1 //===- llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.h --------------*- C++ -*-===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 // 9 // This file contains support for writing Microsoft CodeView debug info. 10 // 11 //===----------------------------------------------------------------------===// 12 13 #ifndef LLVM_LIB_CODEGEN_ASMPRINTER_CODEVIEWDEBUG_H 14 #define LLVM_LIB_CODEGEN_ASMPRINTER_CODEVIEWDEBUG_H 15 16 #include "llvm/ADT/APSInt.h" 17 #include "llvm/ADT/ArrayRef.h" 18 #include "llvm/ADT/DenseMap.h" 19 #include "llvm/ADT/DenseSet.h" 20 #include "llvm/ADT/MapVector.h" 21 #include "llvm/ADT/PointerUnion.h" 22 #include "llvm/ADT/SetVector.h" 23 #include "llvm/ADT/SmallSet.h" 24 #include "llvm/ADT/SmallVector.h" 25 #include "llvm/CodeGen/DbgEntityHistoryCalculator.h" 26 #include "llvm/CodeGen/DebugHandlerBase.h" 27 #include "llvm/CodeGen/MachineJumpTableInfo.h" 28 #include "llvm/DebugInfo/CodeView/CodeView.h" 29 #include "llvm/DebugInfo/CodeView/GlobalTypeTableBuilder.h" 30 #include "llvm/DebugInfo/CodeView/TypeIndex.h" 31 #include "llvm/IR/DebugLoc.h" 32 #include "llvm/Support/Allocator.h" 33 #include "llvm/Support/Compiler.h" 34 #include <cstdint> 35 #include <map> 36 #include <string> 37 #include <tuple> 38 #include <unordered_map> 39 #include <utility> 40 #include <vector> 41 42 namespace llvm { 43 44 struct ClassInfo; 45 class StringRef; 46 class AsmPrinter; 47 class Function; 48 class GlobalVariable; 49 class MCSectionCOFF; 50 class MCStreamer; 51 class MCSymbol; 52 class MachineFunction; 53 54 /// Collects and handles line tables information in a CodeView format. 55 class LLVM_LIBRARY_VISIBILITY CodeViewDebug : public DebugHandlerBase { 56 public: 57 struct LocalVarDef { 58 /// Indicates that variable data is stored in memory relative to the 59 /// specified register. 60 int InMemory : 1; 61 62 /// Offset of variable data in memory. 63 int DataOffset : 31; 64 65 /// Non-zero if this is a piece of an aggregate. 66 uint32_t IsSubfield : 1; 67 68 /// Offset into aggregate. 69 uint32_t StructOffset : 15; 70 71 /// Register containing the data or the register base of the memory 72 /// location containing the data. 73 uint32_t CVRegister : 16; 74 75 uint64_t static toOpaqueValue(const LocalVarDef DR) { 76 uint64_t Val = 0; 77 std::memcpy(&Val, &DR, sizeof(Val)); 78 return Val; 79 } 80 81 LocalVarDef static createFromOpaqueValue(uint64_t Val) { 82 LocalVarDef DR; 83 std::memcpy(&DR, &Val, sizeof(Val)); 84 return DR; 85 } 86 }; 87 88 static_assert(sizeof(uint64_t) == sizeof(LocalVarDef)); 89 90 private: 91 MCStreamer &OS; 92 BumpPtrAllocator Allocator; 93 codeview::GlobalTypeTableBuilder TypeTable; 94 95 /// Whether to emit type record hashes into .debug$H. 96 bool EmitDebugGlobalHashes = false; 97 98 /// The codeview CPU type used by the translation unit. 99 codeview::CPUType TheCPU; 100 101 /// The AsmPrinter used for emitting compiler metadata. When only compiler 102 /// info is being emitted, DebugHandlerBase::Asm may be null. 103 AsmPrinter *CompilerInfoAsm = nullptr; 104 105 static LocalVarDef createDefRangeMem(uint16_t CVRegister, int Offset); 106 107 /// Similar to DbgVariable in DwarfDebug, but not dwarf-specific. 