Sdim// Strip the symbol table of its names.
155224145Sdim//
156218893Sdimstatic void StripSymtab(ValueSymbolTable &ST, bool PreserveDbgInfo) {
157224145Sdim  for (ValueSymbolTable::iterator VI = ST.begin(), VE = ST.end(); VI != VE; ) {
158224145Sdim    Value *V = VI->getValue();
159205407Srdivacky    ++VI;
160224145Sdim    if (!isa<GlobalValue>(V) || cast<GlobalValue>(V)->hasLocalLinkage()) {
161224145Sdim      if (!PreserveDbgInfo || !V->getName().startswith("llvm.dbg"))
162224145Sdim        // Set name to "", removing from symbol table!
163224145Sdim        V->setName("");
164224145Sdim    }
165224145Sdim  }
166224145Sdim}
167205407Srdivacky
168224145Sdim// Strip the symbol table of its names.
169224145Sdimstatic void StripTypeSymtab(TypeSymbolTable &ST, bool PreserveDbgInfo) {
170224145Sdim  for (TypeSymbolTable::iterator TI = ST.begin(), E = ST.end(); TI != E; ) {
171224145Sdim    if (PreserveDbgInfo && StringRef(TI->first).startswith("llvm.dbg"))
172224145Sdim      ++TI;
173224145Sdim    else
174224145Sdim      ST.remove(TI++);
175224145Sdim  }
176224145Sdim}
177224145Sdim
178224145Sdim/// Find values that are marked as llvm.used.
179205407Srdivackystatic void findUsedValues(GlobalVariable *LLVMUsed,
180224145Sdim                           SmallPtrSet<const GlobalValue*, 8> &UsedValues) {
181224145Sdim  if (LLVMUsed == 0) return;
182224145Sdim  UsedValues.insert(LLVMUsed);
183224145Sdim
184205407Srdivacky  ConstantArray *Inits = dyn_cast<ConstantArray>(LLVMUsed->getInitializer());
185224145Sdim  if (Inits == 0) return;
186224145Sdim
187205407Srdivacky  for (unsigned i = 0, e = Inits->getNumOperands(); i != e; ++i)
188224145Sdim    if (GlobalValue *GV =
189224145Sdim          dyn_cast<GlobalValue>(Inits->getOperand(i)->stripPointerCasts()))
190224145Sdim      UsedValues.insert(GV);
191224145Sdim}
192224145Sdim
193224145Sdim/// StripSymbolNames - Strip symbol names.
194224145Sdimstatic bool StripSymbolNames(Module &M, bool PreserveDbgInfo) {
195205407Srdivacky
196224145Sdim  SmallPtrSet<const GlobalValue*, 8> llvmUsedValues;
197224145Sdim  findUsedValues(M.getGlobalVariable("llvm.used"), llvmUsedValues);
198224145Sdim  findUsedValues(M.getGlobalVariable("llvm.compiler.used"), llvmUsedValues);
199224145Sdim
200205407Srdivacky  for (Module::global_iterator I = M.global_begin(), E = M.global_end();
201205407Srdivacky       I != E; ++I) {
202224145Sdim    if (I->hasLocalLinkage() && llvmUsedValues.count(I) == 0)
203224145Sdim      if (!PreserveDbgInfo || !I->getName().startswith("llvm.dbg"))
204206083Srdivacky        I->setName("");     // Internal symbols can't participate in linkage
205224145Sdim  }
206224145Sdim
207205407Srdivacky  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I) {
208224145Sdim    if (I->hasLocalLinkage() && llvmUsedValues.count(I) == 0)
209224145Sdim      if (!PreserveDbgInfo || !I->getName().startswith("llvm.dbg"))
210224145Sdim        I->setName("");     // Internal symbols can't participate in linkage
211224145Sdim    StripSymtab(I->getValueSymbolTable(), PreserveDbgInfo);
212224145Sdim  }
213224145Sdim
214224145Sdim  // Remove all names from types.
215224145Sdim  StripTypeSymtab(M.getTypeSymbolTable(), PreserveDbgInfo);
216224145Sdim
217205407Srdivacky  return true;
218224145Sdim}
219205407Srdivacky
220224145Sdim// StripDebugInfo - Strip debug info in the module if it exists.
221224145Sdim// To do this, we remove llvm.dbg.func.start, llvm.dbg.stoppoint, and
222224145Sdim// llvm.dbg.region.end calls, and any globals they point to if now dead.
223205407Srdivackystatic bool StripDebugInfo(Module &M) {
224224145Sdim
225224145Sdim  bool Changed = false;
226224145Sdim
227224145Sdim  // Remove all of the calls to the debugger intrinsics, and remove them from
228224145Sdim  // the module.
