[Object, MachO] Introduce MachOObjectFile::load_commands() range iterator.

Summary:
Now users don't have to manually deal with getFirstLoadCommandInfo() /
getNextLoadCommandInfo(), calculate the number of load segments, etc.

No functionality change.

Test Plan: regression test suite

Reviewers: rafael, lhames, loladiro

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D10144

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@238983 91177308-0d34-0410-b5e6-96231b3b80d8

include/llvm/Object/MachO.h

@@ -190,6 +190,8 @@ public:
     const char *Ptr;      // Where in memory the load command is.
     MachO::load_command C; // The command itself.
   };
+  typedef SmallVector<LoadCommandInfo, 4> LoadCommandList;
+  typedef LoadCommandList::const_iterator load_command_iterator;
 
   MachOObjectFile(MemoryBufferRef Object, bool IsLittleEndian, bool Is64Bits,
                   std::error_code &EC);
@@ -270,10 +272,14 @@ public:
 
   dice_iterator begin_dices() const;
   dice_iterator end_dices() const;
-  
+
+  load_command_iterator begin_load_commands() const;
+  load_command_iterator end_load_commands() const;
+  iterator_range<load_command_iterator> load_commands() const;
+
   /// For use iterating over all exported symbols.
   iterator_range<export_iterator> exports() const;
-  
+
   /// For use examining a trie not in a MachOObjectFile.
   static iterator_range<export_iterator> exports(ArrayRef<uint8_t> Trie);
 
@@ -326,10 +332,6 @@ public:
   unsigned getAnyRelocationType(const MachO::any_relocation_info &RE) const;
   SectionRef getAnyRelocationSection(const MachO::any_relocation_info &RE) const;
 
-  // Walk load commands.
-  LoadCommandInfo getFirstLoadCommandInfo() const;
-  LoadCommandInfo getNextLoadCommandInfo(const LoadCommandInfo &L) const;
-
   // MachO specific structures.
   MachO::section getSection(DataRefImpl DRI) const;
   MachO::section_64 getSection64(DataRefImpl DRI) const;
@@ -427,10 +429,15 @@ public:
   }
 
 private:
+  // Walk load commands.
+  LoadCommandInfo getFirstLoadCommandInfo() const;
+  LoadCommandInfo getNextLoadCommandInfo(const LoadCommandInfo &L) const;
+
   typedef SmallVector<const char*, 1> SectionList;
   SectionList Sections;
   typedef SmallVector<const char*, 1> LibraryList;
   LibraryList Libraries;
+  LoadCommandList LoadCommands;
   typedef SmallVector<StringRef, 1> LibraryShortName;
   mutable LibraryShortName LibrariesShortNames;
   const char *SymtabLoadCmd;

lib/Object/MachOObjectFile.cpp

@@ -194,8 +194,9 @@ MachOObjectFile::MachOObjectFile(MemoryBufferRef Object, bool IsLittleEndian,
   MachO::LoadCommandType SegmentLoadType = is64Bit() ?
     MachO::LC_SEGMENT_64 : MachO::LC_SEGMENT;
 
-  MachOObjectFile::LoadCommandInfo Load = getFirstLoadCommandInfo();
+  LoadCommandInfo Load = getFirstLoadCommandInfo();
-  for (unsigned I = 0; ; ++I) {
+  for (unsigned I = 0; I < LoadCommandCount; ++I) {
+    LoadCommands.push_back(Load);
     if (Load.C.cmd == MachO::LC_SYMTAB) {
       // Multiple symbol tables
       if (SymtabLoadCmd) {
@@ -260,12 +261,10 @@ MachOObjectFile::MachOObjectFile(MemoryBufferRef Object, bool IsLittleEndian,
                Load.C.cmd == MachO::LC_LOAD_UPWARD_DYLIB) {
       Libraries.push_back(Load.Ptr);
     }
-
+    if (I < LoadCommandCount - 1)
-    if (I == LoadCommandCount - 1)
-      break;
-    else
       Load = getNextLoadCommandInfo(Load);
   }
+  assert(LoadCommands.size() == LoadCommandCount);
 }
 
 void MachOObjectFile::moveSymbolNext(DataRefImpl &Symb) const {
@@ -1863,6 +1862,22 @@ iterator_range<bind_iterator> MachOObjectFile::weakBindTable() const {
                    MachOBindEntry::Kind::Weak);
 }
 