108 struct LocalVariable { 109 const DILocalVariable *DIVar = nullptr; 110 MapVector<LocalVarDef, 111 SmallVector<std::pair<const MCSymbol *, const MCSymbol *>, 1>> 112 DefRanges; 113 bool UseReferenceType = false; 114 std::optional<APSInt> ConstantValue; 115 }; 116 117 struct CVGlobalVariable { 118 const DIGlobalVariable *DIGV; 119 PointerUnion<const GlobalVariable *, const DIExpression *> GVInfo; 120 }; 121 122 struct InlineSite { 123 SmallVector<LocalVariable, 1> InlinedLocals; 124 SmallVector<const DILocation *, 1> ChildSites; 125 const DISubprogram *Inlinee = nullptr; 126 127 /// The ID of the inline site or function used with .cv_loc. Not a type 128 /// index. 129 unsigned SiteFuncId = 0; 130 }; 131 132 // Combines information from DILexicalBlock and LexicalScope. 133 struct LexicalBlock { 134 SmallVector<LocalVariable, 1> Locals; 135 SmallVector<CVGlobalVariable, 1> Globals; 136 SmallVector<LexicalBlock *, 1> Children; 137 const MCSymbol *Begin; 138 const MCSymbol *End; 139 StringRef Name; 140 }; 141 142 struct JumpTableInfo { 143 codeview::JumpTableEntrySize EntrySize; 144 const MCSymbol *Base; 145 uint64_t BaseOffset; 146 const MCSymbol *Branch; 147 const MCSymbol *Table; 148 size_t TableSize; 149 std::vector<const MCSymbol *> Cases; 150 }; 151 152 // For each function, store a vector of labels to its instructions, as well as 153 // to the end of the function. 154 struct FunctionInfo { 155 FunctionInfo() = default; 156 157 // Uncopyable. 158 FunctionInfo(const FunctionInfo &FI) = delete; 159 160 /// Map from inlined call site to inlined instructions and child inlined 161 /// call sites. Listed in program order. 162 std::unordered_map<const DILocation *, InlineSite> InlineSites; 163 164 /// Ordered list of top-level inlined call sites. 165 SmallVector<const DILocation *, 1> ChildSites; 166 167 /// Set of all functions directly inlined into this one. 168 SmallSet<codeview::TypeIndex, 1> Inlinees; 169 170 SmallVector<LocalVariable, 1> Locals; 171 SmallVector<CVGlobalVariable, 1> Globals; 172 173 std::unordered_map<const DILexicalBlockBase*, LexicalBlock> LexicalBlocks; 174 175 // Lexical blocks containing local variables. 176 SmallVector<LexicalBlock *, 1> ChildBlocks; 177 178 std::vector<std::pair<MCSymbol *, MDNode *>> Annotations; 179 std::vector<std::tuple<const MCSymbol *, const MCSymbol *, const DIType *>> 180 HeapAllocSites; 181 182 std::vector<JumpTableInfo> JumpTables; 183 184 const MCSymbol *Begin = nullptr; 185 const MCSymbol *End = nullptr; 186 unsigned FuncId = 0; 187 unsigned LastFileId = 0; 188 189 /// Number of bytes allocated in the prologue for all local stack objects. 190 unsigned FrameSize = 0; 191 192 /// Number of bytes of parameters on the stack. 193 unsigned ParamSize = 0; 194 195 /// Number of bytes pushed to save CSRs. 196 unsigned CSRSize = 0; 197 198 /// Adjustment to apply on x86 when using the VFRAME frame pointer. 199 int OffsetAdjustment = 0; 200 201 /// Two-bit value indicating which register is the designated frame pointer 202 /// register for local variables. Included in S_FRAMEPROC. 203 codeview::EncodedFramePtrReg EncodedLocalFramePtrReg = 204 codeview::EncodedFramePtrReg::None; 205 206 /// Two-bit value indicating which register is the designated frame pointer 207 /// register for stack parameters. Included in S_FRAMEPROC. 208 codeview::EncodedFramePtrReg EncodedParamFramePtrReg = 209 codeview::EncodedFramePtrReg::None; 210 211 codeview::FrameProcedureOptions FrameProcOpts; 212 213 bool HasStackRealignment = false; 214 215 bool HaveLineInfo = false; 216 217 bool HasFramePointer = false; 218 }; 219 FunctionInfo *CurFn = nullptr; 220 221 codeview::SourceLanguage CurrentSourceLanguage = 222 codeview::SourceLanguage::Masm; 223 224 // This map records the constant offset in DIExpression of the 225 // DIGlobalVariableExpression referencing the DIGlobalVariable. 