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/Metadata.h"
  18 #include "llvm/Module.h"
  19 #include "llvm/TypeSymbolTable.h"
  20 #include "llvm/ValueSymbolTable.h"
  21 #include "llvm/Instructions.h"
  22 #include <algorithm>
  23 using namespace llvm;
  24
  25 static bool isSingleValueType(const std::pair<const llvm::Type*,
  26                               unsigned int> &P) {
  27   return P.first->isSingleValueType();
  28 }
  29
  30 static bool isIntegerValue(const std::pair<const Value*, unsigned> &V) {
  31   return isa<IntegerType>(V.first->getType());
  32 }
  33
  34 static bool CompareByFrequency(const std::pair<const llvm::Type*,
  35                                unsigned int> &P1,
  36                                const std::pair<const llvm::Type*,
  37                                unsigned int> &P2) {
  38   return P1.second > P2.second;
  39 }
  40
  41 /// ValueEnumerator - Enumerate module-level information.
  42 ValueEnumerator::ValueEnumerator(const Module *M) {
  43   // Enumerate the global variables.
  44   for (Module::const_global_iterator I = M->global_begin(),
  45          E = M->global_end(); I != E; ++I)
  46     EnumerateValue(I);
  47
  48   // Enumerate the functions.
  49   for (Module::const_iterator I = M->begin(), E = M->end(); I != E; ++I) {
  50     EnumerateValue(I);
  51     EnumerateAttributes(cast<Function>(I)->getAttributes());
  52   }
  53
  54   // Enumerate the aliases.
  55   for (Module::const_alias_iterator I = M->alias_begin(), E = M->alias_end();
  56        I != E; ++I)
  57     EnumerateValue(I);
  58
  59   // Remember what is the cutoff between globalvalue's and other constants.
  60   unsigned FirstConstant = Values.size();
  61
  62   // Enumerate the global variable initializers.
  63   for (Module::const_global_iterator I = M->global_begin(),
  64          E = M->global_end(); I != E; ++I)
  65     if (I->hasInitializer())
  66       EnumerateValue(I->getInitializer());
  67
  68   // Enumerate the aliasees.
  69   for (Module::const_alias_iterator I = M->alias_begin(), E = M->alias_end();
  70        I != E; ++I)
  71     EnumerateValue(I->getAliasee());
  72
  73   // Enumerate types used by the type symbol table.
  74   EnumerateTypeSymbolTable(M->getTypeSymbolTable());
  75
  76   // Insert constants that are named at module level into the slot pool so that
  77   // the module symbol table can refer to them...
  78   EnumerateValueSymbolTable(M->getValueSymbolTable());
  79
  80   // Enumerate types used by function bodies and argument lists.
  81   for (Module::const_iterator F = M->begin(), E = M->end(); F != E; ++F) {
  82
  83     for (Function::const_arg_iterator I = F->arg_begin(), E = F->arg_end();
  84          I != E; ++I)
  85       EnumerateType(I->getType());
  86
  87     for (Function::const_iterator BB = F->begin(), E = F->end(); BB != E; ++BB)
  88       for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E;++I){
  89         for (User::const_op_iterator OI = I->op_begin(), E = I->op_end();
  90              OI != E; ++OI)
  91           EnumerateOperandType(*OI);
  92         EnumerateType(I->getType());
  93         if (const CallInst *CI = dyn_cast<CallInst>(I))
  94           EnumerateAttributes(CI->getAttributes());
  95         else if (const InvokeInst *II = dyn_cast<InvokeInst>(I))
  96           EnumerateAttributes(II->getAttributes());
  97       }
  98   }
  99
 100   // Optimize constant ordering.
 101   OptimizeConstants(FirstConstant, Values.size());
 102
 103   // Sort the type table by frequency so that most commonly used types are early
 104   // in the table (have low bit-width).
 105   std::stable_sort(Types.begin(), Types.end(), CompareByFrequency);
 106
 107   // Partition the Type ID's so that the single-value types occur before the
 108   // aggregate types.  This allows the aggregate types to be dropped from the
 109   // type table after parsing the global variable initializers.
 110   std::partition(Types.begin(), Types.end(), isSingleValueType);
 111
 112   // Now that we rearranged the type table, rebuild TypeMap.
 113   for (unsigned i = 0, e = Types.size(); i != e; ++i)
 114     TypeMap[Types[i].first] = i+1;
 115 }
 116
 117 unsigned ValueEnumerator::getValueID(const Value *V) const {
 118   if (isa<MetadataBase>(V)) {
 119     ValueMapType::const_iterator I = MDValueMap.find(V);
 120     assert(I != MDValueMap.end() && "Value not in slotcalculator!");
 121     return I->second-1;
 122   }
 123
 124   ValueMapType::const_iterator I = ValueMap.find(V);
 125   assert(I != ValueMap.end() && "Value not in slotcalculator!");
 126   return I->second-1;
 127 }
 128
 129 // Optimize constant ordering.