+MachOObjectFile::load_command_iterator
+MachOObjectFile::begin_load_commands() const {
+  return LoadCommands.begin();
+}
+
+MachOObjectFile::load_command_iterator
+MachOObjectFile::end_load_commands() const {
+  return LoadCommands.end();
+}
+
+iterator_range<MachOObjectFile::load_command_iterator>
+MachOObjectFile::load_commands() const {
+  return iterator_range<load_command_iterator>(begin_load_commands(),
+                                               end_load_commands());
+}
+
 StringRef
 MachOObjectFile::getSectionFinalSegmentName(DataRefImpl Sec) const {
   ArrayRef<char> Raw = getSectionRawFinalSegmentName(Sec);

tools/llvm-objdump/MachODump.cpp

@@ -281,8 +281,7 @@ static uint64_t DumpDataInCode(const uint8_t *bytes, uint64_t Length,
   return Size;
 }
 
-static void getSectionsAndSymbols(const MachO::mach_header Header,
+static void getSectionsAndSymbols(MachOObjectFile *MachOObj,
-                                  MachOObjectFile *MachOObj,
                                   std::vector<SectionRef> &Sections,
                                   std::vector<SymbolRef> &Symbols,
                                   SmallVectorImpl<uint64_t> &FoundFns,
@@ -300,10 +299,8 @@ static void getSectionsAndSymbols(const MachO::mach_header Header,
     Sections.push_back(Section);
   }
 
-  MachOObjectFile::LoadCommandInfo Command =
-      MachOObj->getFirstLoadCommandInfo();
   bool BaseSegmentAddressSet = false;
-  for (unsigned i = 0;; ++i) {
+  for (const auto &Command : MachOObj->load_commands()) {
     if (Command.C.cmd == MachO::LC_FUNCTION_STARTS) {
       // We found a function starts segment, parse the addresses for later
       // consumption.
@@ -319,11 +316,6 @@ static void getSectionsAndSymbols(const MachO::mach_header Header,
         BaseSegmentAddress = SLC.vmaddr;
       }
     }
-
-    if (i == Header.ncmds - 1)
-      break;
-    else
-      Command = MachOObj->getNextLoadCommandInfo(Command);
   }
 }
 
@@ -386,9 +378,7 @@ static void PrintIndirectSymbolTable(MachOObjectFile *O, bool verbose,
 }
 
 static void PrintIndirectSymbols(MachOObjectFile *O, bool verbose) {
-  uint32_t LoadCommandCount = O->getHeader().ncmds;
+  for (const auto &Load : O->load_commands()) {
-  MachOObjectFile::LoadCommandInfo Load = O->getFirstLoadCommandInfo();
-  for (unsigned I = 0;; ++I) {
     if (Load.C.cmd == MachO::LC_SEGMENT_64) {
       MachO::segment_command_64 Seg = O->getSegment64LoadCommand(Load);
       for (unsigned J = 0; J < Seg.nsects; ++J) {
@@ -446,10 +436,6 @@ static void PrintIndirectSymbols(MachOObjectFile *O, bool verbose) {
         }
       }
     }
-    if (I == LoadCommandCount - 1)
-      break;
-    else
-      Load = O->getNextLoadCommandInfo(Load);
   }
 }
 
@@ -553,9 +539,8 @@ static void PrintLinkOptHints(MachOObjectFile *O) {
 }
 
 static void PrintDylibs(MachOObjectFile *O, bool JustId) {
-  uint32_t LoadCommandCount = O->getHeader().ncmds;
+  unsigned Index = 0;
-  MachOObjectFile::LoadCommandInfo Load = O->getFirstLoadCommandInfo();
+  for (const auto &Load : O->load_commands()) {
-  for (unsigned I = 0;; ++I) {
     if ((JustId && Load.C.cmd == MachO::LC_ID_DYLIB) ||
         (!JustId && (Load.C.cmd == MachO::LC_ID_DYLIB ||
                      Load.C.cmd == MachO::LC_LOAD_DYLIB ||
@@ -595,13 +580,9 @@ static void PrintDylibs(MachOObjectFile *O, bool JustId) {
           outs() << "LC_LOAD_UPWARD_DYLIB ";
         else
           outs() << "LC_??? ";
-        outs() << "command " << I << "\n";
+        outs() << "command " << Index++ << "\n";
       }
     }
-    if (I == LoadCommandCount - 1)
-      break;
-    else
-      Load = O->getNextLoadCommandInfo(Load);
   }
 }
 
