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 | #include "llvm/Demangle/Demangle.h" |
19 | |
20 | using namespace llvm; |
21 | using namespace lld; |
22 | using namespace lld::macho; |
23 | |
24 | Symbol *SymbolTable::find(CachedHashStringRef cachedName) { |
25 | auto it = symMap.find(Val: cachedName); |
26 | if (it == symMap.end()) |
27 | return nullptr; |
28 | return symVector[it->second]; |
29 | } |
30 | |
31 | std::pair<Symbol *, bool> SymbolTable::insert(StringRef name, |
32 | const InputFile *file) { |
33 | auto p = symMap.insert(KV: {CachedHashStringRef(name), (int)symVector.size()}); |
34 | |
35 | Symbol *sym; |
36 | if (!p.second) { |
37 | // Name already present in the symbol table. |
38 | sym = symVector[p.first->second]; |
39 | } else { |
40 | // Name is a new symbol. |
41 | sym = reinterpret_cast<Symbol *>(make<SymbolUnion>()); |
42 | symVector.push_back(x: sym); |
43 | } |
44 | |
45 | sym->isUsedInRegularObj |= !file || isa<ObjFile>(Val: file); |
46 | return {sym, p.second}; |
47 | } |
48 | |
49 | namespace { |
50 | struct DuplicateSymbolDiag { |
51 | // Pair containing source location and source file |
52 | const std::pair<std::string, std::string> src1; |
53 | const std::pair<std::string, std::string> src2; |
54 | const Symbol *sym; |
55 | |
56 | DuplicateSymbolDiag(const std::pair<std::string, std::string> src1, |
57 | const std::pair<std::string, std::string> src2, |
58 | const Symbol *sym) |
59 | : src1(src1), src2(src2), sym(sym) {} |
60 | }; |
61 | SmallVector<DuplicateSymbolDiag> dupSymDiags; |
62 | } // namespace |
63 | |
64 | // Move symbols at \p fromOff in \p fromIsec into \p toIsec, unless that symbol |
65 | // is \p skip. |
66 | static void transplantSymbolsAtOffset(InputSection *fromIsec, |
67 | InputSection *toIsec, Defined *skip, |
68 | uint64_t fromOff, uint64_t toOff) { |
69 | // Ensure the symbols will still be in address order after our insertions. |
70 | auto insertIt = llvm::upper_bound(Range&: toIsec->symbols, Value&: toOff, |
71 | C: [](uint64_t off, const Symbol *s) { |
72 | return cast<Defined>(Val: s)->value < off; |
73 | }); |
74 | llvm::erase_if(C&: fromIsec->symbols, P: [&](Symbol *s) { |
75 | auto *d = cast<Defined>(Val: s); |
76 | if (d->value != fromOff) |
77 | return false; |
78 | if (d != skip) { |
79 | // This repeated insertion will be quadratic unless insertIt is the end |
80 | // iterator. However, that is typically the case for files that have |
81 | // .subsections_via_symbols set. |
82 | insertIt = toIsec->symbols.insert(I: insertIt, Elt: d); |
83 | d->originalIsec = toIsec; |
84 | d->value = toOff; |
85 | // We don't want to have more than one unwindEntry at a given address, so |
86 | // drop the redundant ones. We We can safely drop the unwindEntries of |
87 | // the symbols in fromIsec since we will be adding another unwindEntry as |
88 | // we finish parsing toIsec's file. (We can assume that toIsec has its |
89 | // own unwindEntry because of the ODR.) |
90 | d->originalUnwindEntry = nullptr; |
91 | } |
92 | return true; |
93 | }); |
94 | } |
95 | |
96 | Defined *SymbolTable::addDefined(StringRef name, InputFile *file, |
97 | InputSection *isec, uint64_t value, |
98 | uint64_t size, bool isWeakDef, |
99 | bool isPrivateExtern, |
100 | bool isReferencedDynamically, bool noDeadStrip, |
101 | bool isWeakDefCanBeHidden) { |
102 | bool overridesWeakDef = false; |
103 | auto [s, wasInserted] = insert(name, file); |
