lib/Bitcode/Writer/ValueEnumerator.cpp

   1 //===-- ValueEnumerator.cpp - Number values and types for bitcode writer --===//
   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 // This file implements the ValueEnumerator class.
  11 //
  12 //===----------------------------------------------------------------------===//
  13
  14 #include "ValueEnumerator.h"
  15 #include "llvm/Constants.h"
  16 #include "llvm/DerivedTypes.h"
  17 #include "llvm/Module.h"
  18 #include "llvm/TypeSymbolTable.h"
  19 #include "llvm/ValueSymbolTable.h"
  20 #include "llvm/Instructions.h"
  21 #include <algorithm>
  22 using namespace llvm;
  23
  24 static bool isSingleValueType(const std::pair<const llvm::Type*,
  25                               unsigned int> &P) {
  26   return P.first->isSingleValueType();
  27 }
  28
  29 static bool isIntegerValue(const std::pair<const Value*, unsigned> &V) {
  30   return V.first->getType()->isIntegerTy();
  31 }
  32
  33 static bool CompareByFrequency(const std::pair<const llvm::Type*,
  34                                unsigned int> &P1,
  35                                const std::pair<const llvm::Type*,
  36                                unsigned int> &P2) {
  37   return P1.second > P2.second;
  38 }
  39
  40 /// ValueEnumerator - Enumerate module-level information.
  41 ValueEnumerator::ValueEnumerator(const Module *M) {
  42   // Enumerate the global variables.
  43   for (Module::const_global_iterator I = M->global_begin(),
  44          E = M->global_end(); I != E; ++I)
  45     EnumerateValue(I);
  46
  47   // Enumerate the functions.
  48   for (Module::const_iterator I = M->begin(), E = M->end(); I != E; ++I) {
  49     EnumerateValue(I);
  50     EnumerateAttributes(cast<Function>(I)->getAttributes());
  51   }
  52
  53   // Enumerate the aliases.
  54   for (Module::const_alias_iterator I = M->alias_begin(), E = M->alias_end();
  55        I != E; ++I)
  56     EnumerateValue(I);
  57
  58   // Remember what is the cutoff between globalvalue's and other constants.
  59   unsigned FirstConstant = Values.size();
  60
  61   // Enumerate the global variable initializers.
  62   for (Module::const_global_iterator I = M->global_begin(),
  63          E = M->global_end(); I != E; ++I)
  64     if (I->hasInitializer())
  65       EnumerateValue(I->getInitializer());
  66
  67   // Enumerate the aliasees.
  68   for (Module::const_alias_iterator I = M->alias_begin(), E = M->alias_end();
  69        I != E; ++I)
  70     EnumerateValue(I->getAliasee());
  71
  72   // Enumerate types used by the type symbol table.
  73   EnumerateTypeSymbolTable(M->getTypeSymbolTable());
  74
  75   // Insert constants and metadata that are named at module level into the slot
  76   // pool so that the module symbol table can refer to them...
  77   EnumerateValueSymbolTable(M->getValueSymbolTable());
  78   EnumerateNamedMetadata(M);
  79
  80   SmallVector<std::pair<unsigned, MDNode*>, 8> MDs;
  81
  82   // Enumerate types used by function bodies and argument lists.
  83   for (Module::const_iterator F = M->begin(), E = M->end(); F != E; ++F) {
  84
  85     for (Function::const_arg_iterator I = F->arg_begin(), E = F->arg_end();
  86          I != E; ++I)
  87       EnumerateType(I->getType());
  88
  89     for (Function::const_iterator BB = F->begin(), E = F->end(); BB != E; ++BB)
  90       for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E;++I){
  91         for (User::const_op_iterator OI = I->op_begin(), E = I->op_end();
  92              OI != E; ++OI) {
  93           if (MDNode *MD = dyn_cast<MDNode>(*OI))
  94             if (MD->isFunctionLocal() && MD->getFunction())
  95               // These will get enumerated during function-incorporation.
  96               continue;
  97           EnumerateOperandType(*OI);
  98         }
  99         EnumerateType(I->getType());
 100         if (const CallInst *CI = dyn_cast<CallInst>(I))
 101           EnumerateAttributes(CI->getAttributes());
 102         else if (const InvokeInst *II = dyn_cast<InvokeInst>(I))
 103           EnumerateAttributes(II->getAttributes());
 104
 105         // Enumerate metadata attached with this instruction.
