LLVM 18.0.0git
Magic.cpp
Go to the documentation of this file.
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
10#include "llvm/ADT/StringRef.h"
11#include "llvm/ADT/Twine.h"
14#include "llvm/Support/Endian.h"
16
17#if !defined(_MSC_VER) && !defined(__MINGW32__)
18#include <unistd.h>
19#else
20#include <io.h>
21#endif
22
23using namespace llvm;
24using namespace llvm::support::endian;
25using namespace llvm::sys::fs;
26
27template <size_t N>
28static bool startswith(StringRef Magic, const char (&S)[N]) {
29 return Magic.startswith(StringRef(S, N - 1));
30}
31
32/// Identify the magic in magic.
34 if (Magic.size() < 4)
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 =
42 if (Magic.size() < MinSize)
44
45 const char *Start = Magic.data() + offsetof(COFF::BigObjHeader, UUID);
46 if (memcmp(Start, COFF::BigObjMagic, sizeof(COFF::BigObjMagic)) == 0)
48 if (memcmp(Start, COFF::ClGlObjMagic, sizeof(COFF::BigObjMagic)) == 0)
51 }
52 // Windows resource file
53 if (Magic.size() >= sizeof(COFF::WinResMagic) &&
54 memcmp(Magic.data(), COFF::WinResMagic, sizeof(COFF::WinResMagic)) == 0)
56 // 0x0000 = COFF unknown machine type
57 if (Magic[1] == 0)
59 if (startswith(Magic, "\0asm"))
61 break;
62 }
63
64 case 0x01:
65 // XCOFF format
66 if (startswith(Magic, "\x01\xDF"))
68 if (startswith(Magic, "\x01\xF7"))
70 break;
71
72 case 0x03:
73 if (startswith(Magic, "\x03\xF0\x00"))
75 break;
76
77 case 0x10:
78 if (startswith(Magic, "\x10\xFF\x10\xAD"))
80 break;
81
82 case 0xDE: // 0x0B17C0DE = BC wraper
83 if (startswith(Magic, "\xDE\xC0\x17\x0B"))
85 break;
86 case 'B':
87 if (startswith(Magic, "BC\xC0\xDE"))
89 break;
90 case '!':
91 if (startswith(Magic, "!<arch>\n") || startswith(Magic, "!<thin>\n"))
93 break;
94 case '<':
95 if (startswith(Magic, "<bigaf>\n"))
97 break;
98 case '\177':
99 if (startswith(Magic, "\177ELF") && Magic.size() >= 18) {
100 bool Data2MSB = Magic[5] == 2;
101 unsigned high = Data2MSB ? 16 : 17;
102 unsigned low = Data2MSB ? 17 : 16;
103 if (Magic[high] == 0) {
104 switch (Magic[low]) {
105 default:
106 return file_magic::elf;
107 case 1:
109 case 2:
111 case 3:
113 case 4:
115 }
116 }
117 // It's still some type of ELF file.
118 return file_magic::elf;
119 }
120 break;
121
122 case 0xCA:
123 if (startswith(Magic, "\xCA\xFE\xBA\xBE") ||
124 startswith(Magic, "\xCA\xFE\xBA\xBF")) {
125 // This is complicated by an overlap with Java class files.
126 // See the Mach-O section in /usr/share/file/magic for details.
