]> git.saurik.com Git - apple/xnu.git/blame - EXTERNAL_HEADERS/mach-o/loader.h
xnu-1504.15.3.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. */
1c79356b
A
200
201/*
202 * The load commands directly follow the mach_header. The total size of all
203 * of the commands is given by the sizeofcmds field in the mach_header. All
204 * load commands must have as their first two fields cmd and cmdsize. The cmd
205 * field is filled in with a constant for that command type. Each command type
206 * has a structure specifically for it. The cmdsize field is the size in bytes
207 * of the particular load command structure plus anything that follows it that
208 * is a part of the load command (i.e. section structures, strings, etc.). To
209 * advance to the next load command the cmdsize can be added to the offset or
91447636
A
210 * pointer of the current load command. The cmdsize for 32-bit architectures
211 * MUST be a multiple of 4 bytes and for 64-bit architectures MUST be a multiple
212 * of 8 bytes (these are forever the maximum alignment of any load commands).
1c79356b
A
213 * The padded bytes must be zero. All tables in the object file must also
214 * follow these rules so the file can be memory mapped. Otherwise the pointers
215 * to these tables will not work well or at all on some machines. With all
216 * padding zeroed like objects will compare byte for byte.
217 */
218struct load_command {
2d21ac55
A
219 uint32_t cmd; /* type of load command */
220 uint32_t cmdsize; /* total size of command in bytes */
1c79356b
A
221};
222
2d21ac55
A
223/*
224 * After MacOS X 10.1 when a new load command is added that is required to be
225 * understood by the dynamic linker for the image to execute properly the
226 * LC_REQ_DYLD bit will be or'ed into the load command constant. If the dynamic
227 * linker sees such a load command it it does not understand will issue a
228 * "unknown load command required for execution" error and refuse to use the
229 * image. Other load commands without this bit that are not understood will
230 * simply be ignored.
231 */
232#define LC_REQ_DYLD 0x80000000
233
1c79356b
A
234/* Constants for the cmd field of all load commands, the type */
235#define LC_SEGMENT 0x1 /* segment of this file to be mapped */
236#define LC_SYMTAB 0x2 /* link-edit stab symbol table info */
237#define LC_SYMSEG 0x3 /* link-edit gdb symbol table info (obsolete) */
238#define LC_THREAD 0x4 /* thread */
239#define LC_UNIXTHREAD 0x5 /* unix thread (includes a stack) */
240#define LC_LOADFVMLIB 0x6 /* load a specified fixed VM shared library */
241#define LC_IDFVMLIB 0x7 /* fixed VM shared library identification */
242#define LC_IDENT 0x8 /* object identification info (obsolete) */
243#define LC_FVMFILE 0x9 /* fixed VM file inclusion (internal use) */
244#define LC_PREPAGE 0xa /* prepage command (internal use) */
245#define LC_DYSYMTAB 0xb /* dynamic link-edit symbol table info */
2d21ac55
A
246#define LC_LOAD_DYLIB 0xc /* load a dynamically linked shared library */
247#define LC_ID_DYLIB 0xd /* dynamically linked shared lib ident */
1c79356b
A
248#define LC_LOAD_DYLINKER 0xe /* load a dynamic linker */
249#define LC_ID_DYLINKER 0xf /* dynamic linker identification */
2d21ac55 250#define LC_PREBOUND_DYLIB 0x10 /* modules prebound for a dynamically */
1c79356b 251 /* linked shared library */
2d21ac55
A
252#define LC_ROUTINES 0x11 /* image routines */
253#define LC_SUB_FRAMEWORK 0x12 /* sub framework */
254#define LC_SUB_UMBRELLA 0x13 /* sub umbrella */
255#define LC_SUB_CLIENT 0x14 /* sub client */
256#define LC_SUB_LIBRARY 0x15 /* sub library */
257#define LC_TWOLEVEL_HINTS 0x16 /* two-level namespace lookup hints */
258#define LC_PREBIND_CKSUM 0x17 /* prebind checksum */
259
260/*
261 * load a dynamically linked shared library that is allowed to be missing
262 * (all symbols are weak imported).
263 */
264#define LC_LOAD_WEAK_DYLIB (0x18 | LC_REQ_DYLD)
265
266#define LC_SEGMENT_64 0x19 /* 64-bit segment of this file to be
267 mapped */
268#define LC_ROUTINES_64 0x1a /* 64-bit image routines */
269#define LC_UUID 0x1b /* the uuid */
270#define LC_RPATH (0x1c | LC_REQ_DYLD) /* runpath additions */
271#define LC_CODE_SIGNATURE 0x1d /* local of code signature */
272#define LC_SEGMENT_SPLIT_INFO 0x1e /* local of info to split segments */
273#define LC_REEXPORT_DYLIB (0x1f | LC_REQ_DYLD) /* load and re-export dylib */
593a1d5f
A
274#define LC_LAZY_LOAD_DYLIB 0x20 /* delay load of dylib until first use */
275#define LC_ENCRYPTION_INFO 0x21 /* encrypted segment information */
b0d623f7
A
276#define LC_DYLD_INFO 0x22 /* compressed dyld information */
277#define LC_DYLD_INFO_ONLY (0x22|LC_REQ_DYLD) /* compressed dyld information only */
1c79356b
A
278
279/*
280 * A variable length string in a load command is represented by an lc_str
281 * union. The strings are stored just after the load command structure and
282 * the offset is from the start of the load command structure. The size
283 * of the string is reflected in the cmdsize field of the load command.
284 * Once again any padded bytes to bring the cmdsize field to a multiple
2d21ac55 285 * of 4 bytes must be zero.
1c79356b
A
286 */
287union lc_str {
2d21ac55
A
288 uint32_t offset; /* offset to the string */
289#ifndef __LP64__
1c79356b 290 char *ptr; /* pointer to the string */
2d21ac55 291#endif
1c79356b
A
292};
293
294/*
295 * The segment load command indicates that a part of this file is to be
296 * mapped into the task's address space. The size of this segment in memory,
297 * vmsize, maybe equal to or larger than the amount to map from this file,
298 * filesize. The file is mapped starting at fileoff to the beginning of
299 * the segment in memory, vmaddr. The rest of the memory of the segment,
300 * if any, is allocated zero fill on demand. The segment's maximum virtual
301 * memory protection and initial virtual memory protection are specified
302 * by the maxprot and initprot fields. If the segment has sections then the
303 * section structures directly follow the segment command and their size is
304 * reflected in cmdsize.
305 */
2d21ac55
A
306struct segment_command { /* for 32-bit architectures */
307 uint32_t cmd; /* LC_SEGMENT */
308 uint32_t cmdsize; /* includes sizeof section structs */
1c79356b 309 char segname[16]; /* segment name */
2d21ac55
A
310 uint32_t vmaddr; /* memory address of this segment */
311 uint32_t vmsize; /* memory size of this segment */
312 uint32_t fileoff; /* file offset of this segment */
313 uint32_t filesize; /* amount to map from the file */
1c79356b
A
314 vm_prot_t maxprot; /* maximum VM protection */
315 vm_prot_t initprot; /* initial VM protection */
2d21ac55
A
316 uint32_t nsects; /* number of sections in segment */
317 uint32_t flags; /* flags */
1c79356b
A
318};
319
91447636
A
320/*
321 * The 64-bit segment load command indicates that a part of this file is to be
322 * mapped into a 64-bit task's address space. If the 64-bit segment has
323 * sections then section_64 structures directly follow the 64-bit segment
324 * command and their size is reflected in cmdsize.
