1.\" Copyright (c) 1991, 1993 2.\" The Regents of the University of California. All rights reserved. 3.\" 4.\" This man page is derived from documentation contributed to Berkeley by 5.\" Donn Seeley at UUNET Technologies, Inc. 6.\" 7.\" Redistribution and use in source and binary forms, with or without 8.\" modification, are permitted provided that the following conditions 9.\" are met: 10.\" 1. Redistributions of source code must retain the above copyright 11.\" notice, this list of conditions and the following disclaimer. 12.\" 2. Redistributions in binary form must reproduce the above copyright 13.\" notice, this list of conditions and the following disclaimer in the 14.\" documentation and/or other materials provided with the distribution. 15.\" 3. Neither the name of the University nor the names of its contributors 16.\" may be used to endorse or promote products derived from this software 17.\" without specific prior written permission. 18.\" 19.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 20.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 23.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29.\" SUCH DAMAGE. 30.\" 31.\" @(#)a.out.5 8.1 (Berkeley) 6/5/93 32.\" $FreeBSD: src/share/man/man5/a.out.5,v 1.10.2.4 2002/04/16 14:50:18 trhodes Exp $ 33.\" 34.Dd December 16, 2021 35.Dt A.OUT 5 36.Os 37.Sh NAME 38.Nm a.out 39.Nd format of executable binary files 40.Sh SYNOPSIS 41.In a.out.h 42.Sh DESCRIPTION 43.Bf -emphasis 44This manual page describes an obsolete executable format. 45See 46.Xr elf 5 47for more information on 48.Dx Ap s 49current executable format. 50.Ef 51.Pp 52The include file 53.In a.out.h 54declares three structures and several macros. 55The structures describe the format of 56executable machine code files 57.Pq Sq binaries 58on the system. 59.Pp 60A binary file consists of up to 7 sections. 61In order, these sections are: 62.Bl -tag -width "text relocations" 63.It exec header 64Contains parameters used by the kernel 65to load a binary file into memory and execute it, 66and by the link editor 67.Xr ld 1 68to combine a binary file with other binary files. 69This section is the only mandatory one. 70.It text segment 71Contains machine code and related data 72that are loaded into memory when a program executes. 73May be loaded read-only. 74.It data segment 75Contains initialized data; always loaded into writable memory. 76.It text relocations 77Contains records used by the link editor 78to update pointers in the text segment when combining binary files. 79.It data relocations 80Like the text relocation section, but for data segment pointers. 81.It symbol table 82Contains records used by the link editor 83to cross reference the addresses of named variables and functions 84.Pq Sq symbols 85between binary files. 86.It string table 87Contains the character strings corresponding to the symbol names. 88.El 89.Pp 90Every binary file begins with an 91.Fa exec 92structure: 93.Bd -literal -offset indent 94struct exec { 95 unsigned long a_midmag; 96 unsigned long a_text; 97 unsigned long a_data; 98 unsigned long a_bss; 99 unsigned long a_syms; 100 unsigned long a_entry; 101 unsigned long a_trsize; 102 unsigned long a_drsize; 103}; 104.Ed 105.Pp 106The fields have the following functions: 107.Bl -tag -width a_trsize 108.It Fa a_midmag 109This field is stored in host byte-order. 110It has a number of sub-components accessed by the macros 111.Fn N_GETFLAG , 112.Fn N_GETMID , 113and 114.Fn N_GETMAGIC , 115and set by the macro 116.Fn N_SETMAGIC . 117.Pp 118The macro 119.Fn N_GETFLAG 120returns a few flags: 121.Bl -tag -width EX_DYNAMIC 122.It Dv EX_DYNAMIC 123indicates that the executable requires the services of the run-time link editor. 124.It Dv EX_PIC 125indicates that the object contains position independent code. 126This flag is 127set by 128.Xr as 1 129when given the 130.Sq -k 131flag and is preserved by 132.Xr ld 1 133if necessary. 134.El 135.Pp 136If both EX_DYNAMIC and EX_PIC are set, the object file is a position independent 137executable image (eg. a shared library), which is to be loaded into the 138process address space by the run-time link editor. 139.Pp 140The macro 141.Fn N_GETMID 142returns the machine-id. 143This indicates which machine(s) the binary is intended to run on. 144.Pp 145.Fn N_GETMAGIC 146specifies the magic number, which uniquely identifies binary files 147and distinguishes different loading conventions. 148The field must contain one of the following values: 149.Bl -tag -width ZMAGIC 150.It Dv OMAGIC 151The text and data segments immediately follow the header 152and are contiguous. 153The kernel loads both text and data segments into writable memory. 154.It Dv NMAGIC 155As with 156.Dv OMAGIC , 157text and data segments immediately follow the header and are contiguous. 158However, the kernel loads the text into read-only memory 159and loads the data into writable memory at the next 160page boundary after the text. 161.It Dv ZMAGIC 162The kernel loads individual pages on demand from the binary. 