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 "InputFiles.h"
11 #include "Chunks.h"
12 #include "Config.h"
13 #include "Driver.h"
14 #include "Error.h"
15 #include "Memory.h"
16 #include "SymbolTable.h"
17 #include "Symbols.h"
18 #include "llvm-c/lto.h"
19 #include "llvm/ADT/SmallVector.h"
20 #include "llvm/ADT/Triple.h"
21 #include "llvm/ADT/Twine.h"
22 #include "llvm/Object/Binary.h"
23 #include "llvm/Object/COFF.h"
24 #include "llvm/Support/COFF.h"
25 #include "llvm/Support/Casting.h"
26 #include "llvm/Support/Endian.h"
27 #include "llvm/Support/Error.h"
28 #include "llvm/Support/ErrorOr.h"
29 #include "llvm/Support/FileSystem.h"
30 #include "llvm/Target/TargetOptions.h"
31 #include <cstring>
32 #include <system_error>
33 #include <utility>
34 
35 using namespace llvm;
36 using namespace llvm::COFF;
37 using namespace llvm::object;
38 using namespace llvm::support::endian;
39 
40 using llvm::Triple;
41 using llvm::support::ulittle32_t;
42 
43 namespace lld {
44 namespace coff {
45 
46 /// Checks that Source is compatible with being a weak alias to Target.
47 /// If Source is Undefined and has no weak alias set, makes it a weak
48 /// alias to Target.
49 static void checkAndSetWeakAlias(SymbolTable *Symtab, InputFile *F,
50                                  SymbolBody *Source, SymbolBody *Target) {
51   auto *U = dyn_cast<Undefined>(Source);
52   if (!U)
53     return;
54   else if (!U->WeakAlias)
55     U->WeakAlias = Target;
56   else if (U->WeakAlias != Target)
57     Symtab->reportDuplicate(Source->symbol(), F);
58 }
59 
60 ArchiveFile::ArchiveFile(MemoryBufferRef M) : InputFile(ArchiveKind, M) {}
61 
62 void ArchiveFile::parse() {
63   // Parse a MemoryBufferRef as an archive file.
64   File = check(Archive::create(MB), toString(this));
65 
66   // Read the symbol table to construct Lazy objects.
67   for (const Archive::Symbol &Sym : File->symbols())
68     Symtab->addLazy(this, Sym);
69 }
70 
71 // Returns a buffer pointing to a member file containing a given symbol.
72 void ArchiveFile::addMember(const Archive::Symbol *Sym) {
73   const Archive::Child &C =
74       check(Sym->getMember(),
75             "could not get the member for symbol " + Sym->getName());
76 
77   // Return an empty buffer if we have already returned the same buffer.
78   if (!Seen.insert(C.getChildOffset()).second)
79     return;
80 
81   Driver->enqueueArchiveMember(C, Sym->getName(), getName());
82 }
83 
84 void ObjectFile::parse() {
85   // Parse a memory buffer as a COFF file.
86   std::unique_ptr<Binary> Bin = check(createBinary(MB), toString(this));
87 
88   if (auto *Obj = dyn_cast<COFFObjectFile>(Bin.get())) {
89     Bin.release();
90     COFFObj.reset(Obj);
91   } else {
92     fatal(toString(this) + " is not a COFF file");
93   }
94 
95   // Read section and symbol tables.
96   initializeChunks();
97   initializeSymbols();
98   initializeSEH();
99 }
100 
101 void ObjectFile::initializeChunks() {
102   uint32_t NumSections = COFFObj->getNumberOfSections();
103   Chunks.reserve(NumSections);
104   SparseChunks.resize(NumSections + 1);
105   for (uint32_t I = 1; I < NumSections + 1; ++I) {
106     const coff_section *Sec;
107     StringRef Name;
108     if (auto EC = COFFObj->getSection(I, Sec))
109       fatal(EC, "getSection failed: #" + Twine(I));
110     if (auto EC = COFFObj->getSectionName(Sec, Name))
111       fatal(EC, "getSectionName failed: #" + Twine(I));
112     if (Name == ".sxdata") {
113       SXData = Sec;
114       continue;
115     }
116     if (Name == ".drectve") {
117       ArrayRef<uint8_t> Data;
118       COFFObj->getSectionContents(Sec, Data);
119       Directives = std::string((const char *)Data.data(), Data.size());
120       continue;
121     }
122 
123     // Object files may have DWARF debug info or MS CodeView debug info
124     // (or both).
