1 //===- SymbolTable.cpp ----------------------------------------------------===//
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 #include "SymbolTable.h"
10 #include "ConcatOutputSection.h"
11 #include "Config.h"
12 #include "InputFiles.h"
13 #include "InputSection.h"
14 #include "Symbols.h"
15 #include "SyntheticSections.h"
16 #include "lld/Common/ErrorHandler.h"
17 #include "lld/Common/Memory.h"
18 
19 using namespace llvm;
20 using namespace lld;
21 using namespace lld::macho;
22 
23 Symbol *SymbolTable::find(CachedHashStringRef cachedName) {
24   auto it = symMap.find(cachedName);
25   if (it == symMap.end())
26     return nullptr;
27   return symVector[it->second];
28 }
29 
30 std::pair<Symbol *, bool> SymbolTable::insert(StringRef name,
31                                               const InputFile *file) {
32   auto p = symMap.insert({CachedHashStringRef(name), (int)symVector.size()});
33 
34   Symbol *sym;
35   if (!p.second) {
36     // Name already present in the symbol table.
37     sym = symVector[p.first->second];
38   } else {
39     // Name is a new symbol.
40     sym = reinterpret_cast<Symbol *>(make<SymbolUnion>());
41     symVector.push_back(sym);
42   }
43 
44   sym->isUsedInRegularObj |= !file || isa<ObjFile>(file);
45   return {sym, p.second};
46 }
47 
48 Defined *SymbolTable::addDefined(StringRef name, InputFile *file,
49                                  InputSection *isec, uint64_t value,
50                                  uint64_t size, bool isWeakDef,
51                                  bool isPrivateExtern, bool isThumb,
52                                  bool isReferencedDynamically, bool noDeadStrip,
53                                  bool isWeakDefCanBeHidden) {
54   Symbol *s;
55   bool wasInserted;
56   bool overridesWeakDef = false;
57   std::tie(s, wasInserted) = insert(name, file);
58 
59   assert(!isWeakDef || (isa<BitcodeFile>(file) && !isec) ||
60          (isa<ObjFile>(file) && file == isec->getFile()));
61 
62   if (!wasInserted) {
63     if (auto *defined = dyn_cast<Defined>(s)) {
64       if (isWeakDef) {
65         // See further comment in createDefined() in InputFiles.cpp
66         if (defined->isWeakDef()) {
67           defined->privateExtern &= isPrivateExtern;
68           defined->weakDefCanBeHidden &= isWeakDefCanBeHidden;
69           defined->referencedDynamically |= isReferencedDynamically;
70           defined->noDeadStrip |= noDeadStrip;
71         }
72         // FIXME: Handle this for bitcode files.
73         if (auto concatIsec = dyn_cast_or_null<ConcatInputSection>(isec))
74           concatIsec->wasCoalesced = true;
75         return defined;
76       }
77 
78       if (defined->isWeakDef()) {
79         // FIXME: Handle this for bitcode files.
80         if (auto concatIsec =
81                 dyn_cast_or_null<ConcatInputSection>(defined->isec)) {
82           concatIsec->wasCoalesced = true;
83           concatIsec->symbols.erase(llvm::find(concatIsec->symbols, defined));
84         }
85       } else {
86         error("duplicate symbol: " + name + "\n>>> defined in " +
87               toString(defined->getFile()) + "\n>>> defined in " +
88               toString(file));
89       }
90 
91     } else if (auto *dysym = dyn_cast<DylibSymbol>(s)) {
92       overridesWeakDef = !isWeakDef && dysym->isWeakDef();
93       dysym->unreference();
94     }
95     // Defined symbols take priority over other types of symbols, so in case
96     // of a name conflict, we fall through to the replaceSymbol() call below.
