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