1.\" Copyright (c) 1991 The Regents of the University of California. 2.\" 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. All advertising materials mentioning features or use of this software 16.\" must display the following acknowledgement: 17.\" This product includes software developed by the University of 18.\" California, Berkeley and its contributors. 19.\" 4. Neither the name of the University nor the names of its contributors 20.\" may be used to endorse or promote products derived from this software 21.\" without specific prior written permission. 22.\" 23.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33.\" SUCH DAMAGE. 34.\" 35.\" @(#)a.out.5 6.3 (Berkeley) 4/29/91 36.\" 37.Dd April 29, 1991 38.Dt A.OUT 5 39.Os 40.Sh NAME 41.Nm a.out 42.Nd format of executable binary files 43.Sh SYNOPSIS 44.Fd #include <a.out.h> 45.Sh DESCRIPTION 46The include file 47.Aq Pa a.out.h 48declares three structures and several macros. 49The structures describe the format of 50executable machine code files 51.Pq Sq binaries 52on the system. 53.Pp 54A binary file consists of up to 7 sections. 55In order, these sections are: 56.Bl -tag -width "text relocations" 57.It exec header 58Contains parameters used by the kernel 59to load a binary file into memory and execute it, 60and by the link editor 61.Xr ld 1 62to combine a binary file with other binary files. 63This section is the only mandatory one. 64.It text segment 65Contains machine code and related data 66that are loaded into memory when a program executes. 67May be loaded read-only. 68.It data segment 69Contains initialized data; always loaded into writable memory. 70.It text relocations 71Contains records used by the link editor 72to update pointers in the text segment when combining binary files. 73.It data relocations 74Like the text relocation section, but for data segment pointers. 75.It symbol table 76Contains records used by the link editor 77to cross reference the addresses of named variables and functions 78.Pq Sq symbols 79between binary files. 80.It string table 81Contains the character strings corresponding to the symbol names. 82.El 83.Pp 84Every binary file begins with an 85.Fa exec 86structure: 87.Bd -literal -offset indent 88struct exec { 89 unsigned short a_mid; 90 unsigned short a_magic; 91 unsigned long a_text; 92 unsigned long a_data; 93 unsigned long a_bss; 94 unsigned long a_syms; 95 unsigned long a_entry; 96 unsigned long a_trsize; 97 unsigned long a_drsize; 98}; 99.Ed 100.Pp 101The fields have the following functions: 102.Bl -tag -width a_trsize 103.It Fa a_mid 104Contains a bit pattern that 105identifies binaries that were built for 106certain sub-classes of an architecture 107.Pq Sq machine IDs 108or variants of the operating system on a given architecture. 109The kernel may not support all machine IDs 110on a given architecture. 111The 112.Fa a_mid 113field is not present on some architectures; 114in this case, the 115.Fa a_magic 116field has type 117.Em unsigned long . 118.It Fa a_magic 119Contains a bit pattern 120.Pq Sq magic number 121that uniquely identifies binary files 122and distinguishes different loading conventions. 123The field must contain one of the following values: 124.Bl -tag -width ZMAGIC 125.It Dv OMAGIC 126The text and data segments immediately follow the header 127and are contiguous. 128The kernel loads both text and data segments into writable memory. 129.It Dv NMAGIC 130As with 131.Dv OMAGIC , 132text and data segments immediately follow the header and are contiguous. 133However, the kernel loads the text into read-only memory 134and loads the data into writable memory at the next 135page boundary after the text. 136.It Dv ZMAGIC 137The kernel loads individual pages on demand from the binary. 138The header, text segment and data segment are all 139padded by the link editor to a multiple of the page size. 140Pages that the kernel loads from the text segment are read-only, 141while pages from the data segment are writable. 142.El 143.It Fa a_text 144Contains the size of the text segment in bytes. 145.It Fa a_data 146Contains the size of the data segment in bytes. 