1.\" $NetBSD: a.out.5,v 1.16 2002/02/13 08:18:08 ross Exp $ 2.\" 3.\" Copyright (c) 1991, 1993 4.\" The Regents of the University of California. All rights reserved. 5.\" 6.\" This man page is derived from documentation contributed to Berkeley by 7.\" Donn Seeley at UUNET Technologies, Inc. 8.\" 9.\" Redistribution and use in source and binary forms, with or without 10.\" modification, are permitted provided that the following conditions 11.\" are met: 12.\" 1. Redistributions of source code must retain the above copyright 13.\" notice, this list of conditions and the following disclaimer. 14.\" 2. Redistributions in binary form must reproduce the above copyright 15.\" notice, this list of conditions and the following disclaimer in the 16.\" documentation and/or other materials provided with the distribution. 17.\" 3. All advertising materials mentioning features or use of this software 18.\" must display the following acknowledgement: 19.\" This product includes software developed by the University of 20.\" California, Berkeley and its contributors. 21.\" 4. Neither the name of the University nor the names of its contributors 22.\" may be used to endorse or promote products derived from this software 23.\" without specific prior written permission. 24.\" 25.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 26.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 27.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 28.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 29.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 30.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 31.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 32.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 33.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 34.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 35.\" SUCH DAMAGE. 36.\" 37.\" @(#)a.out.5 8.1 (Berkeley) 6/5/93 38.\" 39.Dd June 5, 1993 40.Dt A.OUT 5 41.Os 42.Sh NAME 43.Nm a.out 44.Nd format of executable binary files 45.Sh SYNOPSIS 46.Fd #include \*[Lt]sys/types.h\*[Gt] 47.Fd #include \*[Lt]a.out.h\*[Gt] 48.Sh DESCRIPTION 49The include file 50.Aq Pa a.out.h 51declares three structures and several macros. 52The structures describe the format of 53executable machine code files 54.Pq Sq binaries 55on the system. 56.Pp 57A binary file consists of up to 7 sections. 58In order, these sections are: 59.Bl -tag -width "text relocations" 60.It exec header 61Contains parameters used by the kernel 62to load a binary file into memory and execute it, 63and by the link editor 64.Xr ld 1 65to combine a binary file with other binary files. 66This section is the only mandatory one. 67.It text segment 68Contains machine code and related data 69that are loaded into memory when a program executes. 70May be loaded read-only. 71.It data segment 72Contains initialized data; always loaded into writable memory. 73.It text relocations 74Contains records used by the link editor 75to update pointers in the text segment when combining binary files. 76.It data relocations 77Like the text relocation section, but for data segment pointers. 78.It symbol table 79Contains records used by the link editor 80to cross reference the addresses of named variables and functions 81.Pq Sq symbols 82between binary files. 83.It string table 84Contains the character strings corresponding to the symbol names. 85.El 86.Pp 87Every binary file begins with an 88.Fa exec 89structure: 90.Bd -literal -offset indent 91struct exec { 92 unsigned long a_midmag; 93 unsigned long a_text; 94 unsigned long a_data; 95 unsigned long a_bss; 96 unsigned long a_syms; 97 unsigned long a_entry; 98 unsigned long a_trsize; 99 unsigned long a_drsize; 100}; 101.Ed 102.Pp 103The fields have the following functions: 104.Bl -tag -width a_trsize 105.It Fa a_midmag 106This field is stored in network byte-order so that binaries for 107for machines with alternative byte orders can be distinguished. 108It has a number of sub-components accessed by the macros 109.Dv N_GETFLAG() , 110.Dv N_GETMID() , and 111.Dv N_GETMAGIC() , 112and set by the macro 113.Dv N_SETMAGIC() . 114.Pp 115The macro 116.Dv N_GETFLAG() 117returns a few flags: 118.Bl -tag -width EX_DYNAMIC 119.It Dv EX_DYNAMIC 120indicates that the executable requires the services of the run-time link editor. 121.It Dv EX_PIC 122indicates that the object contains position independent code. This flag is 123set by 124.Xr as 1 125when given the 126.Sq -k 127flag and is preserved by 128.Xr ld 1 129if necessary. 130.El 131.Pp 132If both EX_DYNAMIC and EX_PIC are set, the object file is a position independent 133executable image (e.g. a shared library), which is to be loaded into the 134process address space by the run-time link editor. 135.Pp 136The macro 137.Dv N_GETMID() 138returns the machine-id. 139This indicates which machine(s) the binary is intended to run on. 140.Pp 141.Dv N_GETMAGIC() 142specifies the magic number, which uniquely identifies binary files 143and distinguishes different loading conventions. 144The field must contain one of the following values: 145.Bl -tag -width ZMAGIC 146.It Dv OMAGIC 147The text and data segments immediately follow the header 148and are contiguous. 