104 | |
105 | assert(!file || !isa<BitcodeFile>(file) || !isec); |
106 | |
107 | if (!wasInserted) { |
108 | if (auto *defined = dyn_cast<Defined>(Val: s)) { |
109 | if (isWeakDef) { |
110 | // See further comment in createDefined() in InputFiles.cpp |
111 | if (defined->isWeakDef()) { |
112 | defined->privateExtern &= isPrivateExtern; |
113 | defined->weakDefCanBeHidden &= isWeakDefCanBeHidden; |
114 | defined->referencedDynamically |= isReferencedDynamically; |
115 | defined->noDeadStrip |= noDeadStrip; |
116 | } |
117 | if (auto concatIsec = dyn_cast_or_null<ConcatInputSection>(Val: isec)) { |
118 | concatIsec->wasCoalesced = true; |
119 | // Any local symbols that alias the coalesced symbol should be moved |
120 | // into the prevailing section. Note that we have sorted the symbols |
121 | // in ObjFile::parseSymbols() such that extern weak symbols appear |
122 | // last, so we don't need to worry about subsequent symbols being |
123 | // added to an already-coalesced section. |
124 | if (defined->isec()) |
125 | transplantSymbolsAtOffset(fromIsec: concatIsec, toIsec: defined->isec(), |
126 | /*skip=*/nullptr, fromOff: value, toOff: defined->value); |
127 | } |
128 | return defined; |
129 | } |
130 | |
131 | if (defined->isWeakDef()) { |
132 | if (auto concatIsec = |
133 | dyn_cast_or_null<ConcatInputSection>(Val: defined->isec())) { |
134 | concatIsec->wasCoalesced = true; |
135 | if (isec) |
136 | transplantSymbolsAtOffset(fromIsec: concatIsec, toIsec: isec, skip: defined, fromOff: defined->value, |
137 | toOff: value); |
138 | } |
139 | } else { |
140 | std::string srcLoc1 = defined->getSourceLocation(); |
141 | std::string srcLoc2 = isec ? isec->getSourceLocation(off: value) : "" ; |
142 | std::string srcFile1 = toString(file: defined->getFile()); |
143 | std::string srcFile2 = toString(file); |
144 | |
145 | dupSymDiags.push_back(Elt: {make_pair(x&: srcLoc1, y&: srcFile1), |
146 | make_pair(x&: srcLoc2, y&: srcFile2), defined}); |
147 | } |
148 | |
149 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
150 | overridesWeakDef = !isWeakDef && dysym->isWeakDef(); |
151 | dysym->unreference(); |
152 | } else if (auto *undef = dyn_cast<Undefined>(Val: s)) { |
153 | if (undef->wasBitcodeSymbol) { |
154 | auto objFile = dyn_cast<ObjFile>(Val: file); |
155 | if (!objFile) { |
156 | // The file must be a native object file, as opposed to potentially |
157 | // being another bitcode file. A situation arises when some symbols |
158 | // are defined thru `module asm` and thus they are not present in the |
159 | // bitcode's symbol table. Consider bitcode modules `A`, `B`, and `C`. |
160 | // LTO compiles only `A` and `C`, since there's no explicit symbol |
161 | // reference to `B` other than a symbol from `A` via `module asm`. |
162 | // After LTO is finished, the missing symbol now appears in the |
163 | // resulting object file for `A`, which prematurely resolves another |
164 | // prevailing symbol with `B` that hasn't been compiled, instead of |
165 | // the resulting object for `C`. Consequently, an incorrect |
166 | // relocation is generated for the prevailing symbol. |
167 | assert(isa<BitcodeFile>(file) && "Bitcode file is expected." ); |
168 | std::string message = |
169 | "The pending prevailing symbol(" + name.str() + |
170 | ") in the bitcode file(" + toString(file: undef->getFile()) + |
171 | ") is overridden by a non-native object (from bitcode): " + |