147.It Fa a_bss 148Contains the number of bytes in the 149.Sq bss segment 150and is used by the kernel to set the initial break 151.Pq Xr brk 2 152after the data segment. 153The kernel loads the program so that this amount of writable memory 154appears to follow the data segment and initially reads as zeroes. 155.It Fa a_syms 156Contains the size in bytes of the symbol table section. 157.It Fa a_entry 158Contains the address in memory of the entry point 159of the program after the kernel has loaded it; 160the kernel starts the execution of the program 161from the machine instruction at this address. 162.It Fa a_trsize 163Contains the size in bytes of the text relocation table. 164.It Fa a_drsize 165Contains the size in bytes of the data relocation table. 166.El 167.Pp 168The 169.Pa a.out.h 170include file defines several macros which use an 171.Fa exec 172structure to test consistency or to locate section offsets in the binary file. 173.Bl -tag -width N_BADMAG(exec) 174.It Fn N_BADMAG exec 175Nonzero if the 176.Fa a_magic 177field does not contain a recognized value. 178.It Fn N_TXTOFF exec 179The byte offset in the binary file of the beginning of the text segment. 180.It Fn N_SYMOFF exec 181The byte offset of the beginning of the symbol table. 182.It Fn N_STROFF exec 183The byte offset of the beginning of the string table. 184.El 185.Pp 186Relocation records have a standard format which 187is described by the 188.Fa relocation_info 189structure: 190.Bd -literal -offset indent 191struct relocation_info { 192 int r_address; 193 unsigned int r_symbolnum : 24, 194 r_pcrel : 1, 195 r_length : 2, 196 r_extern : 1, 197 : 4; 198}; 199.Ed 200.Pp 201The 202.Fa relocation_info 203fields are used as follows: 204.Bl -tag -width r_symbolnum 205.It Fa r_address 206Contains the byte offset of a pointer that needs to be link-edited. 207Text relocation offsets are reckoned from the start of the text segment, 208and data relocation offsets from the start of the data segment. 209The link editor adds the value that is already stored at this offset 210into the new value that it computes using this relocation record. 211.It Fa r_symbolnum 212Contains the ordinal number of a symbol structure 213in the symbol table (it is 214.Em not 215a byte offset). 216After the link editor resolves the absolute address for this symbol, 217it adds that address to the pointer that is undergoing relocation. 218(If the 219.Fa r_extern 220bit is clear, the situation is different; see below.) 221.It Fa r_pcrel 222If this is set, 223the link editor assumes that it is updating a pointer 224that is part of a machine code instruction using pc-relative addressing. 225The address of the relocated pointer is implicitly added 226to its value when the running program uses it. 227.It Fa r_length 228Contains the log base 2 of the length of the pointer in bytes; 2290 for 1-byte displacements, 1 for 2-byte displacements, 2302 for 4-byte displacements. 231.It Fa r_extern 232Set if this relocation requires an external reference; 233the link editor must use a symbol address to update the pointer. 234When the 235.Fa r_extern 236bit is clear, the relocation is 237.Sq local ; 238the link editor updates the pointer to reflect 239changes in the load addresses of the various segments, 240rather than changes in the value of a symbol. 241In this case, the content of the 242.Fa r_symbolnum 243field is an 244.Fa n_type 245value (see below); 246this type field tells the link editor 247what segment the relocated pointer points into. 248.El 249.Pp 250Symbols map names to addresses (or more generally, strings to values). 251Since the link-editor adjusts addresses, 252a symbol's name must be used to stand for its address 253until an absolute value has been assigned. 254Symbols consist of a fixed-length record in the symbol table 255and a variable-length name in the string table. 256The symbol table is an array of 257.Fa nlist 258structures: 259.Bd -literal -offset indent 260struct nlist { 261 union { 262 char *n_name; 263 long n_strx; 264 } n_un; 265 unsigned char n_type; 266 char n_other; 267 short n_desc; 268 unsigned long n_value; 269}; 270.Ed 271.Pp 272The fields are used as follows: 273.Bl -tag -width n_un.n_strx 274.It Fa n_un.n_strx 275Contains a byte offset into the string table 276for the name of this symbol. 