1 //===- SymbolTable.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 // Symbol table is a bag of all known symbols. We put all symbols of
11 // all input files to the symbol table. The symbol table is basically
12 // a hash table with the logic to resolve symbol name conflicts using
13 // the symbol types.
14 //
15 //===----------------------------------------------------------------------===//
16 
17 #include "SymbolTable.h"
18 #include "Config.h"
19 #include "Error.h"
20 #include "LinkerScript.h"
21 #include "Memory.h"
22 #include "Symbols.h"
23 #include "llvm/ADT/STLExtras.h"
24 
25 using namespace llvm;
26 using namespace llvm::object;
27 using namespace llvm::ELF;
28 
29 using namespace lld;
30 using namespace lld::elf;
31 
32 SymbolTable *elf::Symtab;
33 
34 static InputFile *getFirstElf() {
35   if (!ObjectFiles.empty())
36     return ObjectFiles[0];
37   if (!SharedFiles.empty())
38     return SharedFiles[0];
39   return nullptr;
40 }
41 
42 // All input object files must be for the same architecture
43 // (e.g. it does not make sense to link x86 object files with
44 // MIPS object files.) This function checks for that error.
45 template <class ELFT> static bool isCompatible(InputFile *F) {
46   if (!isa<ELFFileBase<ELFT>>(F) && !isa<BitcodeFile>(F))
47     return true;
48 
49   if (F->EKind == Config->EKind && F->EMachine == Config->EMachine) {
50     if (Config->EMachine != EM_MIPS)
51       return true;
52     if (isMipsN32Abi(F) == Config->MipsN32Abi)
53       return true;
54   }
55 
56   if (!Config->Emulation.empty())
57     error(toString(F) + " is incompatible with " + Config->Emulation);
58   else
59     error(toString(F) + " is incompatible with " + toString(getFirstElf()));
60   return false;
61 }
62 
63 // Add symbols in File to the symbol table.
64 template <class ELFT> void SymbolTable::addFile(InputFile *File) {
65   if (!isCompatible<ELFT>(File))
66     return;
67 
68   // Binary file
69   if (auto *F = dyn_cast<BinaryFile>(File)) {
70     BinaryFiles.push_back(F);
71     F->parse<ELFT>();
72     return;
73   }
74 
75   // .a file
76   if (auto *F = dyn_cast<ArchiveFile>(File)) {
77     F->parse<ELFT>();
78     return;
79   }
80 
81   // Lazy object file
82   if (auto *F = dyn_cast<LazyObjFile>(File)) {
83     F->parse<ELFT>();
84     return;
85   }
86 
87   if (Config->Trace)
88     message(toString(File));
89 
90   // .so file
91   if (auto *F = dyn_cast<SharedFile<ELFT>>(File)) {
92     // DSOs are uniquified not by filename but by soname.
93     F->parseSoName();
94     if (ErrorCount || !SoNames.insert(F->SoName).second)
95       return;
96     SharedFiles.push_back(F);
97     F->parseRest();
98     return;
99   }
100 
101   // LLVM bitcode file
102   if (auto *F = dyn_cast<BitcodeFile>(File)) {
103     BitcodeFiles.push_back(F);
104     F->parse<ELFT>(ComdatGroups);
105     return;
106   }
107 
108   // Regular object file
109   ObjectFiles.push_back(File);
110   cast<ObjFile<ELFT>>(File)->parse(ComdatGroups);
111 }
112 
113 // This function is where all the optimizations of link-time
114 // optimization happens. When LTO is in use, some input files are
115 // not in native object file format but in the LLVM bitcode format.
116 // This function compiles bitcode files into a few big native files
117 // using LLVM functions and replaces bitcode symbols with the results.
118 // Because all bitcode files that consist of a program are passed
119 // to the compiler at once, it can do whole-program optimization.
120 template <class ELFT> void SymbolTable::addCombinedLTOObject() {
121   if (BitcodeFiles.empty())
122     return;
123 
124   // Compile bitcode files and replace bitcode symbols.
125   LTO.reset(new BitcodeCompiler);
126   for (BitcodeFile *F : BitcodeFiles)
127     LTO->add(*F);
128 
129   for (InputFile *File : LTO->compile()) {
130     DenseSet<CachedHashStringRef> DummyGroups;
131     cast<ObjFile<ELFT>>(File)->parse(DummyGroups);
132     ObjectFiles.push_back(File);
133   }
134 }
135 
136 template <class ELFT>
137 DefinedRegular *SymbolTable::addAbsolute(StringRef Name, uint8_t Visibility,
138                                          uint8_t Binding) {
139   Symbol *Sym = addRegular<ELFT>(Name, Visibility, STT_NOTYPE, 0, 0, Binding,
140                                  nullptr, nullptr);
141   return cast<DefinedRegular>(Sym->body());
142 }
143 
144 // Set a flag for --trace-symbol so that we can print out a log message
145 // if a new symbol with the same name is inserted into the symbol table.
