InputFiles.cpp revision 293846 
1//===- InputFiles.cpp -----------------------------------------------------===//
2//
3//                             The LLVM Linker
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9
10#include "Chunks.h"
11#include "Error.h"
12#include "InputFiles.h"
13#include "Symbols.h"
14#include "llvm/ADT/STLExtras.h"
15#include "llvm/IR/LLVMContext.h"
16#include "llvm/LTO/LTOModule.h"
17#include "llvm/Object/COFF.h"
18#include "llvm/Support/COFF.h"
19#include "llvm/Support/Debug.h"
20#include "llvm/Support/Endian.h"
21#include "llvm/Support/raw_ostream.h"
22
23using namespace llvm::COFF;
24using namespace llvm::object;
25using namespace llvm::support::endian;
26using llvm::Triple;
27using llvm::support::ulittle32_t;
28using llvm::sys::fs::file_magic;
29using llvm::sys::fs::identify_magic;
30
31namespace lld {
32namespace coff {
33
34int InputFile::NextIndex = 0;
35
36// Returns the last element of a path, which is supposed to be a filename.
37static StringRef getBasename(StringRef Path) {
38  size_t Pos = Path.find_last_of("\\/");
39  if (Pos == StringRef::npos)
40    return Path;
41  return Path.substr(Pos + 1);
42}
43
44// Returns a string in the format of "foo.obj" or "foo.obj(bar.lib)".
45std::string InputFile::getShortName() {
46  if (ParentName == "")
47    return getName().lower();
48  std::string Res = (getBasename(ParentName) + "(" +
49                     getBasename(getName()) + ")").str();
50  return StringRef(Res).lower();
51}
52
53void ArchiveFile::parse() {
54  // Parse a MemoryBufferRef as an archive file.
55  auto ArchiveOrErr = Archive::create(MB);
56  error(ArchiveOrErr, "Failed to parse static library");
57  File = std::move(*ArchiveOrErr);
58
59  // Allocate a buffer for Lazy objects.
60  size_t NumSyms = File->getNumberOfSymbols();
61  LazySymbols.reserve(NumSyms);
62
63  // Read the symbol table to construct Lazy objects.
64  for (const Archive::Symbol &Sym : File->symbols())
65    LazySymbols.emplace_back(this, Sym);
66
67  // Seen is a map from member files to boolean values. Initially
68  // all members are mapped to false, which indicates all these files
69  // are not read yet.
70  for (auto &ChildOrErr : File->children()) {
71    error(ChildOrErr, "Failed to parse static library");
72    const Archive::Child &Child = *ChildOrErr;
73    Seen[Child.getChildOffset()].clear();
74  }
75}
76
77// Returns a buffer pointing to a member file containing a given symbol.
78// This function is thread-safe.
79MemoryBufferRef ArchiveFile::getMember(const Archive::Symbol *Sym) {
80  auto COrErr = Sym->getMember();
81  error(COrErr, Twine("Could not get the member for symbol ") + Sym->getName());
82  const Archive::Child &C = *COrErr;
83
84  // Return an empty buffer if we have already returned the same buffer.
85  if (Seen[C.getChildOffset()].test_and_set())
86    return MemoryBufferRef();
87  ErrorOr<MemoryBufferRef> Ret = C.getMemoryBufferRef();
88  error(Ret, Twine("Could not get the buffer for the member defining symbol ") +
89                 Sym->getName());
90  return *Ret;
91}
92
93void ObjectFile::parse() {
94  // Parse a memory buffer as a COFF file.
95  auto BinOrErr = createBinary(MB);
96  error(BinOrErr, "Failed to parse object file");
97  std::unique_ptr<Binary> Bin = std::move(*BinOrErr);
98
99  if (auto *Obj = dyn_cast<COFFObjectFile>(Bin.get())) {
100    Bin.release();
101    COFFObj.reset(Obj);
102  } else {
103    error(Twine(getName()) + " is not a COFF file.");
104  }
105
106  // Read section and symbol tables.
107  initializeChunks();
108  initializeSymbols();
109  initializeSEH();
110}
111
112void ObjectFile::initializeChunks() {
113  uint32_t NumSections = COFFObj->getNumberOfSections();
114  Chunks.reserve(NumSections);
115  SparseChunks.resize(NumSections + 1);
116  for (uint32_t I = 1; I < NumSections + 1; ++I) {
117    const coff_section *Sec;
118    StringRef Name;
119    std::error_code EC = COFFObj->getSection(I, Sec);
120    error(EC, Twine("getSection failed: #") + Twine(I));
121    EC = COFFObj->getSectionName(Sec, Name);
122    error(EC, Twine("getSectionName failed: #") + Twine(I));
123    if (Name == ".sxdata") {
124      SXData = Sec;
125      continue;
126    }
127    if (Name == ".drectve") {
128      ArrayRef<uint8_t> Data;
129      COFFObj->getSectionContents(Sec, Data);
130      Directives = std::string((const char *)Data.data(), Data.size());
131      continue;
132    }
133    // Skip non-DWARF debug info. MSVC linker converts the sections into
134    // a PDB file, but we don't support that.
