[ELF] Change rawData to content() and data() to contentMaybeDecompress()

Clarify data() which may trigger decompression and make it feasible to refactor
the member variable rawData.

Author: MaskRay

lld/ELF/AArch64ErrataFix.cpp

@@ -350,7 +350,7 @@ static uint64_t scanCortexA53Errata843419(InputSection *isec, uint64_t &off,
   }
 
   uint64_t patchOff = 0;
-  const uint8_t *buf = isec->rawData.begin();
+  const uint8_t *buf = isec->content().begin();
   const ulittle32_t *instBuf = reinterpret_cast<const ulittle32_t *>(buf + off);
   uint32_t instr1 = *instBuf++;
   uint32_t instr2 = *instBuf++;
@@ -409,7 +409,7 @@ uint64_t Patch843419Section::getLDSTAddr() const {
 void Patch843419Section::writeTo(uint8_t *buf) {
   // Copy the instruction that we will be replacing with a branch in the
   // patchee Section.
-  write32le(buf, read32le(patchee->rawData.begin() + patcheeOffset));
+  write32le(buf, read32le(patchee->content().begin() + patcheeOffset));
 
   // Apply any relocation transferred from the original patchee section.
   target->relocateAlloc(*this, buf);
@@ -591,8 +591,8 @@ AArch64Err843419Patcher::patchInputSectionDescription(
     while (codeSym != mapSyms.end()) {
       auto dataSym = std::next(codeSym);
       uint64_t off = (*codeSym)->value;
-      uint64_t limit =
-          (dataSym == mapSyms.end()) ? isec->rawData.size() : (*dataSym)->value;
+      uint64_t limit = (dataSym == mapSyms.end()) ? isec->content().size()
+                                                  : (*dataSym)->value;
 
       while (off < limit) {
         uint64_t startAddr = isec->getVA(off);

lld/ELF/ARMErrataFix.cpp

@@ -266,7 +266,7 @@ static ScanResult scanCortexA8Errata657417(InputSection *isec, uint64_t &off,
   }
 
   ScanResult scanRes = {0, 0, nullptr};
-  const uint8_t *buf = isec->rawData.begin();
+  const uint8_t *buf = isec->content().begin();
   // ARMv7-A Thumb 32-bit instructions are encoded 2 consecutive
   // little-endian halfwords.
   const ulittle16_t *instBuf = reinterpret_cast<const ulittle16_t *>(buf + off);
@@ -498,8 +498,8 @@ ARMErr657417Patcher::patchInputSectionDescription(
     while (thumbSym != mapSyms.end()) {
       auto nonThumbSym = std::next(thumbSym);
       uint64_t off = (*thumbSym)->value;
-      uint64_t limit = (nonThumbSym == mapSyms.end()) ? isec->rawData.size()
-                                                      : (*nonThumbSym)->value;
+      uint64_t limit = nonThumbSym == mapSyms.end() ? isec->content().size()
+                                                    : (*nonThumbSym)->value;
 
       while (off < limit) {
         ScanResult sr = scanCortexA8Errata657417(isec, off, limit);

lld/ELF/Arch/PPC64.cpp

@@ -1561,7 +1561,7 @@ void PPC64::relocateAlloc(InputSectionBase &sec, uint8_t *buf) const {
         // recursive calls even if the function is preemptible. This is not
         // wrong in the common case where the function is not preempted at
         // runtime. Just ignore.
-        if ((rel.offset + 8 > sec.rawData.size() ||
+        if ((rel.offset + 8 > sec.content().size() ||
              read32(loc + 4) != 0x60000000) &&
             rel.sym->file != sec.file) {
           // Use substr(6) to remove the "__plt_" prefix.

