1 //===- llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.h --------------*- C++ -*-===//
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 contains support for writing Microsoft CodeView debug info.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #ifndef LLVM_LIB_CODEGEN_ASMPRINTER_CODEVIEWDEBUG_H
15 #define LLVM_LIB_CODEGEN_ASMPRINTER_CODEVIEWDEBUG_H
16 
17 #include "DbgEntityHistoryCalculator.h"
18 #include "DebugHandlerBase.h"
19 #include "llvm/ADT/ArrayRef.h"
20 #include "llvm/ADT/DenseMap.h"
21 #include "llvm/ADT/DenseSet.h"
22 #include "llvm/ADT/MapVector.h"
23 #include "llvm/ADT/SetVector.h"
24 #include "llvm/ADT/SmallVector.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   /// Represents the most general definition range.
58   struct LocalVarDefRange {
59     /// Indicates that variable data is stored in memory relative to the
60     /// specified register.
61     int InMemory : 1;
62 
63     /// Offset of variable data in memory.
64     int DataOffset : 31;
65 
66     /// Non-zero if this is a piece of an aggregate.
67     uint16_t IsSubfield : 1;
68 
69     /// Offset into aggregate.
70     uint16_t StructOffset : 15;
71 
72     /// Register containing the data or the register base of the memory
73     /// location containing the data.
74     uint16_t CVRegister;
75 
76     /// Compares all location fields. This includes all fields except the label
77     /// ranges.
78     bool isDifferentLocation(LocalVarDefRange &O) {
79       return InMemory != O.InMemory || DataOffset != O.DataOffset ||
80              IsSubfield != O.IsSubfield || StructOffset != O.StructOffset ||
81              CVRegister != O.CVRegister;
82     }
83 
84     SmallVector<std::pair<const MCSymbol *, const MCSymbol *>, 1> Ranges;
85   };
86 
87   static LocalVarDefRange createDefRangeMem(uint16_t CVRegister, int Offset);
88   static LocalVarDefRange createDefRangeGeneral(uint16_t CVRegister,
89                                                 bool InMemory, int Offset,
90                                                 bool IsSubfield,
91                                                 uint16_t StructOffset);
92 
93   /// Similar to DbgVariable in DwarfDebug, but not dwarf-specific.
94   struct LocalVariable {
95     const DILocalVariable *DIVar = nullptr;
96     SmallVector<LocalVarDefRange, 1> DefRanges;
97     bool UseReferenceType = false;
98   };
99 
100   struct InlineSite {
101     SmallVector<LocalVariable, 1> InlinedLocals;
102     SmallVector<const DILocation *, 1> ChildSites;
103     const DISubprogram *Inlinee = nullptr;
104 
105     /// The ID of the inline site or function used with .cv_loc. Not a type
106     /// index.
107     unsigned SiteFuncId = 0;
108   };
109 
110   // Combines information from DILexicalBlock and LexicalScope.
111   struct LexicalBlock {
112     SmallVector<LocalVariable, 1> Locals;
113     SmallVector<LexicalBlock *, 1> Children;
114     const MCSymbol *Begin;
115     const MCSymbol *End;
116     StringRef Name;
117   };
118 
119   // For each function, store a vector of labels to its instructions, as well as
120   // to the end of the function.
121   struct FunctionInfo {
122     FunctionInfo() = default;
123 
124     // Uncopyable.
125     FunctionInfo(const FunctionInfo &FI) = delete;
126 
127     /// Map from inlined call site to inlined instructions and child inlined
128     /// call sites. Listed in program order.
129     std::unordered_map<const DILocation *, InlineSite> InlineSites;
130 
131     /// Ordered list of top-level inlined call sites.
132     SmallVector<const DILocation *, 1> ChildSites;
133 
134     SmallVector<LocalVariable, 1> Locals;
135 
136     std::unordered_map<const DILexicalBlockBase*, LexicalBlock> LexicalBlocks;
137 
138     // Lexical blocks containing local variables.
