xref: /llvm-project/llvm/lib/BinaryFormat/Magic.cpp (revision 581d79a440a6019d0c5b8218aa47403727a838d0)
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 
11 #include "llvm/BinaryFormat/COFF.h"
12 #include "llvm/BinaryFormat/ELF.h"
13 #include "llvm/BinaryFormat/MachO.h"
14 #include "llvm/Support/Endian.h"
15 #include "llvm/Support/FileSystem.h"
16 #include "llvm/Support/MemoryBuffer.h"
17 
18 #if !defined(_MSC_VER) && !defined(__MINGW32__)
19 #include <unistd.h>
20 #else
21 #include <io.h>
22 #endif
23 
24 using namespace llvm;
25 using namespace llvm::support::endian;
26 using namespace llvm::sys::fs;
27 
28 template <size_t N>
29 static bool startswith(StringRef Magic, const char (&S)[N]) {
30   return Magic.startswith(StringRef(S, N - 1));
31 }
32 
33 /// Identify the magic in magic.
34 file_magic llvm::identify_magic(StringRef Magic) {
35   if (Magic.size() < 4)
36     return file_magic::unknown;
37   switch ((unsigned char)Magic[0]) {
38   case 0x00: {
39     // COFF bigobj, CL.exe's LTO object file, or short import library file
40     if (startswith(Magic, "\0\0\xFF\xFF")) {
41       size_t MinSize =
42           offsetof(COFF::BigObjHeader, UUID) + sizeof(COFF::BigObjMagic);
43       if (Magic.size() < MinSize)
44         return file_magic::coff_import_library;
45 
46       const char *Start = Magic.data() + offsetof(COFF::BigObjHeader, UUID);
47       if (memcmp(Start, COFF::BigObjMagic, sizeof(COFF::BigObjMagic)) == 0)
48         return file_magic::coff_object;
49       if (memcmp(Start, COFF::ClGlObjMagic, sizeof(COFF::BigObjMagic)) == 0)
50         return file_magic::coff_cl_gl_object;
51       return file_magic::coff_import_library;
52     }
53     // Windows resource file
54     if (Magic.size() >= sizeof(COFF::WinResMagic) &&
55         memcmp(Magic.data(), COFF::WinResMagic, sizeof(COFF::WinResMagic)) == 0)
56       return file_magic::windows_resource;
57     // 0x0000 = COFF unknown machine type
58     if (Magic[1] == 0)
59       return file_magic::coff_object;
60     if (startswith(Magic, "\0asm"))
61       return file_magic::wasm_object;
62     break;
63   }
64   case 0xDE: // 0x0B17C0DE = BC wraper
65     if (startswith(Magic, "\xDE\xC0\x17\x0B"))
66       return file_magic::bitcode;
67     break;
68   case 'B':
69     if (startswith(Magic, "BC\xC0\xDE"))
70       return file_magic::bitcode;
71     break;
72   case '!':
73     if (startswith(Magic, "!<arch>\n") || startswith(Magic, "!<thin>\n"))
74       return file_magic::archive;
75     break;
76 
77   case '\177':
78     if (startswith(Magic, "\177ELF") && Magic.size() >= 18) {
79       bool Data2MSB = Magic[5] == 2;
80       unsigned high = Data2MSB ? 16 : 17;
81       unsigned low = Data2MSB ? 17 : 16;
82       if (Magic[high] == 0) {
83         switch (Magic[low]) {
84         default:
85           return file_magic::elf;
86         case 1:
87           return file_magic::elf_relocatable;
88         case 2:
89           return file_magic::elf_executable;
90         case 3:
91           return file_magic::elf_shared_object;
92         case 4:
93           return file_magic::elf_core;
94         }
95       }
96       // It's still some type of ELF file.
97       return file_magic::elf;
98     }
99     break;
100 
101   case 0xCA:
102     if (startswith(Magic, "\xCA\xFE\xBA\xBE") ||
103         startswith(Magic, "\xCA\xFE\xBA\xBF")) {
104       // This is complicated by an overlap with Java class files.
105       // See the Mach-O section in /usr/share/file/magic for details.