125     //
126     // DWARF sections don't need any special handling from the perspective
127     // of the linker; they are just a data section containing relocations.
128     // We can just link them to complete debug info.
129     //
130     // CodeView needs a linker support. We need to interpret and debug
131     // info, and then write it to a separate .pdb file.
132 
133     // Ignore debug info unless /debug is given.
134     if (!Config->Debug && Name.startswith(".debug"))
135       continue;
136 
137     // CodeView sections are stored to a different vector because they are
138     // not linked in the regular manner.
139     if (Name == ".debug" || Name.startswith(".debug$")) {
140       DebugChunks.push_back(new (Alloc) SectionChunk(this, Sec));
141       continue;
142     }
143 
144     if (Sec->Characteristics & llvm::COFF::IMAGE_SCN_LNK_REMOVE)
145       continue;
146     auto *C = new (Alloc) SectionChunk(this, Sec);
147     Chunks.push_back(C);
148     SparseChunks[I] = C;
149   }
150 }
151 
152 void ObjectFile::initializeSymbols() {
153   uint32_t NumSymbols = COFFObj->getNumberOfSymbols();
154   SymbolBodies.reserve(NumSymbols);
155   SparseSymbolBodies.resize(NumSymbols);
156   SmallVector<std::pair<SymbolBody *, uint32_t>, 8> WeakAliases;
157   int32_t LastSectionNumber = 0;
158   for (uint32_t I = 0; I < NumSymbols; ++I) {
159     // Get a COFFSymbolRef object.
160     ErrorOr<COFFSymbolRef> SymOrErr = COFFObj->getSymbol(I);
161     if (!SymOrErr)
162       fatal(SymOrErr.getError(), "broken object file: " + toString(this));
163     COFFSymbolRef Sym = *SymOrErr;
164 
165     const void *AuxP = nullptr;
166     if (Sym.getNumberOfAuxSymbols())
167       AuxP = COFFObj->getSymbol(I + 1)->getRawPtr();
168     bool IsFirst = (LastSectionNumber != Sym.getSectionNumber());
169 
170     SymbolBody *Body = nullptr;
171     if (Sym.isUndefined()) {
172       Body = createUndefined(Sym);
173     } else if (Sym.isWeakExternal()) {
174       Body = createUndefined(Sym);
175       uint32_t TagIndex =
176           static_cast<const coff_aux_weak_external *>(AuxP)->TagIndex;
177       WeakAliases.emplace_back(Body, TagIndex);
178     } else {
179       Body = createDefined(Sym, AuxP, IsFirst);
180     }
181     if (Body) {
182       SymbolBodies.push_back(Body);
183       SparseSymbolBodies[I] = Body;
184     }
185     I += Sym.getNumberOfAuxSymbols();
186     LastSectionNumber = Sym.getSectionNumber();
187   }
188   for (auto WeakAlias : WeakAliases)
189     checkAndSetWeakAlias(Symtab, this, WeakAlias.first,
190                          SparseSymbolBodies[WeakAlias.second]);
191 }
192 
193 SymbolBody *ObjectFile::createUndefined(COFFSymbolRef Sym) {
194   StringRef Name;
195   COFFObj->getSymbolName(Sym, Name);
196   return Symtab->addUndefined(Name, this, Sym.isWeakExternal())->body();
197 }
198 
199 SymbolBody *ObjectFile::createDefined(COFFSymbolRef Sym, const void *AuxP,
200                                       bool IsFirst) {
201   StringRef Name;
202   if (Sym.isCommon()) {
203     auto *C = new (Alloc) CommonChunk(Sym);
204     Chunks.push_back(C);
205     COFFObj->getSymbolName(Sym, Name);
206     Symbol *S =
207         Symtab->addCommon(this, Name, Sym.getValue(), Sym.getGeneric(), C);
208     return S->body();
209   }
210   if (Sym.isAbsolute()) {
211     COFFObj->getSymbolName(Sym, Name);
212     // Skip special symbols.
