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 "Symbols.h"
  15 #include "SyntheticSections.h"
  16 #include "Target.h"
  17 #include "UnwindInfoSection.h"
  18 #include "Writer.h"
  19 #include "lld/Common/Memory.h"
  20 #include "llvm/Support/Endian.h"
  21 #include "llvm/Support/xxhash.h"
  22
  23 using namespace llvm;
  24 using namespace llvm::MachO;
  25 using namespace llvm::support;
  26 using namespace lld;
  27 using namespace lld::macho;
  28
  29 // Verify ConcatInputSection's size on 64-bit builds. The size of std::vector
  30 // can differ based on STL debug levels (e.g. iterator debugging on MSVC's STL),
  31 // so account for that.
  32 static_assert(sizeof(void *) != 8 || sizeof(ConcatInputSection) ==
  33                                          sizeof(std::vector<Reloc>) + 104,
  34               "Try to minimize ConcatInputSection's size, we create many "
  35               "instances of it");
  36
  37 std::vector<ConcatInputSection *> macho::inputSections;
  38
  39 uint64_t InputSection::getFileSize() const {
  40   return isZeroFill(getFlags()) ? 0 : getSize();
  41 }
  42
  43 uint64_t InputSection::getVA(uint64_t off) const {
  44   return parent->addr + getOffset(off);
  45 }
  46
  47 static uint64_t resolveSymbolVA(const Symbol *sym, uint8_t type) {
  48   const RelocAttrs &relocAttrs = target->getRelocAttrs(type);
  49   if (relocAttrs.hasAttr(RelocAttrBits::BRANCH))
  50     return sym->resolveBranchVA();
  51   if (relocAttrs.hasAttr(RelocAttrBits::GOT))
  52     return sym->resolveGotVA();
  53   if (relocAttrs.hasAttr(RelocAttrBits::TLV))
  54     return sym->resolveTlvVA();
  55   return sym->getVA();
  56 }
  57
  58 const Defined *InputSection::getContainingSymbol(uint64_t off) const {
  59   auto *nextSym = llvm::upper_bound(
  60       symbols, off, [](uint64_t a, const Defined *b) { return a < b->value; });
  61   if (nextSym == symbols.begin())
  62     return nullptr;
  63   return *std::prev(nextSym);
  64 }
  65
  66 std::string InputSection::getLocation(uint64_t off) const {
  67   // First, try to find a symbol that's near the offset. Use it as a reference
  68   // point.
  69   if (auto *sym = getContainingSymbol(off))
  70     return (toString(getFile()) + ":(symbol " + toString(*sym) + "+0x" +
  71             Twine::utohexstr(off - sym->value) + ")")
  72         .str();
  73
  74   // If that fails, use the section itself as a reference point.
  75   for (const Subsection &subsec : section.subsections) {
  76     if (subsec.isec == this) {
  77       off += subsec.offset;
  78       break;
  79     }
  80   }
  81
  82   return (toString(getFile()) + ":(" + getName() + "+0x" +
  83           Twine::utohexstr(off) + ")")
  84       .str();
  85 }
  86
  87 std::string InputSection::getSourceLocation(uint64_t off) const {
  88   auto *obj = dyn_cast_or_null<ObjFile>(getFile());
  89   if (!obj)
  90     return {};
  91
  92   DWARFCache *dwarf = obj->getDwarf();
  93   if (!dwarf)
  94     return std::string();
  95
  96   for (const Subsection &subsec : section.subsections) {
  97     if (subsec.isec == this) {
  98       off += subsec.offset;
  99       break;
 100     }
 101   }
 102
 103   auto createMsg = [&](StringRef path, unsigned line) {
 104     std::string filename = sys::path::filename(path).str();
 105     std::string lineStr = (":" + Twine(line)).str();
 106     if (filename == path)
 107       return filename + lineStr;
 108     return (filename + lineStr + " (" + path + lineStr + ")").str();
 109   };
 110
 111   // First, look up a function for a given offset.
 112   if (Optional<DILineInfo> li = dwarf->getDILineInfo(
 113           section.addr + off, object::SectionedAddress::UndefSection))
 114     return createMsg(li->FileName, li->Line);
 115
 116   // If it failed, look up again as a variable.
 117   if (const Defined *sym = getContainingSymbol(off)) {
 118     // Symbols are generally prefixed with an underscore, which is not included
 119     // in the debug information.
 120     StringRef symName = sym->getName();
 121     if (!symName.empty() && symName[0] == '_')
 122       symName = symName.substr(1);
 123
 124     if (Optional<std::pair<std::string, unsigned>> fileLine =
 125             dwarf->getVariableLoc(symName))
 126       return createMsg(fileLine->first, fileLine->second);
 127   }
 128
 129   // Try to get the source file's name from the DWARF information.
