lld/MachO/InputSection.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-exception
   6 //
   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 "UnwindInfoSection.h"
  19 #include "Writer.h"
  20
  21 #include "lld/Common/ErrorHandler.h"
  22 #include "lld/Common/Memory.h"
  23 #include "llvm/Support/Endian.h"
  24 #include "llvm/Support/xxhash.h"
  25
  26 using namespace llvm;
  27 using namespace llvm::MachO;
  28 using namespace llvm::support;
  29 using namespace lld;
  30 using namespace lld::macho;
  31
  32 // Verify ConcatInputSection's size on 64-bit builds. The size of std::vector
  33 // can differ based on STL debug levels (e.g. iterator debugging on MSVC's STL),
  34 // so account for that.
  35 static_assert(sizeof(void *) != 8 ||
  36                   sizeof(ConcatInputSection) == sizeof(std::vector<Reloc>) + 88,
  37               "Try to minimize ConcatInputSection's size, we create many "
  38               "instances of it");
  39
  40 std::vector<ConcatInputSection *> macho::inputSections;
  41 int macho::inputSectionsOrder = 0;
  42
  43 // Call this function to add a new InputSection and have it routed to the
  44 // appropriate container. Depending on its type and current config, it will
  45 // either be added to 'inputSections' vector or to a synthetic section.
  46 void lld::macho::addInputSection(InputSection *inputSection) {
  47   if (auto *isec = dyn_cast<ConcatInputSection>(inputSection)) {
  48     if (isec->isCoalescedWeak())
  49       return;
  50     if (config->emitRelativeMethodLists &&
  51         ObjCMethListSection::isMethodList(isec)) {
  52       if (in.objcMethList->inputOrder == UnspecifiedInputOrder)
  53         in.objcMethList->inputOrder = inputSectionsOrder++;
  54       in.objcMethList->addInput(isec);
  55       isec->parent = in.objcMethList;
  56       return;
  57     }
  58     if (config->emitInitOffsets &&
  59         sectionType(isec->getFlags()) == S_MOD_INIT_FUNC_POINTERS) {
  60       in.initOffsets->addInput(isec);
  61       return;
  62     }
  63     isec->outSecOff = inputSectionsOrder++;
  64     auto *osec = ConcatOutputSection::getOrCreateForInput(isec);
  65     isec->parent = osec;
  66     inputSections.push_back(isec);
  67   } else if (auto *isec = dyn_cast<CStringInputSection>(inputSection)) {
  68     if (isec->getName() == section_names::objcMethname) {
  69       if (in.objcMethnameSection->inputOrder == UnspecifiedInputOrder)
  70         in.objcMethnameSection->inputOrder = inputSectionsOrder++;
  71       in.objcMethnameSection->addInput(isec);
  72     } else {
  73       if (in.cStringSection->inputOrder == UnspecifiedInputOrder)
  74         in.cStringSection->inputOrder = inputSectionsOrder++;
  75       in.cStringSection->addInput(isec);
  76     }
  77   } else if (auto *isec = dyn_cast<WordLiteralInputSection>(inputSection)) {
  78     if (in.wordLiteralSection->inputOrder == UnspecifiedInputOrder)
  79       in.wordLiteralSection->inputOrder = inputSectionsOrder++;
  80     in.wordLiteralSection->addInput(isec);
  81   } else {
  82     llvm_unreachable("unexpected input section kind");
  83   }
  84
  85   assert(inputSectionsOrder <= UnspecifiedInputOrder);
  86 }
  87
  88 uint64_t InputSection::getFileSize() const {
  89   return isZeroFill(getFlags()) ? 0 : getSize();
  90 }
  91
  92 uint64_t InputSection::getVA(uint64_t off) const {
  93   return parent->addr + getOffset(off);
  94 }
  95
  96 static uint64_t resolveSymbolVA(const Symbol *sym, uint8_t type) {
  97   const RelocAttrs &relocAttrs = target->getRelocAttrs(type);
  98   if (relocAttrs.hasAttr(RelocAttrBits::BRANCH))
  99     return sym->resolveBranchVA();
 100   if (relocAttrs.hasAttr(RelocAttrBits::GOT))
 101     return sym->resolveGotVA();
 102   if (relocAttrs.hasAttr(RelocAttrBits::TLV))
 103     return sym->resolveTlvVA();
 104   return sym->getVA();
 105 }
 106
 107 const Defined *InputSection::getContainingSymbol(uint64_t off) const {
 108   auto *nextSym = llvm::upper_bound(
 109       symbols, off, [](uint64_t a, const Defined *b) { return a < b->value; });
 110   if (nextSym == symbols.begin())
 111     return nullptr;
 112   return *std::prev(nextSym);
 113 }
 114
 115 std::string InputSection::getLocation(uint64_t off) const {
 116   // First, try to find a symbol that's near the offset. Use it as a reference
 117   // point.