325 */
2d21ac55 326struct segment_command_64 { /* for 64-bit architectures */
91447636
A
327 uint32_t cmd; /* LC_SEGMENT_64 */
328 uint32_t cmdsize; /* includes sizeof section_64 structs */
329 char segname[16]; /* segment name */
330 uint64_t vmaddr; /* memory address of this segment */
331 uint64_t vmsize; /* memory size of this segment */
332 uint64_t fileoff; /* file offset of this segment */
333 uint64_t filesize; /* amount to map from the file */
334 vm_prot_t maxprot; /* maximum VM protection */
335 vm_prot_t initprot; /* initial VM protection */
336 uint32_t nsects; /* number of sections in segment */
337 uint32_t flags; /* flags */
338};
339
1c79356b
A
340/* Constants for the flags field of the segment_command */
341#define SG_HIGHVM 0x1 /* the file contents for this segment is for
342 the high part of the VM space, the low part
343 is zero filled (for stacks in core files) */
344#define SG_FVMLIB 0x2 /* this segment is the VM that is allocated by
345 a fixed VM library, for overlap checking in
346 the link editor */
347#define SG_NORELOC 0x4 /* this segment has nothing that was relocated
348 in it and nothing relocated to it, that is
349 it maybe safely replaced without relocation*/
2d21ac55
A
350#define SG_PROTECTED_VERSION_1 0x8 /* This segment is protected. If the
351 segment starts at file offset 0, the
352 first page of the segment is not
353 protected. All other pages of the
354 segment are protected. */
1c79356b
A
355
356/*
357 * A segment is made up of zero or more sections. Non-MH_OBJECT files have
358 * all of their segments with the proper sections in each, and padded to the
359 * specified segment alignment when produced by the link editor. The first
360 * segment of a MH_EXECUTE and MH_FVMLIB format file contains the mach_header
2d21ac55 361 * and load commands of the object file before its first section. The zero
1c79356b
A
362 * fill sections are always last in their segment (in all formats). This
363 * allows the zeroed segment padding to be mapped into memory where zero fill
91447636
A
364 * sections might be. The gigabyte zero fill sections, those with the section
365 * type S_GB_ZEROFILL, can only be in a segment with sections of this type.
366 * These segments are then placed after all other segments.
1c79356b 367 *
2d21ac55 368 * The MH_OBJECT format has all of its sections in one segment for
1c79356b
A
369 * compactness. There is no padding to a specified segment boundary and the
370 * mach_header and load commands are not part of the segment.
371 *
372 * Sections with the same section name, sectname, going into the same segment,
373 * segname, are combined by the link editor. The resulting section is aligned
374 * to the maximum alignment of the combined sections and is the new section's
375 * alignment. The combined sections are aligned to their original alignment in
376 * the combined section. Any padded bytes to get the specified alignment are
377 * zeroed.
378 *
379 * The format of the relocation entries referenced by the reloff and nreloc
380 * fields of the section structure for mach object files is described in the
381 * header file <reloc.h>.
382 */
2d21ac55 383struct section { /* for 32-bit architectures */
1c79356b
A
384 char sectname[16]; /* name of this section */
385 char segname[16]; /* segment this section goes in */
2d21ac55
A
386 uint32_t addr; /* memory address of this section */
387 uint32_t size; /* size in bytes of this section */
388 uint32_t offset; /* file offset of this section */
389 uint32_t align; /* section alignment (power of 2) */
390 uint32_t reloff; /* file offset of relocation entries */
391 uint32_t nreloc; /* number of relocation entries */
392 uint32_t flags; /* flags (section type and attributes)*/
393 uint32_t reserved1; /* reserved (for offset or index) */
394 uint32_t reserved2; /* reserved (for count or sizeof) */
1c79356b
A
395};
396
91447636
A
397struct section_64 { /* for 64-bit architectures */
398 char sectname[16]; /* name of this section */
399 char segname[16]; /* segment this section goes in */
400 uint64_t addr; /* memory address of this section */
401 uint64_t size; /* size in bytes of this section */
402 uint32_t offset; /* file offset of this section */
403 uint32_t align; /* section alignment (power of 2) */
404 uint32_t reloff; /* file offset of relocation entries */
405 uint32_t nreloc; /* number of relocation entries */
406 uint32_t flags; /* flags (section type and attributes)*/
407 uint32_t reserved1; /* reserved (for offset or index) */
408 uint32_t reserved2; /* reserved (for count or sizeof) */
409 uint32_t reserved3; /* reserved */
410};
411
1c79356b
A
412/*
413 * The flags field of a section structure is separated into two parts a section
414 * type and section attributes. The section types are mutually exclusive (it
415 * can only have one type) but the section attributes are not (it may have more
416 * than one attribute).
417 */
418#define SECTION_TYPE 0x000000ff /* 256 section types */
419#define SECTION_ATTRIBUTES 0xffffff00 /* 24 section attributes */
420
421/* Constants for the type of a section */
422#define S_REGULAR 0x0 /* regular section */
423#define S_ZEROFILL 0x1 /* zero fill on demand section */
424#define S_CSTRING_LITERALS 0x2 /* section with only literal C strings*/
425#define S_4BYTE_LITERALS 0x3 /* section with only 4 byte literals */
426#define S_8BYTE_LITERALS 0x4 /* section with only 8 byte literals */
427#define S_LITERAL_POINTERS 0x5 /* section with only pointers to */
428 /* literals */
429/*
430 * For the two types of symbol pointers sections and the symbol stubs section
431 * they have indirect symbol table entries. For each of the entries in the
432 * section the indirect symbol table entries, in corresponding order in the
433 * indirect symbol table, start at the index stored in the reserved1 field
434 * of the section structure. Since the indirect symbol table entries
435 * correspond to the entries in the section the number of indirect symbol table
436 * entries is inferred from the size of the section divided by the size of the
437 * entries in the section. For symbol pointers sections the size of the entries
438 * in the section is 4 bytes and for symbol stubs sections the byte size of the
439 * stubs is stored in the reserved2 field of the section structure.
440 */
441#define S_NON_LAZY_SYMBOL_POINTERS 0x6 /* section with only non-lazy
442 symbol pointers */
443#define S_LAZY_SYMBOL_POINTERS 0x7 /* section with only lazy symbol
444 pointers */
445#define S_SYMBOL_STUBS 0x8 /* section with only symbol
446 stubs, byte size of stub in
447 the reserved2 field */
448#define S_MOD_INIT_FUNC_POINTERS 0x9 /* section with only function
449 pointers for initialization*/
2d21ac55
A
450#define S_MOD_TERM_FUNC_POINTERS 0xa /* section with only function
451 pointers for termination */
452#define S_COALESCED 0xb /* section contains symbols that
453 are to be coalesced */
454#define S_GB_ZEROFILL 0xc /* zero fill on demand section
455 (that can be larger than 4
456 gigabytes) */
457#define S_INTERPOSING 0xd /* section with only pairs of
458 function pointers for
459 interposing */
593a1d5f
A
460#define S_16BYTE_LITERALS 0xe /* section with only 16 byte
461 literals */
462#define S_DTRACE_DOF 0xf /* section contains
463 DTrace Object Format */
464#define S_LAZY_DYLIB_SYMBOL_POINTERS 0x10 /* section with only lazy
465 symbol pointers to lazy
466 loaded dylibs */
1c79356b
A
467/*
468 * Constants for the section attributes part of the flags field of a section
469 * structure.
