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/ADT/STLExtras.h" 16 #include "llvm/ADT/SmallPtrSet.h" 17 #include "llvm/IR/Constants.h" 18 #include "llvm/IR/DerivedTypes.h" 19 #include "llvm/IR/Instructions.h" 20 #include "llvm/IR/Module.h" 21 #include "llvm/IR/UseListOrder.h" 22 #include "llvm/IR/ValueSymbolTable.h" 23 #include "llvm/Support/Debug.h" 24 #include "llvm/Support/raw_ostream.h" 25 #include <algorithm> 26 using namespace llvm; 27 28 namespace { 29 struct OrderMap { 30 DenseMap<const Value *, std::pair<unsigned, bool>> IDs; 31 unsigned LastGlobalConstantID; 32 unsigned LastGlobalValueID; 33 34 OrderMap() : LastGlobalConstantID(0), LastGlobalValueID(0) {} 35 36 bool isGlobalConstant(unsigned ID) const { 37 return ID <= LastGlobalConstantID; 38 } 39 bool isGlobalValue(unsigned ID) const { 40 return ID <= LastGlobalValueID && !isGlobalConstant(ID); 41 } 42 43 unsigned size() const { return IDs.size(); } 44 std::pair<unsigned, bool> &operator[](const Value *V) { return IDs[V]; } 45 std::pair<unsigned, bool> lookup(const Value *V) const { 46 return IDs.lookup(V); 47 } 48 void index(const Value *V) { 49 // Explicitly sequence get-size and insert-value operations to avoid UB. 50 unsigned ID = IDs.size() + 1; 51 IDs[V].first = ID; 52 } 53 }; 54 } 55 56 static void orderValue(const Value *V, OrderMap &OM) { 57 if (OM.lookup(V).first) 58 return; 59 60 if (const Constant *C = dyn_cast<Constant>(V)) 61 if (C->getNumOperands() && !isa<GlobalValue>(C)) 62 for (const Value *Op : C->operands()) 63 if (!isa<BasicBlock>(Op) && !isa<GlobalValue>(Op)) 64 orderValue(Op, OM); 65 66 // Note: we cannot cache this lookup above, since inserting into the map 67 // changes the map's size, and thus affects the other IDs. 68 OM.index(V); 69 } 70 71 static OrderMap orderModule(const Module &M) { 72 // This needs to match the order used by ValueEnumerator::ValueEnumerator() 73 // and ValueEnumerator::incorporateFunction(). 74 OrderMap OM; 75 76 // In the reader, initializers of GlobalValues are set *after* all the 77 // globals have been read. Rather than awkwardly modeling this behaviour 78 // directly in predictValueUseListOrderImpl(), just assign IDs to 79 // initializers of GlobalValues before GlobalValues themselves to model this 80 // implicitly. 81 for (const GlobalVariable &G : M.globals()) 82 if (G.hasInitializer()) 83 if (!isa<GlobalValue>(G.getInitializer())) 84 orderValue(G.getInitializer(), OM); 85 for (const GlobalAlias &A : M.aliases()) 86 if (!isa<GlobalValue>(A.getAliasee())) 87 orderValue(A.getAliasee(), OM); 88 for (const Function &F : M) { 89 if (F.hasPrefixData()) 90 if (!isa<GlobalValue>(F.getPrefixData())) 91 orderValue(F.getPrefixData(), OM); 92 if (F.hasPrologueData()) 93 if (!isa<GlobalValue>(F.getPrologueData())) 94 orderValue(F.getPrologueData(), OM); 95 } 96 OM.LastGlobalConstantID = OM.size(); 97 98 // Initializers of GlobalValues are processed in 99 // BitcodeReader::ResolveGlobalAndAliasInits(). Match the order there rather 100 // than ValueEnumerator, and match the code in predictValueUseListOrderImpl() 101 // by giving IDs in reverse order. 102 // 103 // Since GlobalValues never reference each other directly (just through 104 // initializers), their relative IDs only matter for determining order of 105 // uses in their initializers. 