277When a program accesses a symbol table with the 278.Xr nlist 3 279function, 280this field is replaced with the 281.Fa n_un.n_name 282field, which is a pointer to the string in memory. 283.It Fa n_type 284Used by the link editor to determine 285how to update the symbol's value. 286The 287.Fa n_type 288field is broken down into three sub-fields using bitmasks. 289The link editor treats symbols with the 290.Dv N_EXT 291type bit set as 292.Sq external 293symbols and permits references to them from other binary files. 294The 295.Dv N_TYPE 296mask selects bits of interest to the link editor: 297.Bl -tag -width N_TEXT 298.It Dv N_UNDF 299An undefined symbol. 300The link editor must locate an external symbol with the same name 301in another binary file to determine the absolute value of this symbol. 302As a special case, if the 303.Fa n_value 304field is nonzero and no binary file in the link-edit defines this symbol, 305the link-editor will resolve this symbol to an address 306in the bss segment, 307reserving an amount of bytes equal to 308.Fa n_value . 309If this symbol is undefined in more than one binary file 310and the binary files do not agree on the size, 311the link editor chooses the greatest size found across all binaries. 312.It Dv N_ABS 313An absolute symbol. 314The link editor does not update an absolute symbol. 315.It Dv N_TEXT 316A text symbol. 317This symbol's value is a text address and 318the link editor will update it when it merges binary files. 319.It Dv N_DATA 320A data symbol; similar to 321.Dv N_TEXT 322but for data addresses. 323The values for text and data symbols are not file offsets but 324addresses; to recover the file offsets, it is necessary 325to identify the loaded address of the beginning of the corresponding 326section and subtract it, then add the offset of the section. 327.It Dv N_BSS 328A bss symbol; like text or data symbols but 329has no corresponding offset in the binary file. 330.It Dv N_FN 331A filename symbol. 332The link editor inserts this symbol before 333the other symbols from a binary file when 334merging binary files. 335The name of the symbol is the filename given to the link editor, 336and its value is the first text address from that binary file. 337Filename symbols are not needed for link-editing or loading, 338but are useful for debuggers. 339.El 340.Pp 341The 342.Dv N_STAB 343mask selects bits of interest to symbolic debuggers 344such as 345.Xr gdb 1 ; 346the values are described in 347.Xr stab 5 . 348.It Fa n_other 349This field is currently unused. 350.It Fa n_desc 351Reserved for use by debuggers; passed untouched by the link editor. 352Different debuggers use this field for different purposes. 353.It Fa n_value 354Contains the value of the symbol. 355For text, data and bss symbols, this is an address; 356for other symbols (such as debugger symbols), 357the value may be arbitrary. 358.El 359.Pp 360The string table consists of an 361.Em unsigned long 362length followed by null-terminated symbol strings. 363The length represents the size of the entire table in bytes, 364so its minimum value (or the offset of the first string) 365is always 4 on 32-bit machines. 366.Sh SEE ALSO 367.Xr ld 1 , 368.Xr execve 2 , 369.Xr nlist 3 , 370.Xr core 5 , 371.Xr dbx 5 , 372.Xr stab 5 373.Sh HISTORY 374The 375.Pa a.out.h 376include file appeared in 377.At v7 . 378.Sh BUGS 379Since not all of the supported architectures use the 380.Fa a_mid 381field, 382it can be difficult to determine what 383architecture a binary will execute on 384without examining its actual machine code. 385Even with a machine identifier, 386the byte order of the 387.Fa exec 388header is machine-dependent. 389.Pp 390Nobody seems to agree on what 391.Em bss 392stands for. 393.Pp 394New binary file formats may be supported in the future, 395and they probably will not be compatible at any level 396with this ancient format. 397