[yaml2obj/obj2yaml] - Add a basic support for extended section indexes.

In some cases a symbol might have section index == SHN_XINDEX.
This is an escape value indicating that the actual section header index
is too large to fit in the containing field.
Then the SHT_SYMTAB_SHNDX section is used. It contains the 32bit values
that stores section indexes.

ELF gABI says that there can be multiple SHT_SYMTAB_SHNDX sections,
i.e. for example one for .symtab and one for .dynsym
(1) https://groups.google.com/forum/#!topic/generic-abi/-XJAV5d8PRg
(2) DT_SYMTAB_SHNDX: http://www.sco.com/developers/gabi/latest/ch5.dynamic.html

In this patch I am only supporting a single SHT_SYMTAB_SHNDX associated
with a .symtab. This is a more or less common case which is used a few tests I saw in LLVM.

I decided not to create the SHT_SYMTAB_SHNDX section as "implicit",
but implement is like a kind of regular section for now.
i.e. tools do not recreate this section or its content, like they do for
symbol table sections, for example. That should allow to write all kind of
possible broken test cases for our needs and keep the output closer to requested.

Differential revision: https://reviews.llvm.org/D65446

llvm-svn: 368272

Author: George Rimar

llvm/include/llvm/Object/ELF.h

@@ -635,8 +635,11 @@ ELFFile<ELFT>::getSHNDXTable(const Elf_Shdr &Section,
   const Elf_Shdr &SymTable = **SymTableOrErr;
   if (SymTable.sh_type != ELF::SHT_SYMTAB &&
       SymTable.sh_type != ELF::SHT_DYNSYM)
-    // TODO: this error is untested.
-    return createError("invalid sh_type");
+    return createError("SHT_SYMTAB_SHNDX section is linked with " +
+                       object::getELFSectionTypeName(getHeader()->e_machine,
+                                                     SymTable.sh_type) +
+                       " section (expected SHT_SYMTAB/SHT_DYNSYM)");
+
   if (V.size() != (SymTable.sh_size / sizeof(Elf_Sym)))
     return createError("SHT_SYMTAB_SHNDX section has sh_size (" +
                        Twine(SymTable.sh_size) +

llvm/include/llvm/ObjectYAML/ELFYAML.h

@@ -128,6 +128,7 @@ struct Section {
     NoBits,
     Verdef,
     Verneed,
+    SymtabShndxSection,
     Symver,
     MipsABIFlags
   };
@@ -274,6 +275,16 @@ struct RelocationSection : Section {
   }
 };
 
+struct SymtabShndxSection : Section {
+  std::vector<uint32_t> Entries;
+
+  SymtabShndxSection() : Section(SectionKind::SymtabShndxSection) {}
+
+  static bool classof(const Section *S) {
+    return S->Kind == SectionKind::SymtabShndxSection;
+  }
+};
+
 // Represents .MIPS.abiflags section
 struct MipsABIFlags : Section {
   llvm::yaml::Hex16 Version;

llvm/lib/ObjectYAML/ELFEmitter.cpp

@@ -140,6 +140,9 @@ template <class ELFT> class ELFState {
                            ContiguousBlobAccumulator &CBA);
   bool writeSectionContent(Elf_Shdr &SHeader, const ELFYAML::Group &Group,
                            ContiguousBlobAccumulator &CBA);
+  bool writeSectionContent(Elf_Shdr &SHeader,
+                           const ELFYAML::SymtabShndxSection &Shndx,
+                           ContiguousBlobAccumulator &CBA);
   bool writeSectionContent(Elf_Shdr &SHeader,
                            const ELFYAML::SymverSection &Section,
                            ContiguousBlobAccumulator &CBA);
@@ -358,6 +361,9 @@ bool ELFState<ELFT>::initSectionHeaders(ELFState<ELFT> &State,
     } else if (auto S = dyn_cast<ELFYAML::RawContentSection>(Sec)) {
       if (!writeSectionContent(SHeader, *S, CBA))
         return false;
+    } else if (auto S = dyn_cast<ELFYAML::SymtabShndxSection>(Sec)) {
+      if (!writeSectionContent(SHeader, *S, CBA))
+        return false;
     } else if (auto S = dyn_cast<ELFYAML::RelocationSection>(Sec)) {
       if (!writeSectionContent(SHeader, *S, CBA))
         return false;
@@ -740,6 +746,21 @@ bool ELFState<ELFT>::writeSectionContent(
   return true;
 }
 