 130   if (obj->compileUnit)
 131     return obj->sourceFile();
 132
 133   return {};
 134 }
 135
 136 void ConcatInputSection::foldIdentical(ConcatInputSection *copy) {
 137   align = std::max(align, copy->align);
 138   copy->live = false;
 139   copy->wasCoalesced = true;
 140   copy->replacement = this;
 141   for (auto &copySym : copy->symbols)
 142     copySym->wasIdenticalCodeFolded = true;
 143
 144   // Merge the sorted vectors of symbols together.
 145   auto it = symbols.begin();
 146   for (auto copyIt = copy->symbols.begin(); copyIt != copy->symbols.end();) {
 147     if (it == symbols.end()) {
 148       symbols.push_back(*copyIt++);
 149       it = symbols.end();
 150     } else if ((*it)->value > (*copyIt)->value) {
 151       std::swap(*it++, *copyIt);
 152     } else {
 153       ++it;
 154     }
 155   }
 156   copy->symbols.clear();
 157
 158   // Remove duplicate compact unwind info for symbols at the same address.
 159   if (symbols.empty())
 160     return;
 161   it = symbols.begin();
 162   uint64_t v = (*it)->value;
 163   for (++it; it != symbols.end(); ++it) {
 164     Defined *d = *it;
 165     if (d->value == v)
 166       d->unwindEntry = nullptr;
 167     else
 168       v = d->value;
 169   }
 170 }
 171
 172 void ConcatInputSection::writeTo(uint8_t *buf) {
 173   assert(!shouldOmitFromOutput());
 174
 175   if (getFileSize() == 0)
 176     return;
 177
 178   memcpy(buf, data.data(), data.size());
 179
 180   std::vector<uint64_t> relocTargets;
 181   if (!optimizationHints.empty())
 182     relocTargets.reserve(relocs.size());
 183
 184   for (size_t i = 0; i < relocs.size(); i++) {
 185     const Reloc &r = relocs[i];
 186     uint8_t *loc = buf + r.offset;
 187     uint64_t referentVA = 0;
 188     if (target->hasAttr(r.type, RelocAttrBits::SUBTRAHEND)) {
 189       const Symbol *fromSym = r.referent.get<Symbol *>();
 190       const Reloc &minuend = relocs[++i];
 191       uint64_t minuendVA;
 192       if (const Symbol *toSym = minuend.referent.dyn_cast<Symbol *>())
 193         minuendVA = toSym->getVA() + minuend.addend;
 194       else {
 195         auto *referentIsec = minuend.referent.get<InputSection *>();
 196         assert(!::shouldOmitFromOutput(referentIsec));
 197         minuendVA = referentIsec->getVA(minuend.addend);
 198       }
 199       referentVA = minuendVA - fromSym->getVA();
 200     } else if (auto *referentSym = r.referent.dyn_cast<Symbol *>()) {
 201       if (target->hasAttr(r.type, RelocAttrBits::LOAD) &&
 202           !referentSym->isInGot())
 203         target->relaxGotLoad(loc, r.type);
 204       // For dtrace symbols, do not handle them as normal undefined symbols
 205       if (referentSym->getName().startswith("___dtrace_")) {
 206         // Change dtrace call site to pre-defined instructions
 207         target->handleDtraceReloc(referentSym, r, loc);
 208         continue;
 209       }
 210       referentVA = resolveSymbolVA(referentSym, r.type) + r.addend;
 211
 212       if (isThreadLocalVariables(getFlags())) {
 213         // References from thread-local variable sections are treated as offsets
 214         // relative to the start of the thread-local data memory area, which
 215         // is initialized via copying all the TLV data sections (which are all
 216         // contiguous).