149The kernel loads both text and data segments into writable memory. 150.It Dv NMAGIC 151As with 152.Dv OMAGIC , 153text and data segments immediately follow the header and are contiguous. 154However, the kernel loads the text into read-only memory 155and loads the data into writable memory at the next 156page boundary after the text. 157.It Dv ZMAGIC 158The kernel loads individual pages on demand from the binary. 159The header, text segment and data segment are all 160padded by the link editor to a multiple of the page size. 161Pages that the kernel loads from the text segment are read-only, 162while pages from the data segment are writable. 163.El 164.It Fa a_text 165Contains the size of the text segment in bytes. 166.It Fa a_data 167Contains the size of the data segment in bytes. 168.It Fa a_bss 169Contains the number of bytes in the 170.Sq bss segment 171and is used by the kernel to set the initial break 172.Pq Xr brk 2 173after the data segment. 174The kernel loads the program so that this amount of writable memory 175appears to follow the data segment and initially reads as zeroes. 176.It Fa a_syms 177Contains the size in bytes of the symbol table section. 178.It Fa a_entry 179Contains the address in memory of the entry point 180of the program after the kernel has loaded it; 181the kernel starts the execution of the program 182from the machine instruction at this address. 183.It Fa a_trsize 184Contains the size in bytes of the text relocation table. 185.It Fa a_drsize 186Contains the size in bytes of the data relocation table. 187.El 188.Pp 189The 190.Pa a.out.h 191include file defines several macros which use an 192.Fa exec 193structure to test consistency or to locate section offsets in the binary file. 194.Bl -tag -width N_BADMAG(exec) 195.It Fn N_BADMAG exec 196Nonzero if the 197.Fa a_magic 198field does not contain a recognized value. 199.It Fn N_TXTOFF exec 200The byte offset in the binary file of the beginning of the text segment. 201.It Fn N_SYMOFF exec 202The byte offset of the beginning of the symbol table. 203.It Fn N_STROFF exec 204The byte offset of the beginning of the string table. 205.El 206.Pp 207Relocation records have a standard format which 208is described by the 209.Fa relocation_info 210structure: 211.Bd -literal -offset indent 212struct relocation_info { 213 int r_address; 214 unsigned int r_symbolnum : 24, 215 r_pcrel : 1, 216 r_length : 2, 217 r_extern : 1, 218 r_baserel : 1, 219 r_jmptable : 1, 220 r_relative : 1, 221 r_copy : 1; 222}; 223.Ed 224.Pp 225The 226.Fa relocation_info 227fields are used as follows: 228.Bl -tag -width r_symbolnum 229.It Fa r_address 230Contains the byte offset of a pointer that needs to be link-edited. 231Text relocation offsets are reckoned from the start of the text segment, 232and data relocation offsets from the start of the data segment. 233The link editor adds the value that is already stored at this offset 234into the new value that it computes using this relocation record. 235.It Fa r_symbolnum 236Contains the ordinal number of a symbol structure 237in the symbol table (it is 238.Em not 239a byte offset). 240After the link editor resolves the absolute address for this symbol, 241it adds that address to the pointer that is undergoing relocation. 242(If the 243.Fa r_extern 244bit is clear, the situation is different; see below.) 245.It Fa r_pcrel 246If this is set, 247the link editor assumes that it is updating a pointer 248that is part of a machine code instruction using pc-relative addressing. 249The address of the relocated pointer is implicitly added 250to its value when the running program uses it. 251.It Fa r_length 252Contains the log base 2 of the length of the pointer in bytes; 2530 for 1-byte displacements, 1 for 2-byte displacements, 2542 for 4-byte displacements. 255.It Fa r_extern 256Set if this relocation requires an external reference; 257the link editor must use a symbol address to update the pointer. 258When the 259.Fa r_extern 260bit is clear, the relocation is 261.Sq local ; 262the link editor updates the pointer to reflect 263changes in the load addresses of the various segments, 264rather than changes in the value of a symbol (except when 265.Fa r_baserel 266is also set, see below). 267In this case, the content of the 268.Fa r_symbolnum 269field is an 270.Fa n_type 271value (see below); 272this type field tells the link editor 273what segment the relocated pointer points into. 274.It Fa r_baserel 275If set, the symbol, as identified by the 276.Fa r_symbolnum 277field, is to be relocated to an offset into the Global Offset Table. 278At run-time, the entry in the Global Offset Table at this offset is set to 279be the address of the symbol. 280.It Fa r_jmptable 281If set, the symbol, as identified by the 282.Fa r_symbolnum 283field, is to be relocated to an offset into the Procedure Linkage Table. 284.It Fa r_relative 285If set, this relocation is relative to the (run-time) load address of the 286image this object file is going to be a part of. This type of relocation 287only occurs in shared objects. 288.It Fa r_copy 289If set, this relocation record identifies a symbol whose contents should 290be copied to the location given in 291.Fa r_address . 