 106         MDs.clear();
 107         I->getAllMetadataOtherThanDebugLoc(MDs);
 108         for (unsigned i = 0, e = MDs.size(); i != e; ++i)
 109           EnumerateMetadata(MDs[i].second);
 110
 111         if (!I->getDebugLoc().isUnknown()) {
 112           MDNode *Scope, *IA;
 113           I->getDebugLoc().getScopeAndInlinedAt(Scope, IA, I->getContext());
 114           if (Scope) EnumerateMetadata(Scope);
 115           if (IA) EnumerateMetadata(IA);
 116         }
 117       }
 118   }
 119
 120   // Optimize constant ordering.
 121   OptimizeConstants(FirstConstant, Values.size());
 122
 123   // Sort the type table by frequency so that most commonly used types are early
 124   // in the table (have low bit-width).
 125   std::stable_sort(Types.begin(), Types.end(), CompareByFrequency);
 126
 127   // Partition the Type ID's so that the single-value types occur before the
 128   // aggregate types.  This allows the aggregate types to be dropped from the
 129   // type table after parsing the global variable initializers.
 130   std::partition(Types.begin(), Types.end(), isSingleValueType);
 131
 132   // Now that we rearranged the type table, rebuild TypeMap.
 133   for (unsigned i = 0, e = Types.size(); i != e; ++i)
 134     TypeMap[Types[i].first] = i+1;
 135 }
 136
 137 unsigned ValueEnumerator::getInstructionID(const Instruction *Inst) const {
 138   InstructionMapType::const_iterator I = InstructionMap.find(Inst);
 139   assert (I != InstructionMap.end() && "Instruction is not mapped!");
 140   return I->second;
 141 }
 142
 143 void ValueEnumerator::setInstructionID(const Instruction *I) {
 144   InstructionMap[I] = InstructionCount++;
 145 }
 146
 147 unsigned ValueEnumerator::getValueID(const Value *V) const {
 148   if (isa<MDNode>(V) || isa<MDString>(V)) {
 149     ValueMapType::const_iterator I = MDValueMap.find(V);
 150     assert(I != MDValueMap.end() && "Value not in slotcalculator!");
 151     return I->second-1;
 152   }
 153
 154   ValueMapType::const_iterator I = ValueMap.find(V);
 155   assert(I != ValueMap.end() && "Value not in slotcalculator!");
 156   return I->second-1;
 157 }
 158
 159 // Optimize constant ordering.
 160 namespace {
 161   struct CstSortPredicate {
 162     ValueEnumerator &VE;
 163     explicit CstSortPredicate(ValueEnumerator &ve) : VE(ve) {}
 164     bool operator()(const std::pair<const Value*, unsigned> &LHS,
 165                     const std::pair<const Value*, unsigned> &RHS) {
 166       // Sort by plane.
 167       if (LHS.first->getType() != RHS.first->getType())
 168         return VE.getTypeID(LHS.first->getType()) <
 169                VE.getTypeID(RHS.first->getType());
 170       // Then by frequency.
 171       return LHS.second > RHS.second;
 172     }
 173   };
 174 }
 175
 176 /// OptimizeConstants - Reorder constant pool for denser encoding.
 177 void ValueEnumerator::OptimizeConstants(unsigned CstStart, unsigned CstEnd) {
 178   if (CstStart == CstEnd || CstStart+1 == CstEnd) return;
 179
 180   CstSortPredicate P(*this);
 181   std::stable_sort(Values.begin()+CstStart, Values.begin()+CstEnd, P);
 182
 183   // Ensure that integer constants are at the start of the constant pool.  This
 184   // is important so that GEP structure indices come before gep constant exprs.
 185   std::partition(Values.begin()+CstStart, Values.begin()+CstEnd,
 186                  isIntegerValue);
 187
 188   // Rebuild the modified portion of ValueMap.
 189   for (; CstStart != CstEnd; ++CstStart)
 190     ValueMap[Values[CstStart].first] = CstStart+1;
 191 }
 192
 193
 194 /// EnumerateTypeSymbolTable - Insert all of the types in the specified symbol
 195 /// table.
 196 void ValueEnumerator::EnumerateTypeSymbolTable(const TypeSymbolTable &TST) {
 197   for (TypeSymbolTable::const_iterator TI = TST.begin(), TE = TST.end();
 198        TI != TE; ++TI)
 199     EnumerateType(TI->second);
 200 }
 201
 202 /// EnumerateValueSymbolTable - Insert all of the values in the specified symbol
 203 /// table into the values table.