146 void SymbolTable::trace(StringRef Name) {
147   Symtab.insert({CachedHashStringRef(Name), {-1, true}});
148 }
149 
150 // Rename SYM as __wrap_SYM. The original symbol is preserved as __real_SYM.
151 // Used to implement --wrap.
152 template <class ELFT> void SymbolTable::addSymbolWrap(StringRef Name) {
153   SymbolBody *B = find(Name);
154   if (!B)
155     return;
156   Symbol *Sym = B->symbol();
157   Symbol *Real = addUndefined<ELFT>(Saver.save("__real_" + Name));
158   Symbol *Wrap = addUndefined<ELFT>(Saver.save("__wrap_" + Name));
159 
160   defsym(Real, Sym);
161   defsym(Sym, Wrap);
162 
163   WrapSymbols.push_back({Wrap, Real});
164 }
165 
166 // Creates alias for symbol. Used to implement --defsym=ALIAS=SYM.
167 template <class ELFT>
168 void SymbolTable::addSymbolAlias(StringRef Alias, StringRef Name) {
169   SymbolBody *B = find(Name);
170   if (!B) {
171     error("-defsym: undefined symbol: " + Name);
172     return;
173   }
174 
175   defsym(addUndefined<ELFT>(Alias), B->symbol());
176 }
177 
178 // Apply symbol renames created by -wrap and -defsym. The renames are created
179 // before LTO in addSymbolWrap() and addSymbolAlias() to have a chance to inform
180 // LTO (if LTO is running) not to include these symbols in IPO. Now that the
181 // symbols are finalized, we can perform the replacement.
182 void SymbolTable::applySymbolRenames() {
183   // This function rotates 3 symbols:
184   //
185   // __real_foo becomes foo
186   // foo        becomes __wrap_foo
187   // __wrap_foo becomes __real_foo
188   //
189   // The last part is special in that we don't want to change what references to
190   // __wrap_foo point to, we just want have __real_foo in the symbol table.
191 
192   // First make a copy of __real_foo
193   std::vector<Symbol> Origs;
194   for (const auto &P : WrapSymbols)
195     Origs.push_back(*P.second);
196 
197   // Replace __real_foo with foo and foo with __wrap_foo
198   for (SymbolRenaming &S : Defsyms) {
199     S.Dst->body()->copyFrom(S.Src->body());
200     S.Dst->File = S.Src->File;
201     S.Dst->Binding = S.Binding;
202   }
203 
204   // Hide one of the copies of __wrap_foo, create a new symbol and copy
205   // __real_foo into it.
206   for (unsigned I = 0, N = WrapSymbols.size(); I < N; ++I) {
207     // We now have two copies of __wrap_foo. Drop one.
208     Symbol *Wrap = WrapSymbols[I].first;
209     Wrap->IsUsedInRegularObj = false;
210 
211     Symbol *Real = &Origs[I];
212     // If __real_foo was undefined, we don't want it in the symbol table.
213     if (!Real->body()->isInCurrentDSO())
214       continue;
215 
216     auto *NewSym = make<Symbol>();
217     memcpy(NewSym, Real, sizeof(Symbol));
218     SymVector.push_back(NewSym);
219   }
220 }
221 
222 static uint8_t getMinVisibility(uint8_t VA, uint8_t VB) {
223   if (VA == STV_DEFAULT)
224     return VB;
225   if (VB == STV_DEFAULT)
226     return VA;
227   return std::min(VA, VB);
228 }
229 
230 // Find an existing symbol or create and insert a new one.
231 std::pair<Symbol *, bool> SymbolTable::insert(StringRef Name) {
232   // <name>@@<version> means the symbol is the default version. In that
233   // case <name>@@<version> will be used to resolve references to <name>.
234   //
235   // Since this is a hot path, the following string search code is
236   // optimized for speed. StringRef::find(char) is much faster than
237   // StringRef::find(StringRef).
238   size_t Pos = Name.find('@');
239   if (Pos != StringRef::npos && Pos + 1 < Name.size() && Name[Pos + 1] == '@')
240     Name = Name.take_front(Pos);
241 
242   auto P = Symtab.insert(
243       {CachedHashStringRef(Name), SymIndex((int)SymVector.size(), false)});
244   SymIndex &V = P.first->second;
245   bool IsNew = P.second;
246 
247   if (V.Idx == -1) {
248     IsNew = true;
249     V = SymIndex((int)SymVector.size(), true);
250   }
251 
252   Symbol *Sym;
253   if (IsNew) {
254     Sym = make<Symbol>();
255     Sym->InVersionScript = false;
256     Sym->Binding = STB_WEAK;
257     Sym->Visibility = STV_DEFAULT;
258     Sym->IsUsedInRegularObj = false;
259     Sym->ExportDynamic = false;
260     Sym->CanInline = true;
261     Sym->Traced = V.Traced;
262     Sym->VersionId = Config->DefaultSymbolVersion;
263     SymVector.push_back(Sym);
264   } else {
265     Sym = SymVector[V.Idx];
266   }
267   return {Sym, IsNew};
268 }
269 
270 // Find an existing symbol or create and insert a new one, then apply the given
271 // attributes.
