401 lines · cpp
1//===- InputSection.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-exception6//7//===----------------------------------------------------------------------===//8 9#include "InputSection.h"10#include "ConcatOutputSection.h"11#include "Config.h"12#include "InputFiles.h"13#include "OutputSegment.h"14#include "Sections.h"15#include "Symbols.h"16#include "SyntheticSections.h"17#include "Target.h"18#include "Writer.h"19 20#include "lld/Common/ErrorHandler.h"21#include "lld/Common/Memory.h"22#include "llvm/Support/xxhash.h"23 24using namespace llvm;25using namespace llvm::MachO;26using namespace llvm::support;27using namespace lld;28using namespace lld::macho;29 30// Verify ConcatInputSection's size on 64-bit builds. The size of std::vector31// can differ based on STL debug levels (e.g. iterator debugging on MSVC's STL),32// so account for that.33static_assert(sizeof(void *) != 8 ||34 sizeof(ConcatInputSection) == sizeof(std::vector<Reloc>) + 88,35 "Try to minimize ConcatInputSection's size, we create many "36 "instances of it");37 38std::vector<ConcatInputSection *> macho::inputSections;39int macho::inputSectionsOrder = 0;40 41// Call this function to add a new InputSection and have it routed to the42// appropriate container. Depending on its type and current config, it will43// either be added to 'inputSections' vector or to a synthetic section.44void lld::macho::addInputSection(InputSection *inputSection) {45 if (auto *isec = dyn_cast<ConcatInputSection>(inputSection)) {46 if (isec->isCoalescedWeak())47 return;48 if (config->emitRelativeMethodLists &&49 ObjCMethListSection::isMethodList(isec)) {50 if (in.objcMethList->inputOrder == UnspecifiedInputOrder)51 in.objcMethList->inputOrder = inputSectionsOrder++;52 in.objcMethList->addInput(isec);53 isec->parent = in.objcMethList;54 return;55 }56 if (config->emitInitOffsets &&57 sectionType(isec->getFlags()) == S_MOD_INIT_FUNC_POINTERS) {58 in.initOffsets->addInput(isec);59 return;60 }61 isec->outSecOff = inputSectionsOrder++;62 auto *osec = ConcatOutputSection::getOrCreateForInput(isec);63 isec->parent = osec;64 inputSections.push_back(isec);65 } else if (auto *isec = dyn_cast<CStringInputSection>(inputSection)) {66 bool useSectionName = config->separateCstringLiteralSections ||67 isec->getName() == section_names::objcMethname;68 auto *osec = in.getOrCreateCStringSection(69 useSectionName ? isec->getName() : section_names::cString);70 if (osec->inputOrder == UnspecifiedInputOrder)71 osec->inputOrder = inputSectionsOrder++;72 osec->addInput(isec);73 } else if (auto *isec = dyn_cast<WordLiteralInputSection>(inputSection)) {74 if (in.wordLiteralSection->inputOrder == UnspecifiedInputOrder)75 in.wordLiteralSection->inputOrder = inputSectionsOrder++;76 in.wordLiteralSection->addInput(isec);77 } else {78 llvm_unreachable("unexpected input section kind");79 }80 81 assert(inputSectionsOrder <= UnspecifiedInputOrder);82}83 84uint64_t InputSection::getFileSize() const {85 return isZeroFill(getFlags()) ? 0 : getSize();86}87 88uint64_t InputSection::getVA(uint64_t off) const {89 return parent->addr + getOffset(off);90}91 92static uint64_t resolveSymbolVA(const Symbol *sym, uint8_t type) {93 const RelocAttrs &relocAttrs = target->getRelocAttrs(type);94 if (relocAttrs.hasAttr(RelocAttrBits::BRANCH))95 return sym->resolveBranchVA();96 if (relocAttrs.hasAttr(RelocAttrBits::GOT))97 return sym->resolveGotVA();98 if (relocAttrs.hasAttr(RelocAttrBits::TLV))99 return sym->resolveTlvVA();100 return sym->getVA();101}102 103const Defined *InputSection::getContainingSymbol(uint64_t off) const {104 auto *nextSym = llvm::upper_bound(105 symbols, off, [](uint64_t a, const Defined *b) { return a < b->value; });106 if (nextSym == symbols.begin())107 return nullptr;108 return *std::prev(nextSym);109}110 111std::string InputSection::getLocation(uint64_t off) const {112 // First, try to find a symbol that's near the offset. Use it as a reference113 // point.114 if (auto *sym = getContainingSymbol(off))115 return (toString(getFile()) + ":(symbol " + toString(*sym) + "+0x" +116 Twine::utohexstr(off - sym->value) + ")")117 .str();118 119 // If that fails, use the section itself as a reference point.120 for (const Subsection &subsec : section.subsections) {121 if (subsec.isec == this) {122 off += subsec.offset;123 break;124 }125 }126 127 return (toString(getFile()) + ":(" + getName() + "+0x" +128 Twine::utohexstr(off) + ")")129 .str();130}131 132std::string