226 DenseMap<const DIGlobalVariable *, uint64_t> CVGlobalVariableOffsets; 227 228 // Map used to separate variables according to the lexical scope they belong 229 // in. This is populated by recordLocalVariable() before 230 // collectLexicalBlocks() separates the variables between the FunctionInfo 231 // and LexicalBlocks. 232 DenseMap<const LexicalScope *, SmallVector<LocalVariable, 1>> ScopeVariables; 233 234 // Map to separate global variables according to the lexical scope they 235 // belong in. A null local scope represents the global scope. 236 typedef SmallVector<CVGlobalVariable, 1> GlobalVariableList; 237 DenseMap<const DIScope*, std::unique_ptr<GlobalVariableList> > ScopeGlobals; 238 239 // Array of global variables which need to be emitted into a COMDAT section. 240 SmallVector<CVGlobalVariable, 1> ComdatVariables; 241 242 // Array of non-COMDAT global variables. 243 SmallVector<CVGlobalVariable, 1> GlobalVariables; 244 245 /// List of static const data members to be emitted as S_CONSTANTs. 246 SmallVector<const DIDerivedType *, 4> StaticConstMembers; 247 248 /// The set of comdat .debug$S sections that we've seen so far. Each section 249 /// must start with a magic version number that must only be emitted once. 250 /// This set tracks which sections we've already opened. 251 DenseSet<MCSectionCOFF *> ComdatDebugSections; 252 253 /// Switch to the appropriate .debug$S section for GVSym. If GVSym, the symbol 254 /// of an emitted global value, is in a comdat COFF section, this will switch 255 /// to a new .debug$S section in that comdat. This method ensures that the 256 /// section starts with the magic version number on first use. If GVSym is 257 /// null, uses the main .debug$S section. 258 void switchToDebugSectionForSymbol(const MCSymbol *GVSym); 259 260 /// The next available function index for use with our .cv_* directives. Not 261 /// to be confused with type indices for LF_FUNC_ID records. 262 unsigned NextFuncId = 0; 263 264 InlineSite &getInlineSite(const DILocation *InlinedAt, 265 const DISubprogram *Inlinee); 266 267 codeview::TypeIndex getFuncIdForSubprogram(const DISubprogram *SP); 268 269 void calculateRanges(LocalVariable &Var, 270 const DbgValueHistoryMap::Entries &Entries); 271 272 /// Remember some debug info about each function. Keep it in a stable order to 273 /// emit at the end of the TU. 274 MapVector<const Function *, std::unique_ptr<FunctionInfo>> FnDebugInfo; 275 276 /// Map from full file path to .cv_file id. Full paths are built from DIFiles 277 /// and are stored in FileToFilepathMap; 278 DenseMap<StringRef, unsigned> FileIdMap; 279 280 /// All inlined subprograms in the order they should be emitted. 281 SmallSetVector<const DISubprogram *, 4> InlinedSubprograms; 282 283 /// Map from a pair of DI metadata nodes and its DI type (or scope) that can 284 /// be nullptr, to CodeView type indices. Primarily indexed by 285 /// {DIType*, DIType*} and {DISubprogram*, DIType*}. 286 /// 287 /// The second entry in the key is needed for methods as DISubroutineType 288 /// representing static method type are shared with non-method function type. 289 DenseMap<std::pair<const DINode *, const DIType *>, codeview::TypeIndex> 290 TypeIndices; 291 292 /// Map from DICompositeType* to complete type index. Non-record types are 293 /// always looked up in the normal TypeIndices map. 294 DenseMap<const DICompositeType *, codeview::TypeIndex> CompleteTypeIndices; 295 296 /// Complete record types to emit after all active type lowerings are 297 /// finished. 