106 for (const Function &F : M) 107 orderValue(&F, OM); 108 for (const GlobalAlias &A : M.aliases()) 109 orderValue(&A, OM); 110 for (const GlobalVariable &G : M.globals()) 111 orderValue(&G, OM); 112 OM.LastGlobalValueID = OM.size(); 113 114 for (const Function &F : M) { 115 if (F.isDeclaration()) 116 continue; 117 // Here we need to match the union of ValueEnumerator::incorporateFunction() 118 // and WriteFunction(). Basic blocks are implicitly declared before 119 // anything else (by declaring their size). 120 for (const BasicBlock &BB : F) 121 orderValue(&BB, OM); 122 for (const Argument &A : F.args()) 123 orderValue(&A, OM); 124 for (const BasicBlock &BB : F) 125 for (const Instruction &I : BB) 126 for (const Value *Op : I.operands()) 127 if ((isa<Constant>(*Op) && !isa<GlobalValue>(*Op)) || 128 isa<InlineAsm>(*Op)) 129 orderValue(Op, OM); 130 for (const BasicBlock &BB : F) 131 for (const Instruction &I : BB) 132 orderValue(&I, OM); 133 } 134 return OM; 135 } 136 137 static void predictValueUseListOrderImpl(const Value *V, const Function *F, 138 unsigned ID, const OrderMap &OM, 139 UseListOrderStack &Stack) { 140 // Predict use-list order for this one. 141 typedef std::pair<const Use *, unsigned> Entry; 142 SmallVector<Entry, 64> List; 143 for (const Use &U : V->uses()) 144 // Check if this user will be serialized. 145 if (OM.lookup(U.getUser()).first) 146 List.push_back(std::make_pair(&U, List.size())); 147 148 if (List.size() < 2) 149 // We may have lost some users. 150 return; 151 152 bool IsGlobalValue = OM.isGlobalValue(ID); 153 std::sort(List.begin(), List.end(), [&](const Entry &L, const Entry &R) { 154 const Use *LU = L.first; 155 const Use *RU = R.first; 156 if (LU == RU) 157 return false; 158 159 auto LID = OM.lookup(LU->getUser()).first; 160 auto RID = OM.lookup(RU->getUser()).first; 161 162 // Global values are processed in reverse order. 163 // 164 // Moreover, initializers of GlobalValues are set *after* all the globals 165 // have been read (despite having earlier IDs). Rather than awkwardly 166 // modeling this behaviour here, orderModule() has assigned IDs to 167 // initializers of GlobalValues before GlobalValues themselves. 168 if (OM.isGlobalValue(LID) && OM.isGlobalValue(RID)) 169 return LID < RID; 170 171 // If ID is 4, then expect: 7 6 5 1 2 3. 172 if (LID < RID) { 173 if (RID <= ID) 174 if (!IsGlobalValue) // GlobalValue uses don't get reversed. 175 return true; 176 return false; 177 } 178 if (RID < LID) { 179 if (LID <= ID) 180 if (!IsGlobalValue) // GlobalValue uses don't get reversed. 181 return false; 182 return true; 183 } 184 185 // LID and RID are equal, so we have different operands of the same user. 186 // Assume operands are added in order for all instructions. 187 if (LID <= ID) 188 if (!IsGlobalValue) // GlobalValue uses don't get reversed. 189 return LU->getOperandNo() < RU->getOperandNo(); 190 return LU->getOperandNo() > RU->getOperandNo(); 191 }); 192 193 if (std::is_sorted( 194 List.begin(), List.end(), 195 [](const Entry &L, const Entry &R) { return L.second < R.second; })) 196 // Order is already correct. 197 return; 198 199 // Store the shuffle. 200 Stack.emplace_back(V, F, List.size()); 201 assert(List.size() == Stack.back().Shuffle.size() && "Wrong size"); 202 for (size_t I = 0, E = List.size(); I != E; ++I) 203 Stack.back().Shuffle[I] = List[I].second; 204 } 205 206 static void predictValueUseListOrder(const Value *V, const Function *F, 207 OrderMap &OM, UseListOrderStack &Stack) { 208 auto &IDPair = OM[V]; 209 assert(IDPair.first && "Unmapped value"); 210 if (IDPair.second) 211 // Already predicted. 