229224145Sdim  if (Function *Declare = M.getFunction("llvm.dbg.declare")) {
230224145Sdim    while (!Declare->use_empty()) {
231224145Sdim      CallInst *CI = cast<CallInst>(Declare->use_back());
232224145Sdim      CI->eraseFromParent();
233205407Srdivacky    }
234224145Sdim    Declare->eraseFromParent();
235224145Sdim    Changed = true;
236224145Sdim  }
237205407Srdivacky
238224145Sdim  if (Function *DbgVal = M.getFunction("llvm.dbg.value")) {
239224145Sdim    while (!DbgVal->use_empty()) {
240224145Sdim      CallInst *CI = cast<CallInst>(DbgVal->use_back());
241224145Sdim      CI->eraseFromParent();
242224145Sdim    }
243205407Srdivacky    DbgVal->eraseFromParent();
244224145Sdim    Changed = true;
245224145Sdim  }
246205407Srdivacky
247224145Sdim  for (Module::named_metadata_iterator NMI = M.named_metadata_begin(),
248224145Sdim         NME = M.named_metadata_end(); NMI != NME;) {
249224145Sdim    NamedMDNode *NMD = NMI;
250224145Sdim    ++NMI;
251224145Sdim    if (NMD->getName().startswith("llvm.dbg.")) {
252224145Sdim      NMD->eraseFromParent();
253205407Srdivacky      Changed = true;
254224145Sdim    }
255224145Sdim  }
256205407Srdivacky
257224145Sdim  unsigned MDDbgKind = M.getMDKindID("dbg");
258224145Sdim  for (Module::iterator MI = M.begin(), ME = M.end(); MI != ME; ++MI)
259224145Sdim    for (Function::iterator FI = MI->begin(), FE = MI->end(); FI != FE;
260224145Sdim         ++FI)
261224145Sdim      for (BasicBlock::iterator BI = FI->begin(), BE = FI->end(); BI != BE;
262224145Sdim           ++BI) {
263224145Sdim        Changed = true; // FIXME: Only set if there was debug metadata.
264224145Sdim        BI->setMetadata(MDDbgKind, 0);
265224145Sdim      }
266205407Srdivacky
267224145Sdim  return Changed;
268224145Sdim}
269205407Srdivacky
270224145Sdimbool StripSymbols::runOnModule(Module &M) {
271224145Sdim  bool Changed = false;
272224145Sdim  Changed |= StripDebugInfo(M);
273224145Sdim  if (!OnlyDebugInfo)
274224145Sdim    Changed |= StripSymbolNames(M, false);
275224145Sdim  return Changed;
276224145Sdim}
277224145Sdim
278224145Sdimbool StripNonDebugSymbols::runOnModule(Module &M) {
279224145Sdim  return StripSymbolNames(M, true);
280224145Sdim}
281224145Sdim
282224145Sdimbool StripDebugDeclare::runOnModule(Module &M) {
283224145Sdim
284224145Sdim  Function *Declare = M.getFunction("llvm.dbg.declare");
285224145Sdim  std::vector<Constant*> DeadConstants;
286224145Sdim
287224145Sdim  if (Declare) {
288224145Sdim    while (!Declare->use_empty()) {
289224145Sdim      CallInst *CI = cast<CallInst>(Declare->use_back());
290205407Srdivacky      Value *Arg1 = CI->getArgOperand(0);
291224145Sdim      Value *Arg2 = CI->getArgOperand(1);
292224145Sdim      assert(CI->use_empty() && "llvm.dbg intrinsic should have void result");
293205407Srdivacky      CI->eraseFromParent();
294224145Sdim      if (Arg1->use_empty()) {
295224145Sdim        if (Constant *C = dyn_cast<Constant>(Arg1))
296224145Sdim          DeadConstants.push_back(C);
297224145Sdim        else
298224145Sdim          RecursivelyDeleteTriviallyDeadInstructions(Arg1);
299224145Sdim      }
300226633Sdim      if (Arg2->use_empty())
301205407Srdivacky        if (Constant *C = dyn_cast<Constant>(Arg2))
302224145Sdim          DeadConstants.push_back(C);
303206083Srdivacky    }
304224145Sdim    Declare->eraseFromParent();
305224145Sdim  }
306224145Sdim
307224145Sdim  while (!DeadConstants.empty()) {
308224145Sdim    Constant *C = DeadConstants.back();
309224145Sdim    DeadConstants.pop_back();
310224145Sdim    if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
311206083Srdivacky      if (GV->hasLocalLinkage())
312224145Sdim        RemoveDeadConstant(GV);
313205407Srdivacky    } else
314224145Sdim      RemoveDeadConstant(C);
315224145Sdim  }
316205407Srdivacky
317224145Sdim  return true;
318224145Sdim}
319224145Sdim
320205407Srdivacky/// getRealLinkageName - If special LLVM prefix that is used to inform the asm
321224145Sdim/// printer to not emit usual symbol prefix before the symbol name is used then
322224145Sdim/// return linkage name after skipping this special LLVM prefix.
