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 symSucceedsOff = [](uint64_t off, const Symbol *s) { |
71 | return cast<Defined>(Val: s)->value > off; |
72 | }; |
73 | assert(std::is_partitioned(toIsec->symbols.begin(), toIsec->symbols.end(), |
74 | [symSucceedsOff, toOff](const Symbol *s) { |
75 | return !symSucceedsOff(toOff, s); |
76 | }) && |
77 | "Symbols in toIsec must be partitioned by toOff."); |
78 | auto insertIt = llvm::upper_bound(Range&: toIsec->symbols, Value&: toOff, C: symSucceedsOff); |
79 | llvm::erase_if(C&: fromIsec->symbols, P: [&](Symbol *s) { |
80 | auto *d = cast<Defined>(Val: s); |
81 | if (d->value != fromOff) |
82 | return false; |
83 | if (d != skip) { |
84 | // This repeated insertion will be quadratic unless insertIt is the end |
85 | // iterator. However, that is typically the case for files that have |
86 | // .subsections_via_symbols set. |
87 | insertIt = toIsec->symbols.insert(I: insertIt, Elt: d); |
88 | d->originalIsec = toIsec; |
89 | d->value = toOff; |
90 | // We don't want to have more than one unwindEntry at a given address, so |
91 | // drop the redundant ones. We We can safely drop the unwindEntries of |
92 | // the symbols in fromIsec since we will be adding another unwindEntry as |
93 | // we finish parsing toIsec's file. (We can assume that toIsec has its |
94 | // own unwindEntry because of the ODR.) |
95 | d->originalUnwindEntry = nullptr; |
96 | } |
97 | return true; |
98 | }); |
99 | } |
100 | |
101 | Defined *SymbolTable::addDefined(StringRef name, InputFile *file, |
102 | InputSection *isec, uint64_t value, |
103 | uint64_t size, bool isWeakDef, |
104 | bool isPrivateExtern, |
105 | bool isReferencedDynamically, bool noDeadStrip, |
106 | bool isWeakDefCanBeHidden) { |
107 | bool overridesWeakDef = false; |
108 | auto [s, wasInserted] = insert(name, file); |
109 | |
110 | assert(!file || !isa<BitcodeFile>(file) || !isec); |
111 | |
112 | if (!wasInserted) { |
113 | if (auto *defined = dyn_cast<Defined>(Val: s)) { |
114 | if (isWeakDef) { |
115 | // See further comment in createDefined() in InputFiles.cpp |
116 | if (defined->isWeakDef()) { |
117 | defined->privateExtern &= isPrivateExtern; |
118 | defined->weakDefCanBeHidden &= isWeakDefCanBeHidden; |
119 | defined->referencedDynamically |= isReferencedDynamically; |
120 | defined->noDeadStrip |= noDeadStrip; |
121 | } |
122 | if (auto concatIsec = dyn_cast_or_null<ConcatInputSection>(Val: isec)) { |
123 | concatIsec->wasCoalesced = true; |
124 | // Any local symbols that alias the coalesced symbol should be moved |
125 | // into the prevailing section. Note that we have sorted the symbols |
126 | // in ObjFile::parseSymbols() such that extern weak symbols appear |
127 | // last, so we don't need to worry about subsequent symbols being |
128 | // added to an already-coalesced section. |
129 | if (defined->isec()) |
130 | transplantSymbolsAtOffset(fromIsec: concatIsec, toIsec: defined->isec(), |
131 | /*skip=*/nullptr, fromOff: value, toOff: defined->value); |
132 | } |
133 | return defined; |
134 | } |
135 | |
136 | if (defined->isWeakDef()) { |
137 | if (auto concatIsec = |
138 | dyn_cast_or_null<ConcatInputSection>(Val: defined->isec())) { |
139 | concatIsec->wasCoalesced = true; |
140 | if (isec) |
141 | transplantSymbolsAtOffset(fromIsec: concatIsec, toIsec: isec, skip: defined, fromOff: defined->value, |
142 | toOff: value); |
143 | } |
144 | } else { |
145 | std::string srcLoc1 = defined->getSourceLocation(); |
146 | std::string srcLoc2 = isec ? isec->getSourceLocation(off: value) : ""; |
147 | std::string srcFile1 = toString(file: defined->getFile()); |
148 | std::string srcFile2 = toString(file); |
149 | |
150 | dupSymDiags.push_back(Elt: {make_pair(x&: srcLoc1, y&: srcFile1), |