172 | toString(file); |
173 | error(msg: message); |
174 | } else if (!objFile->builtFromBitcode) { |
175 | // Ideally, this should be an object file compiled from a bitcode |
176 | // file. However, this might not hold true if a LC linker option is |
177 | // used. In case LTO internalizes a prevailing hidden weak symbol, |
178 | // there's a situation where an unresolved prevailing symbol might be |
179 | // linked with the corresponding one from a native library, which is |
180 | // loaded later after LTO. Although this could potentially result in |
181 | // an ODR violation, we choose to permit this scenario as a warning. |
182 | std::string message = "The pending prevailing symbol(" + name.str() + |
183 | ") in the bitcode file(" + |
184 | toString(file: undef->getFile()) + |
185 | ") is overridden by a post-processed native " |
186 | "object (from native archive): " + |
187 | toString(file); |
188 | warn(msg: message); |
189 | } else { |
190 | // Preserve the original bitcode file name (instead of using the |
191 | // object file name). |
192 | file = undef->getFile(); |
193 | } |
194 | } |
195 | } |
196 | // Defined symbols take priority over other types of symbols, so in case |
197 | // of a name conflict, we fall through to the replaceSymbol() call below. |
198 | } |
199 | |
200 | // With -flat_namespace, all extern symbols in dylibs are interposable. |
201 | // FIXME: Add support for `-interposable` (PR53680). |
202 | bool interposable = config->namespaceKind == NamespaceKind::flat && |
203 | config->outputType != MachO::MH_EXECUTE && |
204 | !isPrivateExtern; |
205 | Defined *defined = replaceSymbol<Defined>( |
206 | s, arg&: name, arg&: file, arg&: isec, arg&: value, arg&: size, arg&: isWeakDef, /*isExternal=*/arg: true, |
207 | arg&: isPrivateExtern, /*includeInSymtab=*/arg: true, arg&: isReferencedDynamically, |
208 | arg&: noDeadStrip, arg&: overridesWeakDef, arg&: isWeakDefCanBeHidden, arg&: interposable); |
209 | return defined; |
210 | } |
211 | |
212 | Defined *SymbolTable::aliasDefined(Defined *src, StringRef target, |
213 | InputFile *newFile, bool makePrivateExtern) { |
214 | bool isPrivateExtern = makePrivateExtern || src->privateExtern; |
215 | return addDefined(name: target, file: newFile, isec: src->isec(), value: src->value, size: src->size, |
216 | isWeakDef: src->isWeakDef(), isPrivateExtern, |
217 | isReferencedDynamically: src->referencedDynamically, noDeadStrip: src->noDeadStrip, |
218 | isWeakDefCanBeHidden: src->weakDefCanBeHidden); |
219 | } |
220 | |
221 | Symbol *SymbolTable::addUndefined(StringRef name, InputFile *file, |
222 | bool isWeakRef) { |
223 | auto [s, wasInserted] = insert(name, file); |
224 | |
225 | RefState refState = isWeakRef ? RefState::Weak : RefState::Strong; |
226 | |
227 | if (wasInserted) |
228 | replaceSymbol<Undefined>(s, arg&: name, arg&: file, arg&: refState, |
229 | /*wasBitcodeSymbol=*/arg: false); |
230 | else if (auto *lazy = dyn_cast<LazyArchive>(Val: s)) |
231 | lazy->fetchArchiveMember(); |
232 | else if (isa<LazyObject>(Val: s)) |
233 | extract(file&: *s->getFile(), reason: s->getName()); |
234 | else if (auto *dynsym = dyn_cast<DylibSymbol>(Val: s)) |
235 | dynsym->reference(newState: refState); |
236 | else if (auto *undefined = dyn_cast<Undefined>(Val: s)) |
237 | undefined->refState = std::max(a: undefined->refState, b: refState); |
238 | return s; |
239 | } |
240 | |