lld/ELF/Arch/RISCV.cpp

@@ -556,7 +556,7 @@ static void relaxCall(const InputSection &sec, size_t i, uint64_t loc,
                       Relocation &r, uint32_t &remove) {
   const bool rvc = config->eflags & EF_RISCV_RVC;
   const Symbol &sym = *r.sym;
-  const uint64_t insnPair = read64le(sec.rawData.data() + r.offset);
+  const uint64_t insnPair = read64le(sec.content().data() + r.offset);
   const uint32_t rd = extractBits(insnPair, 32 + 11, 32 + 7);
   const uint64_t dest =
       (r.expr == R_PLT_PC ? sym.getPltVA() : sym.getVA()) + r.addend;
@@ -584,7 +584,7 @@ static void relaxTlsLe(const InputSection &sec, size_t i, uint64_t loc,
   uint64_t val = r.sym->getVA(r.addend);
   if (hi20(val) != 0)
     return;
-  uint32_t insn = read32le(sec.rawData.data() + r.offset);
+  uint32_t insn = read32le(sec.content().data() + r.offset);
   switch (r.type) {
   case R_RISCV_TPREL_HI20:
   case R_RISCV_TPREL_ADD:
@@ -728,7 +728,7 @@ void elf::riscvFinalizeRelax(int passes) {
         continue;
 
       auto &rels = sec->relocations;
-      ArrayRef<uint8_t> old = sec->rawData;
+      ArrayRef<uint8_t> old = sec->content();
       size_t newSize =
           old.size() - aux.relocDeltas[sec->relocations.size() - 1];
       size_t writesIdx = 0;

lld/ELF/Arch/X86_64.cpp

@@ -253,7 +253,7 @@ bool X86_64::deleteFallThruJmpInsn(InputSection &is, InputFile *file,
   Relocation &r = is.relocations[rIndex];
 
   // Check if the relocation corresponds to a direct jmp.
-  const uint8_t *secContents = is.rawData.data();
+  const uint8_t *secContents = is.content().data();
   // If it is not a direct jmp instruction, there is nothing to do here.
   if (*(secContents + r.offset - 1) != 0xe9)
     return false;

lld/ELF/DWARF.cpp

@@ -44,17 +44,17 @@ template <class ELFT> LLDDwarfObj<ELFT>::LLDDwarfObj(ObjFile<ELFT> *obj) {
                 .Case(".debug_str_offsets", &strOffsetsSection)
                 .Case(".debug_line", &lineSection)
                 .Default(nullptr)) {
-      m->Data = toStringRef(sec->data());
+      m->Data = toStringRef(sec->contentMaybeDecompress());
       m->sec = sec;
       continue;
     }
 
     if (sec->name == ".debug_abbrev")
-      abbrevSection = toStringRef(sec->data());
+      abbrevSection = toStringRef(sec->contentMaybeDecompress());
     else if (sec->name == ".debug_str")
-      strSection = toStringRef(sec->data());
+      strSection = toStringRef(sec->contentMaybeDecompress());
     else if (sec->name == ".debug_line_str")
-      lineStrSection = toStringRef(sec->data());
+      lineStrSection = toStringRef(sec->contentMaybeDecompress());
     else if (sec->name == ".debug_info" &&
              !(objSections[i].sh_flags & ELF::SHF_GROUP)) {
       // In DWARF v5, -fdebug-types-section places type units in .debug_info
@@ -66,7 +66,7 @@ template <class ELFT> LLDDwarfObj<ELFT>::LLDDwarfObj(ObjFile<ELFT> *obj) {
       // need to perform a lightweight parsing. We drop the SHF_GROUP flag when
       // the InputSection was created, so we need to retrieve sh_flags from the
       // associated ELF section header.
-      infoSection.Data = toStringRef(sec->data());
+      infoSection.Data = toStringRef(sec->contentMaybeDecompress());
       infoSection.sec = sec;
     }
   }

lld/ELF/Driver.cpp

@@ -2123,7 +2123,7 @@ static void readSymbolPartitionSection(InputSectionBase *s) {
   if (!isa<Defined>(sym) || !sym->includeInDynsym())
     return;
 