298 SmallVector<const DICompositeType *, 4> DeferredCompleteTypes; 299 300 /// Number of type lowering frames active on the stack. 301 unsigned TypeEmissionLevel = 0; 302 303 codeview::TypeIndex VBPType; 304 305 const DISubprogram *CurrentSubprogram = nullptr; 306 307 // The UDTs we have seen while processing types; each entry is a pair of type 308 // index and type name. 309 std::vector<std::pair<std::string, const DIType *>> LocalUDTs; 310 std::vector<std::pair<std::string, const DIType *>> GlobalUDTs; 311 312 using FileToFilepathMapTy = std::map<const DIFile *, std::string>; 313 FileToFilepathMapTy FileToFilepathMap; 314 315 StringRef getFullFilepath(const DIFile *File); 316 317 unsigned maybeRecordFile(const DIFile *F); 318 319 void maybeRecordLocation(const DebugLoc &DL, const MachineFunction *MF); 320 321 void clear(); 322 323 void setCurrentSubprogram(const DISubprogram *SP) { 324 CurrentSubprogram = SP; 325 LocalUDTs.clear(); 326 } 327 328 /// Emit the magic version number at the start of a CodeView type or symbol 329 /// section. Appears at the front of every .debug$S or .debug$T or .debug$P 330 /// section. 331 void emitCodeViewMagicVersion(); 332 333 void emitTypeInformation(); 334 335 void emitTypeGlobalHashes(); 336 337 void emitObjName(); 338 339 void emitCompilerInformation(); 340 341 void emitSecureHotPatchInformation(); 342 343 void emitBuildInfo(); 344 345 void emitInlineeLinesSubsection(); 346 347 void emitDebugInfoForThunk(const Function *GV, 348 FunctionInfo &FI, 349 const MCSymbol *Fn); 350 351 void emitDebugInfoForFunction(const Function *GV, FunctionInfo &FI); 352 353 void emitDebugInfoForRetainedTypes(); 354 355 void emitDebugInfoForUDTs( 356 const std::vector<std::pair<std::string, const DIType *>> &UDTs); 357 358 void collectDebugInfoForGlobals(); 359 void emitDebugInfoForGlobals(); 360 void emitGlobalVariableList(ArrayRef<CVGlobalVariable> Globals); 361 void emitConstantSymbolRecord(const DIType *DTy, APSInt &Value, 362 const std::string &QualifiedName); 363 void emitDebugInfoForGlobal(const CVGlobalVariable &CVGV); 364 void emitStaticConstMemberList(); 365 366 /// Opens a subsection of the given kind in a .debug$S codeview section. 367 /// Returns an end label for use with endCVSubsection when the subsection is 368 /// finished. 369 MCSymbol *beginCVSubsection(codeview::DebugSubsectionKind Kind); 370 void endCVSubsection(MCSymbol *EndLabel); 371 372 /// Opens a symbol record of the given kind. Returns an end label for use with 373 /// endSymbolRecord. 374 MCSymbol *beginSymbolRecord(codeview::SymbolKind Kind); 375 void endSymbolRecord(MCSymbol *SymEnd); 376 377 /// Emits an S_END, S_INLINESITE_END, or S_PROC_ID_END record. These records 378 /// are empty, so we emit them with a simpler assembly sequence that doesn't 379 /// involve labels. 380 void emitEndSymbolRecord(codeview::SymbolKind EndKind); 381 382 void emitInlinedCallSite(const FunctionInfo &FI, const DILocation *InlinedAt, 383 const InlineSite &Site); 384 385 void emitInlinees(const SmallSet<codeview::TypeIndex, 1> &Inlinees); 386 387 using InlinedEntity = DbgValueHistoryMap::InlinedEntity; 388 389 void collectGlobalVariableInfo(); 390 void collectVariableInfo(const DISubprogram *SP); 391 392 void collectVariableInfoFromMFTable(DenseSet<InlinedEntity> &Processed); 393 394 // Construct the lexical block tree for a routine, pruning emptpy lexical 395 // scopes, and populate it with local variables. 