 204 void ValueEnumerator::EnumerateValueSymbolTable(const ValueSymbolTable &VST) {
 205   for (ValueSymbolTable::const_iterator VI = VST.begin(), VE = VST.end();
 206        VI != VE; ++VI)
 207     EnumerateValue(VI->getValue());
 208 }
 209
 210 /// EnumerateNamedMetadata - Insert all of the values referenced by
 211 /// named metadata in the specified module.
 212 void ValueEnumerator::EnumerateNamedMetadata(const Module *M) {
 213   for (Module::const_named_metadata_iterator I = M->named_metadata_begin(),
 214        E = M->named_metadata_end(); I != E; ++I)
 215     EnumerateNamedMDNode(I);
 216 }
 217
 218 void ValueEnumerator::EnumerateNamedMDNode(const NamedMDNode *MD) {
 219   for (unsigned i = 0, e = MD->getNumOperands(); i != e; ++i)
 220     EnumerateMetadata(MD->getOperand(i));
 221 }
 222
 223 /// EnumerateMDNodeOperands - Enumerate all non-function-local values
 224 /// and types referenced by the given MDNode.
 225 void ValueEnumerator::EnumerateMDNodeOperands(const MDNode *N) {
 226   for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) {
 227     if (Value *V = N->getOperand(i)) {
 228       if (isa<MDNode>(V) || isa<MDString>(V))
 229         EnumerateMetadata(V);
 230       else if (!isa<Instruction>(V) && !isa<Argument>(V))
 231         EnumerateValue(V);
 232     } else
 233       EnumerateType(Type::getVoidTy(N->getContext()));
 234   }
 235 }
 236
 237 void ValueEnumerator::EnumerateMetadata(const Value *MD) {
 238   assert((isa<MDNode>(MD) || isa<MDString>(MD)) && "Invalid metadata kind");
 239
 240   // Enumerate the type of this value.
 241   EnumerateType(MD->getType());
 242
 243   const MDNode *N = dyn_cast<MDNode>(MD);
 244
 245   // In the module-level pass, skip function-local nodes themselves, but
 246   // do walk their operands.
 247   if (N && N->isFunctionLocal() && N->getFunction()) {
 248     EnumerateMDNodeOperands(N);
 249     return;
 250   }
 251
 252   // Check to see if it's already in!
 253   unsigned &MDValueID = MDValueMap[MD];
 254   if (MDValueID) {
 255     // Increment use count.
 256     MDValues[MDValueID-1].second++;
 257     return;
 258   }
 259   MDValues.push_back(std::make_pair(MD, 1U));
 260   MDValueID = MDValues.size();
 261
 262   // Enumerate all non-function-local operands.
 263   if (N)
 264     EnumerateMDNodeOperands(N);
 265 }
 266
 267 /// EnumerateFunctionLocalMetadataa - Incorporate function-local metadata
 268 /// information reachable from the given MDNode.
 269 void ValueEnumerator::EnumerateFunctionLocalMetadata(const MDNode *N) {
 270   assert(N->isFunctionLocal() && N->getFunction() &&
 271          "EnumerateFunctionLocalMetadata called on non-function-local mdnode!");
 272
 273   // Enumerate the type of this value.
 274   EnumerateType(N->getType());
 275
 276   // Check to see if it's already in!
 277   unsigned &MDValueID = MDValueMap[N];
 278   if (MDValueID) {
 279     // Increment use count.
 280     MDValues[MDValueID-1].second++;
 281     return;
 282   }
 283   MDValues.push_back(std::make_pair(N, 1U));
 284   MDValueID = MDValues.size();
 285
 286   // To incoroporate function-local information visit all function-local
 287   // MDNodes and all function-local values they reference.
 288   for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i)
 289     if (Value *V = N->getOperand(i)) {
 290       if (MDNode *O = dyn_cast<MDNode>(V)) {
 291         if (O->isFunctionLocal() && O->getFunction())
 292           EnumerateFunctionLocalMetadata(O);
 293       } else if (isa<Instruction>(V) || isa<Argument>(V))
 294         EnumerateValue(V);
 295     }
 296
 297   // Also, collect all function-local MDNodes for easy access.
