lib/Analysis/IPA/GlobalsModRef.cpp

   1 //===- GlobalsModRef.cpp - Simple Mod/Ref Analysis for Globals ------------===//
   2 //
   3 //                     The LLVM Compiler Infrastructure
   4 //
   5 // This file was developed by the LLVM research group and is distributed under
   6 // the University of Illinois Open Source License. See LICENSE.TXT for details.
   7 //
   8 //===----------------------------------------------------------------------===//
   9 //
  10 // This simple pass provides alias and mod/ref information for global values
  11 // that do not have their address taken, and keeps track of whether functions
  12 // read or write memory (are "pure").  For this simple (but very common) case,
  13 // we can provide pretty accurate and useful information.
  14 //
  15 //===----------------------------------------------------------------------===//
  16
  17 #include "llvm/Analysis/Passes.h"
  18 #include "llvm/Module.h"
  19 #include "llvm/Pass.h"
  20 #include "llvm/Instructions.h"
  21 #include "llvm/Constants.h"
  22 #include "llvm/Analysis/AliasAnalysis.h"
  23 #include "llvm/Analysis/CallGraph.h"
  24 #include "llvm/Support/InstIterator.h"
  25 #include "llvm/Support/CommandLine.h"
  26 #include "llvm/ADT/Statistic.h"
  27 #include "llvm/ADT/SCCIterator.h"
  28 #include <set>
  29 using namespace llvm;
  30
  31 namespace {
  32   Statistic<>
  33   NumNonAddrTakenGlobalVars("globalsmodref-aa",
  34                             "Number of global vars without address taken");
  35   Statistic<>
  36   NumNonAddrTakenFunctions("globalsmodref-aa",
  37                            "Number of functions without address taken");
  38   Statistic<>
  39   NumNoMemFunctions("globalsmodref-aa",
  40                     "Number of functions that do not access memory");
  41   Statistic<>
  42   NumReadMemFunctions("globalsmodref-aa",
  43                       "Number of functions that only read memory");
  44
  45   /// FunctionRecord - One instance of this structure is stored for every
  46   /// function in the program.  Later, the entries for these functions are
  47   /// removed if the function is found to call an external function (in which
  48   /// case we know nothing about it.
  49   struct FunctionRecord {
  50     /// GlobalInfo - Maintain mod/ref info for all of the globals without
  51     /// addresses taken that are read or written (transitively) by this
  52     /// function.
  53     std::map<GlobalValue*, unsigned> GlobalInfo;
  54
  55     unsigned getInfoForGlobal(GlobalValue *GV) const {
  56       std::map<GlobalValue*, unsigned>::const_iterator I = GlobalInfo.find(GV);
  57       if (I != GlobalInfo.end())
  58         return I->second;
  59       return 0;
  60     }
  61
  62     /// FunctionEffect - Capture whether or not this function reads or writes to
  63     /// ANY memory.  If not, we can do a lot of aggressive analysis on it.
  64     unsigned FunctionEffect;
  65
  66     FunctionRecord() : FunctionEffect(0) {}
  67   };
  68
  69   /// GlobalsModRef - The actual analysis pass.
  70   class GlobalsModRef : public ModulePass, public AliasAnalysis {
  71     /// NonAddressTakenGlobals - The globals that do not have their addresses
  72     /// taken.
  73     std::set<GlobalValue*> NonAddressTakenGlobals;
  74
  75     /// FunctionInfo - For each function, keep track of what globals are
  76     /// modified or read.
