llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.h

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