241 | Symbol *SymbolTable::addCommon(StringRef name, InputFile *file, uint64_t size, |
242 | uint32_t align, bool isPrivateExtern) { |
243 | auto [s, wasInserted] = insert(name, file); |
244 | |
245 | if (!wasInserted) { |
246 | if (auto *common = dyn_cast<CommonSymbol>(Val: s)) { |
247 | if (size < common->size) |
248 | return s; |
249 | } else if (isa<Defined>(Val: s)) { |
250 | return s; |
251 | } |
252 | // Common symbols take priority over all non-Defined symbols, so in case of |
253 | // a name conflict, we fall through to the replaceSymbol() call below. |
254 | } |
255 | |
256 | replaceSymbol<CommonSymbol>(s, arg&: name, arg&: file, arg&: size, arg&: align, arg&: isPrivateExtern); |
257 | return s; |
258 | } |
259 | |
260 | Symbol *SymbolTable::addDylib(StringRef name, DylibFile *file, bool isWeakDef, |
261 | bool isTlv) { |
262 | auto [s, wasInserted] = insert(name, file); |
263 | |
264 | RefState refState = RefState::Unreferenced; |
265 | if (!wasInserted) { |
266 | if (auto *defined = dyn_cast<Defined>(Val: s)) { |
267 | if (isWeakDef && !defined->isWeakDef()) |
268 | defined->overridesWeakDef = true; |
269 | } else if (auto *undefined = dyn_cast<Undefined>(Val: s)) { |
270 | refState = undefined->refState; |
271 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
272 | refState = dysym->getRefState(); |
273 | } |
274 | } |
275 | |
276 | bool isDynamicLookup = file == nullptr; |
277 | if (wasInserted || isa<Undefined>(Val: s) || |
278 | (isa<DylibSymbol>(Val: s) && |
279 | ((!isWeakDef && s->isWeakDef()) || |
280 | (!isDynamicLookup && cast<DylibSymbol>(Val: s)->isDynamicLookup())))) { |
281 | if (auto *dynsym = dyn_cast<DylibSymbol>(Val: s)) |
282 | dynsym->unreference(); |
283 | replaceSymbol<DylibSymbol>(s, arg&: file, arg&: name, arg&: isWeakDef, arg&: refState, arg&: isTlv); |
284 | } |
285 | |
286 | return s; |
287 | } |
288 | |
289 | Symbol *SymbolTable::addDynamicLookup(StringRef name) { |
290 | return addDylib(name, /*file=*/nullptr, /*isWeakDef=*/false, /*isTlv=*/false); |
291 | } |
292 | |
293 | Symbol *SymbolTable::addLazyArchive(StringRef name, ArchiveFile *file, |
294 | const object::Archive::Symbol &sym) { |
295 | auto [s, wasInserted] = insert(name, file); |
296 | |
297 | if (wasInserted) { |
298 | replaceSymbol<LazyArchive>(s, arg&: file, arg: sym); |
299 | } else if (isa<Undefined>(Val: s)) { |
300 | file->fetch(sym); |
301 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
302 | if (dysym->isWeakDef()) { |
303 | if (dysym->getRefState() != RefState::Unreferenced) |
304 | file->fetch(sym); |
305 | else |
306 | replaceSymbol<LazyArchive>(s, arg&: file, arg: sym); |
307 | } |
308 | } |
309 | return s; |
310 | } |
311 | |
312 | Symbol *SymbolTable::addLazyObject(StringRef name, InputFile &file) { |
313 | auto [s, wasInserted] = insert(name, file: &file); |
314 | |
315 | if (wasInserted) { |
316 | replaceSymbol<LazyObject>(s, arg&: file, arg&: name); |
317 | } else if (isa<Undefined>(Val: s)) { |
318 | extract(file, reason: name); |
319 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
320 | if (dysym->isWeakDef()) { |
321 | if (dysym->getRefState() != RefState::Unreferenced) |
322 | extract(file, reason: name); |
323 | else |
324 | replaceSymbol<LazyObject>(s, arg&: file, arg&: name); |
325 | } |
326 | } |
327 | return s; |
328 | } |
329 | |
330 | Defined *SymbolTable::addSynthetic(StringRef name, InputSection *isec, |
