lib/Transforms/IPO/StripSymbols.cpp

   1 //===- StripSymbols.cpp - Strip symbols and debug info from a module ------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // The StripSymbols transformation implements code stripping. Specifically, it
  10 // can delete:
  11 //
  12 //   * names for virtual registers
  13 //   * symbols for internal globals and functions
  14 //   * debug information
  15 //
  16 // Note that this transformation makes code much less readable, so it should
  17 // only be used in situations where the 'strip' utility would be used, such as
  18 // reducing code size or making it harder to reverse engineer code.
  19 //
  20 //===----------------------------------------------------------------------===//
  21
  22 #include "llvm/ADT/SmallPtrSet.h"
  23 #include "llvm/Transforms/Utils/Local.h"
  24 #include "llvm/IR/Constants.h"
  25 #include "llvm/IR/DebugInfo.h"
  26 #include "llvm/IR/DerivedTypes.h"
  27 #include "llvm/IR/Instructions.h"
  28 #include "llvm/IR/Module.h"
  29 #include "llvm/IR/TypeFinder.h"
  30 #include "llvm/IR/ValueSymbolTable.h"
  31 #include "llvm/Pass.h"
  32 #include "llvm/Transforms/IPO.h"
  33 using namespace llvm;
  34
  35 namespace {
  36   class StripSymbols : public ModulePass {
  37     bool OnlyDebugInfo;
  38   public:
  39     static char ID; // Pass identification, replacement for typeid
  40     explicit StripSymbols(bool ODI = false)
  41       : ModulePass(ID), OnlyDebugInfo(ODI) {
  42         initializeStripSymbolsPass(*PassRegistry::getPassRegistry());
  43       }
  44
  45     bool runOnModule(Module &M) override;
  46
  47     void getAnalysisUsage(AnalysisUsage &AU) const override {
  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         initializeStripNonDebugSymbolsPass(*PassRegistry::getPassRegistry());
  58       }
  59
  60     bool runOnModule(Module &M) override;
  61
  62     void getAnalysisUsage(AnalysisUsage &AU) const override {
  63       AU.setPreservesAll();
  64     }
  65   };
  66
  67   class StripDebugDeclare : public ModulePass {
  68   public:
  69     static char ID; // Pass identification, replacement for typeid
  70     explicit StripDebugDeclare()
  71       : ModulePass(ID) {
  72         initializeStripDebugDeclarePass(*PassRegistry::getPassRegistry());
  73       }
  74
  75     bool runOnModule(Module &M) override;
  76
  77     void getAnalysisUsage(AnalysisUsage &AU) const override {
  78       AU.setPreservesAll();
  79     }
  80   };
  81
  82   class StripDeadDebugInfo : public ModulePass {
  83   public:
  84     static char ID; // Pass identification, replacement for typeid
  85     explicit StripDeadDebugInfo()
  86       : ModulePass(ID) {
  87         initializeStripDeadDebugInfoPass(*PassRegistry::getPassRegistry());
  88       }
  89
  90     bool runOnModule(Module &M) override;
  91
  92     void getAnalysisUsage(AnalysisUsage &AU) const override {
  93       AU.setPreservesAll();
  94     }
  95   };
  96 }
  97
  98 char StripSymbols::ID = 0;
  99 INITIALIZE_PASS(StripSymbols, "strip",
 100                 "Strip all symbols from a module", false, false)
 101
 102 ModulePass *llvm::createStripSymbolsPass(bool OnlyDebugInfo) {
 103   return new StripSymbols(OnlyDebugInfo);
 104 }
 105
 106 char StripNonDebugSymbols::ID = 0;
 107 INITIALIZE_PASS(StripNonDebugSymbols, "strip-nondebug",
 108                 "Strip all symbols, except dbg symbols, from a module",
 109                 false, false)
 110
 111 ModulePass *llvm::createStripNonDebugSymbolsPass() {
 112   return new StripNonDebugSymbols();
 113 }
 114
 115 char StripDebugDeclare::ID = 0;
 116 INITIALIZE_PASS(StripDebugDeclare, "strip-debug-declare",
 117                 "Strip all llvm.dbg.declare intrinsics", false, false)
 118
 119 ModulePass *llvm::createStripDebugDeclarePass() {
 120   return new StripDebugDeclare();
 121 }
 122
 123 char StripDeadDebugInfo::ID = 0;
 124 INITIALIZE_PASS(StripDeadDebugInfo, "strip-dead-debug-info",
 125                 "Strip debug info for unused symbols", false, false)
 126
 127 ModulePass *llvm::createStripDeadDebugInfoPass() {
 128   return new StripDeadDebugInfo();
 129 }
 130
 131 /// OnlyUsedBy - Return true if V is only used by Usr.