212 return; 213 214 // Do the actual prediction. 215 IDPair.second = true; 216 if (!V->use_empty() && std::next(V->use_begin()) != V->use_end()) 217 predictValueUseListOrderImpl(V, F, IDPair.first, OM, Stack); 218 219 // Recursive descent into constants. 220 if (const Constant *C = dyn_cast<Constant>(V)) 221 if (C->getNumOperands()) // Visit GlobalValues. 222 for (const Value *Op : C->operands()) 223 if (isa<Constant>(Op)) // Visit GlobalValues. 224 predictValueUseListOrder(Op, F, OM, Stack); 225 } 226 227 static UseListOrderStack predictUseListOrder(const Module &M) { 228 OrderMap OM = orderModule(M); 229 230 // Use-list orders need to be serialized after all the users have been added 231 // to a value, or else the shuffles will be incomplete. Store them per 232 // function in a stack. 233 // 234 // Aside from function order, the order of values doesn't matter much here. 235 UseListOrderStack Stack; 236 237 // We want to visit the functions backward now so we can list function-local 238 // constants in the last Function they're used in. Module-level constants 239 // have already been visited above. 240 for (auto I = M.rbegin(), E = M.rend(); I != E; ++I) { 241 const Function &F = *I; 242 if (F.isDeclaration()) 243 continue; 244 for (const BasicBlock &BB : F) 245 predictValueUseListOrder(&BB, &F, OM, Stack); 246 for (const Argument &A : F.args()) 247 predictValueUseListOrder(&A, &F, OM, Stack); 248 for (const BasicBlock &BB : F) 249 for (const Instruction &I : BB) 250 for (const Value *Op : I.operands()) 251 if (isa<Constant>(*Op) || isa<InlineAsm>(*Op)) // Visit GlobalValues. 252 predictValueUseListOrder(Op, &F, OM, Stack); 253 for (const BasicBlock &BB : F) 254 for (const Instruction &I : BB) 255 predictValueUseListOrder(&I, &F, OM, Stack); 256 } 257 258 // Visit globals last, since the module-level use-list block will be seen 259 // before the function bodies are processed. 260 for (const GlobalVariable &G : M.globals()) 261 predictValueUseListOrder(&G, nullptr, OM, Stack); 262 for (const Function &F : M) 263 predictValueUseListOrder(&F, nullptr, OM, Stack); 264 for (const GlobalAlias &A : M.aliases()) 265 predictValueUseListOrder(&A, nullptr, OM, Stack); 266 for (const GlobalVariable &G : M.globals()) 267 if (G.hasInitializer()) 268 predictValueUseListOrder(G.getInitializer(), nullptr, OM, Stack); 269 for (const GlobalAlias &A : M.aliases()) 270 predictValueUseListOrder(A.getAliasee(), nullptr, OM, Stack); 271 for (const Function &F : M) { 272 if (F.hasPrefixData()) 273 predictValueUseListOrder(F.getPrefixData(), nullptr, OM, Stack); 274 if (F.hasPrologueData()) 275 predictValueUseListOrder(F.getPrologueData(), nullptr, OM, Stack); 276 } 277 278 return Stack; 279 } 280 281 static bool isIntOrIntVectorValue(const std::pair<const Value*, unsigned> &V) { 282 return V.first->getType()->isIntOrIntVectorTy(); 283 } 284 285 ValueEnumerator::ValueEnumerator(const Module &M) 286 : HasMDString(false), HasMDLocation(false) { 287 if (shouldPreserveBitcodeUseListOrder()) 288 UseListOrders = predictUseListOrder(M); 289 290 // Enumerate the global variables. 291 for (Module::const_global_iterator I = M.global_begin(), E = M.global_end(); 292 I != E; ++I) 293 EnumerateValue(I); 294 295 // Enumerate the functions. 296 for (Module::const_iterator I = M.begin(), E = M.end(); I != E; ++I) { 297 EnumerateValue(I); 298 EnumerateAttributes(cast<Function>(I)->getAttributes()); 299 } 300 301 // Enumerate the aliases. 