2 * Copyright (c) 1999-2010 Apple Inc. All Rights Reserved.
4 * @APPLE_LICENSE_HEADER_START@
6 * This file contains Original Code and/or Modifications of Original Code
7 * as defined in and that are subject to the Apple Public Source License
8 * Version 2.0 (the 'License'). You may not use this file except in
9 * compliance with the License. Please obtain a copy of the License at
10 * http://www.opensource.apple.com/apsl/ and read it before using this
13 * The Original Code and all software distributed under the License are
14 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
15 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
16 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
18 * Please see the License for the specific language governing rights and
19 * limitations under the License.
21 * @APPLE_LICENSE_HEADER_END@
23 #ifndef _MACHO_LOADER_H_
24 #define _MACHO_LOADER_H_
27 * This file describes the format of mach object files.
32 * <mach/machine.h> is needed here for the cpu_type_t and cpu_subtype_t types
33 * and contains the constants for the possible values of these types.
35 #include <mach/machine.h>
38 * <mach/vm_prot.h> is needed here for the vm_prot_t type and contains the
39 * constants that are or'ed together for the possible values of this type.
41 #include <mach/vm_prot.h>
44 * <machine/thread_status.h> is expected to define the flavors of the thread
45 * states and the structures of those flavors for each machine.
47 #include <mach/machine/thread_status.h>
48 #include <architecture/byte_order.h>
51 * The 32-bit mach header appears at the very beginning of the object file for
52 * 32-bit architectures.
55 uint32_t magic
; /* mach magic number identifier */
56 cpu_type_t cputype
; /* cpu specifier */
57 cpu_subtype_t cpusubtype
; /* machine specifier */
58 uint32_t filetype
; /* type of file */
59 uint32_t ncmds
; /* number of load commands */
60 uint32_t sizeofcmds
; /* the size of all the load commands */
61 uint32_t flags
; /* flags */
64 /* Constant for the magic field of the mach_header (32-bit architectures) */
65 #define MH_MAGIC 0xfeedface /* the mach magic number */
66 #define MH_CIGAM 0xcefaedfe /* NXSwapInt(MH_MAGIC) */
69 * The 64-bit mach header appears at the very beginning of object files for
70 * 64-bit architectures.
72 struct mach_header_64
{
73 uint32_t magic
; /* mach magic number identifier */
74 cpu_type_t cputype
; /* cpu specifier */
75 cpu_subtype_t cpusubtype
; /* machine specifier */
76 uint32_t filetype
; /* type of file */
77 uint32_t ncmds
; /* number of load commands */
78 uint32_t sizeofcmds
; /* the size of all the load commands */
79 uint32_t flags
; /* flags */
80 uint32_t reserved
; /* reserved */
83 /* Constant for the magic field of the mach_header_64 (64-bit architectures) */
84 #define MH_MAGIC_64 0xfeedfacf /* the 64-bit mach magic number */
85 #define MH_CIGAM_64 0xcffaedfe /* NXSwapInt(MH_MAGIC_64) */
88 * The layout of the file depends on the filetype. For all but the MH_OBJECT
89 * file type the segments are padded out and aligned on a segment alignment
90 * boundary for efficient demand pageing. The MH_EXECUTE, MH_FVMLIB, MH_DYLIB,
91 * MH_DYLINKER and MH_BUNDLE file types also have the headers included as part
92 * of their first segment.
94 * The file type MH_OBJECT is a compact format intended as output of the
95 * assembler and input (and possibly output) of the link editor (the .o
96 * format). All sections are in one unnamed segment with no segment padding.
97 * This format is used as an executable format when the file is so small the
98 * segment padding greatly increases its size.
100 * The file type MH_PRELOAD is an executable format intended for things that
101 * are not executed under the kernel (proms, stand alones, kernels, etc). The
102 * format can be executed under the kernel but may demand paged it and not
103 * preload it before execution.
105 * A core file is in MH_CORE format and can be any in an arbritray legal
108 * Constants for the filetype field of the mach_header
110 #define MH_OBJECT 0x1 /* relocatable object file */
111 #define MH_EXECUTE 0x2 /* demand paged executable file */
112 #define MH_FVMLIB 0x3 /* fixed VM shared library file */
113 #define MH_CORE 0x4 /* core file */
114 #define MH_PRELOAD 0x5 /* preloaded executable file */
115 #define MH_DYLIB 0x6 /* dynamically bound shared library */
116 #define MH_DYLINKER 0x7 /* dynamic link editor */
117 #define MH_BUNDLE 0x8 /* dynamically bound bundle file */
118 #define MH_DYLIB_STUB 0x9 /* shared library stub for static */
119 /* linking only, no section contents */
120 #define MH_DSYM 0xa /* companion file with only debug */
122 #define MH_KEXT_BUNDLE 0xb /* x86_64 kexts */
124 /* Constants for the flags field of the mach_header */
125 #define MH_NOUNDEFS 0x1 /* the object file has no undefined
127 #define MH_INCRLINK 0x2 /* the object file is the output of an
128 incremental link against a base file
129 and can't be link edited again */
130 #define MH_DYLDLINK 0x4 /* the object file is input for the
131 dynamic linker and can't be staticly
133 #define MH_BINDATLOAD 0x8 /* the object file's undefined
134 references are bound by the dynamic
135 linker when loaded. */
136 #define MH_PREBOUND 0x10 /* the file has its dynamic undefined
137 references prebound. */
138 #define MH_SPLIT_SEGS 0x20 /* the file has its read-only and
139 read-write segments split */
140 #define MH_LAZY_INIT 0x40 /* the shared library init routine is
141 to be run lazily via catching memory
142 faults to its writeable segments
144 #define MH_TWOLEVEL 0x80 /* the image is using two-level name
146 #define MH_FORCE_FLAT 0x100 /* the executable is forcing all images
147 to use flat name space bindings */
148 #define MH_NOMULTIDEFS 0x200 /* this umbrella guarantees no multiple
149 defintions of symbols in its
150 sub-images so the two-level namespace
151 hints can always be used. */
152 #define MH_NOFIXPREBINDING 0x400 /* do not have dyld notify the
153 prebinding agent about this
155 #define MH_PREBINDABLE 0x800 /* the binary is not prebound but can
156 have its prebinding redone. only used
157 when MH_PREBOUND is not set. */
158 #define MH_ALLMODSBOUND 0x1000 /* indicates that this binary binds to
159 all two-level namespace modules of
160 its dependent libraries. only used
161 when MH_PREBINDABLE and MH_TWOLEVEL
163 #define MH_SUBSECTIONS_VIA_SYMBOLS 0x2000/* safe to divide up the sections into
164 sub-sections via symbols for dead
166 #define MH_CANONICAL 0x4000 /* the binary has been canonicalized
167 via the unprebind operation */
168 #define MH_WEAK_DEFINES 0x8000 /* the final linked image contains
169 external weak symbols */
170 #define MH_BINDS_TO_WEAK 0x10000 /* the final linked image uses
173 #define MH_ALLOW_STACK_EXECUTION 0x20000/* When this bit is set, all stacks
174 in the task will be given stack
175 execution privilege. Only used in
176 MH_EXECUTE filetypes. */
177 #define MH_ROOT_SAFE 0x40000 /* When this bit is set, the binary
178 declares it is safe for use in
179 processes with uid zero */
181 #define MH_SETUID_SAFE 0x80000 /* When this bit is set, the binary
182 declares it is safe for use in
183 processes when issetugid() is true */
185 #define MH_NO_REEXPORTED_DYLIBS 0x100000 /* When this bit is set on a dylib,
186 the static linker does not need to
187 examine dependent dylibs to see
188 if any are re-exported */
189 #define MH_PIE 0x200000 /* When this bit is set, the OS will
190 load the main executable at a
191 random address. Only used in
192 MH_EXECUTE filetypes. */
193 #define MH_DEAD_STRIPPABLE_DYLIB 0x400000 /* Only for use on dylibs. When
194 linking against a dylib that
195 has this bit set, the static linker
196 will automatically not create a
197 LC_LOAD_DYLIB load command to the
198 dylib if no symbols are being
199 referenced from the dylib. */
200 #define MH_HAS_TLV_DESCRIPTORS 0x800000 /* Contains a section of type
201 S_THREAD_LOCAL_VARIABLES */
203 #define MH_NO_HEAP_EXECUTION 0x1000000 /* When this bit is set, the OS will
204 run the main executable with
205 a non-executable heap even on
206 platforms (e.g. i386) that don't
207 require it. Only used in MH_EXECUTE
210 #define MH_APP_EXTENSION_SAFE 0x02000000 /* The code was linked for use in an
211 application extension. */
214 * The load commands directly follow the mach_header. The total size of all
215 * of the commands is given by the sizeofcmds field in the mach_header. All
216 * load commands must have as their first two fields cmd and cmdsize. The cmd
217 * field is filled in with a constant for that command type. Each command type
218 * has a structure specifically for it. The cmdsize field is the size in bytes
219 * of the particular load command structure plus anything that follows it that
220 * is a part of the load command (i.e. section structures, strings, etc.). To
221 * advance to the next load command the cmdsize can be added to the offset or
222 * pointer of the current load command. The cmdsize for 32-bit architectures
223 * MUST be a multiple of 4 bytes and for 64-bit architectures MUST be a multiple
224 * of 8 bytes (these are forever the maximum alignment of any load commands).