331 | uint64_t value, bool isPrivateExtern, |
332 | bool includeInSymtab, |
333 | bool referencedDynamically) { |
334 | assert(!isec || !isec->getFile()); // See makeSyntheticInputSection(). |
335 | Defined *s = addDefined(name, /*file=*/nullptr, isec, value, /*size=*/0, |
336 | /*isWeakDef=*/false, isPrivateExtern, |
337 | isReferencedDynamically: referencedDynamically, /*noDeadStrip=*/false, |
338 | /*isWeakDefCanBeHidden=*/false); |
339 | s->includeInSymtab = includeInSymtab; |
340 | return s; |
341 | } |
342 | |
343 | enum class Boundary { |
344 | Start, |
345 | End, |
346 | }; |
347 | |
348 | static Defined *createBoundarySymbol(const Undefined &sym) { |
349 | return symtab->addSynthetic( |
350 | name: sym.getName(), /*isec=*/nullptr, /*value=*/-1, /*isPrivateExtern=*/true, |
351 | /*includeInSymtab=*/false, /*referencedDynamically=*/false); |
352 | } |
353 | |
354 | static void handleSectionBoundarySymbol(const Undefined &sym, StringRef segSect, |
355 | Boundary which) { |
356 | auto [segName, sectName] = segSect.split(Separator: '$'); |
357 | |
358 | // Attach the symbol to any InputSection that will end up in the right |
359 | // OutputSection -- it doesn't matter which one we pick. |
360 | // Don't bother looking through inputSections for a matching |
361 | // ConcatInputSection -- we need to create ConcatInputSection for |
362 | // non-existing sections anyways, and that codepath works even if we should |
363 | // already have a ConcatInputSection with the right name. |
364 | |
365 | OutputSection *osec = nullptr; |
366 | // This looks for __TEXT,__cstring etc. |
367 | for (SyntheticSection *ssec : syntheticSections) |
368 | if (ssec->segname == segName && ssec->name == sectName) { |
369 | osec = ssec->isec->parent; |
370 | break; |
371 | } |
372 | |
373 | if (!osec) { |
374 | ConcatInputSection *isec = makeSyntheticInputSection(segName, sectName); |
375 | |
376 | // This runs after markLive() and is only called for Undefineds that are |
377 | // live. Marking the isec live ensures an OutputSection is created that the |
378 | // start/end symbol can refer to. |
379 | assert(sym.isLive()); |
380 | assert(isec->live); |
381 | |
382 | // This runs after gatherInputSections(), so need to explicitly set parent |
383 | // and add to inputSections. |
384 | osec = isec->parent = ConcatOutputSection::getOrCreateForInput(isec); |
385 | inputSections.push_back(x: isec); |
386 | } |
387 | |
388 | if (which == Boundary::Start) |
389 | osec->sectionStartSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
390 | else |
391 | osec->sectionEndSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
392 | } |
393 | |
394 | static void handleSegmentBoundarySymbol(const Undefined &sym, StringRef segName, |
395 | Boundary which) { |
396 | OutputSegment *seg = getOrCreateOutputSegment(name: segName); |
397 | if (which == Boundary::Start) |
398 | seg->segmentStartSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
399 | else |
400 | seg->segmentEndSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
401 | } |
402 | |
403 | // Try to find a definition for an undefined symbol. |
404 | // Returns true if a definition was found and no diagnostics are needed. |
405 | static bool recoverFromUndefinedSymbol(const Undefined &sym) { |
406 | // Handle start/end symbols. |
407 | StringRef name = sym.getName(); |
408 | if (name.consume_front(Prefix: "section$start$" )) { |
