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