]> git.saurik.com Git - apple/xnu.git/blame - EXTERNAL_HEADERS/mach-o/loader.h
xnu-1699.22.73.tar.gz
[apple/xnu.git] / EXTERNAL_HEADERS / mach-o / loader.h
CommitLineData
1c79356b 1/*
593a1d5f 2 * Copyright (c) 1999-2008 Apple Inc. All Rights Reserved.
1c79356b 3 *
b0d623f7 4 * @APPLE_LICENSE_HEADER_START@
1c79356b 5 *
2d21ac55
A
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
b0d623f7
A
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
11 * file.
2d21ac55
A
12 *
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
8f6c56a5
A
15 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
16 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
2d21ac55
A
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.
8f6c56a5 20 *
b0d623f7 21 * @APPLE_LICENSE_HEADER_END@
1c79356b
A
22 */
23#ifndef _MACHO_LOADER_H_
24#define _MACHO_LOADER_H_
25
26/*
27 * This file describes the format of mach object files.
28 */
2d21ac55 29#include <stdint.h>
1c79356b
A
30
31/*
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.
34 */
35#include <mach/machine.h>
36
37/*
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.
40 */
41#include <mach/vm_prot.h>
42
43/*
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.
46 */
47#include <mach/machine/thread_status.h>
2d21ac55 48#include <architecture/byte_order.h>
1c79356b
A
49
50/*
2d21ac55
A
51 * The 32-bit mach header appears at the very beginning of the object file for
52 * 32-bit architectures.
1c79356b
A
53 */
54struct mach_header {
91447636 55 uint32_t magic; /* mach magic number identifier */
1c79356b
A
56 cpu_type_t cputype; /* cpu specifier */
57 cpu_subtype_t cpusubtype; /* machine specifier */
91447636
A
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 */
62};
63
2d21ac55
A
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) */
67
91447636
A
68/*
69 * The 64-bit mach header appears at the very beginning of object files for
70 * 64-bit architectures.
71 */
72struct 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 */
1c79356b
A
81};
82
91447636 83/* Constant for the magic field of the mach_header_64 (64-bit architectures) */
2d21ac55
A
84#define MH_MAGIC_64 0xfeedfacf /* the 64-bit mach magic number */
85#define MH_CIGAM_64 0xcffaedfe /* NXSwapInt(MH_MAGIC_64) */
91447636 86
1c79356b
A
87/*
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.
93 *
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
2d21ac55 98 * segment padding greatly increases its size.
1c79356b
A
99 *
100 * The file type MH_PRELOAD is an executable format intended for things that
2d21ac55 101 * are not executed under the kernel (proms, stand alones, kernels, etc). The
1c79356b
A
102 * format can be executed under the kernel but may demand paged it and not
103 * preload it before execution.
104 *
105 * A core file is in MH_CORE format and can be any in an arbritray legal
106 * Mach-O file.
107 *
108 * Constants for the filetype field of the mach_header
109 */
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 */
2d21ac55 115#define MH_DYLIB 0x6 /* dynamically bound shared library */
1c79356b 116#define MH_DYLINKER 0x7 /* dynamic link editor */
2d21ac55
A
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 */
121 /* sections */
b0d623f7 122#define MH_KEXT_BUNDLE 0xb /* x86_64 kexts */
1c79356b
A
123
124/* Constants for the flags field of the mach_header */
125#define MH_NOUNDEFS 0x1 /* the object file has no undefined
2d21ac55 126 references */
1c79356b
A
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
132 link edited again */
133#define MH_BINDATLOAD 0x8 /* the object file's undefined
134 references are bound by the dynamic
135 linker when loaded. */
2d21ac55 136#define MH_PREBOUND 0x10 /* the file has its dynamic undefined
1c79356b 137 references prebound. */
2d21ac55
A
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
143 (obsolete) */
144#define MH_TWOLEVEL 0x80 /* the image is using two-level name
145 space bindings */
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
154 executable */
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
162 are both set. */
163#define MH_SUBSECTIONS_VIA_SYMBOLS 0x2000/* safe to divide up the sections into
164 sub-sections via symbols for dead
165 code stripping */
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
171 weak symbols */
172
0c530ab8
A
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. */
b0d623f7
A
177#define MH_DEAD_STRIPPABLE_DYLIB 0x400000 /* Only for use on dylibs. When
178 linking against a dylib that
179 has this bit set, the static linker
180 will automatically not create a
181 LC_LOAD_DYLIB load command to the
182 dylib if no symbols are being
183 referenced from the dylib. */
2d21ac55
A
184#define MH_ROOT_SAFE 0x40000 /* When this bit is set, the binary
185 declares it is safe for use in
186 processes with uid zero */
187
188#define MH_SETUID_SAFE 0x80000 /* When this bit is set, the binary
189 declares it is safe for use in
190 processes when issetugid() is true */
191
192#define MH_NO_REEXPORTED_DYLIBS 0x100000 /* When this bit is set on a dylib,
193 the static linker does not need to
194 examine dependent dylibs to see
195 if any are re-exported */
593a1d5f
A
196#define MH_PIE 0x200000 /* When this bit is set, the OS will
197 load the main executable at a
198 random address. Only used in
199 MH_EXECUTE filetypes. */
6d2010ae
A
200#define MH_NO_HEAP_EXECUTION 0x1000000 /* When this bit is set, the OS will
201 run the main executable with
202 a non-executable heap even on
203 platforms (e.g. i386) that don't
204 require it. Only used in MH_EXECUTE
205 filetypes. */
1c79356b
A
206
207/*
208 * The load commands directly follow the mach_header. The total size of all
209 * of the commands is given by the sizeofcmds field in the mach_header. All
210 * load commands must have as their first two fields cmd and cmdsize. The cmd
211 * field is filled in with a constant for that command type. Each command type
212 * has a structure specifically for it. The cmdsize field is the size in bytes
213 * of the particular load command structure plus anything that follows it that
214 * is a part of the load command (i.e. section structures, strings, etc.). To
215 * advance to the next load command the cmdsize can be added to the offset or
91447636
A
216 * pointer of the current load command. The cmdsize for 32-bit architectures
217 * MUST be a multiple of 4 bytes and for 64-bit architectures MUST be a multiple
218 * of 8 bytes (these are forever the maximum alignment of any load commands).
1c79356b
A
219 * The padded bytes must be zero. All tables in the object file must also
220 * follow these rules so the file can be memory mapped. Otherwise the pointers
221 * to these tables will not work well or at all on some machines. With all
222 * padding zeroed like objects will compare byte for byte.
223 */
224struct load_command {
2d21ac55
A
225 uint32_t cmd; /* type of load command */
226 uint32_t cmdsize; /* total size of command in bytes */
1c79356b
A
227};
228
2d21ac55
A
229/*
230 * After MacOS X 10.1 when a new load command is added that is required to be
231 * understood by the dynamic linker for the image to execute properly the
232 * LC_REQ_DYLD bit will be or'ed into the load command constant. If the dynamic
233 * linker sees such a load command it it does not understand will issue a
234 * "unknown load command required for execution" error and refuse to use the
235 * image. Other load commands without this bit that are not understood will
236 * simply be ignored.