 118   if (auto *sym = getContainingSymbol(off))
 119     return (toString(getFile()) + ":(symbol " + toString(*sym) + "+0x" +
 120             Twine::utohexstr(off - sym->value) + ")")
 121         .str();
 122
 123   // If that fails, use the section itself as a reference point.
 124   for (const Subsection &subsec : section.subsections) {
 125     if (subsec.isec == this) {
 126       off += subsec.offset;
 127       break;
 128     }
 129   }
 130
 131   return (toString(getFile()) + ":(" + getName() + "+0x" +
 132           Twine::utohexstr(off) + ")")
 133       .str();
 134 }
 135
 136 std::string InputSection::getSourceLocation(uint64_t off) const {
 137   auto *obj = dyn_cast_or_null<ObjFile>(getFile());
 138   if (!obj)
 139     return {};
 140
 141   DWARFCache *dwarf = obj->getDwarf();
 142   if (!dwarf)
 143     return std::string();
 144
 145   for (const Subsection &subsec : section.subsections) {
 146     if (subsec.isec == this) {
 147       off += subsec.offset;
 148       break;
 149     }
 150   }
 151
 152   auto createMsg = [&](StringRef path, unsigned line) {
 153     std::string filename = sys::path::filename(path).str();
 154     std::string lineStr = (":" + Twine(line)).str();
 155     if (filename == path)
 156       return filename + lineStr;
 157     return (filename + lineStr + " (" + path + lineStr + ")").str();
 158   };
 159
 160   // First, look up a function for a given offset.
 161   if (std::optional<DILineInfo> li = dwarf->getDILineInfo(
 162           section.addr + off, object::SectionedAddress::UndefSection))
 163     return createMsg(li->FileName, li->Line);
 164
 165   // If it failed, look up again as a variable.
 166   if (const Defined *sym = getContainingSymbol(off)) {
 167     // Symbols are generally prefixed with an underscore, which is not included
 168     // in the debug information.
 169     StringRef symName = sym->getName();
 170     symName.consume_front("_");
 171
 172     if (std::optional<std::pair<std::string, unsigned>> fileLine =
 173             dwarf->getVariableLoc(symName))
 174       return createMsg(fileLine->first, fileLine->second);
 175   }
 176
 177   // Try to get the source file's name from the DWARF information.
 178   if (obj->compileUnit)
 179     return obj->sourceFile();
 180
 181   return {};
 182 }
 183
 184 const Reloc *InputSection::getRelocAt(uint32_t off) const {
 185   auto it = llvm::find_if(
 186       relocs, [=](const macho::Reloc &r) { return r.offset == off; });
 187   if (it == relocs.end())
 188     return nullptr;
 189   return &*it;
 190 }
 191
 192 void ConcatInputSection::foldIdentical(ConcatInputSection *copy,
 193                                        Symbol::ICFFoldKind foldKind) {
 194   align = std::max(align, copy->align);
 195   copy->live = false;
 196   copy->wasCoalesced = true;
 197   copy->replacement = this;
 198   for (auto &copySym : copy->symbols)
 199     copySym->identicalCodeFoldingKind = foldKind;
 200
 201   symbols.insert(symbols.end(), copy->symbols.begin(), copy->symbols.end());
 202   copy->symbols.clear();
 203
 204   // Remove duplicate compact unwind info for symbols at the same address.