470 */
471#define SECTION_ATTRIBUTES_USR 0xff000000 /* User setable attributes */
472#define S_ATTR_PURE_INSTRUCTIONS 0x80000000 /* section contains only true
473 machine instructions */
2d21ac55
A
474#define S_ATTR_NO_TOC 0x40000000 /* section contains coalesced
475 symbols that are not to be
476 in a ranlib table of
477 contents */
478#define S_ATTR_STRIP_STATIC_SYMS 0x20000000 /* ok to strip static symbols
479 in this section in files
480 with the MH_DYLDLINK flag */
481#define S_ATTR_NO_DEAD_STRIP 0x10000000 /* no dead stripping */
482#define S_ATTR_LIVE_SUPPORT 0x08000000 /* blocks are live if they
483 reference live blocks */
484#define S_ATTR_SELF_MODIFYING_CODE 0x04000000 /* Used with i386 code stubs
485 written on by dyld */
486/*
487 * If a segment contains any sections marked with S_ATTR_DEBUG then all
488 * sections in that segment must have this attribute. No section other than
489 * a section marked with this attribute may reference the contents of this
490 * section. A section with this attribute may contain no symbols and must have
491 * a section type S_REGULAR. The static linker will not copy section contents
492 * from sections with this attribute into its output file. These sections
493 * generally contain DWARF debugging info.
494 */
495#define S_ATTR_DEBUG 0x02000000 /* a debug section */
1c79356b
A
496#define SECTION_ATTRIBUTES_SYS 0x00ffff00 /* system setable attributes */
497#define S_ATTR_SOME_INSTRUCTIONS 0x00000400 /* section contains some
498 machine instructions */
499#define S_ATTR_EXT_RELOC 0x00000200 /* section has external
500 relocation entries */
501#define S_ATTR_LOC_RELOC 0x00000100 /* section has local
502 relocation entries */
503
504
505/*
506 * The names of segments and sections in them are mostly meaningless to the
507 * link-editor. But there are few things to support traditional UNIX
508 * executables that require the link-editor and assembler to use some names
509 * agreed upon by convention.
510 *
511 * The initial protection of the "__TEXT" segment has write protection turned
512 * off (not writeable).
513 *
514 * The link-editor will allocate common symbols at the end of the "__common"
515 * section in the "__DATA" segment. It will create the section and segment
516 * if needed.
517 */
518
519/* The currently known segment names and the section names in those segments */
520
521#define SEG_PAGEZERO "__PAGEZERO" /* the pagezero segment which has no */
522 /* protections and catches NULL */
523 /* references for MH_EXECUTE files */
524
525
526#define SEG_TEXT "__TEXT" /* the tradition UNIX text segment */
527#define SECT_TEXT "__text" /* the real text part of the text */
528 /* section no headers, and no padding */
529#define SECT_FVMLIB_INIT0 "__fvmlib_init0" /* the fvmlib initialization */
530 /* section */
531#define SECT_FVMLIB_INIT1 "__fvmlib_init1" /* the section following the */
532 /* fvmlib initialization */
533 /* section */
534
535#define SEG_DATA "__DATA" /* the tradition UNIX data segment */
536#define SECT_DATA "__data" /* the real initialized data section */
537 /* no padding, no bss overlap */
538#define SECT_BSS "__bss" /* the real uninitialized data section*/
539 /* no padding */
540#define SECT_COMMON "__common" /* the section common symbols are */
541 /* allocated in by the link editor */
542
543#define SEG_OBJC "__OBJC" /* objective-C runtime segment */
544#define SECT_OBJC_SYMBOLS "__symbol_table" /* symbol table */
545#define SECT_OBJC_MODULES "__module_info" /* module information */
546#define SECT_OBJC_STRINGS "__selector_strs" /* string table */
547#define SECT_OBJC_REFS "__selector_refs" /* string table */
548
2d21ac55 549#define SEG_ICON "__ICON" /* the icon segment */
1c79356b
A
550#define SECT_ICON_HEADER "__header" /* the icon headers */
551#define SECT_ICON_TIFF "__tiff" /* the icons in tiff format */
552
553#define SEG_LINKEDIT "__LINKEDIT" /* the segment containing all structs */
554 /* created and maintained by the link */
555 /* editor. Created with -seglinkedit */
556 /* option to ld(1) for MH_EXECUTE and */
557 /* FVMLIB file types only */
558
559#define SEG_UNIXSTACK "__UNIXSTACK" /* the unix stack segment */
560
2d21ac55
A
561#define SEG_IMPORT "__IMPORT" /* the segment for the self (dyld) */
562 /* modifing code stubs that has read, */
563 /* write and execute permissions */
564
1c79356b
A
565/*
566 * Fixed virtual memory shared libraries are identified by two things. The
567 * target pathname (the name of the library as found for execution), and the
568 * minor version number. The address of where the headers are loaded is in
2d21ac55 569 * header_addr. (THIS IS OBSOLETE and no longer supported).
1c79356b
A
570 */
571struct fvmlib {
572 union lc_str name; /* library's target pathname */
2d21ac55
A
573 uint32_t minor_version; /* library's minor version number */
574 uint32_t header_addr; /* library's header address */
1c79356b
A
575};
576
577/*
578 * A fixed virtual shared library (filetype == MH_FVMLIB in the mach header)
579 * contains a fvmlib_command (cmd == LC_IDFVMLIB) to identify the library.
580 * An object that uses a fixed virtual shared library also contains a
581 * fvmlib_command (cmd == LC_LOADFVMLIB) for each library it uses.
2d21ac55 582 * (THIS IS OBSOLETE and no longer supported).
1c79356b
A
583 */
584struct fvmlib_command {
2d21ac55
A
585 uint32_t cmd; /* LC_IDFVMLIB or LC_LOADFVMLIB */
586 uint32_t cmdsize; /* includes pathname string */
1c79356b
A
587 struct fvmlib fvmlib; /* the library identification */
588};
589
590/*
591 * Dynamicly linked shared libraries are identified by two things. The
592 * pathname (the name of the library as found for execution), and the
593 * compatibility version number. The pathname must match and the compatibility
594 * number in the user of the library must be greater than or equal to the
595 * library being used. The time stamp is used to record the time a library was
596 * built and copied into user so it can be use to determined if the library used
597 * at runtime is exactly the same as used to built the program.
598 */
599struct dylib {
600 union lc_str name; /* library's path name */
2d21ac55
A
601 uint32_t timestamp; /* library's build time stamp */
602 uint32_t current_version; /* library's current version number */
603 uint32_t compatibility_version; /* library's compatibility vers number*/
1c79356b
A
604};
605
606/*
2d21ac55 607 * A dynamically linked shared library (filetype == MH_DYLIB in the mach header)
1c79356b 608 * contains a dylib_command (cmd == LC_ID_DYLIB) to identify the library.