135    if (Name == ".debug" || Name.startswith(".debug$"))
136      continue;
137    // We want to preserve DWARF debug sections only when /debug is on.
138    if (!Config->Debug && Name.startswith(".debug"))
139      continue;
140    if (Sec->Characteristics & llvm::COFF::IMAGE_SCN_LNK_REMOVE)
141      continue;
142    auto *C = new (Alloc) SectionChunk(this, Sec);
143    Chunks.push_back(C);
144    SparseChunks[I] = C;
145  }
146}
147
148void ObjectFile::initializeSymbols() {
149  uint32_t NumSymbols = COFFObj->getNumberOfSymbols();
150  SymbolBodies.reserve(NumSymbols);
151  SparseSymbolBodies.resize(NumSymbols);
152  llvm::SmallVector<Undefined *, 8> WeakAliases;
153  int32_t LastSectionNumber = 0;
154  for (uint32_t I = 0; I < NumSymbols; ++I) {
155    // Get a COFFSymbolRef object.
156    auto SymOrErr = COFFObj->getSymbol(I);
157    error(SymOrErr, Twine("broken object file: ") + getName());
158
159    COFFSymbolRef Sym = *SymOrErr;
160
161    const void *AuxP = nullptr;
162    if (Sym.getNumberOfAuxSymbols())
163      AuxP = COFFObj->getSymbol(I + 1)->getRawPtr();
164    bool IsFirst = (LastSectionNumber != Sym.getSectionNumber());
165
166    SymbolBody *Body = nullptr;
167    if (Sym.isUndefined()) {
168      Body = createUndefined(Sym);
169    } else if (Sym.isWeakExternal()) {
170      Body = createWeakExternal(Sym, AuxP);
171      WeakAliases.push_back((Undefined *)Body);
172    } else {
173      Body = createDefined(Sym, AuxP, IsFirst);
174    }
175    if (Body) {
176      SymbolBodies.push_back(Body);
177      SparseSymbolBodies[I] = Body;
178    }
179    I += Sym.getNumberOfAuxSymbols();
180    LastSectionNumber = Sym.getSectionNumber();
181  }
182  for (Undefined *U : WeakAliases)
183    U->WeakAlias = SparseSymbolBodies[(uintptr_t)U->WeakAlias];
184}
185
186Undefined *ObjectFile::createUndefined(COFFSymbolRef Sym) {
187  StringRef Name;
188  COFFObj->getSymbolName(Sym, Name);
189  return new (Alloc) Undefined(Name);
190}
191
192Undefined *ObjectFile::createWeakExternal(COFFSymbolRef Sym, const void *AuxP) {
193  StringRef Name;
194  COFFObj->getSymbolName(Sym, Name);
195  auto *U = new (Alloc) Undefined(Name);
196  auto *Aux = (const coff_aux_weak_external *)AuxP;
197  U->WeakAlias = (Undefined *)(uintptr_t)Aux->TagIndex;
198  return U;
199}
200
201Defined *ObjectFile::createDefined(COFFSymbolRef Sym, const void *AuxP,
202                                   bool IsFirst) {
203  StringRef Name;
204  if (Sym.isCommon()) {
205    auto *C = new (Alloc) CommonChunk(Sym);
206    Chunks.push_back(C);
207    return new (Alloc) DefinedCommon(this, Sym, C);
208  }
209  if (Sym.isAbsolute()) {
210    COFFObj->getSymbolName(Sym, Name);
211    // Skip special symbols.
212    if (Name == "@comp.id")
213      return nullptr;
214    // COFF spec 5.10.1. The .sxdata section.
215    if (Name == "@feat.00") {
216      if (Sym.getValue() & 1)
217        SEHCompat = true;
218      return nullptr;
219    }
220    return new (Alloc) DefinedAbsolute(Name, Sym);
221  }
222  if (Sym.getSectionNumber() == llvm::COFF::IMAGE_SYM_DEBUG)
223    return nullptr;
224
225  // Nothing else to do without a section chunk.