127 if (Magic.size() >= 8 && Magic[7] < 43)
129 }
130 break;
131
132 // The two magic numbers for mach-o are:
133 // 0xfeedface - 32-bit mach-o
134 // 0xfeedfacf - 64-bit mach-o
135 case 0xFE:
136 case 0xCE:
137 case 0xCF: {
138 uint16_t type = 0;
139 if (startswith(Magic, "\xFE\xED\xFA\xCE") ||
140 startswith(Magic, "\xFE\xED\xFA\xCF")) {
141 /* Native endian */
142 size_t MinSize;
143 if (Magic[3] == char(0xCE))
144 MinSize = sizeof(MachO::mach_header);
145 else
146 MinSize = sizeof(MachO::mach_header_64);
147 if (Magic.size() >= MinSize)
148 type = Magic[12] << 24 | Magic[13] << 12 | Magic[14] << 8 | Magic[15];
149 } else if (startswith(Magic, "\xCE\xFA\xED\xFE") ||
150 startswith(Magic, "\xCF\xFA\xED\xFE")) {
151 /* Reverse endian */
152 size_t MinSize;
153 if (Magic[0] == char(0xCE))
154 MinSize = sizeof(MachO::mach_header);
155 else
156 MinSize = sizeof(MachO::mach_header_64);
157 if (Magic.size() >= MinSize)
158 type = Magic[15] << 24 | Magic[14] << 12 | Magic[13] << 8 | Magic[12];
159 }
160 switch (type) {
161 default:
162 break;
163 case 1:
165 case 2:
167 case 3:
169 case 4:
171 case 5:
173 case 6:
175 case 7:
177 case 8:
179 case 9:
181 case 10:
183 case 11:
185 case 12:
187 }
188 break;
189 }
190 case 0xF0: // PowerPC Windows
191 case 0x83: // Alpha 32-bit
192 case 0x84: // Alpha 64-bit
193 case 0x66: // MPS R4000 Windows
194 case 0x50: // mc68K
195 if (startswith(Magic, "\x50\xed\x55\xba"))
197 [[fallthrough]];
198
199 case 0x4c: // 80386 Windows
200 case 0xc4: // ARMNT Windows
201 if (Magic[1] == 0x01)
203 [[fallthrough]];
204
205 case 0x90: // PA-RISC Windows
206 case 0x68: // mc68K Windows
207 if (Magic[1] == 0x02)
209 break;
210
211 case 'M': // Possible MS-DOS stub on Windows PE file, MSF/PDB file or a
212 // Minidump file.
213 if (startswith(Magic, "MZ") && Magic.size() >= 0x3c + 4) {
214 uint32_t off = read32le(Magic.data() + 0x3c);
215 // PE/COFF file, either EXE or DLL.
216 if (Magic.substr(off).startswith(
219 }
220 if (Magic.startswith("Microsoft C/C++ MSF 7.00\r\n"))
221 return file_magic::pdb;
222 if (startswith(Magic, "MDMP"))
224 break;
225
226 case 0x64: // x86-64 or ARM64 Windows.
227 if (Magic[1] == char(0x86) || Magic[1] == char(0xaa))
229 break;
230
231 case 0x2d: // YAML '-' MachO TBD.
232 if (startswith(Magic, "--- !tapi") || startswith(Magic, "---\narchs:"))
234 break;
235 case 0x7b: // JSON '{' MachO TBD.
237 break;
238
239 case 'D': // DirectX container file - DXBC
240 if (startswith(Magic, "DXBC"))
242 break;
243
244 case 0x41: // ARM64EC windows
245 if (Magic[1] == char(0xA6))
247 break;
248
249 case 0x4e: // ARM64X windows
250 if (Magic[1] == char(0xA6))
252 break;
253
254 default:
255 break;
256 }
257 return file_magic::unknown;
258}
259
260std::error_code llvm::identify_magic(const Twine &Path, file_magic &Result) {
261 auto FileOrError = MemoryBuffer::getFile(Path, /*IsText=*/false,
262 /*RequiresNullTerminator=*/false);
263 if (!FileOrError)
264 return FileOrError.getError();
265
266 std::unique_ptr<MemoryBuffer> FileBuffer = std::move(*FileOrError);
267 Result = identify_magic(FileBuffer->getBuffer());
268
269 return std::error_code();
270}
#define offsetof(TYPE, MEMBER)
static bool startswith(StringRef Magic, const char(&S)[N])
Definition: Magic.cpp:28
Merge contiguous icmps into a memcmp
Definition: MergeICmps.cpp:911
std::pair< llvm::MachO::Target, std::string > UUID
static ErrorOr< std::unique_ptr< MemoryBuffer > > getFile(const Twine &Filename, bool IsText=false, bool RequiresNullTerminator=true, bool IsVolatile=false, std::optional< Align > Alignment=std::nullopt)
Open the specified file as a MemoryBuffer, returning a new MemoryBuffer if successful,...