2d21ac55
A
609 * An object that uses a dynamically linked shared library also contains a
610 * dylib_command (cmd == LC_LOAD_DYLIB, LC_LOAD_WEAK_DYLIB, or
611 * LC_REEXPORT_DYLIB) for each library it uses.
1c79356b
A
612 */
613struct dylib_command {
2d21ac55
A
614 uint32_t cmd; /* LC_ID_DYLIB, LC_LOAD_{,WEAK_}DYLIB,
615 LC_REEXPORT_DYLIB */
616 uint32_t cmdsize; /* includes pathname string */
1c79356b
A
617 struct dylib dylib; /* the library identification */
618};
619
620/*
2d21ac55
A
621 * A dynamically linked shared library may be a subframework of an umbrella
622 * framework. If so it will be linked with "-umbrella umbrella_name" where
623 * Where "umbrella_name" is the name of the umbrella framework. A subframework
624 * can only be linked against by its umbrella framework or other subframeworks
625 * that are part of the same umbrella framework. Otherwise the static link
626 * editor produces an error and states to link against the umbrella framework.
627 * The name of the umbrella framework for subframeworks is recorded in the
628 * following structure.
629 */
630struct sub_framework_command {
631 uint32_t cmd; /* LC_SUB_FRAMEWORK */
632 uint32_t cmdsize; /* includes umbrella string */
633 union lc_str umbrella; /* the umbrella framework name */
634};
635
636/*
637 * For dynamically linked shared libraries that are subframework of an umbrella
638 * framework they can allow clients other than the umbrella framework or other
639 * subframeworks in the same umbrella framework. To do this the subframework
640 * is built with "-allowable_client client_name" and an LC_SUB_CLIENT load
641 * command is created for each -allowable_client flag. The client_name is
642 * usually a framework name. It can also be a name used for bundles clients
643 * where the bundle is built with "-client_name client_name".
644 */
645struct sub_client_command {
646 uint32_t cmd; /* LC_SUB_CLIENT */
647 uint32_t cmdsize; /* includes client string */
648 union lc_str client; /* the client name */
649};
650
651/*
652 * A dynamically linked shared library may be a sub_umbrella of an umbrella
653 * framework. If so it will be linked with "-sub_umbrella umbrella_name" where
654 * Where "umbrella_name" is the name of the sub_umbrella framework. When
655 * staticly linking when -twolevel_namespace is in effect a twolevel namespace
656 * umbrella framework will only cause its subframeworks and those frameworks
657 * listed as sub_umbrella frameworks to be implicited linked in. Any other
658 * dependent dynamic libraries will not be linked it when -twolevel_namespace
659 * is in effect. The primary library recorded by the static linker when
660 * resolving a symbol in these libraries will be the umbrella framework.
661 * Zero or more sub_umbrella frameworks may be use by an umbrella framework.
662 * The name of a sub_umbrella framework is recorded in the following structure.
663 */
664struct sub_umbrella_command {
665 uint32_t cmd; /* LC_SUB_UMBRELLA */
666 uint32_t cmdsize; /* includes sub_umbrella string */
667 union lc_str sub_umbrella; /* the sub_umbrella framework name */
668};
669
670/*
671 * A dynamically linked shared library may be a sub_library of another shared
672 * library. If so it will be linked with "-sub_library library_name" where
673 * Where "library_name" is the name of the sub_library shared library. When
674 * staticly linking when -twolevel_namespace is in effect a twolevel namespace
675 * shared library will only cause its subframeworks and those frameworks
676 * listed as sub_umbrella frameworks and libraries listed as sub_libraries to
677 * be implicited linked in. Any other dependent dynamic libraries will not be
678 * linked it when -twolevel_namespace is in effect. The primary library
679 * recorded by the static linker when resolving a symbol in these libraries
680 * will be the umbrella framework (or dynamic library). Zero or more sub_library
681 * shared libraries may be use by an umbrella framework or (or dynamic library).
682 * The name of a sub_library framework is recorded in the following structure.
683 * For example /usr/lib/libobjc_profile.A.dylib would be recorded as "libobjc".
684 */
685struct sub_library_command {
686 uint32_t cmd; /* LC_SUB_LIBRARY */
687 uint32_t cmdsize; /* includes sub_library string */
688 union lc_str sub_library; /* the sub_library name */
689};
690
691/*
692 * A program (filetype == MH_EXECUTE) that is
693 * prebound to its dynamic libraries has one of these for each library that
1c79356b
A
694 * the static linker used in prebinding. It contains a bit vector for the
695 * modules in the library. The bits indicate which modules are bound (1) and
696 * which are not (0) from the library. The bit for module 0 is the low bit
697 * of the first byte. So the bit for the Nth module is:
698 * (linked_modules[N/8] >> N%8) & 1
699 */
700struct prebound_dylib_command {
2d21ac55
A
701 uint32_t cmd; /* LC_PREBOUND_DYLIB */
702 uint32_t cmdsize; /* includes strings */
1c79356b 703 union lc_str name; /* library's path name */
2d21ac55 704 uint32_t nmodules; /* number of modules in library */
1c79356b
A
705 union lc_str linked_modules; /* bit vector of linked modules */
706};
707
708/*
709 * A program that uses a dynamic linker contains a dylinker_command to identify
710 * the name of the dynamic linker (LC_LOAD_DYLINKER). And a dynamic linker
711 * contains a dylinker_command to identify the dynamic linker (LC_ID_DYLINKER).
712 * A file can have at most one of these.
713 */
714struct dylinker_command {
2d21ac55
A
715 uint32_t cmd; /* LC_ID_DYLINKER or LC_LOAD_DYLINKER */
716 uint32_t cmdsize; /* includes pathname string */
1c79356b
A
717 union lc_str name; /* dynamic linker's path name */
718};
719
720/*
721 * Thread commands contain machine-specific data structures suitable for
722 * use in the thread state primitives. The machine specific data structures
723 * follow the struct thread_command as follows.
724 * Each flavor of machine specific data structure is preceded by an unsigned
2d21ac55 725 * long constant for the flavor of that data structure, an uint32_t
1c79356b
A
726 * that is the count of longs of the size of the state data structure and then
727 * the state data structure follows. This triple may be repeated for many
728 * flavors. The constants for the flavors, counts and state data structure
729 * definitions are expected to be in the header file <machine/thread_status.h>.
730 * These machine specific data structures sizes must be multiples of
2d21ac55 731 * 4 bytes The cmdsize reflects the total size of the thread_command
1c79356b
A
732 * and all of the sizes of the constants for the flavors, counts and state
733 * data structures.
734 *
735 * For executable objects that are unix processes there will be one
736 * thread_command (cmd == LC_UNIXTHREAD) created for it by the link-editor.
737 * This is the same as a LC_THREAD, except that a stack is automatically
738 * created (based on the shell's limit for the stack size). Command arguments
739 * and environment variables are copied onto that stack.
