1 //===- llvm/BinaryFormat/Magic.cpp - File magic identification --*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 
9 #include "llvm/BinaryFormat/Magic.h"
10 #include "llvm/ADT/StringRef.h"
11 #include "llvm/ADT/Twine.h"
12 #include "llvm/BinaryFormat/COFF.h"
13 #include "llvm/BinaryFormat/MachO.h"
14 #include "llvm/Support/Endian.h"
15 #include "llvm/Support/MemoryBuffer.h"
16 
17 #if !defined(_MSC_VER) && !defined(__MINGW32__)
18 #include <unistd.h>
19 #else
20 #include <io.h>
21 #endif
22 
23 using namespace llvm;
24 using namespace llvm::support::endian;
25 using namespace llvm::sys::fs;
26 
27 template <size_t N>
28 static bool startswith(StringRef Magic, const char (&S)[N]) {
29   return Magic.startswith(StringRef(S, N - 1));
30 }
31 
32 /// Identify the magic in magic.
33 file_magic llvm::identify_magic(StringRef Magic) {
34   if (Magic.size() < 4)
35     return file_magic::unknown;
36   switch ((unsigned char)Magic[0]) {
37   case 0x00: {
38     // COFF bigobj, CL.exe's LTO object file, or short import library file
39     if (startswith(Magic, "\0\0\xFF\xFF")) {
40       size_t MinSize =
41           offsetof(COFF::BigObjHeader, UUID) + sizeof(COFF::BigObjMagic);
42       if (Magic.size() < MinSize)
43         return file_magic::coff_import_library;
44 
45       const char *Start = Magic.data() + offsetof(COFF::BigObjHeader, UUID);
46       if (memcmp(Start, COFF::BigObjMagic, sizeof(COFF::BigObjMagic)) == 0)
47         return file_magic::coff_object;
48       if (memcmp(Start, COFF::ClGlObjMagic, sizeof(COFF::BigObjMagic)) == 0)
49         return file_magic::coff_cl_gl_object;
50       return file_magic::coff_import_library;
51     }
52     // Windows resource file
53     if (Magic.size() >= sizeof(COFF::WinResMagic) &&
54         memcmp(Magic.data(), COFF::WinResMagic, sizeof(COFF::WinResMagic)) == 0)
55       return file_magic::windows_resource;
56     // 0x0000 = COFF unknown machine type
57     if (Magic[1] == 0)
58       return file_magic::coff_object;
59     if (startswith(Magic, "\0asm"))
60       return file_magic::wasm_object;
61     break;
62   }
63 
64   case 0x01:
65     // XCOFF format
66     if (startswith(Magic, "\x01\xDF"))
67       return file_magic::xcoff_object_32;
68     if (startswith(Magic, "\x01\xF7"))
69       return file_magic::xcoff_object_64;
70     break;
71 
72   case 0x03:
73     if (startswith(Magic, "\x03\xF0\x00"))
74       return file_magic::goff_object;
75     break;
76 
77   case 0xDE: // 0x0B17C0DE = BC wraper
78     if (startswith(Magic, "\xDE\xC0\x17\x0B"))
79       return file_magic::bitcode;
80     break;
81   case 'B':
82     if (startswith(Magic, "BC\xC0\xDE"))
83       return file_magic::bitcode;
84     break;
85   case '!':
86     if (startswith(Magic, "!<arch>\n") || startswith(Magic, "!<thin>\n"))
87       return file_magic::archive;
88     break;
89   case '<':
90     if (startswith(Magic, "<bigaf>\n"))
91       return file_magic::archive;
92     break;
93   case '\177':
94     if (startswith(Magic, "\177ELF") && Magic.size() >= 18) {
95       bool Data2MSB = Magic[5] == 2;
96       unsigned high = Data2MSB ? 16 : 17;
97       unsigned low = Data2MSB ? 17 : 16;
98       if (Magic[high] == 0) {
99         switch (Magic[low]) {
100         default:
101           return file_magic::elf;
102         case 1:
103           return file_magic::elf_relocatable;
104         case 2:
105           return file_magic::elf_executable;
106         case 3:
107           return file_magic::elf_shared_object;
108         case 4:
109           return file_magic::elf_core;
110         }
111       }
112       // It's still some type of ELF file.
113       return file_magic::elf;
114     }
115     break;
116 
117   case 0xCA:
118     if (startswith(Magic, "\xCA\xFE\xBA\xBE") ||
119         startswith(Magic, "\xCA\xFE\xBA\xBF")) {
120       // This is complicated by an overlap with Java class files.
121       // See the Mach-O section in /usr/share/file/magic for details.