  77     std::map<Function*, FunctionRecord> FunctionInfo;
  78
  79   public:
  80     bool runOnModule(Module &M) {
  81       InitializeAliasAnalysis(this);                 // set up super class
  82       AnalyzeGlobals(M);                          // find non-addr taken globals
  83       AnalyzeCallGraph(getAnalysis<CallGraph>(), M); // Propagate on CG
  84       return false;
  85     }
  86
  87     virtual void getAnalysisUsage(AnalysisUsage &AU) const {
  88       AliasAnalysis::getAnalysisUsage(AU);
  89       AU.addRequired<CallGraph>();
  90       AU.setPreservesAll();                         // Does not transform code
  91     }
  92
  93     //------------------------------------------------
  94     // Implement the AliasAnalysis API
  95     //
  96     AliasResult alias(const Value *V1, unsigned V1Size,
  97                       const Value *V2, unsigned V2Size);
  98     ModRefResult getModRefInfo(CallSite CS, Value *P, unsigned Size);
  99     ModRefResult getModRefInfo(CallSite CS1, CallSite CS2) {
 100       return AliasAnalysis::getModRefInfo(CS1,CS2);
 101     }
 102     bool hasNoModRefInfoForCalls() const { return false; }
 103
 104     /// getModRefBehavior - Return the behavior of the specified function if
 105     /// called from the specified call site.  The call site may be null in which
 106     /// case the most generic behavior of this function should be returned.
 107     virtual ModRefBehavior getModRefBehavior(Function *F, CallSite CS,
 108                                          std::vector<PointerAccessInfo> *Info) {
 109       if (FunctionRecord *FR = getFunctionInfo(F))
 110         if (FR->FunctionEffect == 0)
 111           return DoesNotAccessMemory;
 112         else if ((FR->FunctionEffect & Mod) == 0)
 113           return OnlyReadsMemory;
 114       return AliasAnalysis::getModRefBehavior(F, CS, Info);
 115     }
 116
 117     virtual void deleteValue(Value *V);
 118     virtual void copyValue(Value *From, Value *To);
 119
 120   private:
 121     /// getFunctionInfo - Return the function info for the function, or null if
 122     /// the function calls an external function (in which case we don't have
 123     /// anything useful to say about it).
 124     FunctionRecord *getFunctionInfo(Function *F) {
 125       std::map<Function*, FunctionRecord>::iterator I = FunctionInfo.find(F);
 126       if (I != FunctionInfo.end())
 127         return &I->second;
 128       return 0;
 129     }
 130
 131     void AnalyzeGlobals(Module &M);
 132     void AnalyzeCallGraph(CallGraph &CG, Module &M);
 133     void AnalyzeSCC(std::vector<CallGraphNode *> &SCC);
 134     bool AnalyzeUsesOfGlobal(Value *V, std::vector<Function*> &Readers,
 135                              std::vector<Function*> &Writers);
 136   };
 137
 138   RegisterOpt<GlobalsModRef> X("globalsmodref-aa",
 139                                "Simple mod/ref analysis for globals");
 140   RegisterAnalysisGroup<AliasAnalysis, GlobalsModRef> Y;
 141 }
 142
 143 Pass *llvm::createGlobalsModRefPass() { return new GlobalsModRef(); }
 144
 145
 146 /// AnalyzeGlobalUses - Scan through the users of all of the internal
 147 /// GlobalValue's in the program.  If none of them have their "Address taken"
 148 /// (really, their address passed to something nontrivial), record this fact,
 149 /// and record the functions that they are used directly in.
 150 void GlobalsModRef::AnalyzeGlobals(Module &M) {
 151   std::vector<Function*> Readers, Writers;
 152   for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I)
 153     if (I->hasInternalLinkage()) {
 154       if (!AnalyzeUsesOfGlobal(I, Readers, Writers)) {
 155         // Remember that we are tracking this global.