-  StringRef partName = reinterpret_cast<const char *>(s->rawData.data());
+  StringRef partName = reinterpret_cast<const char *>(s->content().data());
   for (Partition &part : partitions) {
     if (part.name == partName) {
       sym->partition = part.getNumber();

lld/ELF/EhFrame.cpp

@@ -42,7 +42,7 @@ class EhReader {
 private:
   template <class P> void failOn(const P *loc, const Twine &msg) {
     fatal("corrupted .eh_frame: " + msg + "\n>>> defined in " +
-          isec->getObjMsg((const uint8_t *)loc - isec->rawData.data()));
+          isec->getObjMsg((const uint8_t *)loc - isec->content().data()));
   }
 
   uint8_t readByte();

lld/ELF/ICF.cpp

@@ -313,7 +313,7 @@ bool ICF<ELFT>::constantEq(const InputSection *secA, ArrayRef<RelTy> ra,
 template <class ELFT>
 bool ICF<ELFT>::equalsConstant(const InputSection *a, const InputSection *b) {
   if (a->flags != b->flags || a->getSize() != b->getSize() ||
-      a->rawData != b->rawData)
+      a->content() != b->content())
     return false;
 
   // If two sections have different output sections, we cannot merge them.
@@ -492,7 +492,7 @@ template <class ELFT> void ICF<ELFT>::run() {
   // Initially, we use hash values to partition sections.
   parallelForEach(sections, [&](InputSection *s) {
     // Set MSB to 1 to avoid collisions with unique IDs.
-    s->eqClass[0] = xxHash64(s->rawData) | (1U << 31);
+    s->eqClass[0] = xxHash64(s->content()) | (1U << 31);
   });
 
   // Perform 2 rounds of relocation hash propagation. 2 is an empirical value to

lld/ELF/InputFiles.cpp

@@ -809,8 +809,9 @@ void ObjFile<ELFT>::initializeSections(bool ignoreComdats,
       // simply handle such sections as non-mergeable ones. Degrading like this
       // is acceptable because section merging is optional.
       if (auto *ms = dyn_cast<MergeInputSection>(s)) {
-        s = makeThreadLocal<InputSection>(ms->file, ms->flags, ms->type,
-                                          ms->alignment, ms->data(), ms->name);
+        s = makeThreadLocal<InputSection>(
+            ms->file, ms->flags, ms->type, ms->alignment,
+            ms->contentMaybeDecompress(), ms->name);
         sections[info] = s;
       }
 
@@ -877,10 +878,10 @@ template <class ELFT> static uint32_t readAndFeatures(const InputSection &sec) {
   using Elf_Note = typename ELFT::Note;
 
   uint32_t featuresSet = 0;
-  ArrayRef<uint8_t> data = sec.rawData;
+  ArrayRef<uint8_t> data = sec.content();
   auto reportFatal = [&](const uint8_t *place, const char *msg) {
     fatal(toString(sec.file) + ":(" + sec.name + "+0x" +
-          Twine::utohexstr(place - sec.rawData.data()) + "): " + msg);
+          Twine::utohexstr(place - sec.content().data()) + "): " + msg);
   };
   while (!data.empty()) {
     // Read one NOTE record.

lld/ELF/InputSection.cpp

@@ -59,7 +59,7 @@ InputSectionBase::InputSectionBase(InputFile *file, uint64_t flags,
   // In order to reduce memory allocation, we assume that mergeable
   // sections are smaller than 4 GiB, which is not an unreasonable
   // assumption as of 2017.
-  if (sectionKind == SectionBase::Merge && rawData.size() > UINT32_MAX)
+  if (sectionKind == SectionBase::Merge && content().size() > UINT32_MAX)
     error(toString(this) + ": section too large");
 
   // The ELF spec states that a value of 0 means the section has
@@ -105,14 +105,15 @@ size_t InputSectionBase::getSize() const {
     return s->getSize();
   if (compressed)
     return size;
-  return rawData.size() - bytesDropped;
+  return content().size() - bytesDropped;
 }
 