InputSection::getSourceLocation(uint64_t off) const {133 auto *obj = dyn_cast_or_null<ObjFile>(getFile());134 if (!obj)135 return {};136 137 DWARFCache *dwarf = obj->getDwarf();138 if (!dwarf)139 return std::string();140 141 for (const Subsection &subsec : section.subsections) {142 if (subsec.isec == this) {143 off += subsec.offset;144 break;145 }146 }147 148 auto createMsg = [&](StringRef path, unsigned line) {149 std::string filename = sys::path::filename(path).str();150 std::string lineStr = (":" + Twine(line)).str();151 if (filename == path)152 return filename + lineStr;153 return (filename + lineStr + " (" + path + lineStr + ")").str();154 };155 156 // First, look up a function for a given offset.157 if (std::optional<DILineInfo> li = dwarf->getDILineInfo(158 section.addr + off, object::SectionedAddress::UndefSection))159 return createMsg(li->FileName, li->Line);160 161 // If it failed, look up again as a variable.162 if (const Defined *sym = getContainingSymbol(off)) {163 // Symbols are generally prefixed with an underscore, which is not included164 // in the debug information.165 StringRef symName = sym->getName();166 symName.consume_front("_");167 168 if (std::optional<std::pair<std::string, unsigned>> fileLine =169 dwarf->getVariableLoc(symName))170 return createMsg(fileLine->first, fileLine->second);171 }172 173 // Try to get the source file's name from the DWARF information.174 if (obj->compileUnit)175 return obj->sourceFile();176 177 return {};178}179 180const Reloc *InputSection::getRelocAt(uint32_t off) const {181 auto it = llvm::find_if(182 relocs, [=](const macho::Reloc &r) { return r.offset == off; });183 if (it == relocs.end())184 return nullptr;185 return &*it;186}187 188void ConcatInputSection::foldIdentical(ConcatInputSection *copy,189 Symbol::ICFFoldKind foldKind) {190 align = std::max(align, copy->align);191 copy->live = false;192 copy->wasCoalesced = true;193 copy->replacement = this;194 for (auto ©Sym : copy->symbols)195 copySym->identicalCodeFoldingKind = foldKind;196 197 symbols.insert(symbols.end(), copy->symbols.begin(), copy->symbols.end());198 copy->symbols.clear();199 200 // Remove duplicate compact unwind info for symbols at the same address.201 if (symbols.empty())202 return;203 for (auto it = symbols.begin() + 1; it != symbols.end(); ++it) {204 assert((*it)->value == 0);205 (*it)->originalUnwindEntry = nullptr;206 }207}208 209void ConcatInputSection::writeTo(uint8_t *buf) {210 assert(!shouldOmitFromOutput());211 212 if (getFileSize() == 0)213 return;214 215 memcpy(buf, data.data(), data.size());216 217 for (size_t i = 0; i < relocs.size(); i++) {218 const Reloc &r = relocs[i];219 uint8_t *loc = buf + r.offset;220 uint64_t referentVA = 0;221 222 const bool needsFixup = config->emitChainedFixups &&223 target->hasAttr(r.type, RelocAttrBits::UNSIGNED);224 if (target->hasAttr(r.type, RelocAttrBits::SUBTRAHEND)) {225 const Symbol *fromSym = cast<Symbol *>(r.referent);226 const Reloc &minuend = relocs[++i];227 uint64_t minuendVA;228 if (const Symbol *toSym = minuend.referent.dyn_cast<Symbol *>())229 minuendVA = toSym->getVA() + minuend.addend;230 else {231 auto *referentIsec = cast<InputSection *>(minuend.referent);232 assert(!::shouldOmitFromOutput(referentIsec));233 minuendVA = referentIsec->getVA(minuend.addend);234 }235 referentVA = minuendVA - fromSym->getVA();236 } else if (auto *referentSym = r.referent.dyn_cast<Symbol *>()) {237 if (target->hasAttr(r.type, RelocAttrBits::LOAD) &&238 !referentSym->isInGot())239 target->relaxGotLoad(loc, r.type);240 // For dtrace symbols, do not handle them as normal undefined symbols241 if (referentSym->getName().starts_with("___dtrace_")) {242 // Change dtrace call site to pre-defined instructions243 target->handleDtraceReloc(referentSym, r, loc);244 continue;245 }246 referentVA = resolveSymbolVA(referentSym, r.type) + r.addend;247 248 if (isThreadLocalVariables(getFlags()) && isa<Defined>(referentSym)) {249 // References from thread-local variable sections are treated as offsets250 // relative to the start of the thread-local data memory area, which251 // is initialized via copying all the TLV data sections (which are all252 // contiguous).253 referentVA -= firstTLVDataSection->addr;254 } else if (needsFixup) {255 writeChainedFixup(loc, referentSym, r.addend);256 continue;257 }258 } else if (auto *referentIsec = r.referent.dyn_cast<InputSection *>()) {259 assert(!::shouldOmitFromOutput(referentIsec));260 referentVA = referentIsec->getVA(r.addend);261 