272 std::pair<Symbol *, bool> SymbolTable::insert(StringRef Name, uint8_t Type,
273                                               uint8_t Visibility,
274                                               bool CanOmitFromDynSym,
275                                               InputFile *File) {
276   bool IsUsedInRegularObj = !File || File->kind() == InputFile::ObjKind;
277   Symbol *S;
278   bool WasInserted;
279   std::tie(S, WasInserted) = insert(Name);
280 
281   // Merge in the new symbol's visibility.
282   S->Visibility = getMinVisibility(S->Visibility, Visibility);
283 
284   if (!CanOmitFromDynSym && (Config->Shared || Config->ExportDynamic))
285     S->ExportDynamic = true;
286 
287   if (IsUsedInRegularObj)
288     S->IsUsedInRegularObj = true;
289 
290   if (!WasInserted && S->body()->Type != SymbolBody::UnknownType &&
291       ((Type == STT_TLS) != S->body()->isTls())) {
292     error("TLS attribute mismatch: " + toString(*S->body()) +
293           "\n>>> defined in " + toString(S->File) + "\n>>> defined in " +
294           toString(File));
295   }
296 
297   return {S, WasInserted};
298 }
299 
300 template <class ELFT> Symbol *SymbolTable::addUndefined(StringRef Name) {
301   return addUndefined<ELFT>(Name, /*IsLocal=*/false, STB_GLOBAL, STV_DEFAULT,
302                             /*Type*/ 0,
303                             /*CanOmitFromDynSym*/ false, /*File*/ nullptr);
304 }
305 
306 static uint8_t getVisibility(uint8_t StOther) { return StOther & 3; }
307 
308 template <class ELFT>
309 Symbol *SymbolTable::addUndefined(StringRef Name, bool IsLocal, uint8_t Binding,
310                                   uint8_t StOther, uint8_t Type,
311                                   bool CanOmitFromDynSym, InputFile *File) {
312   Symbol *S;
313   bool WasInserted;
314   uint8_t Visibility = getVisibility(StOther);
315   std::tie(S, WasInserted) =
316       insert(Name, Type, Visibility, CanOmitFromDynSym, File);
317   // An undefined symbol with non default visibility must be satisfied
318   // in the same DSO.
319   if (WasInserted ||
320       (isa<SharedSymbol>(S->body()) && Visibility != STV_DEFAULT)) {
321     S->Binding = Binding;
322     replaceBody<Undefined>(S, File, Name, IsLocal, StOther, Type);
323     return S;
324   }
325   if (Binding != STB_WEAK) {
326     SymbolBody *B = S->body();
327     if (!B->isInCurrentDSO())
328       S->Binding = Binding;
329     if (auto *SS = dyn_cast<SharedSymbol>(B))
330       SS->getFile<ELFT>()->IsUsed = true;
331   }
332   if (auto *L = dyn_cast<Lazy>(S->body())) {
333     // An undefined weak will not fetch archive members. See comment on Lazy in
334     // Symbols.h for the details.
335     if (S->isWeak())
336       L->Type = Type;
337     else if (InputFile *F = L->fetch())
338       addFile<ELFT>(F);
339   }
340   return S;
341 }
342 
343 // Using .symver foo,foo@@VER unfortunately creates two symbols: foo and
344 // foo@@VER. We want to effectively ignore foo, so give precedence to
345 // foo@@VER.
346 // FIXME: If users can transition to using
347 // .symver foo,foo@@@VER
348 // we can delete this hack.
349 static int compareVersion(Symbol *S, StringRef Name) {
350   bool A = Name.contains("@@");
351   bool B = S->body()->getName().contains("@@");
352   if (A && !B)
353     return 1;
354   if (!A && B)
355     return -1;
356   return 0;
357 }
358 
359 // We have a new defined symbol with the specified binding. Return 1 if the new
360 // symbol should win, -1 if the new symbol should lose, or 0 if both symbols are
361 // strong defined symbols.
362 static int compareDefined(Symbol *S, bool WasInserted, uint8_t Binding,
363                           StringRef Name) {
364   if (WasInserted)
365     return 1;
366   SymbolBody *Body = S->body();
367   if (!Body->isInCurrentDSO())
368     return 1;
369 
370   if (int R = compareVersion(S, Name))
371     return R;
372 
373   if (Binding == STB_WEAK)
374     return -1;
375   if (S->isWeak())
376     return 1;
377   return 0;
378 }
379 
380 // We have a new non-common defined symbol with the specified binding. Return 1
381 // if the new symbol should win, -1 if the new symbol should lose, or 0 if there
382 // is a conflict. If the new symbol wins, also update the binding.