122       if (Magic.size() >= 8 && Magic[7] < 43)
123         return file_magic::macho_universal_binary;
124     }
125     break;
126 
127   // The two magic numbers for mach-o are:
128   // 0xfeedface - 32-bit mach-o
129   // 0xfeedfacf - 64-bit mach-o
130   case 0xFE:
131   case 0xCE:
132   case 0xCF: {
133     uint16_t type = 0;
134     if (startswith(Magic, "\xFE\xED\xFA\xCE") ||
135         startswith(Magic, "\xFE\xED\xFA\xCF")) {
136       /* Native endian */
137       size_t MinSize;
138       if (Magic[3] == char(0xCE))
139         MinSize = sizeof(MachO::mach_header);
140       else
141         MinSize = sizeof(MachO::mach_header_64);
142       if (Magic.size() >= MinSize)
143         type = Magic[12] << 24 | Magic[13] << 12 | Magic[14] << 8 | Magic[15];
144     } else if (startswith(Magic, "\xCE\xFA\xED\xFE") ||
145                startswith(Magic, "\xCF\xFA\xED\xFE")) {
146       /* Reverse endian */
147       size_t MinSize;
148       if (Magic[0] == char(0xCE))
149         MinSize = sizeof(MachO::mach_header);
150       else
151         MinSize = sizeof(MachO::mach_header_64);
152       if (Magic.size() >= MinSize)
153         type = Magic[15] << 24 | Magic[14] << 12 | Magic[13] << 8 | Magic[12];
154     }
155     switch (type) {
156     default:
157       break;
158     case 1:
159       return file_magic::macho_object;
160     case 2:
161       return file_magic::macho_executable;
162     case 3:
163       return file_magic::macho_fixed_virtual_memory_shared_lib;
164     case 4:
165       return file_magic::macho_core;
166     case 5:
167       return file_magic::macho_preload_executable;
168     case 6:
169       return file_magic::macho_dynamically_linked_shared_lib;
170     case 7:
171       return file_magic::macho_dynamic_linker;
172     case 8:
173       return file_magic::macho_bundle;
174     case 9:
175       return file_magic::macho_dynamically_linked_shared_lib_stub;
176     case 10:
177       return file_magic::macho_dsym_companion;
178     case 11:
179       return file_magic::macho_kext_bundle;
180     }
181     break;
182   }
183   case 0xF0: // PowerPC Windows
184   case 0x83: // Alpha 32-bit
185   case 0x84: // Alpha 64-bit
186   case 0x66: // MPS R4000 Windows
187   case 0x50: // mc68K
188   case 0x4c: // 80386 Windows
189   case 0xc4: // ARMNT Windows
190     if (Magic[1] == 0x01)
191       return file_magic::coff_object;
192     LLVM_FALLTHROUGH;
193 
194   case 0x90: // PA-RISC Windows
195   case 0x68: // mc68K Windows
196     if (Magic[1] == 0x02)
197       return file_magic::coff_object;
198     break;
199 
200   case 'M': // Possible MS-DOS stub on Windows PE file, MSF/PDB file or a
201             // Minidump file.
202     if (startswith(Magic, "MZ") && Magic.size() >= 0x3c + 4) {
203       uint32_t off = read32le(Magic.data() + 0x3c);
204       // PE/COFF file, either EXE or DLL.
205       if (Magic.substr(off).startswith(
206               StringRef(COFF::PEMagic, sizeof(COFF::PEMagic))))
207         return file_magic::pecoff_executable;
208     }
209     if (Magic.startswith("Microsoft C/C++ MSF 7.00\r\n"))
210       return file_magic::pdb;
211     if (startswith(Magic, "MDMP"))
212       return file_magic::minidump;
213     break;
214 
215   case 0x64: // x86-64 or ARM64 Windows.
216     if (Magic[1] == char(0x86) || Magic[1] == char(0xaa))
217       return file_magic::coff_object;
218     break;
219 
220   case 0x2d: // YAML '-'
221     if (startswith(Magic, "--- !tapi") || startswith(Magic, "---\narchs:"))
222       return file_magic::tapi_file;
223     break;
224 
225   default:
226     break;
227   }
228   return file_magic::unknown;
229 }
230 
231 std::error_code llvm::identify_magic(const Twine &Path, file_magic &Result) {
232   auto FileOrError = MemoryBuffer::getFile(Path, /*IsText=*/false,
233                                            /*RequiresNullTerminator=*/false);
234   if (!FileOrError)
235     return FileOrError.getError();
236 
237   std::unique_ptr<MemoryBuffer> FileBuffer = std::move(*FileOrError);
238   Result = identify_magic(FileBuffer->getBuffer());
239 
240   return std::error_code();
241 }
242