302 for (Module::const_alias_iterator I = M.alias_begin(), E = M.alias_end(); 303 I != E; ++I) 304 EnumerateValue(I); 305 306 // Remember what is the cutoff between globalvalue's and other constants. 307 unsigned FirstConstant = Values.size(); 308 309 // Enumerate the global variable initializers. 310 for (Module::const_global_iterator I = M.global_begin(), E = M.global_end(); 311 I != E; ++I) 312 if (I->hasInitializer()) 313 EnumerateValue(I->getInitializer()); 314 315 // Enumerate the aliasees. 316 for (Module::const_alias_iterator I = M.alias_begin(), E = M.alias_end(); 317 I != E; ++I) 318 EnumerateValue(I->getAliasee()); 319 320 // Enumerate the prefix data constants. 321 for (Module::const_iterator I = M.begin(), E = M.end(); I != E; ++I) 322 if (I->hasPrefixData()) 323 EnumerateValue(I->getPrefixData()); 324 325 // Enumerate the prologue data constants. 326 for (Module::const_iterator I = M.begin(), E = M.end(); I != E; ++I) 327 if (I->hasPrologueData()) 328 EnumerateValue(I->getPrologueData()); 329 330 // Enumerate the metadata type. 331 // 332 // TODO: Move this to ValueEnumerator::EnumerateOperandType() once bitcode 333 // only encodes the metadata type when it's used as a value. 334 EnumerateType(Type::getMetadataTy(M.getContext())); 335 336 // Insert constants and metadata that are named at module level into the slot 337 // pool so that the module symbol table can refer to them... 338 EnumerateValueSymbolTable(M.getValueSymbolTable()); 339 EnumerateNamedMetadata(M); 340 341 SmallVector<std::pair<unsigned, MDNode *>, 8> MDs; 342 343 // Enumerate types used by function bodies and argument lists. 344 for (const Function &F : M) { 345 for (const Argument &A : F.args()) 346 EnumerateType(A.getType()); 347 348 for (const BasicBlock &BB : F) 349 for (const Instruction &I : BB) { 350 for (const Use &Op : I.operands()) { 351 auto *MD = dyn_cast<MetadataAsValue>(&Op); 352 if (!MD) { 353 EnumerateOperandType(Op); 354 continue; 355 } 356 357 // Local metadata is enumerated during function-incorporation. 358 if (isa<LocalAsMetadata>(MD->getMetadata())) 359 continue; 360 361 EnumerateMetadata(MD->getMetadata()); 362 } 363 EnumerateType(I.getType()); 364 if (const CallInst *CI = dyn_cast<CallInst>(&I)) 365 EnumerateAttributes(CI->getAttributes()); 366 else if (const InvokeInst *II = dyn_cast<InvokeInst>(&I)) 367 EnumerateAttributes(II->getAttributes()); 368 369 // Enumerate metadata attached with this instruction. 370 MDs.clear(); 371 I.getAllMetadataOtherThanDebugLoc(MDs); 372 for (unsigned i = 0, e = MDs.size(); i != e; ++i) 373 EnumerateMetadata(MDs[i].second); 374 375 if (!I.getDebugLoc().isUnknown()) { 376 MDNode *Scope, *IA; 377 I.getDebugLoc().getScopeAndInlinedAt(Scope, IA, I.getContext()); 378 if (Scope) EnumerateMetadata(Scope); 379 if (IA) EnumerateMetadata(IA); 380 } 381 } 382 } 383 384 // Optimize constant ordering. 385 OptimizeConstants(FirstConstant, Values.size()); 386 } 387 388 unsigned ValueEnumerator::getInstructionID(const Instruction *Inst) const { 389 InstructionMapType::const_iterator I = InstructionMap.find(Inst); 390 assert(I != InstructionMap.end() && "Instruction is not mapped!"); 391 return I->second; 392 } 393 394 unsigned ValueEnumerator::getComdatID(const Comdat *C) const { 395 unsigned ComdatID = Comdats.idFor(C); 396 assert(ComdatID && "Comdat not found!"); 397 return ComdatID; 398 } 399 400 void ValueEnumerator::setInstructionID(const