 205   if (symbols.empty())
 206     return;
 207   for (auto it = symbols.begin() + 1; it != symbols.end(); ++it) {
 208     assert((*it)->value == 0);
 209     (*it)->originalUnwindEntry = nullptr;
 210   }
 211 }
 212
 213 void ConcatInputSection::writeTo(uint8_t *buf) {
 214   assert(!shouldOmitFromOutput());
 215
 216   if (getFileSize() == 0)
 217     return;
 218
 219   memcpy(buf, data.data(), data.size());
 220
 221   for (size_t i = 0; i < relocs.size(); i++) {
 222     const Reloc &r = relocs[i];
 223     uint8_t *loc = buf + r.offset;
 224     uint64_t referentVA = 0;
 225
 226     const bool needsFixup = config->emitChainedFixups &&
 227                             target->hasAttr(r.type, RelocAttrBits::UNSIGNED);
 228     if (target->hasAttr(r.type, RelocAttrBits::SUBTRAHEND)) {
 229       const Symbol *fromSym = r.referent.get<Symbol *>();
 230       const Reloc &minuend = relocs[++i];
 231       uint64_t minuendVA;
 232       if (const Symbol *toSym = minuend.referent.dyn_cast<Symbol *>())
 233         minuendVA = toSym->getVA() + minuend.addend;
 234       else {
 235         auto *referentIsec = minuend.referent.get<InputSection *>();
 236         assert(!::shouldOmitFromOutput(referentIsec));
 237         minuendVA = referentIsec->getVA(minuend.addend);
 238       }
 239       referentVA = minuendVA - fromSym->getVA();
 240     } else if (auto *referentSym = r.referent.dyn_cast<Symbol *>()) {
 241       if (target->hasAttr(r.type, RelocAttrBits::LOAD) &&
 242           !referentSym->isInGot())
 243         target->relaxGotLoad(loc, r.type);
 244       // For dtrace symbols, do not handle them as normal undefined symbols
 245       if (referentSym->getName().starts_with("___dtrace_")) {
 246         // Change dtrace call site to pre-defined instructions
 247         target->handleDtraceReloc(referentSym, r, loc);
 248         continue;
 249       }
 250       referentVA = resolveSymbolVA(referentSym, r.type) + r.addend;
 251
 252       if (isThreadLocalVariables(getFlags()) && isa<Defined>(referentSym)) {
 253         // References from thread-local variable sections are treated as offsets
 254         // relative to the start of the thread-local data memory area, which
 255         // is initialized via copying all the TLV data sections (which are all
 256         // contiguous).
 257         referentVA -= firstTLVDataSection->addr;
 258       } else if (needsFixup) {
 259         writeChainedFixup(loc, referentSym, r.addend);
 260         continue;
 261       }
 262     } else if (auto *referentIsec = r.referent.dyn_cast<InputSection *>()) {
 263       assert(!::shouldOmitFromOutput(referentIsec));
 264       referentVA = referentIsec->getVA(r.addend);
 265
 266       if (needsFixup) {
 267         writeChainedRebase(loc, referentVA);
 268         continue;
 269       }
 270     }
 271     target->relocateOne(loc, r, referentVA, getVA() + r.offset);
 272   }
 273 }
 274
 275 ConcatInputSection *macho::makeSyntheticInputSection(StringRef segName,
 276                                                      StringRef sectName,
 277                                                      uint32_t flags,
 278                                                      ArrayRef<uint8_t> data,
 279                                                      uint32_t align) {
 280   Section &section =
 281       *make<Section>(/*file=*/nullptr, segName, sectName, flags, /*addr=*/0);
 282   auto isec = make<ConcatInputSection>(section, data, align);
 283   // Since this is an explicitly created 'fake' input section,
 284   // it should not be dead stripped.