163The header, text segment and data segment are all 164padded by the link editor to a multiple of the page size. 165Pages that the kernel loads from the text segment are read-only, 166while pages from the data segment are writable. 167.El 168.It Fa a_text 169Contains the size of the text segment in bytes. 170.It Fa a_data 171Contains the size of the data segment in bytes. 172.It Fa a_bss 173Contains the number of bytes in the 174.Sq bss segment . 175The kernel loads the program so that this amount of writable memory 176appears to follow the data segment and initially reads as zeroes. 177.Po 178.Em bss 179= block started by symbol 180.Pc 181.It Fa a_syms 182Contains the size in bytes of the symbol table section. 183.It Fa a_entry 184Contains the address in memory of the entry point 185of the program after the kernel has loaded it; 186the kernel starts the execution of the program 187from the machine instruction at this address. 188.It Fa a_trsize 189Contains the size in bytes of the text relocation table. 190.It Fa a_drsize 191Contains the size in bytes of the data relocation table. 192.El 193.Pp 194The 195.In a.out.h 196include file defines several macros which use an 197.Fa exec 198structure to test consistency or to locate section offsets in the binary file. 199.Bl -tag -width N_BADMAG(exec) 200.It Fn N_BADMAG exec 201Nonzero if the 202.Fa a_magic 203field does not contain a recognized value. 204.It Fn N_TXTOFF exec 205The byte offset in the binary file of the beginning of the text segment. 206.It Fn N_SYMOFF exec 207The byte offset of the beginning of the symbol table. 208.It Fn N_STROFF exec 209The byte offset of the beginning of the string table. 210.El 211.Pp 212Relocation records have a standard format which 213is described by the 214.Fa relocation_info 215structure: 216.Bd -literal -offset indent 217struct relocation_info { 218 int r_address; 219 unsigned int r_symbolnum : 24, 220 r_pcrel : 1, 221 r_length : 2, 222 r_extern : 1, 223 r_baserel : 1, 224 r_jmptable : 1, 225 r_relative : 1, 226 r_copy : 1; 227}; 228.Ed 229.Pp 230The 231.Fa relocation_info 232fields are used as follows: 233.Bl -tag -width r_symbolnum 234.It Fa r_address 235Contains the byte offset of a pointer that needs to be link-edited. 236Text relocation offsets are reckoned from the start of the text segment, 237and data relocation offsets from the start of the data segment. 238The link editor adds the value that is already stored at this offset 239into the new value that it computes using this relocation record. 240.It Fa r_symbolnum 241Contains the ordinal number of a symbol structure 242in the symbol table (it is 243.Em not 244a byte offset). 245After the link editor resolves the absolute address for this symbol, 246it adds that address to the pointer that is undergoing relocation. 247(If the 248.Fa r_extern 249bit is clear, the situation is different; see below.) 250.It Fa r_pcrel 251If this is set, 252the link editor assumes that it is updating a pointer 253that is part of a machine code instruction using pc-relative addressing. 254The address of the relocated pointer is implicitly added 255to its value when the running program uses it. 256.It Fa r_length 257Contains the log base 2 of the length of the pointer in bytes; 2580 for 1-byte displacements, 1 for 2-byte displacements, 2592 for 4-byte displacements. 260.It Fa r_extern 261Set if this relocation requires an external reference; 262the link editor must use a symbol address to update the pointer. 263When the 264.Fa r_extern 265bit is clear, the relocation is 266.Sq local ; 267the link editor updates the pointer to reflect 268changes in the load addresses of the various segments, 269rather than changes in the value of a symbol (except when 270.Fa r_baserel 271is also set (see below). 272In this case, the content of the 273.Fa r_symbolnum 274field is an 275.Fa n_type 276value (see below); 277this type field tells the link editor 278what segment the relocated pointer points into. 279.It Fa r_baserel 280If set, the symbol, as identified by the 281.Fa r_symbolnum 282field, is to be relocated to an offset into the Global Offset Table. 283At run-time, the entry in the Global Offset Table at this offset is set to 284be the address of the symbol. 285.It Fa r_jmptable 286If set, the symbol, as identified by the 287.Fa r_symbolnum 288field, is to be relocated to an offset into the Procedure Linkage Table. 289.It Fa r_relative 290If set, this relocation is relative to the (run-time) load address of the 291image this object file is going to be a part of. 292This type of relocation 293only occurs in shared objects. 294.It Fa r_copy 295If set, this relocation record identifies a symbol whose contents should 296be copied to the location given in 297.Fa r_address . 298The copying is done by the run-time link-editor from a suitable data 299item in a shared object. 300.El 301.Pp 302Symbols map names to addresses (or more generally, strings to values). 