106       if (Magic.size() >= 8 && Magic[7] < 43)
107         return file_magic::macho_universal_binary;
108     }
109     break;
110 
111   // The two magic numbers for mach-o are:
112   // 0xfeedface - 32-bit mach-o
113   // 0xfeedfacf - 64-bit mach-o
114   case 0xFE:
115   case 0xCE:
116   case 0xCF: {
117     uint16_t type = 0;
118     if (startswith(Magic, "\xFE\xED\xFA\xCE") ||
119         startswith(Magic, "\xFE\xED\xFA\xCF")) {
120       /* Native endian */
121       size_t MinSize;
122       if (Magic[3] == char(0xCE))
123         MinSize = sizeof(MachO::mach_header);
124       else
125         MinSize = sizeof(MachO::mach_header_64);
126       if (Magic.size() >= MinSize)
127         type = Magic[12] << 24 | Magic[13] << 12 | Magic[14] << 8 | Magic[15];
128     } else if (startswith(Magic, "\xCE\xFA\xED\xFE") ||
129                startswith(Magic, "\xCF\xFA\xED\xFE")) {
130       /* Reverse endian */
131       size_t MinSize;
132       if (Magic[0] == char(0xCE))
133         MinSize = sizeof(MachO::mach_header);
134       else
135         MinSize = sizeof(MachO::mach_header_64);
136       if (Magic.size() >= MinSize)
137         type = Magic[15] << 24 | Magic[14] << 12 | Magic[13] << 8 | Magic[12];
138     }
139     switch (type) {
140     default:
141       break;
142     case 1:
143       return file_magic::macho_object;
144     case 2:
145       return file_magic::macho_executable;
146     case 3:
147       return file_magic::macho_fixed_virtual_memory_shared_lib;
148     case 4:
149       return file_magic::macho_core;
150     case 5:
151       return file_magic::macho_preload_executable;
152     case 6:
153       return file_magic::macho_dynamically_linked_shared_lib;
154     case 7:
155       return file_magic::macho_dynamic_linker;
156     case 8:
157       return file_magic::macho_bundle;
158     case 9:
159       return file_magic::macho_dynamically_linked_shared_lib_stub;
160     case 10:
161       return file_magic::macho_dsym_companion;
162     case 11:
163       return file_magic::macho_kext_bundle;
164     }
165     break;
166   }
167   case 0xF0: // PowerPC Windows
168   case 0x83: // Alpha 32-bit
169   case 0x84: // Alpha 64-bit
170   case 0x66: // MPS R4000 Windows
171   case 0x50: // mc68K
172   case 0x4c: // 80386 Windows
173   case 0xc4: // ARMNT Windows
174     if (Magic[1] == 0x01)
175       return file_magic::coff_object;
176     LLVM_FALLTHROUGH;
177 
178   case 0x90: // PA-RISC Windows
179   case 0x68: // mc68K Windows
180     if (Magic[1] == 0x02)
181       return file_magic::coff_object;
182     break;
183 
184   case 'M': // Possible MS-DOS stub on Windows PE file, MSF/PDB file or a
185             // Minidump file.
186     if (startswith(Magic, "MZ") && Magic.size() >= 0x3c + 4) {
187       uint32_t off = read32le(Magic.data() + 0x3c);
188       // PE/COFF file, either EXE or DLL.
189       if (Magic.substr(off).startswith(
190               StringRef(COFF::PEMagic, sizeof(COFF::PEMagic))))
191         return file_magic::pecoff_executable;
192     }
193     if (Magic.startswith("Microsoft C/C++ MSF 7.00\r\n"))
194       return file_magic::pdb;
195     if (startswith(Magic, "MDMP"))
196       return file_magic::minidump;
197     break;
198 
199   case 0x64: // x86-64 or ARM64 Windows.
200     if (Magic[1] == char(0x86) || Magic[1] == char(0xaa))
201       return file_magic::coff_object;
202     break;
203 
204   default:
205     break;
206   }
207   return file_magic::unknown;
208 }
209 
210 std::error_code llvm::identify_magic(const Twine &Path, file_magic &Result) {
211   auto FileOrError = MemoryBuffer::getFile(Path, -1LL, false);
212   if (!FileOrError)
213     return FileOrError.getError();
214 
215   std::unique_ptr<MemoryBuffer> FileBuffer = std::move(*FileOrError);
216   Result = identify_magic(FileBuffer->getBuffer());
217 
218   return std::error_code();
219 }
220