 template <class ELFT>
 static void decompressAux(const InputSectionBase &sec, uint8_t *out,
                           size_t size) {
-  auto *hdr = reinterpret_cast<const typename ELFT::Chdr *>(sec.rawData.data());
-  auto compressed = sec.rawData.slice(sizeof(typename ELFT::Chdr));
+  auto *hdr =
+      reinterpret_cast<const typename ELFT::Chdr *>(sec.content().data());
+  auto compressed = sec.content().slice(sizeof(typename ELFT::Chdr));
   if (Error e = hdr->ch_type == ELFCOMPRESS_ZLIB
                     ? compression::zlib::decompress(compressed, out, size)
                     : compression::zstd::decompress(compressed, out, size))
@@ -170,7 +171,7 @@ uint64_t SectionBase::getOffset(uint64_t offset) const {
     // Second, InputSection::copyRelocations on .eh_frame. Some pieces may be
     // discarded due to GC/ICF. We should compute the output section offset.
     const EhInputSection *es = cast<EhInputSection>(this);
-    if (!es->rawData.empty())
+    if (!es->content().empty())
       if (InputSection *isec = es->getParent())
         return isec->outSecOff + es->getParentOffset(offset);
     return offset;
@@ -209,12 +210,12 @@ template <typename ELFT> void InputSectionBase::parseCompressedHeader() {
   flags &= ~(uint64_t)SHF_COMPRESSED;
 
   // New-style header
-  if (rawData.size() < sizeof(typename ELFT::Chdr)) {
+  if (content().size() < sizeof(typename ELFT::Chdr)) {
     error(toString(this) + ": corrupted compressed section");
     return;
   }
 
-  auto *hdr = reinterpret_cast<const typename ELFT::Chdr *>(rawData.data());
+  auto *hdr = reinterpret_cast<const typename ELFT::Chdr *>(content().data());
   if (hdr->ch_type == ELFCOMPRESS_ZLIB) {
     if (!compression::zlib::isAvailable())
       error(toString(this) + " is compressed with ELFCOMPRESS_ZLIB, but lld is "
@@ -356,7 +357,7 @@ template <class ELFT, class RelTy>
 void InputSection::copyRelocations(uint8_t *buf, ArrayRef<RelTy> rels) {
   const TargetInfo &target = *elf::target;
   InputSectionBase *sec = getRelocatedSection();
-  (void)sec->data(); // uncompress if needed
+  (void)sec->contentMaybeDecompress(); // uncompress if needed
 
   for (const RelTy &rel : rels) {
     RelType type = rel.getType(config->isMips64EL);
@@ -409,7 +410,7 @@ void InputSection::copyRelocations(uint8_t *buf, ArrayRef<RelTy> rels) {
       }
 
       int64_t addend = getAddend<ELFT>(rel);
-      const uint8_t *bufLoc = sec->rawData.begin() + rel.r_offset;
+      const uint8_t *bufLoc = sec->content().begin() + rel.r_offset;
       if (!RelTy::IsRela)
         addend = target.getImplicitAddend(bufLoc, type);
 
@@ -1103,8 +1104,8 @@ template <class ELFT> void InputSection::writeTo(uint8_t *buf) {
   // If this is a compressed section, uncompress section contents directly
   // to the buffer.
   if (compressed) {
-    auto *hdr = reinterpret_cast<const typename ELFT::Chdr *>(rawData.data());
-    auto compressed = rawData.slice(sizeof(typename ELFT::Chdr));
+    auto *hdr = reinterpret_cast<const typename ELFT::Chdr *>(content().data());
+    auto compressed = content().slice(sizeof(typename ELFT::Chdr));
     size_t size = this->size;
     if (Error e = hdr->ch_type == ELFCOMPRESS_ZLIB
                       ? compression::zlib::decompress(compressed, buf, size)
@@ -1118,8 +1119,8 @@ template <class ELFT> void InputSection::writeTo(uint8_t *buf) {
 
   // Copy section contents from source object file to output file
   // and then apply relocations.
-  memcpy(buf, rawData.data(), rawData.size());
-  relocate<ELFT>(buf, buf + rawData.size());
+  memcpy(buf, content().data(), content().size());
+  relocate<ELFT>(buf, buf + content().size());
 }
 
 void InputSection::replace(InputSection *other) {
@@ -1166,7 +1167,7 @@ template <class ELFT> void EhInputSection::split() {
 
 template <class ELFT, class RelTy>
 void EhInputSection::split(ArrayRef<RelTy> rels) {
-  ArrayRef<uint8_t> d = rawData;
+  ArrayRef<uint8_t> d = content();
   const char *msg = nullptr;
   unsigned relI = 0;
   while (!d.empty()) {
@@ -1190,7 +1191,7 @@ void EhInputSection::split(ArrayRef<RelTy> rels) {
 
