1.\" Copyright (c) 1991, 1993 2.\" The Regents of the University of California. All rights reserved. 3.\" 4.\" This manual 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.\" $FreeBSD$ 33.\" 34.Dd June 10, 2010 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 43The include file 44.In a.out.h 45declares three structures and several macros. 46The structures describe the format of 47executable machine code files 48.Pq Sq binaries 49on the system. 50.Pp 51A binary file consists of up to 7 sections. 52In order, these sections are: 53.Bl -tag -width "text relocations" 54.It exec header 55Contains parameters used by the kernel 56to load a binary file into memory and execute it, 57and by the link editor 58.Xr ld 1 59to combine a binary file with other binary files. 60This section is the only mandatory one. 61.It text segment 62Contains machine code and related data 63that are loaded into memory when a program executes. 64May be loaded read-only. 65.It data segment 66Contains initialized data; always loaded into writable memory. 67.It text relocations 68Contains records used by the link editor 69to update pointers in the text segment when combining binary files. 70.It data relocations 71Like the text relocation section, but for data segment pointers. 72.It symbol table 73Contains records used by the link editor 74to cross reference the addresses of named variables and functions 75.Pq Sq symbols 76between binary files. 77.It string table 78Contains the character strings corresponding to the symbol names. 79.El 80.Pp 81Every binary file begins with an 82.Fa exec 83structure: 84.Bd -literal -offset indent 85struct exec { 86 unsigned long a_midmag; 87 unsigned long a_text; 88 unsigned long a_data; 89 unsigned long a_bss; 90 unsigned long a_syms; 91 unsigned long a_entry; 92 unsigned long a_trsize; 93 unsigned long a_drsize; 94}; 95.Ed 96.Pp 97The fields have the following functions: 98.Bl -tag -width a_trsize 99.It Fa a_midmag 100This field is stored in host byte-order. 101It has a number of sub-components accessed by the macros 102.Fn N_GETFLAG , 103.Fn N_GETMID , 104and 105.Fn N_GETMAGIC , 106and set by the macro 107.Fn N_SETMAGIC . 108.Pp 109The macro 110.Fn N_GETFLAG 111returns a few flags: 112.Bl -tag -width EX_DYNAMIC 113.It Dv EX_DYNAMIC 114indicates that the executable requires the services of the run-time link editor. 115.It Dv EX_PIC 116indicates that the object contains position independent code. 117This flag is 118set by 119.Xr as 1 120when given the 121.Sq -k 122flag and is preserved by 123.Xr ld 1 124if necessary. 125.El 126.Pp 127If both EX_DYNAMIC and EX_PIC are set, the object file is a position independent 128executable image (e.g.\& a shared library), which is to be loaded into the 129process address space by the run-time link editor. 130.Pp 131The macro 132.Fn N_GETMID 133returns the machine-id. 134This indicates which machine(s) the binary is intended to run on. 135.Pp 136.Fn N_GETMAGIC 137specifies the magic number, which uniquely identifies binary files 138and distinguishes different loading conventions. 139The field must contain one of the following values: 140.Bl -tag -width ZMAGIC 141.It Dv OMAGIC 142The text and data segments immediately follow the header 143and are contiguous. 144The kernel loads both text and data segments into writable memory. 145.It Dv NMAGIC 146As with 147.Dv OMAGIC , 148text and data segments immediately follow the header and are contiguous. 149However, the kernel loads the text into read-only memory 150and loads the data into writable memory at the next 151page boundary after the text. 152.It Dv ZMAGIC 153The kernel loads individual pages on demand from the binary. 154The header, text segment and data segment are all 155padded by the link editor to a multiple of the page size. 156Pages that the kernel loads from the text segment are read-only, 157while pages from the data segment are writable. 158.El 159.It Fa a_text 160Contains the size of the text segment in bytes. 161.It Fa a_data 162Contains the size of the data segment in bytes. 163.It Fa a_bss 164Contains the number of bytes in the 165.Sq bss segment 166and is used by the kernel to set the initial break 167.Pq Xr brk 2 168after the data segment. 169The kernel loads the program so that this amount of writable memory 170appears to follow the data segment and initially reads as zeroes. 171.Em ( bss 172= block started by symbol) 173.It Fa a_syms 174Contains the size in bytes of the symbol table section. 175.It Fa a_entry 176Contains the address in memory of the entry point 177of the program after the kernel has loaded it; 178the kernel starts the execution of the program 179from the machine instruction at this address. 