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.Dd June 10, 2010 32.Dt A.OUT 5 33.Os 34.Sh NAME 35.Nm a.out 36.Nd format of executable binary files 37.Sh SYNOPSIS 38.In a.out.h 39.Sh DESCRIPTION 40The include file 41.In a.out.h 42declares three structures and several macros. 43The structures describe the format of 44executable machine code files 45.Pq Sq binaries 46on the system. 47.Pp 48A binary file consists of up to 7 sections. 49In order, these sections are: 50.Bl -tag -width "text relocations" 51.It exec header 52Contains parameters used by the kernel 53to load a binary file into memory and execute it, 54and by the link editor 55.Xr ld 1 56to combine a binary file with other binary files. 57This section is the only mandatory one. 58.It text segment 59Contains machine code and related data 60that are loaded into memory when a program executes. 61May be loaded read-only. 62.It data segment 63Contains initialized data; always loaded into writable memory. 64.It text relocations 65Contains records used by the link editor 66to update pointers in the text segment when combining binary files. 67.It data relocations 68Like the text relocation section, but for data segment pointers. 69.It symbol table 70Contains records used by the link editor 71to cross reference the addresses of named variables and functions 72.Pq Sq symbols 73between binary files. 74.It string table 75Contains the character strings corresponding to the symbol names. 76.El 77.Pp 78Every binary file begins with an 79.Fa exec 80structure: 81.Bd -literal -offset indent 82struct exec { 83 unsigned long a_midmag; 84 unsigned long a_text; 85 unsigned long a_data; 86 unsigned long a_bss; 87 unsigned long a_syms; 88 unsigned long a_entry; 89 unsigned long a_trsize; 90 unsigned long a_drsize; 91}; 92.Ed 93.Pp 94The fields have the following functions: 95.Bl -tag -width a_trsize 96.It Fa a_midmag 97This field is stored in host byte-order. 98It has a number of sub-components accessed by the macros 99.Fn N_GETFLAG , 100.Fn N_GETMID , 101and 102.Fn N_GETMAGIC , 103and set by the macro 104.Fn N_SETMAGIC . 105.Pp 106The macro 107.Fn N_GETFLAG 108returns a few flags: 109.Bl -tag -width EX_DYNAMIC 110.It Dv EX_DYNAMIC 111indicates that the executable requires the services of the run-time link editor. 112.It Dv EX_PIC 113indicates that the object contains position independent code. 114This flag is 115set by 116.Xr as 1 117when given the 118.Sq -k 119flag and is preserved by 120.Xr ld 1 121if necessary. 122.El 123.Pp 124If both EX_DYNAMIC and EX_PIC are set, the object file is a position independent 125executable image (e.g.\& a shared library), which is to be loaded into the 126process address space by the run-time link editor. 127.Pp 128The macro 129.Fn N_GETMID 130returns the machine-id. 131This indicates which machine(s) the binary is intended to run on. 132.Pp 133.Fn N_GETMAGIC 134specifies the magic number, which uniquely identifies binary files 135and distinguishes different loading conventions. 136The field must contain one of the following values: 137.Bl -tag -width ZMAGIC 138.It Dv OMAGIC 139The text and data segments immediately follow the header 140and are contiguous. 141The kernel loads both text and data segments into writable memory. 142.It Dv NMAGIC 143As with 144.Dv OMAGIC , 145text and data segments immediately follow the header and are contiguous. 146However, the kernel loads the text into read-only memory 147and loads the data into writable memory at the next 148page boundary after the text. 149.It Dv ZMAGIC 150The kernel loads individual pages on demand from the binary. 151The header, text segment and data segment are all 152padded by the link editor to a multiple of the page size. 153Pages that the kernel loads from the text segment are read-only, 154while pages from the data segment are writable. 155.El 156.It Fa a_text 157Contains the size of the text segment in bytes. 158.It Fa a_data 159Contains the size of the data segment in bytes. 160.It Fa a_bss 161Contains the number of bytes in the 162.Sq bss segment 163and is used by the kernel to set the initial break 164.Pq Xr brk 2 165after the data segment. 166The kernel loads the program so that this amount of writable memory 167appears to follow the data segment and initially reads as zeroes. 168.Em ( bss 169= block started by symbol) 170.It Fa a_syms 171Contains the size in bytes of the symbol table section. 172.It Fa a_entry 173Contains the address in memory of the entry point 174of the program after the kernel has loaded it; 175the kernel starts the execution of the program 176from the machine instruction at this address. 