[MachO] Allow the LC_IDENT load command

[vsk]

Summary

xnu coredumps include an LC_IDENT load command. It's helpful to be able
to just ignore these. IIUC an interested client can grab the identifier
using the MachOObjectFile::load_commands() API.

The status quo is that llvm bails out when it finds an LC_IDENT because
the command is obsolete (see isLoadCommandObsolete).

[JDevlieghere]

LGTM

[jasonmolenda]

I discussed this with KevinE in the past and he was adamant that he didn't want LC_IDENT being used for this. I don't have strong feelings. I think it was an old load command used for something else altogether decades ago and it started getting used in kernel corefiles and he wasn't thrilled with that. There is a Mach-O LC_NOTE load command that allows for the kernel version string to be stored in the kernel file ("kern ver str") but it hasn't been adopted by darwin kernel dumpers yet.

to be clear: I don't object to ignoring this in the llvm tools, but it was an intentional decision in the past. I don't think that should weigh especially heavily on us today though.

llvm/lib/Object/MachOObjectFile.cpp

for (unsigned I = 0; I < LoadCommandCount; ++I) {
     } else if (Load.C.cmd == MachO::LC_THREAD) {
       if ((Err = checkThreadCommand(*this, Load, I, "LC_THREAD")))
         return;
     // Note: LC_TWOLEVEL_HINTS is really obsolete and is not supported.
     } else if (Load.C.cmd == MachO::LC_TWOLEVEL_HINTS) {
        if ((Err = checkTwoLevelHintsCommand(*this, Load, I,
                                             &TwoLevelHintsLoadCmd, Elements)))
          return;
+    } else if (Load.C.cmd == MachO::LC_IDENT) {
+      // Note: LC_IDENT is ignored.
+      continue;
     } else if (isLoadCommandObsolete(Load.C.cmd)) {
       Err = malformedError("load command " + Twine(I) + " for cmd value of: " +
                            Twine(Load.C.cmd) + " is obsolete and not "
                            "supported");
       return;
     }
     // TODO: generate a error for unknown load commands by default.  But still
     // need work out an approach to allow or not allow unknown values like this

llvm/test/Object/macho-invalid.test

 INVALID-IDFVMLIB-OBSOLETE: macho-invalid-idfvmlib-obsolete': truncated or malformed object (load command 0 for cmd value of: 7 is obsolete and not supported)
 
 RUN: not llvm-objdump --macho --private-headers %p/Inputs/macho-invalid-loadfvmlib-obsolete 2>&1 | FileCheck --check-prefix=INVALID-LOADFVMLIB-OBSOLETE %s
 INVALID-LOADFVMLIB-OBSOLETE: macho-invalid-loadfvmlib-obsolete': truncated or malformed object (load command 0 for cmd value of: 6 is obsolete and not supported)
 
 RUN: not llvm-objdump --macho --private-headers %p/Inputs/macho-invalid-prebound_dylib-obsolete 2>&1 | FileCheck --check-prefix=INVALID-PREBOUND_DYLIB-OBSOLETE %s
 INVALID-PREBOUND_DYLIB-OBSOLETE: macho-invalid-prebound_dylib-obsolete': truncated or malformed object (load command 0 for cmd value of: 16 is obsolete and not supported)
 
-RUN: not llvm-objdump --macho --private-headers %p/Inputs/macho-invalid-ident-obsolete 2>&1 | FileCheck --check-prefix=INVALID-IDENT-OBSOLETE %s
-INVALID-IDENT-OBSOLETE: macho-invalid-ident-obsolete': truncated or malformed object (load command 0 for cmd value of: 8 is obsolete and not supported)
-
 RUN: not llvm-objdump --macho --private-headers %p/Inputs/macho-invalid-fvmfile-obsolete 2>&1 | FileCheck --check-prefix=INVALID-FVMFILE-OBSOLETE %s
 INVALID-FVMFILE-OBSOLETE: macho-invalid-fvmfile-obsolete': truncated or malformed object (load command 0 for cmd value of: 9 is obsolete and not supported)
 
 RUN: not llvm-objdump --macho --private-headers %p/Inputs/macho-invalid-prepage-obsolete 2>&1 | FileCheck --check-prefix=INVALID-PREPAGE-OBSOLETE %s
 INVALID-PREPAGE-OBSOLETE: macho-invalid-prepage-obsolete': truncated or malformed object (load command 0 for cmd value of: 10 is obsolete and not supported)
 
 RUN: not llvm-objdump --macho --private-headers %p/Inputs/macho-invalid-symtab-overlap 2>&1 | FileCheck --check-prefix=INVALID-SYMTAB-OVERLAP %s
 INVALID-SYMTAB-OVERLAP: macho-invalid-symtab-overlap': truncated or malformed object (symbol table at offset 8 with a size of 12, overlaps Mach-O headers at offset 0 with a size of 52)

llvm/test/ObjectYAML/MachO/lc-ident.yaml

+# RUN: yaml2obj %s | obj2yaml | FileCheck %s
+
+--- !mach-o
+FileHeader:
+  magic:           0xFEEDFACF
+  cputype:         0x01000007
+  cpusubtype:      0x80000003
+  filetype:        0x00000002
+  ncmds:           1
+  sizeofcmds:      16
+  flags:           0x00218085
+  reserved:        0x00000000
+LoadCommands:
+  - cmd:             LC_IDENT
+    cmdsize:         16
+    PayloadBytes:    [ 0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7 ]
+...
+
+# CHECK: LoadCommands:
+# CHECK-NEXT:   - cmd:             LC_IDENT
+# CHECK-NEXT:     cmdsize:         16
+# CHECK-NEXT:     PayloadBytes:    [ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07 ]