4 @include configdoc.texi
5 @c (configdoc.texi is generated by the Makefile)
12 * Ld: (ld). The GNU linker.
18 This file documents the GNU linker LD.
20 Copyright (C) 1991, 1992, 1993, 1994 Free Software Foundation, Inc.
22 Permission is granted to make and distribute verbatim copies of
23 this manual provided the copyright notice and this permission notice
24 are preserved on all copies.
26 Permission is granted to copy and distribute modified versions of this
27 manual under the conditions for verbatim copying, provided also that
28 the entire resulting derived work is distributed under the terms of a
29 permission notice identical to this one.
31 Permission is granted to copy and distribute translations of this manual
32 into another language, under the above conditions for modified versions.
35 Permission is granted to process this file through Tex and print the
36 results, provided the printed document carries copying permission
37 notice identical to this one except for the removal of this paragraph
38 (this paragraph not being relevant to the printed manual).
44 @setchapternewpage odd
45 @settitle Using LD, the GNU linker
48 @subtitle The GNU linker
50 @subtitle @code{ld} version 2
51 @subtitle January 1994
52 @author Steve Chamberlain and Roland Pesch
53 @author Cygnus Support
58 \hfill Cygnus Support\par
59 \hfill steve\@cygnus.com, pesch\@cygnus.com\par
60 \hfill {\it Using LD, the GNU linker}\par
61 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
62 \hfill and Roland Pesch (pesch\@cygnus.com)\par
64 \global\parindent=0pt % Steve likes it this way.
67 @vskip 0pt plus 1filll
68 Copyright @copyright{} 1991, 1992, 1993, 1994 Free Software Foundation, Inc.
70 Permission is granted to make and distribute verbatim copies of
71 this manual provided the copyright notice and this permission notice
72 are preserved on all copies.
74 Permission is granted to copy and distribute modified versions of this
75 manual under the conditions for verbatim copying, provided also that
76 the entire resulting derived work is distributed under the terms of a
77 permission notice identical to this one.
79 Permission is granted to copy and distribute translations of this manual
80 into another language, under the above conditions for modified versions.
83 @c FIXME: Talk about importance of *order* of args, cmds to linker!
88 This file documents the GNU linker ld.
92 * Invocation:: Invocation
93 * Commands:: Command Language
95 * Machine Dependent:: Machine Dependent Features
99 * H8/300:: ld and the H8/300
102 * Hitachi:: ld and other Hitachi micros
105 * i960:: ld and the Intel 960 family
108 @ifclear SingleFormat
111 @c Following blank line required for remaining bug in makeinfo conds/menus
113 * MRI:: MRI Compatible Script Files
122 @cindex what is this?
123 @code{ld} combines a number of object and archive files, relocates
124 their data and ties up symbol references. Usually the last step in
125 compiling a program is to run @code{ld}.
127 @code{ld} accepts Linker Command Language files written in
128 a superset of AT&T's Link Editor Command Language syntax,
129 to provide explicit and total control over the linking process.
131 @ifclear SingleFormat
132 This version of @code{ld} uses the general purpose BFD libraries
133 to operate on object files. This allows @code{ld} to read, combine, and
134 write object files in many different formats---for example, COFF or
135 @code{a.out}. Different formats may be linked together to produce any
136 available kind of object file. @xref{BFD}, for more information.
139 Aside from its flexibility, the GNU linker is more helpful than other
140 linkers in providing diagnostic information. Many linkers abandon
141 execution immediately upon encountering an error; whenever possible,
142 @code{ld} continues executing, allowing you to identify other errors
143 (or, in some cases, to get an output file in spite of the error).
148 The GNU linker @code{ld} is meant to cover a broad range of situations,
149 and to be as compatible as possible with other linkers. As a result,
150 you have many choices to control its behavior.
154 * Options:: Command Line Options
155 * Environment:: Environment Variables
159 @section Command Line Options
164 Here is a summary of the options you can use on the @code{ld} command
167 @c FIXME! -relax only avail h8/300, i960. Conditionals screwed in examples.