237 */
238#define LC_REQ_DYLD 0x80000000
239
1c79356b
A
240/* Constants for the cmd field of all load commands, the type */
241#define LC_SEGMENT 0x1 /* segment of this file to be mapped */
242#define LC_SYMTAB 0x2 /* link-edit stab symbol table info */
243#define LC_SYMSEG 0x3 /* link-edit gdb symbol table info (obsolete) */
244#define LC_THREAD 0x4 /* thread */
245#define LC_UNIXTHREAD 0x5 /* unix thread (includes a stack) */
246#define LC_LOADFVMLIB 0x6 /* load a specified fixed VM shared library */
247#define LC_IDFVMLIB 0x7 /* fixed VM shared library identification */
248#define LC_IDENT 0x8 /* object identification info (obsolete) */
249#define LC_FVMFILE 0x9 /* fixed VM file inclusion (internal use) */
250#define LC_PREPAGE 0xa /* prepage command (internal use) */
251#define LC_DYSYMTAB 0xb /* dynamic link-edit symbol table info */
2d21ac55
A
252#define LC_LOAD_DYLIB 0xc /* load a dynamically linked shared library */
253#define LC_ID_DYLIB 0xd /* dynamically linked shared lib ident */
1c79356b
A
254#define LC_LOAD_DYLINKER 0xe /* load a dynamic linker */
255#define LC_ID_DYLINKER 0xf /* dynamic linker identification */
2d21ac55 256#define LC_PREBOUND_DYLIB 0x10 /* modules prebound for a dynamically */
1c79356b 257 /* linked shared library */
2d21ac55
A
258#define LC_ROUTINES 0x11 /* image routines */
259#define LC_SUB_FRAMEWORK 0x12 /* sub framework */
260#define LC_SUB_UMBRELLA 0x13 /* sub umbrella */
261#define LC_SUB_CLIENT 0x14 /* sub client */
262#define LC_SUB_LIBRARY 0x15 /* sub library */
263#define LC_TWOLEVEL_HINTS 0x16 /* two-level namespace lookup hints */
264#define LC_PREBIND_CKSUM 0x17 /* prebind checksum */
265
266/*
267 * load a dynamically linked shared library that is allowed to be missing
268 * (all symbols are weak imported).
269 */
270#define LC_LOAD_WEAK_DYLIB (0x18 | LC_REQ_DYLD)
271
272#define LC_SEGMENT_64 0x19 /* 64-bit segment of this file to be
273 mapped */
274#define LC_ROUTINES_64 0x1a /* 64-bit image routines */
275#define LC_UUID 0x1b /* the uuid */
276#define LC_RPATH (0x1c | LC_REQ_DYLD) /* runpath additions */
277#define LC_CODE_SIGNATURE 0x1d /* local of code signature */
278#define LC_SEGMENT_SPLIT_INFO 0x1e /* local of info to split segments */
279#define LC_REEXPORT_DYLIB (0x1f | LC_REQ_DYLD) /* load and re-export dylib */
593a1d5f
A
280#define LC_LAZY_LOAD_DYLIB 0x20 /* delay load of dylib until first use */
281#define LC_ENCRYPTION_INFO 0x21 /* encrypted segment information */
b0d623f7
A
282#define LC_DYLD_INFO 0x22 /* compressed dyld information */
283#define LC_DYLD_INFO_ONLY (0x22|LC_REQ_DYLD) /* compressed dyld information only */
1c79356b
A
284
285/*
286 * A variable length string in a load command is represented by an lc_str
287 * union. The strings are stored just after the load command structure and
288 * the offset is from the start of the load command structure. The size
289 * of the string is reflected in the cmdsize field of the load command.
290 * Once again any padded bytes to bring the cmdsize field to a multiple
2d21ac55 291 * of 4 bytes must be zero.
1c79356b
A
292 */
293union lc_str {
2d21ac55
A
294 uint32_t offset; /* offset to the string */
295#ifndef __LP64__
1c79356b 296 char *ptr; /* pointer to the string */
2d21ac55 297#endif
1c79356b
A
298};
299
300/*
301 * The segment load command indicates that a part of this file is to be
302 * mapped into the task's address space. The size of this segment in memory,
303 * vmsize, maybe equal to or larger than the amount to map from this file,
304 * filesize. The file is mapped starting at fileoff to the beginning of
305 * the segment in memory, vmaddr. The rest of the memory of the segment,
306 * if any, is allocated zero fill on demand. The segment's maximum virtual
307 * memory protection and initial virtual memory protection are specified
308 * by the maxprot and initprot fields. If the segment has sections then the
309 * section structures directly follow the segment command and their size is
310 * reflected in cmdsize.
311 */
2d21ac55
A
312struct segment_command { /* for 32-bit architectures */
313 uint32_t cmd; /* LC_SEGMENT */
314 uint32_t cmdsize; /* includes sizeof section structs */
1c79356b 315 char segname[16]; /* segment name */
2d21ac55
A
316 uint32_t vmaddr; /* memory address of this segment */
317 uint32_t vmsize; /* memory size of this segment */
318 uint32_t fileoff; /* file offset of this segment */
319 uint32_t filesize; /* amount to map from the file */
1c79356b
A
320 vm_prot_t maxprot; /* maximum VM protection */
321 vm_prot_t initprot; /* initial VM protection */
2d21ac55
A
322 uint32_t nsects; /* number of sections in segment */
323 uint32_t flags; /* flags */
1c79356b
A
324};
325
91447636
A
326/*
327 * The 64-bit segment load command indicates that a part of this file is to be
328 * mapped into a 64-bit task's address space. If the 64-bit segment has
329 * sections then section_64 structures directly follow the 64-bit segment
330 * command and their size is reflected in cmdsize.
331 */
2d21ac55 332struct segment_command_64 { /* for 64-bit architectures */
91447636
A
333 uint32_t cmd; /* LC_SEGMENT_64 */
334 uint32_t cmdsize; /* includes sizeof section_64 structs */
335 char segname[16]; /* segment name */
336 uint64_t vmaddr; /* memory address of this segment */
337 uint64_t vmsize; /* memory size of this segment */
338 uint64_t fileoff; /* file offset of this segment */
339 uint64_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 */
344};
345
1c79356b
A
346/* Constants for the flags field of the segment_command */
347#define SG_HIGHVM 0x1 /* the file contents for this segment is for
348 the high part of the VM space, the low part
349 is zero filled (for stacks in core files) */
350#define SG_FVMLIB 0x2 /* this segment is the VM that is allocated by
351 a fixed VM library, for overlap checking in
352 the link editor */
353#define SG_NORELOC 0x4 /* this segment has nothing that was relocated
354 in it and nothing relocated to it, that is
355 it maybe safely replaced without relocation*/
2d21ac55
A
356#define SG_PROTECTED_VERSION_1 0x8 /* This segment is protected. If the
357 segment starts at file offset 0, the
358 first page of the segment is not
359 protected. All other pages of the
360 segment are protected. */
1c79356b
A
361
362/*
363 * A segment is made up of zero or more sections. Non-MH_OBJECT files have
364 * all of their segments with the proper sections in each, and padded to the
365 * specified segment alignment when produced by the link editor. The first
366 * segment of a MH_EXECUTE and MH_FVMLIB format file contains the mach_header
2d21ac55 367 * and load commands of the object file before its first section. The zero
1c79356b
A
368 * fill sections are always last in their segment (in all formats). This
369 * allows the zeroed segment padding to be mapped into memory where zero fill
91447636
A
370 * sections might be. The gigabyte zero fill sections, those with the section
371 * type S_GB_ZEROFILL, can only be in a segment with sections of this type.
372 * These segments are then placed after all other segments.
1c79356b 373 *
2d21ac55 374 * The MH_OBJECT format has all of its sections in one segment for
1c79356b
A
375 * compactness. There is no padding to a specified segment boundary and the
376 * mach_header and load commands are not part of the segment.
377 *
378 * Sections with the same section name, sectname, going into the same segment,
379 * segname, are combined by the link editor. The resulting section is aligned
380 * to the maximum alignment of the combined sections and is the new section's
381 * alignment. The combined sections are aligned to their original alignment in
382 * the combined section. Any padded bytes to get the specified alignment are
383 * zeroed.
384 *
385 * The format of the relocation entries referenced by the reloff and nreloc
386 * fields of the section structure for mach object files is described in the
387 * header file <reloc.h>.