     // Find the first relocation that points to [off,off+size). Relocations
     // have been sorted by r_offset.
-    const uint64_t off = d.data() - rawData.data();
+    const uint64_t off = d.data() - content().data();
     while (relI != rels.size() && rels[relI].r_offset < off)
       ++relI;
     unsigned firstRel = -1;
@@ -1201,7 +1202,7 @@ void EhInputSection::split(ArrayRef<RelTy> rels) {
   }
   if (msg)
     errorOrWarn("corrupted .eh_frame: " + Twine(msg) + "\n>>> defined in " +
-                getObjMsg(d.data() - rawData.data()));
+                getObjMsg(d.data() - content().data()));
 }
 
 // Return the offset in an output section for a given input offset.
@@ -1286,13 +1287,13 @@ void MergeInputSection::splitIntoPieces() {
   assert(pieces.empty());
 
   if (flags & SHF_STRINGS)
-    splitStrings(toStringRef(data()), entsize);
+    splitStrings(toStringRef(contentMaybeDecompress()), entsize);
   else
-    splitNonStrings(data(), entsize);
+    splitNonStrings(contentMaybeDecompress(), entsize);
 }
 
 SectionPiece &MergeInputSection::getSectionPiece(uint64_t offset) {
-  if (rawData.size() <= offset)
+  if (content().size() <= offset)
     fatal(toString(this) + ": offset is outside the section");
   return partition_point(
       pieces, [=](SectionPiece p) { return p.inputOff <= offset; })[-1];

lld/ELF/InputSection.h

@@ -164,7 +164,8 @@ class InputSectionBase : public SectionBase {
     }
   }
 
-  ArrayRef<uint8_t> data() const {
+  ArrayRef<uint8_t> content() const { return rawData; }
+  ArrayRef<uint8_t> contentMaybeDecompress() const {
     if (compressed)
       decompress();
     return rawData;
@@ -228,9 +229,9 @@ class InputSectionBase : public SectionBase {
 
 
   template <typename T> llvm::ArrayRef<T> getDataAs() const {
-    size_t s = rawData.size();
+    size_t s = content().size();
     assert(s % sizeof(T) == 0);
-    return llvm::makeArrayRef<T>((const T *)rawData.data(), s / sizeof(T));
+    return llvm::makeArrayRef<T>((const T *)content().data(), s / sizeof(T));
   }
 
 protected:
@@ -288,8 +289,8 @@ class MergeInputSection : public InputSectionBase {
   llvm::CachedHashStringRef getData(size_t i) const {
     size_t begin = pieces[i].inputOff;
     size_t end =
-        (pieces.size() - 1 == i) ? rawData.size() : pieces[i + 1].inputOff;
-    return {toStringRef(rawData.slice(begin, end - begin)), pieces[i].hash};
+        (pieces.size() - 1 == i) ? content().size() : pieces[i + 1].inputOff;
+    return {toStringRef(content().slice(begin, end - begin)), pieces[i].hash};
   }
 
   // Returns the SectionPiece at a given input section offset.
@@ -313,7 +314,7 @@ struct EhSectionPiece {
       : inputOff(off), sec(sec), size(size), firstRelocation(firstRelocation) {}
 
   ArrayRef<uint8_t> data() const {
-    return {sec->rawData.data() + this->inputOff, size};
+    return {sec->content().data() + this->inputOff, size};
   }
 
   size_t inputOff;

lld/ELF/MarkLive.cpp

@@ -75,7 +75,7 @@ template <class ELFT> class MarkLive {
 template <class ELFT>
 static uint64_t getAddend(InputSectionBase &sec,
                           const typename ELFT::Rel &rel) {
-  return target->getImplicitAddend(sec.rawData.begin() + rel.r_offset,
+  return target->getImplicitAddend(sec.content().begin() + rel.r_offset,
                                    rel.getType(config->isMips64EL));
 }