@@ -2132,9 +2113,7 @@ static int SymbolizerGetOpInfo(void *DisInfo, uint64_t Pc, uint64_t Offset,
 // it returns a pointer to that string.  Else it returns nullptr.
 static const char *GuessCstringPointer(uint64_t ReferenceValue,
                                        struct DisassembleInfo *info) {
-  uint32_t LoadCommandCount = info->O->getHeader().ncmds;
+  for (const auto &Load : info->O->load_commands()) {
-  MachOObjectFile::LoadCommandInfo Load = info->O->getFirstLoadCommandInfo();
-  for (unsigned I = 0;; ++I) {
     if (Load.C.cmd == MachO::LC_SEGMENT_64) {
       MachO::segment_command_64 Seg = info->O->getSegment64LoadCommand(Load);
       for (unsigned J = 0; J < Seg.nsects; ++J) {
@@ -2178,10 +2157,6 @@ static const char *GuessCstringPointer(uint64_t ReferenceValue,
         }
       }
     }
-    if (I == LoadCommandCount - 1)
-      break;
-    else
-      Load = info->O->getNextLoadCommandInfo(Load);
   }
   return nullptr;
 }
@@ -2192,11 +2167,9 @@ static const char *GuessCstringPointer(uint64_t ReferenceValue,
 // symbol name being referenced by the stub or pointer.
 static const char *GuessIndirectSymbol(uint64_t ReferenceValue,
                                        struct DisassembleInfo *info) {
-  uint32_t LoadCommandCount = info->O->getHeader().ncmds;
-  MachOObjectFile::LoadCommandInfo Load = info->O->getFirstLoadCommandInfo();
   MachO::dysymtab_command Dysymtab = info->O->getDysymtabLoadCommand();
   MachO::symtab_command Symtab = info->O->getSymtabLoadCommand();
-  for (unsigned I = 0;; ++I) {
+  for (const auto &Load : info->O->load_commands()) {
     if (Load.C.cmd == MachO::LC_SEGMENT_64) {
       MachO::segment_command_64 Seg = info->O->getSegment64LoadCommand(Load);
       for (unsigned J = 0; J < Seg.nsects; ++J) {
@@ -2266,10 +2239,6 @@ static const char *GuessIndirectSymbol(uint64_t ReferenceValue,
         }
       }
     }
-    if (I == LoadCommandCount - 1)
-      break;
-    else
-      Load = info->O->getNextLoadCommandInfo(Load);
   }
   return nullptr;
 }
@@ -2356,9 +2325,7 @@ static uint64_t GuessPointerPointer(uint64_t ReferenceValue,
   selref = false;
   msgref = false;
   cfstring = false;
-  uint32_t LoadCommandCount = info->O->getHeader().ncmds;
+  for (const auto &Load : info->O->load_commands()) {
-  MachOObjectFile::LoadCommandInfo Load = info->O->getFirstLoadCommandInfo();
-  for (unsigned I = 0;; ++I) {
     if (Load.C.cmd == MachO::LC_SEGMENT_64) {
       MachO::segment_command_64 Seg = info->O->getSegment64LoadCommand(Load);
       for (unsigned J = 0; J < Seg.nsects; ++J) {
@@ -2403,10 +2370,6 @@ static uint64_t GuessPointerPointer(uint64_t ReferenceValue,
       }
     }
     // TODO: Look for LC_SEGMENT for 32-bit Mach-O files.
-    if (I == LoadCommandCount - 1)
-      break;
-    else
-      Load = info->O->getNextLoadCommandInfo(Load);
   }
   return 0;
 }
@@ -6075,7 +6038,7 @@ static void DisassembleMachO(StringRef Filename, MachOObjectFile *MachOOF,
   SmallVector<uint64_t, 8> FoundFns;
   uint64_t BaseSegmentAddress;
 