388 */
2d21ac55 389struct section { /* for 32-bit architectures */
1c79356b
A
390 char sectname[16]; /* name of this section */
391 char segname[16]; /* segment this section goes in */
2d21ac55
A
392 uint32_t addr; /* memory address of this section */
393 uint32_t size; /* size in bytes of this section */
394 uint32_t offset; /* file offset of this section */
395 uint32_t align; /* section alignment (power of 2) */
396 uint32_t reloff; /* file offset of relocation entries */
397 uint32_t nreloc; /* number of relocation entries */
398 uint32_t flags; /* flags (section type and attributes)*/
399 uint32_t reserved1; /* reserved (for offset or index) */
400 uint32_t reserved2; /* reserved (for count or sizeof) */
1c79356b
A
401};
402
91447636
A
403struct section_64 { /* for 64-bit architectures */
404 char sectname[16]; /* name of this section */
405 char segname[16]; /* segment this section goes in */
406 uint64_t addr; /* memory address of this section */
407 uint64_t size; /* size in bytes of this section */
408 uint32_t offset; /* file offset of this section */
409 uint32_t align; /* section alignment (power of 2) */
410 uint32_t reloff; /* file offset of relocation entries */
411 uint32_t nreloc; /* number of relocation entries */
412 uint32_t flags; /* flags (section type and attributes)*/
413 uint32_t reserved1; /* reserved (for offset or index) */
414 uint32_t reserved2; /* reserved (for count or sizeof) */
415 uint32_t reserved3; /* reserved */
416};
417
1c79356b
A
418/*
419 * The flags field of a section structure is separated into two parts a section
420 * type and section attributes. The section types are mutually exclusive (it
421 * can only have one type) but the section attributes are not (it may have more
422 * than one attribute).
423 */
424#define SECTION_TYPE 0x000000ff /* 256 section types */
425#define SECTION_ATTRIBUTES 0xffffff00 /* 24 section attributes */
426
427/* Constants for the type of a section */
428#define S_REGULAR 0x0 /* regular section */
429#define S_ZEROFILL 0x1 /* zero fill on demand section */
430#define S_CSTRING_LITERALS 0x2 /* section with only literal C strings*/
431#define S_4BYTE_LITERALS 0x3 /* section with only 4 byte literals */
432#define S_8BYTE_LITERALS 0x4 /* section with only 8 byte literals */
433#define S_LITERAL_POINTERS 0x5 /* section with only pointers to */
434 /* literals */
435/*
436 * For the two types of symbol pointers sections and the symbol stubs section
437 * they have indirect symbol table entries. For each of the entries in the
438 * section the indirect symbol table entries, in corresponding order in the
439 * indirect symbol table, start at the index stored in the reserved1 field
440 * of the section structure. Since the indirect symbol table entries
441 * correspond to the entries in the section the number of indirect symbol table
442 * entries is inferred from the size of the section divided by the size of the
443 * entries in the section. For symbol pointers sections the size of the entries
444 * in the section is 4 bytes and for symbol stubs sections the byte size of the
445 * stubs is stored in the reserved2 field of the section structure.
446 */
447#define S_NON_LAZY_SYMBOL_POINTERS 0x6 /* section with only non-lazy
448 symbol pointers */
449#define S_LAZY_SYMBOL_POINTERS 0x7 /* section with only lazy symbol
450 pointers */
451#define S_SYMBOL_STUBS 0x8 /* section with only symbol
452 stubs, byte size of stub in
453 the reserved2 field */
454#define S_MOD_INIT_FUNC_POINTERS 0x9 /* section with only function
455 pointers for initialization*/
2d21ac55
A
456#define S_MOD_TERM_FUNC_POINTERS 0xa /* section with only function
457 pointers for termination */
458#define S_COALESCED 0xb /* section contains symbols that
459 are to be coalesced */
460#define S_GB_ZEROFILL 0xc /* zero fill on demand section
461 (that can be larger than 4
462 gigabytes) */
463#define S_INTERPOSING 0xd /* section with only pairs of
464 function pointers for
465 interposing */
593a1d5f
A
466#define S_16BYTE_LITERALS 0xe /* section with only 16 byte
467 literals */
468#define S_DTRACE_DOF 0xf /* section contains
469 DTrace Object Format */
470#define S_LAZY_DYLIB_SYMBOL_POINTERS 0x10 /* section with only lazy
471 symbol pointers to lazy
472 loaded dylibs */
1c79356b
A
473/*
474 * Constants for the section attributes part of the flags field of a section
475 * structure.
476 */
477#define SECTION_ATTRIBUTES_USR 0xff000000 /* User setable attributes */
478#define S_ATTR_PURE_INSTRUCTIONS 0x80000000 /* section contains only true
479 machine instructions */
2d21ac55
A
480#define S_ATTR_NO_TOC 0x40000000 /* section contains coalesced
481 symbols that are not to be
482 in a ranlib table of
483 contents */
484#define S_ATTR_STRIP_STATIC_SYMS 0x20000000 /* ok to strip static symbols
485 in this section in files
486 with the MH_DYLDLINK flag */
487#define S_ATTR_NO_DEAD_STRIP 0x10000000 /* no dead stripping */
488#define S_ATTR_LIVE_SUPPORT 0x08000000 /* blocks are live if they
489 reference live blocks */
490#define S_ATTR_SELF_MODIFYING_CODE 0x04000000 /* Used with i386 code stubs
491 written on by dyld */
492/*
493 * If a segment contains any sections marked with S_ATTR_DEBUG then all
494 * sections in that segment must have this attribute. No section other than
495 * a section marked with this attribute may reference the contents of this
496 * section. A section with this attribute may contain no symbols and must have
497 * a section type S_REGULAR. The static linker will not copy section contents
498 * from sections with this attribute into its output file. These sections
499 * generally contain DWARF debugging info.
500 */
501#define S_ATTR_DEBUG 0x02000000 /* a debug section */
1c79356b
A
502#define SECTION_ATTRIBUTES_SYS 0x00ffff00 /* system setable attributes */
503#define S_ATTR_SOME_INSTRUCTIONS 0x00000400 /* section contains some
504 machine instructions */
505#define S_ATTR_EXT_RELOC 0x00000200 /* section has external
506 relocation entries */
507#define S_ATTR_LOC_RELOC 0x00000100 /* section has local
508 relocation entries */
509
510
511/*
512 * The names of segments and sections in them are mostly meaningless to the
513 * link-editor. But there are few things to support traditional UNIX
514 * executables that require the link-editor and assembler to use some names
515 * agreed upon by convention.
516 *
517 * The initial protection of the "__TEXT" segment has write protection turned
518 * off (not writeable).
519 *
520 * The link-editor will allocate common symbols at the end of the "__common"
521 * section in the "__DATA" segment. It will create the section and segment
522 * if needed.