323243830Sdimstatic StringRef getRealLinkageName(StringRef LinkageName) {
324224145Sdim  char One = '\1';
325224145Sdim  if (LinkageName.startswith(StringRef(&One, 1)))
326224145Sdim    return LinkageName.substr(1);
327224145Sdim  return LinkageName;
328205407Srdivacky}
329224145Sdim
330224145Sdimbool StripDeadDebugInfo::runOnModule(Module &M) {
331224145Sdim  bool Changed = false;
332224145Sdim
333224145Sdim  // Debugging infomration is encoded in llvm IR using metadata. This is designed
334224145Sdim  // such a way that debug info for symbols preserved even if symbols are
335224145Sdim  // optimized away by the optimizer. This special pass removes debug info for
336224145Sdim  // such symbols.
337224145Sdim
338205407Srdivacky  // llvm.dbg.gv keeps track of debug info for global variables.
339205407Srdivacky  if (NamedMDNode *NMD = M.getNamedMetadata("llvm.dbg.gv")) {
340205407Srdivacky    SmallVector<MDNode *, 8> MDs;
341224145Sdim    for (unsigned i = 0, e = NMD->getNumOperands(); i != e; ++i)
342205407Srdivacky      if (DIGlobalVariable(NMD->getOperand(i)).Verify())
343224145Sdim        MDs.push_back(NMD->getOperand(i));
344224145Sdim      else
345224145Sdim        Changed = true;
346205407Srdivacky    NMD->eraseFromParent();
347224145Sdim    NMD = NULL;
348224145Sdim
349224145Sdim    for (SmallVector<MDNode *, 8>::iterator I = MDs.begin(),
350224145Sdim           E = MDs.end(); I != E; ++I) {
351224145Sdim      if (M.getGlobalVariable(DIGlobalVariable(*I).getGlobal()->getName(),
352224145Sdim                              true)) {
353224145Sdim        if (!NMD)
354224145Sdim          NMD = M.getOrInsertNamedMetadata("llvm.dbg.gv");
355205407Srdivacky        NMD->addOperand(*I);
356239462Sdim      }
357239462Sdim      else
358239462Sdim        Changed = true;
359239462Sdim    }
360239462Sdim  }
361239462Sdim
362239462Sdim  // llvm.dbg.sp keeps track of debug info for subprograms.
363239462Sdim  if (NamedMDNode *NMD = M.getNamedMetadata("llvm.dbg.sp")) {
364239462Sdim    SmallVector<MDNode *, 8> MDs;
365239462Sdim    for (unsigned i = 0, e = NMD->getNumOperands(); i != e; ++i)
366239462Sdim      if (DISubprogram(NMD->getOperand(i)).Verify())
367239462Sdim        MDs.push_back(NMD->getOperand(i));
368239462Sdim      else
369239462Sdim        Changed = true;
370239462Sdim    NMD->eraseFromParent();
371224145Sdim    NMD = NULL;
372224145Sdim
373224145Sdim    for (SmallVector<MDNode *, 8>::iterator I = MDs.begin(),
374224145Sdim           E = MDs.end(); I != E; ++I) {
375224145Sdim      bool FnIsLive = false;
376224145Sdim      if (Function *F = DISubprogram(*I).getFunction())
377224145Sdim        if (M.getFunction(F->getName()))
378224145Sdim          FnIsLive = true;
379224145Sdim      if (FnIsLive) {
380205407Srdivacky          if (!NMD)
381224145Sdim            NMD = M.getOrInsertNamedMetadata("llvm.dbg.sp");
382224145Sdim          NMD->addOperand(*I);
383224145Sdim      } else {
384224145Sdim        // Remove llvm.dbg.lv.fnname named mdnode which may have been used
385224145Sdim        // to hold debug info for dead function's local variables.
386224145Sdim        StringRef FName = DISubprogram(*I).getLinkageName();
387224145Sdim        if (FName.empty())
388205407Srdivacky          FName = DISubprogram(*I).getName();
389224145Sdim        if (NamedMDNode *LVNMD =
390224145Sdim            M.getNamedMetadata(Twine("llvm.dbg.lv.",
391224145Sdim                                     getRealLinkageName(FName))))
392224145Sdim          LVNMD->eraseFromParent();
393224145Sdim      }
394224145Sdim    }
395205407Srdivacky  }
396224145Sdim
397224145Sdim  return Changed;
398205407Srdivacky}
399224145Sdim