1 //===- PDB.cpp ------------------------------------------------------------===//
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
14 #include "SymbolTable.h"
16 #include "llvm/DebugInfo/CodeView/CVDebugRecord.h"
17 #include "llvm/DebugInfo/CodeView/CVTypeVisitor.h"
18 #include "llvm/DebugInfo/CodeView/DebugSubsectionRecord.h"
19 #include "llvm/DebugInfo/CodeView/DebugSubsectionVisitor.h"
20 #include "llvm/DebugInfo/CodeView/LazyRandomTypeCollection.h"
21 #include "llvm/DebugInfo/CodeView/TypeDumpVisitor.h"
22 #include "llvm/DebugInfo/CodeView/TypeIndexDiscovery.h"
23 #include "llvm/DebugInfo/CodeView/TypeStreamMerger.h"
24 #include "llvm/DebugInfo/CodeView/TypeTableBuilder.h"
25 #include "llvm/DebugInfo/MSF/MSFBuilder.h"
26 #include "llvm/DebugInfo/MSF/MSFCommon.h"
27 #include "llvm/DebugInfo/PDB/Native/DbiStream.h"
28 #include "llvm/DebugInfo/PDB/Native/DbiStreamBuilder.h"
29 #include "llvm/DebugInfo/PDB/Native/InfoStream.h"
30 #include "llvm/DebugInfo/PDB/Native/InfoStreamBuilder.h"
31 #include "llvm/DebugInfo/PDB/Native/PDBFile.h"
32 #include "llvm/DebugInfo/PDB/Native/PDBFileBuilder.h"
33 #include "llvm/DebugInfo/PDB/Native/DbiModuleDescriptorBuilder.h"
34 #include "llvm/DebugInfo/PDB/Native/PDBStringTableBuilder.h"
35 #include "llvm/DebugInfo/PDB/Native/PDBTypeServerHandler.h"
36 #include "llvm/DebugInfo/PDB/Native/TpiStream.h"
37 #include "llvm/DebugInfo/PDB/Native/TpiStreamBuilder.h"
38 #include "llvm/Object/COFF.h"
39 #include "llvm/Support/BinaryByteStream.h"
40 #include "llvm/Support/Endian.h"
41 #include "llvm/Support/FileOutputBuffer.h"
42 #include "llvm/Support/Path.h"
43 #include "llvm/Support/ScopedPrinter.h"
47 using namespace lld::coff;
49 using namespace llvm::codeview;
51 using llvm::object::coff_section;
53 static ExitOnError ExitOnErr;
55 // Returns a list of all SectionChunks.
56 static void addSectionContribs(SymbolTable *Symtab, pdb::DbiStreamBuilder &DbiBuilder) {
57 for (Chunk *C : Symtab->getChunks())
58 if (auto *SC = dyn_cast<SectionChunk>(C))
59 DbiBuilder.addSectionContrib(SC->File->ModuleDBI, SC->Header);
62 static SectionChunk *findByName(std::vector<SectionChunk *> &Sections,
64 for (SectionChunk *C : Sections)
65 if (C->getSectionName() == Name)
70 static ArrayRef<uint8_t> consumeDebugMagic(ArrayRef<uint8_t> Data,
72 // First 4 bytes are section magic.
74 fatal(SecName + " too short");
75 if (support::endian::read32le(Data.data()) != COFF::DEBUG_SECTION_MAGIC)
76 fatal(SecName + " has an invalid magic");
80 static ArrayRef<uint8_t> getDebugSection(ObjectFile *File, StringRef SecName) {
81 if (SectionChunk *Sec = findByName(File->getDebugChunks(), SecName))
82 return consumeDebugMagic(Sec->getContents(), SecName);
86 static void addTypeInfo(pdb::TpiStreamBuilder &TpiBuilder,
87 TypeTableBuilder &TypeTable) {
88 // Start the TPI or IPI stream header.
89 TpiBuilder.setVersionHeader(pdb::PdbTpiV80);
91 // Flatten the in memory type table.