383 static int compareDefinedNonCommon(Symbol *S, bool WasInserted, uint8_t Binding,
384                                    bool IsAbsolute, uint64_t Value,
385                                    StringRef Name) {
386   if (int Cmp = compareDefined(S, WasInserted, Binding, Name)) {
387     if (Cmp > 0)
388       S->Binding = Binding;
389     return Cmp;
390   }
391   SymbolBody *B = S->body();
392   if (isa<DefinedCommon>(B)) {
393     // Non-common symbols take precedence over common symbols.
394     if (Config->WarnCommon)
395       warn("common " + S->body()->getName() + " is overridden");
396     return 1;
397   } else if (auto *R = dyn_cast<DefinedRegular>(B)) {
398     if (R->Section == nullptr && Binding == STB_GLOBAL && IsAbsolute &&
399         R->Value == Value)
400       return -1;
401   }
402   return 0;
403 }
404 
405 Symbol *SymbolTable::addCommon(StringRef N, uint64_t Size, uint32_t Alignment,
406                                uint8_t Binding, uint8_t StOther, uint8_t Type,
407                                InputFile *File) {
408   Symbol *S;
409   bool WasInserted;
410   std::tie(S, WasInserted) = insert(N, Type, getVisibility(StOther),
411                                     /*CanOmitFromDynSym*/ false, File);
412   int Cmp = compareDefined(S, WasInserted, Binding, N);
413   if (Cmp > 0) {
414     S->Binding = Binding;
415     replaceBody<DefinedCommon>(S, File, N, Size, Alignment, StOther, Type);
416   } else if (Cmp == 0) {
417     auto *C = dyn_cast<DefinedCommon>(S->body());
418     if (!C) {
419       // Non-common symbols take precedence over common symbols.
420       if (Config->WarnCommon)
421         warn("common " + S->body()->getName() + " is overridden");
422       return S;
423     }
424 
425     if (Config->WarnCommon)
426       warn("multiple common of " + S->body()->getName());
427 
428     Alignment = C->Alignment = std::max(C->Alignment, Alignment);
429     if (Size > C->Size)
430       replaceBody<DefinedCommon>(S, File, N, Size, Alignment, StOther, Type);
431   }
432   return S;
433 }
434 
435 static void warnOrError(const Twine &Msg) {
436   if (Config->AllowMultipleDefinition)
437     warn(Msg);
438   else
439     error(Msg);
440 }
441 
442 static void reportDuplicate(SymbolBody *Sym, InputFile *NewFile) {
443   warnOrError("duplicate symbol: " + toString(*Sym) + "\n>>> defined in " +
444               toString(Sym->getFile()) + "\n>>> defined in " +
445               toString(NewFile));
446 }
447 
448 template <class ELFT>
449 static void reportDuplicate(SymbolBody *Sym, InputSectionBase *ErrSec,
450                             typename ELFT::uint ErrOffset) {
451   DefinedRegular *D = dyn_cast<DefinedRegular>(Sym);
452   if (!D || !D->Section || !ErrSec) {
453     reportDuplicate(Sym, ErrSec ? ErrSec->File : nullptr);
454     return;
455   }
456 
457   // Construct and print an error message in the form of:
458   //
459   //   ld.lld: error: duplicate symbol: foo
460   //   >>> defined at bar.c:30
461   //   >>>            bar.o (/home/alice/src/bar.o)
462   //   >>> defined at baz.c:563
463   //   >>>            baz.o in archive libbaz.a
464   auto *Sec1 = cast<InputSectionBase>(D->Section);
465   std::string Src1 = Sec1->getSrcMsg<ELFT>(D->Value);
466   std::string Obj1 = Sec1->getObjMsg<ELFT>(D->Value);
467   std::string Src2 = ErrSec->getSrcMsg<ELFT>(ErrOffset);
468   std::string Obj2 = ErrSec->getObjMsg<ELFT>(ErrOffset);
469 
470   std::string Msg = "duplicate symbol: " + toString(*Sym) + "\n>>> defined at ";
471   if (!Src1.empty())
472     Msg += Src1 + "\n>>>            ";
473   Msg += Obj1 + "\n>>> defined at ";
474   if (!Src2.empty())
475     Msg += Src2 + "\n>>>            ";
476   Msg += Obj2;
477   warnOrError(Msg);
478 }
479 
480 template <typename ELFT>
481 Symbol *SymbolTable::addRegular(StringRef Name, uint8_t StOther, uint8_t Type,
482                                 uint64_t Value, uint64_t Size, uint8_t Binding,
483                                 SectionBase *Section, InputFile *File) {
484   Symbol *S;
485   bool WasInserted;
486   std::tie(S, WasInserted) = insert(Name, Type, getVisibility(StOther),
487                                     /*CanOmitFromDynSym*/ false, File);
488   int Cmp = compareDefinedNonCommon(S, WasInserted, Binding, Section == nullptr,
489                                     Value, Name);
490   if (Cmp > 0)
491     replaceBody<DefinedRegular>(S, File, Name, /*IsLocal=*/false, StOther, Type,
492                                 Value, Size, Section);
493   else if (Cmp == 0)
494     reportDuplicate<ELFT>(S->body(),
495                           dyn_cast_or_null<InputSectionBase>(Section), Value);
496   return S;
497 }
498 
499 template <typename ELFT>
500 void SymbolTable::addShared(StringRef Name, SharedFile<ELFT> *File,
501                             const typename ELFT::Sym &Sym,
502                             const typename ELFT::Verdef *Verdef) {
503   // DSO symbols do not affect visibility in the output, so we pass STV_DEFAULT
504   // as the visibility, which will leave the visibility in the symbol table
505   // unchanged.
