[mach-o] Extend LC_DATA_IN_CODE support to x86_64

[andrewrk]

LGTM

[LemonBoy]

Friendly ping after ~1week

lld/lib/ReaderWriter/MachO/ArchHandler.cpp

 int32_t ArchHandler::readS32(const uint8_t *addr, bool isBig) {
   return read32(addr, isBig);
 }
 
 uint32_t ArchHandler::readU32(const uint8_t *addr, bool isBig) {
   return read32(addr, isBig);
 }
 
-  int64_t ArchHandler::readS64(const uint8_t *addr, bool isBig) {
   return read64(addr, isBig);
 }
 
 bool ArchHandler::isDwarfCIE(bool isBig, const DefinedAtom *atom) {
   assert(atom->contentType() == DefinedAtom::typeCFI);
   if (atom->rawContent().size() < sizeof(uint32_t))
     return false;
   uint32_t size = read32(atom->rawContent().data(), isBig);

lld/lib/ReaderWriter/MachO/ArchHandler_x86_64.cpp

public:
   void appendSectionRelocations(const DefinedAtom &atom,
                                 uint64_t atomSectionOffset,
                                 const Reference &ref,
                                 FindSymbolIndexForAtom symbolIndexForAtom,
                                 FindSectionIndexForAtom sectionIndexForAtom,
                                 FindAddressForAtom addressForAtom,
                                 normalized::Relocations &relocs) override;
 
+  bool isDataInCodeTransition(Reference::KindValue refKind) override {
+    return refKind == modeCode || refKind == modeData;
+  }
+
+  Reference::KindValue dataInCodeTransitionStart(
+                                        const MachODefinedAtom &atom) override {
+    return modeData;
+  }
+
+  Reference::KindValue dataInCodeTransitionEnd(
+                                        const MachODefinedAtom &atom) override {
+    return modeCode;
+  }
+
 private:
   static const Registry::KindStrings _sKindStrings[];
   static const StubInfo              _sStubInfo;
 
   enum X86_64Kind: Reference::KindValue {
     invalid,               /// for error condition
 
+    modeCode,              /// Content starting at this offset is code.
+    modeData,              /// Content starting at this offset is data.
+
     // Kinds found in mach-o .o files:
     branch32,              /// ex: call _foo
     ripRel32,              /// ex: movq _foo(%rip), %rax
     ripRel32Minus1,        /// ex: movb $0x12, _foo(%rip)
     ripRel32Minus2,        /// ex: movw $0x1234, _foo(%rip)
     ripRel32Minus4,        /// ex: movl $0x12345678, _foo(%rip)
     ripRel32Anon,          /// ex: movq L1(%rip), %rax
     ripRel32Minus1Anon,    /// ex: movb $0x12, L1(%rip)
private:
 
   void applyFixupRelocatable(const Reference &ref, uint8_t *location,
                              uint64_t fixupAddress,
                              uint64_t targetAddress,
                              uint64_t inAtomAddress);
 };
 
 const Registry::KindStrings ArchHandler_x86_64::_sKindStrings[] = {
-  LLD_KIND_STRING_ENTRY(invalid), LLD_KIND_STRING_ENTRY(branch32),
-  LLD_KIND_STRING_ENTRY(ripRel32), LLD_KIND_STRING_ENTRY(ripRel32Minus1),
-  LLD_KIND_STRING_ENTRY(ripRel32Minus2), LLD_KIND_STRING_ENTRY(ripRel32Minus4),
+  LLD_KIND_STRING_ENTRY(invalid),
+  LLD_KIND_STRING_ENTRY(modeCode),
+  LLD_KIND_STRING_ENTRY(modeData),
+  LLD_KIND_STRING_ENTRY(branch32),
+  LLD_KIND_STRING_ENTRY(ripRel32),
+  LLD_KIND_STRING_ENTRY(ripRel32Minus1),
+  LLD_KIND_STRING_ENTRY(ripRel32Minus2),
+  LLD_KIND_STRING_ENTRY(ripRel32Minus4),
   LLD_KIND_STRING_ENTRY(ripRel32Anon),
   LLD_KIND_STRING_ENTRY(ripRel32Minus1Anon),
   LLD_KIND_STRING_ENTRY(ripRel32Minus2Anon),
   LLD_KIND_STRING_ENTRY(ripRel32Minus4Anon),
   LLD_KIND_STRING_ENTRY(ripRel32GotLoad),
   LLD_KIND_STRING_ENTRY(ripRel32GotLoadNowLea),
-  LLD_KIND_STRING_ENTRY(ripRel32Got), LLD_KIND_STRING_ENTRY(ripRel32Tlv),
+  LLD_KIND_STRING_ENTRY(ripRel32Got),
+  LLD_KIND_STRING_ENTRY(ripRel32Tlv),
   LLD_KIND_STRING_ENTRY(lazyPointer),
   LLD_KIND_STRING_ENTRY(lazyImmediateLocation),
-  LLD_KIND_STRING_ENTRY(pointer64), LLD_KIND_STRING_ENTRY(pointer64Anon),
-  LLD_KIND_STRING_ENTRY(delta32), LLD_KIND_STRING_ENTRY(delta64),
-  LLD_KIND_STRING_ENTRY(delta32Anon), LLD_KIND_STRING_ENTRY(delta64Anon),
+  LLD_KIND_STRING_ENTRY(pointer64),
+  LLD_KIND_STRING_ENTRY(pointer64Anon),
+  LLD_KIND_STRING_ENTRY(delta32),
+  LLD_KIND_STRING_ENTRY(delta64),
+  LLD_KIND_STRING_ENTRY(delta32Anon),
+  LLD_KIND_STRING_ENTRY(delta64Anon),
   LLD_KIND_STRING_ENTRY(negDelta64),
   LLD_KIND_STRING_ENTRY(negDelta32),
-  LLD_KIND_STRING_ENTRY(imageOffset), LLD_KIND_STRING_ENTRY(imageOffsetGot),
+  LLD_KIND_STRING_ENTRY(imageOffset),
+  LLD_KIND_STRING_ENTRY(imageOffsetGot),
   LLD_KIND_STRING_ENTRY(unwindFDEToFunction),
   LLD_KIND_STRING_ENTRY(unwindInfoToEhFrame),
   LLD_KIND_STRING_ENTRY(tlvInitSectionOffset),
   LLD_KIND_STRING_END
 };
 