226  auto *SC = cast_or_null<SectionChunk>(SparseChunks[Sym.getSectionNumber()]);
227  if (!SC)
228    return nullptr;
229
230  // Handle section definitions
231  if (IsFirst && AuxP) {
232    auto *Aux = reinterpret_cast<const coff_aux_section_definition *>(AuxP);
233    if (Aux->Selection == IMAGE_COMDAT_SELECT_ASSOCIATIVE)
234      if (auto *ParentSC = cast_or_null<SectionChunk>(
235              SparseChunks[Aux->getNumber(Sym.isBigObj())]))
236        ParentSC->addAssociative(SC);
237    SC->Checksum = Aux->CheckSum;
238  }
239
240  auto *B = new (Alloc) DefinedRegular(this, Sym, SC);
241  if (SC->isCOMDAT() && Sym.getValue() == 0 && !AuxP)
242    SC->setSymbol(B);
243
244  return B;
245}
246
247void ObjectFile::initializeSEH() {
248  if (!SEHCompat || !SXData)
249    return;
250  ArrayRef<uint8_t> A;
251  COFFObj->getSectionContents(SXData, A);
252  if (A.size() % 4 != 0)
253    error(".sxdata must be an array of symbol table indices");
254  auto *I = reinterpret_cast<const ulittle32_t *>(A.data());
255  auto *E = reinterpret_cast<const ulittle32_t *>(A.data() + A.size());
256  for (; I != E; ++I)
257    SEHandlers.insert(SparseSymbolBodies[*I]);
258}
259
260MachineTypes ObjectFile::getMachineType() {
261  if (COFFObj)
262    return static_cast<MachineTypes>(COFFObj->getMachine());
263  return IMAGE_FILE_MACHINE_UNKNOWN;
264}
265
266StringRef ltrim1(StringRef S, const char *Chars) {
267  if (!S.empty() && strchr(Chars, S[0]))
268    return S.substr(1);
269  return S;
270}
271
272void ImportFile::parse() {
273  const char *Buf = MB.getBufferStart();
274  const char *End = MB.getBufferEnd();
275  const auto *Hdr = reinterpret_cast<const coff_import_header *>(Buf);
276
277  // Check if the total size is valid.
278  if ((size_t)(End - Buf) != (sizeof(*Hdr) + Hdr->SizeOfData))
279    error("broken import library");
280
281  // Read names and create an __imp_ symbol.
282  StringRef Name = StringAlloc.save(StringRef(Buf + sizeof(*Hdr)));
283  StringRef ImpName = StringAlloc.save(Twine("__imp_") + Name);
284  const char *NameStart = Buf + sizeof(coff_import_header) + Name.size() + 1;
285  DLLName = StringRef(NameStart);
286  StringRef ExtName;
287  switch (Hdr->getNameType()) {
288  case IMPORT_ORDINAL:
289    ExtName = "";
290    break;
291  case IMPORT_NAME:
292    ExtName = Name;
293    break;
294  case IMPORT_NAME_NOPREFIX:
295    ExtName = ltrim1(Name, "?@_");
296    break;
297  case IMPORT_NAME_UNDECORATE:
298    ExtName = ltrim1(Name, "?@_");
299    ExtName = ExtName.substr(0, ExtName.find('@'));
300    break;
301  }
302  ImpSym = new (Alloc) DefinedImportData(DLLName, ImpName, ExtName, Hdr);
303  SymbolBodies.push_back(ImpSym);
304
305  // If type is function, we need to create a thunk which jump to an
306  // address pointed by the __imp_ symbol. (This allows you to call
307  // DLL functions just like regular non-DLL functions.)
308  if (Hdr->getType() != llvm::COFF::IMPORT_CODE)
309    return;
310  ThunkSym = new (Alloc) DefinedImportThunk(Name, ImpSym, Hdr->Machine);
311  SymbolBodies.push_back(ThunkSym);
312}
313
314void BitcodeFile::parse() {
315  // Usually parse() is thread-safe, but bitcode file is an exception.
316  std::lock_guard<std::mutex> Lock(Mu);
317
318  ErrorOr<std::unique_ptr<LTOModule>> ModOrErr =
319      LTOModule::createFromBuffer(llvm::getGlobalContext(), MB.getBufferStart(),
320                                  MB.getBufferSize(), llvm::TargetOptions());
321  error(ModOrErr, "Could not create lto module");
322  M = std::move(*ModOrErr);
323
324  llvm::StringSaver Saver(Alloc);
325  for (unsigned I = 0, E = M->getSymbolCount(); I != E; ++I) {
326    lto_symbol_attributes Attrs = M->getSymbolAttributes(I);
327    if ((Attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
328      continue;
329
330    StringRef SymName = Saver.save(M->getSymbolName(I));
331    int SymbolDef = Attrs & LTO_SYMBOL_DEFINITION_MASK;
332    if (SymbolDef == LTO_SYMBOL_DEFINITION_UNDEFINED) {
333      SymbolBodies.push_back(new (Alloc) Undefined(SymName));
334    } else {
335      bool Replaceable =
336          (SymbolDef == LTO_SYMBOL_DEFINITION_TENTATIVE || // common
337           (Attrs & LTO_SYMBOL_COMDAT) ||                  // comdat
338           (SymbolDef == LTO_SYMBOL_DEFINITION_WEAK &&     // weak external
339            (Attrs & LTO_SYMBOL_ALIAS)));
340      SymbolBodies.push_back(new (Alloc) DefinedBitcode(this, SymName,
341                                                        Replaceable));
342    }
343  }
344
345  Directives = M->getLinkerOpts();
346}
347
348MachineTypes BitcodeFile::getMachineType() {
349  if (!M)
350    return IMAGE_FILE_MACHINE_UNKNOWN;
351  switch (Triple(M->getTargetTriple()).getArch()) {
352  case Triple::x86_64:
353    return AMD64;
354  case Triple::x86:
355    return I386;
356  case Triple::arm:
357    return ARMNT;
358  default:
359    return IMAGE_FILE_MACHINE_UNKNOWN;
360  }
361}
362
363std::mutex BitcodeFile::Mu;
364
365} // namespace coff
366} // namespace lld
367