169 ld [ -o @var{output} ] @var{objfile}@dots{}
170 [ -A@var{architecture} ] [ -b @var{input-format} ] [ -Bstatic ]
171 [ -c @var{MRI-commandfile} ] [ -d | -dc | -dp ]
172 [ -defsym @var{symbol}=@var{expression} ]
173 [ -dynamic-linker @var{file} ]
174 [ -e @var{entry} ] [ -F ] [ -F @var{format} ]
175 [ -format @var{input-format} ] [ -g ] [ -G @var{size} ] [ -help ]
176 [ -i ] [ -l@var{archive} ] [ -L@var{searchdir} ] [ -M ]
177 [ -Map @var{mapfile} ] [ -m @var{emulation} ] [ -N | -n ]
178 [ -noinhibit-exec ] [ -oformat @var{output-format} ]
179 [ -R @var{filename} ] [ -relax ] [ -retain-symbols-file @var{filename} ]
180 [ -r | -Ur ] [ -rpath @var{dir} ] [ -S ] [ -s ] [ -sort-common ]
181 [ -stats ] [ -T @var{commandfile} ]
182 [ -Ttext @var{org} ] [ -Tdata @var{org} ]
183 [ -Tbss @var{org} ] [ -t ] [ -traditional-format ]
184 [ -u @var{symbol}] [-V] [-v] [ -version ]
185 [ -warn-common ] [ -y @var{symbol} ] [ -X ] [-x ]
186 [ -( [ archives ] -) ] [ --start-group [ archives ] --end-group ]
189 This plethora of command-line options may seem intimidating, but in
190 actual practice few of them are used in any particular context.
191 @cindex standard Unix system
192 For instance, a frequent use of @code{ld} is to link standard Unix
193 object files on a standard, supported Unix system. On such a system, to
194 link a file @code{hello.o}:
197 ld -o @var{output} /lib/crt0.o hello.o -lc
200 This tells @code{ld} to produce a file called @var{output} as the
201 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
202 the library @code{libc.a}, which will come from the standard search
203 directories. (See the discussion of the @samp{-l} option below.)
205 The command-line options to @code{ld} may be specified in any order, and
206 may be repeated at will. Repeating most options with a
207 different argument will either have no further effect, or override prior
208 occurrences (those further to the left on the command line) of that
211 @ifclear SingleFormat
212 The exceptions---which may meaningfully be used more than once---are
213 @samp{-A}, @samp{-b} (or its synonym @samp{-format}), @samp{-defsym},
214 @samp{-L}, @samp{-l}, @samp{-R}, and @samp{-u}.
217 The exceptions---which may meaningfully be used more than once---are
218 @samp{-A}, @samp{-defsym}, @samp{-L}, @samp{-l}, @samp{-R}, and @samp{-u}.
222 The list of object files to be linked together, shown as @var{objfile}@dots{},
223 may follow, precede, or be mixed in with command-line options, except that
224 an @var{objfile} argument may not be placed between an option and
227 Usually the linker is invoked with at least one object file, but you can
228 specify other forms of binary input files using @samp{-l}, @samp{-R},
229 and the script command language. If @emph{no} binary input files at all
230 are specified, the linker does not produce any output, and issues the
231 message @samp{No input files}.
233 For options whose names are a single letter,
234 option arguments must either follow the option letter without intervening
235 whitespace, or be given as separate arguments immediately following the
236 option that requires them.
238 For options whose names are multiple letters, either one dash or two can
239 precede the option name; for example, @samp{--oformat} and
240 @samp{-oformat} are equivalent. Arguments to multiple-letter options
241 must either be separated from the option name by an equals sign, or be
242 given as separate arguments immediately following the option that
243 requires them. For example, @samp{--oformat srec} and
244 @samp{--oformat=srec} are equivalent. Unique abbreviations of the names
245 of multiple-letter options are accepted.
249 @cindex architectures
251 @item -A@var{architecture}
252 In the current release of @code{ld}, this option is useful only for the
253 Intel 960 family of architectures. In that @code{ld} configuration, the
254 @var{architecture} argument identifies the particular architecture in
255 the 960 family, enabling some safeguards and modifying the
256 archive-library search path. @xref{i960,,@code{ld} and the Intel 960
257 family}, for details.
259 Future releases of @code{ld} may support similar functionality for
260 other architecture families.
263 @ifclear SingleFormat
264 @cindex binary input format
265 @kindex -b @var{format}
267 @item -b @var{input-format}
269 @code{ld} may be configured to support more than one kind of object
270 file. If your @code{ld} is configured this way, you can use the
271 @samp{-b} option to specify the binary format for input object files
272 that follow this option on the command line. Even when @code{ld} is
273 configured to support alternative object formats, you don't usually need
274 to specify this, as @code{ld} should be configured to expect as a
275 default input format the most usual format on each machine.
276 @var{input-format} is a text string, the name of a particular format
277 supported by the BFD libraries. (You can list the available binary
278 formats with @samp{objdump -i}.) @w{@samp{-format @var{input-format}}}
279 has the same effect, as does the script command @code{TARGET}.
282 You may want to use this option if you are linking files with an unusual
283 binary format. You can also use @samp{-b} to switch formats explicitly (when
284 linking object files of different formats), by including
285 @samp{-b @var{input-format}} before each group of object files in a
288 The default format is taken from the environment variable
293 You can also define the input
294 format from a script, using the command @code{TARGET}; see @ref{Option
300 Do not link against shared libraries. This option is accepted for
301 command-line compatibility with the SunOS linker.