92 TypeTable.ForEachRecord([&](TypeIndex TI, ArrayRef<uint8_t> Rec) {
94 TpiBuilder.addTypeRecord(Rec, None);
98 static void mergeDebugT(ObjectFile *File,
99 TypeTableBuilder &IDTable,
100 TypeTableBuilder &TypeTable,
101 SmallVectorImpl<TypeIndex> &TypeIndexMap,
102 pdb::PDBTypeServerHandler &Handler) {
103 ArrayRef<uint8_t> Data = getDebugSection(File, ".debug$T");
107 BinaryByteStream Stream(Data, support::little);
109 BinaryStreamReader Reader(Stream);
110 Handler.addSearchPath(sys::path::parent_path(File->getName()));
111 if (auto EC = Reader.readArray(Types, Reader.getLength()))
112 fatal(EC, "Reader::readArray failed");
113 if (auto Err = mergeTypeAndIdRecords(IDTable, TypeTable,
114 TypeIndexMap, &Handler, Types))
115 fatal(Err, "codeview::mergeTypeStreams failed");
118 static bool remapTypeIndex(TypeIndex &TI, ArrayRef<TypeIndex> TypeIndexMap) {
121 if (TI.toArrayIndex() >= TypeIndexMap.size())
123 TI = TypeIndexMap[TI.toArrayIndex()];
127 static bool remapTypesInSymbolRecord(ObjectFile *File,
128 MutableArrayRef<uint8_t> Contents,
129 ArrayRef<TypeIndex> TypeIndexMap,
130 ArrayRef<TiReference> TypeRefs) {
131 for (const TiReference &Ref : TypeRefs) {
132 unsigned ByteSize = Ref.Count * sizeof(TypeIndex);
133 if (Contents.size() < Ref.Offset + ByteSize) {
134 log("ignoring short symbol record");
137 MutableArrayRef<TypeIndex> TIs(
138 reinterpret_cast<TypeIndex *>(Contents.data() + Ref.Offset), Ref.Count);
139 for (TypeIndex &TI : TIs)
140 if (!remapTypeIndex(TI, TypeIndexMap)) {
141 log("ignoring symbol record in " + File->getName() +
142 " with bad type index 0x" + utohexstr(TI.getIndex()));
149 /// MSVC translates S_PROC_ID_END to S_END.
150 uint16_t canonicalizeSymbolKind(SymbolKind Kind) {
151 if (Kind == SymbolKind::S_PROC_ID_END)
152 return SymbolKind::S_END;
156 /// Copy the symbol record. In a PDB, symbol records must be 4 byte aligned.
157 /// The object file may not be aligned.
158 static MutableArrayRef<uint8_t> copySymbolForPdb(const CVSymbol &Sym,
159 BumpPtrAllocator &Alloc) {
160 size_t Size = alignTo(Sym.length(), alignOf(CodeViewContainer::Pdb));
161 assert(Size >= 4 && "record too short");
162 assert(Size <= MaxRecordLength && "record too long");
163 void *Mem = Alloc.Allocate(Size, 4);
165 // Copy the symbol record and zero out any padding bytes.
166 MutableArrayRef<uint8_t> NewData(reinterpret_cast<uint8_t *>(Mem), Size);
167 memcpy(NewData.data(), Sym.data().data(), Sym.length());
168 memset(NewData.data() + Sym.length(), 0, Size - Sym.length());
170 // Update the record prefix length. It should point to the beginning of the
171 // next record. MSVC does some canonicalization of the record kind, so we do
173 auto *Prefix = reinterpret_cast<RecordPrefix *>(Mem);
174 Prefix->RecordKind = canonicalizeSymbolKind(Sym.kind());
175 Prefix->RecordLen = Size - 2;
179 static void mergeSymbolRecords(BumpPtrAllocator &Alloc, ObjectFile *File,
180 ArrayRef<TypeIndex> TypeIndexMap,
181 BinaryStreamRef SymData) {
182 // FIXME: Improve error recovery by warning and skipping records when
185 BinaryStreamReader Reader(SymData);
186 ExitOnErr(Reader.readArray(Syms, Reader.getLength()));
187 for (const CVSymbol &Sym : Syms) {
188 // Discover type index references in the record. Skip it if we don't know
190 SmallVector<TiReference, 32> TypeRefs;
191 if (!discoverTypeIndices(Sym, TypeRefs)) {
192 log("ignoring unknown symbol record with kind 0x" + utohexstr(Sym.kind()));
196 // Copy the symbol record so we can mutate it.