+template <class ELFT>
+bool ELFState<ELFT>::writeSectionContent(
+    Elf_Shdr &SHeader, const ELFYAML::SymtabShndxSection &Shndx,
+    ContiguousBlobAccumulator &CBA) {
+  raw_ostream &OS =
+      CBA.getOSAndAlignedOffset(SHeader.sh_offset, SHeader.sh_addralign);
+
+  for (uint32_t E : Shndx.Entries)
+    support::endian::write<uint32_t>(OS, E, ELFT::TargetEndianness);
+
+  SHeader.sh_entsize = Shndx.EntSize ? (uint64_t)*Shndx.EntSize : 4;
+  SHeader.sh_size = Shndx.Entries.size() * SHeader.sh_entsize;
+  return true;
+}
+
 template <class ELFT>
 bool ELFState<ELFT>::writeSectionContent(Elf_Shdr &SHeader,
                                          const ELFYAML::Group &Section,

llvm/lib/ObjectYAML/ELFYAML.cpp

@@ -563,7 +563,7 @@ void ScalarEnumerationTraits<ELFYAML::ELF_SHN>::enumeration(
   ECase(SHN_HEXAGON_SCOMMON_4);
   ECase(SHN_HEXAGON_SCOMMON_8);
 #undef ECase
-  IO.enumFallback<Hex32>(Value);
+  IO.enumFallback<Hex16>(Value);
 }
 
 void ScalarEnumerationTraits<ELFYAML::ELF_STB>::enumeration(
@@ -897,8 +897,6 @@ StringRef MappingTraits<ELFYAML::Symbol>::validate(IO &IO,
                                                    ELFYAML::Symbol &Symbol) {
   if (Symbol.Index && Symbol.Section.data())
     return "Index and Section cannot both be specified for Symbol";
-  if (Symbol.Index && *Symbol.Index == ELFYAML::ELF_SHN(ELF::SHN_XINDEX))
-    return "Large indexes are not supported";
   if (Symbol.NameIndex && !Symbol.Name.empty())
     return "Name and NameIndex cannot both be specified for Symbol";
   return StringRef();
@@ -969,6 +967,11 @@ static void groupSectionMapping(IO &IO, ELFYAML::Group &Group) {
   IO.mapRequired("Members", Group.Members);
 }
 
+static void sectionMapping(IO &IO, ELFYAML::SymtabShndxSection &Section) {
+  commonSectionMapping(IO, Section);
+  IO.mapRequired("Entries", Section.Entries);
+}
+
 void MappingTraits<ELFYAML::SectionOrType>::mapping(
     IO &IO, ELFYAML::SectionOrType &sectionOrType) {
   IO.mapRequired("SectionOrType", sectionOrType.sectionNameOrType);
@@ -1049,6 +1052,11 @@ void MappingTraits<std::unique_ptr<ELFYAML::Section>>::mapping(
       Section.reset(new ELFYAML::VerneedSection());
     sectionMapping(IO, *cast<ELFYAML::VerneedSection>(Section.get()));
     break;
+  case ELF::SHT_SYMTAB_SHNDX:
+    if (!IO.outputting())
+      Section.reset(new ELFYAML::SymtabShndxSection());
+    sectionMapping(IO, *cast<ELFYAML::SymtabShndxSection>(Section.get()));
+    break;
   default:
     if (!IO.outputting())
       Section.reset(new ELFYAML::RawContentSection());

llvm/test/Object/invalid.test

@@ -207,7 +207,7 @@ Sections:
 # RUN: yaml2obj %s --docnum=11 -o %t11
 # RUN: not llvm-readobj --symbols %t11 2>&1 | FileCheck --check-prefix=INVALID-XINDEX-SIZE %s
 
-# INVALID-XINDEX-SIZE: error: {{.*}}: SHT_SYMTAB_SHNDX section has sh_size (24) which is not equal to the number of symbols (6)
+# INVALID-XINDEX-SIZE: error: {{.*}}: SHT_SYMTAB_SHNDX section has sh_size (24) which is not equal to the number of symbols (2)
 
 --- !ELF
 FileHeader:
@@ -218,8 +218,7 @@ FileHeader:
 Sections:
   - Name: .symtab_shndx
     Type: SHT_SYMTAB_SHNDX
-    Size: 0x18
-    EntSize: 4
+    Entries: [ 0, 1 ]
     Link: .symtab
 