303 @kindex -c @var{MRI-cmdfile}
304 @cindex compatibility, MRI
305 @item -c @var{MRI-commandfile}
306 For compatibility with linkers produced by MRI, @code{ld} accepts script
307 files written in an alternate, restricted command language, described in
308 @ref{MRI,,MRI Compatible Script Files}. Introduce MRI script files with
309 the option @samp{-c}; use the @samp{-T} option to run linker
310 scripts written in the general-purpose @code{ld} scripting language.
311 If @var{MRI-cmdfile} does not exist, @code{ld} looks for it in the directories
312 specified by any @samp{-L} options.
314 @cindex common allocation
321 These three options are equivalent; multiple forms are supported for
322 compatibility with other linkers. They
323 assign space to common symbols even if a relocatable output file is
324 specified (with @samp{-r}). The script command
325 @code{FORCE_COMMON_ALLOCATION} has the same effect. @xref{Option
328 @cindex symbols, from command line
329 @kindex -defsym @var{symbol}=@var{exp}
330 @item -defsym @var{symbol}=@var{expression}
331 Create a global symbol in the output file, containing the absolute
332 address given by @var{expression}. You may use this option as many
333 times as necessary to define multiple symbols in the command line. A
334 limited form of arithmetic is supported for the @var{expression} in this
335 context: you may give a hexadecimal constant or the name of an existing
336 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
337 constants or symbols. If you need more elaborate expressions, consider
338 using the linker command language from a script (@pxref{Assignment, ,
339 Assignment: Symbol Definitions}). @emph{Note:} there should be no
340 white space between @var{symbol}, the equals sign (``@key{=}''), and
344 @cindex dynamic linker, from command line
345 @kindex -dynamic-linker @var{file}
346 @item -dynamic-linker @var{file}
347 Set the name of the dynamic linker. This is only meaningful when
348 generating dynamically linked ELF executables. The default dynamic
349 linker is normally correct; don't use this unless you know what you are
353 @cindex entry point, from command line
354 @kindex -e @var{entry}
356 Use @var{entry} as the explicit symbol for beginning execution of your
357 program, rather than the default entry point. @xref{Entry Point}, for a
358 discussion of defaults and other ways of specifying the
361 @ifclear SingleFormat
364 @itemx -F@var{format}
365 Ignored. Some older linkers used this option throughout a compilation
366 toolchain for specifying object-file format for both input and output
367 object files. The mechanisms @code{ld} uses for this purpose (the
368 @samp{-b} or @samp{-format} options for input files, @samp{-oformat}
369 option or the @code{TARGET} command in linker scripts for output files,
370 the @code{GNUTARGET} environment variable) are more flexible, but
371 @code{ld} accepts the @samp{-F} option for compatibility with scripts
372 written to call the old linker.
375 @item -format @var{input-format}
376 Synonym for @samp{-b @var{input-format}}.
381 Ignored. Provided for compatibility with other tools.
386 @itemx -G @var{value}
387 Set the maximum size of objects to be optimized using the GP register to
388 @var{size} under MIPS ECOFF. Ignored for other object file formats.
394 Print a summary of the command-line options on the standard output and exit.
397 @cindex incremental link
399 Perform an incremental link (same as option @samp{-r}).
401 @cindex archive files, from cmd line
402 @kindex -l@var{archive}
404 Add archive file @var{archive} to the list of files to link. This
405 option may be used any number of times. @code{ld} will search its
406 path-list for occurrences of @code{lib@var{ar}.a} for every @var{archive}
409 @cindex search directory, from cmd line
411 @item -L@var{searchdir}
412 @itemx -L @var{searchdir}
413 Add path @var{searchdir} to the list of paths that @code{ld} will search
414 for archive libraries and @code{ld} control scripts. You may use this
415 option any number of times.
418 The default set of paths searched (without being specified with
419 @samp{-L}) depends on which emulation mode @code{ld} is using, and in
420 some cases also on how it was configured. @xref{Environment}.
423 The paths can also be specified in a link script with the
424 @code{SEARCH_DIR} command.
429 Print (to the standard output) a link map---diagnostic information
430 about where symbols are mapped by @code{ld}, and information on global
431 common storage allocation.
435 @item -Map @var{mapfile}
436 Print to the file @var{mapfile} a link map---diagnostic information
437 about where symbols are mapped by @code{ld}, and information on global
438 common storage allocation.
441 @kindex -m @var{emulation}
442 @item -m@var{emulation}
443 @itemx -m @var{emulation}
444 Emulate the @var{emulation} linker. You can list the available
445 emulations with the @samp{-V} option. The
446 default depends on how your @code{ld} was configured.