262 if (needsFixup) {263 writeChainedRebase(loc, referentVA);264 continue;265 }266 }267 target->relocateOne(loc, r, referentVA, getVA() + r.offset);268 }269}270 271ConcatInputSection *macho::makeSyntheticInputSection(StringRef segName,272 StringRef sectName,273 uint32_t flags,274 ArrayRef<uint8_t> data,275 uint32_t align) {276 Section §ion =277 *make<Section>(/*file=*/nullptr, segName, sectName, flags, /*addr=*/0);278 auto isec = make<ConcatInputSection>(section, data, align);279 // Since this is an explicitly created 'fake' input section,280 // it should not be dead stripped.281 isec->live = true;282 section.subsections.push_back({0, isec});283 return isec;284}285 286void CStringInputSection::splitIntoPieces() {287 size_t off = 0;288 StringRef s = toStringRef(data);289 while (!s.empty()) {290 size_t end = s.find(0);291 if (end == StringRef::npos)292 fatal(getLocation(off) + ": string is not null terminated");293 uint32_t hash = deduplicateLiterals ? xxh3_64bits(s.take_front(end)) : 0;294 pieces.emplace_back(off, hash);295 size_t size = end + 1; // include null terminator296 s = s.substr(size);297 off += size;298 }299}300 301StringPiece &CStringInputSection::getStringPiece(uint64_t off) {302 if (off >= data.size())303 fatal(toString(this) + ": offset is outside the section");304 305 auto it =306 partition_point(pieces, [=](StringPiece p) { return p.inSecOff <= off; });307 return it[-1];308}309 310const StringPiece &CStringInputSection::getStringPiece(uint64_t off) const {311 return const_cast<CStringInputSection *>(this)->getStringPiece(off);312}313 314size_t CStringInputSection::getStringPieceIndex(uint64_t off) const {315 if (off >= data.size())316 fatal(toString(this) + ": offset is outside the section");317 318 auto it =319 partition_point(pieces, [=](StringPiece p) { return p.inSecOff <= off; });320 return std::distance(pieces.begin(), it) - 1;321}322 323uint64_t CStringInputSection::getOffset(uint64_t off) const {324 const StringPiece &piece = getStringPiece(off);325 uint64_t addend = off - piece.inSecOff;326 return piece.outSecOff + addend;327}328 329WordLiteralInputSection::WordLiteralInputSection(const Section §ion,330 ArrayRef<uint8_t> data,331 uint32_t align)332 : InputSection(WordLiteralKind, section, data, align) {333 switch (sectionType(getFlags())) {334 case S_4BYTE_LITERALS:335 power2LiteralSize = 2;336 break;337 case S_8BYTE_LITERALS:338 power2LiteralSize = 3;339 break;340 case S_16BYTE_LITERALS:341 power2LiteralSize = 4;342 break;343 default:344 llvm_unreachable("invalid literal section type");345 }346 347 live.resize(data.size() >> power2LiteralSize, !config->deadStrip);348}349 350uint64_t WordLiteralInputSection::getOffset(uint64_t off) const {351 if (off >= data.size())352 fatal(toString(this) + ": offset is outside the section");353 354 auto *osec = cast<WordLiteralSection>(parent);355 const uintptr_t buf = reinterpret_cast<uintptr_t>(data.data());356 switch (sectionType(getFlags())) {357 case S_4BYTE_LITERALS:358 return osec->getLiteral4Offset(buf + (off & ~3LLU)) | (off & 3);359 case S_8BYTE_LITERALS:360 return osec->getLiteral8Offset(buf + (off & ~7LLU)) | (off & 7);361 case S_16BYTE_LITERALS:362 return osec->getLiteral16Offset(buf + (off & ~15LLU)) | (off & 15);363 default:364 llvm_unreachable("invalid literal section type");365 }366}367 368bool macho::isCodeSection(const InputSection *isec) {369 return sections::isCodeSection(isec->getName(), isec->getSegName(),370 isec->getFlags());371}372 373bool macho::isCfStringSection(const InputSection *isec) {374 return isec->getName() == section_names::cfString &&375 isec->getSegName() == segment_names::data;376}377 378bool macho::isClassRefsSection(const InputSection *isec) {379 return isec->getName() == section_names::objcClassRefs &&380 isec->getSegName() == segment_names::data;381}382 383bool macho::isSelRefsSection(const InputSection *isec) {384 return isec->getName() == section_names::objcSelrefs &&385 isec->getSegName() == segment_names::data;386}387 388bool macho::isEhFrameSection(const InputSection *isec) {389 return isec->getName() == section_names::ehFrame &&390 isec->getSegName() == segment_names::text;391}392 393bool macho::isGccExceptTabSection(const InputSection *isec) {394 return isec->getName() == section_names::gccExceptTab &&395 isec->getSegName() == segment_names::text;396}397 398std::string lld::toString(const InputSection *isec) {399 return (toString(isec->getFile()) + ":(" + isec->getName() + ")").str();400}401