[llvm] Teach LLVM about filesets

[JDevlieghere]

Summary

Teach LLVM about filesets. Filesets were added in macOS 11 (Big Sur) to combine multiple Mach-O files. They consist of a load command (LC_FILESET_ENTRY) consisting of a fileset_entry_command.

struct fileset_entry_command {
    uint32_t     cmd;        /* LC_FILESET_ENTRY */
    uint32_t     cmdsize;    /* includes entry_id string */
    uint64_t     vmaddr;     /* memory address of the entry */
    uint64_t     fileoff;    /* file offset of the entry */
    union lc_str entry_id;   /* contained entry id */
    uint32_t     reserved;   /* reserved */
};

[1] https://github.com/apple-oss-distributions/xnu/blob/e7776783b89a353188416a9a346c6cdb4928faad/EXTERNAL_HEADERS/mach-o/loader.h#L326

[jasonmolenda]

Looks good to me.

llvm/include/llvm/BinaryFormat/MachO.h

enum HeaderFileType {
   MH_FVMLIB = 0x3u,
   MH_CORE = 0x4u,
   MH_PRELOAD = 0x5u,
   MH_DYLIB = 0x6u,
   MH_DYLINKER = 0x7u,
   MH_BUNDLE = 0x8u,
   MH_DYLIB_STUB = 0x9u,
   MH_DSYM = 0xAu,
-  MH_KEXT_BUNDLE = 0xBu
+  MH_KEXT_BUNDLE = 0xBu,
+  MH_FILESET = 0xCu,
 };
 
 enum {
   // Constant bits for the "flags" field in llvm::MachO::mach_header and
   // llvm::MachO::mach_header_64
   MH_NOUNDEFS = 0x00000001u,
   MH_INCRLINK = 0x00000002u,
   MH_DYLDLINK = 0x00000004u,

 };
 
 struct linker_option_command {
   uint32_t cmd;
   uint32_t cmdsize;
   uint32_t count;
 };
 
+struct fileset_entry_command {
+  uint32_t cmd;
+  uint32_t cmdsize;
+  uint64_t vmaddr;
+  uint64_t fileoff;
+  uint32_t entry_id;
+};
+
 // The symseg_command is obsolete and no longer supported.
 struct symseg_command {
   uint32_t cmd;
   uint32_t cmdsize;
   uint32_t offset;
   uint32_t size;
 };
 

 }
 
 inline void swapStruct(linker_option_command &C) {
   sys::swapByteOrder(C.cmd);
   sys::swapByteOrder(C.cmdsize);
   sys::swapByteOrder(C.count);
 }
 
+inline void swapStruct(fileset_entry_command &C) {
+  sys::swapByteOrder(C.cmd);
+  sys::swapByteOrder(C.cmdsize);
+  sys::swapByteOrder(C.vmaddr);
+  sys::swapByteOrder(C.fileoff);
+  sys::swapByteOrder(C.entry_id);
+}
+
 inline void swapStruct(version_min_command &C) {
   sys::swapByteOrder(C.cmd);
   sys::swapByteOrder(C.cmdsize);
   sys::swapByteOrder(C.version);
   sys::swapByteOrder(C.sdk);
 }
 
 inline void swapStruct(note_command &C) {

llvm/include/llvm/BinaryFormat/MachO.def

 HANDLE_LOAD_COMMAND(LC_LINKER_OPTION, 0x0000002Du, linker_option_command)
 HANDLE_LOAD_COMMAND(LC_LINKER_OPTIMIZATION_HINT, 0x0000002Eu, linkedit_data_command)
 HANDLE_LOAD_COMMAND(LC_VERSION_MIN_TVOS, 0x0000002Fu, version_min_command)
 HANDLE_LOAD_COMMAND(LC_VERSION_MIN_WATCHOS, 0x00000030u, version_min_command)
 HANDLE_LOAD_COMMAND(LC_NOTE, 0x00000031u, note_command)
 HANDLE_LOAD_COMMAND(LC_BUILD_VERSION, 0x00000032u, build_version_command)
 HANDLE_LOAD_COMMAND(LC_DYLD_EXPORTS_TRIE, 0x80000033u, linkedit_data_command)
 HANDLE_LOAD_COMMAND(LC_DYLD_CHAINED_FIXUPS, 0x80000034u, linkedit_data_command)
+HANDLE_LOAD_COMMAND(LC_FILESET_ENTRY, 0x80000035u, fileset_entry_command)
 
 #endif
 
 #ifdef LOAD_COMMAND_STRUCT
 
 LOAD_COMMAND_STRUCT(dyld_info_command)
 LOAD_COMMAND_STRUCT(dylib_command)
 LOAD_COMMAND_STRUCT(dylinker_command)

 LOAD_COMMAND_STRUCT(symseg_command)
 LOAD_COMMAND_STRUCT(symtab_command)
 LOAD_COMMAND_STRUCT(thread_command)
 LOAD_COMMAND_STRUCT(twolevel_hints_command)
 LOAD_COMMAND_STRUCT(uuid_command)
 LOAD_COMMAND_STRUCT(version_min_command)
 LOAD_COMMAND_STRUCT(note_command)
 LOAD_COMMAND_STRUCT(build_version_command)
+LOAD_COMMAND_STRUCT(fileset_entry_command)
 
 #endif
 
 #undef HANDLE_LOAD_COMMAND
 #undef LOAD_COMMAND_STRUCT

llvm/lib/ObjectYAML/MachOYAML.cpp

 void MappingTraits<MachO::build_version_command>::mapping(
     IO &IO, MachO::build_version_command &LoadCommand) {
   IO.mapRequired("platform", LoadCommand.platform);
   IO.mapRequired("minos", LoadCommand.minos);
   IO.mapRequired("sdk", LoadCommand.sdk);
   IO.mapRequired("ntools", LoadCommand.ntools);
 }
 
+void MappingTraits<MachO::fileset_entry_command>::mapping(
+    IO &IO, MachO::fileset_entry_command &LoadCommand) {
+  IO.mapRequired("vmaddr", LoadCommand.vmaddr);
+  IO.mapRequired("fileoff", LoadCommand.fileoff);
+  IO.mapRequired("id", LoadCommand.entry_id);
+}
+
 } // end namespace yaml
 
 } // end namespace llvm