Instruction *I) { 401 InstructionMap[I] = InstructionCount++; 402 } 403 404 unsigned ValueEnumerator::getValueID(const Value *V) const { 405 if (auto *MD = dyn_cast<MetadataAsValue>(V)) 406 return getMetadataID(MD->getMetadata()); 407 408 ValueMapType::const_iterator I = ValueMap.find(V); 409 assert(I != ValueMap.end() && "Value not in slotcalculator!"); 410 return I->second-1; 411 } 412 413 void ValueEnumerator::dump() const { 414 print(dbgs(), ValueMap, "Default"); 415 dbgs() << '\n'; 416 print(dbgs(), MDValueMap, "MetaData"); 417 dbgs() << '\n'; 418 } 419 420 void ValueEnumerator::print(raw_ostream &OS, const ValueMapType &Map, 421 const char *Name) const { 422 423 OS << "Map Name: " << Name << "\n"; 424 OS << "Size: " << Map.size() << "\n"; 425 for (ValueMapType::const_iterator I = Map.begin(), 426 E = Map.end(); I != E; ++I) { 427 428 const Value *V = I->first; 429 if (V->hasName()) 430 OS << "Value: " << V->getName(); 431 else 432 OS << "Value: [null]\n"; 433 V->dump(); 434 435 OS << " Uses(" << std::distance(V->use_begin(),V->use_end()) << "):"; 436 for (const Use &U : V->uses()) { 437 if (&U != &*V->use_begin()) 438 OS << ","; 439 if(U->hasName()) 440 OS << " " << U->getName(); 441 else 442 OS << " [null]"; 443 444 } 445 OS << "\n\n"; 446 } 447 } 448 449 void ValueEnumerator::print(raw_ostream &OS, const MetadataMapType &Map, 450 const char *Name) const { 451 452 OS << "Map Name: " << Name << "\n"; 453 OS << "Size: " << Map.size() << "\n"; 454 for (auto I = Map.begin(), E = Map.end(); I != E; ++I) { 455 const Metadata *MD = I->first; 456 OS << "Metadata: slot = " << I->second << "\n"; 457 MD->print(OS); 458 } 459 } 460 461 /// OptimizeConstants - Reorder constant pool for denser encoding. 462 void ValueEnumerator::OptimizeConstants(unsigned CstStart, unsigned CstEnd) { 463 if (CstStart == CstEnd || CstStart+1 == CstEnd) return; 464 465 if (shouldPreserveBitcodeUseListOrder()) 466 // Optimizing constants makes the use-list order difficult to predict. 467 // Disable it for now when trying to preserve the order. 468 return; 469 470 std::stable_sort(Values.begin() + CstStart, Values.begin() + CstEnd, 471 [this](const std::pair<const Value *, unsigned> &LHS, 472 const std::pair<const Value *, unsigned> &RHS) { 473 // Sort by plane. 474 if (LHS.first->getType() != RHS.first->getType()) 475 return getTypeID(LHS.first->getType()) < getTypeID(RHS.first->getType()); 476 // Then by frequency. 477 return LHS.second > RHS.second; 478 }); 479 480 // Ensure that integer and vector of integer constants are at the start of the 481 // constant pool. This is important so that GEP structure indices come before 482 // gep constant exprs. 483 std::partition(Values.begin()+CstStart, Values.begin()+CstEnd, 484 isIntOrIntVectorValue); 485 486 // Rebuild the modified portion of ValueMap. 487 for (; CstStart != CstEnd; ++CstStart) 488 ValueMap[Values[CstStart].first] = CstStart+1; 489 } 490 491 492 /// EnumerateValueSymbolTable - Insert all of the values in the specified symbol 493 /// table into the values table. 494 void ValueEnumerator::EnumerateValueSymbolTable(const ValueSymbolTable &VST) { 495 for (ValueSymbolTable::const_iterator VI = VST.begin(), VE = VST.end(); 496 VI != VE; ++VI) 497 EnumerateValue(VI->getValue()); 498 } 499 500 /// Insert all of the values referenced by named metadata in the specified 501 /// module. 