97   }
98 
99   Defined *defined = replaceSymbol<Defined>(
100       s, name, file, isec, value, size, isWeakDef, /*isExternal=*/true,
101       isPrivateExtern, isThumb, isReferencedDynamically, noDeadStrip,
102       overridesWeakDef, isWeakDefCanBeHidden);
103   return defined;
104 }
105 
106 Symbol *SymbolTable::addUndefined(StringRef name, InputFile *file,
107                                   bool isWeakRef) {
108   Symbol *s;
109   bool wasInserted;
110   std::tie(s, wasInserted) = insert(name, file);
111 
112   RefState refState = isWeakRef ? RefState::Weak : RefState::Strong;
113 
114   if (wasInserted)
115     replaceSymbol<Undefined>(s, name, file, refState);
116   else if (auto *lazy = dyn_cast<LazySymbol>(s))
117     lazy->fetchArchiveMember();
118   else if (auto *dynsym = dyn_cast<DylibSymbol>(s))
119     dynsym->reference(refState);
120   else if (auto *undefined = dyn_cast<Undefined>(s))
121     undefined->refState = std::max(undefined->refState, refState);
122   return s;
123 }
124 
125 Symbol *SymbolTable::addCommon(StringRef name, InputFile *file, uint64_t size,
126                                uint32_t align, bool isPrivateExtern) {
127   Symbol *s;
128   bool wasInserted;
129   std::tie(s, wasInserted) = insert(name, file);
130 
131   if (!wasInserted) {
132     if (auto *common = dyn_cast<CommonSymbol>(s)) {
133       if (size < common->size)
134         return s;
135     } else if (isa<Defined>(s)) {
136       return s;
137     }
138     // Common symbols take priority over all non-Defined symbols, so in case of
139     // a name conflict, we fall through to the replaceSymbol() call below.
140   }
141 
142   replaceSymbol<CommonSymbol>(s, name, file, size, align, isPrivateExtern);
143   return s;
144 }
145 
146 Symbol *SymbolTable::addDylib(StringRef name, DylibFile *file, bool isWeakDef,
147                               bool isTlv) {
148   Symbol *s;
149   bool wasInserted;
150   std::tie(s, wasInserted) = insert(name, file);
151 
152   RefState refState = RefState::Unreferenced;
153   if (!wasInserted) {
154     if (auto *defined = dyn_cast<Defined>(s)) {
155       if (isWeakDef && !defined->isWeakDef())
156         defined->overridesWeakDef = true;
157     } else if (auto *undefined = dyn_cast<Undefined>(s)) {
158       refState = undefined->refState;
159     } else if (auto *dysym = dyn_cast<DylibSymbol>(s)) {
160       refState = dysym->getRefState();
161     }
162   }
163 
164   bool isDynamicLookup = file == nullptr;
165   if (wasInserted || isa<Undefined>(s) ||
166       (isa<DylibSymbol>(s) &&
167        ((!isWeakDef && s->isWeakDef()) ||
168         (!isDynamicLookup && cast<DylibSymbol>(s)->isDynamicLookup())))) {
169     if (auto *dynsym = dyn_cast<DylibSymbol>(s))
170       dynsym->unreference();
171     replaceSymbol<DylibSymbol>(s, file, name, isWeakDef, refState, isTlv);
172   }
173 
174   return s;
175 }
176 
177 Symbol *SymbolTable::addDynamicLookup(StringRef name) {
178   return addDylib(name, /*file=*/nullptr, /*isWeakDef=*/false, /*isTlv=*/false);
179 }
180 
181 Symbol *SymbolTable::addLazy(StringRef name, ArchiveFile *file,
182                              const object::Archive::Symbol &sym) {
183   Symbol *s;
184   bool wasInserted;
185   std::tie(s, wasInserted) = insert(name, file);
186 
187   if (wasInserted) {
188     replaceSymbol<LazySymbol>(s, file, sym);
189   } else if (isa<Undefined>(s)) {
190     file->fetch(sym);
191   } else if (auto *dysym = dyn_cast<DylibSymbol>(s)) {
192     if (dysym->isWeakDef()) {
193       if (dysym->getRefState() != RefState::Unreferenced)
194         file->fetch(sym);
195       else
196         replaceSymbol<LazySymbol>(s, file, sym);
197     }
198   }
199   return s;
200 }
201 
202 Defined *SymbolTable::addSynthetic(StringRef name, InputSection *isec,
203                                    uint64_t value, bool isPrivateExtern,
204                                    bool includeInSymtab,
205                                    bool referencedDynamically) {
206   Defined *s =
207       addDefined(name, nullptr, isec, value, /*size=*/0,
208                  /*isWeakDef=*/false, isPrivateExtern,
209                  /*isThumb=*/false, referencedDynamically,
210                  /*noDeadStrip=*/false, /*isWeakDefCanBeHidden=*/false);
211   s->includeInSymtab = includeInSymtab;
212   return s;
213 }
214 
215 enum class Boundary {
216   Start,
217   End,
218 };
219 
220 static Defined *createBoundarySymbol(const Undefined &sym) {
221   return symtab->addSynthetic(
222       sym.getName(), /*isec=*/nullptr, /*value=*/-1, /*isPrivateExtern=*/true,
223       /*includeInSymtab=*/false, /*referencedDynamically=*/false);
224 }
225 
226 static void handleSectionBoundarySymbol(const Undefined &sym, StringRef segSect,
227                                         Boundary which) {
228   StringRef segName, sectName;
229   std::tie(segName, sectName) = segSect.split('$');
230 
231   // Attach the symbol to any InputSection that will end up in the right
232   // OutputSection -- it doesn't matter which one we pick.