303Since the link-editor adjusts addresses, 304a symbol's name must be used to stand for its address 305until an absolute value has been assigned. 306Symbols consist of a fixed-length record in the symbol table 307and a variable-length name in the string table. 308The symbol table is an array of 309.Fa nlist 310structures: 311.Bd -literal -offset indent 312struct nlist { 313 union { 314 char *n_name; 315 long n_strx; 316 } n_un; 317 unsigned char n_type; 318 char n_other; 319 short n_desc; 320 unsigned long n_value; 321}; 322.Ed 323.Pp 324The fields are used as follows: 325.Bl -tag -width n_un.n_strx 326.It Fa n_un.n_strx 327Contains a byte offset into the string table 328for the name of this symbol. 329When a program accesses a symbol table with the 330.Xr nlist 3 331function, 332this field is replaced with the 333.Fa n_un.n_name 334field, which is a pointer to the string in memory. 335.It Fa n_type 336Used by the link editor to determine 337how to update the symbol's value. 338The 339.Fa n_type 340field is broken down into three sub-fields using bitmasks. 341The link editor treats symbols with the 342.Dv N_EXT 343type bit set as 344.Sq external 345symbols and permits references to them from other binary files. 346The 347.Dv N_TYPE 348mask selects bits of interest to the link editor: 349.Bl -tag -width N_TEXT 350.It Dv N_UNDF 351An undefined symbol. 352The link editor must locate an external symbol with the same name 353in another binary file to determine the absolute value of this symbol. 354As a special case, if the 355.Fa n_value 356field is nonzero and no binary file in the link-edit defines this symbol, 357the link-editor will resolve this symbol to an address 358in the bss segment, 359reserving an amount of bytes equal to 360.Fa n_value . 361If this symbol is undefined in more than one binary file 362and the binary files do not agree on the size, 363the link editor chooses the greatest size found across all binaries. 364.It Dv N_ABS 365An absolute symbol. 366The link editor does not update an absolute symbol. 367.It Dv N_TEXT 368A text symbol. 369This symbol's value is a text address and 370the link editor will update it when it merges binary files. 371.It Dv N_DATA 372A data symbol; similar to 373.Dv N_TEXT 374but for data addresses. 375The values for text and data symbols are not file offsets but 376addresses; to recover the file offsets, it is necessary 377to identify the loaded address of the beginning of the corresponding 378section and subtract it, then add the offset of the section. 379.It Dv N_BSS 380A bss symbol; like text or data symbols but 381has no corresponding offset in the binary file. 382.It Dv N_FN 383A filename symbol. 384The link editor inserts this symbol before 385the other symbols from a binary file when 386merging binary files. 387The name of the symbol is the filename given to the link editor, 388and its value is the first text address from that binary file. 389Filename symbols are not needed for link-editing or loading, 390but are useful for debuggers. 391.El 392.Pp 393The 394.Dv N_STAB 395mask selects bits of interest to symbolic debuggers 396such as 397.Xr gdb 1 ; 398the values are described in 399.Xr stab 5 . 400.It Fa n_other 401This field provides information on the nature of the symbol independent of 402the symbol's location in terms of segments as determined by the 403.Fa n_type 404field. 405Currently, the lower 4 bits of the 406.Fa n_other 407field hold one of two values: 408.Dv AUX_FUNC 409and 410.Dv AUX_OBJECT 411(see 412.In link.h 413for their definitions). 414.Dv AUX_FUNC 415associates the symbol with a callable function, while 416.Dv AUX_OBJECT 417associates the symbol with data, irrespective of their locations in 418either the text or the data segment. 419This field is intended to be used by 420.Xr ld 1 421for the construction of dynamic executables. 422.It Fa n_desc 423Reserved for use by debuggers; passed untouched by the link editor. 424Different debuggers use this field for different purposes. 425.It Fa n_value 426Contains the value of the symbol. 427For text, data and bss symbols, this is an address; 428for other symbols (such as debugger symbols), 429the value may be arbitrary. 430.El 431.Pp 432The string table consists of an 433.Em unsigned long 434length followed by null-terminated symbol strings. 435The length represents the size of the entire table in bytes, 436so its minimum value (or the offset of the first string) 437is always 4 on 32-bit machines. 438.Sh SEE ALSO 439.Xr as 1 , 440.Xr gdb 1 , 441.Xr ld 1 , 442.Xr execve 2 , 443.Xr nlist 3 , 444.Xr core 5 , 445.Xr elf 5 , 446.Xr link 5 , 447.Xr stab 5 448.Sh HISTORY 449The 450.In a.out.h 451include file appeared in 452.At v7 . 453.Sh BUGS 454Since not all of the supported architectures use the 455.Fa a_midmag 456field, 457it can be difficult to determine what 458architecture a binary will execute on 459without examining its actual machine code. 460Even with a machine identifier, 461the byte order of the 462.Fa exec 463header is machine-dependent. 464