 298   FunctionLocalMDs.push_back(N);
 299 }
 300
 301 void ValueEnumerator::EnumerateValue(const Value *V) {
 302   assert(!V->getType()->isVoidTy() && "Can't insert void values!");
 303   assert(!isa<MDNode>(V) && !isa<MDString>(V) &&
 304          "EnumerateValue doesn't handle Metadata!");
 305
 306   // Check to see if it's already in!
 307   unsigned &ValueID = ValueMap[V];
 308   if (ValueID) {
 309     // Increment use count.
 310     Values[ValueID-1].second++;
 311     return;
 312   }
 313
 314   // Enumerate the type of this value.
 315   EnumerateType(V->getType());
 316
 317   if (const Constant *C = dyn_cast<Constant>(V)) {
 318     if (isa<GlobalValue>(C)) {
 319       // Initializers for globals are handled explicitly elsewhere.
 320     } else if (isa<ConstantArray>(C) && cast<ConstantArray>(C)->isString()) {
 321       // Do not enumerate the initializers for an array of simple characters.
 322       // The initializers just polute the value table, and we emit the strings
 323       // specially.
 324     } else if (C->getNumOperands()) {
 325       // If a constant has operands, enumerate them.  This makes sure that if a
 326       // constant has uses (for example an array of const ints), that they are
 327       // inserted also.
 328
 329       // We prefer to enumerate them with values before we enumerate the user
 330       // itself.  This makes it more likely that we can avoid forward references
 331       // in the reader.  We know that there can be no cycles in the constants
 332       // graph that don't go through a global variable.
 333       for (User::const_op_iterator I = C->op_begin(), E = C->op_end();
 334            I != E; ++I)
 335         if (!isa<BasicBlock>(*I)) // Don't enumerate BB operand to BlockAddress.
 336           EnumerateValue(*I);
 337
 338       // Finally, add the value.  Doing this could make the ValueID reference be
 339       // dangling, don't reuse it.
 340       Values.push_back(std::make_pair(V, 1U));
 341       ValueMap[V] = Values.size();
 342       return;
 343     }
 344   }
 345
 346   // Add the value.
 347   Values.push_back(std::make_pair(V, 1U));
 348   ValueID = Values.size();
 349 }
 350
 351
 352 void ValueEnumerator::EnumerateType(const Type *Ty) {
 353   unsigned &TypeID = TypeMap[Ty];
 354
 355   if (TypeID) {
 356     // If we've already seen this type, just increase its occurrence count.
 357     Types[TypeID-1].second++;
 358     return;
 359   }
 360
 361   // First time we saw this type, add it.
 362   Types.push_back(std::make_pair(Ty, 1U));
 363   TypeID = Types.size();
 364
 365   // Enumerate subtypes.
 366   for (Type::subtype_iterator I = Ty->subtype_begin(), E = Ty->subtype_end();
 367        I != E; ++I)
 368     EnumerateType(*I);
 369 }
 370
 371 // Enumerate the types for the specified value.  If the value is a constant,
 372 // walk through it, enumerating the types of the constant.
 373 void ValueEnumerator::EnumerateOperandType(const Value *V) {
 374   EnumerateType(V->getType());
 375
 376   if (const Constant *C = dyn_cast<Constant>(V)) {
 377     // If this constant is already enumerated, ignore it, we know its type must
 378     // be enumerated.
 379     if (ValueMap.count(V)) return;
 380
 381     // This constant may have operands, make sure to enumerate the types in
 382     // them.
 383     for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i) {
 384       const User *Op = C->getOperand(i);
 385
 386       // Don't enumerate basic blocks here, this happens as operands to
 387       // blockaddress.
 388       if (isa<BasicBlock>(Op)) continue;
 389
 390       EnumerateOperandType(Op);
 391     }
 392
 393     if (const MDNode *N = dyn_cast<MDNode>(V)) {
 394       for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i)
 395         if (Value *Elem = N->getOperand(i))
 396           EnumerateOperandType(Elem);
 397     }
 398   } else if (isa<MDString>(V) || isa<MDNode>(V))
 399     EnumerateMetadata(V);
 400 }
 401
 402 void ValueEnumerator::EnumerateAttributes(const AttrListPtr &PAL) {
 403   if (PAL.isEmpty()) return;  // null is always 0.
 404   // Do a lookup.
 405   unsigned &Entry = AttributeMap[PAL.getRawPointer()];
 406   if (Entry == 0) {
 407     // Never saw this before, add it.