506   Symbol *S;
507   bool WasInserted;
508   std::tie(S, WasInserted) = insert(Name, Sym.getType(), STV_DEFAULT,
509                                     /*CanOmitFromDynSym*/ true, File);
510   // Make sure we preempt DSO symbols with default visibility.
511   if (Sym.getVisibility() == STV_DEFAULT)
512     S->ExportDynamic = true;
513 
514   SymbolBody *Body = S->body();
515   // An undefined symbol with non default visibility must be satisfied
516   // in the same DSO.
517   if (WasInserted || ((Body->isUndefined() || Body->isLazy()) &&
518                       Body->getVisibility() == STV_DEFAULT)) {
519     replaceBody<SharedSymbol>(S, File, Name, Sym.st_other, Sym.getType(), &Sym,
520                               Verdef);
521     if (!S->isWeak())
522       File->IsUsed = true;
523   }
524 }
525 
526 Symbol *SymbolTable::addBitcode(StringRef Name, uint8_t Binding,
527                                 uint8_t StOther, uint8_t Type,
528                                 bool CanOmitFromDynSym, BitcodeFile *F) {
529   Symbol *S;
530   bool WasInserted;
531   std::tie(S, WasInserted) =
532       insert(Name, Type, getVisibility(StOther), CanOmitFromDynSym, F);
533   int Cmp = compareDefinedNonCommon(S, WasInserted, Binding,
534                                     /*IsAbs*/ false, /*Value*/ 0, Name);
535   if (Cmp > 0)
536     replaceBody<DefinedRegular>(S, F, Name, /*IsLocal=*/false, StOther, Type, 0,
537                                 0, nullptr);
538   else if (Cmp == 0)
539     reportDuplicate(S->body(), F);
540   return S;
541 }
542 
543 SymbolBody *SymbolTable::find(StringRef Name) {
544   auto It = Symtab.find(CachedHashStringRef(Name));
545   if (It == Symtab.end())
546     return nullptr;
547   SymIndex V = It->second;
548   if (V.Idx == -1)
549     return nullptr;
550   return SymVector[V.Idx]->body();
551 }
552 
553 void SymbolTable::defsym(Symbol *Dst, Symbol *Src) {
554   // We want to tell LTO not to inline Dst symbol because LTO doesn't
555   // know the final symbol contents after renaming.
556   Dst->CanInline = false;
557 
558   // Tell LTO not to eliminate this symbol.
559   Src->IsUsedInRegularObj = true;
560 
561   Defsyms.push_back({Dst, Src, Dst->Binding});
562 }
563 
564 template <class ELFT>
565 Symbol *SymbolTable::addLazyArchive(StringRef Name, ArchiveFile *F,
566                                     const object::Archive::Symbol Sym) {
567   Symbol *S;
568   bool WasInserted;
569   std::tie(S, WasInserted) = insert(Name);
570   if (WasInserted) {
571     replaceBody<LazyArchive>(S, F, Sym, SymbolBody::UnknownType);
572     return S;
573   }
574   if (!S->body()->isUndefined())
575     return S;
576 
577   // An undefined weak will not fetch archive members. See comment on Lazy in
578   // Symbols.h for the details.
579   if (S->isWeak()) {
580     replaceBody<LazyArchive>(S, F, Sym, S->body()->Type);
581     return S;
582   }
583   std::pair<MemoryBufferRef, uint64_t> MBInfo = F->getMember(&Sym);
584   if (!MBInfo.first.getBuffer().empty())
585     addFile<ELFT>(createObjectFile(MBInfo.first, F->getName(), MBInfo.second));
586   return S;
587 }
588 
589 template <class ELFT>
590 void SymbolTable::addLazyObject(StringRef Name, LazyObjFile &Obj) {
591   Symbol *S;
592   bool WasInserted;
593   std::tie(S, WasInserted) = insert(Name);
594   if (WasInserted) {
595     replaceBody<LazyObject>(S, &Obj, Name, SymbolBody::UnknownType);
596     return;
597   }
598   if (!S->body()->isUndefined())
599     return;
600 
601   // See comment for addLazyArchive above.
602   if (S->isWeak())
603     replaceBody<LazyObject>(S, &Obj, Name, S->body()->Type);
604   else if (InputFile *F = Obj.fetch())
605     addFile<ELFT>(F);
606 }
607 
608 // If we already saw this symbol, force loading its file.