213     if (Name == "@comp.id")
214       return nullptr;
215     // COFF spec 5.10.1. The .sxdata section.
216     if (Name == "@feat.00") {
217       if (Sym.getValue() & 1)
218         SEHCompat = true;
219       return nullptr;
220     }
221     if (Sym.isExternal())
222       return Symtab->addAbsolute(Name, Sym)->body();
223     else
224       return new (Alloc) DefinedAbsolute(Name, Sym);
225   }
226   int32_t SectionNumber = Sym.getSectionNumber();
227   if (SectionNumber == llvm::COFF::IMAGE_SYM_DEBUG)
228     return nullptr;
229 
230   // Reserved sections numbers don't have contents.
231   if (llvm::COFF::isReservedSectionNumber(SectionNumber))
232     fatal("broken object file: " + toString(this));
233 
234   // This symbol references a section which is not present in the section
235   // header.
236   if ((uint32_t)SectionNumber >= SparseChunks.size())
237     fatal("broken object file: " + toString(this));
238 
239   // Nothing else to do without a section chunk.
240   auto *SC = cast_or_null<SectionChunk>(SparseChunks[SectionNumber]);
241   if (!SC)
242     return nullptr;
243 
244   // Handle section definitions
245   if (IsFirst && AuxP) {
246     auto *Aux = reinterpret_cast<const coff_aux_section_definition *>(AuxP);
247     if (Aux->Selection == IMAGE_COMDAT_SELECT_ASSOCIATIVE)
248       if (auto *ParentSC = cast_or_null<SectionChunk>(
249               SparseChunks[Aux->getNumber(Sym.isBigObj())]))
250         ParentSC->addAssociative(SC);
251     SC->Checksum = Aux->CheckSum;
252   }
253 
254   DefinedRegular *B;
255   if (Sym.isExternal()) {
256     COFFObj->getSymbolName(Sym, Name);
257     Symbol *S =
258         Symtab->addRegular(this, Name, SC->isCOMDAT(), Sym.getGeneric(), SC);
259     B = cast<DefinedRegular>(S->body());
260   } else
261     B = new (Alloc) DefinedRegular(this, /*Name*/ "", SC->isCOMDAT(),
262                                    /*IsExternal*/ false, Sym.getGeneric(), SC);
263   if (SC->isCOMDAT() && Sym.getValue() == 0 && !AuxP)
264     SC->setSymbol(B);
265 
266   return B;
267 }
268 
269 void ObjectFile::initializeSEH() {
270   if (!SEHCompat || !SXData)
271     return;
272   ArrayRef<uint8_t> A;
273   COFFObj->getSectionContents(SXData, A);
274   if (A.size() % 4 != 0)
275     fatal(".sxdata must be an array of symbol table indices");
276   auto *I = reinterpret_cast<const ulittle32_t *>(A.data());
277   auto *E = reinterpret_cast<const ulittle32_t *>(A.data() + A.size());
278   for (; I != E; ++I)
279     SEHandlers.insert(SparseSymbolBodies[*I]);
280 }
281 
282 MachineTypes ObjectFile::getMachineType() {
283   if (COFFObj)
284     return static_cast<MachineTypes>(COFFObj->getMachine());
285   return IMAGE_FILE_MACHINE_UNKNOWN;
286 }
287 
288 StringRef ltrim1(StringRef S, const char *Chars) {
289   if (!S.empty() && strchr(Chars, S[0]))
290     return S.substr(1);
291   return S;
292 }
293 
294 void ImportFile::parse() {
295   const char *Buf = MB.getBufferStart();
296   const char *End = MB.getBufferEnd();
297   const auto *Hdr = reinterpret_cast<const coff_import_header *>(Buf);
298 
299   // Check if the total size is valid.
300   if ((size_t)(End - Buf) != (sizeof(*Hdr) + Hdr->SizeOfData))
301     fatal("broken import library");
302 
303   // Read names and create an __imp_ symbol.