 285   isec->live = true;
 286   section.subsections.push_back({0, isec});
 287   return isec;
 288 }
 289
 290 void CStringInputSection::splitIntoPieces() {
 291   size_t off = 0;
 292   StringRef s = toStringRef(data);
 293   while (!s.empty()) {
 294     size_t end = s.find(0);
 295     if (end == StringRef::npos)
 296       fatal(getLocation(off) + ": string is not null terminated");
 297     uint32_t hash = deduplicateLiterals ? xxh3_64bits(s.take_front(end)) : 0;
 298     pieces.emplace_back(off, hash);
 299     size_t size = end + 1; // include null terminator
 300     s = s.substr(size);
 301     off += size;
 302   }
 303 }
 304
 305 StringPiece &CStringInputSection::getStringPiece(uint64_t off) {
 306   if (off >= data.size())
 307     fatal(toString(this) + ": offset is outside the section");
 308
 309   auto it =
 310       partition_point(pieces, [=](StringPiece p) { return p.inSecOff <= off; });
 311   return it[-1];
 312 }
 313
 314 const StringPiece &CStringInputSection::getStringPiece(uint64_t off) const {
 315   return const_cast<CStringInputSection *>(this)->getStringPiece(off);
 316 }
 317
 318 size_t CStringInputSection::getStringPieceIndex(uint64_t off) const {
 319   if (off >= data.size())
 320     fatal(toString(this) + ": offset is outside the section");
 321
 322   auto it =
 323       partition_point(pieces, [=](StringPiece p) { return p.inSecOff <= off; });
 324   return std::distance(pieces.begin(), it) - 1;
 325 }
 326
 327 uint64_t CStringInputSection::getOffset(uint64_t off) const {
 328   const StringPiece &piece = getStringPiece(off);
 329   uint64_t addend = off - piece.inSecOff;
 330   return piece.outSecOff + addend;
 331 }
 332
 333 WordLiteralInputSection::WordLiteralInputSection(const Section &section,
 334                                                  ArrayRef<uint8_t> data,
 335                                                  uint32_t align)
 336     : InputSection(WordLiteralKind, section, data, align) {
 337   switch (sectionType(getFlags())) {
 338   case S_4BYTE_LITERALS:
 339     power2LiteralSize = 2;
 340     break;
 341   case S_8BYTE_LITERALS:
 342     power2LiteralSize = 3;
 343     break;
 344   case S_16BYTE_LITERALS:
 345     power2LiteralSize = 4;
 346     break;
 347   default:
 348     llvm_unreachable("invalid literal section type");
 349   }
 350
 351   live.resize(data.size() >> power2LiteralSize, !config->deadStrip);
 352 }
 353
 354 uint64_t WordLiteralInputSection::getOffset(uint64_t off) const {
 355   auto *osec = cast<WordLiteralSection>(parent);
 356   const uintptr_t buf = reinterpret_cast<uintptr_t>(data.data());
 357   switch (sectionType(getFlags())) {
 358   case S_4BYTE_LITERALS:
 359     return osec->getLiteral4Offset(buf + (off & ~3LLU)) | (off & 3);
 360   case S_8BYTE_LITERALS:
 361     return osec->getLiteral8Offset(buf + (off & ~7LLU)) | (off & 7);
 362   case S_16BYTE_LITERALS:
 363     return osec->getLiteral16Offset(buf + (off & ~15LLU)) | (off & 15);
 364   default:
 365     llvm_unreachable("invalid literal section type");
 366   }
 367 }
 368
 369 bool macho::isCodeSection(const InputSection *isec) {
 370   return sections::isCodeSection(isec->getName(), isec->getSegName(),
 371                                  isec->getFlags());
 372 }
 373
 374 bool macho::isCfStringSection(const InputSection *isec) {
 375   return isec->getName() == section_names::cfString &&
 376          isec->getSegName() == segment_names::data;
 377 }
 378
 379 bool macho::isClassRefsSection(const InputSection *isec) {
 380   return isec->getName() == section_names::objcClassRefs &&
 381          isec->getSegName() == segment_names::data;
 382 }
 383
 384 bool macho::isSelRefsSection(const InputSection *isec) {
 385   return isec->getName() == section_names::objcSelrefs &&
 386          isec->getSegName() == segment_names::data;
 387 }
 388
 389 bool macho::isEhFrameSection(const InputSection *isec) {
 390   return isec->getName() == section_names::ehFrame &&
 391          isec->getSegName() == segment_names::text;
 392 }
 393
 394 bool macho::isGccExceptTabSection(const InputSection *isec) {
 395   return isec->getName() == section_names::gccExceptTab &&
 396          isec->getSegName() == segment_names::text;
 397 }
 398
 399 std::string lld::toString(const InputSection *isec) {
 400   return (toString(isec->getFile()) + ":(" + isec->getName() + ")").str();
 401 }