180.It Fa a_trsize 181Contains the size in bytes of the text relocation table. 182.It Fa a_drsize 183Contains the size in bytes of the data relocation table. 184.El 185.Pp 186The 187.In a.out.h 188include file defines several macros which use an 189.Fa exec 190structure to test consistency or to locate section offsets in the binary file. 191.Bl -tag -width N_BADMAG(exec) 192.It Fn N_BADMAG exec 193Nonzero if the 194.Fa a_magic 195field does not contain a recognized value. 196.It Fn N_TXTOFF exec 197The byte offset in the binary file of the beginning of the text segment. 198.It Fn N_SYMOFF exec 199The byte offset of the beginning of the symbol table. 200.It Fn N_STROFF exec 201The byte offset of the beginning of the string table. 202.El 203.Pp 204Relocation records have a standard format which 205is described by the 206.Fa relocation_info 207structure: 208.Bd -literal -offset indent 209struct relocation_info { 210 int r_address; 211 unsigned int r_symbolnum : 24, 212 r_pcrel : 1, 213 r_length : 2, 214 r_extern : 1, 215 r_baserel : 1, 216 r_jmptable : 1, 217 r_relative : 1, 218 r_copy : 1; 219}; 220.Ed 221.Pp 222The 223.Fa relocation_info 224fields are used as follows: 225.Bl -tag -width r_symbolnum 226.It Fa r_address 227Contains the byte offset of a pointer that needs to be link-edited. 228Text relocation offsets are reckoned from the start of the text segment, 229and data relocation offsets from the start of the data segment. 230The link editor adds the value that is already stored at this offset 231into the new value that it computes using this relocation record. 232.It Fa r_symbolnum 233Contains the ordinal number of a symbol structure 234in the symbol table (it is 235.Em not 236a byte offset). 237After the link editor resolves the absolute address for this symbol, 238it adds that address to the pointer that is undergoing relocation. 239(If the 240.Fa r_extern 241bit is clear, the situation is different; see below.) 242.It Fa r_pcrel 243If this is set, 244the link editor assumes that it is updating a pointer 245that is part of a machine code instruction using pc-relative addressing. 246The address of the relocated pointer is implicitly added 247to its value when the running program uses it. 248.It Fa r_length 249Contains the log base 2 of the length of the pointer in bytes; 2500 for 1-byte displacements, 1 for 2-byte displacements, 2512 for 4-byte displacements. 252.It Fa r_extern 253Set if this relocation requires an external reference; 254the link editor must use a symbol address to update the pointer. 255When the 256.Fa r_extern 257bit is clear, the relocation is 258.Sq local ; 259the link editor updates the pointer to reflect 260changes in the load addresses of the various segments, 261rather than changes in the value of a symbol (except when 262.Fa r_baserel 263is also set (see below). 264In this case, the content of the 265.Fa r_symbolnum 266field is an 267.Fa n_type 268value (see below); 269this type field tells the link editor 270what segment the relocated pointer points into. 271.It Fa r_baserel 272If set, the symbol, as identified by the 273.Fa r_symbolnum 274field, is to be relocated to an offset into the Global Offset Table. 275At run-time, the entry in the Global Offset Table at this offset is set to 276be the address of the symbol. 277.It Fa r_jmptable 278If set, the symbol, as identified by the 279.Fa r_symbolnum 280field, is to be relocated to an offset into the Procedure Linkage Table. 281.It Fa r_relative 282If set, this relocation is relative to the (run-time) load address of the 283image this object file is going to be a part of. 284This type of relocation 285only occurs in shared objects. 286.It Fa r_copy 287If set, this relocation record identifies a symbol whose contents should 288be copied to the location given in 289.Fa r_address . 290The copying is done by the run-time link-editor from a suitable data 291item in a shared object. 292.El 293.Pp 294Symbols map names to addresses (or more generally, strings to values). 295Since the link-editor adjusts addresses, 296a symbol's name must be used to stand for its address 297until an absolute value has been assigned. 298Symbols consist of a fixed-length record in the symbol table 299and a variable-length name in the string table. 300The symbol table is an array of 301.Fa nlist 302structures: 303.Bd -literal -offset indent 304struct nlist { 305 union { 306 const char *n_name; 307 long n_strx; 308 } n_un; 309 unsigned char n_type; 310 char n_other; 311 short n_desc; 312 unsigned long n_value; 313}; 314.Ed 315.Pp 316The fields are used as follows: 317.Bl -tag -width n_un.n_strx 318.It Fa n_un.n_strx 319Contains a byte offset into the string table 320for the name of this symbol. 