409 | handleSectionBoundarySymbol(sym, segSect: name, which: Boundary::Start); |
410 | return true; |
411 | } |
412 | if (name.consume_front(Prefix: "section$end$" )) { |
413 | handleSectionBoundarySymbol(sym, segSect: name, which: Boundary::End); |
414 | return true; |
415 | } |
416 | if (name.consume_front(Prefix: "segment$start$" )) { |
417 | handleSegmentBoundarySymbol(sym, segName: name, which: Boundary::Start); |
418 | return true; |
419 | } |
420 | if (name.consume_front(Prefix: "segment$end$" )) { |
421 | handleSegmentBoundarySymbol(sym, segName: name, which: Boundary::End); |
422 | return true; |
423 | } |
424 | |
425 | // Leave dtrace symbols, since we will handle them when we do the relocation |
426 | if (name.starts_with(Prefix: "___dtrace_" )) |
427 | return true; |
428 | |
429 | // Handle -U. |
430 | if (config->explicitDynamicLookups.count(Key: sym.getName())) { |
431 | symtab->addDynamicLookup(name: sym.getName()); |
432 | return true; |
433 | } |
434 | |
435 | // Handle -undefined. |
436 | if (config->undefinedSymbolTreatment == |
437 | UndefinedSymbolTreatment::dynamic_lookup || |
438 | config->undefinedSymbolTreatment == UndefinedSymbolTreatment::suppress) { |
439 | symtab->addDynamicLookup(name: sym.getName()); |
440 | return true; |
441 | } |
442 | |
443 | // We do not return true here, as we still need to print diagnostics. |
444 | if (config->undefinedSymbolTreatment == UndefinedSymbolTreatment::warning) |
445 | symtab->addDynamicLookup(name: sym.getName()); |
446 | |
447 | return false; |
448 | } |
449 | |
450 | namespace { |
451 | struct UndefinedDiag { |
452 | struct SectionAndOffset { |
453 | const InputSection *isec; |
454 | uint64_t offset; |
455 | }; |
456 | |
457 | std::vector<SectionAndOffset> codeReferences; |
458 | std::vector<std::string> otherReferences; |
459 | }; |
460 | |
461 | MapVector<const Undefined *, UndefinedDiag> undefs; |
462 | } // namespace |
463 | |
464 | void macho::reportPendingDuplicateSymbols() { |
465 | for (const auto &duplicate : dupSymDiags) { |
466 | if (!config->deadStripDuplicates || duplicate.sym->isLive()) { |
467 | std::string message = |
468 | "duplicate symbol: " + toString(*duplicate.sym) + "\n>>> defined in " ; |
469 | if (!duplicate.src1.first.empty()) |
470 | message += duplicate.src1.first + "\n>>> " ; |
471 | message += duplicate.src1.second + "\n>>> defined in " ; |
472 | if (!duplicate.src2.first.empty()) |
473 | message += duplicate.src2.first + "\n>>> " ; |
474 | error(msg: message + duplicate.src2.second); |
475 | } |
476 | } |
477 | } |
478 | |
479 | // Check whether the definition name def is a mangled function name that matches |
480 | // the reference name ref. |
481 | static bool canSuggestExternCForCXX(StringRef ref, StringRef def) { |
482 | llvm::ItaniumPartialDemangler d; |
483 | std::string name = def.str(); |
484 | if (d.partialDemangle(MangledName: name.c_str())) |
485 | return false; |
486 | char *buf = d.getFunctionName(Buf: nullptr, N: nullptr); |
487 | if (!buf) |
488 | return false; |
489 | bool ret = ref == buf; |
490 | free(ptr: buf); |
491 | return ret; |
492 | } |
493 | |
494 | // Suggest an alternative spelling of an "undefined symbol" diagnostic. Returns |
495 | // the suggested symbol, which is either in the symbol table, or in the same |
496 | // file of sym. |
497 | static const Symbol *getAlternativeSpelling(const Undefined &sym, |
498 | std::string &preHint, |
499 | std::string &postHint) { |