502 void ValueEnumerator::EnumerateNamedMetadata(const Module &M) { 503 for (Module::const_named_metadata_iterator I = M.named_metadata_begin(), 504 E = M.named_metadata_end(); 505 I != E; ++I) 506 EnumerateNamedMDNode(I); 507 } 508 509 void ValueEnumerator::EnumerateNamedMDNode(const NamedMDNode *MD) { 510 for (unsigned i = 0, e = MD->getNumOperands(); i != e; ++i) 511 EnumerateMetadata(MD->getOperand(i)); 512 } 513 514 /// EnumerateMDNodeOperands - Enumerate all non-function-local values 515 /// and types referenced by the given MDNode. 516 void ValueEnumerator::EnumerateMDNodeOperands(const MDNode *N) { 517 for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) { 518 Metadata *MD = N->getOperand(i); 519 if (!MD) 520 continue; 521 assert(!isa<LocalAsMetadata>(MD) && "MDNodes cannot be function-local"); 522 EnumerateMetadata(MD); 523 } 524 } 525 526 void ValueEnumerator::EnumerateMetadata(const Metadata *MD) { 527 assert( 528 (isa<MDNode>(MD) || isa<MDString>(MD) || isa<ConstantAsMetadata>(MD)) && 529 "Invalid metadata kind"); 530 531 // Insert a dummy ID to block the co-recursive call to 532 // EnumerateMDNodeOperands() from re-visiting MD in a cyclic graph. 533 // 534 // Return early if there's already an ID. 535 if (!MDValueMap.insert(std::make_pair(MD, 0)).second) 536 return; 537 538 // Visit operands first to minimize RAUW. 539 if (auto *N = dyn_cast<MDNode>(MD)) 540 EnumerateMDNodeOperands(N); 541 else if (auto *C = dyn_cast<ConstantAsMetadata>(MD)) 542 EnumerateValue(C->getValue()); 543 544 HasMDString |= isa<MDString>(MD); 545 HasMDLocation |= isa<MDLocation>(MD); 546 547 // Replace the dummy ID inserted above with the correct one. MDValueMap may 548 // have changed by inserting operands, so we need a fresh lookup here. 549 MDs.push_back(MD); 550 MDValueMap[MD] = MDs.size(); 551 } 552 553 /// EnumerateFunctionLocalMetadataa - Incorporate function-local metadata 554 /// information reachable from the metadata. 555 void ValueEnumerator::EnumerateFunctionLocalMetadata( 556 const LocalAsMetadata *Local) { 557 // Check to see if it's already in! 558 unsigned &MDValueID = MDValueMap[Local]; 559 if (MDValueID) 560 return; 561 562 MDs.push_back(Local); 563 MDValueID = MDs.size(); 564 565 EnumerateValue(Local->getValue()); 566 567 // Also, collect all function-local metadata for easy access. 568 FunctionLocalMDs.push_back(Local); 569 } 570 571 void ValueEnumerator::EnumerateValue(const Value *V) { 572 assert(!V->getType()->isVoidTy() && "Can't insert void values!"); 573 assert(!isa<MetadataAsValue>(V) && "EnumerateValue doesn't handle Metadata!"); 574 575 // Check to see if it's already in! 576 unsigned &ValueID = ValueMap[V]; 577 if (ValueID) { 578 // Increment use count. 579 Values[ValueID-1].second++; 580 return; 581 } 582 583 if (auto *GO = dyn_cast<GlobalObject>(V)) 584 if (const Comdat *C = GO->getComdat()) 585 Comdats.insert(C); 586 587 // Enumerate the type of this value. 588 EnumerateType(V->getType()); 589 590 if (const Constant *C = dyn_cast<Constant>(V)) { 591 if (isa<GlobalValue>(C)) { 592 // Initializers for globals are handled explicitly elsewhere. 593 } else if (C->getNumOperands()) { 594 // If a constant has operands, enumerate them. This makes sure that if a 595 // constant has uses (for example an array of const ints), that they are 596 // inserted also. 597 598 // We prefer to enumerate them with values before we enumerate the user 599 // itself. This makes it more likely that we can avoid forward references 600 // in the reader. We know that there can be no cycles in the constants 601 // graph that don't go through a global variable. 