StringRef - Represent a constant reference to a string, i.e.
Definition: StringRef.h:50
Twine - A lightweight data structure for efficiently representing the concatenation of temporary valu...
Definition: Twine.h:81
static const char ClGlObjMagic[]
Definition: COFF.h:42
static const char WinResMagic[]
Definition: COFF.h:48
static const char BigObjMagic[]
Definition: COFF.h:37
static const char PEMagic[]
Definition: COFF.h:35
uint32_t read32le(const void *P)
Definition: Endian.h:377
This is an optimization pass for GlobalISel generic memory operations.
Definition: AddressRanges.h:18
file_magic identify_magic(StringRef magic)
Identify the type of a binary file based on how magical it is.
Definition: Magic.cpp:33
#define N
file_magic - An "enum class" enumeration of file types based on magic (the first N bytes of the file)...
Definition: Magic.h:20
@ coff_import_library
COFF import library.
Definition: Magic.h:47
@ pdb
Windows PDB debug info file.
Definition: Magic.h:53
@ elf_relocatable
ELF Relocatable object file.
Definition: Magic.h:26
@ archive
ar style archive file
Definition: Magic.h:24
@ elf_shared_object
ELF dynamically linked shared lib.
Definition: Magic.h:28
@ goff_object
GOFF object file.
Definition: Magic.h:30
@ minidump
Windows minidump file.
Definition: Magic.h:44
@ macho_dynamically_linked_shared_lib
Mach-O dynlinked shared lib.
Definition: Magic.h:36
@ xcoff_object_64
64-bit XCOFF object file
Definition: Magic.h:51
@ elf_executable
ELF Executable image.
Definition: Magic.h:27
@ macho_dynamically_linked_shared_lib_stub
Mach-O Shared lib stub.
Definition: Magic.h:39
@ macho_preload_executable
Mach-O Preloaded Executable.
Definition: Magic.h:35
@ macho_file_set
Mach-O file set binary.
Definition: Magic.h:43
@ dxcontainer_object
DirectX container file.
Definition: Magic.h:57
@ macho_kext_bundle
Mach-O kext bundle file.
Definition: Magic.h:41
@ pecoff_executable
PECOFF executable file.
Definition: Magic.h:48
@ offload_binary
LLVM offload object file.
Definition: Magic.h:56
@ macho_universal_binary
Mach-O universal binary.
Definition: Magic.h:42
@ bitcode
Bitcode file.
Definition: Magic.h:23
@ macho_core
Mach-O Core File.
Definition: Magic.h:34
@ wasm_object
WebAssembly Object file.
Definition: Magic.h:52
@ xcoff_object_32
32-bit XCOFF object file
Definition: Magic.h:50
@ windows_resource
Windows compiled resource file (.res)
Definition: Magic.h:49
@ elf_core
ELF core image.
Definition: Magic.h:29
@ macho_object
Mach-O Object file.
Definition: Magic.h:31
@ coff_object
COFF object file.
Definition: Magic.h:46
@ elf
ELF Unknown type.
Definition: Magic.h:25
@ macho_bundle
Mach-O Bundle file.
Definition: Magic.h:38
@ coff_cl_gl_object
Microsoft cl.exe's intermediate code file.
Definition: Magic.h:45
@ cuda_fatbinary
CUDA Fatbinary object file.
Definition: Magic.h:55
@ macho_executable
Mach-O Executable.
Definition: Magic.h:32
@ macho_dsym_companion
Mach-O dSYM companion file.
Definition: Magic.h:40
@ unknown
Unrecognized file.
Definition: Magic.h:22
@ macho_fixed_virtual_memory_shared_lib
Mach-O Shared Lib, FVM.
Definition: Magic.h:33
@ macho_dynamic_linker
The Mach-O dynamic linker.
Definition: Magic.h:37
@ tapi_file
Text-based Dynamic Library Stub file.
Definition: Magic.h:54