 132 static bool OnlyUsedBy(Value *V, Value *Usr) {
 133   for (User *U : V->users())
 134     if (U != Usr)
 135       return false;
 136
 137   return true;
 138 }
 139
 140 static void RemoveDeadConstant(Constant *C) {
 141   assert(C->use_empty() && "Constant is not dead!");
 142   SmallPtrSet<Constant*, 4> Operands;
 143   for (Value *Op : C->operands())
 144     if (OnlyUsedBy(Op, C))
 145       Operands.insert(cast<Constant>(Op));
 146   if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
 147     if (!GV->hasLocalLinkage()) return;   // Don't delete non-static globals.
 148     GV->eraseFromParent();
 149   }
 150   else if (!isa<Function>(C))
 151     if (isa<CompositeType>(C->getType()))
 152       C->destroyConstant();
 153
 154   // If the constant referenced anything, see if we can delete it as well.
 155   for (Constant *O : Operands)
 156     RemoveDeadConstant(O);
 157 }
 158
 159 // Strip the symbol table of its names.
 160 //
 161 static void StripSymtab(ValueSymbolTable &ST, bool PreserveDbgInfo) {
 162   for (ValueSymbolTable::iterator VI = ST.begin(), VE = ST.end(); VI != VE; ) {
 163     Value *V = VI->getValue();
 164     ++VI;
 165     if (!isa<GlobalValue>(V) || cast<GlobalValue>(V)->hasLocalLinkage()) {
 166       if (!PreserveDbgInfo || !V->getName().startswith("llvm.dbg"))
 167         // Set name to "", removing from symbol table!
 168         V->setName("");
 169     }
 170   }
 171 }
 172
 173 // Strip any named types of their names.
 174 static void StripTypeNames(Module &M, bool PreserveDbgInfo) {
 175   TypeFinder StructTypes;
 176   StructTypes.run(M, false);
 177
 178   for (unsigned i = 0, e = StructTypes.size(); i != e; ++i) {
 179     StructType *STy = StructTypes[i];
 180     if (STy->isLiteral() || STy->getName().empty()) continue;
 181
 182     if (PreserveDbgInfo && STy->getName().startswith("llvm.dbg"))
 183       continue;
 184
 185     STy->setName("");
 186   }
 187 }
 188
 189 /// Find values that are marked as llvm.used.
 190 static void findUsedValues(GlobalVariable *LLVMUsed,
 191                            SmallPtrSetImpl<const GlobalValue*> &UsedValues) {
 192   if (!LLVMUsed) return;
 193   UsedValues.insert(LLVMUsed);
 194
 195   ConstantArray *Inits = cast<ConstantArray>(LLVMUsed->getInitializer());
 196
 197   for (unsigned i = 0, e = Inits->getNumOperands(); i != e; ++i)
 198     if (GlobalValue *GV =
 199           dyn_cast<GlobalValue>(Inits->getOperand(i)->stripPointerCasts()))
 200       UsedValues.insert(GV);
 201 }
 202
 203 /// StripSymbolNames - Strip symbol names.
 204 static bool StripSymbolNames(Module &M, bool PreserveDbgInfo) {
 205
 206   SmallPtrSet<const GlobalValue*, 8> llvmUsedValues;
 207   findUsedValues(M.getGlobalVariable("llvm.used"), llvmUsedValues);
 208   findUsedValues(M.getGlobalVariable("llvm.compiler.used"), llvmUsedValues);
 209
 210   for (Module::global_iterator I = M.global_begin(), E = M.global_end();
 211        I != E; ++I) {
 212     if (I->hasLocalLinkage() && llvmUsedValues.count(&*I) == 0)
 213       if (!PreserveDbgInfo || !I->getName().startswith("llvm.dbg"))
 214         I->setName("");     // Internal symbols can't participate in linkage
 215   }
 216
 217   for (Function &I : M) {
 218     if (I.hasLocalLinkage() && llvmUsedValues.count(&I) == 0)
 219       if (!PreserveDbgInfo || !I.getName().startswith("llvm.dbg"))
 220         I.setName(""); // Internal symbols can't participate in linkage
 221     if (auto *Symtab = I.getValueSymbolTable())
 222       StripSymtab(*Symtab, PreserveDbgInfo);
 223   }
 224
 225   // Remove all names from types.