151 | make_pair(x&: srcLoc2, y&: srcFile2), defined}); |
152 | } |
153 | |
154 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
155 | overridesWeakDef = !isWeakDef && dysym->isWeakDef(); |
156 | dysym->unreference(); |
157 | } else if (auto *undef = dyn_cast<Undefined>(Val: s)) { |
158 | if (undef->wasBitcodeSymbol) { |
159 | auto objFile = dyn_cast<ObjFile>(Val: file); |
160 | if (!objFile) { |
161 | // The file must be a native object file, as opposed to potentially |
162 | // being another bitcode file. A situation arises when some symbols |
163 | // are defined thru `module asm` and thus they are not present in the |
164 | // bitcode's symbol table. Consider bitcode modules `A`, `B`, and `C`. |
165 | // LTO compiles only `A` and `C`, since there's no explicit symbol |
166 | // reference to `B` other than a symbol from `A` via `module asm`. |
167 | // After LTO is finished, the missing symbol now appears in the |
168 | // resulting object file for `A`, which prematurely resolves another |
169 | // prevailing symbol with `B` that hasn't been compiled, instead of |
170 | // the resulting object for `C`. Consequently, an incorrect |
171 | // relocation is generated for the prevailing symbol. |
172 | assert(isa<BitcodeFile>(file) && "Bitcode file is expected."); |
173 | std::string message = |
174 | "The pending prevailing symbol("+ name.str() + |
175 | ") in the bitcode file("+ toString(file: undef->getFile()) + |
176 | ") is overridden by a non-native object (from bitcode): "+ |
177 | toString(file); |
178 | error(msg: message); |
179 | } else if (!objFile->builtFromBitcode) { |
180 | // Ideally, this should be an object file compiled from a bitcode |
181 | // file. However, this might not hold true if a LC linker option is |
182 | // used. In case LTO internalizes a prevailing hidden weak symbol, |
183 | // there's a situation where an unresolved prevailing symbol might be |
184 | // linked with the corresponding one from a native library, which is |
185 | // loaded later after LTO. Although this could potentially result in |
186 | // an ODR violation, we choose to permit this scenario as a warning. |
187 | std::string message = "The pending prevailing symbol("+ name.str() + |
188 | ") in the bitcode file("+ |
189 | toString(file: undef->getFile()) + |
190 | ") is overridden by a post-processed native " |
191 | "object (from native archive): "+ |
192 | toString(file); |
193 | warn(msg: message); |
194 | } else { |
195 | // Preserve the original bitcode file name (instead of using the |
196 | // object file name). |
197 | file = undef->getFile(); |
198 | } |
199 | } |
200 | } |
201 | // Defined symbols take priority over other types of symbols, so in case |
202 | // of a name conflict, we fall through to the replaceSymbol() call below. |
203 | } |
204 | |
205 | // With -flat_namespace, all extern symbols in dylibs are interposable. |
206 | bool interposable = ((config->namespaceKind == NamespaceKind::flat && |
207 | config->outputType != MachO::MH_EXECUTE) || |
208 | config->interposable) && |
209 | !isPrivateExtern; |
210 | Defined *defined = replaceSymbol<Defined>( |
211 | s, arg&: name, arg&: file, arg&: isec, arg&: value, arg&: size, arg&: isWeakDef, /*isExternal=*/arg: true, |
212 | arg&: isPrivateExtern, /*includeInSymtab=*/arg: true, arg&: isReferencedDynamically, |
213 | arg&: noDeadStrip, arg&: overridesWeakDef, arg&: isWeakDefCanBeHidden, arg&: interposable); |
214 | return defined; |
215 | } |
216 | |
217 | Defined *SymbolTable::aliasDefined(Defined *src, StringRef target, |
218 | InputFile *newFile, bool makePrivateExtern) { |
219 | bool isPrivateExtern = makePrivateExtern || src->privateExtern; |
220 | return addDefined(name: target, file: newFile, isec: src->isec(), value: src->value, size: src->size, |