609 template <class ELFT> void SymbolTable::fetchIfLazy(StringRef Name) {
610   if (SymbolBody *B = find(Name)) {
611     // Mark the symbol not to be eliminated by LTO
612     // even if it is a bitcode symbol.
613     B->symbol()->IsUsedInRegularObj = true;
614     if (auto *L = dyn_cast_or_null<Lazy>(B))
615       if (InputFile *File = L->fetch())
616         addFile<ELFT>(File);
617   }
618 }
619 
620 // This function takes care of the case in which shared libraries depend on
621 // the user program (not the other way, which is usual). Shared libraries
622 // may have undefined symbols, expecting that the user program provides
623 // the definitions for them. An example is BSD's __progname symbol.
624 // We need to put such symbols to the main program's .dynsym so that
625 // shared libraries can find them.
626 // Except this, we ignore undefined symbols in DSOs.
627 template <class ELFT> void SymbolTable::scanShlibUndefined() {
628   for (InputFile *F : SharedFiles) {
629     for (StringRef U : cast<SharedFile<ELFT>>(F)->getUndefinedSymbols()) {
630       SymbolBody *Sym = find(U);
631       if (!Sym || !Sym->isDefined())
632         continue;
633       Sym->symbol()->ExportDynamic = true;
634 
635       // If -dynamic-list is given, the default version is set to
636       // VER_NDX_LOCAL, which prevents a symbol to be exported via .dynsym.
637       // Set to VER_NDX_GLOBAL so the symbol will be handled as if it were
638       // specified by -dynamic-list.
639       Sym->symbol()->VersionId = VER_NDX_GLOBAL;
640     }
641   }
642 }
643 
644 // Initialize DemangledSyms with a map from demangled symbols to symbol
645 // objects. Used to handle "extern C++" directive in version scripts.
646 //
647 // The map will contain all demangled symbols. That can be very large,
648 // and in LLD we generally want to avoid do anything for each symbol.
649 // Then, why are we doing this? Here's why.
650 //
651 // Users can use "extern C++ {}" directive to match against demangled
652 // C++ symbols. For example, you can write a pattern such as
653 // "llvm::*::foo(int, ?)". Obviously, there's no way to handle this
654 // other than trying to match a pattern against all demangled symbols.
655 // So, if "extern C++" feature is used, we need to demangle all known
656 // symbols.
657 StringMap<std::vector<SymbolBody *>> &SymbolTable::getDemangledSyms() {
658   if (!DemangledSyms) {
659     DemangledSyms.emplace();
660     for (Symbol *Sym : SymVector) {
661       SymbolBody *B = Sym->body();
662       if (!B->isInCurrentDSO())
663         continue;
664       if (Optional<std::string> S = demangle(B->getName()))
665         (*DemangledSyms)[*S].push_back(B);
666       else
667         (*DemangledSyms)[B->getName()].push_back(B);
668     }
669   }
670   return *DemangledSyms;
671 }
672 
673 std::vector<SymbolBody *> SymbolTable::findByVersion(SymbolVersion Ver) {
674   if (Ver.IsExternCpp)
675     return getDemangledSyms().lookup(Ver.Name);
676   if (SymbolBody *B = find(Ver.Name))
677     if (B->isInCurrentDSO())
678       return {B};
679   return {};
680 }
681 
682 std::vector<SymbolBody *> SymbolTable::findAllByVersion(SymbolVersion Ver) {
683   std::vector<SymbolBody *> Res;
684   StringMatcher M(Ver.Name);
685 
686   if (Ver.IsExternCpp) {
687     for (auto &P : getDemangledSyms())
688       if (M.match(P.first()))
689         Res.insert(Res.end(), P.second.begin(), P.second.end());
690     return Res;
691   }
692 
693   for (Symbol *Sym : SymVector) {
694     SymbolBody *B = Sym->body();
695     if (B->isInCurrentDSO() && M.match(B->getName()))
696       Res.push_back(B);
697   }
698   return Res;
699 }
700 
701 // If there's only one anonymous version definition in a version
702 // script file, the script does not actually define any symbol version,
703 // but just specifies symbols visibilities.
704 void SymbolTable::handleAnonymousVersion() {
705   for (SymbolVersion &Ver : Config->VersionScriptGlobals)
706     assignExactVersion(Ver, VER_NDX_GLOBAL, "global");
707   for (SymbolVersion &Ver : Config->VersionScriptGlobals)
708     assignWildcardVersion(Ver, VER_NDX_GLOBAL);
709   for (SymbolVersion &Ver : Config->VersionScriptLocals)
710     assignExactVersion(Ver, VER_NDX_LOCAL, "local");
711   for (SymbolVersion &Ver : Config->VersionScriptLocals)
712     assignWildcardVersion(Ver, VER_NDX_LOCAL);
713 }
714 
715 // Handles -dynamic-list.