 226   StripTypeNames(M, PreserveDbgInfo);
 227
 228   return true;
 229 }
 230
 231 bool StripSymbols::runOnModule(Module &M) {
 232   if (skipModule(M))
 233     return false;
 234
 235   bool Changed = false;
 236   Changed |= StripDebugInfo(M);
 237   if (!OnlyDebugInfo)
 238     Changed |= StripSymbolNames(M, false);
 239   return Changed;
 240 }
 241
 242 bool StripNonDebugSymbols::runOnModule(Module &M) {
 243   if (skipModule(M))
 244     return false;
 245
 246   return StripSymbolNames(M, true);
 247 }
 248
 249 bool StripDebugDeclare::runOnModule(Module &M) {
 250   if (skipModule(M))
 251     return false;
 252
 253   Function *Declare = M.getFunction("llvm.dbg.declare");
 254   std::vector<Constant*> DeadConstants;
 255
 256   if (Declare) {
 257     while (!Declare->use_empty()) {
 258       CallInst *CI = cast<CallInst>(Declare->user_back());
 259       Value *Arg1 = CI->getArgOperand(0);
 260       Value *Arg2 = CI->getArgOperand(1);
 261       assert(CI->use_empty() && "llvm.dbg intrinsic should have void result");
 262       CI->eraseFromParent();
 263       if (Arg1->use_empty()) {
 264         if (Constant *C = dyn_cast<Constant>(Arg1))
 265           DeadConstants.push_back(C);
 266         else
 267           RecursivelyDeleteTriviallyDeadInstructions(Arg1);
 268       }
 269       if (Arg2->use_empty())
 270         if (Constant *C = dyn_cast<Constant>(Arg2))
 271           DeadConstants.push_back(C);
 272     }
 273     Declare->eraseFromParent();
 274   }
 275
 276   while (!DeadConstants.empty()) {
 277     Constant *C = DeadConstants.back();
 278     DeadConstants.pop_back();
 279     if (GlobalVariable *GV = dyn_cast<GlobalVariable>(C)) {
 280       if (GV->hasLocalLinkage())
 281         RemoveDeadConstant(GV);
 282     } else
 283       RemoveDeadConstant(C);
 284   }
 285
 286   return true;
 287 }
 288
 289 /// Remove any debug info for global variables/functions in the given module for
 290 /// which said global variable/function no longer exists (i.e. is null).
 291 ///
 292 /// Debugging information is encoded in llvm IR using metadata. This is designed
 293 /// such a way that debug info for symbols preserved even if symbols are
 294 /// optimized away by the optimizer. This special pass removes debug info for
 295 /// such symbols.
 296 bool StripDeadDebugInfo::runOnModule(Module &M) {
 297   if (skipModule(M))
 298     return false;
 299
 300   bool Changed = false;
 301
 302   LLVMContext &C = M.getContext();
 303
 304   // Find all debug info in F. This is actually overkill in terms of what we
 305   // want to do, but we want to try and be as resilient as possible in the face
 306   // of potential debug info changes by using the formal interfaces given to us
 307   // as much as possible.
 308   DebugInfoFinder F;
 309   F.processModule(M);
 310
 311   // For each compile unit, find the live set of global variables/functions and
 312   // replace the current list of potentially dead global variables/functions
 313   // with the live list.
 314   SmallVector<Metadata *, 64> LiveGlobalVariables;
 315   DenseSet<DIGlobalVariableExpression *> VisitedSet;
 316
 317   std::set<DIGlobalVariableExpression *> LiveGVs;
 318   for (GlobalVariable &GV : M.globals()) {
 319     SmallVector<DIGlobalVariableExpression *, 1> GVEs;
 320     GV.getDebugInfo(GVEs);
 321     for (auto *GVE : GVEs)
 322       LiveGVs.insert(GVE);
 323   }
 324
 325   std::set<DICompileUnit *> LiveCUs;
 326   // Any CU referenced from a subprogram is live.
 327   for (DISubprogram *SP : F.subprograms()) {
 328     if (SP->getUnit())
 329       LiveCUs.insert(SP->getUnit());
 330   }
 331
 332   bool HasDeadCUs = false;
 333   for (DICompileUnit *DIC : F.compile_units()) {
 334     // Create our live global variable list.
 335     bool GlobalVariableChange = false;
 336     for (auto *DIG : DIC->getGlobalVariables()) {
 337       if (DIG->getExpression() && DIG->getExpression()->isConstant())
 338         LiveGVs.insert(DIG);
 339
 340       // Make sure we only visit each global variable only once.
 341       if (!VisitedSet.insert(DIG).second)
 342         continue;
 343
 344       // If a global variable references DIG, the global variable is live.
 345       if (LiveGVs.count(DIG))
 346         LiveGlobalVariables.push_back(DIG);
 347       else
 348         GlobalVariableChange = true;
 349     }
 350
 351     if (!LiveGlobalVariables.empty())
 352       LiveCUs.insert(DIC);
 353     else if (!LiveCUs.count(DIC))
 354       HasDeadCUs = true;
 355
 356     // If we found dead global variables, replace the current global
 357     // variable list with our new live global variable list.
 358     if (GlobalVariableChange) {
 359       DIC->replaceGlobalVariables(MDTuple::get(C, LiveGlobalVariables));
 360       Changed = true;
 361     }
 362
 363     // Reset lists for the next iteration.
 364     LiveGlobalVariables.clear();
 365   }
 366
 367   if (HasDeadCUs) {
 368     // Delete the old node and replace it with a new one
 369     NamedMDNode *NMD = M.getOrInsertNamedMetadata("llvm.dbg.cu");
 370     NMD->clearOperands();
 371     if (!LiveCUs.empty()) {
 372       for (DICompileUnit *CU : LiveCUs)
 373         NMD->addOperand(CU);
 374     }
 375     Changed = true;
 376   }
 377
 378   return Changed;
 379 }