 156         NonAddressTakenGlobals.insert(I);
 157         ++NumNonAddrTakenFunctions;
 158       }
 159       Readers.clear(); Writers.clear();
 160     }
 161
 162   for (Module::global_iterator I = M.global_begin(), E = M.global_end();
 163        I != E; ++I)
 164     if (I->hasInternalLinkage()) {
 165       if (!AnalyzeUsesOfGlobal(I, Readers, Writers)) {
 166         // Remember that we are tracking this global, and the mod/ref fns
 167         NonAddressTakenGlobals.insert(I);
 168         for (unsigned i = 0, e = Readers.size(); i != e; ++i)
 169           FunctionInfo[Readers[i]].GlobalInfo[I] |= Ref;
 170
 171         if (!I->isConstant())  // No need to keep track of writers to constants
 172           for (unsigned i = 0, e = Writers.size(); i != e; ++i)
 173             FunctionInfo[Writers[i]].GlobalInfo[I] |= Mod;
 174         ++NumNonAddrTakenGlobalVars;
 175       }
 176       Readers.clear(); Writers.clear();
 177     }
 178 }
 179
 180 /// AnalyzeUsesOfGlobal - Look at all of the users of the specified global value
 181 /// derived pointer.  If this is used by anything complex (i.e., the address
 182 /// escapes), return true.  Also, while we are at it, keep track of those
 183 /// functions that read and write to the value.
 184 bool GlobalsModRef::AnalyzeUsesOfGlobal(Value *V,
 185                                         std::vector<Function*> &Readers,
 186                                         std::vector<Function*> &Writers) {
 187   if (!isa<PointerType>(V->getType())) return true;
 188
 189   for (Value::use_iterator UI = V->use_begin(), E = V->use_end(); UI != E; ++UI)
 190     if (LoadInst *LI = dyn_cast<LoadInst>(*UI)) {
 191       Readers.push_back(LI->getParent()->getParent());
 192     } else if (StoreInst *SI = dyn_cast<StoreInst>(*UI)) {
 193       if (V == SI->getOperand(0)) return true;  // Storing the pointer
 194       Writers.push_back(SI->getParent()->getParent());
 195     } else if (GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(*UI)) {
 196       if (AnalyzeUsesOfGlobal(GEP, Readers, Writers)) return true;
 197     } else if (CallInst *CI = dyn_cast<CallInst>(*UI)) {
 198       // Make sure that this is just the function being called, not that it is
 199       // passing into the function.
 200       for (unsigned i = 1, e = CI->getNumOperands(); i != e; ++i)
 201         if (CI->getOperand(i) == V) return true;
 202     } else if (InvokeInst *II = dyn_cast<InvokeInst>(*UI)) {
 203       // Make sure that this is just the function being called, not that it is
 204       // passing into the function.
 205       for (unsigned i = 3, e = II->getNumOperands(); i != e; ++i)
 206         if (II->getOperand(i) == V) return true;
 207     } else if (ConstantExpr *CE = dyn_cast<ConstantExpr>(*UI)) {
 208       if (CE->getOpcode() == Instruction::GetElementPtr ||
 209           CE->getOpcode() == Instruction::Cast) {
 210         if (AnalyzeUsesOfGlobal(CE, Readers, Writers))
 211           return true;
 212       } else {
 213         return true;
 214       }
 215     } else if (GlobalValue *GV = dyn_cast<GlobalValue>(*UI)) {
 216       if (AnalyzeUsesOfGlobal(GV, Readers, Writers)) return true;
 217     } else {
 218       return true;
 219     }
 220   return false;
 221 }
 222
 223 /// AnalyzeCallGraph - At this point, we know the functions where globals are
 224 /// immediately stored to and read from.  Propagate this information up the call
 225 /// graph to all callers and compute the mod/ref info for all memory for each
 226 /// function.
 227 void GlobalsModRef::AnalyzeCallGraph(CallGraph &CG, Module &M) {
 228   // We do a bottom-up SCC traversal of the call graph.  In other words, we
 229   // visit all callees before callers (leaf-first).
 230   for (scc_iterator<CallGraph*> I = scc_begin(&CG), E = scc_end(&CG); I!=E; ++I)
 231     if ((*I).size() != 1) {
 232       AnalyzeSCC(*I);
 233     } else if (Function *F = (*I)[0]->getFunction()) {
 234       if (!F->isExternal()) {
 235         // Nonexternal function.
