[obj2yaml/yaml2obj] Add section group support.

[shankarke]

Summary

This adds support of reading and writing section groups.

[atanasyan]

include/llvm/Object/ELFYAML.h

88

Can we remove now the Info member field from the RelocationSection class?

lib/Object/ELFYAML.cpp

526

Is it possible/correct to get or create a group section with missed Members list? Should we made this field mandatory?

tools/obj2yaml/elf2yaml.cpp

306

I would write this loop with the if/else statement. It's a bit shorter and emphasizes the fact that we have two kinds of group section members:

for (int i = 0; i < count; i++) {
  if (groupMembers[i] == llvm::ELF::GRP_COMDAT) {
    s.sectionNameOrType = "GRP_COMDAT";
  } else {
    const Elf_Shdr *sHdr = Obj.getSection(groupMembers[i]);
    ErrorOr<StringRef> sectionName = Obj.getSectionName(sHdr);
    if (std::error_code ec = sectionName.getError())
      return ec;
    s.sectionNameOrType = *sectionName;
  }
  S->Members.push_back(s);
}

tools/yaml2obj/yaml2elf.cpp

391

Why does this error message mention the relocation section type?

[shankarke]

include/llvm/Object/ELFYAML.h

88

Good catch. Fixed.

lib/Object/ELFYAML.cpp

526

Yes we need to make it mandatory. I was thinking if we would want to create invalid tests cases using these tools.

tools/obj2yaml/elf2yaml.cpp

306

Sure.

tools/yaml2obj/yaml2elf.cpp

391

Copy/paste error. Will fix.

[shankarke]

Updated with comments from atanasyan.

[atanasyan]

LGTM

[shankarke]

https://llvm.org/svn/llvm-project/llvm/trunk@230124

include/llvm/Object/ELFYAML.h

struct Symbol {
   llvm::yaml::Hex64 Size;
   uint8_t Other;
 };
 struct LocalGlobalWeakSymbols {
   std::vector<Symbol> Local;
   std::vector<Symbol> Global;
   std::vector<Symbol> Weak;
 };
+
+struct SectionOrType {
+  StringRef sectionNameOrType;
+};
+
 struct Section {
-  enum class SectionKind { RawContent, Relocation };
+  enum class SectionKind { Group, RawContent, Relocation };
   SectionKind Kind;
   StringRef Name;
   ELF_SHT Type;
   ELF_SHF Flags;
   llvm::yaml::Hex64 Address;
   StringRef Link;
+  StringRef Info;
   llvm::yaml::Hex64 AddressAlign;
   Section(SectionKind Kind) : Kind(Kind) {}
   virtual ~Section();
 };
 struct RawContentSection : Section {
   yaml::BinaryRef Content;
   llvm::yaml::Hex64 Size;
   RawContentSection() : Section(SectionKind::RawContent) {}
   static bool classof(const Section *S) {
     return S->Kind == SectionKind::RawContent;
   }
 };
+
+struct Group : Section {
+  // Members of a group contain a flag and a list of section indices
+  // that are part of the group.
+  std::vector<SectionOrType> Members;
+  Group() : Section(SectionKind::Group) {}
+  static bool classof(const Section *S) {
+    return S->Kind == SectionKind::Group;
+  }
+};
+
 struct Relocation {
   llvm::yaml::Hex64 Offset;
   int64_t Addend;
   ELF_REL Type;
   StringRef Symbol;
 };
 struct RelocationSection : Section {
-  StringRef Info;
   std::vector<Relocation> Relocations;
   RelocationSection() : Section(SectionKind::Relocation) {}
   static bool classof(const Section *S) {
     return S->Kind == SectionKind::Relocation;
   }
 };
 struct Object {
   FileHeader Header;
   std::vector<std::unique_ptr<Section>> Sections;
   // Although in reality the symbols reside in a section, it is a lot
   // cleaner and nicer if we read them from the YAML as a separate
   // top-level key, which automatically ensures that invariants like there
   // being a single SHT_SYMTAB section are upheld.
   LocalGlobalWeakSymbols Symbols;
 };
 
 } // end namespace ELFYAML
 } // end namespace llvm
 
 LLVM_YAML_IS_SEQUENCE_VECTOR(std::unique_ptr<llvm::ELFYAML::Section>)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::Symbol)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::Relocation)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::SectionOrType)
 
 namespace llvm {
 namespace yaml {
 
 template <>
 struct ScalarEnumerationTraits<ELFYAML::ELF_ET> {
   static void enumeration(IO &IO, ELFYAML::ELF_ET &Value);
 };
struct MappingTraits<std::unique_ptr<ELFYAML::Section>> {
   static StringRef validate(IO &io, std::unique_ptr<ELFYAML::Section> &Section);
 };
 
 template <>
 struct MappingTraits<ELFYAML::Object> {
   static void mapping(IO &IO, ELFYAML::Object &Object);
 };
 