523 */
524
525/* The currently known segment names and the section names in those segments */
526
527#define SEG_PAGEZERO "__PAGEZERO" /* the pagezero segment which has no */
528 /* protections and catches NULL */
529 /* references for MH_EXECUTE files */
530
531
532#define SEG_TEXT "__TEXT" /* the tradition UNIX text segment */
533#define SECT_TEXT "__text" /* the real text part of the text */
534 /* section no headers, and no padding */
535#define SECT_FVMLIB_INIT0 "__fvmlib_init0" /* the fvmlib initialization */
536 /* section */
537#define SECT_FVMLIB_INIT1 "__fvmlib_init1" /* the section following the */
538 /* fvmlib initialization */
539 /* section */
540
541#define SEG_DATA "__DATA" /* the tradition UNIX data segment */
542#define SECT_DATA "__data" /* the real initialized data section */
543 /* no padding, no bss overlap */
544#define SECT_BSS "__bss" /* the real uninitialized data section*/
545 /* no padding */
546#define SECT_COMMON "__common" /* the section common symbols are */
547 /* allocated in by the link editor */
548
549#define SEG_OBJC "__OBJC" /* objective-C runtime segment */
550#define SECT_OBJC_SYMBOLS "__symbol_table" /* symbol table */
551#define SECT_OBJC_MODULES "__module_info" /* module information */
552#define SECT_OBJC_STRINGS "__selector_strs" /* string table */
553#define SECT_OBJC_REFS "__selector_refs" /* string table */
554
2d21ac55 555#define SEG_ICON "__ICON" /* the icon segment */
1c79356b
A
556#define SECT_ICON_HEADER "__header" /* the icon headers */
557#define SECT_ICON_TIFF "__tiff" /* the icons in tiff format */
558
559#define SEG_LINKEDIT "__LINKEDIT" /* the segment containing all structs */
560 /* created and maintained by the link */
561 /* editor. Created with -seglinkedit */
562 /* option to ld(1) for MH_EXECUTE and */
563 /* FVMLIB file types only */
564
565#define SEG_UNIXSTACK "__UNIXSTACK" /* the unix stack segment */
566
2d21ac55
A
567#define SEG_IMPORT "__IMPORT" /* the segment for the self (dyld) */
568 /* modifing code stubs that has read, */
569 /* write and execute permissions */
570
1c79356b
A
571/*
572 * Fixed virtual memory shared libraries are identified by two things. The
573 * target pathname (the name of the library as found for execution), and the
574 * minor version number. The address of where the headers are loaded is in
2d21ac55 575 * header_addr. (THIS IS OBSOLETE and no longer supported).
1c79356b
A
576 */
577struct fvmlib {
578 union lc_str name; /* library's target pathname */
2d21ac55
A
579 uint32_t minor_version; /* library's minor version number */
580 uint32_t header_addr; /* library's header address */
1c79356b
A
581};
582
583/*
584 * A fixed virtual shared library (filetype == MH_FVMLIB in the mach header)
585 * contains a fvmlib_command (cmd == LC_IDFVMLIB) to identify the library.
586 * An object that uses a fixed virtual shared library also contains a
587 * fvmlib_command (cmd == LC_LOADFVMLIB) for each library it uses.
2d21ac55 588 * (THIS IS OBSOLETE and no longer supported).
1c79356b
A
589 */
590struct fvmlib_command {
2d21ac55
A
591 uint32_t cmd; /* LC_IDFVMLIB or LC_LOADFVMLIB */
592 uint32_t cmdsize; /* includes pathname string */
1c79356b
A
593 struct fvmlib fvmlib; /* the library identification */
594};
595
596/*
597 * Dynamicly linked shared libraries are identified by two things. The
598 * pathname (the name of the library as found for execution), and the
599 * compatibility version number. The pathname must match and the compatibility
600 * number in the user of the library must be greater than or equal to the
601 * library being used. The time stamp is used to record the time a library was
602 * built and copied into user so it can be use to determined if the library used
603 * at runtime is exactly the same as used to built the program.
604 */
605struct dylib {
606 union lc_str name; /* library's path name */
2d21ac55
A
607 uint32_t timestamp; /* library's build time stamp */
608 uint32_t current_version; /* library's current version number */
609 uint32_t compatibility_version; /* library's compatibility vers number*/
1c79356b
A
610};
611
612/*
2d21ac55 613 * A dynamically linked shared library (filetype == MH_DYLIB in the mach header)
1c79356b 614 * contains a dylib_command (cmd == LC_ID_DYLIB) to identify the library.
2d21ac55
A
615 * An object that uses a dynamically linked shared library also contains a
616 * dylib_command (cmd == LC_LOAD_DYLIB, LC_LOAD_WEAK_DYLIB, or
617 * LC_REEXPORT_DYLIB) for each library it uses.
1c79356b
A
618 */
619struct dylib_command {
2d21ac55
A
620 uint32_t cmd; /* LC_ID_DYLIB, LC_LOAD_{,WEAK_}DYLIB,
621 LC_REEXPORT_DYLIB */
622 uint32_t cmdsize; /* includes pathname string */
1c79356b
A
623 struct dylib dylib; /* the library identification */
624};
625
626/*
2d21ac55
A
627 * A dynamically linked shared library may be a subframework of an umbrella
628 * framework. If so it will be linked with "-umbrella umbrella_name" where
629 * Where "umbrella_name" is the name of the umbrella framework. A subframework
630 * can only be linked against by its umbrella framework or other subframeworks
631 * that are part of the same umbrella framework. Otherwise the static link
632 * editor produces an error and states to link against the umbrella framework.
633 * The name of the umbrella framework for subframeworks is recorded in the
634 * following structure.
635 */
636struct sub_framework_command {
637 uint32_t cmd; /* LC_SUB_FRAMEWORK */
638 uint32_t cmdsize; /* includes umbrella string */
639 union lc_str umbrella; /* the umbrella framework name */
640};
641
642/*
643 * For dynamically linked shared libraries that are subframework of an umbrella
644 * framework they can allow clients other than the umbrella framework or other
645 * subframeworks in the same umbrella framework. To do this the subframework
646 * is built with "-allowable_client client_name" and an LC_SUB_CLIENT load
647 * command is created for each -allowable_client flag. The client_name is
648 * usually a framework name. It can also be a name used for bundles clients
649 * where the bundle is built with "-client_name client_name".
650 */
651struct sub_client_command {
652 uint32_t cmd; /* LC_SUB_CLIENT */
653 uint32_t cmdsize; /* includes client string */
654 union lc_str client; /* the client name */
655};
656
657/*
658 * A dynamically linked shared library may be a sub_umbrella of an umbrella
659 * framework. If so it will be linked with "-sub_umbrella umbrella_name" where
660 * Where "umbrella_name" is the name of the sub_umbrella framework. When
661 * staticly linking when -twolevel_namespace is in effect a twolevel namespace
662 * umbrella framework will only cause its subframeworks and those frameworks
663 * listed as sub_umbrella frameworks to be implicited linked in. Any other
664 * dependent dynamic libraries will not be linked it when -twolevel_namespace
665 * is in effect. The primary library recorded by the static linker when
666 * resolving a symbol in these libraries will be the umbrella framework.
667 * Zero or more sub_umbrella frameworks may be use by an umbrella framework.
668 * The name of a sub_umbrella framework is recorded in the following structure.
669 */
670struct sub_umbrella_command {
671 uint32_t cmd; /* LC_SUB_UMBRELLA */
672 uint32_t cmdsize; /* includes sub_umbrella string */
673 union lc_str sub_umbrella; /* the sub_umbrella framework name */
674};
675
676/*
677 * A dynamically linked shared library may be a sub_library of another shared
678 * library. If so it will be linked with "-sub_library library_name" where
679 * Where "library_name" is the name of the sub_library shared library. When
680 * staticly linking when -twolevel_namespace is in effect a twolevel namespace
681 * shared library will only cause its subframeworks and those frameworks
682 * listed as sub_umbrella frameworks and libraries listed as sub_libraries to
683 * be implicited linked in. Any other dependent dynamic libraries will not be
684 * linked it when -twolevel_namespace is in effect. The primary library
685 * recorded by the static linker when resolving a symbol in these libraries
686 * will be the umbrella framework (or dynamic library). Zero or more sub_library
687 * shared libraries may be use by an umbrella framework or (or dynamic library).
688 * The name of a sub_library framework is recorded in the following structure.
689 * For example /usr/lib/libobjc_profile.A.dylib would be recorded as "libobjc".