 const ArchHandler::StubInfo ArchHandler_x86_64::_sStubInfo = {
   "dyld_stub_binder",
case ripRel32GotLoadNowLea:
     *loc32 = targetAddress - (fixupAddress + 4) + ref.addend();
     return;
   case negDelta64:
     *loc64 = fixupAddress - targetAddress + ref.addend();
     return;
   case negDelta32:
     *loc32 = fixupAddress - targetAddress + ref.addend();
     return;
+  case modeCode:
+  case modeData:
   case lazyPointer:
     // Do nothing
     return;
   case lazyImmediateLocation:
     *loc32 = ref.addend();
     return;
   case imageOffset:
   case imageOffsetGot:
void ArchHandler_x86_64::applyFixupRelocatable(const Reference &ref,
   case lazyImmediateLocation:
     llvm_unreachable("lazy reference kind implies Stubs pass was run");
     return;
   case imageOffset:
   case imageOffsetGot:
   case unwindInfoToEhFrame:
     llvm_unreachable("fixup implies __unwind_info");
     return;
+  case modeCode:
+  case modeData:
   case unwindFDEToFunction:
     // Do nothing for now
     return;
   case invalid:
     // Fall into llvm_unreachable().
     break;
   }
   llvm_unreachable("unknown x86_64 Reference Kind");
 }
 
 void ArchHandler_x86_64::appendSectionRelocations(
                                    const DefinedAtom &atom,
                                    uint64_t atomSectionOffset,
                                    const Reference &ref,
                                    FindSymbolIndexForAtom symbolIndexForAtom,
                                    FindSectionIndexForAtom sectionIndexForAtom,
                                    FindAddressForAtom addressForAtom,
                                    normalized::Relocations &relocs) {
   if (ref.kindNamespace() != Reference::KindNamespace::mach_o)
     return;
   assert(ref.kindArch() == Reference::KindArch::x86_64);
   uint32_t sectionOffset = atomSectionOffset + ref.offsetInAtom();
   switch (static_cast<X86_64Kind>(ref.kindValue())) {
+  case modeCode:
+  case modeData:
+    return;
   case branch32:
     appendReloc(relocs, sectionOffset, symbolIndexForAtom(*ref.target()), 0,
                 X86_64_RELOC_BRANCH | rPcRel | rExtern | rLength4);
     return;
   case ripRel32:
     appendReloc(relocs, sectionOffset, symbolIndexForAtom(*ref.target()), 0,
                 X86_64_RELOC_SIGNED | rPcRel | rExtern | rLength4 );
     return;

lld/test/mach-o/parse-data-in-code-x86_64.yaml

+# RUN: ld64.lld -arch x86_64 -r -print_atoms %s -o %t  | FileCheck %s \
+# RUN: && ld64.lld -arch x86_64 -r -print_atoms %t -o %t2  | FileCheck %s
+#
+# Test parsing LC_DATA_IN_CODE
+#
+#
+
+--- !mach-o
+arch:            x86_64
+file-type:       MH_OBJECT
+flags:           [ MH_SUBSECTIONS_VIA_SYMBOLS ]
+sections:
+  - segment:         __TEXT
+    section:         __text
+    type:            S_REGULAR
+    attributes:      [ S_ATTR_PURE_INSTRUCTIONS, S_ATTR_SOME_INSTRUCTIONS ]
+    address:         0x0000000000000000
+    content:         [ 0x90, 0x90, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00,
+                       0x00, 0x00, 0x90, 0x90, 0x90, 0x90, 0x03, 0x00,
+                       0x00, 0x00 ]
+local-symbols:
+  - name:            _func1
+    type:            N_SECT
+    sect:            1
+    value:           0x0000000000000000
+  - name:            _func2
+    type:            N_SECT
+    sect:            1
+    value:           0x000000000000000B
+dataInCode:
+  - offset:          0x00000002
+    length:          0x0008
+    kind:            DICE_KIND_JUMP_TABLE32
+  - offset:          0x0000000E
+    length:          0x0004
+    kind:            DICE_KIND_JUMP_TABLE32
+...
+
+
+
+# CHECK: defined-atoms:
+# CHECK:   - name:            _func1
+# CHECK:     references:
+# CHECK:       - kind:            modeData
+# CHECK:         offset:          2
+# CHECK:         addend:          4
+# CHECK:       - kind:            modeCode
+# CHECK:         offset:          10
+# CHECK:   - name:            _func2
+# CHECK:     references:
+# CHECK:       - kind:            modeData
+# CHECK:         offset:          3
+# CHECK:         addend:          4
+# CHECK-NOT:   - kind:            modeData
+
+
+
+
+#
+#_func1:
+#  nop
+#  nop
+#  .data_region jt32
+#  .long 1
+#  .long 2
+#  .end_data_region
+#  nop
+#
+#
+# _func2:
+#  nop
+#  nop
+#  nop
+#  .data_region jt32
+#  .long 3
+#  .end_data_region
+#