304   StringRef Name = StringAlloc.save(StringRef(Buf + sizeof(*Hdr)));
305   StringRef ImpName = StringAlloc.save("__imp_" + Name);
306   const char *NameStart = Buf + sizeof(coff_import_header) + Name.size() + 1;
307   DLLName = StringRef(NameStart);
308   StringRef ExtName;
309   switch (Hdr->getNameType()) {
310   case IMPORT_ORDINAL:
311     ExtName = "";
312     break;
313   case IMPORT_NAME:
314     ExtName = Name;
315     break;
316   case IMPORT_NAME_NOPREFIX:
317     ExtName = ltrim1(Name, "?@_");
318     break;
319   case IMPORT_NAME_UNDECORATE:
320     ExtName = ltrim1(Name, "?@_");
321     ExtName = ExtName.substr(0, ExtName.find('@'));
322     break;
323   }
324 
325   this->Hdr = Hdr;
326   ExternalName = ExtName;
327 
328   ImpSym = cast<DefinedImportData>(
329       Symtab->addImportData(ImpName, this)->body());
330 
331   // If type is function, we need to create a thunk which jump to an
332   // address pointed by the __imp_ symbol. (This allows you to call
333   // DLL functions just like regular non-DLL functions.)
334   if (Hdr->getType() != llvm::COFF::IMPORT_CODE)
335     return;
336   ThunkSym = cast<DefinedImportThunk>(
337       Symtab->addImportThunk(Name, ImpSym, Hdr->Machine)->body());
338 }
339 
340 void BitcodeFile::parse() {
341   Obj = check(lto::InputFile::create(MemoryBufferRef(
342       MB.getBuffer(), Saver.save(ParentName + MB.getBufferIdentifier()))));
343   for (const lto::InputFile::Symbol &ObjSym : Obj->symbols()) {
344     StringRef SymName = Saver.save(ObjSym.getName());
345     Symbol *Sym;
346     if (ObjSym.isUndefined()) {
347       Sym = Symtab->addUndefined(SymName, this, false);
348     } else if (ObjSym.isCommon()) {
349       Sym = Symtab->addCommon(this, SymName, ObjSym.getCommonSize());
350     } else if (ObjSym.isWeak() && ObjSym.isIndirect()) {
351       // Weak external.
352       Sym = Symtab->addUndefined(SymName, this, true);
353       std::string Fallback = ObjSym.getCOFFWeakExternalFallback();
354       SymbolBody *Alias = Symtab->addUndefined(Saver.save(Fallback));
355       checkAndSetWeakAlias(Symtab, this, Sym->body(), Alias);
356     } else {
357       bool IsCOMDAT = ObjSym.getComdatIndex() != -1;
358       Sym = Symtab->addRegular(this, SymName, IsCOMDAT);
359     }
360     SymbolBodies.push_back(Sym->body());
361   }
362   Directives = Obj->getCOFFLinkerOpts();
363 }
364 
365 MachineTypes BitcodeFile::getMachineType() {
366   switch (Triple(Obj->getTargetTriple()).getArch()) {
367   case Triple::x86_64:
368     return AMD64;
369   case Triple::x86:
370     return I386;
371   case Triple::arm:
372     return ARMNT;
373   default:
374     return IMAGE_FILE_MACHINE_UNKNOWN;
375   }
376 }
377 } // namespace coff
378 } // namespace lld
379 
380 // Returns the last element of a path, which is supposed to be a filename.
381 static StringRef getBasename(StringRef Path) {
382   size_t Pos = Path.find_last_of("\\/");
383   if (Pos == StringRef::npos)
384     return Path;
385   return Path.substr(Pos + 1);
386 }
387 
388 // Returns a string in the format of "foo.obj" or "foo.obj(bar.lib)".
389 std::string lld::toString(coff::InputFile *File) {
390   if (!File)
391     return "(internal)";
392   if (File->ParentName.empty())
393     return File->getName().lower();
394 
395   std::string Res =
396       (getBasename(File->ParentName) + "(" + getBasename(File->getName()) + ")")
397           .str();
398   return StringRef(Res).lower();
399 }
400