225 * The padded bytes must be zero. All tables in the object file must also
226 * follow these rules so the file can be memory mapped. Otherwise the pointers
227 * to these tables will not work well or at all on some machines. With all
228 * padding zeroed like objects will compare byte for byte.
230 struct load_command
{
231 uint32_t cmd
; /* type of load command */
232 uint32_t cmdsize
; /* total size of command in bytes */
236 * After MacOS X 10.1 when a new load command is added that is required to be
237 * understood by the dynamic linker for the image to execute properly the
238 * LC_REQ_DYLD bit will be or'ed into the load command constant. If the dynamic
239 * linker sees such a load command it it does not understand will issue a
240 * "unknown load command required for execution" error and refuse to use the
241 * image. Other load commands without this bit that are not understood will
244 #define LC_REQ_DYLD 0x80000000
246 /* Constants for the cmd field of all load commands, the type */
247 #define LC_SEGMENT 0x1 /* segment of this file to be mapped */
248 #define LC_SYMTAB 0x2 /* link-edit stab symbol table info */
249 #define LC_SYMSEG 0x3 /* link-edit gdb symbol table info (obsolete) */
250 #define LC_THREAD 0x4 /* thread */
251 #define LC_UNIXTHREAD 0x5 /* unix thread (includes a stack) */
252 #define LC_LOADFVMLIB 0x6 /* load a specified fixed VM shared library */
253 #define LC_IDFVMLIB 0x7 /* fixed VM shared library identification */
254 #define LC_IDENT 0x8 /* object identification info (obsolete) */
255 #define LC_FVMFILE 0x9 /* fixed VM file inclusion (internal use) */
256 #define LC_PREPAGE 0xa /* prepage command (internal use) */
257 #define LC_DYSYMTAB 0xb /* dynamic link-edit symbol table info */
258 #define LC_LOAD_DYLIB 0xc /* load a dynamically linked shared library */
259 #define LC_ID_DYLIB 0xd /* dynamically linked shared lib ident */
260 #define LC_LOAD_DYLINKER 0xe /* load a dynamic linker */
261 #define LC_ID_DYLINKER 0xf /* dynamic linker identification */
262 #define LC_PREBOUND_DYLIB 0x10 /* modules prebound for a dynamically */
263 /* linked shared library */
264 #define LC_ROUTINES 0x11 /* image routines */
265 #define LC_SUB_FRAMEWORK 0x12 /* sub framework */
266 #define LC_SUB_UMBRELLA 0x13 /* sub umbrella */
267 #define LC_SUB_CLIENT 0x14 /* sub client */
268 #define LC_SUB_LIBRARY 0x15 /* sub library */
269 #define LC_TWOLEVEL_HINTS 0x16 /* two-level namespace lookup hints */
270 #define LC_PREBIND_CKSUM 0x17 /* prebind checksum */
273 * load a dynamically linked shared library that is allowed to be missing
274 * (all symbols are weak imported).
276 #define LC_LOAD_WEAK_DYLIB (0x18 | LC_REQ_DYLD)
278 #define LC_SEGMENT_64 0x19 /* 64-bit segment of this file to be
280 #define LC_ROUTINES_64 0x1a /* 64-bit image routines */
281 #define LC_UUID 0x1b /* the uuid */
282 #define LC_RPATH (0x1c | LC_REQ_DYLD) /* runpath additions */
283 #define LC_CODE_SIGNATURE 0x1d /* local of code signature */
284 #define LC_SEGMENT_SPLIT_INFO 0x1e /* local of info to split segments */
285 #define LC_REEXPORT_DYLIB (0x1f | LC_REQ_DYLD) /* load and re-export dylib */
286 #define LC_LAZY_LOAD_DYLIB 0x20 /* delay load of dylib until first use */
287 #define LC_ENCRYPTION_INFO 0x21 /* encrypted segment information */
288 #define LC_DYLD_INFO 0x22 /* compressed dyld information */
289 #define LC_DYLD_INFO_ONLY (0x22|LC_REQ_DYLD) /* compressed dyld information only */
290 #define LC_LOAD_UPWARD_DYLIB (0x23 | LC_REQ_DYLD) /* load upward dylib */
291 #define LC_VERSION_MIN_MACOSX 0x24 /* build for MacOSX min OS version */
292 #define LC_VERSION_MIN_IPHONEOS 0x25 /* build for iPhoneOS min OS version */
293 #define LC_FUNCTION_STARTS 0x26 /* compressed table of function start addresses */
294 #define LC_DYLD_ENVIRONMENT 0x27 /* string for dyld to treat
295 like environment variable */
296 #define LC_MAIN (0x28|LC_REQ_DYLD) /* replacement for LC_UNIXTHREAD */
297 #define LC_DATA_IN_CODE 0x29 /* table of non-instructions in __text */
298 #define LC_SOURCE_VERSION 0x2A /* source version used to build binary */
299 #define LC_DYLIB_CODE_SIGN_DRS 0x2B /* Code signing DRs copied from linked dylibs */
300 #define LC_ENCRYPTION_INFO_64 0x2C /* 64-bit encrypted segment information */
301 #define LC_LINKER_OPTION 0x2D /* linker options in MH_OBJECT files */
302 #define LC_LINKER_OPTIMIZATION_HINT 0x2E /* optimization hints in MH_OBJECT files */
303 #define LC_VERSION_MIN_WATCHOS 0x30 /* build for Watch min OS version */
306 * A variable length string in a load command is represented by an lc_str
307 * union. The strings are stored just after the load command structure and
308 * the offset is from the start of the load command structure. The size
309 * of the string is reflected in the cmdsize field of the load command.
310 * Once again any padded bytes to bring the cmdsize field to a multiple
311 * of 4 bytes must be zero.
314 uint32_t offset
; /* offset to the string */
316 char *ptr
; /* pointer to the string */
321 * The segment load command indicates that a part of this file is to be
322 * mapped into the task's address space. The size of this segment in memory,
323 * vmsize, maybe equal to or larger than the amount to map from this file,
324 * filesize. The file is mapped starting at fileoff to the beginning of
325 * the segment in memory, vmaddr. The rest of the memory of the segment,
326 * if any, is allocated zero fill on demand. The segment's maximum virtual
327 * memory protection and initial virtual memory protection are specified
328 * by the maxprot and initprot fields. If the segment has sections then the
329 * section structures directly follow the segment command and their size is
330 * reflected in cmdsize.
332 struct segment_command
{ /* for 32-bit architectures */
333 uint32_t cmd
; /* LC_SEGMENT */
334 uint32_t cmdsize
; /* includes sizeof section structs */
335 char segname
[16]; /* segment name */
336 uint32_t vmaddr
; /* memory address of this segment */
337 uint32_t vmsize
; /* memory size of this segment */
338 uint32_t fileoff
; /* file offset of this segment */
339 uint32_t filesize
; /* amount to map from the file */
340 vm_prot_t maxprot
; /* maximum VM protection */
341 vm_prot_t initprot
; /* initial VM protection */
342 uint32_t nsects
; /* number of sections in segment */
343 uint32_t flags
; /* flags */
347 * The 64-bit segment load command indicates that a part of this file is to be
348 * mapped into a 64-bit task's address space. If the 64-bit segment has
349 * sections then section_64 structures directly follow the 64-bit segment
350 * command and their size is reflected in cmdsize.