 130 namespace {
 131   struct CstSortPredicate {
 132     ValueEnumerator &VE;
 133     explicit CstSortPredicate(ValueEnumerator &ve) : VE(ve) {}
 134     bool operator()(const std::pair<const Value*, unsigned> &LHS,
 135                     const std::pair<const Value*, unsigned> &RHS) {
 136       // Sort by plane.
 137       if (LHS.first->getType() != RHS.first->getType())
 138         return VE.getTypeID(LHS.first->getType()) <
 139                VE.getTypeID(RHS.first->getType());
 140       // Then by frequency.
 141       return LHS.second > RHS.second;
 142     }
 143   };
 144 }
 145
 146 /// OptimizeConstants - Reorder constant pool for denser encoding.
 147 void ValueEnumerator::OptimizeConstants(unsigned CstStart, unsigned CstEnd) {
 148   if (CstStart == CstEnd || CstStart+1 == CstEnd) return;
 149
 150   CstSortPredicate P(*this);
 151   std::stable_sort(Values.begin()+CstStart, Values.begin()+CstEnd, P);
 152
 153   // Ensure that integer constants are at the start of the constant pool.  This
 154   // is important so that GEP structure indices come before gep constant exprs.
 155   std::partition(Values.begin()+CstStart, Values.begin()+CstEnd,
 156                  isIntegerValue);
 157
 158   // Rebuild the modified portion of ValueMap.
 159   for (; CstStart != CstEnd; ++CstStart)
 160     ValueMap[Values[CstStart].first] = CstStart+1;
 161 }
 162
 163
 164 /// EnumerateTypeSymbolTable - Insert all of the types in the specified symbol
 165 /// table.
 166 void ValueEnumerator::EnumerateTypeSymbolTable(const TypeSymbolTable &TST) {
 167   for (TypeSymbolTable::const_iterator TI = TST.begin(), TE = TST.end();
 168        TI != TE; ++TI)
 169     EnumerateType(TI->second);
 170 }
 171
 172 /// EnumerateValueSymbolTable - Insert all of the values in the specified symbol
 173 /// table into the values table.
 174 void ValueEnumerator::EnumerateValueSymbolTable(const ValueSymbolTable &VST) {
 175   for (ValueSymbolTable::const_iterator VI = VST.begin(), VE = VST.end();
 176        VI != VE; ++VI)
 177     EnumerateValue(VI->getValue());
 178 }
 179
 180 void ValueEnumerator::EnumerateMetadata(const MetadataBase *MD) {
 181   // Check to see if it's already in!
 182   unsigned &MDValueID = MDValueMap[MD];
 183   if (MDValueID) {
 184     // Increment use count.
 185     MDValues[MDValueID-1].second++;
 186     return;
 187   }
 188
 189   // Enumerate the type of this value.
 190   EnumerateType(MD->getType());
 191
 192   if (const MDNode *N = dyn_cast<MDNode>(MD)) {
 193     MDValues.push_back(std::make_pair(MD, 1U));
 194     MDValueMap[MD] = MDValues.size();
 195     MDValueID = MDValues.size();
 196     for (MDNode::const_elem_iterator I = N->elem_begin(), E = N->elem_end();
 197          I != E; ++I) {
 198       if (*I)
 199         EnumerateValue(*I);
 200       else
 201         EnumerateType(Type::getVoidTy(MD->getContext()));
 202     }
 203     return;
 204   } else if (const NamedMDNode *N = dyn_cast<NamedMDNode>(MD)) {
 205     for(NamedMDNode::const_elem_iterator I = N->elem_begin(),
 206           E = N->elem_end(); I != E; ++I) {
 207       MetadataBase *M = *I;
 208       EnumerateValue(M);
 209     }
 210     MDValues.push_back(std::make_pair(MD, 1U));
 211     MDValueMap[MD] = Values.size();
 212     return;
 213   }
 214
 215   // Add the value.
 216   MDValues.push_back(std::make_pair(MD, 1U));
 217   MDValueID = MDValues.size();
 218 }
 219
 220 void ValueEnumerator::EnumerateValue(const Value *V) {
 221   assert(V->getType() != Type::getVoidTy(V->getContext()) &&
 222          "Can't insert void values!");
 223   if (const MetadataBase *MB = dyn_cast<MetadataBase>(V))
 224     return EnumerateMetadata(MB);
 225
 226   // Check to see if it's already in!
 227   unsigned &ValueID = ValueMap[V];
 228   if (ValueID) {
 229     // Increment use count.
 230     Values[ValueID-1].second++;
 231     return;
 232   }
 233
 234   // Enumerate the type of this value.
 235   EnumerateType(V->getType());
 236
 237   if (const Constant *C = dyn_cast<Constant>(V)) {
 238     if (isa<GlobalValue>(C)) {
 239       // Initializers for globals are handled explicitly elsewhere.
 240     } else if (isa<ConstantArray>(C) && cast<ConstantArray>(C)->isString()) {
 241       // Do not enumerate the initializers for an array of simple characters.
 242       // The initializers just polute the value table, and we emit the strings
 243       // specially.
 244     } else if (C->getNumOperands()) {
 245       // If a constant has operands, enumerate them.  This makes sure that if a
 246       // constant has uses (for example an array of const ints), that they are
 247       // inserted also.