+template <> struct MappingTraits<ELFYAML::SectionOrType> {
+  static void mapping(IO &IO, ELFYAML::SectionOrType &sectionOrType);
+};
+
 } // end namespace yaml
 } // end namespace llvm
 
 #endif

lib/Object/ELFYAML.cpp

 
 static void commonSectionMapping(IO &IO, ELFYAML::Section &Section) {
   IO.mapOptional("Name", Section.Name, StringRef());
   IO.mapRequired("Type", Section.Type);
   IO.mapOptional("Flags", Section.Flags, ELFYAML::ELF_SHF(0));
   IO.mapOptional("Address", Section.Address, Hex64(0));
   IO.mapOptional("Link", Section.Link, StringRef());
   IO.mapOptional("AddressAlign", Section.AddressAlign, Hex64(0));
+  IO.mapOptional("Info", Section.Info, StringRef());
 }
 
 static void sectionMapping(IO &IO, ELFYAML::RawContentSection &Section) {
   commonSectionMapping(IO, Section);
   IO.mapOptional("Content", Section.Content);
   IO.mapOptional("Size", Section.Size, Hex64(Section.Content.binary_size()));
 }
 
 static void sectionMapping(IO &IO, ELFYAML::RelocationSection &Section) {
   commonSectionMapping(IO, Section);
-  IO.mapOptional("Info", Section.Info, StringRef());
   IO.mapOptional("Relocations", Section.Relocations);
 }
 
+static void groupSectionMapping(IO &IO, ELFYAML::Group &group) {
+  commonSectionMapping(IO, group);
+  IO.mapRequired("Members", group.Members);
+}
+
+void MappingTraits<ELFYAML::SectionOrType>::mapping(
+    IO &IO, ELFYAML::SectionOrType &sectionOrType) {
+  IO.mapRequired("SectionOrType", sectionOrType.sectionNameOrType);
+}
+
 void MappingTraits<std::unique_ptr<ELFYAML::Section>>::mapping(
     IO &IO, std::unique_ptr<ELFYAML::Section> &Section) {
   ELFYAML::ELF_SHT sectionType;
   if (IO.outputting())
     sectionType = Section->Type;
   else
     IO.mapRequired("Type", sectionType);
 
   switch (sectionType) {
   case ELF::SHT_REL:
   case ELF::SHT_RELA:
     if (!IO.outputting())
       Section.reset(new ELFYAML::RelocationSection());
     sectionMapping(IO, *cast<ELFYAML::RelocationSection>(Section.get()));
     break;
+  case ELF::SHT_GROUP:
+    if (!IO.outputting())
+      Section.reset(new ELFYAML::Group());
+    groupSectionMapping(IO, *cast<ELFYAML::Group>(Section.get()));
+    break;
   default:
     if (!IO.outputting())
       Section.reset(new ELFYAML::RawContentSection());
     sectionMapping(IO, *cast<ELFYAML::RawContentSection>(Section.get()));
   }
 }
 
 StringRef MappingTraits<std::unique_ptr<ELFYAML::Section>>::validate(

test/Object/obj2yaml-sectiongroup.test

+# Checks that the tool is able to read section groups with ELF.
+RUN: obj2yaml %p/Inputs/sectionGroup.elf.x86-64 > %t1.sectiongroup.yaml
+RUN: FileCheck %s --check-prefix ELF-GROUP < %t1.sectiongroup.yaml
+RUN: yaml2obj -format=elf %t1.sectiongroup.yaml -o %t2.o.elf
+RUN: llvm-readobj -sections %t2.o.elf | FileCheck %s -check-prefix=SECTIONS
+#ELF-GROUP:  - Name:            .group
+#ELF-GROUP:    Type:            SHT_GROUP
+#ELF-GROUP:    Link:            .symtab
+#ELF-GROUP:    Info:            a
+#ELF-GROUP:    Members:
+#ELF-GROUP:      - SectionOrType:   GRP_COMDAT
+#ELF-GROUP:      - SectionOrType:   .rodata.a
+#SECTIONS: Format: ELF64-x86-64
+#SECTIONS: Arch: x86_64
+#SECTIONS: AddressSize: 64bit
+#SECTIONS:   Section {
+#SECTIONS:     Index: 1
+#SECTIONS:     Name: .group (21)
+#SECTIONS:     Type: SHT_GROUP (0x11)
+#SECTIONS:     Flags [ (0x0)
+#SECTIONS:     ]
+#SECTIONS:     Address: 0x0
+#SECTIONS:     Size: 8
+#SECTIONS:     AddressAlignment: 4
+#SECTIONS:     EntrySize: 4
+#SECTIONS:   }