352 struct segment_command_64
{ /* for 64-bit architectures */
353 uint32_t cmd
; /* LC_SEGMENT_64 */
354 uint32_t cmdsize
; /* includes sizeof section_64 structs */
355 char segname
[16]; /* segment name */
356 uint64_t vmaddr
; /* memory address of this segment */
357 uint64_t vmsize
; /* memory size of this segment */
358 uint64_t fileoff
; /* file offset of this segment */
359 uint64_t filesize
; /* amount to map from the file */
360 vm_prot_t maxprot
; /* maximum VM protection */
361 vm_prot_t initprot
; /* initial VM protection */
362 uint32_t nsects
; /* number of sections in segment */
363 uint32_t flags
; /* flags */
366 /* Constants for the flags field of the segment_command */
367 #define SG_HIGHVM 0x1 /* the file contents for this segment is for
368 the high part of the VM space, the low part
369 is zero filled (for stacks in core files) */
370 #define SG_FVMLIB 0x2 /* this segment is the VM that is allocated by
371 a fixed VM library, for overlap checking in
373 #define SG_NORELOC 0x4 /* this segment has nothing that was relocated
374 in it and nothing relocated to it, that is
375 it maybe safely replaced without relocation*/
376 #define SG_PROTECTED_VERSION_1 0x8 /* This segment is protected. If the
377 segment starts at file offset 0, the
378 first page of the segment is not
379 protected. All other pages of the
380 segment are protected. */
383 * A segment is made up of zero or more sections. Non-MH_OBJECT files have
384 * all of their segments with the proper sections in each, and padded to the
385 * specified segment alignment when produced by the link editor. The first
386 * segment of a MH_EXECUTE and MH_FVMLIB format file contains the mach_header
387 * and load commands of the object file before its first section. The zero
388 * fill sections are always last in their segment (in all formats). This
389 * allows the zeroed segment padding to be mapped into memory where zero fill
390 * sections might be. The gigabyte zero fill sections, those with the section
391 * type S_GB_ZEROFILL, can only be in a segment with sections of this type.
392 * These segments are then placed after all other segments.
394 * The MH_OBJECT format has all of its sections in one segment for
395 * compactness. There is no padding to a specified segment boundary and the
396 * mach_header and load commands are not part of the segment.
398 * Sections with the same section name, sectname, going into the same segment,
399 * segname, are combined by the link editor. The resulting section is aligned
400 * to the maximum alignment of the combined sections and is the new section's
401 * alignment. The combined sections are aligned to their original alignment in
402 * the combined section. Any padded bytes to get the specified alignment are
405 * The format of the relocation entries referenced by the reloff and nreloc
406 * fields of the section structure for mach object files is described in the
407 * header file <reloc.h>.
409 struct section
{ /* for 32-bit architectures */
410 char sectname
[16]; /* name of this section */
411 char segname
[16]; /* segment this section goes in */
412 uint32_t addr
; /* memory address of this section */
413 uint32_t size
; /* size in bytes of this section */
414 uint32_t offset
; /* file offset of this section */
415 uint32_t align
; /* section alignment (power of 2) */
416 uint32_t reloff
; /* file offset of relocation entries */
417 uint32_t nreloc
; /* number of relocation entries */
418 uint32_t flags
; /* flags (section type and attributes)*/
419 uint32_t reserved1
; /* reserved (for offset or index) */
420 uint32_t reserved2
; /* reserved (for count or sizeof) */
423 struct section_64
{ /* for 64-bit architectures */
424 char sectname
[16]; /* name of this section */
425 char segname
[16]; /* segment this section goes in */
426 uint64_t addr
; /* memory address of this section */
427 uint64_t size
; /* size in bytes of this section */
428 uint32_t offset
; /* file offset of this section */
429 uint32_t align
; /* section alignment (power of 2) */
430 uint32_t reloff
; /* file offset of relocation entries */
431 uint32_t nreloc
; /* number of relocation entries */
432 uint32_t flags
; /* flags (section type and attributes)*/
433 uint32_t reserved1
; /* reserved (for offset or index) */
434 uint32_t reserved2
; /* reserved (for count or sizeof) */
435 uint32_t reserved3
; /* reserved */
439 * The flags field of a section structure is separated into two parts a section
440 * type and section attributes. The section types are mutually exclusive (it
441 * can only have one type) but the section attributes are not (it may have more
442 * than one attribute).
444 #define SECTION_TYPE 0x000000ff /* 256 section types */
445 #define SECTION_ATTRIBUTES 0xffffff00 /* 24 section attributes */
447 /* Constants for the type of a section */
448 #define S_REGULAR 0x0 /* regular section */
449 #define S_ZEROFILL 0x1 /* zero fill on demand section */
450 #define S_CSTRING_LITERALS 0x2 /* section with only literal C strings*/
451 #define S_4BYTE_LITERALS 0x3 /* section with only 4 byte literals */
452 #define S_8BYTE_LITERALS 0x4 /* section with only 8 byte literals */
453 #define S_LITERAL_POINTERS 0x5 /* section with only pointers to */
456 * For the two types of symbol pointers sections and the symbol stubs section
457 * they have indirect symbol table entries. For each of the entries in the
458 * section the indirect symbol table entries, in corresponding order in the
459 * indirect symbol table, start at the index stored in the reserved1 field
460 * of the section structure. Since the indirect symbol table entries
461 * correspond to the entries in the section the number of indirect symbol table
462 * entries is inferred from the size of the section divided by the size of the
463 * entries in the section. For symbol pointers sections the size of the entries
464 * in the section is 4 bytes and for symbol stubs sections the byte size of the
465 * stubs is stored in the reserved2 field of the section structure.
467 #define S_NON_LAZY_SYMBOL_POINTERS 0x6 /* section with only non-lazy
469 #define S_LAZY_SYMBOL_POINTERS 0x7 /* section with only lazy symbol
471 #define S_SYMBOL_STUBS 0x8 /* section with only symbol
472 stubs, byte size of stub in
473 the reserved2 field */
474 #define S_MOD_INIT_FUNC_POINTERS 0x9 /* section with only function
475 pointers for initialization*/
476 #define S_MOD_TERM_FUNC_POINTERS 0xa /* section with only function
477 pointers for termination */
478 #define S_COALESCED 0xb /* section contains symbols that
479 are to be coalesced */
480 #define S_GB_ZEROFILL 0xc /* zero fill on demand section
481 (that can be larger than 4
483 #define S_INTERPOSING 0xd /* section with only pairs of
484 function pointers for
486 #define S_16BYTE_LITERALS 0xe /* section with only 16 byte
488 #define S_DTRACE_DOF 0xf /* section contains
489 DTrace Object Format */
490 #define S_LAZY_DYLIB_SYMBOL_POINTERS 0x10 /* section with only lazy
491 symbol pointers to lazy
494 * Section types to support thread local variables
496 #define S_THREAD_LOCAL_REGULAR 0x11 /* template of initial
498 #define S_THREAD_LOCAL_ZEROFILL 0x12 /* template of initial
500 #define S_THREAD_LOCAL_VARIABLES 0x13 /* TLV descriptors */
501 #define S_THREAD_LOCAL_VARIABLE_POINTERS 0x14 /* pointers to TLV
503 #define S_THREAD_LOCAL_INIT_FUNCTION_POINTERS 0x15 /* functions to call
508 * Constants for the section attributes part of the flags field of a section
511 #define SECTION_ATTRIBUTES_USR 0xff000000 /* User setable attributes */
512 #define S_ATTR_PURE_INSTRUCTIONS 0x80000000 /* section contains only true
513 machine instructions */
514 #define S_ATTR_NO_TOC 0x40000000 /* section contains coalesced
515 symbols that are not to be
518 #define S_ATTR_STRIP_STATIC_SYMS 0x20000000 /* ok to strip static symbols
519 in this section in files
520 with the MH_DYLDLINK flag */
521 #define S_ATTR_NO_DEAD_STRIP 0x10000000 /* no dead stripping */
522 #define S_ATTR_LIVE_SUPPORT 0x08000000 /* blocks are live if they
523 reference live blocks */
524 #define S_ATTR_SELF_MODIFYING_CODE 0x04000000 /* Used with i386 code stubs
525 written on by dyld */
527 * If a segment contains any sections marked with S_ATTR_DEBUG then all
528 * sections in that segment must have this attribute. No section other than
529 * a section marked with this attribute may reference the contents of this
530 * section. A section with this attribute may contain no symbols and must have
531 * a section type S_REGULAR. The static linker will not copy section contents
532 * from sections with this attribute into its output file. These sections
533 * generally contain DWARF debugging info.