177.It Fa a_trsize 178Contains the size in bytes of the text relocation table. 179.It Fa a_drsize 180Contains the size in bytes of the data relocation table. 181.El 182.Pp 183The 184.In a.out.h 185include file defines several macros which use an 186.Fa exec 187structure to test consistency or to locate section offsets in the binary file. 188.Bl -tag -width N_BADMAG(exec) 189.It Fn N_BADMAG exec 190Nonzero if the 191.Fa a_magic 192field does not contain a recognized value. 193.It Fn N_TXTOFF exec 194The byte offset in the binary file of the beginning of the text segment. 195.It Fn N_SYMOFF exec 196The byte offset of the beginning of the symbol table. 197.It Fn N_STROFF exec 198The byte offset of the beginning of the string table. 199.El 200.Pp 201Relocation records have a standard format which 202is described by the 203.Fa relocation_info 204structure: 205.Bd -literal -offset indent 206struct relocation_info { 207 int r_address; 208 unsigned int r_symbolnum : 24, 209 r_pcrel : 1, 210 r_length : 2, 211 r_extern : 1, 212 r_baserel : 1, 213 r_jmptable : 1, 214 r_relative : 1, 215 r_copy : 1; 216}; 217.Ed 218.Pp 219The 220.Fa relocation_info 221fields are used as follows: 222.Bl -tag -width r_symbolnum 223.It Fa r_address 224Contains the byte offset of a pointer that needs to be link-edited. 225Text relocation offsets are reckoned from the start of the text segment, 226and data relocation offsets from the start of the data segment. 227The link editor adds the value that is already stored at this offset 228into the new value that it computes using this relocation record. 229.It Fa r_symbolnum 230Contains the ordinal number of a symbol structure 231in the symbol table (it is 232.Em not 233a byte offset). 234After the link editor resolves the absolute address for this symbol, 235it adds that address to the pointer that is undergoing relocation. 236(If the 237.Fa r_extern 238bit is clear, the situation is different; see below.) 239.It Fa r_pcrel 240If this is set, 241the link editor assumes that it is updating a pointer 242that is part of a machine code instruction using pc-relative addressing. 243The address of the relocated pointer is implicitly added 244to its value when the running program uses it. 245.It Fa r_length 246Contains the log base 2 of the length of the pointer in bytes; 2470 for 1-byte displacements, 1 for 2-byte displacements, 2482 for 4-byte displacements. 249.It Fa r_extern 250Set if this relocation requires an external reference; 251the link editor must use a symbol address to update the pointer. 252When the 253.Fa r_extern 254bit is clear, the relocation is 255.Sq local ; 256the link editor updates the pointer to reflect 257changes in the load addresses of the various segments, 258rather than changes in the value of a symbol (except when 259.Fa r_baserel 260is also set (see below). 261In this case, the content of the 262.Fa r_symbolnum 263field is an 264.Fa n_type 265value (see below); 266this type field tells the link editor 267what segment the relocated pointer points into. 268.It Fa r_baserel 269If set, the symbol, as identified by the 270.Fa r_symbolnum 271field, is to be relocated to an offset into the Global Offset Table. 272At run-time, the entry in the Global Offset Table at this offset is set to 273be the address of the symbol. 274.It Fa r_jmptable 275If set, the symbol, as identified by the 276.Fa r_symbolnum 277field, is to be relocated to an offset into the Procedure Linkage Table. 278.It Fa r_relative 279If set, this relocation is relative to the (run-time) load address of the 280image this object file is going to be a part of. 281This type of relocation 282only occurs in shared objects. 283.It Fa r_copy 284If set, this relocation record identifies a symbol whose contents should 285be copied to the location given in 286.Fa r_address . 287The copying is done by the run-time link-editor from a suitable data 288item in a shared object. 289.El 290.Pp 291Symbols map names to addresses (or more generally, strings to values). 292Since the link-editor adjusts addresses, 293a symbol's name must be used to stand for its address 294until an absolute value has been assigned. 295Symbols consist of a fixed-length record in the symbol table 296and a variable-length name in the string table. 297The symbol table is an array of 298.Fa nlist 299structures: 300.Bd -literal -offset indent 301struct nlist { 302 union { 303 const char *n_name; 304 long n_strx; 305 } n_un; 306 unsigned char n_type; 307 char n_other; 308 short n_desc; 309 unsigned long n_value; 310}; 311.Ed 312.Pp 313The fields are used as follows: 314.Bl -tag -width n_un.n_strx 315.It Fa n_un.n_strx 316Contains a byte offset into the string table 317for the name of this symbol. 