690 */
691struct sub_library_command {
692 uint32_t cmd; /* LC_SUB_LIBRARY */
693 uint32_t cmdsize; /* includes sub_library string */
694 union lc_str sub_library; /* the sub_library name */
695};
696
697/*
698 * A program (filetype == MH_EXECUTE) that is
699 * prebound to its dynamic libraries has one of these for each library that
1c79356b
A
700 * the static linker used in prebinding. It contains a bit vector for the
701 * modules in the library. The bits indicate which modules are bound (1) and
702 * which are not (0) from the library. The bit for module 0 is the low bit
703 * of the first byte. So the bit for the Nth module is:
704 * (linked_modules[N/8] >> N%8) & 1
705 */
706struct prebound_dylib_command {
2d21ac55
A
707 uint32_t cmd; /* LC_PREBOUND_DYLIB */
708 uint32_t cmdsize; /* includes strings */
1c79356b 709 union lc_str name; /* library's path name */
2d21ac55 710 uint32_t nmodules; /* number of modules in library */
1c79356b
A
711 union lc_str linked_modules; /* bit vector of linked modules */
712};
713
714/*
715 * A program that uses a dynamic linker contains a dylinker_command to identify
716 * the name of the dynamic linker (LC_LOAD_DYLINKER). And a dynamic linker
717 * contains a dylinker_command to identify the dynamic linker (LC_ID_DYLINKER).
718 * A file can have at most one of these.
719 */
720struct dylinker_command {
2d21ac55
A
721 uint32_t cmd; /* LC_ID_DYLINKER or LC_LOAD_DYLINKER */
722 uint32_t cmdsize; /* includes pathname string */
1c79356b
A
723 union lc_str name; /* dynamic linker's path name */
724};
725
726/*
727 * Thread commands contain machine-specific data structures suitable for
728 * use in the thread state primitives. The machine specific data structures
729 * follow the struct thread_command as follows.
730 * Each flavor of machine specific data structure is preceded by an unsigned
2d21ac55 731 * long constant for the flavor of that data structure, an uint32_t
1c79356b
A
732 * that is the count of longs of the size of the state data structure and then
733 * the state data structure follows. This triple may be repeated for many
734 * flavors. The constants for the flavors, counts and state data structure
735 * definitions are expected to be in the header file <machine/thread_status.h>.
736 * These machine specific data structures sizes must be multiples of
2d21ac55 737 * 4 bytes The cmdsize reflects the total size of the thread_command
1c79356b
A
738 * and all of the sizes of the constants for the flavors, counts and state
739 * data structures.
740 *
741 * For executable objects that are unix processes there will be one
742 * thread_command (cmd == LC_UNIXTHREAD) created for it by the link-editor.
743 * This is the same as a LC_THREAD, except that a stack is automatically
744 * created (based on the shell's limit for the stack size). Command arguments
745 * and environment variables are copied onto that stack.
746 */
747struct thread_command {
2d21ac55
A
748 uint32_t cmd; /* LC_THREAD or LC_UNIXTHREAD */
749 uint32_t cmdsize; /* total size of this command */
750 /* uint32_t flavor flavor of thread state */
751 /* uint32_t count count of longs in thread state */
1c79356b
A
752 /* struct XXX_thread_state state thread state for this flavor */
753 /* ... */
754};
755
2d21ac55
A
756/*
757 * The routines command contains the address of the dynamic shared library
758 * initialization routine and an index into the module table for the module
759 * that defines the routine. Before any modules are used from the library the
760 * dynamic linker fully binds the module that defines the initialization routine
761 * and then calls it. This gets called before any module initialization
762 * routines (used for C++ static constructors) in the library.
763 */
764struct routines_command { /* for 32-bit architectures */
765 uint32_t cmd; /* LC_ROUTINES */
766 uint32_t cmdsize; /* total size of this command */
767 uint32_t init_address; /* address of initialization routine */
768 uint32_t init_module; /* index into the module table that */
769 /* the init routine is defined in */
770 uint32_t reserved1;
771 uint32_t reserved2;
772 uint32_t reserved3;
773 uint32_t reserved4;
774 uint32_t reserved5;
775 uint32_t reserved6;
776};
777
778/*
779 * The 64-bit routines command. Same use as above.
780 */
781struct routines_command_64 { /* for 64-bit architectures */
782 uint32_t cmd; /* LC_ROUTINES_64 */
783 uint32_t cmdsize; /* total size of this command */
784 uint64_t init_address; /* address of initialization routine */
785 uint64_t init_module; /* index into the module table that */
786 /* the init routine is defined in */
787 uint64_t reserved1;
788 uint64_t reserved2;
789 uint64_t reserved3;
790 uint64_t reserved4;
791 uint64_t reserved5;
792 uint64_t reserved6;
793};
794
1c79356b
A
795/*
796 * The symtab_command contains the offsets and sizes of the link-edit 4.3BSD
797 * "stab" style symbol table information as described in the header files
798 * <nlist.h> and <stab.h>.
799 */
800struct symtab_command {
2d21ac55
A
801 uint32_t cmd; /* LC_SYMTAB */
802 uint32_t cmdsize; /* sizeof(struct symtab_command) */
803 uint32_t symoff; /* symbol table offset */
804 uint32_t nsyms; /* number of symbol table entries */
805 uint32_t stroff; /* string table offset */
806 uint32_t strsize; /* string table size in bytes */
1c79356b
A
807};
808
809/*
810 * This is the second set of the symbolic information which is used to support
2d21ac55 811 * the data structures for the dynamically link editor.
1c79356b
A
812 *
813 * The original set of symbolic information in the symtab_command which contains
814 * the symbol and string tables must also be present when this load command is
815 * present. When this load command is present the symbol table is organized
816 * into three groups of symbols:
817 * local symbols (static and debugging symbols) - grouped by module
818 * defined external symbols - grouped by module (sorted by name if not lib)
2d21ac55
A
819 * undefined external symbols (sorted by name if MH_BINDATLOAD is not set,
820 * and in order the were seen by the static
821 * linker if MH_BINDATLOAD is set)
1c79356b
A
822 * In this load command there are offsets and counts to each of the three groups
823 * of symbols.
824 *
825 * This load command contains a the offsets and sizes of the following new
826 * symbolic information tables:
827 * table of contents
828 * module table
829 * reference symbol table
830 * indirect symbol table
831 * The first three tables above (the table of contents, module table and
2d21ac55 832 * reference symbol table) are only present if the file is a dynamically linked
1c79356b
A
833 * shared library. For executable and object modules, which are files
834 * containing only one module, the information that would be in these three
835 * tables is determined as follows:
836 * table of contents - the defined external symbols are sorted by name
837 * module table - the file contains only one module so everything in the
838 * file is part of the module.
839 * reference symbol table - is the defined and undefined external symbols
840 *
2d21ac55 841 * For dynamically linked shared library files this load command also contains
1c79356b
A
842 * offsets and sizes to the pool of relocation entries for all sections
843 * separated into two groups:
844 * external relocation entries
845 * local relocation entries
846 * For executable and object modules the relocation entries continue to hang
847 * off the section structures.
848 */
849struct dysymtab_command {
2d21ac55
A
850 uint32_t cmd; /* LC_DYSYMTAB */
851 uint32_t cmdsize; /* sizeof(struct dysymtab_command) */
1c79356b
A
852
853 /*
854 * The symbols indicated by symoff and nsyms of the LC_SYMTAB load command
855 * are grouped into the following three groups:
856 * local symbols (further grouped by the module they are from)
857 * defined external symbols (further grouped by the module they are from)
858 * undefined symbols
859 *
860 * The local symbols are used only for debugging. The dynamic binding
861 * process may have to use them to indicate to the debugger the local
862 * symbols for a module that is being bound.
863 *
864 * The last two groups are used by the dynamic binding process to do the
865 * binding (indirectly through the module table and the reference symbol
2d21ac55 866 * table when this is a dynamically linked shared library file).