396 void collectLexicalBlockInfo(SmallVectorImpl<LexicalScope *> &Scopes, 397 SmallVectorImpl<LexicalBlock *> &Blocks, 398 SmallVectorImpl<LocalVariable> &Locals, 399 SmallVectorImpl<CVGlobalVariable> &Globals); 400 void collectLexicalBlockInfo(LexicalScope &Scope, 401 SmallVectorImpl<LexicalBlock *> &ParentBlocks, 402 SmallVectorImpl<LocalVariable> &ParentLocals, 403 SmallVectorImpl<CVGlobalVariable> &ParentGlobals); 404 405 /// Records information about a local variable in the appropriate scope. In 406 /// particular, locals from inlined code live inside the inlining site. 407 void recordLocalVariable(LocalVariable &&Var, const LexicalScope *LS); 408 409 /// Emits local variables in the appropriate order. 410 void emitLocalVariableList(const FunctionInfo &FI, 411 ArrayRef<LocalVariable> Locals); 412 413 /// Emits an S_LOCAL record and its associated defined ranges. 414 void emitLocalVariable(const FunctionInfo &FI, const LocalVariable &Var); 415 416 /// Emits a sequence of lexical block scopes and their children. 417 void emitLexicalBlockList(ArrayRef<LexicalBlock *> Blocks, 418 const FunctionInfo& FI); 419 420 /// Emit a lexical block scope and its children. 421 void emitLexicalBlock(const LexicalBlock &Block, const FunctionInfo& FI); 422 423 /// Translates the DIType to codeview if necessary and returns a type index 424 /// for it. 425 codeview::TypeIndex getTypeIndex(const DIType *Ty, 426 const DIType *ClassTy = nullptr); 427 428 codeview::TypeIndex 429 getTypeIndexForThisPtr(const DIDerivedType *PtrTy, 430 const DISubroutineType *SubroutineTy); 431 432 codeview::TypeIndex getTypeIndexForReferenceTo(const DIType *Ty); 433 434 codeview::TypeIndex getMemberFunctionType(const DISubprogram *SP, 435 const DICompositeType *Class); 436 437 codeview::TypeIndex getScopeIndex(const DIScope *Scope); 438 439 codeview::TypeIndex getVBPTypeIndex(); 440 441 void addToUDTs(const DIType *Ty); 442 443 void addUDTSrcLine(const DIType *Ty, codeview::TypeIndex TI); 444 445 codeview::TypeIndex lowerType(const DIType *Ty, const DIType *ClassTy); 446 codeview::TypeIndex lowerTypeAlias(const DIDerivedType *Ty); 447 codeview::TypeIndex lowerTypeArray(const DICompositeType *Ty); 448 codeview::TypeIndex lowerTypeString(const DIStringType *Ty); 449 codeview::TypeIndex lowerTypeBasic(const DIBasicType *Ty); 450 codeview::TypeIndex lowerTypePointer( 451 const DIDerivedType *Ty, 452 codeview::PointerOptions PO = codeview::PointerOptions::None); 453 codeview::TypeIndex lowerTypeMemberPointer( 454 const DIDerivedType *Ty, 455 codeview::PointerOptions PO = codeview::PointerOptions::None); 456 codeview::TypeIndex lowerTypeModifier(const DIDerivedType *Ty); 457 codeview::TypeIndex lowerTypeFunction(const DISubroutineType *Ty); 458 codeview::TypeIndex lowerTypeVFTableShape(const DIDerivedType *Ty); 459 codeview::TypeIndex lowerTypeMemberFunction( 460 const DISubroutineType *Ty, const DIType *ClassTy, int ThisAdjustment, 461 bool IsStaticMethod, 462 codeview::FunctionOptions FO = codeview::FunctionOptions::None); 463 codeview::TypeIndex lowerTypeEnum(const DICompositeType *Ty); 464 codeview::TypeIndex lowerTypeClass(const DICompositeType *Ty); 465 codeview::TypeIndex lowerTypeUnion(const DICompositeType *Ty); 466 467 /// Symbol records should point to complete types, but type records should 468 /// always point to incomplete types to avoid cycles in the type graph. Only 469 /// use this entry point when generating symbol records. The complete and 470 /// incomplete type indices only differ for record types. All other types use 471 /// the same index. 