 248
 249       // We prefer to enumerate them with values before we enumerate the user
 250       // itself.  This makes it more likely that we can avoid forward references
 251       // in the reader.  We know that there can be no cycles in the constants
 252       // graph that don't go through a global variable.
 253       for (User::const_op_iterator I = C->op_begin(), E = C->op_end();
 254            I != E; ++I)
 255         EnumerateValue(*I);
 256
 257       // Finally, add the value.  Doing this could make the ValueID reference be
 258       // dangling, don't reuse it.
 259       Values.push_back(std::make_pair(V, 1U));
 260       ValueMap[V] = Values.size();
 261       return;
 262     }
 263   }
 264
 265   // Add the value.
 266   Values.push_back(std::make_pair(V, 1U));
 267   ValueID = Values.size();
 268 }
 269
 270
 271 void ValueEnumerator::EnumerateType(const Type *Ty) {
 272   unsigned &TypeID = TypeMap[Ty];
 273
 274   if (TypeID) {
 275     // If we've already seen this type, just increase its occurrence count.
 276     Types[TypeID-1].second++;
 277     return;
 278   }
 279
 280   // First time we saw this type, add it.
 281   Types.push_back(std::make_pair(Ty, 1U));
 282   TypeID = Types.size();
 283
 284   // Enumerate subtypes.
 285   for (Type::subtype_iterator I = Ty->subtype_begin(), E = Ty->subtype_end();
 286        I != E; ++I)
 287     EnumerateType(*I);
 288 }
 289
 290 // Enumerate the types for the specified value.  If the value is a constant,
 291 // walk through it, enumerating the types of the constant.
 292 void ValueEnumerator::EnumerateOperandType(const Value *V) {
 293   EnumerateType(V->getType());
 294   if (const Constant *C = dyn_cast<Constant>(V)) {
 295     // If this constant is already enumerated, ignore it, we know its type must
 296     // be enumerated.
 297     if (ValueMap.count(V)) return;
 298
 299     // This constant may have operands, make sure to enumerate the types in
 300     // them.
 301     for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i)
 302       EnumerateOperandType(C->getOperand(i));
 303
 304     if (const MDNode *N = dyn_cast<MDNode>(V)) {
 305       for (unsigned i = 0, e = N->getNumElements(); i != e; ++i) {
 306         Value *Elem = N->getElement(i);
 307         if (Elem)
 308           EnumerateOperandType(Elem);
 309       }
 310     }
 311   } else if (isa<MDString>(V) || isa<MDNode>(V))
 312     EnumerateValue(V);
 313 }
 314
 315 void ValueEnumerator::EnumerateAttributes(const AttrListPtr &PAL) {
 316   if (PAL.isEmpty()) return;  // null is always 0.
 317   // Do a lookup.
 318   unsigned &Entry = AttributeMap[PAL.getRawPointer()];
 319   if (Entry == 0) {
 320     // Never saw this before, add it.
 321     Attributes.push_back(PAL);
 322     Entry = Attributes.size();
 323   }
 324 }
 325
 326
 327 void ValueEnumerator::incorporateFunction(const Function &F) {
 328   NumModuleValues = Values.size();
 329
 330   // Adding function arguments to the value table.
 331   for(Function::const_arg_iterator I = F.arg_begin(), E = F.arg_end();
 332       I != E; ++I)
 333     EnumerateValue(I);
 334
 335   FirstFuncConstantID = Values.size();
 336
 337   // Add all function-level constants to the value table.
 338   for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) {
 339     for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I)
 340       for (User::const_op_iterator OI = I->op_begin(), E = I->op_end();
 341            OI != E; ++OI) {
 342         if ((isa<Constant>(*OI) && !isa<GlobalValue>(*OI)) ||
 343             isa<InlineAsm>(*OI))
 344           EnumerateValue(*OI);
 345       }
 346     BasicBlocks.push_back(BB);
 347     ValueMap[BB] = BasicBlocks.size();
 348   }
 349
 350   // Optimize the constant layout.
 351   OptimizeConstants(FirstFuncConstantID, Values.size());
 352
 353   // Add the function's parameter attributes so they are available for use in
 354   // the function's instruction.
 355   EnumerateAttributes(F.getAttributes());
 356
 357   FirstInstID = Values.size();
 358
 359   // Add all of the instructions.
 360   for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) {
 361     for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I) {
 362       if (I->getType() != Type::getVoidTy(F.getContext()))
 363         EnumerateValue(I);
 364     }
 365   }
 366 }
 367
 368 void ValueEnumerator::purgeFunction() {
 369   /// Remove purged values from the ValueMap.
 370   for (unsigned i = NumModuleValues, e = Values.size(); i != e; ++i)
 371     ValueMap.erase(Values[i].first);
 372   for (unsigned i = 0, e = BasicBlocks.size(); i != e; ++i)
 373     ValueMap.erase(BasicBlocks[i]);
 374
 375   Values.resize(NumModuleValues);
 376   BasicBlocks.clear();
 377 }
 378