197 MutableArrayRef<uint8_t> NewData = copySymbolForPdb(Sym, Alloc);
199 // Re-map all the type index references.
200 MutableArrayRef<uint8_t> Contents =
201 NewData.drop_front(sizeof(RecordPrefix));
202 if (!remapTypesInSymbolRecord(File, Contents, TypeIndexMap, TypeRefs))
205 // FIXME: Fill in "Parent" and "End" fields by maintaining a stack of
208 // Add the symbol to the module.
209 File->ModuleDBI->addSymbol(CVSymbol(Sym.kind(), NewData));
213 // Allocate memory for a .debug$S section and relocate it.
214 static ArrayRef<uint8_t> relocateDebugChunk(BumpPtrAllocator &Alloc,
215 SectionChunk *DebugChunk) {
216 uint8_t *Buffer = Alloc.Allocate<uint8_t>(DebugChunk->getSize());
217 assert(DebugChunk->OutputSectionOff == 0 &&
218 "debug sections should not be in output sections");
219 DebugChunk->writeTo(Buffer);
220 return consumeDebugMagic(makeArrayRef(Buffer, DebugChunk->getSize()),
224 // Add all object files to the PDB. Merge .debug$T sections into IpiData and
226 static void addObjectsToPDB(BumpPtrAllocator &Alloc, SymbolTable *Symtab,
227 pdb::PDBFileBuilder &Builder,
228 TypeTableBuilder &TypeTable,
229 TypeTableBuilder &IDTable) {
230 // Follow type servers. If the same type server is encountered more than
231 // once for this instance of `PDBTypeServerHandler` (for example if many
232 // object files reference the same TypeServer), the types from the
233 // TypeServer will only be visited once.
234 pdb::PDBTypeServerHandler Handler;
236 // PDBs use a single global string table for filenames in the file checksum
238 auto PDBStrTab = std::make_shared<DebugStringTableSubsection>();
240 // Visit all .debug$T sections to add them to Builder.
241 for (ObjectFile *File : Symtab->ObjectFiles) {
242 // Add a module descriptor for every object file. We need to put an absolute
243 // path to the object into the PDB. If this is a plain object, we make its
244 // path absolute. If it's an object in an archive, we make the archive path
246 bool InArchive = !File->ParentName.empty();
247 SmallString<128> Path = InArchive ? File->ParentName : File->getName();
248 sys::fs::make_absolute(Path);
249 StringRef Name = InArchive ? File->getName() : StringRef(Path);
250 File->ModuleDBI = &ExitOnErr(Builder.getDbiBuilder().addModuleInfo(Name));
251 File->ModuleDBI->setObjFileName(Path);
253 // Before we can process symbol substreams from .debug$S, we need to process
254 // type information, file checksums, and the string table. Add type info to
255 // the PDB first, so that we can get the map from object file type and item
256 // indices to PDB type and item indices.
257 SmallVector<TypeIndex, 128> TypeIndexMap;
258 mergeDebugT(File, IDTable, TypeTable, TypeIndexMap, Handler);
260 // Now do all line info.
261 for (SectionChunk *DebugChunk : File->getDebugChunks()) {
262 if (!DebugChunk->isLive() || DebugChunk->getSectionName() != ".debug$S")
265 ArrayRef<uint8_t> RelocatedDebugContents =
266 relocateDebugChunk(Alloc, DebugChunk);
267 if (RelocatedDebugContents.empty())
270 DebugSubsectionArray Subsections;
271 BinaryStreamReader Reader(RelocatedDebugContents, support::little);
272 ExitOnErr(Reader.readArray(Subsections, RelocatedDebugContents.size()));
274 DebugStringTableSubsectionRef CVStrTab;
275 DebugChecksumsSubsectionRef Checksums;
276 for (const DebugSubsectionRecord &SS : Subsections) {
278 case DebugSubsectionKind::StringTable:
279 ExitOnErr(CVStrTab.initialize(SS.getRecordData()));
281 case DebugSubsectionKind::FileChecksums:
282 ExitOnErr(Checksums.initialize(SS.getRecordData()));
284 case DebugSubsectionKind::Lines:
285 // We can add the relocated line table directly to the PDB without
286 // modification because the file checksum offsets will stay the same.