-  getSectionsAndSymbols(Header, MachOOF, Sections, Symbols, FoundFns,
+  getSectionsAndSymbols(MachOOF, Sections, Symbols, FoundFns,
                         BaseSegmentAddress);
 
   // Sort the symbols by address, just in case they didn't come in that way.
@@ -8367,15 +8330,12 @@ static void PrintLinkEditDataCommand(MachO::linkedit_data_command ld,
     outs() << "\n";
 }
 
-static void PrintLoadCommands(const MachOObjectFile *Obj, uint32_t ncmds,
+static void PrintLoadCommands(const MachOObjectFile *Obj, uint32_t filetype,
-                              uint32_t filetype, uint32_t cputype,
+                              uint32_t cputype, bool verbose) {
-                              bool verbose) {
-  if (ncmds == 0)
-    return;
   StringRef Buf = Obj->getData();
-  MachOObjectFile::LoadCommandInfo Command = Obj->getFirstLoadCommandInfo();
+  unsigned Index = 0;
-  for (unsigned i = 0;; ++i) {
+  for (const auto &Command : Obj->load_commands()) {
-    outs() << "Load command " << i << "\n";
+    outs() << "Load command " << Index++ << "\n";
     if (Command.C.cmd == MachO::LC_SEGMENT) {
       MachO::segment_command SLC = Obj->getSegmentLoadCommand(Command);
       const char *sg_segname = SLC.segname;
@@ -8494,14 +8454,10 @@ static void PrintLoadCommands(const MachOObjectFile *Obj, uint32_t ncmds,
       // TODO: get and print the raw bytes of the load command.
     }
     // TODO: print all the other kinds of load commands.
-    if (i == ncmds - 1)
-      break;
-    else
-      Command = Obj->getNextLoadCommandInfo(Command);
   }
 }
 
-static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds,
+static void getAndPrintMachHeader(const MachOObjectFile *Obj,
                                   uint32_t &filetype, uint32_t &cputype,
                                   bool verbose) {
   if (Obj->is64Bit()) {
@@ -8509,7 +8465,6 @@ static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds,
     H_64 = Obj->getHeader64();
     PrintMachHeader(H_64.magic, H_64.cputype, H_64.cpusubtype, H_64.filetype,
                     H_64.ncmds, H_64.sizeofcmds, H_64.flags, verbose);
-    ncmds = H_64.ncmds;
     filetype = H_64.filetype;
     cputype = H_64.cputype;
   } else {
@@ -8517,7 +8472,6 @@ static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds,
     H = Obj->getHeader();
     PrintMachHeader(H.magic, H.cputype, H.cpusubtype, H.filetype, H.ncmds,
                     H.sizeofcmds, H.flags, verbose);
-    ncmds = H.ncmds;
     filetype = H.filetype;
     cputype = H.cputype;
   }
@@ -8525,11 +8479,10 @@ static void getAndPrintMachHeader(const MachOObjectFile *Obj, uint32_t &ncmds,
 
 void llvm::printMachOFileHeader(const object::ObjectFile *Obj) {
   const MachOObjectFile *file = dyn_cast<const MachOObjectFile>(Obj);
-  uint32_t ncmds = 0;
   uint32_t filetype = 0;
   uint32_t cputype = 0;
-  getAndPrintMachHeader(file, ncmds, filetype, cputype, !NonVerbose);
+  getAndPrintMachHeader(file, filetype, cputype, !NonVerbose);
-  PrintLoadCommands(file, ncmds, filetype, cputype, !NonVerbose);
+  PrintLoadCommands(file, filetype, cputype, !NonVerbose);
 }
 
 //===----------------------------------------------------------------------===//

tools/llvm-size/llvm-size.cpp

@@ -122,12 +122,10 @@ static void PrintDarwinSectionSizes(MachOObjectFile *MachO) {
     fmt << "0x";
   fmt << "%" << radix_fmt;
 