 217         if (isa<Defined>(referentSym))
 218           referentVA -= firstTLVDataSection->addr;
 219       }
 220     } else if (auto *referentIsec = r.referent.dyn_cast<InputSection *>()) {
 221       assert(!::shouldOmitFromOutput(referentIsec));
 222       referentVA = referentIsec->getVA(r.addend);
 223     }
 224     target->relocateOne(loc, r, referentVA, getVA() + r.offset);
 225
 226     if (!optimizationHints.empty())
 227       relocTargets.push_back(referentVA);
 228   }
 229
 230   if (!optimizationHints.empty())
 231     target->applyOptimizationHints(buf, this, relocTargets);
 232 }
 233
 234 ConcatInputSection *macho::makeSyntheticInputSection(StringRef segName,
 235                                                      StringRef sectName,
 236                                                      uint32_t flags,
 237                                                      ArrayRef<uint8_t> data,
 238                                                      uint32_t align) {
 239   Section &section =
 240       *make<Section>(/*file=*/nullptr, segName, sectName, flags, /*addr=*/0);
 241   auto isec = make<ConcatInputSection>(section, data, align);
 242   section.subsections.push_back({0, isec});
 243   return isec;
 244 }
 245
 246 void CStringInputSection::splitIntoPieces() {
 247   size_t off = 0;
 248   StringRef s = toStringRef(data);
 249   while (!s.empty()) {
 250     size_t end = s.find(0);
 251     if (end == StringRef::npos)
 252       fatal(getLocation(off) + ": string is not null terminated");
 253     size_t size = end + 1;
 254     uint32_t hash = config->dedupLiterals ? xxHash64(s.substr(0, size)) : 0;
 255     pieces.emplace_back(off, hash);
 256     s = s.substr(size);
 257     off += size;
 258   }
 259 }
 260
 261 StringPiece &CStringInputSection::getStringPiece(uint64_t off) {
 262   if (off >= data.size())
 263     fatal(toString(this) + ": offset is outside the section");
 264
 265   auto it =
 266       partition_point(pieces, [=](StringPiece p) { return p.inSecOff <= off; });
 267   return it[-1];
 268 }
 269
 270 const StringPiece &CStringInputSection::getStringPiece(uint64_t off) const {
 271   return const_cast<CStringInputSection *>(this)->getStringPiece(off);
 272 }
 273
 274 uint64_t CStringInputSection::getOffset(uint64_t off) const {
 275   const StringPiece &piece = getStringPiece(off);
 276   uint64_t addend = off - piece.inSecOff;
 277   return piece.outSecOff + addend;
 278 }
 279
 280 WordLiteralInputSection::WordLiteralInputSection(const Section &section,
 281                                                  ArrayRef<uint8_t> data,
 282                                                  uint32_t align)
 283     : InputSection(WordLiteralKind, section, data, align) {
 284   switch (sectionType(getFlags())) {
 285   case S_4BYTE_LITERALS:
 286     power2LiteralSize = 2;
 287     break;
 288   case S_8BYTE_LITERALS:
 289     power2LiteralSize = 3;
 290     break;
 291   case S_16BYTE_LITERALS:
 292     power2LiteralSize = 4;
 293     break;
 294   default:
 295     llvm_unreachable("invalid literal section type");
 296   }
 297
 298   live.resize(data.size() >> power2LiteralSize, !config->deadStrip);
 299 }
 300
 301 uint64_t WordLiteralInputSection::getOffset(uint64_t off) const {
 302   auto *osec = cast<WordLiteralSection>(parent);
 303   const uintptr_t buf = reinterpret_cast<uintptr_t>(data.data());
 304   switch (sectionType(getFlags())) {
 305   case S_4BYTE_LITERALS:
 306     return osec->getLiteral4Offset(buf + (off & ~3LLU)) | (off & 3);
 307   case S_8BYTE_LITERALS:
 308     return osec->getLiteral8Offset(buf + (off & ~7LLU)) | (off & 7);
 309   case S_16BYTE_LITERALS:
 310     return osec->getLiteral16Offset(buf + (off & ~15LLU)) | (off & 15);
 311   default:
 312     llvm_unreachable("invalid literal section type");
 313   }
 314 }
 315
 316 bool macho::isCodeSection(const InputSection *isec) {
 317   uint32_t type = sectionType(isec->getFlags());
 318   if (type != S_REGULAR && type != S_COALESCED)
 319     return false;
 320
 321   uint32_t attr = isec->getFlags() & SECTION_ATTRIBUTES_USR;
 322   if (attr == S_ATTR_PURE_INSTRUCTIONS)
 323     return true;
 324
 325   if (isec->getSegName() == segment_names::text)
 326     return StringSwitch<bool>(isec->getName())
 327         .Cases(section_names::textCoalNt, section_names::staticInit, true)
 328         .Default(false);
 329
 330   return false;
 331 }
 332
 333 bool macho::isCfStringSection(const InputSection *isec) {
 334   return isec->getName() == section_names::cfString &&
 335          isec->getSegName() == segment_names::data;
 336 }
 337
 338 bool macho::isClassRefsSection(const InputSection *isec) {
 339   return isec->getName() == section_names::objcClassRefs &&
 340          isec->getSegName() == segment_names::data;
 341 }
 342
 343 bool macho::isEhFrameSection(const InputSection *isec) {
 344   return isec->getName() == section_names::ehFrame &&
 345          isec->getSegName() == segment_names::text;
 346 }
 347
 348 bool macho::isGccExceptTabSection(const InputSection *isec) {
 349   return isec->getName() == section_names::gccExceptTab &&
 350          isec->getSegName() == segment_names::text;
 351 }
 352
 353 std::string lld::toString(const InputSection *isec) {
 354   return (toString(isec->getFile()) + ":(" + isec->getName() + ")").str();
 355 }