287 File->ModuleDBI->addDebugSubsection(SS);
289 case DebugSubsectionKind::Symbols:
290 mergeSymbolRecords(Alloc, File, TypeIndexMap, SS.getRecordData());
293 // FIXME: Process the rest of the subsections.
298 if (Checksums.valid()) {
299 // Make a new file checksum table that refers to offsets in the PDB-wide
300 // string table. Generally the string table subsection appears after the
301 // checksum table, so we have to do this after looping over all the
303 if (!CVStrTab.valid())
304 fatal(".debug$S sections must have both a string table subsection "
305 "and a checksum subsection table or neither");
307 make_unique<DebugChecksumsSubsection>(*PDBStrTab);
308 for (FileChecksumEntry &FC : Checksums) {
309 StringRef FileName = ExitOnErr(CVStrTab.getString(FC.FileNameOffset));
310 ExitOnErr(Builder.getDbiBuilder().addModuleSourceFile(
311 *File->ModuleDBI, FileName));
312 NewChecksums->addChecksum(FileName, FC.Kind, FC.Checksum);
314 File->ModuleDBI->addDebugSubsection(std::move(NewChecksums));
319 Builder.getStringTableBuilder().setStrings(*PDBStrTab);
321 // Construct TPI stream contents.
322 addTypeInfo(Builder.getTpiBuilder(), TypeTable);
324 // Construct IPI stream contents.
325 addTypeInfo(Builder.getIpiBuilder(), IDTable);
328 // Creates a PDB file.
329 void coff::createPDB(StringRef Path, SymbolTable *Symtab,
330 ArrayRef<uint8_t> SectionTable,
331 const llvm::codeview::DebugInfo *DI) {
332 BumpPtrAllocator Alloc;
333 pdb::PDBFileBuilder Builder(Alloc);
334 ExitOnErr(Builder.initialize(4096)); // 4096 is blocksize
336 // Create streams in MSF for predefined streams, namely
337 // PDB, TPI, DBI and IPI.
338 for (int I = 0; I < (int)pdb::kSpecialStreamCount; ++I)
339 ExitOnErr(Builder.getMsfBuilder().addStream(0));
341 // Add an Info stream.
342 auto &InfoBuilder = Builder.getInfoBuilder();
343 InfoBuilder.setAge(DI ? DI->PDB70.Age : 0);
345 pdb::PDB_UniqueId uuid{};
347 memcpy(&uuid, &DI->PDB70.Signature, sizeof(uuid));
348 InfoBuilder.setGuid(uuid);
349 // Should be the current time, but set 0 for reproducibilty.
350 InfoBuilder.setSignature(0);
351 InfoBuilder.setVersion(pdb::PdbRaw_ImplVer::PdbImplVC70);
353 // Add an empty DPI stream.
354 pdb::DbiStreamBuilder &DbiBuilder = Builder.getDbiBuilder();
355 DbiBuilder.setVersionHeader(pdb::PdbDbiV110);
357 TypeTableBuilder TypeTable(BAlloc);
358 TypeTableBuilder IDTable(BAlloc);
359 addObjectsToPDB(Alloc, Symtab, Builder, TypeTable, IDTable);
361 // Add Section Contributions.
362 addSectionContribs(Symtab, DbiBuilder);
364 // Add Section Map stream.
365 ArrayRef<object::coff_section> Sections = {
366 (const object::coff_section *)SectionTable.data(),
367 SectionTable.size() / sizeof(object::coff_section)};
368 std::vector<pdb::SecMapEntry> SectionMap =
369 pdb::DbiStreamBuilder::createSectionMap(Sections);
370 DbiBuilder.setSectionMap(SectionMap);
372 ExitOnErr(DbiBuilder.addModuleInfo("* Linker *"));
374 // Add COFF section header stream.
376 DbiBuilder.addDbgStream(pdb::DbgHeaderType::SectionHdr, SectionTable));
379 ExitOnErr(Builder.commit(Path));