 236         AnalyzeSCC(*I);
 237       } else {
 238         // Otherwise external function.  Handle intrinsics and other special
 239         // cases here.
 240         if (getAnalysis<AliasAnalysis>().doesNotAccessMemory(F))
 241           // If it does not access memory, process the function, causing us to
 242           // realize it doesn't do anything (the body is empty).
 243           AnalyzeSCC(*I);
 244         else {
 245           // Otherwise, don't process it.  This will cause us to conservatively
 246           // assume the worst.
 247         }
 248       }
 249     } else {
 250       // Do not process the external node, assume the worst.
 251     }
 252 }
 253
 254 void GlobalsModRef::AnalyzeSCC(std::vector<CallGraphNode *> &SCC) {
 255   assert(!SCC.empty() && "SCC with no functions?");
 256   FunctionRecord &FR = FunctionInfo[SCC[0]->getFunction()];
 257
 258   bool CallsExternal = false;
 259   unsigned FunctionEffect = 0;
 260
 261   // Collect the mod/ref properties due to called functions.  We only compute
 262   // one mod-ref set
 263   for (unsigned i = 0, e = SCC.size(); i != e && !CallsExternal; ++i)
 264     for (CallGraphNode::iterator CI = SCC[i]->begin(), E = SCC[i]->end();
 265          CI != E; ++CI)
 266       if (Function *Callee = CI->second->getFunction()) {
 267         if (FunctionRecord *CalleeFR = getFunctionInfo(Callee)) {
 268           // Propagate function effect up.
 269           FunctionEffect |= CalleeFR->FunctionEffect;
 270
 271           // Incorporate callee's effects on globals into our info.
 272           for (std::map<GlobalValue*, unsigned>::iterator GI =
 273                  CalleeFR->GlobalInfo.begin(), E = CalleeFR->GlobalInfo.end();
 274                GI != E; ++GI)
 275             FR.GlobalInfo[GI->first] |= GI->second;
 276
 277         } else {
 278           // Okay, if we can't say anything about it, maybe some other alias
 279           // analysis can.
 280           ModRefBehavior MRB =
 281             AliasAnalysis::getModRefBehavior(Callee, CallSite());
 282           if (MRB != DoesNotAccessMemory) {
 283             // FIXME: could make this more aggressive for functions that just
 284             // read memory.  We should just say they read all globals.
 285             CallsExternal = true;
 286             break;
 287           }
 288         }
 289       } else {
 290         CallsExternal = true;
 291         break;
 292       }
 293
 294   // If this SCC calls an external function, we can't say anything about it, so
 295   // remove all SCC functions from the FunctionInfo map.
 296   if (CallsExternal) {
 297     for (unsigned i = 0, e = SCC.size(); i != e; ++i)
 298       FunctionInfo.erase(SCC[i]->getFunction());
 299     return;
 300   }
 301
 302   // Otherwise, unless we already know that this function mod/refs memory, scan
 303   // the function bodies to see if there are any explicit loads or stores.
 304   if (FunctionEffect != ModRef) {
 305     for (unsigned i = 0, e = SCC.size(); i != e && FunctionEffect != ModRef;++i)
 306       for (inst_iterator II = inst_begin(SCC[i]->getFunction()),
 307              E = inst_end(SCC[i]->getFunction());
 308            II != E && FunctionEffect != ModRef; ++II)
 309         if (isa<LoadInst>(*II))
 310           FunctionEffect |= Ref;
 311         else if (isa<StoreInst>(*II))
 312           FunctionEffect |= Mod;
 313         else if (isa<MallocInst>(*II) || isa<FreeInst>(*II))
 314           FunctionEffect |= ModRef;
 315   }
 316
 317   if ((FunctionEffect & Mod) == 0)
 318     ++NumReadMemFunctions;
 319   if (FunctionEffect == 0)
 320     ++NumNoMemFunctions;
 321   FR.FunctionEffect = FunctionEffect;
 322
 323   // Finally, now that we know the full effect on this SCC, clone the
 324   // information to each function in the SCC.