535 #define S_ATTR_DEBUG 0x02000000 /* a debug section */
536 #define SECTION_ATTRIBUTES_SYS 0x00ffff00 /* system setable attributes */
537 #define S_ATTR_SOME_INSTRUCTIONS 0x00000400 /* section contains some
538 machine instructions */
539 #define S_ATTR_EXT_RELOC 0x00000200 /* section has external
540 relocation entries */
541 #define S_ATTR_LOC_RELOC 0x00000100 /* section has local
542 relocation entries */
546 * The names of segments and sections in them are mostly meaningless to the
547 * link-editor. But there are few things to support traditional UNIX
548 * executables that require the link-editor and assembler to use some names
549 * agreed upon by convention.
551 * The initial protection of the "__TEXT" segment has write protection turned
552 * off (not writeable).
554 * The link-editor will allocate common symbols at the end of the "__common"
555 * section in the "__DATA" segment. It will create the section and segment
559 /* The currently known segment names and the section names in those segments */
561 #define SEG_PAGEZERO "__PAGEZERO" /* the pagezero segment which has no */
562 /* protections and catches NULL */
563 /* references for MH_EXECUTE files */
566 #define SEG_TEXT "__TEXT" /* the tradition UNIX text segment */
567 #define SECT_TEXT "__text" /* the real text part of the text */
568 /* section no headers, and no padding */
569 #define SECT_FVMLIB_INIT0 "__fvmlib_init0" /* the fvmlib initialization */
571 #define SECT_FVMLIB_INIT1 "__fvmlib_init1" /* the section following the */
572 /* fvmlib initialization */
575 #define SEG_DATA "__DATA" /* the tradition UNIX data segment */
576 #define SECT_DATA "__data" /* the real initialized data section */
577 /* no padding, no bss overlap */
578 #define SECT_BSS "__bss" /* the real uninitialized data section*/
580 #define SECT_COMMON "__common" /* the section common symbols are */
581 /* allocated in by the link editor */
583 #define SEG_OBJC "__OBJC" /* objective-C runtime segment */
584 #define SECT_OBJC_SYMBOLS "__symbol_table" /* symbol table */
585 #define SECT_OBJC_MODULES "__module_info" /* module information */
586 #define SECT_OBJC_STRINGS "__selector_strs" /* string table */
587 #define SECT_OBJC_REFS "__selector_refs" /* string table */
589 #define SEG_ICON "__ICON" /* the icon segment */
590 #define SECT_ICON_HEADER "__header" /* the icon headers */
591 #define SECT_ICON_TIFF "__tiff" /* the icons in tiff format */
593 #define SEG_LINKEDIT "__LINKEDIT" /* the segment containing all structs */
594 /* created and maintained by the link */
595 /* editor. Created with -seglinkedit */
596 /* option to ld(1) for MH_EXECUTE and */
597 /* FVMLIB file types only */
599 #define SEG_UNIXSTACK "__UNIXSTACK" /* the unix stack segment */
601 #define SEG_IMPORT "__IMPORT" /* the segment for the self (dyld) */
602 /* modifing code stubs that has read, */
603 /* write and execute permissions */
606 * Fixed virtual memory shared libraries are identified by two things. The
607 * target pathname (the name of the library as found for execution), and the
608 * minor version number. The address of where the headers are loaded is in
609 * header_addr. (THIS IS OBSOLETE and no longer supported).
612 union lc_str name
; /* library's target pathname */
613 uint32_t minor_version
; /* library's minor version number */
614 uint32_t header_addr
; /* library's header address */
618 * A fixed virtual shared library (filetype == MH_FVMLIB in the mach header)
619 * contains a fvmlib_command (cmd == LC_IDFVMLIB) to identify the library.
620 * An object that uses a fixed virtual shared library also contains a
621 * fvmlib_command (cmd == LC_LOADFVMLIB) for each library it uses.
622 * (THIS IS OBSOLETE and no longer supported).
624 struct fvmlib_command
{
625 uint32_t cmd
; /* LC_IDFVMLIB or LC_LOADFVMLIB */
626 uint32_t cmdsize
; /* includes pathname string */
627 struct fvmlib fvmlib
; /* the library identification */
631 * Dynamicly linked shared libraries are identified by two things. The
632 * pathname (the name of the library as found for execution), and the
633 * compatibility version number. The pathname must match and the compatibility
634 * number in the user of the library must be greater than or equal to the
635 * library being used. The time stamp is used to record the time a library was
636 * built and copied into user so it can be use to determined if the library used
637 * at runtime is exactly the same as used to built the program.
640 union lc_str name
; /* library's path name */
641 uint32_t timestamp
; /* library's build time stamp */
642 uint32_t current_version
; /* library's current version number */
643 uint32_t compatibility_version
; /* library's compatibility vers number*/
647 * A dynamically linked shared library (filetype == MH_DYLIB in the mach header)
648 * contains a dylib_command (cmd == LC_ID_DYLIB) to identify the library.
649 * An object that uses a dynamically linked shared library also contains a
650 * dylib_command (cmd == LC_LOAD_DYLIB, LC_LOAD_WEAK_DYLIB, or
651 * LC_REEXPORT_DYLIB) for each library it uses.
653 struct dylib_command
{
654 uint32_t cmd
; /* LC_ID_DYLIB, LC_LOAD_{,WEAK_}DYLIB,
656 uint32_t cmdsize
; /* includes pathname string */
657 struct dylib dylib
; /* the library identification */
661 * A dynamically linked shared library may be a subframework of an umbrella
662 * framework. If so it will be linked with "-umbrella umbrella_name" where
663 * Where "umbrella_name" is the name of the umbrella framework. A subframework
664 * can only be linked against by its umbrella framework or other subframeworks
665 * that are part of the same umbrella framework. Otherwise the static link
666 * editor produces an error and states to link against the umbrella framework.
667 * The name of the umbrella framework for subframeworks is recorded in the
668 * following structure.
670 struct sub_framework_command
{
671 uint32_t cmd
; /* LC_SUB_FRAMEWORK */
672 uint32_t cmdsize
; /* includes umbrella string */
673 union lc_str umbrella
; /* the umbrella framework name */
677 * For dynamically linked shared libraries that are subframework of an umbrella
678 * framework they can allow clients other than the umbrella framework or other
679 * subframeworks in the same umbrella framework. To do this the subframework
680 * is built with "-allowable_client client_name" and an LC_SUB_CLIENT load
681 * command is created for each -allowable_client flag. The client_name is
682 * usually a framework name. It can also be a name used for bundles clients
683 * where the bundle is built with "-client_name client_name".
685 struct sub_client_command
{
686 uint32_t cmd
; /* LC_SUB_CLIENT */
687 uint32_t cmdsize
; /* includes client string */
688 union lc_str client
; /* the client name */
692 * A dynamically linked shared library may be a sub_umbrella of an umbrella
693 * framework. If so it will be linked with "-sub_umbrella umbrella_name" where
694 * Where "umbrella_name" is the name of the sub_umbrella framework. When
695 * staticly linking when -twolevel_namespace is in effect a twolevel namespace
696 * umbrella framework will only cause its subframeworks and those frameworks
697 * listed as sub_umbrella frameworks to be implicited linked in. Any other
698 * dependent dynamic libraries will not be linked it when -twolevel_namespace
699 * is in effect. The primary library recorded by the static linker when
700 * resolving a symbol in these libraries will be the umbrella framework.
701 * Zero or more sub_umbrella frameworks may be use by an umbrella framework.
702 * The name of a sub_umbrella framework is recorded in the following structure.