1c79356b 867 */
2d21ac55
A
868 uint32_t ilocalsym; /* index to local symbols */
869 uint32_t nlocalsym; /* number of local symbols */
1c79356b 870
2d21ac55
A
871 uint32_t iextdefsym;/* index to externally defined symbols */
872 uint32_t nextdefsym;/* number of externally defined symbols */
1c79356b 873
2d21ac55
A
874 uint32_t iundefsym; /* index to undefined symbols */
875 uint32_t nundefsym; /* number of undefined symbols */
1c79356b
A
876
877 /*
878 * For the for the dynamic binding process to find which module a symbol
879 * is defined in the table of contents is used (analogous to the ranlib
880 * structure in an archive) which maps defined external symbols to modules
2d21ac55 881 * they are defined in. This exists only in a dynamically linked shared
1c79356b
A
882 * library file. For executable and object modules the defined external
883 * symbols are sorted by name and is use as the table of contents.
884 */
2d21ac55
A
885 uint32_t tocoff; /* file offset to table of contents */
886 uint32_t ntoc; /* number of entries in table of contents */
1c79356b
A
887
888 /*
889 * To support dynamic binding of "modules" (whole object files) the symbol
890 * table must reflect the modules that the file was created from. This is
891 * done by having a module table that has indexes and counts into the merged
892 * tables for each module. The module structure that these two entries
2d21ac55 893 * refer to is described below. This exists only in a dynamically linked
1c79356b
A
894 * shared library file. For executable and object modules the file only
895 * contains one module so everything in the file belongs to the module.
896 */
2d21ac55
A
897 uint32_t modtaboff; /* file offset to module table */
898 uint32_t nmodtab; /* number of module table entries */
1c79356b
A
899
900 /*
901 * To support dynamic module binding the module structure for each module
902 * indicates the external references (defined and undefined) each module
903 * makes. For each module there is an offset and a count into the
904 * reference symbol table for the symbols that the module references.
2d21ac55 905 * This exists only in a dynamically linked shared library file. For
1c79356b
A
906 * executable and object modules the defined external symbols and the
907 * undefined external symbols indicates the external references.
908 */
2d21ac55
A
909 uint32_t extrefsymoff; /* offset to referenced symbol table */
910 uint32_t nextrefsyms; /* number of referenced symbol table entries */
1c79356b
A
911
912 /*
913 * The sections that contain "symbol pointers" and "routine stubs" have
914 * indexes and (implied counts based on the size of the section and fixed
915 * size of the entry) into the "indirect symbol" table for each pointer
916 * and stub. For every section of these two types the index into the
917 * indirect symbol table is stored in the section header in the field
918 * reserved1. An indirect symbol table entry is simply a 32bit index into
919 * the symbol table to the symbol that the pointer or stub is referring to.
920 * The indirect symbol table is ordered to match the entries in the section.
921 */
2d21ac55
A
922 uint32_t indirectsymoff; /* file offset to the indirect symbol table */
923 uint32_t nindirectsyms; /* number of indirect symbol table entries */
1c79356b
A
924
925 /*
926 * To support relocating an individual module in a library file quickly the
927 * external relocation entries for each module in the library need to be
928 * accessed efficiently. Since the relocation entries can't be accessed
929 * through the section headers for a library file they are separated into
930 * groups of local and external entries further grouped by module. In this
931 * case the presents of this load command who's extreloff, nextrel,
932 * locreloff and nlocrel fields are non-zero indicates that the relocation
933 * entries of non-merged sections are not referenced through the section
934 * structures (and the reloff and nreloc fields in the section headers are
935 * set to zero).
936 *
937 * Since the relocation entries are not accessed through the section headers
938 * this requires the r_address field to be something other than a section
939 * offset to identify the item to be relocated. In this case r_address is
940 * set to the offset from the vmaddr of the first LC_SEGMENT command.
2d21ac55
A
941 * For MH_SPLIT_SEGS images r_address is set to the the offset from the
942 * vmaddr of the first read-write LC_SEGMENT command.
1c79356b
A
943 *
944 * The relocation entries are grouped by module and the module table
945 * entries have indexes and counts into them for the group of external
946 * relocation entries for that the module.
947 *
948 * For sections that are merged across modules there must not be any
949 * remaining external relocation entries for them (for merged sections
950 * remaining relocation entries must be local).
951 */
2d21ac55
A
952 uint32_t extreloff; /* offset to external relocation entries */
953 uint32_t nextrel; /* number of external relocation entries */
1c79356b
A
954
955 /*
956 * All the local relocation entries are grouped together (they are not
957 * grouped by their module since they are only used if the object is moved
958 * from it staticly link edited address).
959 */
2d21ac55
A
960 uint32_t locreloff; /* offset to local relocation entries */
961 uint32_t nlocrel; /* number of local relocation entries */
1c79356b
A
962
963};
964
965/*
966 * An indirect symbol table entry is simply a 32bit index into the symbol table
967 * to the symbol that the pointer or stub is refering to. Unless it is for a
968 * non-lazy symbol pointer section for a defined symbol which strip(1) as
969 * removed. In which case it has the value INDIRECT_SYMBOL_LOCAL. If the
970 * symbol was also absolute INDIRECT_SYMBOL_ABS is or'ed with that.
971 */
972#define INDIRECT_SYMBOL_LOCAL 0x80000000
973#define INDIRECT_SYMBOL_ABS 0x40000000
974
975
976/* a table of contents entry */
977struct dylib_table_of_contents {
2d21ac55 978 uint32_t symbol_index; /* the defined external symbol
1c79356b 979 (index into the symbol table) */
2d21ac55 980 uint32_t module_index; /* index into the module table this symbol
1c79356b
A
981 is defined in */
982};
983
984/* a module table entry */
985struct dylib_module {
2d21ac55 986 uint32_t module_name; /* the module name (index into string table) */
1c79356b 987
2d21ac55
A
988 uint32_t iextdefsym; /* index into externally defined symbols */
989 uint32_t nextdefsym; /* number of externally defined symbols */
990 uint32_t irefsym; /* index into reference symbol table */
991 uint32_t nrefsym; /* number of reference symbol table entries */
992 uint32_t ilocalsym; /* index into symbols for local symbols */
993 uint32_t nlocalsym; /* number of local symbols */
1c79356b 994
2d21ac55
A
995 uint32_t iextrel; /* index into external relocation entries */
996 uint32_t nextrel; /* number of external relocation entries */
1c79356b 997
2d21ac55
A
998 uint32_t iinit_iterm; /* low 16 bits are the index into the init
999 section, high 16 bits are the index into
1000 the term section */
1001 uint32_t ninit_nterm; /* low 16 bits are the number of init section
1002 entries, high 16 bits are the number of
1003 term section entries */
1c79356b 1004
2d21ac55 1005 uint32_t /* for this module address of the start of */
1c79356b 1006 objc_module_info_addr; /* the (__OBJC,__module_info) section */
2d21ac55 1007 uint32_t /* for this module size of */
1c79356b
A
1008 objc_module_info_size; /* the (__OBJC,__module_info) section */
1009};
1010
91447636
A
1011/* a 64-bit module table entry */
1012struct dylib_module_64 {
2d21ac55 1013 uint32_t module_name; /* the module name (index into string table) */
91447636 1014
2d21ac55
A
1015 uint32_t iextdefsym; /* index into externally defined symbols */
1016 uint32_t nextdefsym; /* number of externally defined symbols */
1017 uint32_t irefsym; /* index into reference symbol table */
1018 uint32_t nrefsym; /* number of reference symbol table entries */
1019 uint32_t ilocalsym; /* index into symbols for local symbols */
1020 uint32_t nlocalsym; /* number of local symbols */
91447636 1021
2d21ac55
A
1022 uint32_t iextrel; /* index into external relocation entries */
1023 uint32_t nextrel; /* number of external relocation entries */
91447636 1024
2d21ac55 1025 uint32_t iinit_iterm; /* low 16 bits are the index into the init
91447636
A
1026 section, high 16 bits are the index into
1027 the term section */
2d21ac55
A
1028 uint32_t ninit_nterm; /* low 16 bits are the number of init section
1029 entries, high 16 bits are the number of
1030 term section entries */
1031
1032 uint32_t /* for this module size of */
1033 objc_module_info_size; /* the (__OBJC,__module_info) section */
1034 uint64_t /* for this module address of the start of */
1035 objc_module_info_addr; /* the (__OBJC,__module_info) section */
91447636
A
1036};
1037
1c79356b
A
1038/*
1039 * The entries in the reference symbol table are used when loading the module
1040 * (both by the static and dynamic link editors) and if the module is unloaded
1041 * or replaced. Therefore all external symbols (defined and undefined) are
1042 * listed in the module's reference table. The flags describe the type of
1043 * reference that is being made. The constants for the flags are defined in
1044 * <mach-o/nlist.h> as they are also used for symbol table entries.