 325   for (unsigned i = 1, e = SCC.size(); i != e; ++i)
 326     FunctionInfo[SCC[i]->getFunction()] = FR;
 327 }
 328
 329
 330
 331 /// getUnderlyingObject - This traverses the use chain to figure out what object
 332 /// the specified value points to.  If the value points to, or is derived from,
 333 /// a global object, return it.
 334 static const GlobalValue *getUnderlyingObject(const Value *V) {
 335   if (!isa<PointerType>(V->getType())) return 0;
 336
 337   // If we are at some type of object... return it.
 338   if (const GlobalValue *GV = dyn_cast<GlobalValue>(V)) return GV;
 339
 340   // Traverse through different addressing mechanisms...
 341   if (const Instruction *I = dyn_cast<Instruction>(V)) {
 342     if (isa<CastInst>(I) || isa<GetElementPtrInst>(I))
 343       return getUnderlyingObject(I->getOperand(0));
 344   } else if (const ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
 345     if (CE->getOpcode() == Instruction::Cast ||
 346         CE->getOpcode() == Instruction::GetElementPtr)
 347       return getUnderlyingObject(CE->getOperand(0));
 348   }
 349   return 0;
 350 }
 351
 352 /// alias - If one of the pointers is to a global that we are tracking, and the
 353 /// other is some random pointer, we know there cannot be an alias, because the
 354 /// address of the global isn't taken.
 355 AliasAnalysis::AliasResult
 356 GlobalsModRef::alias(const Value *V1, unsigned V1Size,
 357                      const Value *V2, unsigned V2Size) {
 358   GlobalValue *GV1 = const_cast<GlobalValue*>(getUnderlyingObject(V1));
 359   GlobalValue *GV2 = const_cast<GlobalValue*>(getUnderlyingObject(V2));
 360
 361   // If the global's address is taken, pretend we don't know it's a pointer to
 362   // the global.
 363   if (GV1 && !NonAddressTakenGlobals.count(GV1)) GV1 = 0;
 364   if (GV2 && !NonAddressTakenGlobals.count(GV2)) GV2 = 0;
 365
 366   if ((GV1 || GV2) && GV1 != GV2)
 367     return NoAlias;
 368
 369   return AliasAnalysis::alias(V1, V1Size, V2, V2Size);
 370 }
 371
 372 AliasAnalysis::ModRefResult
 373 GlobalsModRef::getModRefInfo(CallSite CS, Value *P, unsigned Size) {
 374   unsigned Known = ModRef;
 375
 376   // If we are asking for mod/ref info of a direct call with a pointer to a
 377   // global we are tracking, return information if we have it.
 378   if (GlobalValue *GV = const_cast<GlobalValue*>(getUnderlyingObject(P)))
 379     if (GV->hasInternalLinkage())
 380       if (Function *F = CS.getCalledFunction())
 381         if (NonAddressTakenGlobals.count(GV))
 382           if (FunctionRecord *FR = getFunctionInfo(F))
 383             Known = FR->getInfoForGlobal(GV);
 384
 385   if (Known == NoModRef)
 386     return NoModRef; // No need to query other mod/ref analyses
 387   return ModRefResult(Known & AliasAnalysis::getModRefInfo(CS, P, Size));
 388 }
 389
 390
 391 //===----------------------------------------------------------------------===//
 392 // Methods to update the analysis as a result of the client transformation.
 393 //
 394 void GlobalsModRef::deleteValue(Value *V) {
 395   if (GlobalValue *GV = dyn_cast<GlobalValue>(V))
 396     NonAddressTakenGlobals.erase(GV);
 397 }
 398
 399 void GlobalsModRef::copyValue(Value *From, Value *To) {
 400 }