704 struct sub_umbrella_command
{
705 uint32_t cmd
; /* LC_SUB_UMBRELLA */
706 uint32_t cmdsize
; /* includes sub_umbrella string */
707 union lc_str sub_umbrella
; /* the sub_umbrella framework name */
711 * A dynamically linked shared library may be a sub_library of another shared
712 * library. If so it will be linked with "-sub_library library_name" where
713 * Where "library_name" is the name of the sub_library shared library. When
714 * staticly linking when -twolevel_namespace is in effect a twolevel namespace
715 * shared library will only cause its subframeworks and those frameworks
716 * listed as sub_umbrella frameworks and libraries listed as sub_libraries to
717 * be implicited linked in. Any other dependent dynamic libraries will not be
718 * linked it when -twolevel_namespace is in effect. The primary library
719 * recorded by the static linker when resolving a symbol in these libraries
720 * will be the umbrella framework (or dynamic library). Zero or more sub_library
721 * shared libraries may be use by an umbrella framework or (or dynamic library).
722 * The name of a sub_library framework is recorded in the following structure.
723 * For example /usr/lib/libobjc_profile.A.dylib would be recorded as "libobjc".
725 struct sub_library_command
{
726 uint32_t cmd
; /* LC_SUB_LIBRARY */
727 uint32_t cmdsize
; /* includes sub_library string */
728 union lc_str sub_library
; /* the sub_library name */
732 * A program (filetype == MH_EXECUTE) that is
733 * prebound to its dynamic libraries has one of these for each library that
734 * the static linker used in prebinding. It contains a bit vector for the
735 * modules in the library. The bits indicate which modules are bound (1) and
736 * which are not (0) from the library. The bit for module 0 is the low bit
737 * of the first byte. So the bit for the Nth module is:
738 * (linked_modules[N/8] >> N%8) & 1
740 struct prebound_dylib_command
{
741 uint32_t cmd
; /* LC_PREBOUND_DYLIB */
742 uint32_t cmdsize
; /* includes strings */
743 union lc_str name
; /* library's path name */
744 uint32_t nmodules
; /* number of modules in library */
745 union lc_str linked_modules
; /* bit vector of linked modules */
749 * A program that uses a dynamic linker contains a dylinker_command to identify
750 * the name of the dynamic linker (LC_LOAD_DYLINKER). And a dynamic linker
751 * contains a dylinker_command to identify the dynamic linker (LC_ID_DYLINKER).
752 * A file can have at most one of these.
753 * This struct is also used for the LC_DYLD_ENVIRONMENT load command and
754 * contains string for dyld to treat like environment variable.
756 struct dylinker_command
{
757 uint32_t cmd
; /* LC_ID_DYLINKER, LC_LOAD_DYLINKER or
758 LC_DYLD_ENVIRONMENT */
759 uint32_t cmdsize
; /* includes pathname string */
760 union lc_str name
; /* dynamic linker's path name */
764 * Thread commands contain machine-specific data structures suitable for
765 * use in the thread state primitives. The machine specific data structures
766 * follow the struct thread_command as follows.
767 * Each flavor of machine specific data structure is preceded by an unsigned
768 * long constant for the flavor of that data structure, an uint32_t
769 * that is the count of longs of the size of the state data structure and then
770 * the state data structure follows. This triple may be repeated for many
771 * flavors. The constants for the flavors, counts and state data structure
772 * definitions are expected to be in the header file <machine/thread_status.h>.
773 * These machine specific data structures sizes must be multiples of
774 * 4 bytes The cmdsize reflects the total size of the thread_command
775 * and all of the sizes of the constants for the flavors, counts and state
778 * For executable objects that are unix processes there will be one
779 * thread_command (cmd == LC_UNIXTHREAD) created for it by the link-editor.
780 * This is the same as a LC_THREAD, except that a stack is automatically
781 * created (based on the shell's limit for the stack size). Command arguments
782 * and environment variables are copied onto that stack.
784 struct thread_command
{
785 uint32_t cmd
; /* LC_THREAD or LC_UNIXTHREAD */
786 uint32_t cmdsize
; /* total size of this command */
787 /* uint32_t flavor flavor of thread state */
788 /* uint32_t count count of longs in thread state */
789 /* struct XXX_thread_state state thread state for this flavor */
794 * The routines command contains the address of the dynamic shared library
795 * initialization routine and an index into the module table for the module
796 * that defines the routine. Before any modules are used from the library the
797 * dynamic linker fully binds the module that defines the initialization routine
798 * and then calls it. This gets called before any module initialization
799 * routines (used for C++ static constructors) in the library.
801 struct routines_command
{ /* for 32-bit architectures */
802 uint32_t cmd
; /* LC_ROUTINES */
803 uint32_t cmdsize
; /* total size of this command */
804 uint32_t init_address
; /* address of initialization routine */
805 uint32_t init_module
; /* index into the module table that */
806 /* the init routine is defined in */
816 * The 64-bit routines command. Same use as above.
818 struct routines_command_64
{ /* for 64-bit architectures */
819 uint32_t cmd
; /* LC_ROUTINES_64 */
820 uint32_t cmdsize
; /* total size of this command */
821 uint64_t init_address
; /* address of initialization routine */
822 uint64_t init_module
; /* index into the module table that */
823 /* the init routine is defined in */
833 * The symtab_command contains the offsets and sizes of the link-edit 4.3BSD
834 * "stab" style symbol table information as described in the header files
835 * <nlist.h> and <stab.h>.
837 struct symtab_command
{
838 uint32_t cmd
; /* LC_SYMTAB */
839 uint32_t cmdsize
; /* sizeof(struct symtab_command) */
840 uint32_t symoff
; /* symbol table offset */
841 uint32_t nsyms
; /* number of symbol table entries */
842 uint32_t stroff
; /* string table offset */
843 uint32_t strsize
; /* string table size in bytes */
847 * This is the second set of the symbolic information which is used to support
848 * the data structures for the dynamically link editor.
850 * The original set of symbolic information in the symtab_command which contains
851 * the symbol and string tables must also be present when this load command is
852 * present. When this load command is present the symbol table is organized
853 * into three groups of symbols:
854 * local symbols (static and debugging symbols) - grouped by module
855 * defined external symbols - grouped by module (sorted by name if not lib)
856 * undefined external symbols (sorted by name if MH_BINDATLOAD is not set,
857 * and in order the were seen by the static
858 * linker if MH_BINDATLOAD is set)
859 * In this load command there are offsets and counts to each of the three groups
862 * This load command contains a the offsets and sizes of the following new
863 * symbolic information tables:
866 * reference symbol table
867 * indirect symbol table
868 * The first three tables above (the table of contents, module table and
869 * reference symbol table) are only present if the file is a dynamically linked
870 * shared library. For executable and object modules, which are files
871 * containing only one module, the information that would be in these three
872 * tables is determined as follows:
873 * table of contents - the defined external symbols are sorted by name
874 * module table - the file contains only one module so everything in the
875 * file is part of the module.
876 * reference symbol table - is the defined and undefined external symbols
878 * For dynamically linked shared library files this load command also contains
879 * offsets and sizes to the pool of relocation entries for all sections
880 * separated into two groups:
881 * external relocation entries
882 * local relocation entries
883 * For executable and object modules the relocation entries continue to hang
884 * off the section structures.
886 struct dysymtab_command
{
887 uint32_t cmd
; /* LC_DYSYMTAB */
888 uint32_t cmdsize
; /* sizeof(struct dysymtab_command) */
891 * The symbols indicated by symoff and nsyms of the LC_SYMTAB load command
892 * are grouped into the following three groups:
893 * local symbols (further grouped by the module they are from)
894 * defined external symbols (further grouped by the module they are from)
897 * The local symbols are used only for debugging. The dynamic binding
898 * process may have to use them to indicate to the debugger the local
899 * symbols for a module that is being bound.
901 * The last two groups are used by the dynamic binding process to do the
902 * binding (indirectly through the module table and the reference symbol
903 * table when this is a dynamically linked shared library file).
905 uint32_t ilocalsym
; /* index to local symbols */
906 uint32_t nlocalsym
; /* number of local symbols */
908 uint32_t iextdefsym
;/* index to externally defined symbols */
909 uint32_t nextdefsym
;/* number of externally defined symbols */
911 uint32_t iundefsym
; /* index to undefined symbols */
912 uint32_t nundefsym
; /* number of undefined symbols */
915 * For the for the dynamic binding process to find which module a symbol
916 * is defined in the table of contents is used (analogous to the ranlib
917 * structure in an archive) which maps defined external symbols to modules
918 * they are defined in. This exists only in a dynamically linked shared
919 * library file. For executable and object modules the defined external
920 * symbols are sorted by name and is use as the table of contents.