740 */
741struct thread_command {
2d21ac55
A
742 uint32_t cmd; /* LC_THREAD or LC_UNIXTHREAD */
743 uint32_t cmdsize; /* total size of this command */
744 /* uint32_t flavor flavor of thread state */
745 /* uint32_t count count of longs in thread state */
1c79356b
A
746 /* struct XXX_thread_state state thread state for this flavor */
747 /* ... */
748};
749
2d21ac55
A
750/*
751 * The routines command contains the address of the dynamic shared library
752 * initialization routine and an index into the module table for the module
753 * that defines the routine. Before any modules are used from the library the
754 * dynamic linker fully binds the module that defines the initialization routine
755 * and then calls it. This gets called before any module initialization
756 * routines (used for C++ static constructors) in the library.
757 */
758struct routines_command { /* for 32-bit architectures */
759 uint32_t cmd; /* LC_ROUTINES */
760 uint32_t cmdsize; /* total size of this command */
761 uint32_t init_address; /* address of initialization routine */
762 uint32_t init_module; /* index into the module table that */
763 /* the init routine is defined in */
764 uint32_t reserved1;
765 uint32_t reserved2;
766 uint32_t reserved3;
767 uint32_t reserved4;
768 uint32_t reserved5;
769 uint32_t reserved6;
770};
771
772/*
773 * The 64-bit routines command. Same use as above.
774 */
775struct routines_command_64 { /* for 64-bit architectures */
776 uint32_t cmd; /* LC_ROUTINES_64 */
777 uint32_t cmdsize; /* total size of this command */
778 uint64_t init_address; /* address of initialization routine */
779 uint64_t init_module; /* index into the module table that */
780 /* the init routine is defined in */
781 uint64_t reserved1;
782 uint64_t reserved2;
783 uint64_t reserved3;
784 uint64_t reserved4;
785 uint64_t reserved5;
786 uint64_t reserved6;
787};
788
1c79356b
A
789/*
790 * The symtab_command contains the offsets and sizes of the link-edit 4.3BSD
791 * "stab" style symbol table information as described in the header files
792 * <nlist.h> and <stab.h>.
793 */
794struct symtab_command {
2d21ac55
A
795 uint32_t cmd; /* LC_SYMTAB */
796 uint32_t cmdsize; /* sizeof(struct symtab_command) */
797 uint32_t symoff; /* symbol table offset */
798 uint32_t nsyms; /* number of symbol table entries */
799 uint32_t stroff; /* string table offset */
800 uint32_t strsize; /* string table size in bytes */
1c79356b
A
801};
802
803/*
804 * This is the second set of the symbolic information which is used to support
2d21ac55 805 * the data structures for the dynamically link editor.
1c79356b
A
806 *
807 * The original set of symbolic information in the symtab_command which contains
808 * the symbol and string tables must also be present when this load command is
809 * present. When this load command is present the symbol table is organized
810 * into three groups of symbols:
811 * local symbols (static and debugging symbols) - grouped by module
812 * defined external symbols - grouped by module (sorted by name if not lib)
2d21ac55
A
813 * undefined external symbols (sorted by name if MH_BINDATLOAD is not set,
814 * and in order the were seen by the static
815 * linker if MH_BINDATLOAD is set)
1c79356b
A
816 * In this load command there are offsets and counts to each of the three groups
817 * of symbols.
818 *
819 * This load command contains a the offsets and sizes of the following new
820 * symbolic information tables:
821 * table of contents
822 * module table
823 * reference symbol table
824 * indirect symbol table
825 * The first three tables above (the table of contents, module table and
2d21ac55 826 * reference symbol table) are only present if the file is a dynamically linked
1c79356b
A
827 * shared library. For executable and object modules, which are files
828 * containing only one module, the information that would be in these three
829 * tables is determined as follows:
830 * table of contents - the defined external symbols are sorted by name
831 * module table - the file contains only one module so everything in the
832 * file is part of the module.
833 * reference symbol table - is the defined and undefined external symbols
834 *
2d21ac55 835 * For dynamically linked shared library files this load command also contains
1c79356b
A
836 * offsets and sizes to the pool of relocation entries for all sections
837 * separated into two groups:
838 * external relocation entries
839 * local relocation entries
840 * For executable and object modules the relocation entries continue to hang
841 * off the section structures.
842 */
843struct dysymtab_command {
2d21ac55
A
844 uint32_t cmd; /* LC_DYSYMTAB */
845 uint32_t cmdsize; /* sizeof(struct dysymtab_command) */
1c79356b
A
846
847 /*
848 * The symbols indicated by symoff and nsyms of the LC_SYMTAB load command
849 * are grouped into the following three groups:
850 * local symbols (further grouped by the module they are from)
851 * defined external symbols (further grouped by the module they are from)
852 * undefined symbols
853 *
854 * The local symbols are used only for debugging. The dynamic binding
855 * process may have to use them to indicate to the debugger the local
856 * symbols for a module that is being bound.
857 *
858 * The last two groups are used by the dynamic binding process to do the
859 * binding (indirectly through the module table and the reference symbol
2d21ac55 860 * table when this is a dynamically linked shared library file).
1c79356b 861 */
2d21ac55
A
862 uint32_t ilocalsym; /* index to local symbols */
863 uint32_t nlocalsym; /* number of local symbols */
1c79356b 864
2d21ac55
A
865 uint32_t iextdefsym;/* index to externally defined symbols */
866 uint32_t nextdefsym;/* number of externally defined symbols */
1c79356b 867
2d21ac55
A
868 uint32_t iundefsym; /* index to undefined symbols */
869 uint32_t nundefsym; /* number of undefined symbols */
1c79356b
A
870
871 /*
872 * For the for the dynamic binding process to find which module a symbol
873 * is defined in the table of contents is used (analogous to the ranlib
874 * structure in an archive) which maps defined external symbols to modules
2d21ac55 875 * they are defined in. This exists only in a dynamically linked shared
1c79356b
A
876 * library file. For executable and object modules the defined external
877 * symbols are sorted by name and is use as the table of contents.
878 */
2d21ac55
A
879 uint32_t tocoff; /* file offset to table of contents */
880 uint32_t ntoc; /* number of entries in table of contents */
1c79356b
A
881
882 /*
883 * To support dynamic binding of "modules" (whole object files) the symbol
884 * table must reflect the modules that the file was created from. This is
885 * done by having a module table that has indexes and counts into the merged
886 * tables for each module. The module structure that these two entries
2d21ac55 887 * refer to is described below. This exists only in a dynamically linked
1c79356b
A
888 * shared library file. For executable and object modules the file only
889 * contains one module so everything in the file belongs to the module.
890 */
2d21ac55
A
891 uint32_t modtaboff; /* file offset to module table */
892 uint32_t nmodtab; /* number of module table entries */
1c79356b
A
893
894 /*
895 * To support dynamic module binding the module structure for each module
896 * indicates the external references (defined and undefined) each module
897 * makes. For each module there is an offset and a count into the
898 * reference symbol table for the symbols that the module references.
2d21ac55 899 * This exists only in a dynamically linked shared library file. For
1c79356b
A
900 * executable and object modules the defined external symbols and the
901 * undefined external symbols indicates the external references.