716 void SymbolTable::handleDynamicList() {
717   for (SymbolVersion &Ver : Config->DynamicList) {
718     std::vector<SymbolBody *> Syms;
719     if (Ver.HasWildcard)
720       Syms = findByVersion(Ver);
721     else
722       Syms = findAllByVersion(Ver);
723 
724     for (SymbolBody *B : Syms) {
725       if (!Config->Shared)
726         B->symbol()->ExportDynamic = true;
727       else if (B->symbol()->includeInDynsym())
728         B->IsPreemptible = true;
729     }
730   }
731 }
732 
733 // Set symbol versions to symbols. This function handles patterns
734 // containing no wildcard characters.
735 void SymbolTable::assignExactVersion(SymbolVersion Ver, uint16_t VersionId,
736                                      StringRef VersionName) {
737   if (Ver.HasWildcard)
738     return;
739 
740   // Get a list of symbols which we need to assign the version to.
741   std::vector<SymbolBody *> Syms = findByVersion(Ver);
742   if (Syms.empty()) {
743     if (Config->NoUndefinedVersion)
744       error("version script assignment of '" + VersionName + "' to symbol '" +
745             Ver.Name + "' failed: symbol not defined");
746     return;
747   }
748 
749   // Assign the version.
750   for (SymbolBody *B : Syms) {
751     // Skip symbols containing version info because symbol versions
752     // specified by symbol names take precedence over version scripts.
753     // See parseSymbolVersion().
754     if (B->getName().contains('@'))
755       continue;
756 
757     Symbol *Sym = B->symbol();
758     if (Sym->InVersionScript)
759       warn("duplicate symbol '" + Ver.Name + "' in version script");
760     Sym->VersionId = VersionId;
761     Sym->InVersionScript = true;
762   }
763 }
764 
765 void SymbolTable::assignWildcardVersion(SymbolVersion Ver, uint16_t VersionId) {
766   if (!Ver.HasWildcard)
767     return;
768 
769   // Exact matching takes precendence over fuzzy matching,
770   // so we set a version to a symbol only if no version has been assigned
771   // to the symbol. This behavior is compatible with GNU.
772   for (SymbolBody *B : findAllByVersion(Ver))
773     if (B->symbol()->VersionId == Config->DefaultSymbolVersion)
774       B->symbol()->VersionId = VersionId;
775 }
776 
777 // This function processes version scripts by updating VersionId
778 // member of symbols.
779 void SymbolTable::scanVersionScript() {
780   // Handle edge cases first.
781   handleAnonymousVersion();
782   handleDynamicList();
783 
784   // Now we have version definitions, so we need to set version ids to symbols.
785   // Each version definition has a glob pattern, and all symbols that match
786   // with the pattern get that version.
787 
788   // First, we assign versions to exact matching symbols,
789   // i.e. version definitions not containing any glob meta-characters.
790   for (VersionDefinition &V : Config->VersionDefinitions)
791     for (SymbolVersion &Ver : V.Globals)
792       assignExactVersion(Ver, V.Id, V.Name);
793 
794   // Next, we assign versions to fuzzy matching symbols,
795   // i.e. version definitions containing glob meta-characters.
796   // Note that because the last match takes precedence over previous matches,
797   // we iterate over the definitions in the reverse order.
798   for (VersionDefinition &V : llvm::reverse(Config->VersionDefinitions))
799     for (SymbolVersion &Ver : V.Globals)
800       assignWildcardVersion(Ver, V.Id);
801 
802   // Symbol themselves might know their versions because symbols
803   // can contain versions in the form of <name>@<version>.
804   // Let them parse and update their names to exclude version suffix.