221 | isWeakDef: src->isWeakDef(), isPrivateExtern, |
222 | isReferencedDynamically: src->referencedDynamically, noDeadStrip: src->noDeadStrip, |
223 | isWeakDefCanBeHidden: src->weakDefCanBeHidden); |
224 | } |
225 | |
226 | Symbol *SymbolTable::addUndefined(StringRef name, InputFile *file, |
227 | bool isWeakRef) { |
228 | auto [s, wasInserted] = insert(name, file); |
229 | |
230 | RefState refState = isWeakRef ? RefState::Weak : RefState::Strong; |
231 | |
232 | if (wasInserted) |
233 | replaceSymbol<Undefined>(s, arg&: name, arg&: file, arg&: refState, |
234 | /*wasBitcodeSymbol=*/arg: false); |
235 | else if (auto *lazy = dyn_cast<LazyArchive>(Val: s)) |
236 | lazy->fetchArchiveMember(); |
237 | else if (isa<LazyObject>(Val: s)) |
238 | extract(file&: *s->getFile(), reason: s->getName()); |
239 | else if (auto *dynsym = dyn_cast<DylibSymbol>(Val: s)) |
240 | dynsym->reference(newState: refState); |
241 | else if (auto *undefined = dyn_cast<Undefined>(Val: s)) |
242 | undefined->refState = std::max(a: undefined->refState, b: refState); |
243 | return s; |
244 | } |
245 | |
246 | Symbol *SymbolTable::addCommon(StringRef name, InputFile *file, uint64_t size, |
247 | uint32_t align, bool isPrivateExtern) { |
248 | auto [s, wasInserted] = insert(name, file); |
249 | |
250 | if (!wasInserted) { |
251 | if (auto *common = dyn_cast<CommonSymbol>(Val: s)) { |
252 | if (size < common->size) |
253 | return s; |
254 | } else if (isa<Defined>(Val: s)) { |
255 | return s; |
256 | } |
257 | // Common symbols take priority over all non-Defined symbols, so in case of |
258 | // a name conflict, we fall through to the replaceSymbol() call below. |
259 | } |
260 | |
261 | replaceSymbol<CommonSymbol>(s, arg&: name, arg&: file, arg&: size, arg&: align, arg&: isPrivateExtern); |
262 | return s; |
263 | } |
264 | |
265 | Symbol *SymbolTable::addDylib(StringRef name, DylibFile *file, bool isWeakDef, |
266 | bool isTlv) { |
267 | auto [s, wasInserted] = insert(name, file); |
268 | |
269 | RefState refState = RefState::Unreferenced; |
270 | if (!wasInserted) { |
271 | if (auto *defined = dyn_cast<Defined>(Val: s)) { |
272 | if (isWeakDef && !defined->isWeakDef()) |
273 | defined->overridesWeakDef = true; |
274 | } else if (auto *undefined = dyn_cast<Undefined>(Val: s)) { |
275 | refState = undefined->refState; |
276 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
277 | refState = dysym->getRefState(); |
278 | } |
279 | } |
280 | |
281 | bool isDynamicLookup = file == nullptr; |
282 | if (wasInserted || isa<Undefined>(Val: s) || |
283 | (isa<DylibSymbol>(Val: s) && |
284 | ((!isWeakDef && s->isWeakDef()) || |
285 | (!isDynamicLookup && cast<DylibSymbol>(Val: s)->isDynamicLookup())))) { |
286 | if (auto *dynsym = dyn_cast<DylibSymbol>(Val: s)) |
287 | dynsym->unreference(); |
288 | replaceSymbol<DylibSymbol>(s, arg&: file, arg&: name, arg&: isWeakDef, arg&: refState, arg&: isTlv); |
289 | } |
290 | |
291 | return s; |
292 | } |
293 | |
294 | Symbol *SymbolTable::addDynamicLookup(StringRef name) { |
295 | return addDylib(name, /*file=*/nullptr, /*isWeakDef=*/false, /*isTlv=*/false); |
296 | } |
297 | |
298 | Symbol *SymbolTable::addLazyArchive(StringRef name, ArchiveFile *file, |
299 | const object::Archive::Symbol &sym) { |
300 | auto [s, wasInserted] = insert(name, file); |
301 | |
302 | if (wasInserted) { |
303 | replaceSymbol<LazyArchive>(s, arg&: file, arg: sym); |
304 | } else if (isa<Undefined>(Val: s)) { |
305 | file->fetch(sym); |
306 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
307 | if (dysym->isWeakDef()) { |
308 | if (dysym->getRefState() != RefState::Unreferenced) |
309 | file->fetch(sym); |
310 | else |
311 | replaceSymbol<LazyArchive>(s, arg&: file, arg: sym); |
312 | } |
