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