139     SmallVector<LexicalBlock *, 1> ChildBlocks;
140 
141     std::vector<std::pair<MCSymbol *, MDNode *>> Annotations;
142 
143     const MCSymbol *Begin = nullptr;
144     const MCSymbol *End = nullptr;
145     unsigned FuncId = 0;
146     unsigned LastFileId = 0;
147     bool HaveLineInfo = false;
148   };
149   FunctionInfo *CurFn = nullptr;
150 
151   // Map used to seperate variables according to the lexical scope they belong
152   // in.  This is populated by recordLocalVariable() before
153   // collectLexicalBlocks() separates the variables between the FunctionInfo
154   // and LexicalBlocks.
155   DenseMap<const LexicalScope *, SmallVector<LocalVariable, 1>> ScopeVariables;
156 
157   /// The set of comdat .debug$S sections that we've seen so far. Each section
158   /// must start with a magic version number that must only be emitted once.
159   /// This set tracks which sections we've already opened.
160   DenseSet<MCSectionCOFF *> ComdatDebugSections;
161 
162   /// Switch to the appropriate .debug$S section for GVSym. If GVSym, the symbol
163   /// of an emitted global value, is in a comdat COFF section, this will switch
164   /// to a new .debug$S section in that comdat. This method ensures that the
165   /// section starts with the magic version number on first use. If GVSym is
166   /// null, uses the main .debug$S section.
167   void switchToDebugSectionForSymbol(const MCSymbol *GVSym);
168 
169   /// The next available function index for use with our .cv_* directives. Not
170   /// to be confused with type indices for LF_FUNC_ID records.
171   unsigned NextFuncId = 0;
172 
173   InlineSite &getInlineSite(const DILocation *InlinedAt,
174                             const DISubprogram *Inlinee);
175 
176   codeview::TypeIndex getFuncIdForSubprogram(const DISubprogram *SP);
177 
178   void calculateRanges(LocalVariable &Var,
179                        const DbgValueHistoryMap::InstrRanges &Ranges);
180 
181   static void collectInlineSiteChildren(SmallVectorImpl<unsigned> &Children,
182                                         const FunctionInfo &FI,
183                                         const InlineSite &Site);
184 
185   /// Remember some debug info about each function. Keep it in a stable order to
186   /// emit at the end of the TU.
187   MapVector<const Function *, std::unique_ptr<FunctionInfo>> FnDebugInfo;
188 
189   /// Map from full file path to .cv_file id. Full paths are built from DIFiles
190   /// and are stored in FileToFilepathMap;
191   DenseMap<StringRef, unsigned> FileIdMap;
192 
193   /// All inlined subprograms in the order they should be emitted.
194   SmallSetVector<const DISubprogram *, 4> InlinedSubprograms;
195 
196   /// Map from a pair of DI metadata nodes and its DI type (or scope) that can
197   /// be nullptr, to CodeView type indices. Primarily indexed by
198   /// {DIType*, DIType*} and {DISubprogram*, DIType*}.
199   ///
200   /// The second entry in the key is needed for methods as DISubroutineType
201   /// representing static method type are shared with non-method function type.
202   DenseMap<std::pair<const DINode *, const DIType *>, codeview::TypeIndex>
203       TypeIndices;
204 
205   /// Map from DICompositeType* to complete type index. Non-record types are
206   /// always looked up in the normal TypeIndices map.
207   DenseMap<const DICompositeType *, codeview::TypeIndex> CompleteTypeIndices;
208 
209   /// Complete record types to emit after all active type lowerings are
210   /// finished.
211   SmallVector<const DICompositeType *, 4> DeferredCompleteTypes;
212 
213   /// Number of type lowering frames active on the stack.
214   unsigned TypeEmissionLevel = 0;
215 
216   codeview::TypeIndex VBPType;
217 
218   const DISubprogram *CurrentSubprogram = nullptr;
219 
220   // The UDTs we have seen while processing types; each entry is a pair of type
221   // index and type name.