313 | } |
314 | return s; |
315 | } |
316 | |
317 | Symbol *SymbolTable::addLazyObject(StringRef name, InputFile &file) { |
318 | auto [s, wasInserted] = insert(name, file: &file); |
319 | |
320 | if (wasInserted) { |
321 | replaceSymbol<LazyObject>(s, arg&: file, arg&: name); |
322 | } else if (isa<Undefined>(Val: s)) { |
323 | extract(file, reason: name); |
324 | } else if (auto *dysym = dyn_cast<DylibSymbol>(Val: s)) { |
325 | if (dysym->isWeakDef()) { |
326 | if (dysym->getRefState() != RefState::Unreferenced) |
327 | extract(file, reason: name); |
328 | else |
329 | replaceSymbol<LazyObject>(s, arg&: file, arg&: name); |
330 | } |
331 | } |
332 | return s; |
333 | } |
334 | |
335 | Defined *SymbolTable::addSynthetic(StringRef name, InputSection *isec, |
336 | uint64_t value, bool isPrivateExtern, |
337 | bool includeInSymtab, |
338 | bool referencedDynamically) { |
339 | assert(!isec || !isec->getFile()); // See makeSyntheticInputSection(). |
340 | Defined *s = addDefined(name, /*file=*/nullptr, isec, value, /*size=*/0, |
341 | /*isWeakDef=*/false, isPrivateExtern, |
342 | isReferencedDynamically: referencedDynamically, /*noDeadStrip=*/false, |
343 | /*isWeakDefCanBeHidden=*/false); |
344 | s->includeInSymtab = includeInSymtab; |
345 | return s; |
346 | } |
347 | |
348 | enum class Boundary { |
349 | Start, |
350 | End, |
351 | }; |
352 | |
353 | static Defined *createBoundarySymbol(const Undefined &sym) { |
354 | return symtab->addSynthetic( |
355 | name: sym.getName(), /*isec=*/nullptr, /*value=*/-1, /*isPrivateExtern=*/true, |
356 | /*includeInSymtab=*/false, /*referencedDynamically=*/false); |
357 | } |
358 | |
359 | static void handleSectionBoundarySymbol(const Undefined &sym, StringRef segSect, |
360 | Boundary which) { |
361 | auto [segName, sectName] = segSect.split(Separator: '$'); |
362 | |
363 | // Attach the symbol to any InputSection that will end up in the right |
364 | // OutputSection -- it doesn't matter which one we pick. |
365 | // Don't bother looking through inputSections for a matching |
366 | // ConcatInputSection -- we need to create ConcatInputSection for |
367 | // non-existing sections anyways, and that codepath works even if we should |
368 | // already have a ConcatInputSection with the right name. |
369 | |
370 | OutputSection *osec = nullptr; |
371 | // This looks for __TEXT,__cstring etc. |
372 | for (SyntheticSection *ssec : syntheticSections) |
373 | if (ssec->segname == segName && ssec->name == sectName) { |
374 | osec = ssec->isec->parent; |
375 | break; |
376 | } |
377 | |
378 | if (!osec) { |
379 | ConcatInputSection *isec = makeSyntheticInputSection(segName, sectName); |
380 | |
381 | // This runs after markLive() and is only called for Undefineds that are |
382 | // live. Marking the isec live ensures an OutputSection is created that the |
383 | // start/end symbol can refer to. |
384 | assert(sym.isLive()); |
385 | assert(isec->live); |
386 | |
387 | // This runs after gatherInputSections(), so need to explicitly set parent |
388 | // and add to inputSections. |
389 | osec = isec->parent = ConcatOutputSection::getOrCreateForInput(isec); |
390 | inputSections.push_back(x: isec); |
391 | } |
392 | |
393 | if (which == Boundary::Start) |
394 | osec->sectionStartSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
395 | else |
396 | osec->sectionEndSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
397 | } |
398 | |
399 | static void handleSegmentBoundarySymbol(const Undefined &sym, StringRef segName, |
400 | Boundary which) { |
401 | OutputSegment *seg = getOrCreateOutputSegment(name: segName); |
402 | if (which == Boundary::Start) |
403 | seg->segmentStartSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
404 | else |