902 */
2d21ac55
A
903 uint32_t extrefsymoff; /* offset to referenced symbol table */
904 uint32_t nextrefsyms; /* number of referenced symbol table entries */
1c79356b
A
905
906 /*
907 * The sections that contain "symbol pointers" and "routine stubs" have
908 * indexes and (implied counts based on the size of the section and fixed
909 * size of the entry) into the "indirect symbol" table for each pointer
910 * and stub. For every section of these two types the index into the
911 * indirect symbol table is stored in the section header in the field
912 * reserved1. An indirect symbol table entry is simply a 32bit index into
913 * the symbol table to the symbol that the pointer or stub is referring to.
914 * The indirect symbol table is ordered to match the entries in the section.
915 */
2d21ac55
A
916 uint32_t indirectsymoff; /* file offset to the indirect symbol table */
917 uint32_t nindirectsyms; /* number of indirect symbol table entries */
1c79356b
A
918
919 /*
920 * To support relocating an individual module in a library file quickly the
921 * external relocation entries for each module in the library need to be
922 * accessed efficiently. Since the relocation entries can't be accessed
923 * through the section headers for a library file they are separated into
924 * groups of local and external entries further grouped by module. In this
925 * case the presents of this load command who's extreloff, nextrel,
926 * locreloff and nlocrel fields are non-zero indicates that the relocation
927 * entries of non-merged sections are not referenced through the section
928 * structures (and the reloff and nreloc fields in the section headers are
929 * set to zero).
930 *
931 * Since the relocation entries are not accessed through the section headers
932 * this requires the r_address field to be something other than a section
933 * offset to identify the item to be relocated. In this case r_address is
934 * set to the offset from the vmaddr of the first LC_SEGMENT command.
2d21ac55
A
935 * For MH_SPLIT_SEGS images r_address is set to the the offset from the
936 * vmaddr of the first read-write LC_SEGMENT command.
1c79356b
A
937 *
938 * The relocation entries are grouped by module and the module table
939 * entries have indexes and counts into them for the group of external
940 * relocation entries for that the module.
941 *
942 * For sections that are merged across modules there must not be any
943 * remaining external relocation entries for them (for merged sections
944 * remaining relocation entries must be local).
945 */
2d21ac55
A
946 uint32_t extreloff; /* offset to external relocation entries */
947 uint32_t nextrel; /* number of external relocation entries */
1c79356b
A
948
949 /*
950 * All the local relocation entries are grouped together (they are not
951 * grouped by their module since they are only used if the object is moved
952 * from it staticly link edited address).
953 */
2d21ac55
A
954 uint32_t locreloff; /* offset to local relocation entries */
955 uint32_t nlocrel; /* number of local relocation entries */
1c79356b
A
956
957};
958
959/*
960 * An indirect symbol table entry is simply a 32bit index into the symbol table
961 * to the symbol that the pointer or stub is refering to. Unless it is for a
962 * non-lazy symbol pointer section for a defined symbol which strip(1) as
963 * removed. In which case it has the value INDIRECT_SYMBOL_LOCAL. If the
964 * symbol was also absolute INDIRECT_SYMBOL_ABS is or'ed with that.
965 */
966#define INDIRECT_SYMBOL_LOCAL 0x80000000
967#define INDIRECT_SYMBOL_ABS 0x40000000
968
969
970/* a table of contents entry */
971struct dylib_table_of_contents {
2d21ac55 972 uint32_t symbol_index; /* the defined external symbol
1c79356b 973 (index into the symbol table) */
2d21ac55 974 uint32_t module_index; /* index into the module table this symbol
1c79356b
A
975 is defined in */
976};
977
978/* a module table entry */
979struct dylib_module {
2d21ac55 980 uint32_t module_name; /* the module name (index into string table) */
1c79356b 981
2d21ac55
A
982 uint32_t iextdefsym; /* index into externally defined symbols */
983 uint32_t nextdefsym; /* number of externally defined symbols */
984 uint32_t irefsym; /* index into reference symbol table */
985 uint32_t nrefsym; /* number of reference symbol table entries */
986 uint32_t ilocalsym; /* index into symbols for local symbols */
987 uint32_t nlocalsym; /* number of local symbols */
1c79356b 988
2d21ac55
A
989 uint32_t iextrel; /* index into external relocation entries */
990 uint32_t nextrel; /* number of external relocation entries */
1c79356b 991
2d21ac55
A
992 uint32_t iinit_iterm; /* low 16 bits are the index into the init
993 section, high 16 bits are the index into
994 the term section */
995 uint32_t ninit_nterm; /* low 16 bits are the number of init section
996 entries, high 16 bits are the number of
997 term section entries */
1c79356b 998
2d21ac55 999 uint32_t /* for this module address of the start of */
1c79356b 1000 objc_module_info_addr; /* the (__OBJC,__module_info) section */
2d21ac55 1001 uint32_t /* for this module size of */
1c79356b
A
1002 objc_module_info_size; /* the (__OBJC,__module_info) section */
1003};
1004
91447636
A
1005/* a 64-bit module table entry */
1006struct dylib_module_64 {
2d21ac55 1007 uint32_t module_name; /* the module name (index into string table) */
91447636 1008
2d21ac55
A
1009 uint32_t iextdefsym; /* index into externally defined symbols */
1010 uint32_t nextdefsym; /* number of externally defined symbols */
1011 uint32_t irefsym; /* index into reference symbol table */
1012 uint32_t nrefsym; /* number of reference symbol table entries */
1013 uint32_t ilocalsym; /* index into symbols for local symbols */
1014 uint32_t nlocalsym; /* number of local symbols */
91447636 1015
2d21ac55
A
1016 uint32_t iextrel; /* index into external relocation entries */
1017 uint32_t nextrel; /* number of external relocation entries */
91447636 1018
2d21ac55 1019 uint32_t iinit_iterm; /* low 16 bits are the index into the init
91447636
A
1020 section, high 16 bits are the index into
1021 the term section */
2d21ac55
A
1022 uint32_t ninit_nterm; /* low 16 bits are the number of init section
1023 entries, high 16 bits are the number of
1024 term section entries */
1025
1026 uint32_t /* for this module size of */
1027 objc_module_info_size; /* the (__OBJC,__module_info) section */
1028 uint64_t /* for this module address of the start of */
1029 objc_module_info_addr; /* the (__OBJC,__module_info) section */
91447636
A
1030};
1031
1c79356b
A
1032/*
1033 * The entries in the reference symbol table are used when loading the module
1034 * (both by the static and dynamic link editors) and if the module is unloaded
1035 * or replaced. Therefore all external symbols (defined and undefined) are
1036 * listed in the module's reference table. The flags describe the type of
1037 * reference that is being made. The constants for the flags are defined in
1038 * <mach-o/nlist.h> as they are also used for symbol table entries.
1039 */
1040struct dylib_reference {
2d21ac55 1041 uint32_t isym:24, /* index into the symbol table */
1c79356b
A
1042 flags:8; /* flags to indicate the type of reference */
1043};
1044
2d21ac55
A
1045/*
1046 * The twolevel_hints_command contains the offset and number of hints in the
1047 * two-level namespace lookup hints table.