602 for (User::const_op_iterator I = C->op_begin(), E = C->op_end(); 603 I != E; ++I) 604 if (!isa<BasicBlock>(*I)) // Don't enumerate BB operand to BlockAddress. 605 EnumerateValue(*I); 606 607 // Finally, add the value. Doing this could make the ValueID reference be 608 // dangling, don't reuse it. 609 Values.push_back(std::make_pair(V, 1U)); 610 ValueMap[V] = Values.size(); 611 return; 612 } 613 } 614 615 // Add the value. 616 Values.push_back(std::make_pair(V, 1U)); 617 ValueID = Values.size(); 618 } 619 620 621 void ValueEnumerator::EnumerateType(Type *Ty) { 622 unsigned *TypeID = &TypeMap[Ty]; 623 624 // We've already seen this type. 625 if (*TypeID) 626 return; 627 628 // If it is a non-anonymous struct, mark the type as being visited so that we 629 // don't recursively visit it. This is safe because we allow forward 630 // references of these in the bitcode reader. 631 if (StructType *STy = dyn_cast<StructType>(Ty)) 632 if (!STy->isLiteral()) 633 *TypeID = ~0U; 634 635 // Enumerate all of the subtypes before we enumerate this type. This ensures 636 // that the type will be enumerated in an order that can be directly built. 637 for (Type *SubTy : Ty->subtypes()) 638 EnumerateType(SubTy); 639 640 // Refresh the TypeID pointer in case the table rehashed. 641 TypeID = &TypeMap[Ty]; 642 643 // Check to see if we got the pointer another way. This can happen when 644 // enumerating recursive types that hit the base case deeper than they start. 645 // 646 // If this is actually a struct that we are treating as forward ref'able, 647 // then emit the definition now that all of its contents are available. 648 if (*TypeID && *TypeID != ~0U) 649 return; 650 651 // Add this type now that its contents are all happily enumerated. 652 Types.push_back(Ty); 653 654 *TypeID = Types.size(); 655 } 656 657 // Enumerate the types for the specified value. If the value is a constant, 658 // walk through it, enumerating the types of the constant. 659 void ValueEnumerator::EnumerateOperandType(const Value *V) { 660 EnumerateType(V->getType()); 661 662 if (auto *MD = dyn_cast<MetadataAsValue>(V)) { 663 assert(!isa<LocalAsMetadata>(MD->getMetadata()) && 664 "Function-local metadata should be left for later"); 665 666 EnumerateMetadata(MD->getMetadata()); 667 return; 668 } 669 670 const Constant *C = dyn_cast<Constant>(V); 671 if (!C) 672 return; 673 674 // If this constant is already enumerated, ignore it, we know its type must 675 // be enumerated. 676 if (ValueMap.count(C)) 677 return; 678 679 // This constant may have operands, make sure to enumerate the types in 680 // them. 681 for (unsigned i = 0, e = C->getNumOperands(); i != e; ++i) { 682 const Value *Op = C->getOperand(i); 683 684 // Don't enumerate basic blocks here, this happens as operands to 685 // blockaddress. 686 if (isa<BasicBlock>(Op)) 687 continue; 688 689 EnumerateOperandType(Op); 690 } 691 } 692 693 void ValueEnumerator::EnumerateAttributes(AttributeSet PAL) { 694 if (PAL.isEmpty()) return; // null is always 0. 695 696 // Do a lookup. 697 unsigned &Entry = AttributeMap[PAL]; 698 if (Entry == 0) { 699 // Never saw this before, add it. 700 Attribute.push_back(PAL); 701 Entry = Attribute.size(); 702 } 703 704 // Do lookups for all attribute groups. 