318When a program accesses a symbol table with the 319.Xr nlist 3 320function, 321this field is replaced with the 322.Fa n_un.n_name 323field, which is a pointer to the string in memory. 324.It Fa n_type 325Used by the link editor to determine 326how to update the symbol's value. 327The 328.Fa n_type 329field is broken down into three sub-fields using bitmasks. 330The link editor treats symbols with the 331.Dv N_EXT 332type bit set as 333.Sq external 334symbols and permits references to them from other binary files. 335The 336.Dv N_TYPE 337mask selects bits of interest to the link editor: 338.Bl -tag -width N_TEXT 339.It Dv N_UNDF 340An undefined symbol. 341The link editor must locate an external symbol with the same name 342in another binary file to determine the absolute value of this symbol. 343As a special case, if the 344.Fa n_value 345field is nonzero and no binary file in the link-edit defines this symbol, 346the link-editor will resolve this symbol to an address 347in the bss segment, 348reserving an amount of bytes equal to 349.Fa n_value . 350If this symbol is undefined in more than one binary file 351and the binary files do not agree on the size, 352the link editor chooses the greatest size found across all binaries. 353.It Dv N_ABS 354An absolute symbol. 355The link editor does not update an absolute symbol. 356.It Dv N_TEXT 357A text symbol. 358This symbol's value is a text address and 359the link editor will update it when it merges binary files. 360.It Dv N_DATA 361A data symbol; similar to 362.Dv N_TEXT 363but for data addresses. 364The values for text and data symbols are not file offsets but 365addresses; to recover the file offsets, it is necessary 366to identify the loaded address of the beginning of the corresponding 367section and subtract it, then add the offset of the section. 368.It Dv N_BSS 369A bss symbol; like text or data symbols but 370has no corresponding offset in the binary file. 371.It Dv N_FN 372A filename symbol. 373The link editor inserts this symbol before 374the other symbols from a binary file when 375merging binary files. 376The name of the symbol is the filename given to the link editor, 377and its value is the first text address from that binary file. 378Filename symbols are not needed for link-editing or loading, 379but are useful for debuggers. 380.El 381.Pp 382The 383.Dv N_STAB 384mask selects bits of interest to symbolic debuggers 385such as 386.Xr gdb 1 Pq Pa ports/devel/gdb ; 387the values are described in 388.Xr stab 5 . 389.It Fa n_other 390This field provides information on the nature of the symbol independent of 391the symbol's location in terms of segments as determined by the 392.Fa n_type 393field. 394Currently, the lower 4 bits of the 395.Fa n_other 396field hold one of two values: 397.Dv AUX_FUNC 398and 399.Dv AUX_OBJECT 400(see 401.In link.h 402for their definitions). 403.Dv AUX_FUNC 404associates the symbol with a callable function, while 405.Dv AUX_OBJECT 406associates the symbol with data, irrespective of their locations in 407either the text or the data segment. 408This field is intended to be used by 409.Xr ld 1 410for the construction of dynamic executables. 411.It Fa n_desc 412Reserved for use by debuggers; passed untouched by the link editor. 413Different debuggers use this field for different purposes. 414.It Fa n_value 415Contains the value of the symbol. 416For text, data and bss symbols, this is an address; 417for other symbols (such as debugger symbols), 418the value may be arbitrary. 419.El 420.Pp 421The string table consists of an 422.Em unsigned long 423length followed by null-terminated symbol strings. 424The length represents the size of the entire table in bytes, 425so its minimum value (or the offset of the first string) 426is always 4 on 32-bit machines. 427.Sh SEE ALSO 428.Xr as 1 , 429.Xr gdb 1 Pq Pa ports/devel/gdb , 430.Xr ld 1 , 431.Xr brk 2 , 432.Xr execve 2 , 433.Xr nlist 3 , 434.Xr core 5 , 435.Xr elf 5 , 436.Xr link 5 , 437.Xr stab 5 438.Sh HISTORY 439The 440.In a.out.h 441include file appeared in 442.At v7 . 443.Sh BUGS 444Since not all of the supported architectures use the 445.Fa a_midmag 446field, 447it can be difficult to determine what 448architecture a binary will execute on 449without examining its actual machine code. 450Even with a machine identifier, 451the byte order of the 452.Fa exec 453header is machine-dependent. 454