922 uint32_t tocoff
; /* file offset to table of contents */
923 uint32_t ntoc
; /* number of entries in table of contents */
926 * To support dynamic binding of "modules" (whole object files) the symbol
927 * table must reflect the modules that the file was created from. This is
928 * done by having a module table that has indexes and counts into the merged
929 * tables for each module. The module structure that these two entries
930 * refer to is described below. This exists only in a dynamically linked
931 * shared library file. For executable and object modules the file only
932 * contains one module so everything in the file belongs to the module.
934 uint32_t modtaboff
; /* file offset to module table */
935 uint32_t nmodtab
; /* number of module table entries */
938 * To support dynamic module binding the module structure for each module
939 * indicates the external references (defined and undefined) each module
940 * makes. For each module there is an offset and a count into the
941 * reference symbol table for the symbols that the module references.
942 * This exists only in a dynamically linked shared library file. For
943 * executable and object modules the defined external symbols and the
944 * undefined external symbols indicates the external references.
946 uint32_t extrefsymoff
; /* offset to referenced symbol table */
947 uint32_t nextrefsyms
; /* number of referenced symbol table entries */
950 * The sections that contain "symbol pointers" and "routine stubs" have
951 * indexes and (implied counts based on the size of the section and fixed
952 * size of the entry) into the "indirect symbol" table for each pointer
953 * and stub. For every section of these two types the index into the
954 * indirect symbol table is stored in the section header in the field
955 * reserved1. An indirect symbol table entry is simply a 32bit index into
956 * the symbol table to the symbol that the pointer or stub is referring to.
957 * The indirect symbol table is ordered to match the entries in the section.
959 uint32_t indirectsymoff
; /* file offset to the indirect symbol table */
960 uint32_t nindirectsyms
; /* number of indirect symbol table entries */
963 * To support relocating an individual module in a library file quickly the
964 * external relocation entries for each module in the library need to be
965 * accessed efficiently. Since the relocation entries can't be accessed
966 * through the section headers for a library file they are separated into
967 * groups of local and external entries further grouped by module. In this
968 * case the presents of this load command who's extreloff, nextrel,
969 * locreloff and nlocrel fields are non-zero indicates that the relocation
970 * entries of non-merged sections are not referenced through the section
971 * structures (and the reloff and nreloc fields in the section headers are
974 * Since the relocation entries are not accessed through the section headers
975 * this requires the r_address field to be something other than a section
976 * offset to identify the item to be relocated. In this case r_address is
977 * set to the offset from the vmaddr of the first LC_SEGMENT command.
978 * For MH_SPLIT_SEGS images r_address is set to the the offset from the
979 * vmaddr of the first read-write LC_SEGMENT command.
981 * The relocation entries are grouped by module and the module table
982 * entries have indexes and counts into them for the group of external
983 * relocation entries for that the module.
985 * For sections that are merged across modules there must not be any
986 * remaining external relocation entries for them (for merged sections
987 * remaining relocation entries must be local).
989 uint32_t extreloff
; /* offset to external relocation entries */
990 uint32_t nextrel
; /* number of external relocation entries */
993 * All the local relocation entries are grouped together (they are not
994 * grouped by their module since they are only used if the object is moved
995 * from it staticly link edited address).
997 uint32_t locreloff
; /* offset to local relocation entries */
998 uint32_t nlocrel
; /* number of local relocation entries */
1003 * An indirect symbol table entry is simply a 32bit index into the symbol table
1004 * to the symbol that the pointer or stub is refering to. Unless it is for a
1005 * non-lazy symbol pointer section for a defined symbol which strip(1) as
1006 * removed. In which case it has the value INDIRECT_SYMBOL_LOCAL. If the
1007 * symbol was also absolute INDIRECT_SYMBOL_ABS is or'ed with that.
1009 #define INDIRECT_SYMBOL_LOCAL 0x80000000
1010 #define INDIRECT_SYMBOL_ABS 0x40000000
1013 /* a table of contents entry */
1014 struct dylib_table_of_contents
{
1015 uint32_t symbol_index
; /* the defined external symbol
1016 (index into the symbol table) */
1017 uint32_t module_index
; /* index into the module table this symbol
1021 /* a module table entry */
1022 struct dylib_module
{
1023 uint32_t module_name
; /* the module name (index into string table) */
1025 uint32_t iextdefsym
; /* index into externally defined symbols */
1026 uint32_t nextdefsym
; /* number of externally defined symbols */
1027 uint32_t irefsym
; /* index into reference symbol table */
1028 uint32_t nrefsym
; /* number of reference symbol table entries */
1029 uint32_t ilocalsym
; /* index into symbols for local symbols */
1030 uint32_t nlocalsym
; /* number of local symbols */
1032 uint32_t iextrel
; /* index into external relocation entries */
1033 uint32_t nextrel
; /* number of external relocation entries */
1035 uint32_t iinit_iterm
; /* low 16 bits are the index into the init
1036 section, high 16 bits are the index into
1038 uint32_t ninit_nterm
; /* low 16 bits are the number of init section
1039 entries, high 16 bits are the number of
1040 term section entries */
1042 uint32_t /* for this module address of the start of */
1043 objc_module_info_addr
; /* the (__OBJC,__module_info) section */
1044 uint32_t /* for this module size of */
1045 objc_module_info_size
; /* the (__OBJC,__module_info) section */
1048 /* a 64-bit module table entry */
1049 struct dylib_module_64
{
1050 uint32_t module_name
; /* the module name (index into string table) */
1052 uint32_t iextdefsym
; /* index into externally defined symbols */
1053 uint32_t nextdefsym
; /* number of externally defined symbols */
1054 uint32_t irefsym
; /* index into reference symbol table */
1055 uint32_t nrefsym
; /* number of reference symbol table entries */
1056 uint32_t ilocalsym
; /* index into symbols for local symbols */
1057 uint32_t nlocalsym
; /* number of local symbols */
1059 uint32_t iextrel
; /* index into external relocation entries */
1060 uint32_t nextrel
; /* number of external relocation entries */
1062 uint32_t iinit_iterm
; /* low 16 bits are the index into the init
1063 section, high 16 bits are the index into
1065 uint32_t ninit_nterm
; /* low 16 bits are the number of init section
1066 entries, high 16 bits are the number of
1067 term section entries */
1069 uint32_t /* for this module size of */
1070 objc_module_info_size
; /* the (__OBJC,__module_info) section */
1071 uint64_t /* for this module address of the start of */
1072 objc_module_info_addr
; /* the (__OBJC,__module_info) section */
1076 * The entries in the reference symbol table are used when loading the module
1077 * (both by the static and dynamic link editors) and if the module is unloaded
1078 * or replaced. Therefore all external symbols (defined and undefined) are
1079 * listed in the module's reference table. The flags describe the type of
1080 * reference that is being made. The constants for the flags are defined in
1081 * <mach-o/nlist.h> as they are also used for symbol table entries.
1083 struct dylib_reference
{
1084 uint32_t isym
:24, /* index into the symbol table */
1085 flags
:8; /* flags to indicate the type of reference */
1089 * The twolevel_hints_command contains the offset and number of hints in the
1090 * two-level namespace lookup hints table.
1092 struct twolevel_hints_command
{
1093 uint32_t cmd
; /* LC_TWOLEVEL_HINTS */
1094 uint32_t cmdsize
; /* sizeof(struct twolevel_hints_command) */
1095 uint32_t offset
; /* offset to the hint table */
1096 uint32_t nhints
; /* number of hints in the hint table */
1100 * The entries in the two-level namespace lookup hints table are twolevel_hint
1101 * structs. These provide hints to the dynamic link editor where to start
1102 * looking for an undefined symbol in a two-level namespace image. The
1103 * isub_image field is an index into the sub-images (sub-frameworks and
1104 * sub-umbrellas list) that made up the two-level image that the undefined
1105 * symbol was found in when it was built by the static link editor. If
1106 * isub-image is 0 the the symbol is expected to be defined in library and not
1107 * in the sub-images. If isub-image is non-zero it is an index into the array
1108 * of sub-images for the umbrella with the first index in the sub-images being
1109 * 1. The array of sub-images is the ordered list of sub-images of the umbrella
1110 * that would be searched for a symbol that has the umbrella recorded as its
1111 * primary library. The table of contents index is an index into the
1112 * library's table of contents. This is used as the starting point of the
1113 * binary search or a directed linear search.