 ## Check that llvm-readobj reports an error if the e_phentsize field is broken.

llvm/test/tools/llvm-readobj/elf-section-types.test

@@ -166,8 +166,7 @@ Sections:
   - Name: symtab_shndx
     Type: SHT_SYMTAB_SHNDX
     Link: .symtab
-    EntSize: 0x4
-    Content: "0000000000000000"
+    Entries: [ 0, 1 ]
   - Name: relr
     Type: SHT_RELR
   - Name: android_rel

llvm/test/tools/obj2yaml/Inputs/shn_xindex.o

@@ -0,0 +1,189 @@
+## Check that obj2yaml is able to dump a normal object which
+## contains the SHT_SYMTAB_SHNDX section and symbols with
+## section index == SHN_XINDEX.
+
+# RUN: yaml2obj --docnum=1 %s -o %t1
+# RUN: obj2yaml %t1 | FileCheck %s --check-prefix=CASE1
+
+# CASE1:      --- !ELF
+# CASE1-NEXT: FileHeader:
+# CASE1-NEXT:   Class:   ELFCLASS64
+# CASE1-NEXT:   Data:    ELFDATA2LSB
+# CASE1-NEXT:   Type:    ET_REL
+# CASE1-NEXT:   Machine: EM_X86_64
+# CASE1-NEXT: Sections:
+# CASE1-NEXT:   - Name:    bar
+# CASE1-NEXT:     Type:    SHT_PROGBITS
+# CASE1-NEXT:   - Name:    .symtab_shndx
+# CASE1-NEXT:     Type:    SHT_SYMTAB_SHNDX
+# CASE1-NEXT:     Link:    .symtab
+# CASE1-NEXT:     EntSize: 0x0000000000000004
+# CASE1-NEXT:     Entries: [ 0, 1, 2 ]
+# CASE1-NEXT: Symbols:
+# CASE1-NEXT:   - Name:  bar
+# CASE1-NEXT:     Type:  STT_SECTION
+# CASE1-NEXT:     Index: SHN_XINDEX
+# CASE1-NEXT:   - Name:  .symtab_shndx
+# CASE1-NEXT:     Type:  STT_SECTION
+# CASE1-NEXT:     Index: SHN_XINDEX
+# CASE1-NEXT: ...
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_REL
+  Machine: EM_X86_64
+Sections:
+  - Name: bar
+    Type: SHT_PROGBITS
+  - Name:    .symtab_shndx
+    Type:    SHT_SYMTAB_SHNDX
+    Entries: [ 0, 1, 2 ]
+    Link:    .symtab
+Symbols:
+  - Type:  STT_SECTION
+    Index: SHN_XINDEX
+  - Type:  STT_SECTION
+    Index: SHN_XINDEX
+
+## Check that yaml2obj is unable to dump an object, which has
+## symbols with section index == SHN_XINDEX, but no SHT_SYMTAB_SHNDX section.
+
+# RUN: yaml2obj --docnum=2 %s -o %t2
+# RUN: not obj2yaml %t2 2>&1 | FileCheck %s -DFILE=%t2 --check-prefix=CASE2
+
+# CASE2: Error reading file: [[FILE]]: extended symbol index (1) is past the end of the SHT_SYMTAB_SHNDX section of size 0
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_REL
+  Machine: EM_X86_64
+Symbols:
+  - Type:  STT_SECTION
+    Index: SHN_XINDEX
+
+## Check that yaml2obj is unable to dump an object, which has symbols with
+## section index == SHN_XINDEX, but SHT_SYMTAB_SHNDX table contains invalid indices
+## that are larger than total number of the sections.
+
+# RUN: yaml2obj --docnum=3 %s -o %t3
+# RUN: not obj2yaml %t3 2>&1 | FileCheck %s -DFILE=%t3 --check-prefix=CASE3
+
+# CASE3: Error reading file: [[FILE]]: invalid section index: 254
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_REL
+  Machine: EM_X86_64
+Sections:
+  - Name: bar
+    Type: SHT_PROGBITS
+  - Name:    .symtab_shndx
+    Type:    SHT_SYMTAB_SHNDX
+    Entries: [ 0, 254 ]
+    Link:    .symtab
+Symbols:
+  - Type:  STT_SECTION
+    Index: SHN_XINDEX
+
+## Check that yaml2obj is unable to dump an object, which has symbols with
+## section index == SHN_XINDEX, but SHT_SYMTAB_SHNDX table contains more
+## entries than the number of symbols in .symtab.
+