tools/obj2yaml/elf2yaml.cpp

 #include "llvm/Support/YAMLTraits.h"
 
 using namespace llvm;
 
 namespace {
 
 template <class ELFT>
 class ELFDumper {
+  typedef object::Elf_Sym_Impl<ELFT> Elf_Sym;
   typedef typename object::ELFFile<ELFT>::Elf_Shdr Elf_Shdr;
   typedef typename object::ELFFile<ELFT>::Elf_Sym_Iter Elf_Sym_Iter;
+  typedef typename object::ELFFile<ELFT>::Elf_Word Elf_Word;
 
   const object::ELFFile<ELFT> &Obj;
 
   std::error_code dumpSymbol(Elf_Sym_Iter Sym, ELFYAML::Symbol &S);
   std::error_code dumpCommonSection(const Elf_Shdr *Shdr, ELFYAML::Section &S);
   std::error_code dumpCommonRelocationSection(const Elf_Shdr *Shdr,
                                               ELFYAML::RelocationSection &S);
   template <class RelT>
   std::error_code dumpRelocation(const Elf_Shdr *Shdr, const RelT *Rel,
                                  ELFYAML::Relocation &R);
 
   ErrorOr<ELFYAML::RelocationSection *> dumpRelSection(const Elf_Shdr *Shdr);
   ErrorOr<ELFYAML::RelocationSection *> dumpRelaSection(const Elf_Shdr *Shdr);
   ErrorOr<ELFYAML::RawContentSection *>
   dumpContentSection(const Elf_Shdr *Shdr);
+  ErrorOr<ELFYAML::Group *> dumpGroup(const Elf_Shdr *Shdr);
 
 public:
   ELFDumper(const object::ELFFile<ELFT> &O);
   ErrorOr<ELFYAML::Object *> dump();
 };
 
 }
 
for (const Elf_Shdr &Sec : Obj.sections()) {
     }
     case ELF::SHT_REL: {
       ErrorOr<ELFYAML::RelocationSection *> S = dumpRelSection(&Sec);
       if (std::error_code EC = S.getError())
         return EC;
       Y->Sections.push_back(std::unique_ptr<ELFYAML::Section>(S.get()));
       break;
     }
-    // FIXME: Support SHT_GROUP section format.
+    case ELF::SHT_GROUP: {
+      ErrorOr<ELFYAML::Group *> G = dumpGroup(&Sec);
+      if (std::error_code EC = G.getError())
+        return EC;
+      Y->Sections.push_back(std::unique_ptr<ELFYAML::Section>(G.get()));
+      break;
+    }
     default: {
       ErrorOr<ELFYAML::RawContentSection *> S = dumpContentSection(&Sec);
       if (std::error_code EC = S.getError())
         return EC;
       Y->Sections.push_back(std::unique_ptr<ELFYAML::Section>(S.get()));
     }
     }
   }
if (std::error_code EC = ContentOrErr.getError())
     return EC;
   S->Content = yaml::BinaryRef(ContentOrErr.get());
   S->Size = S->Content.binary_size();
 
   return S.release();
 }
 
 template <class ELFT>
+ErrorOr<ELFYAML::Group *> ELFDumper<ELFT>::dumpGroup(const Elf_Shdr *Shdr) {
+  auto S = make_unique<ELFYAML::Group>();
+
+  if (std::error_code EC = dumpCommonSection(Shdr, *S))
+    return EC;
+  // Get sh_info which is the signature.
+  const Elf_Sym *symbol = Obj.getSymbol(Shdr->sh_info);
+  const Elf_Shdr *symtab = Obj.getSection(Shdr->sh_link);
+  auto sectionContents = Obj.getSectionContents(Shdr);
+  if (std::error_code ec = sectionContents.getError())
+    return ec;
+  ErrorOr<StringRef> symbolName = Obj.getSymbolName(symtab, symbol);
+  if (std::error_code EC = symbolName.getError())
+    return EC;
+  S->Info = *symbolName;
+  const Elf_Word *groupMembers =
+      reinterpret_cast<const Elf_Word *>(sectionContents->data());
+  const long count = (Shdr->sh_size) / sizeof(Elf_Word);
+  ELFYAML::SectionOrType s;
+  for (int i = 0; i < count; i++) {
+    if (groupMembers[i] == llvm::ELF::GRP_COMDAT) {
+      s.sectionNameOrType = "GRP_COMDAT";
+    } else {
+      const Elf_Shdr *sHdr = Obj.getSection(groupMembers[i]);
+      ErrorOr<StringRef> sectionName = Obj.getSectionName(sHdr);
+      if (std::error_code ec = sectionName.getError())
+        return ec;
+      s.sectionNameOrType = *sectionName;
+    }
+    S->Members.push_back(s);
+  }
+  return S.release();
+}
+
+template <class ELFT>
 static std::error_code elf2yaml(raw_ostream &Out,
                                 const object::ELFFile<ELFT> &Obj) {
   ELFDumper<ELFT> Dumper(Obj);
   ErrorOr<ELFYAML::Object *> YAMLOrErr = Dumper.dump();
   if (std::error_code EC = YAMLOrErr.getError())
     return EC;
 