405 | seg->segmentEndSymbols.push_back(NewVal: createBoundarySymbol(sym)); |
406 | } |
407 | |
408 | // Try to find a definition for an undefined symbol. |
409 | // Returns true if a definition was found and no diagnostics are needed. |
410 | static bool recoverFromUndefinedSymbol(const Undefined &sym) { |
411 | // Handle start/end symbols. |
412 | StringRef name = sym.getName(); |
413 | if (name.consume_front(Prefix: "section$start$")) { |
414 | handleSectionBoundarySymbol(sym, segSect: name, which: Boundary::Start); |
415 | return true; |
416 | } |
417 | if (name.consume_front(Prefix: "section$end$")) { |
418 | handleSectionBoundarySymbol(sym, segSect: name, which: Boundary::End); |
419 | return true; |
420 | } |
421 | if (name.consume_front(Prefix: "segment$start$")) { |
422 | handleSegmentBoundarySymbol(sym, segName: name, which: Boundary::Start); |
423 | return true; |
424 | } |
425 | if (name.consume_front(Prefix: "segment$end$")) { |
426 | handleSegmentBoundarySymbol(sym, segName: name, which: Boundary::End); |
427 | return true; |
428 | } |
429 | |
430 | // Leave dtrace symbols, since we will handle them when we do the relocation |
431 | if (name.starts_with(Prefix: "___dtrace_")) |
432 | return true; |
433 | |
434 | // Handle -U. |
435 | if (config->explicitDynamicLookups.count(Key: sym.getName())) { |
436 | symtab->addDynamicLookup(name: sym.getName()); |
437 | return true; |
438 | } |
439 | |
440 | // Handle -undefined. |
441 | if (config->undefinedSymbolTreatment == |
442 | UndefinedSymbolTreatment::dynamic_lookup || |
443 | config->undefinedSymbolTreatment == UndefinedSymbolTreatment::suppress) { |
444 | symtab->addDynamicLookup(name: sym.getName()); |
445 | return true; |
446 | } |
447 | |
448 | // We do not return true here, as we still need to print diagnostics. |
449 | if (config->undefinedSymbolTreatment == UndefinedSymbolTreatment::warning) |
450 | symtab->addDynamicLookup(name: sym.getName()); |
451 | |
452 | return false; |
453 | } |
454 | |
455 | namespace { |
456 | struct UndefinedDiag { |
457 | struct SectionAndOffset { |
458 | const InputSection *isec; |
459 | uint64_t offset; |
460 | }; |
461 | |
462 | std::vector<SectionAndOffset> codeReferences; |
463 | std::vector<std::string> otherReferences; |
464 | }; |
465 | |
466 | MapVector<const Undefined *, UndefinedDiag> undefs; |
467 | } // namespace |
468 | |
469 | void macho::reportPendingDuplicateSymbols() { |
470 | for (const auto &duplicate : dupSymDiags) { |
471 | if (!config->deadStripDuplicates || duplicate.sym->isLive()) { |
472 | std::string message = |
473 | "duplicate symbol: "+ toString(*duplicate.sym) + "\n>>> defined in "; |
474 | if (!duplicate.src1.first.empty()) |
475 | message += duplicate.src1.first + "\n>>> "; |
476 | message += duplicate.src1.second + "\n>>> defined in "; |
477 | if (!duplicate.src2.first.empty()) |
478 | message += duplicate.src2.first + "\n>>> "; |
479 | error(msg: message + duplicate.src2.second); |
480 | } |
481 | } |
482 | } |
483 | |
484 | // Check whether the definition name def is a mangled function name that matches |
485 | // the reference name ref. |
486 | static bool canSuggestExternCForCXX(StringRef ref, StringRef def) { |
487 | llvm::ItaniumPartialDemangler d; |
488 | std::string name = def.str(); |
489 | if (d.partialDemangle(MangledName: name.c_str())) |
490 | return false; |
491 | char *buf = d.getFunctionName(Buf: nullptr, N: nullptr); |
492 | if (!buf) |
493 | return false; |
494 | bool ret = ref == buf; |
495 | free(ptr: buf); |
496 | return ret; |
497 | } |
498 | |
499 | // Suggest an alternative spelling of an "undefined symbol" diagnostic. Returns |
500 | // the suggested symbol, which is either in the symbol table, or in the same |
501 | // file of sym. |
502 | static const Symbol *getAlternativeSpelling(const Undefined &sym, |