292The copying is done by the run-time link-editor from a suitable data 293item in a shared object. 294.El 295.Pp 296Symbols map names to addresses (or more generally, strings to values). 297Since the link-editor adjusts addresses, 298a symbol's name must be used to stand for its address 299until an absolute value has been assigned. 300Symbols consist of a fixed-length record in the symbol table 301and a variable-length name in the string table. 302The symbol table is an array of 303.Fa nlist 304structures: 305.Bd -literal -offset indent 306struct nlist { 307 union { 308 char *n_name; 309 long n_strx; 310 } n_un; 311 unsigned char n_type; 312 char n_other; 313 short n_desc; 314 unsigned long n_value; 315}; 316.Ed 317.Pp 318The fields are used as follows: 319.Bl -tag -width n_un.n_strx 320.It Fa n_un.n_strx 321Contains a byte offset into the string table 322for the name of this symbol. 323When a program accesses a symbol table with the 324.Xr nlist 3 325function, 326this field is replaced with the 327.Fa n_un.n_name 328field, which is a pointer to the string in memory. 329.It Fa n_type 330Used by the link editor to determine 331how to update the symbol's value. 332The 333.Fa n_type 334field is broken down into three sub-fields using bitmasks. 335The link editor treats symbols with the 336.Dv N_EXT 337type bit set as 338.Sq external 339symbols and permits references to them from other binary files. 340The 341.Dv N_TYPE 342mask selects bits of interest to the link editor: 343.Bl -tag -width N_TEXT 344.It Dv N_UNDF 345An undefined symbol. 346The link editor must locate an external symbol with the same name 347in another binary file to determine the absolute value of this symbol. 348As a special case, if the 349.Fa n_value 350field is nonzero and no binary file in the link-edit defines this symbol, 351the link-editor will resolve this symbol to an address 352in the bss segment, 353reserving an amount of bytes equal to 354.Fa n_value . 355If this symbol is undefined in more than one binary file 356and the binary files do not agree on the size, 357the link editor chooses the greatest size found across all binaries. 358.It Dv N_ABS 359An absolute symbol. 360The link editor does not update an absolute symbol. 361.It Dv N_TEXT 362A text symbol. 363This symbol's value is a text address and 364the link editor will update it when it merges binary files. 365.It Dv N_DATA 366A data symbol; similar to 367.Dv N_TEXT 368but for data addresses. 369The values for text and data symbols are not file offsets but 370addresses; to recover the file offsets, it is necessary 371to identify the loaded address of the beginning of the corresponding 372section and subtract it, then add the offset of the section. 373.It Dv N_BSS 374A bss symbol; like text or data symbols but 375has no corresponding offset in the binary file. 376.It Dv N_FN 377A filename symbol. 378The link editor inserts this symbol before 379the other symbols from a binary file when 380merging binary files. 381The name of the symbol is the filename given to the link editor, 382and its value is the first text address from that binary file. 383Filename symbols are not needed for link-editing or loading, 384but are useful for debuggers. 385.El 386.Pp 387The 388.Dv N_STAB 389mask selects bits of interest to symbolic debuggers 390such as 391.Xr gdb 1 ; 392the values are described in 393.Xr stab 5 . 394.It Fa n_other 395This field provides information on the nature of the symbol independent of 396the symbol's location in terms of segments as determined by the 397.Fa n_type 398field. Currently, the lower 4 bits of the 399.Fa n_other 400field hold one of two values: 401.Dv AUX_FUNC 402and 403.Dv AUX_OBJECT 404.Po 405see 406.Aq Pa link.h 407for their definitions 408.Pc . 409.Dv AUX_FUNC 410associates the symbol with a callable function, while 411.Dv AUX_OBJECT 412associates the symbol with data, irrespective of their locations in 413either the text or the data segment. 414This field is intended to be used by 415.Xr ld 1 416for the construction of dynamic executables. 417.It Fa n_desc 418Reserved for use by debuggers; passed untouched by the link editor. 419Different debuggers use this field for different purposes. 420.It Fa n_value 421Contains the value of the symbol. 422For text, data and bss symbols, this is an address; 423for other symbols (such as debugger symbols), 424the value may be arbitrary. 425.El 426.Pp 427The string table consists of an 428.Em unsigned long 429length followed by null-terminated symbol strings. 430The length represents the size of the entire table in bytes, 431so its minimum value (or the offset of the first string) 432is always 4 on 32-bit machines. 433.Sh SEE ALSO 434.Xr as 1 , 435.Xr gdb 1 , 436.Xr ld 1 , 437.Xr brk 2 , 438.Xr execve 2 , 439.Xr nlist 3 , 440.Xr core 5 , 441.Xr elf 5 , 442.Xr link 5 , 443.Xr stab 5 444.Sh HISTORY 445The 446.Pa a.out.h 447include file appeared in 448.At v7 . 449.Sh BUGS 450Nobody seems to agree on what 451.Em bss 452stands for. 453.Pp 454New binary file formats may be supported in the future, 455and they probably will not be compatible at any level 456with this ancient format. 457