1115 struct twolevel_hint
{
1117 isub_image
:8, /* index into the sub images */
1118 itoc
:24; /* index into the table of contents */
1122 * The prebind_cksum_command contains the value of the original check sum for
1123 * prebound files or zero. When a prebound file is first created or modified
1124 * for other than updating its prebinding information the value of the check sum
1125 * is set to zero. When the file has it prebinding re-done and if the value of
1126 * the check sum is zero the original check sum is calculated and stored in
1127 * cksum field of this load command in the output file. If when the prebinding
1128 * is re-done and the cksum field is non-zero it is left unchanged from the
1131 struct prebind_cksum_command
{
1132 uint32_t cmd
; /* LC_PREBIND_CKSUM */
1133 uint32_t cmdsize
; /* sizeof(struct prebind_cksum_command) */
1134 uint32_t cksum
; /* the check sum or zero */
1138 * The uuid load command contains a single 128-bit unique random number that
1139 * identifies an object produced by the static link editor.
1141 struct uuid_command
{
1142 uint32_t cmd
; /* LC_UUID */
1143 uint32_t cmdsize
; /* sizeof(struct uuid_command) */
1144 uint8_t uuid
[16]; /* the 128-bit uuid */
1148 * The rpath_command contains a path which at runtime should be added to
1149 * the current run path used to find @rpath prefixed dylibs.
1151 struct rpath_command
{
1152 uint32_t cmd
; /* LC_RPATH */
1153 uint32_t cmdsize
; /* includes string */
1154 union lc_str path
; /* path to add to run path */
1158 * The linkedit_data_command contains the offsets and sizes of a blob
1159 * of data in the __LINKEDIT segment.
1161 struct linkedit_data_command
{
1162 uint32_t cmd
; /* LC_CODE_SIGNATURE, LC_SEGMENT_SPLIT_INFO,
1163 LC_FUNCTION_STARTS, LC_DATA_IN_CODE,
1164 LC_DYLIB_CODE_SIGN_DRS or
1165 LC_LINKER_OPTIMIZATION_HINT. */
1166 uint32_t cmdsize
; /* sizeof(struct linkedit_data_command) */
1167 uint32_t dataoff
; /* file offset of data in __LINKEDIT segment */
1168 uint32_t datasize
; /* file size of data in __LINKEDIT segment */
1172 * The encryption_info_command contains the file offset and size of an
1173 * of an encrypted segment.
1175 struct encryption_info_command
{
1176 uint32_t cmd
; /* LC_ENCRYPTION_INFO */
1177 uint32_t cmdsize
; /* sizeof(struct encryption_info_command) */
1178 uint32_t cryptoff
; /* file offset of encrypted range */
1179 uint32_t cryptsize
; /* file size of encrypted range */
1180 uint32_t cryptid
; /* which enryption system,
1181 0 means not-encrypted yet */
1185 * The encryption_info_command_64 contains the file offset and size of an
1186 * of an encrypted segment (for use in x86_64 targets).
1188 struct encryption_info_command_64
{
1189 uint32_t cmd
; /* LC_ENCRYPTION_INFO_64 */
1190 uint32_t cmdsize
; /* sizeof(struct encryption_info_command_64) */
1191 uint32_t cryptoff
; /* file offset of encrypted range */
1192 uint32_t cryptsize
; /* file size of encrypted range */
1193 uint32_t cryptid
; /* which enryption system,
1194 0 means not-encrypted yet */
1195 uint32_t pad
; /* padding to make this struct's size a multiple
1200 * The version_min_command contains the min OS version on which this
1201 * binary was built to run.
1203 struct version_min_command
{
1204 uint32_t cmd
; /* LC_VERSION_MIN_MACOSX or
1205 LC_VERSION_MIN_IPHONEOS
1206 LC_VERSION_MIN_WATCHOS */
1207 uint32_t cmdsize
; /* sizeof(struct min_version_command) */
1208 uint32_t version
; /* X.Y.Z is encoded in nibbles xxxx.yy.zz */
1209 uint32_t sdk
; /* X.Y.Z is encoded in nibbles xxxx.yy.zz */
1213 * The dyld_info_command contains the file offsets and sizes of
1214 * the new compressed form of the information dyld needs to
1215 * load the image. This information is used by dyld on Mac OS X
1216 * 10.6 and later. All information pointed to by this command
1217 * is encoded using byte streams, so no endian swapping is needed
1220 struct dyld_info_command
{
1221 uint32_t cmd
; /* LC_DYLD_INFO or LC_DYLD_INFO_ONLY */
1222 uint32_t cmdsize
; /* sizeof(struct dyld_info_command) */
1225 * Dyld rebases an image whenever dyld loads it at an address different
1226 * from its preferred address. The rebase information is a stream
1227 * of byte sized opcodes whose symbolic names start with REBASE_OPCODE_.
1228 * Conceptually the rebase information is a table of tuples:
1229 * <seg-index, seg-offset, type>
1230 * The opcodes are a compressed way to encode the table by only
1231 * encoding when a column changes. In addition simple patterns
1232 * like "every n'th offset for m times" can be encoded in a few
1235 uint32_t rebase_off
; /* file offset to rebase info */
1236 uint32_t rebase_size
; /* size of rebase info */
1239 * Dyld binds an image during the loading process, if the image
1240 * requires any pointers to be initialized to symbols in other images.
1241 * The bind information is a stream of byte sized
1242 * opcodes whose symbolic names start with BIND_OPCODE_.
1243 * Conceptually the bind information is a table of tuples:
1244 * <seg-index, seg-offset, type, symbol-library-ordinal, symbol-name, addend>
1245 * The opcodes are a compressed way to encode the table by only
1246 * encoding when a column changes. In addition simple patterns
1247 * like for runs of pointers initialzed to the same value can be
1248 * encoded in a few bytes.
1250 uint32_t bind_off
; /* file offset to binding info */
1251 uint32_t bind_size
; /* size of binding info */
1254 * Some C++ programs require dyld to unique symbols so that all
1255 * images in the process use the same copy of some code/data.
1256 * This step is done after binding. The content of the weak_bind
1257 * info is an opcode stream like the bind_info. But it is sorted
1258 * alphabetically by symbol name. This enable dyld to walk
1259 * all images with weak binding information in order and look
1260 * for collisions. If there are no collisions, dyld does
1261 * no updating. That means that some fixups are also encoded
1262 * in the bind_info. For instance, all calls to "operator new"
1263 * are first bound to libstdc++.dylib using the information
1264 * in bind_info. Then if some image overrides operator new
1265 * that is detected when the weak_bind information is processed
1266 * and the call to operator new is then rebound.
1268 uint32_t weak_bind_off
; /* file offset to weak binding info */
1269 uint32_t weak_bind_size
; /* size of weak binding info */
1272 * Some uses of external symbols do not need to be bound immediately.
1273 * Instead they can be lazily bound on first use. The lazy_bind
1274 * are contains a stream of BIND opcodes to bind all lazy symbols.
1275 * Normal use is that dyld ignores the lazy_bind section when
1276 * loading an image. Instead the static linker arranged for the
1277 * lazy pointer to initially point to a helper function which
1278 * pushes the offset into the lazy_bind area for the symbol
1279 * needing to be bound, then jumps to dyld which simply adds
1280 * the offset to lazy_bind_off to get the information on what
1283 uint32_t lazy_bind_off
; /* file offset to lazy binding info */
1284 uint32_t lazy_bind_size
; /* size of lazy binding infs */
1287 * The symbols exported by a dylib are encoded in a trie. This
1288 * is a compact representation that factors out common prefixes.
1289 * It also reduces LINKEDIT pages in RAM because it encodes all
1290 * information (name, address, flags) in one small, contiguous range.
1291 * The export area is a stream of nodes. The first node sequentially
1292 * is the start node for the trie.
1294 * Nodes for a symbol start with a uleb128 that is the length of
1295 * the exported symbol information for the string so far.
1296 * If there is no exported symbol, the node starts with a zero byte.
1297 * If there is exported info, it follows the length.
1299 * First is a uleb128 containing flags. Normally, it is followed by
1300 * a uleb128 encoded offset which is location of the content named
1301 * by the symbol from the mach_header for the image. If the flags
1302 * is EXPORT_SYMBOL_FLAGS_REEXPORT, then following the flags is
1303 * a uleb128 encoded library ordinal, then a zero terminated
1304 * UTF8 string. If the string is zero length, then the symbol
1305 * is re-export from the specified dylib with the same name.