472 codeview::TypeIndex getCompleteTypeIndex(const DIType *Ty); 473 474 codeview::TypeIndex lowerCompleteTypeClass(const DICompositeType *Ty); 475 codeview::TypeIndex lowerCompleteTypeUnion(const DICompositeType *Ty); 476 477 struct TypeLoweringScope; 478 479 void emitDeferredCompleteTypes(); 480 481 void collectMemberInfo(ClassInfo &Info, const DIDerivedType *DDTy); 482 ClassInfo collectClassInfo(const DICompositeType *Ty); 483 484 /// Common record member lowering functionality for record types, which are 485 /// structs, classes, and unions. Returns the field list index and the member 486 /// count. 487 std::tuple<codeview::TypeIndex, codeview::TypeIndex, unsigned, bool> 488 lowerRecordFieldList(const DICompositeType *Ty); 489 490 /// Inserts {{Node, ClassTy}, TI} into TypeIndices and checks for duplicates. 491 codeview::TypeIndex recordTypeIndexForDINode(const DINode *Node, 492 codeview::TypeIndex TI, 493 const DIType *ClassTy = nullptr); 494 495 /// Collect the names of parent scopes, innermost to outermost. Return the 496 /// innermost subprogram scope if present. Ensure that parent type scopes are 497 /// inserted into the type table. 498 const DISubprogram * 499 collectParentScopeNames(const DIScope *Scope, 500 SmallVectorImpl<StringRef> &ParentScopeNames); 501 std::string getFullyQualifiedName(const DIScope *Scope, StringRef Name); 502 std::string getFullyQualifiedName(const DIScope *Scope); 503 504 unsigned getPointerSizeInBytes(); 505 506 void discoverJumpTableBranches(const MachineFunction *MF, bool isThumb); 507 void collectDebugInfoForJumpTables(const MachineFunction *MF, bool isThumb); 508 void emitDebugInfoForJumpTables(const FunctionInfo &FI); 509 510 protected: 511 /// Gather pre-function debug information. 512 void beginFunctionImpl(const MachineFunction *MF) override; 513 514 /// Gather post-function debug information. 515 void endFunctionImpl(const MachineFunction *) override; 516 517 /// Check if the current module is in Fortran. 518 bool moduleIsInFortran() { 519 return CurrentSourceLanguage == codeview::SourceLanguage::Fortran; 520 } 521 522 public: 523 CodeViewDebug(AsmPrinter *AP); 524 525 void beginModule(Module *M) override; 526 527 /// Emit the COFF section that holds the line table information. 528 void endModule() override; 529 530 /// Process beginning of an instruction. 531 void beginInstruction(const MachineInstr *MI) override; 532 }; 533 534 template <> struct DenseMapInfo<CodeViewDebug::LocalVarDef> { 535 536 static inline CodeViewDebug::LocalVarDef getEmptyKey() { 537 return CodeViewDebug::LocalVarDef::createFromOpaqueValue(~0ULL); 538 } 539 540 static inline CodeViewDebug::LocalVarDef getTombstoneKey() { 541 return CodeViewDebug::LocalVarDef::createFromOpaqueValue(~0ULL - 1ULL); 542 } 543 544 static unsigned getHashValue(const CodeViewDebug::LocalVarDef &DR) { 545 return CodeViewDebug::LocalVarDef::toOpaqueValue(DR) * 37ULL; 546 } 547 548 static bool isEqual(const CodeViewDebug::LocalVarDef &LHS, 549 const CodeViewDebug::LocalVarDef &RHS) { 550 return CodeViewDebug::LocalVarDef::toOpaqueValue(LHS) == 551 CodeViewDebug::LocalVarDef::toOpaqueValue(RHS); 552 } 553 }; 554 555 } // end namespace llvm 556 557 #endif // LLVM_LIB_CODEGEN_ASMPRINTER_CODEVIEWDEBUG_H 558