233   // Don't bother looking through inputSections for a matching
234   // ConcatInputSection -- we need to create ConcatInputSection for
235   // non-existing sections anyways, and that codepath works even if we should
236   // already have a ConcatInputSection with the right name.
237 
238   OutputSection *osec = nullptr;
239   // This looks for __TEXT,__cstring etc.
240   for (SyntheticSection *ssec : syntheticSections)
241     if (ssec->segname == segName && ssec->name == sectName) {
242       osec = ssec->isec->parent;
243       break;
244     }
245 
246   if (!osec) {
247     ConcatInputSection *isec = make<ConcatInputSection>(segName, sectName);
248 
249     // This runs after markLive() and is only called for Undefineds that are
250     // live. Marking the isec live ensures an OutputSection is created that the
251     // start/end symbol can refer to.
252     assert(sym.isLive());
253     isec->live = true;
254 
255     // This runs after gatherInputSections(), so need to explicitly set parent
256     // and add to inputSections.
257     osec = isec->parent = ConcatOutputSection::getOrCreateForInput(isec);
258     inputSections.push_back(isec);
259   }
260 
261   if (which == Boundary::Start)
262     osec->sectionStartSymbols.push_back(createBoundarySymbol(sym));
263   else
264     osec->sectionEndSymbols.push_back(createBoundarySymbol(sym));
265 }
266 
267 static void handleSegmentBoundarySymbol(const Undefined &sym, StringRef segName,
268                                         Boundary which) {
269   OutputSegment *seg = getOrCreateOutputSegment(segName);
270   if (which == Boundary::Start)
271     seg->segmentStartSymbols.push_back(createBoundarySymbol(sym));
272   else
273     seg->segmentEndSymbols.push_back(createBoundarySymbol(sym));
274 }
275 
276 void lld::macho::treatUndefinedSymbol(const Undefined &sym, StringRef source) {
277   // Handle start/end symbols.
278   StringRef name = sym.getName();
279   if (name.consume_front("section$start$"))
280     return handleSectionBoundarySymbol(sym, name, Boundary::Start);
281   if (name.consume_front("section$end$"))
282     return handleSectionBoundarySymbol(sym, name, Boundary::End);
283   if (name.consume_front("segment$start$"))
284     return handleSegmentBoundarySymbol(sym, name, Boundary::Start);
285   if (name.consume_front("segment$end$"))
286     return handleSegmentBoundarySymbol(sym, name, Boundary::End);
287 
288   // Handle -U.
289   if (config->explicitDynamicLookups.count(sym.getName())) {
290     symtab->addDynamicLookup(sym.getName());
291     return;
292   }
293 
294   // Handle -undefined.
295   auto message = [source, &sym]() {
296     std::string message = "undefined symbol";
297     if (config->archMultiple)
298       message += (" for arch " + getArchitectureName(config->arch())).str();
299     message += ": " + toString(sym);
300     if (!source.empty())
301       message += "\n>>> referenced by " + source.str();
302     else
303       message += "\n>>> referenced by " + toString(sym.getFile());
304     return message;
305   };
306   switch (config->undefinedSymbolTreatment) {
307   case UndefinedSymbolTreatment::error:
308     error(message());
309     break;
310   case UndefinedSymbolTreatment::warning:
311     warn(message());
312     LLVM_FALLTHROUGH;
313   case UndefinedSymbolTreatment::dynamic_lookup:
314   case UndefinedSymbolTreatment::suppress:
315     symtab->addDynamicLookup(sym.getName());
316     break;
317   case UndefinedSymbolTreatment::unknown:
318     llvm_unreachable("unknown -undefined TREATMENT");
319   }
320 }
321 
322 SymbolTable *macho::symtab;
323