+# RUN: yaml2obj --docnum=4 %s -o %t4
+# RUN: not obj2yaml %t4 2>&1 | FileCheck %s -DFILE=%t4 --check-prefix=CASE4
+
+## FIXME: The error message below needs rewording. Size should not be equal to the number of symbols.
+## CASE4: Error reading file: [[FILE]]: SHT_SYMTAB_SHNDX section has sh_size (48) which is not equal to the number of symbols (3)
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_REL
+  Machine: EM_X86_64
+Sections:
+  - Name: bar
+    Type: SHT_PROGBITS
+  - Name:    .symtab_shndx
+    Type:    SHT_SYMTAB_SHNDX
+    Entries: [ 0, 1, 2 ]
+    Link:    .symtab
+Symbols:
+  - Type:  STT_SECTION
+    Index: SHN_XINDEX
+
+## ELF gABI allows having multiple SHT_SYMTAB_SHNDX sections.
+## We only support having one associated with .symtab now.
+
+# RUN: yaml2obj --docnum=5 %s -o %t5
+# RUN: not obj2yaml %t5 2>&1 | FileCheck %s -DFILE=%t5 --check-prefix=CASE5
+
+# CASE5: Error reading file: [[FILE]]: multiple SHT_SYMTAB_SHNDX sections are not supported
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_REL
+  Machine: EM_X86_64
+Sections:
+  - Name:    .symtab_shndx1
+    Type:    SHT_SYMTAB_SHNDX
+    Entries: [ 0 ]
+    EntSize: 4
+    Link:    .symtab
+  - Name:    .symtab_shndx2
+    Type:    SHT_SYMTAB_SHNDX
+    Entries: [ 0 ]
+    Link:    .symtab
+
+## Check that yaml2obj can't dump the object if SHT_SYMTAB_SHNDX is
+## not associated with a SHT_SYMTAB section (this case is illegal).
+
+# RUN: yaml2obj --docnum=6 %s -o %t6
+# RUN: not obj2yaml %t6 2>&1 | FileCheck %s -DFILE=%t6 --check-prefix=CASE6
+
+# CASE6: Error reading file: [[FILE]]: SHT_SYMTAB_SHNDX section is linked with SHT_PROGBITS section (expected SHT_SYMTAB/SHT_DYNSYM)
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_REL
+  Machine: EM_X86_64
+Sections:
+  - Name:    .symtab_shndx
+    Type:    SHT_SYMTAB_SHNDX
+    Entries: [ 0 ]
+    Link:    .foo
+  - Name:    .foo
+    Type:    SHT_PROGBITS
+
+## Check that yaml2obj can't dump the object if SHT_SYMTAB_SHNDX is
+## associated with a SHT_DYNSYM section (not implemented yet).
+
+# RUN: yaml2obj --docnum=7 %s -o %t7
+# RUN: not obj2yaml %t7 2>&1 | FileCheck %s -DFILE=%t7 --check-prefix=CASE7
+
+# CASE7: Error reading file: [[FILE]]: only SHT_SYMTAB_SHNDX associated with SHT_SYMTAB are supported
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_REL
+  Machine: EM_X86_64
+Sections:
+  - Name:    .symtab_shndx
+    Type:    SHT_SYMTAB_SHNDX
+    Entries: [ 0, 1 ]
+    Link:    .dynsym
+DynamicSymbols:
+  - Type:  STT_SECTION
+    Index: SHN_XINDEX

llvm/test/tools/obj2yaml/elf-sht-symtab-shndx.yaml

@@ -18,7 +18,7 @@
 # CHECK-NEXT:     Index:   SHN_HEXAGON_SCOMMON_1
 # CHECK-NEXT:     Binding: STB_GLOBAL
 # CHECK-NEXT:   - Name:    unknown_index
-# CHECK-NEXT:     Index:   0x0000FFFE
+# CHECK-NEXT:     Index:   0xFFFE
 # CHECK-NEXT:     Binding: STB_GLOBAL
 
 !ELF
@@ -47,8 +47,3 @@ Symbols:
   - Name:     unknown_index
     Index:    0xfffe
     Binding:  STB_GLOBAL
-
-## shn_xindex.o contains a symbol with st_shndx == SHN_XINDEX.
-## We do not support it at this moment.
-# RUN: not obj2yaml %S/Inputs/shn_xindex.o 2>&1 | FileCheck %s --check-prefix=ERR
-# ERR: Error reading file: {{.*}}shn_xindex.o: SHN_XINDEX symbols are not supported