-  uint32_t LoadCommandCount = MachO->getHeader().ncmds;
   uint32_t Filetype = MachO->getHeader().filetype;
-  MachOObjectFile::LoadCommandInfo Load = MachO->getFirstLoadCommandInfo();
 
   uint64_t total = 0;
-  for (unsigned I = 0;; ++I) {
+  for (const auto &Load : MachO->load_commands()) {
     if (Load.C.cmd == MachO::LC_SEGMENT_64) {
       MachO::segment_command_64 Seg = MachO->getSegment64LoadCommand(Load);
       outs() << "Segment " << Seg.segname << ": "
@@ -181,10 +179,6 @@ static void PrintDarwinSectionSizes(MachOObjectFile *MachO) {
       if (Seg.nsects != 0)
         outs() << "\ttotal " << format(fmt.str().c_str(), sec_total) << "\n";
     }
-    if (I == LoadCommandCount - 1)
-      break;
-    else
-      Load = MachO->getNextLoadCommandInfo(Load);
   }
   outs() << "total " << format(fmt.str().c_str(), total) << "\n";
 }
@@ -194,14 +188,11 @@ static void PrintDarwinSectionSizes(MachOObjectFile *MachO) {
 /// This is when used when @c OutputFormat is berkeley with a Mach-O file and
 /// produces the same output as darwin's size(1) default output.
 static void PrintDarwinSegmentSizes(MachOObjectFile *MachO) {
-  uint32_t LoadCommandCount = MachO->getHeader().ncmds;
-  MachOObjectFile::LoadCommandInfo Load = MachO->getFirstLoadCommandInfo();
-
   uint64_t total_text = 0;
   uint64_t total_data = 0;
   uint64_t total_objc = 0;
   uint64_t total_others = 0;
-  for (unsigned I = 0;; ++I) {
+  for (const auto &Load : MachO->load_commands()) {
     if (Load.C.cmd == MachO::LC_SEGMENT_64) {
       MachO::segment_command_64 Seg = MachO->getSegment64LoadCommand(Load);
       if (MachO->getHeader().filetype == MachO::MH_OBJECT) {
@@ -255,10 +246,6 @@ static void PrintDarwinSegmentSizes(MachOObjectFile *MachO) {
           total_others += Seg.vmsize;
       }
     }
-    if (I == LoadCommandCount - 1)
-      break;
-    else
-      Load = MachO->getNextLoadCommandInfo(Load);
   }
   uint64_t total = total_text + total_data + total_objc + total_others;
 

tools/macho-dump/macho-dump.cpp

@@ -340,7 +340,7 @@ DumpDylibID(const MachOObjectFile &Obj,
 }
 
 static int DumpLoadCommand(const MachOObjectFile &Obj,
-                           MachOObjectFile::LoadCommandInfo &LCI) {
+                           const MachOObjectFile::LoadCommandInfo &LCI) {
   switch (LCI.C.cmd) {
   case MachO::LC_SEGMENT:
     return DumpSegmentCommand(Obj, LCI);
@@ -369,9 +369,8 @@ static int DumpLoadCommand(const MachOObjectFile &Obj,
   }
 }
 
-
 static int DumpLoadCommand(const MachOObjectFile &Obj, unsigned Index,
-                           MachOObjectFile::LoadCommandInfo &LCI) {
+                           const MachOObjectFile::LoadCommandInfo &LCI) {
   outs() << "  # Load Command " << Index << "\n"
          << " (('command', " << LCI.C.cmd << ")\n"
          << "  ('size', " << LCI.C.cmdsize << ")\n";
@@ -423,16 +422,11 @@ int main(int argc, char **argv) {
 
   // Print the load commands.
   int Res = 0;
-  MachOObjectFile::LoadCommandInfo Command =
+  unsigned Index = 0;
-    InputObject->getFirstLoadCommandInfo();
   outs() << "('load_commands', [\n";
-  for (unsigned i = 0; ; ++i) {
+  for (const auto &Load : InputObject->load_commands()) {
-    if (DumpLoadCommand(*InputObject, i, Command))
+    if (DumpLoadCommand(*InputObject, Index++, Load))
       break;
-
-    if (i == Header->ncmds - 1)
-      break;
-    Command = InputObject->getNextLoadCommandInfo(Command);
   }
   outs() << "])\n";