503 | std::string &preHint, |
504 | std::string &postHint) { |
505 | DenseMap<StringRef, const Symbol *> map; |
506 | if (sym.getFile() && sym.getFile()->kind() == InputFile::ObjKind) { |
507 | // Build a map of local defined symbols. |
508 | for (const Symbol *s : sym.getFile()->symbols) |
509 | if (auto *defined = dyn_cast_or_null<Defined>(Val: s)) |
510 | if (!defined->isExternal()) |
511 | map.try_emplace(Key: s->getName(), Args&: s); |
512 | } |
513 | |
514 | auto suggest = [&](StringRef newName) -> const Symbol * { |
515 | // If defined locally. |
516 | if (const Symbol *s = map.lookup(Val: newName)) |
517 | return s; |
518 | |
519 | // If in the symbol table and not undefined. |
520 | if (const Symbol *s = symtab->find(name: newName)) |
521 | if (!isa<Undefined>(Val: s)) |
522 | return s; |
523 | |
524 | return nullptr; |
525 | }; |
526 | |
527 | // This loop enumerates all strings of Levenshtein distance 1 as typo |
528 | // correction candidates and suggests the one that exists as a non-undefined |
529 | // symbol. |
530 | StringRef name = sym.getName(); |
531 | for (size_t i = 0, e = name.size(); i != e + 1; ++i) { |
532 | // Insert a character before name[i]. |
533 | std::string newName = (name.substr(Start: 0, N: i) + "0"+ name.substr(Start: i)).str(); |
534 | for (char c = '0'; c <= 'z'; ++c) { |
535 | newName[i] = c; |
536 | if (const Symbol *s = suggest(newName)) |
537 | return s; |
538 | } |
539 | if (i == e) |
540 | break; |
541 | |
542 | // Substitute name[i]. |
543 | newName = std::string(name); |
544 | for (char c = '0'; c <= 'z'; ++c) { |
545 | newName[i] = c; |
546 | if (const Symbol *s = suggest(newName)) |
547 | return s; |
548 | } |
549 | |
550 | // Transpose name[i] and name[i+1]. This is of edit distance 2 but it is |
551 | // common. |
552 | if (i + 1 < e) { |
553 | newName[i] = name[i + 1]; |
554 | newName[i + 1] = name[i]; |
555 | if (const Symbol *s = suggest(newName)) |
556 | return s; |
557 | } |
558 | |
559 | // Delete name[i]. |
560 | newName = (name.substr(Start: 0, N: i) + name.substr(Start: i + 1)).str(); |
561 | if (const Symbol *s = suggest(newName)) |
562 | return s; |
563 | } |
564 | |
565 | // Case mismatch, e.g. Foo vs FOO. |
566 | for (auto &it : map) |
567 | if (name.equals_insensitive(RHS: it.first)) |
568 | return it.second; |
569 | for (Symbol *sym : symtab->getSymbols()) |
570 | if (!isa<Undefined>(Val: sym) && name.equals_insensitive(RHS: sym->getName())) |
571 | return sym; |
572 | |
573 | // The reference may be a mangled name while the definition is not. Suggest a |
574 | // missing extern "C". |
575 | if (name.starts_with(Prefix: "__Z")) { |
576 | std::string buf = name.str(); |
577 | llvm::ItaniumPartialDemangler d; |
578 | if (!d.partialDemangle(MangledName: buf.c_str())) |
579 | if (char *buf = d.getFunctionName(Buf: nullptr, N: nullptr)) { |
580 | const Symbol *s = suggest((Twine("_") + buf).str()); |
581 | free(ptr: buf); |
582 | if (s) { |
583 | preHint = ": extern \"C\" "; |
584 | return s; |
585 | } |
586 | } |
587 | } else { |
588 | StringRef nameWithoutUnderscore = name; |
589 | nameWithoutUnderscore.consume_front(Prefix: "_"); |
590 | const Symbol *s = nullptr; |
591 | for (auto &it : map) |
592 | if (canSuggestExternCForCXX(ref: nameWithoutUnderscore, def: it.first)) { |
593 | s = it.second; |
594 | break; |
595 | } |
596 | if (!s) |
597 | for (Symbol *sym : symtab->getSymbols()) |
598 | if (canSuggestExternCForCXX(ref: nameWithoutUnderscore, def: sym->getName())) { |
599 | s = sym; |
600 | break; |
601 | } |
602 | if (s) { |
603 | preHint = " to declare "; |
604 | postHint = " as extern \"C\"?"; |
605 | return s; |
606 | } |
607 | } |
608 | |
609 | return nullptr; |
610 | } |
611 | |