705 for (unsigned i = 0, e = PAL.getNumSlots(); i != e; ++i) { 706 AttributeSet AS = PAL.getSlotAttributes(i); 707 unsigned &Entry = AttributeGroupMap[AS]; 708 if (Entry == 0) { 709 AttributeGroups.push_back(AS); 710 Entry = AttributeGroups.size(); 711 } 712 } 713 } 714 715 void ValueEnumerator::incorporateFunction(const Function &F) { 716 InstructionCount = 0; 717 NumModuleValues = Values.size(); 718 NumModuleMDs = MDs.size(); 719 720 // Adding function arguments to the value table. 721 for (Function::const_arg_iterator I = F.arg_begin(), E = F.arg_end(); 722 I != E; ++I) 723 EnumerateValue(I); 724 725 FirstFuncConstantID = Values.size(); 726 727 // Add all function-level constants to the value table. 728 for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) { 729 for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I) 730 for (User::const_op_iterator OI = I->op_begin(), E = I->op_end(); 731 OI != E; ++OI) { 732 if ((isa<Constant>(*OI) && !isa<GlobalValue>(*OI)) || 733 isa<InlineAsm>(*OI)) 734 EnumerateValue(*OI); 735 } 736 BasicBlocks.push_back(BB); 737 ValueMap[BB] = BasicBlocks.size(); 738 } 739 740 // Optimize the constant layout. 741 OptimizeConstants(FirstFuncConstantID, Values.size()); 742 743 // Add the function's parameter attributes so they are available for use in 744 // the function's instruction. 745 EnumerateAttributes(F.getAttributes()); 746 747 FirstInstID = Values.size(); 748 749 SmallVector<LocalAsMetadata *, 8> FnLocalMDVector; 750 // Add all of the instructions. 751 for (Function::const_iterator BB = F.begin(), E = F.end(); BB != E; ++BB) { 752 for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I!=E; ++I) { 753 for (User::const_op_iterator OI = I->op_begin(), E = I->op_end(); 754 OI != E; ++OI) { 755 if (auto *MD = dyn_cast<MetadataAsValue>(&*OI)) 756 if (auto *Local = dyn_cast<LocalAsMetadata>(MD->getMetadata())) 757 // Enumerate metadata after the instructions they might refer to. 758 FnLocalMDVector.push_back(Local); 759 } 760 761 if (!I->getType()->isVoidTy()) 762 EnumerateValue(I); 763 } 764 } 765 766 // Add all of the function-local metadata. 767 for (unsigned i = 0, e = FnLocalMDVector.size(); i != e; ++i) 768 EnumerateFunctionLocalMetadata(FnLocalMDVector[i]); 769 } 770 771 void ValueEnumerator::purgeFunction() { 772 /// Remove purged values from the ValueMap. 773 for (unsigned i = NumModuleValues, e = Values.size(); i != e; ++i) 774 ValueMap.erase(Values[i].first); 775 for (unsigned i = NumModuleMDs, e = MDs.size(); i != e; ++i) 776 MDValueMap.erase(MDs[i]); 777 for (unsigned i = 0, e = BasicBlocks.size(); i != e; ++i) 778 ValueMap.erase(BasicBlocks[i]); 779 780 Values.resize(NumModuleValues); 781 MDs.resize(NumModuleMDs); 782 BasicBlocks.clear(); 783 FunctionLocalMDs.clear(); 784 } 785 786 static void IncorporateFunctionInfoGlobalBBIDs(const Function *F, 787 DenseMap<const BasicBlock*, unsigned> &IDMap) { 788 unsigned Counter = 0; 789 for (Function::const_iterator BB = F->begin(), E = F->end(); BB != E; ++BB) 790 IDMap[BB] = ++Counter; 791 } 792 793 /// getGlobalBasicBlockID - This returns the function-specific ID for the 794 /// specified basic block. This is relatively expensive information, so it 795 /// should only be used by rare constructs such as address-of-label. 796 unsigned ValueEnumerator::getGlobalBasicBlockID(const BasicBlock *BB) const { 797 unsigned &Idx = GlobalBasicBlockIDs[BB]; 798 if (Idx != 0) 799 return Idx-1; 800 801 IncorporateFunctionInfoGlobalBBIDs(BB->getParent(), GlobalBasicBlockIDs); 802 return getGlobalBasicBlockID(BB); 803 } 804