1048 */
1049struct twolevel_hints_command {
1050 uint32_t cmd; /* LC_TWOLEVEL_HINTS */
1051 uint32_t cmdsize; /* sizeof(struct twolevel_hints_command) */
1052 uint32_t offset; /* offset to the hint table */
1053 uint32_t nhints; /* number of hints in the hint table */
1054};
1055
1056/*
1057 * The entries in the two-level namespace lookup hints table are twolevel_hint
1058 * structs. These provide hints to the dynamic link editor where to start
1059 * looking for an undefined symbol in a two-level namespace image. The
1060 * isub_image field is an index into the sub-images (sub-frameworks and
1061 * sub-umbrellas list) that made up the two-level image that the undefined
1062 * symbol was found in when it was built by the static link editor. If
1063 * isub-image is 0 the the symbol is expected to be defined in library and not
1064 * in the sub-images. If isub-image is non-zero it is an index into the array
1065 * of sub-images for the umbrella with the first index in the sub-images being
1066 * 1. The array of sub-images is the ordered list of sub-images of the umbrella
1067 * that would be searched for a symbol that has the umbrella recorded as its
1068 * primary library. The table of contents index is an index into the
1069 * library's table of contents. This is used as the starting point of the
1070 * binary search or a directed linear search.
1071 */
1072struct twolevel_hint {
1073 uint32_t
1074 isub_image:8, /* index into the sub images */
1075 itoc:24; /* index into the table of contents */
1076};
1077
1078/*
1079 * The prebind_cksum_command contains the value of the original check sum for
1080 * prebound files or zero. When a prebound file is first created or modified
1081 * for other than updating its prebinding information the value of the check sum
1082 * is set to zero. When the file has it prebinding re-done and if the value of
1083 * the check sum is zero the original check sum is calculated and stored in
1084 * cksum field of this load command in the output file. If when the prebinding
1085 * is re-done and the cksum field is non-zero it is left unchanged from the
1086 * input file.
1087 */
1088struct prebind_cksum_command {
1089 uint32_t cmd; /* LC_PREBIND_CKSUM */
1090 uint32_t cmdsize; /* sizeof(struct prebind_cksum_command) */
1091 uint32_t cksum; /* the check sum or zero */
1092};
1093
1094/*
1095 * The uuid load command contains a single 128-bit unique random number that
1096 * identifies an object produced by the static link editor.
1097 */
1098struct uuid_command {
1099 uint32_t cmd; /* LC_UUID */
1100 uint32_t cmdsize; /* sizeof(struct uuid_command) */
1101 uint8_t uuid[16]; /* the 128-bit uuid */
1102};
1103
1104/*
1105 * The rpath_command contains a path which at runtime should be added to
1106 * the current run path used to find @rpath prefixed dylibs.
1107 */
1108struct rpath_command {
1109 uint32_t cmd; /* LC_RPATH */
1110 uint32_t cmdsize; /* includes string */
1111 union lc_str path; /* path to add to run path */
1112};
1113
1114/*
1115 * The linkedit_data_command contains the offsets and sizes of a blob
1116 * of data in the __LINKEDIT segment.
1117 */
1118struct linkedit_data_command {
1119 uint32_t cmd; /* LC_CODE_SIGNATURE or LC_SEGMENT_SPLIT_INFO */
1120 uint32_t cmdsize; /* sizeof(struct linkedit_data_command) */
1121 uint32_t dataoff; /* file offset of data in __LINKEDIT segment */
1122 uint32_t datasize; /* file size of data in __LINKEDIT segment */
1123};
1124
593a1d5f
A
1125/*
1126 * The encryption_info_command contains the file offset and size of an
1127 * of an encrypted segment.
1128 */
1129struct encryption_info_command {
1130 uint32_t cmd; /* LC_ENCRYPTION_INFO */
1131 uint32_t cmdsize; /* sizeof(struct encryption_info_command) */
1132 uint32_t cryptoff; /* file offset of encrypted range */
1133 uint32_t cryptsize; /* file size of encrypted range */
1134 uint32_t cryptid; /* which enryption system,
1135 0 means not-encrypted yet */
1136};
1137
b0d623f7
A
1138/*
1139 * The dyld_info_command contains the file offsets and sizes of
1140 * the new compressed form of the information dyld needs to
1141 * load the image. This information is used by dyld on Mac OS X
1142 * 10.6 and later. All information pointed to by this command
1143 * is encoded using byte streams, so no endian swapping is needed
1144 * to interpret it.
1145 */
1146struct dyld_info_command {
1147 uint32_t cmd; /* LC_DYLD_INFO or LC_DYLD_INFO_ONLY */
1148 uint32_t cmdsize; /* sizeof(struct dyld_info_command) */
1149
1150 /*
1151 * Dyld rebases an image whenever dyld loads it at an address different
1152 * from its preferred address. The rebase information is a stream
1153 * of byte sized opcodes whose symbolic names start with REBASE_OPCODE_.
1154 * Conceptually the rebase information is a table of tuples:
1155 * <seg-index, seg-offset, type>
1156 * The opcodes are a compressed way to encode the table by only
1157 * encoding when a column changes. In addition simple patterns
1158 * like "every n'th offset for m times" can be encoded in a few
1159 * bytes.
1160 */
1161 uint32_t rebase_off; /* file offset to rebase info */
1162 uint32_t rebase_size; /* size of rebase info */
1163
1164 /*
1165 * Dyld binds an image during the loading process, if the image
1166 * requires any pointers to be initialized to symbols in other images.
1167 * The rebase information is a stream of byte sized
1168 * opcodes whose symbolic names start with BIND_OPCODE_.
1169 * Conceptually the bind information is a table of tuples:
1170 * <seg-index, seg-offset, type, symbol-library-ordinal, symbol-name, addend>
1171 * The opcodes are a compressed way to encode the table by only
1172 * encoding when a column changes. In addition simple patterns
1173 * like for runs of pointers initialzed to the same value can be
1174 * encoded in a few bytes.
1175 */
1176 uint32_t bind_off; /* file offset to binding info */
1177 uint32_t bind_size; /* size of binding info */
1178
1179 /*
1180 * Some C++ programs require dyld to unique symbols so that all
1181 * images in the process use the same copy of some code/data.
1182 * This step is done after binding. The content of the weak_bind
1183 * info is an opcode stream like the bind_info. But it is sorted
1184 * alphabetically by symbol name. This enable dyld to walk
1185 * all images with weak binding information in order and look
1186 * for collisions. If there are no collisions, dyld does
1187 * no updating. That means that some fixups are also encoded
1188 * in the bind_info. For instance, all calls to "operator new"
1189 * are first bound to libstdc++.dylib using the information
1190 * in bind_info. Then if some image overrides operator new
1191 * that is detected when the weak_bind information is processed
1192 * and the call to operator new is then rebound.
1193 */
1194 uint32_t weak_bind_off; /* file offset to weak binding info */
1195 uint32_t weak_bind_size; /* size of weak binding info */
1196
1197 /*
1198 * Some uses of external symbols do not need to be bound immediately.
1199 * Instead they can be lazily bound on first use. The lazy_bind
1200 * are contains a stream of BIND opcodes to bind all lazy symbols.
1201 * Normal use is that dyld ignores the lazy_bind section when
1202 * loading an image. Instead the static linker arranged for the
1203 * lazy pointer to initially point to a helper function which
1204 * pushes the offset into the lazy_bind area for the symbol
1205 * needing to be bound, then jumps to dyld which simply adds
1206 * the offset to lazy_bind_off to get the information on what
1207 * to bind.