   std::unique_ptr<ELFYAML::Object> YAML(YAMLOrErr.get());

tools/yaml2obj/yaml2elf.cpp

class ELFState {
   void addSymbols(const std::vector<ELFYAML::Symbol> &Symbols,
                   std::vector<Elf_Sym> &Syms, unsigned SymbolBinding);
   void writeSectionContent(Elf_Shdr &SHeader,
                            const ELFYAML::RawContentSection &Section,
                            ContiguousBlobAccumulator &CBA);
   bool writeSectionContent(Elf_Shdr &SHeader,
                            const ELFYAML::RelocationSection &Section,
                            ContiguousBlobAccumulator &CBA);
+  bool writeSectionContent(Elf_Shdr &SHeader, const ELFYAML::Group &Group,
+                           ContiguousBlobAccumulator &CBA);
 
   // - SHT_NULL entry (placed first, i.e. 0'th entry)
   // - symbol table (.symtab) (placed third to last)
   // - string table (.strtab) (placed second to last)
   // - section header string table (.shstrtab) (placed last)
   unsigned getDotSymTabSecNo() const { return Doc.Sections.size() + 1; }
   unsigned getDotStrTabSecNo() const { return Doc.Sections.size() + 2; }
   unsigned getDotShStrTabSecNo() const { return Doc.Sections.size() + 3; }
else if (auto S = dyn_cast<ELFYAML::RelocationSection>(Sec.get())) {
         errs() << "error: Unknown section referenced: '" << S->Info
                << "' at YAML section '" << S->Name << "'.\n";
         return false;
       }
       SHeader.sh_info = Index;
 
       if (!writeSectionContent(SHeader, *S, CBA))
         return false;
+    } else if (auto S = dyn_cast<ELFYAML::Group>(Sec.get())) {
+      unsigned SymIdx;
+      if (SymN2I.lookup(S->Info, SymIdx)) {
+        errs() << "error: Unknown symbol referenced: '" << S->Info
+               << "' at YAML section '" << S->Name << "'.\n";
+        return false;
+      }
+      SHeader.sh_info = SymIdx;
+      if (!writeSectionContent(SHeader, *S, CBA))
+        return false;
     } else
       llvm_unreachable("Unknown section type");
 
     SHeaders.push_back(SHeader);
   }
   return true;
 }
 
if (IsRela) {
       REntry.r_offset = Rel.Offset;
       REntry.setSymbolAndType(SymIdx, Rel.Type, isMips64EL(Doc));
       OS.write((const char *)&REntry, sizeof(REntry));
     }
   }
   return true;
 }
 
+template <class ELFT>
+bool ELFState<ELFT>::writeSectionContent(Elf_Shdr &SHeader,
+                                         const ELFYAML::Group &Section,
+                                         ContiguousBlobAccumulator &CBA) {
+  typedef typename object::ELFFile<ELFT>::Elf_Word Elf_Word;
+  if (Section.Type != llvm::ELF::SHT_GROUP) {
+    errs() << "error: Invalid section type.\n";
+    return false;
+  }
+
+  SHeader.sh_entsize = sizeof(Elf_Word);
+  SHeader.sh_size = SHeader.sh_entsize * Section.Members.size();
+
+  auto &OS = CBA.getOSAndAlignedOffset(SHeader.sh_offset);
+
+  for (auto member : Section.Members) {
+    Elf_Word SIdx;
+    unsigned int sectionIndex = 0;
+    if (member.sectionNameOrType == "GRP_COMDAT")
+      sectionIndex = llvm::ELF::GRP_COMDAT;
+    else if (SN2I.lookup(member.sectionNameOrType, sectionIndex)) {
+      errs() << "error: Unknown section referenced: '"
+             << member.sectionNameOrType << "' at YAML section' "
+             << Section.Name << "\n";
+      return false;
+    }
+    SIdx = sectionIndex;
+    OS.write((const char *)&SIdx, sizeof(SIdx));
+  }
+  return true;
+}
+
 template <class ELFT> bool ELFState<ELFT>::buildSectionIndex() {
   SN2I.addName(".symtab", getDotSymTabSecNo());
   SN2I.addName(".strtab", getDotStrTabSecNo());
   SN2I.addName(".shstrtab", getDotShStrTabSecNo());
 
   for (unsigned i = 0, e = Doc.Sections.size(); i != e; ++i) {
     StringRef Name = Doc.Sections[i]->Name;
     if (Name.empty())