222   std::vector<std::pair<std::string, const DIType *>> LocalUDTs;
223   std::vector<std::pair<std::string, const DIType *>> GlobalUDTs;
224 
225   using FileToFilepathMapTy = std::map<const DIFile *, std::string>;
226   FileToFilepathMapTy FileToFilepathMap;
227 
228   StringRef getFullFilepath(const DIFile *File);
229 
230   unsigned maybeRecordFile(const DIFile *F);
231 
232   void maybeRecordLocation(const DebugLoc &DL, const MachineFunction *MF);
233 
234   void clear();
235 
236   void setCurrentSubprogram(const DISubprogram *SP) {
237     CurrentSubprogram = SP;
238     LocalUDTs.clear();
239   }
240 
241   /// Emit the magic version number at the start of a CodeView type or symbol
242   /// section. Appears at the front of every .debug$S or .debug$T or .debug$P
243   /// section.
244   void emitCodeViewMagicVersion();
245 
246   void emitTypeInformation();
247 
248   void emitTypeGlobalHashes();
249 
250   void emitCompilerInformation();
251 
252   void emitInlineeLinesSubsection();
253 
254   void emitDebugInfoForThunk(const Function *GV,
255                              FunctionInfo &FI,
256                              const MCSymbol *Fn);
257 
258   void emitDebugInfoForFunction(const Function *GV, FunctionInfo &FI);
259 
260   void emitDebugInfoForGlobals();
261 
262   void emitDebugInfoForRetainedTypes();
263 
264   void
265   emitDebugInfoForUDTs(ArrayRef<std::pair<std::string, const DIType *>> UDTs);
266 
267   void emitDebugInfoForGlobal(const DIGlobalVariable *DIGV,
268                               const GlobalVariable *GV, MCSymbol *GVSym);
269 
270   /// Opens a subsection of the given kind in a .debug$S codeview section.
271   /// Returns an end label for use with endCVSubsection when the subsection is
272   /// finished.
273   MCSymbol *beginCVSubsection(codeview::DebugSubsectionKind Kind);
274 
275   void endCVSubsection(MCSymbol *EndLabel);
276 
277   void emitInlinedCallSite(const FunctionInfo &FI, const DILocation *InlinedAt,
278                            const InlineSite &Site);
279 
280   using InlinedVariable = DbgValueHistoryMap::InlinedVariable;
281 
282   void collectVariableInfo(const DISubprogram *SP);
283 
284   void collectVariableInfoFromMFTable(DenseSet<InlinedVariable> &Processed);
285 
286   // Construct the lexical block tree for a routine, pruning emptpy lexical
287   // scopes, and populate it with local variables.
288   void collectLexicalBlockInfo(SmallVectorImpl<LexicalScope *> &Scopes,
289                                SmallVectorImpl<LexicalBlock *> &Blocks,
290                                SmallVectorImpl<LocalVariable> &Locals);
291   void collectLexicalBlockInfo(LexicalScope &Scope,
292                                SmallVectorImpl<LexicalBlock *> &ParentBlocks,
293                                SmallVectorImpl<LocalVariable> &ParentLocals);
294 
295   /// Records information about a local variable in the appropriate scope. In
296   /// particular, locals from inlined code live inside the inlining site.
297   void recordLocalVariable(LocalVariable &&Var, const LexicalScope *LS);
298 
299   /// Emits local variables in the appropriate order.
300   void emitLocalVariableList(ArrayRef<LocalVariable> Locals);
301 
302   /// Emits an S_LOCAL record and its associated defined ranges.
303   void emitLocalVariable(const LocalVariable &Var);
304 
305   /// Emits a sequence of lexical block scopes and their children.
306   void emitLexicalBlockList(ArrayRef<LexicalBlock *> Blocks,
307                             const FunctionInfo& FI);
308 
309   /// Emit a lexical block scope and its children.
310   void emitLexicalBlock(const LexicalBlock &Block, const FunctionInfo& FI);
311 
312   /// Translates the DIType to codeview if necessary and returns a type index
313   /// for it.