1045 */
1046struct dylib_reference {
2d21ac55 1047 uint32_t isym:24, /* index into the symbol table */
1c79356b
A
1048 flags:8; /* flags to indicate the type of reference */
1049};
1050
2d21ac55
A
1051/*
1052 * The twolevel_hints_command contains the offset and number of hints in the
1053 * two-level namespace lookup hints table.
1054 */
1055struct twolevel_hints_command {
1056 uint32_t cmd; /* LC_TWOLEVEL_HINTS */
1057 uint32_t cmdsize; /* sizeof(struct twolevel_hints_command) */
1058 uint32_t offset; /* offset to the hint table */
1059 uint32_t nhints; /* number of hints in the hint table */
1060};
1061
1062/*
1063 * The entries in the two-level namespace lookup hints table are twolevel_hint
1064 * structs. These provide hints to the dynamic link editor where to start
1065 * looking for an undefined symbol in a two-level namespace image. The
1066 * isub_image field is an index into the sub-images (sub-frameworks and
1067 * sub-umbrellas list) that made up the two-level image that the undefined
1068 * symbol was found in when it was built by the static link editor. If
1069 * isub-image is 0 the the symbol is expected to be defined in library and not
1070 * in the sub-images. If isub-image is non-zero it is an index into the array
1071 * of sub-images for the umbrella with the first index in the sub-images being
1072 * 1. The array of sub-images is the ordered list of sub-images of the umbrella
1073 * that would be searched for a symbol that has the umbrella recorded as its
1074 * primary library. The table of contents index is an index into the
1075 * library's table of contents. This is used as the starting point of the
1076 * binary search or a directed linear search.
1077 */
1078struct twolevel_hint {
1079 uint32_t
1080 isub_image:8, /* index into the sub images */
1081 itoc:24; /* index into the table of contents */
1082};
1083
1084/*
1085 * The prebind_cksum_command contains the value of the original check sum for
1086 * prebound files or zero. When a prebound file is first created or modified
1087 * for other than updating its prebinding information the value of the check sum
1088 * is set to zero. When the file has it prebinding re-done and if the value of
1089 * the check sum is zero the original check sum is calculated and stored in
1090 * cksum field of this load command in the output file. If when the prebinding
1091 * is re-done and the cksum field is non-zero it is left unchanged from the
1092 * input file.
1093 */
1094struct prebind_cksum_command {
1095 uint32_t cmd; /* LC_PREBIND_CKSUM */
1096 uint32_t cmdsize; /* sizeof(struct prebind_cksum_command) */
1097 uint32_t cksum; /* the check sum or zero */
1098};
1099
1100/*
1101 * The uuid load command contains a single 128-bit unique random number that
1102 * identifies an object produced by the static link editor.
1103 */
1104struct uuid_command {
1105 uint32_t cmd; /* LC_UUID */
1106 uint32_t cmdsize; /* sizeof(struct uuid_command) */
1107 uint8_t uuid[16]; /* the 128-bit uuid */
1108};
1109
1110/*
1111 * The rpath_command contains a path which at runtime should be added to
1112 * the current run path used to find @rpath prefixed dylibs.
1113 */
1114struct rpath_command {
1115 uint32_t cmd; /* LC_RPATH */
1116 uint32_t cmdsize; /* includes string */
1117 union lc_str path; /* path to add to run path */
1118};
1119
1120/*
1121 * The linkedit_data_command contains the offsets and sizes of a blob
1122 * of data in the __LINKEDIT segment.
1123 */
1124struct linkedit_data_command {
1125 uint32_t cmd; /* LC_CODE_SIGNATURE or LC_SEGMENT_SPLIT_INFO */
1126 uint32_t cmdsize; /* sizeof(struct linkedit_data_command) */
1127 uint32_t dataoff; /* file offset of data in __LINKEDIT segment */
1128 uint32_t datasize; /* file size of data in __LINKEDIT segment */
1129};
1130
593a1d5f
A
1131/*
1132 * The encryption_info_command contains the file offset and size of an
1133 * of an encrypted segment.
1134 */
1135struct encryption_info_command {
1136 uint32_t cmd; /* LC_ENCRYPTION_INFO */
1137 uint32_t cmdsize; /* sizeof(struct encryption_info_command) */
1138 uint32_t cryptoff; /* file offset of encrypted range */
1139 uint32_t cryptsize; /* file size of encrypted range */
1140 uint32_t cryptid; /* which enryption system,
1141 0 means not-encrypted yet */
1142};
1143
b0d623f7
A
1144/*
1145 * The dyld_info_command contains the file offsets and sizes of
1146 * the new compressed form of the information dyld needs to
1147 * load the image. This information is used by dyld on Mac OS X
1148 * 10.6 and later. All information pointed to by this command
1149 * is encoded using byte streams, so no endian swapping is needed
1150 * to interpret it.
1151 */
1152struct dyld_info_command {
1153 uint32_t cmd; /* LC_DYLD_INFO or LC_DYLD_INFO_ONLY */
1154 uint32_t cmdsize; /* sizeof(struct dyld_info_command) */
1155
1156 /*
1157 * Dyld rebases an image whenever dyld loads it at an address different
1158 * from its preferred address. The rebase information is a stream
1159 * of byte sized opcodes whose symbolic names start with REBASE_OPCODE_.
1160 * Conceptually the rebase information is a table of tuples:
1161 * <seg-index, seg-offset, type>
1162 * The opcodes are a compressed way to encode the table by only
1163 * encoding when a column changes. In addition simple patterns
1164 * like "every n'th offset for m times" can be encoded in a few
1165 * bytes.
1166 */
1167 uint32_t rebase_off; /* file offset to rebase info */
1168 uint32_t rebase_size; /* size of rebase info */
1169
1170 /*
1171 * Dyld binds an image during the loading process, if the image
1172 * requires any pointers to be initialized to symbols in other images.
1173 * The rebase information is a stream of byte sized
1174 * opcodes whose symbolic names start with BIND_OPCODE_.
1175 * Conceptually the bind information is a table of tuples:
1176 * <seg-index, seg-offset, type, symbol-library-ordinal, symbol-name, addend>
1177 * The opcodes are a compressed way to encode the table by only
1178 * encoding when a column changes. In addition simple patterns
1179 * like for runs of pointers initialzed to the same value can be
1180 * encoded in a few bytes.
1181 */
1182 uint32_t bind_off; /* file offset to binding info */
1183 uint32_t bind_size; /* size of binding info */
1184
1185 /*
1186 * Some C++ programs require dyld to unique symbols so that all
1187 * images in the process use the same copy of some code/data.
1188 * This step is done after binding. The content of the weak_bind
1189 * info is an opcode stream like the bind_info. But it is sorted
1190 * alphabetically by symbol name. This enable dyld to walk
1191 * all images with weak binding information in order and look
1192 * for collisions. If there are no collisions, dyld does
1193 * no updating. That means that some fixups are also encoded
1194 * in the bind_info. For instance, all calls to "operator new"
1195 * are first bound to libstdc++.dylib using the information
1196 * in bind_info. Then if some image overrides operator new
1197 * that is detected when the weak_bind information is processed
1198 * and the call to operator new is then rebound.