805   for (Symbol *Sym : SymVector)
806     Sym->body()->parseSymbolVersion();
807 }
808 
809 template void SymbolTable::addSymbolWrap<ELF32LE>(StringRef);
810 template void SymbolTable::addSymbolWrap<ELF32BE>(StringRef);
811 template void SymbolTable::addSymbolWrap<ELF64LE>(StringRef);
812 template void SymbolTable::addSymbolWrap<ELF64BE>(StringRef);
813 
814 template Symbol *SymbolTable::addUndefined<ELF32LE>(StringRef);
815 template Symbol *SymbolTable::addUndefined<ELF32BE>(StringRef);
816 template Symbol *SymbolTable::addUndefined<ELF64LE>(StringRef);
817 template Symbol *SymbolTable::addUndefined<ELF64BE>(StringRef);
818 
819 template Symbol *SymbolTable::addUndefined<ELF32LE>(StringRef, bool, uint8_t,
820                                                     uint8_t, uint8_t, bool,
821                                                     InputFile *);
822 template Symbol *SymbolTable::addUndefined<ELF32BE>(StringRef, bool, uint8_t,
823                                                     uint8_t, uint8_t, bool,
824                                                     InputFile *);
825 template Symbol *SymbolTable::addUndefined<ELF64LE>(StringRef, bool, uint8_t,
826                                                     uint8_t, uint8_t, bool,
827                                                     InputFile *);
828 template Symbol *SymbolTable::addUndefined<ELF64BE>(StringRef, bool, uint8_t,
829                                                     uint8_t, uint8_t, bool,
830                                                     InputFile *);
831 
832 template void SymbolTable::addSymbolAlias<ELF32LE>(StringRef, StringRef);
833 template void SymbolTable::addSymbolAlias<ELF32BE>(StringRef, StringRef);
834 template void SymbolTable::addSymbolAlias<ELF64LE>(StringRef, StringRef);
835 template void SymbolTable::addSymbolAlias<ELF64BE>(StringRef, StringRef);
836 
837 template void SymbolTable::addCombinedLTOObject<ELF32LE>();
838 template void SymbolTable::addCombinedLTOObject<ELF32BE>();
839 template void SymbolTable::addCombinedLTOObject<ELF64LE>();
840 template void SymbolTable::addCombinedLTOObject<ELF64BE>();
841 
842 template Symbol *SymbolTable::addRegular<ELF32LE>(StringRef, uint8_t, uint8_t,
843                                                   uint64_t, uint64_t, uint8_t,
844                                                   SectionBase *, InputFile *);
845 template Symbol *SymbolTable::addRegular<ELF32BE>(StringRef, uint8_t, uint8_t,
846                                                   uint64_t, uint64_t, uint8_t,
847                                                   SectionBase *, InputFile *);
848 template Symbol *SymbolTable::addRegular<ELF64LE>(StringRef, uint8_t, uint8_t,
849                                                   uint64_t, uint64_t, uint8_t,
850                                                   SectionBase *, InputFile *);
851 template Symbol *SymbolTable::addRegular<ELF64BE>(StringRef, uint8_t, uint8_t,
852                                                   uint64_t, uint64_t, uint8_t,
853                                                   SectionBase *, InputFile *);
854 
855 template DefinedRegular *SymbolTable::addAbsolute<ELF32LE>(StringRef, uint8_t,
856                                                            uint8_t);
857 template DefinedRegular *SymbolTable::addAbsolute<ELF32BE>(StringRef, uint8_t,
858                                                            uint8_t);
859 template DefinedRegular *SymbolTable::addAbsolute<ELF64LE>(StringRef, uint8_t,
860                                                            uint8_t);
861 template DefinedRegular *SymbolTable::addAbsolute<ELF64BE>(StringRef, uint8_t,
862                                                            uint8_t);
863 
864 template Symbol *
865 SymbolTable::addLazyArchive<ELF32LE>(StringRef, ArchiveFile *,
866                                      const object::Archive::Symbol);
867 template Symbol *
868 SymbolTable::addLazyArchive<ELF32BE>(StringRef, ArchiveFile *,
869                                      const object::Archive::Symbol);
870 template Symbol *
871 SymbolTable::addLazyArchive<ELF64LE>(StringRef, ArchiveFile *,
872                                      const object::Archive::Symbol);
873 template Symbol *
874 SymbolTable::addLazyArchive<ELF64BE>(StringRef, ArchiveFile *,
875                                      const object::Archive::Symbol);
876 
877 template void SymbolTable::addLazyObject<ELF32LE>(StringRef, LazyObjFile &);
878 template void SymbolTable::addLazyObject<ELF32BE>(StringRef, LazyObjFile &);
879 template void SymbolTable::addLazyObject<ELF64LE>(StringRef, LazyObjFile &);
880 template void SymbolTable::addLazyObject<ELF64BE>(StringRef, LazyObjFile &);
881 
882 template void SymbolTable::addShared<ELF32LE>(StringRef, SharedFile<ELF32LE> *,
883                                               const typename ELF32LE::Sym &,
884                                               const typename ELF32LE::Verdef *);
885 template void SymbolTable::addShared<ELF32BE>(StringRef, SharedFile<ELF32BE> *,
886                                               const typename ELF32BE::Sym &,
887                                               const typename ELF32BE::Verdef *);
888 template void SymbolTable::addShared<ELF64LE>(StringRef, SharedFile<ELF64LE> *,
889                                               const typename ELF64LE::Sym &,
890                                               const typename ELF64LE::Verdef *);
891 template void SymbolTable::addShared<ELF64BE>(StringRef, SharedFile<ELF64BE> *,
892                                               const typename ELF64BE::Sym &,
893                                               const typename ELF64BE::Verdef *);
894 
895 template void SymbolTable::fetchIfLazy<ELF32LE>(StringRef);
896 template void SymbolTable::fetchIfLazy<ELF32BE>(StringRef);
897 template void SymbolTable::fetchIfLazy<ELF64LE>(StringRef);
898 template void SymbolTable::fetchIfLazy<ELF64BE>(StringRef);
899 
900 template void SymbolTable::scanShlibUndefined<ELF32LE>();
901 template void SymbolTable::scanShlibUndefined<ELF32BE>();
902 template void SymbolTable::scanShlibUndefined<ELF64LE>();
903 template void SymbolTable::scanShlibUndefined<ELF64BE>();
904