314   codeview::TypeIndex getTypeIndex(DITypeRef TypeRef,
315                                    DITypeRef ClassTyRef = DITypeRef());
316 
317   codeview::TypeIndex getTypeIndexForReferenceTo(DITypeRef TypeRef);
318 
319   codeview::TypeIndex getMemberFunctionType(const DISubprogram *SP,
320                                             const DICompositeType *Class);
321 
322   codeview::TypeIndex getScopeIndex(const DIScope *Scope);
323 
324   codeview::TypeIndex getVBPTypeIndex();
325 
326   void addToUDTs(const DIType *Ty);
327 
328   void addUDTSrcLine(const DIType *Ty, codeview::TypeIndex TI);
329 
330   codeview::TypeIndex lowerType(const DIType *Ty, const DIType *ClassTy);
331   codeview::TypeIndex lowerTypeAlias(const DIDerivedType *Ty);
332   codeview::TypeIndex lowerTypeArray(const DICompositeType *Ty);
333   codeview::TypeIndex lowerTypeBasic(const DIBasicType *Ty);
334   codeview::TypeIndex lowerTypePointer(
335       const DIDerivedType *Ty,
336       codeview::PointerOptions PO = codeview::PointerOptions::None);
337   codeview::TypeIndex lowerTypeMemberPointer(
338       const DIDerivedType *Ty,
339       codeview::PointerOptions PO = codeview::PointerOptions::None);
340   codeview::TypeIndex lowerTypeModifier(const DIDerivedType *Ty);
341   codeview::TypeIndex lowerTypeFunction(const DISubroutineType *Ty);
342   codeview::TypeIndex lowerTypeVFTableShape(const DIDerivedType *Ty);
343   codeview::TypeIndex lowerTypeMemberFunction(const DISubroutineType *Ty,
344                                               const DIType *ClassTy,
345                                               int ThisAdjustment,
346                                               bool IsStaticMethod);
347   codeview::TypeIndex lowerTypeEnum(const DICompositeType *Ty);
348   codeview::TypeIndex lowerTypeClass(const DICompositeType *Ty);
349   codeview::TypeIndex lowerTypeUnion(const DICompositeType *Ty);
350 
351   /// Symbol records should point to complete types, but type records should
352   /// always point to incomplete types to avoid cycles in the type graph. Only
353   /// use this entry point when generating symbol records. The complete and
354   /// incomplete type indices only differ for record types. All other types use
355   /// the same index.
356   codeview::TypeIndex getCompleteTypeIndex(DITypeRef TypeRef);
357 
358   codeview::TypeIndex lowerCompleteTypeClass(const DICompositeType *Ty);
359   codeview::TypeIndex lowerCompleteTypeUnion(const DICompositeType *Ty);
360 
361   struct TypeLoweringScope;
362 
363   void emitDeferredCompleteTypes();
364 
365   void collectMemberInfo(ClassInfo &Info, const DIDerivedType *DDTy);
366   ClassInfo collectClassInfo(const DICompositeType *Ty);
367 
368   /// Common record member lowering functionality for record types, which are
369   /// structs, classes, and unions. Returns the field list index and the member
370   /// count.
371   std::tuple<codeview::TypeIndex, codeview::TypeIndex, unsigned, bool>
372   lowerRecordFieldList(const DICompositeType *Ty);
373 
374   /// Inserts {{Node, ClassTy}, TI} into TypeIndices and checks for duplicates.
375   codeview::TypeIndex recordTypeIndexForDINode(const DINode *Node,
376                                                codeview::TypeIndex TI,
377                                                const DIType *ClassTy = nullptr);
378 
379   unsigned getPointerSizeInBytes();
380 
381 protected:
382   /// Gather pre-function debug information.
383   void beginFunctionImpl(const MachineFunction *MF) override;
384 
385   /// Gather post-function debug information.
386   void endFunctionImpl(const MachineFunction *) override;
387 
388 public:
389   CodeViewDebug(AsmPrinter *AP);
390 
391   void setSymbolSize(const MCSymbol *, uint64_t) override {}
392 
393   /// Emit the COFF section that holds the line table information.
394   void endModule() override;
395 
396   /// Process beginning of an instruction.
397   void beginInstruction(const MachineInstr *MI) override;
398 };
399 
400 } // end namespace llvm
401 
402 #endif // LLVM_LIB_CODEGEN_ASMPRINTER_CODEVIEWDEBUG_H
403