1199 */
1200 uint32_t weak_bind_off; /* file offset to weak binding info */
1201 uint32_t weak_bind_size; /* size of weak binding info */
1202
1203 /*
1204 * Some uses of external symbols do not need to be bound immediately.
1205 * Instead they can be lazily bound on first use. The lazy_bind
1206 * are contains a stream of BIND opcodes to bind all lazy symbols.
1207 * Normal use is that dyld ignores the lazy_bind section when
1208 * loading an image. Instead the static linker arranged for the
1209 * lazy pointer to initially point to a helper function which
1210 * pushes the offset into the lazy_bind area for the symbol
1211 * needing to be bound, then jumps to dyld which simply adds
1212 * the offset to lazy_bind_off to get the information on what
1213 * to bind.
1214 */
1215 uint32_t lazy_bind_off; /* file offset to lazy binding info */
1216 uint32_t lazy_bind_size; /* size of lazy binding infs */
1217
1218 /*
1219 * The symbols exported by a dylib are encoded in a trie. This
1220 * is a compact representation that factors out common prefixes.
1221 * It also reduces LINKEDIT pages in RAM because it encodes all
1222 * information (name, address, flags) in one small, contiguous range.
1223 * The export area is a stream of nodes. The first node sequentially
1224 * is the start node for the trie.
1225 *
1226 * Nodes for a symbol start with a byte that is the length of
1227 * the exported symbol information for the string so far.
1228 * If there is no exported symbol, the byte is zero. If there
1229 * is exported info, it follows the length byte. The exported
1230 * info normally consists of a flags and offset both encoded
1231 * in uleb128. The offset is location of the content named
1232 * by the symbol. It is the offset from the mach_header for
1233 * the image.
1234 *
1235 * After the initial byte and optional exported symbol information
1236 * is a byte of how many edges (0-255) that this node has leaving
1237 * it, followed by each edge.
1238 * Each edge is a zero terminated cstring of the addition chars
1239 * in the symbol, followed by a uleb128 offset for the node that
1240 * edge points to.
1241 *
1242 */
1243 uint32_t export_off; /* file offset to lazy binding info */
1244 uint32_t export_size; /* size of lazy binding infs */
1245};
1246
1247/*
1248 * The following are used to encode rebasing information
1249 */
1250#define REBASE_TYPE_POINTER 1
1251#define REBASE_TYPE_TEXT_ABSOLUTE32 2
1252#define REBASE_TYPE_TEXT_PCREL32 3
1253
1254#define REBASE_OPCODE_MASK 0xF0
1255#define REBASE_IMMEDIATE_MASK 0x0F
1256#define REBASE_OPCODE_DONE 0x00
1257#define REBASE_OPCODE_SET_TYPE_IMM 0x10
1258#define REBASE_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB 0x20
1259#define REBASE_OPCODE_ADD_ADDR_ULEB 0x30
1260#define REBASE_OPCODE_ADD_ADDR_IMM_SCALED 0x40
1261#define REBASE_OPCODE_DO_REBASE_IMM_TIMES 0x50
1262#define REBASE_OPCODE_DO_REBASE_ULEB_TIMES 0x60
1263#define REBASE_OPCODE_DO_REBASE_ADD_ADDR_ULEB 0x70
1264#define REBASE_OPCODE_DO_REBASE_ULEB_TIMES_SKIPPING_ULEB 0x80
1265
1266
1267/*
1268 * The following are used to encode binding information
1269 */
1270#define BIND_TYPE_POINTER 1
1271#define BIND_TYPE_TEXT_ABSOLUTE32 2
1272#define BIND_TYPE_TEXT_PCREL32 3
1273
1274#define BIND_SPECIAL_DYLIB_SELF 0
1275#define BIND_SPECIAL_DYLIB_MAIN_EXECUTABLE -1
1276#define BIND_SPECIAL_DYLIB_FLAT_LOOKUP -2
1277
1278#define BIND_SYMBOL_FLAGS_WEAK_IMPORT 0x1
1279#define BIND_SYMBOL_FLAGS_NON_WEAK_DEFINITION 0x8
1280
1281#define BIND_OPCODE_MASK 0xF0
1282#define BIND_IMMEDIATE_MASK 0x0F
1283#define BIND_OPCODE_DONE 0x00
1284#define BIND_OPCODE_SET_DYLIB_ORDINAL_IMM 0x10
1285#define BIND_OPCODE_SET_DYLIB_ORDINAL_ULEB 0x20
1286#define BIND_OPCODE_SET_DYLIB_SPECIAL_IMM 0x30
1287#define BIND_OPCODE_SET_SYMBOL_TRAILING_FLAGS_IMM 0x40
1288#define BIND_OPCODE_SET_TYPE_IMM 0x50
1289#define BIND_OPCODE_SET_ADDEND_SLEB 0x60
1290#define BIND_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB 0x70
1291#define BIND_OPCODE_ADD_ADDR_ULEB 0x80
1292#define BIND_OPCODE_DO_BIND 0x90
1293#define BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB 0xA0
1294#define BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED 0xB0
1295#define BIND_OPCODE_DO_BIND_ULEB_TIMES_SKIPPING_ULEB 0xC0
1296
1297
1298/*
1299 * The following are used on the flags byte of a terminal node
1300 * in the export information.
1301 */
1302#define EXPORT_SYMBOL_FLAGS_KIND_MASK 0x03
1303#define EXPORT_SYMBOL_FLAGS_KIND_REGULAR 0x00
1304#define EXPORT_SYMBOL_FLAGS_KIND_THREAD_LOCAL 0x01
1305#define EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION 0x04
1306#define EXPORT_SYMBOL_FLAGS_INDIRECT_DEFINITION 0x08
1307#define EXPORT_SYMBOL_FLAGS_HAS_SPECIALIZATIONS 0x10
1308
1c79356b
A
1309/*
1310 * The symseg_command contains the offset and size of the GNU style
1311 * symbol table information as described in the header file <symseg.h>.
1312 * The symbol roots of the symbol segments must also be aligned properly
1313 * in the file. So the requirement of keeping the offsets aligned to a
2d21ac55 1314 * multiple of a 4 bytes translates to the length field of the symbol
1c79356b
A
1315 * roots also being a multiple of a long. Also the padding must again be
1316 * zeroed. (THIS IS OBSOLETE and no longer supported).
1317 */
1318struct symseg_command {
2d21ac55
A
1319 uint32_t cmd; /* LC_SYMSEG */
1320 uint32_t cmdsize; /* sizeof(struct symseg_command) */
1321 uint32_t offset; /* symbol segment offset */
1322 uint32_t size; /* symbol segment size in bytes */
1c79356b
A
1323};
1324
1325/*
1326 * The ident_command contains a free format string table following the
1327 * ident_command structure. The strings are null terminated and the size of
2d21ac55 1328 * the command is padded out with zero bytes to a multiple of 4 bytes/
1c79356b
A
1329 * (THIS IS OBSOLETE and no longer supported).
1330 */
1331struct ident_command {
2d21ac55
A
1332 uint32_t cmd; /* LC_IDENT */
1333 uint32_t cmdsize; /* strings that follow this command */
1c79356b
A
1334};
1335
1336/*
1337 * The fvmfile_command contains a reference to a file to be loaded at the
2d21ac55 1338 * specified virtual address. (Presently, this command is reserved for
1c79356b
A
1339 * internal use. The kernel ignores this command when loading a program into
1340 * memory).
1341 */
1342struct fvmfile_command {
2d21ac55
A
1343 uint32_t cmd; /* LC_FVMFILE */
1344 uint32_t cmdsize; /* includes pathname string */
1c79356b 1345 union lc_str name; /* files pathname */
2d21ac55 1346 uint32_t header_addr; /* files virtual address */
1c79356b
A
1347};
1348
55e303ae 1349#endif /* _MACHO_LOADER_H_ */