1306 * If the flags is EXPORT_SYMBOL_FLAGS_STUB_AND_RESOLVER, then following
1307 * the flags is two uleb128s: the stub offset and the resolver offset.
1308 * The stub is used by non-lazy pointers. The resolver is used
1309 * by lazy pointers and must be called to get the actual address to use.
1311 * After the optional exported symbol information is a byte of
1312 * how many edges (0-255) that this node has leaving it,
1313 * followed by each edge.
1314 * Each edge is a zero terminated UTF8 of the addition chars
1315 * in the symbol, followed by a uleb128 offset for the node that
1319 uint32_t export_off
; /* file offset to lazy binding info */
1320 uint32_t export_size
; /* size of lazy binding infs */
1324 * The following are used to encode rebasing information
1326 #define REBASE_TYPE_POINTER 1
1327 #define REBASE_TYPE_TEXT_ABSOLUTE32 2
1328 #define REBASE_TYPE_TEXT_PCREL32 3
1330 #define REBASE_OPCODE_MASK 0xF0
1331 #define REBASE_IMMEDIATE_MASK 0x0F
1332 #define REBASE_OPCODE_DONE 0x00
1333 #define REBASE_OPCODE_SET_TYPE_IMM 0x10
1334 #define REBASE_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB 0x20
1335 #define REBASE_OPCODE_ADD_ADDR_ULEB 0x30
1336 #define REBASE_OPCODE_ADD_ADDR_IMM_SCALED 0x40
1337 #define REBASE_OPCODE_DO_REBASE_IMM_TIMES 0x50
1338 #define REBASE_OPCODE_DO_REBASE_ULEB_TIMES 0x60
1339 #define REBASE_OPCODE_DO_REBASE_ADD_ADDR_ULEB 0x70
1340 #define REBASE_OPCODE_DO_REBASE_ULEB_TIMES_SKIPPING_ULEB 0x80
1344 * The following are used to encode binding information
1346 #define BIND_TYPE_POINTER 1
1347 #define BIND_TYPE_TEXT_ABSOLUTE32 2
1348 #define BIND_TYPE_TEXT_PCREL32 3
1350 #define BIND_SPECIAL_DYLIB_SELF 0
1351 #define BIND_SPECIAL_DYLIB_MAIN_EXECUTABLE -1
1352 #define BIND_SPECIAL_DYLIB_FLAT_LOOKUP -2
1354 #define BIND_SYMBOL_FLAGS_WEAK_IMPORT 0x1
1355 #define BIND_SYMBOL_FLAGS_NON_WEAK_DEFINITION 0x8
1357 #define BIND_OPCODE_MASK 0xF0
1358 #define BIND_IMMEDIATE_MASK 0x0F
1359 #define BIND_OPCODE_DONE 0x00
1360 #define BIND_OPCODE_SET_DYLIB_ORDINAL_IMM 0x10
1361 #define BIND_OPCODE_SET_DYLIB_ORDINAL_ULEB 0x20
1362 #define BIND_OPCODE_SET_DYLIB_SPECIAL_IMM 0x30
1363 #define BIND_OPCODE_SET_SYMBOL_TRAILING_FLAGS_IMM 0x40
1364 #define BIND_OPCODE_SET_TYPE_IMM 0x50
1365 #define BIND_OPCODE_SET_ADDEND_SLEB 0x60
1366 #define BIND_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB 0x70
1367 #define BIND_OPCODE_ADD_ADDR_ULEB 0x80
1368 #define BIND_OPCODE_DO_BIND 0x90
1369 #define BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB 0xA0
1370 #define BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED 0xB0
1371 #define BIND_OPCODE_DO_BIND_ULEB_TIMES_SKIPPING_ULEB 0xC0
1375 * The following are used on the flags byte of a terminal node
1376 * in the export information.
1378 #define EXPORT_SYMBOL_FLAGS_KIND_MASK 0x03
1379 #define EXPORT_SYMBOL_FLAGS_KIND_REGULAR 0x00
1380 #define EXPORT_SYMBOL_FLAGS_KIND_THREAD_LOCAL 0x01
1381 #define EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION 0x04
1382 #define EXPORT_SYMBOL_FLAGS_REEXPORT 0x08
1383 #define EXPORT_SYMBOL_FLAGS_STUB_AND_RESOLVER 0x10
1386 * The linker_option_command contains linker options embedded in object files.
1388 struct linker_option_command
{
1389 uint32_t cmd
; /* LC_LINKER_OPTION only used in MH_OBJECT filetypes */
1391 uint32_t count
; /* number of strings */
1392 /* concatenation of zero terminated UTF8 strings.
1393 Zero filled at end to align */
1397 * The symseg_command contains the offset and size of the GNU style
1398 * symbol table information as described in the header file <symseg.h>.
1399 * The symbol roots of the symbol segments must also be aligned properly
1400 * in the file. So the requirement of keeping the offsets aligned to a
1401 * multiple of a 4 bytes translates to the length field of the symbol
1402 * roots also being a multiple of a long. Also the padding must again be
1403 * zeroed. (THIS IS OBSOLETE and no longer supported).
1405 struct symseg_command
{
1406 uint32_t cmd
; /* LC_SYMSEG */
1407 uint32_t cmdsize
; /* sizeof(struct symseg_command) */
1408 uint32_t offset
; /* symbol segment offset */
1409 uint32_t size
; /* symbol segment size in bytes */
1413 * The ident_command contains a free format string table following the
1414 * ident_command structure. The strings are null terminated and the size of
1415 * the command is padded out with zero bytes to a multiple of 4 bytes/
1416 * (THIS IS OBSOLETE and no longer supported).
1418 struct ident_command
{
1419 uint32_t cmd
; /* LC_IDENT */
1420 uint32_t cmdsize
; /* strings that follow this command */
1424 * The fvmfile_command contains a reference to a file to be loaded at the
1425 * specified virtual address. (Presently, this command is reserved for
1426 * internal use. The kernel ignores this command when loading a program into
1429 struct fvmfile_command
{
1430 uint32_t cmd
; /* LC_FVMFILE */
1431 uint32_t cmdsize
; /* includes pathname string */
1432 union lc_str name
; /* files pathname */
1433 uint32_t header_addr
; /* files virtual address */
1438 * The entry_point_command is a replacement for thread_command.
1439 * It is used for main executables to specify the location (file offset)
1440 * of main(). If -stack_size was used at link time, the stacksize
1441 * field will contain the stack size need for the main thread.
1443 struct entry_point_command
{
1444 uint32_t cmd
; /* LC_MAIN only used in MH_EXECUTE filetypes */
1445 uint32_t cmdsize
; /* 24 */
1446 uint64_t entryoff
; /* file (__TEXT) offset of main() */
1447 uint64_t stacksize
;/* if not zero, initial stack size */
1452 * The source_version_command is an optional load command containing
1453 * the version of the sources used to build the binary.
1455 struct source_version_command
{
1456 uint32_t cmd
; /* LC_SOURCE_VERSION */
1457 uint32_t cmdsize
; /* 16 */
1458 uint64_t version
; /* A.B.C.D.E packed as a24.b10.c10.d10.e10 */
1463 * The LC_DATA_IN_CODE load commands uses a linkedit_data_command
1464 * to point to an array of data_in_code_entry entries. Each entry
1465 * describes a range of data in a code section.
1467 struct data_in_code_entry
{
1468 uint32_t offset
; /* from mach_header to start of data range*/
1469 uint16_t length
; /* number of bytes in data range */
1470 uint16_t kind
; /* a DICE_KIND_* value */
1472 #define DICE_KIND_DATA 0x0001
1473 #define DICE_KIND_JUMP_TABLE8 0x0002
1474 #define DICE_KIND_JUMP_TABLE16 0x0003
1475 #define DICE_KIND_JUMP_TABLE32 0x0004
1476 #define DICE_KIND_ABS_JUMP_TABLE32 0x0005
1481 * Sections of type S_THREAD_LOCAL_VARIABLES contain an array
1482 * of tlv_descriptor structures.
1484 struct tlv_descriptor
1486 void* (*thunk
)(struct tlv_descriptor
*);
1488 unsigned long offset
;
1491 #endif /* _MACHO_LOADER_H_ */