612 | static void reportUndefinedSymbol(const Undefined &sym, |
613 | const UndefinedDiag &locations, |
614 | bool correctSpelling) { |
615 | std::string message = "undefined symbol"; |
616 | if (config->archMultiple) |
617 | message += (" for arch "+ getArchitectureName(Arch: config->arch())).str(); |
618 | message += ": "+ toString(sym); |
619 | |
620 | const size_t maxUndefinedReferences = 3; |
621 | size_t i = 0; |
622 | for (const std::string &loc : locations.otherReferences) { |
623 | if (i >= maxUndefinedReferences) |
624 | break; |
625 | message += "\n>>> referenced by "+ loc; |
626 | ++i; |
627 | } |
628 | |
629 | for (const UndefinedDiag::SectionAndOffset &loc : locations.codeReferences) { |
630 | if (i >= maxUndefinedReferences) |
631 | break; |
632 | message += "\n>>> referenced by "; |
633 | std::string src = loc.isec->getSourceLocation(off: loc.offset); |
634 | if (!src.empty()) |
635 | message += src + "\n>>> "; |
636 | message += loc.isec->getLocation(off: loc.offset); |
637 | ++i; |
638 | } |
639 | |
640 | size_t totalReferences = |
641 | locations.otherReferences.size() + locations.codeReferences.size(); |
642 | if (totalReferences > i) |
643 | message += |
644 | ("\n>>> referenced "+ Twine(totalReferences - i) + " more times") |
645 | .str(); |
646 | |
647 | if (correctSpelling) { |
648 | std::string preHint = ": ", postHint; |
649 | if (const Symbol *corrected = |
650 | getAlternativeSpelling(sym, preHint, postHint)) { |
651 | message += |
652 | "\n>>> did you mean"+ preHint + toString(*corrected) + postHint; |
653 | if (corrected->getFile()) |
654 | message += "\n>>> defined in: "+ toString(file: corrected->getFile()); |
655 | } |
656 | } |
657 | |
658 | if (config->undefinedSymbolTreatment == UndefinedSymbolTreatment::error) |
659 | error(msg: message); |
660 | else if (config->undefinedSymbolTreatment == |
661 | UndefinedSymbolTreatment::warning) |
662 | warn(msg: message); |
663 | else |
664 | assert(false && "diagnostics make sense for -undefined error|warning only"); |
665 | } |
666 | |
667 | void macho::reportPendingUndefinedSymbols() { |
668 | // Enable spell corrector for the first 2 diagnostics. |
669 | for (const auto &[i, undef] : llvm::enumerate(First&: undefs)) |
670 | reportUndefinedSymbol(sym: *undef.first, locations: undef.second, correctSpelling: i < 2); |
671 | |
672 | // This function is called multiple times during execution. Clear the printed |
673 | // diagnostics to avoid printing the same things again the next time. |
674 | undefs.clear(); |
675 | } |
676 | |
677 | void macho::treatUndefinedSymbol(const Undefined &sym, StringRef source) { |
678 | if (recoverFromUndefinedSymbol(sym)) |
679 | return; |
680 | |
681 | undefs[&sym].otherReferences.push_back(x: source.str()); |
682 | } |
683 | |
684 | void macho::treatUndefinedSymbol(const Undefined &sym, const InputSection *isec, |
685 | uint64_t offset) { |
686 | if (recoverFromUndefinedSymbol(sym)) |
687 | return; |
688 | |
689 | undefs[&sym].codeReferences.push_back(x: {.isec: isec, .offset: offset}); |
690 | } |
691 | |
692 | std::unique_ptr<SymbolTable> macho::symtab; |
693 |
Definitions
- find
- insert
- DuplicateSymbolDiag
- DuplicateSymbolDiag
- dupSymDiags
- transplantSymbolsAtOffset
- addDefined
- aliasDefined
- addUndefined
- addCommon
- addDylib
- addDynamicLookup
- addLazyArchive
- addLazyObject
- addSynthetic
- Boundary
- createBoundarySymbol
- handleSectionBoundarySymbol
- handleSegmentBoundarySymbol
- recoverFromUndefinedSymbol
- UndefinedDiag
- SectionAndOffset
- undefs
- reportPendingDuplicateSymbols
- canSuggestExternCForCXX
- getAlternativeSpelling
- reportUndefinedSymbol
- reportPendingUndefinedSymbols
- treatUndefinedSymbol
- treatUndefinedSymbol
Improve your Profiling and Debugging skills
Find out more