500 | DenseMap<StringRef, const Symbol *> map; |
501 | if (sym.getFile() && sym.getFile()->kind() == InputFile::ObjKind) { |
502 | // Build a map of local defined symbols. |
503 | for (const Symbol *s : sym.getFile()->symbols) |
504 | if (auto *defined = dyn_cast_or_null<Defined>(Val: s)) |
505 | if (!defined->isExternal()) |
506 | map.try_emplace(Key: s->getName(), Args&: s); |
507 | } |
508 | |
509 | auto suggest = [&](StringRef newName) -> const Symbol * { |
510 | // If defined locally. |
511 | if (const Symbol *s = map.lookup(Val: newName)) |
512 | return s; |
513 | |
514 | // If in the symbol table and not undefined. |
515 | if (const Symbol *s = symtab->find(name: newName)) |
516 | if (dyn_cast<Undefined>(Val: s) == nullptr) |
517 | return s; |
518 | |
519 | return nullptr; |
520 | }; |
521 | |
522 | // This loop enumerates all strings of Levenshtein distance 1 as typo |
523 | // correction candidates and suggests the one that exists as a non-undefined |
524 | // symbol. |
525 | StringRef name = sym.getName(); |
526 | for (size_t i = 0, e = name.size(); i != e + 1; ++i) { |
527 | // Insert a character before name[i]. |
528 | std::string newName = (name.substr(Start: 0, N: i) + "0" + name.substr(Start: i)).str(); |
529 | for (char c = '0'; c <= 'z'; ++c) { |
530 | newName[i] = c; |
531 | if (const Symbol *s = suggest(newName)) |
532 | return s; |
533 | } |
534 | if (i == e) |
535 | break; |
536 | |
537 | // Substitute name[i]. |
538 | newName = std::string(name); |
539 | for (char c = '0'; c <= 'z'; ++c) { |
540 | newName[i] = c; |
541 | if (const Symbol *s = suggest(newName)) |
542 | return s; |
543 | } |
544 | |
545 | // Transpose name[i] and name[i+1]. This is of edit distance 2 but it is |
546 | // common. |
547 | if (i + 1 < e) { |
548 | newName[i] = name[i + 1]; |
549 | newName[i + 1] = name[i]; |
550 | if (const Symbol *s = suggest(newName)) |
551 | return s; |
552 | } |
553 | |
554 | // Delete name[i]. |
555 | newName = (name.substr(Start: 0, N: i) + name.substr(Start: i + 1)).str(); |
556 | if (const Symbol *s = suggest(newName)) |
557 | return s; |
558 | } |
559 | |
560 | // Case mismatch, e.g. Foo vs FOO. |
561 | for (auto &it : map) |
562 | if (name.equals_insensitive(RHS: it.first)) |
563 | return it.second; |
564 | for (Symbol *sym : symtab->getSymbols()) |
565 | if (dyn_cast<Undefined>(Val: sym) == nullptr && |
566 | name.equals_insensitive(RHS: sym->getName())) |
567 | return sym; |
568 | |
569 | // The reference may be a mangled name while the definition is not. Suggest a |
570 | // missing extern "C". |
571 | if (name.starts_with(Prefix: "__Z" )) { |
572 | std::string buf = name.str(); |
573 | llvm::ItaniumPartialDemangler d; |
574 | if (!d.partialDemangle(MangledName: buf.c_str())) |
575 | if (char *buf = d.getFunctionName(Buf: nullptr, N: nullptr)) { |
576 | const Symbol *s = suggest((Twine("_" ) + buf).str()); |
577 | free(ptr: buf); |
578 | if (s) { |
579 | preHint = ": extern \"C\" " ; |
580 | return s; |
581 | } |
582 | } |
583 | } else { |
584 | StringRef nameWithoutUnderscore = name; |
585 | nameWithoutUnderscore.consume_front(Prefix: "_" ); |
586 | const Symbol *s = nullptr; |
587 | for (auto &it : map) |
588 | if (canSuggestExternCForCXX(ref: nameWithoutUnderscore, def: it.first)) { |
589 | s = it.second; |
590 | break; |
591 | } |
592 | if (!s) |
593 | for (Symbol *sym : symtab->getSymbols()) |