1208 */
1209 uint32_t lazy_bind_off; /* file offset to lazy binding info */
1210 uint32_t lazy_bind_size; /* size of lazy binding infs */
1211
1212 /*
1213 * The symbols exported by a dylib are encoded in a trie. This
1214 * is a compact representation that factors out common prefixes.
1215 * It also reduces LINKEDIT pages in RAM because it encodes all
1216 * information (name, address, flags) in one small, contiguous range.
1217 * The export area is a stream of nodes. The first node sequentially
1218 * is the start node for the trie.
1219 *
1220 * Nodes for a symbol start with a byte that is the length of
1221 * the exported symbol information for the string so far.
1222 * If there is no exported symbol, the byte is zero. If there
1223 * is exported info, it follows the length byte. The exported
1224 * info normally consists of a flags and offset both encoded
1225 * in uleb128. The offset is location of the content named
1226 * by the symbol. It is the offset from the mach_header for
1227 * the image.
1228 *
1229 * After the initial byte and optional exported symbol information
1230 * is a byte of how many edges (0-255) that this node has leaving
1231 * it, followed by each edge.
1232 * Each edge is a zero terminated cstring of the addition chars
1233 * in the symbol, followed by a uleb128 offset for the node that
1234 * edge points to.
1235 *
1236 */
1237 uint32_t export_off; /* file offset to lazy binding info */
1238 uint32_t export_size; /* size of lazy binding infs */
1239};
1240
1241/*
1242 * The following are used to encode rebasing information
1243 */
1244#define REBASE_TYPE_POINTER 1
1245#define REBASE_TYPE_TEXT_ABSOLUTE32 2
1246#define REBASE_TYPE_TEXT_PCREL32 3
1247
1248#define REBASE_OPCODE_MASK 0xF0
1249#define REBASE_IMMEDIATE_MASK 0x0F
1250#define REBASE_OPCODE_DONE 0x00
1251#define REBASE_OPCODE_SET_TYPE_IMM 0x10
1252#define REBASE_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB 0x20
1253#define REBASE_OPCODE_ADD_ADDR_ULEB 0x30
1254#define REBASE_OPCODE_ADD_ADDR_IMM_SCALED 0x40
1255#define REBASE_OPCODE_DO_REBASE_IMM_TIMES 0x50
1256#define REBASE_OPCODE_DO_REBASE_ULEB_TIMES 0x60
1257#define REBASE_OPCODE_DO_REBASE_ADD_ADDR_ULEB 0x70
1258#define REBASE_OPCODE_DO_REBASE_ULEB_TIMES_SKIPPING_ULEB 0x80
1259
1260
1261/*
1262 * The following are used to encode binding information
1263 */
1264#define BIND_TYPE_POINTER 1
1265#define BIND_TYPE_TEXT_ABSOLUTE32 2
1266#define BIND_TYPE_TEXT_PCREL32 3
1267
1268#define BIND_SPECIAL_DYLIB_SELF 0
1269#define BIND_SPECIAL_DYLIB_MAIN_EXECUTABLE -1
1270#define BIND_SPECIAL_DYLIB_FLAT_LOOKUP -2
1271
1272#define BIND_SYMBOL_FLAGS_WEAK_IMPORT 0x1
1273#define BIND_SYMBOL_FLAGS_NON_WEAK_DEFINITION 0x8
1274
1275#define BIND_OPCODE_MASK 0xF0
1276#define BIND_IMMEDIATE_MASK 0x0F
1277#define BIND_OPCODE_DONE 0x00
1278#define BIND_OPCODE_SET_DYLIB_ORDINAL_IMM 0x10
1279#define BIND_OPCODE_SET_DYLIB_ORDINAL_ULEB 0x20
1280#define BIND_OPCODE_SET_DYLIB_SPECIAL_IMM 0x30
1281#define BIND_OPCODE_SET_SYMBOL_TRAILING_FLAGS_IMM 0x40
1282#define BIND_OPCODE_SET_TYPE_IMM 0x50
1283#define BIND_OPCODE_SET_ADDEND_SLEB 0x60
1284#define BIND_OPCODE_SET_SEGMENT_AND_OFFSET_ULEB 0x70
1285#define BIND_OPCODE_ADD_ADDR_ULEB 0x80
1286#define BIND_OPCODE_DO_BIND 0x90
1287#define BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB 0xA0
1288#define BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED 0xB0
1289#define BIND_OPCODE_DO_BIND_ULEB_TIMES_SKIPPING_ULEB 0xC0
1290
1291
1292/*
1293 * The following are used on the flags byte of a terminal node
1294 * in the export information.
1295 */
1296#define EXPORT_SYMBOL_FLAGS_KIND_MASK 0x03
1297#define EXPORT_SYMBOL_FLAGS_KIND_REGULAR 0x00
1298#define EXPORT_SYMBOL_FLAGS_KIND_THREAD_LOCAL 0x01
1299#define EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION 0x04
1300#define EXPORT_SYMBOL_FLAGS_INDIRECT_DEFINITION 0x08
1301#define EXPORT_SYMBOL_FLAGS_HAS_SPECIALIZATIONS 0x10
1302
1c79356b
A
1303/*
1304 * The symseg_command contains the offset and size of the GNU style
1305 * symbol table information as described in the header file <symseg.h>.
1306 * The symbol roots of the symbol segments must also be aligned properly
1307 * in the file. So the requirement of keeping the offsets aligned to a
2d21ac55 1308 * multiple of a 4 bytes translates to the length field of the symbol
1c79356b
A
1309 * roots also being a multiple of a long. Also the padding must again be
1310 * zeroed. (THIS IS OBSOLETE and no longer supported).
1311 */
1312struct symseg_command {
2d21ac55
A
1313 uint32_t cmd; /* LC_SYMSEG */
1314 uint32_t cmdsize; /* sizeof(struct symseg_command) */
1315 uint32_t offset; /* symbol segment offset */
1316 uint32_t size; /* symbol segment size in bytes */
1c79356b
A
1317};
1318
1319/*
1320 * The ident_command contains a free format string table following the
1321 * ident_command structure. The strings are null terminated and the size of
2d21ac55 1322 * the command is padded out with zero bytes to a multiple of 4 bytes/
1c79356b
A
1323 * (THIS IS OBSOLETE and no longer supported).
1324 */
1325struct ident_command {
2d21ac55
A
1326 uint32_t cmd; /* LC_IDENT */
1327 uint32_t cmdsize; /* strings that follow this command */
1c79356b
A
1328};
1329
1330/*
1331 * The fvmfile_command contains a reference to a file to be loaded at the
2d21ac55 1332 * specified virtual address. (Presently, this command is reserved for
1c79356b
A
1333 * internal use. The kernel ignores this command when loading a program into
1334 * memory).
1335 */
1336struct fvmfile_command {
2d21ac55
A
1337 uint32_t cmd; /* LC_FVMFILE */
1338 uint32_t cmdsize; /* includes pathname string */
1c79356b 1339 union lc_str name; /* files pathname */
2d21ac55 1340 uint32_t header_addr; /* files virtual address */
1c79356b
A
1341};
1342
55e303ae 1343#endif /* _MACHO_LOADER_H_ */