 408     Attributes.push_back(PAL);
 409     Entry = Attributes.size();
 410   }
 411 }
 412
 413
 414 void ValueEnumerator::incorporateFunction(const Function &F) {
 415   InstructionCount = 0;
 416   NumModuleValues = Values.size();
 417   NumModuleMDValues = MDValues.size();
 418
 419   // Adding function arguments to the value table.
 420   for (Function::const_arg_iterator I = F.arg_begin(), E = F.arg_end();
 421        I != E; ++I)
 422     EnumerateValue(I);
 423
 424   FirstFuncConstantID = Values.size();
 425
 426   // Add all function-level constants to the value table.
 427   for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) {
 428     for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I)
 429       for (User::const_op_iterator OI = I->op_begin(), E = I->op_end();
 430            OI != E; ++OI) {
 431         if ((isa<Constant>(*OI) && !isa<GlobalValue>(*OI)) ||
 432             isa<InlineAsm>(*OI))
 433           EnumerateValue(*OI);
 434       }
 435     BasicBlocks.push_back(BB);
 436     ValueMap[BB] = BasicBlocks.size();
 437   }
 438
 439   // Optimize the constant layout.
 440   OptimizeConstants(FirstFuncConstantID, Values.size());
 441
 442   // Add the function's parameter attributes so they are available for use in
 443   // the function's instruction.
 444   EnumerateAttributes(F.getAttributes());
 445
 446   FirstInstID = Values.size();
 447
 448   SmallVector<MDNode *, 8> FnLocalMDVector;
 449   // Add all of the instructions.
 450   for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) {
 451     for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I) {
 452       for (User::const_op_iterator OI = I->op_begin(), E = I->op_end();
 453            OI != E; ++OI) {
 454         if (MDNode *MD = dyn_cast<MDNode>(*OI))
 455           if (MD->isFunctionLocal() && MD->getFunction())
 456             // Enumerate metadata after the instructions they might refer to.
 457             FnLocalMDVector.push_back(MD);
 458       }
 459
 460       SmallVector<std::pair<unsigned, MDNode*>, 8> MDs;
 461       I->getAllMetadataOtherThanDebugLoc(MDs);
 462       for (unsigned i = 0, e = MDs.size(); i != e; ++i) {
 463         MDNode *N = MDs[i].second;
 464         if (N->isFunctionLocal() && N->getFunction())
 465           FnLocalMDVector.push_back(N);
 466       }
 467
 468       if (!I->getType()->isVoidTy())
 469         EnumerateValue(I);
 470     }
 471   }
 472
 473   // Add all of the function-local metadata.
 474   for (unsigned i = 0, e = FnLocalMDVector.size(); i != e; ++i)
 475     EnumerateFunctionLocalMetadata(FnLocalMDVector[i]);
 476 }
 477
 478 void ValueEnumerator::purgeFunction() {
 479   /// Remove purged values from the ValueMap.
 480   for (unsigned i = NumModuleValues, e = Values.size(); i != e; ++i)
 481     ValueMap.erase(Values[i].first);
 482   for (unsigned i = NumModuleMDValues, e = MDValues.size(); i != e; ++i)
 483     MDValueMap.erase(MDValues[i].first);
 484   for (unsigned i = 0, e = BasicBlocks.size(); i != e; ++i)
 485     ValueMap.erase(BasicBlocks[i]);
 486
 487   Values.resize(NumModuleValues);
 488   MDValues.resize(NumModuleMDValues);
 489   BasicBlocks.clear();
 490   FunctionLocalMDs.clear();
 491 }
 492
 493 static void IncorporateFunctionInfoGlobalBBIDs(const Function *F,
 494                                  DenseMap<const BasicBlock*, unsigned> &IDMap) {
 495   unsigned Counter = 0;
 496   for (Function::const_iterator BB = F->begin(), E = F->end(); BB != E; ++BB)
 497     IDMap[BB] = ++Counter;
 498 }
 499
 500 /// getGlobalBasicBlockID - This returns the function-specific ID for the
 501 /// specified basic block.  This is relatively expensive information, so it
 502 /// should only be used by rare constructs such as address-of-label.
 503 unsigned ValueEnumerator::getGlobalBasicBlockID(const BasicBlock *BB) const {
 504   unsigned &Idx = GlobalBasicBlockIDs[BB];
 505   if (Idx != 0)
 506     return Idx-1;
 507
 508   IncorporateFunctionInfoGlobalBBIDs(BB->getParent(), GlobalBasicBlockIDs);
 509   return getGlobalBasicBlockID(BB);
 510 }
 511