321When a program accesses a symbol table with the 322.Xr nlist 3 323function, 324this field is replaced with the 325.Fa n_un.n_name 326field, which is a pointer to the string in memory. 327.It Fa n_type 328Used by the link editor to determine 329how to update the symbol's value. 330The 331.Fa n_type 332field is broken down into three sub-fields using bitmasks. 333The link editor treats symbols with the 334.Dv N_EXT 335type bit set as 336.Sq external 337symbols and permits references to them from other binary files. 338The 339.Dv N_TYPE 340mask selects bits of interest to the link editor: 341.Bl -tag -width N_TEXT 342.It Dv N_UNDF 343An undefined symbol. 344The link editor must locate an external symbol with the same name 345in another binary file to determine the absolute value of this symbol. 346As a special case, if the 347.Fa n_value 348field is nonzero and no binary file in the link-edit defines this symbol, 349the link-editor will resolve this symbol to an address 350in the bss segment, 351reserving an amount of bytes equal to 352.Fa n_value . 353If this symbol is undefined in more than one binary file 354and the binary files do not agree on the size, 355the link editor chooses the greatest size found across all binaries. 356.It Dv N_ABS 357An absolute symbol. 358The link editor does not update an absolute symbol. 359.It Dv N_TEXT 360A text symbol. 361This symbol's value is a text address and 362the link editor will update it when it merges binary files. 363.It Dv N_DATA 364A data symbol; similar to 365.Dv N_TEXT 366but for data addresses. 367The values for text and data symbols are not file offsets but 368addresses; to recover the file offsets, it is necessary 369to identify the loaded address of the beginning of the corresponding 370section and subtract it, then add the offset of the section. 371.It Dv N_BSS 372A bss symbol; like text or data symbols but 373has no corresponding offset in the binary file. 374.It Dv N_FN 375A filename symbol. 376The link editor inserts this symbol before 377the other symbols from a binary file when 378merging binary files. 379The name of the symbol is the filename given to the link editor, 380and its value is the first text address from that binary file. 381Filename symbols are not needed for link-editing or loading, 382but are useful for debuggers. 383.El 384.Pp 385The 386.Dv N_STAB 387mask selects bits of interest to symbolic debuggers 388such as 389.Xr gdb 1 Pq Pa ports/devel/gdb ; 390the values are described in 391.Xr stab 5 . 392.It Fa n_other 393This field provides information on the nature of the symbol independent of 394the symbol's location in terms of segments as determined by the 395.Fa n_type 396field. 397Currently, the lower 4 bits of the 398.Fa n_other 399field hold one of two values: 400.Dv AUX_FUNC 401and 402.Dv AUX_OBJECT 403(see 404.In link.h 405for their definitions). 406.Dv AUX_FUNC 407associates the symbol with a callable function, while 408.Dv AUX_OBJECT 409associates the symbol with data, irrespective of their locations in 410either the text or the data segment. 411This field is intended to be used by 412.Xr ld 1 413for the construction of dynamic executables. 414.It Fa n_desc 415Reserved for use by debuggers; passed untouched by the link editor. 416Different debuggers use this field for different purposes. 417.It Fa n_value 418Contains the value of the symbol. 419For text, data and bss symbols, this is an address; 420for other symbols (such as debugger symbols), 421the value may be arbitrary. 422.El 423.Pp 424The string table consists of an 425.Em unsigned long 426length followed by null-terminated symbol strings. 427The length represents the size of the entire table in bytes, 428so its minimum value (or the offset of the first string) 429is always 4 on 32-bit machines. 430.Sh SEE ALSO 431.Xr as 1 , 432.Xr gdb 1 Pq Pa ports/devel/gdb , 433.Xr ld 1 , 434.Xr brk 2 , 435.Xr execve 2 , 436.Xr nlist 3 , 437.Xr core 5 , 438.Xr elf 5 , 439.Xr link 5 , 440.Xr stab 5 441.Sh HISTORY 442The 443.In a.out.h 444include file appeared in 445.At v7 . 446.Sh BUGS 447Since not all of the supported architectures use the 448.Fa a_midmag 449field, 450it can be difficult to determine what 451architecture a binary will execute on 452without examining its actual machine code. 453Even with a machine identifier, 454the byte order of the 455.Fa exec 456header is machine-dependent. 457