594 | if (canSuggestExternCForCXX(ref: nameWithoutUnderscore, def: sym->getName())) { |
595 | s = sym; |
596 | break; |
597 | } |
598 | if (s) { |
599 | preHint = " to declare " ; |
600 | postHint = " as extern \"C\"?" ; |
601 | return s; |
602 | } |
603 | } |
604 | |
605 | return nullptr; |
606 | } |
607 | |
608 | static void reportUndefinedSymbol(const Undefined &sym, |
609 | const UndefinedDiag &locations, |
610 | bool correctSpelling) { |
611 | std::string message = "undefined symbol" ; |
612 | if (config->archMultiple) |
613 | message += (" for arch " + getArchitectureName(Arch: config->arch())).str(); |
614 | message += ": " + toString(sym); |
615 | |
616 | const size_t maxUndefinedReferences = 3; |
617 | size_t i = 0; |
618 | for (const std::string &loc : locations.otherReferences) { |
619 | if (i >= maxUndefinedReferences) |
620 | break; |
621 | message += "\n>>> referenced by " + loc; |
622 | ++i; |
623 | } |
624 | |
625 | for (const UndefinedDiag::SectionAndOffset &loc : locations.codeReferences) { |
626 | if (i >= maxUndefinedReferences) |
627 | break; |
628 | message += "\n>>> referenced by " ; |
629 | std::string src = loc.isec->getSourceLocation(off: loc.offset); |
630 | if (!src.empty()) |
631 | message += src + "\n>>> " ; |
632 | message += loc.isec->getLocation(off: loc.offset); |
633 | ++i; |
634 | } |
635 | |
636 | size_t totalReferences = |
637 | locations.otherReferences.size() + locations.codeReferences.size(); |
638 | if (totalReferences > i) |
639 | message += |
640 | ("\n>>> referenced " + Twine(totalReferences - i) + " more times" ) |
641 | .str(); |
642 | |
643 | if (correctSpelling) { |
644 | std::string preHint = ": " , postHint; |
645 | if (const Symbol *corrected = |
646 | getAlternativeSpelling(sym, preHint, postHint)) { |
647 | message += |
648 | "\n>>> did you mean" + preHint + toString(*corrected) + postHint; |
649 | if (corrected->getFile()) |
650 | message += "\n>>> defined in: " + toString(file: corrected->getFile()); |
651 | } |
652 | } |
653 | |
654 | if (config->undefinedSymbolTreatment == UndefinedSymbolTreatment::error) |
655 | error(msg: message); |
656 | else if (config->undefinedSymbolTreatment == |
657 | UndefinedSymbolTreatment::warning) |
658 | warn(msg: message); |
659 | else |
660 | assert(false && "diagnostics make sense for -undefined error|warning only" ); |
661 | } |
662 | |
663 | void macho::reportPendingUndefinedSymbols() { |
664 | // Enable spell corrector for the first 2 diagnostics. |
665 | for (const auto &[i, undef] : llvm::enumerate(First&: undefs)) |
666 | reportUndefinedSymbol(sym: *undef.first, locations: undef.second, correctSpelling: i < 2); |
667 | |
668 | // This function is called multiple times during execution. Clear the printed |
669 | // diagnostics to avoid printing the same things again the next time. |
670 | undefs.clear(); |
671 | } |
672 | |
673 | void macho::treatUndefinedSymbol(const Undefined &sym, StringRef source) { |
674 | if (recoverFromUndefinedSymbol(sym)) |
675 | return; |
676 | |
677 | undefs[&sym].otherReferences.push_back(x: source.str()); |
678 | } |
679 | |
680 | void macho::treatUndefinedSymbol(const Undefined &sym, const InputSection *isec, |
681 | uint64_t offset) { |
682 | if (recoverFromUndefinedSymbol(sym)) |
683 | return; |
684 | |
685 | undefs[&sym].codeReferences.push_back(x: {.isec: isec, .offset: offset}); |
686 | } |
687 | |
688 | std::unique_ptr<SymbolTable> macho::symtab; |
689 | |