449 @cindex read/write from cmd line
452 Set the text and data sections to be readable and writable. Also, do
453 not page-align the data segment. If the output format supports Unix
454 style magic numbers, mark the output as @code{OMAGIC}.
457 @cindex read-only text
460 Set the text segment to be read only, and mark the output as
461 @code{NMAGIC} if possible.
463 @item -noinhibit-exec
464 @cindex output file after errors
465 @kindex -noinhibit-exec
466 Retain the executable output file whenever it is still usable.
467 Normally, the linker will not produce an output file if it encounters
468 errors during the link process; it exits without writing an output file
469 when it issues any error whatsoever.
471 @item -o @var{output}
472 @kindex -o @var{output}
473 @cindex naming the output file
474 Use @var{output} as the name for the program produced by @code{ld}; if this
475 option is not specified, the name @file{a.out} is used by default. The
476 script command @code{OUTPUT} can also specify the output file name.
478 @ifclear SingleFormat
480 @item -oformat @var{output-format}
481 @code{ld} may be configured to support more than one kind of object
482 file. If your @code{ld} is configured this way, you can use the
483 @samp{-oformat} option to specify the binary format for the output
484 object file. Even when @code{ld} is configured to support alternative
485 object formats, you don't usually need to specify this, as @code{ld}
486 should be configured to produce as a default output format the most
487 usual format on each machine. @var{output-format} is a text string, the
488 name of a particular format supported by the BFD libraries. (You can
489 list the available binary formats with @samp{objdump -i}.) The script
490 command @code{OUTPUT_FORMAT} can also specify the output format, but
491 this option overrides it. @xref{BFD}.
494 @item -R @var{filename}
495 @kindex -R @var{file}
496 @cindex symbol-only input
497 Read symbol names and their addresses from @var{filename}, but do not
498 relocate it or include it in the output. This allows your output file
499 to refer symbolically to absolute locations of memory defined in other
503 @cindex synthesizing linker
504 @cindex relaxing addressing modes
506 An option with machine dependent effects.
508 Currently this option is only supported on the H8/300 and the Intel 960.
511 @xref{H8/300,,@code{ld} and the H8/300}.
514 @xref{i960,, @code{ld} and the Intel 960 family}.
517 On some platforms, the @samp{-relax} option performs global optimizations that
518 become possible when the linker resolves addressing in the program, such
519 as relaxing address modes and synthesizing new instructions in the
523 On platforms where this is not supported, @samp{-relax} is accepted, but
527 @item -retain-symbols-file @var{filename}
528 @cindex retaining specified symbols
529 @cindex stripping all but some symbols
530 @cindex symbols, retaining selectively
531 Retain @emph{only} the symbols listed in the file @var{filename},
532 discarding all others. @var{filename} is simply a flat file, with one
533 symbol name per line. This option is especially useful in environments
537 where a large global symbol table is accumulated gradually, to conserve
540 @samp{-retain-symbols-file} does @emph{not} discard undefined symbols,
541 or symbols needed for relocations.
543 You may only specify @samp{-retain-symbols-file} once in the command
544 line. It overrides @samp{-s} and @samp{-S}.
547 @item -rpath @var{dir}
548 @cindex runtime library search path
550 Add a directory to the runtime library search path. This is only
551 meaningful when linking an ELF executable with shared objects. All
552 -rpath arguments are concatenated and passed to the runtime linker,
553 which uses them to locate shared objects at runtime.
557 @cindex relocatable output
560 Generate relocatable output---i.e., generate an output file that can in
561 turn serve as input to @code{ld}. This is often called @dfn{partial
562 linking}. As a side effect, in environments that support standard Unix
563 magic numbers, this option also sets the output file's magic number to
566 If this option is not specified, an absolute file is produced. When
567 linking C++ programs, this option @emph{will not} resolve references to
568 constructors; to do that, use @samp{-Ur}.
570 This option does the same thing as @samp{-i}.
573 @cindex strip debugger symbols
575 Omit debugger symbol information (but not all symbols) from the output file.
578 @cindex strip all symbols
580 Omit all symbol information from the output file.
583 Normally, when @code{ld} places the global common symbols in the
584 appropriate output sections, it sorts them by size. First come all the
585 one byte symbols, then all the two bytes, then all the four bytes, and
586 then everything else. This is to prevent gaps between symbols due to
587 alignment constraints. This option disables that sorting.
590 Compute and display statistics about the operation of the linker,
591 such as execution time and memory usage.
593 @item -Tbss @var{org}
594 @kindex -Tbss @var{org}
595 @itemx -Tdata @var{org}
596 @kindex -Tdata @var{org}
597 @itemx -Ttext @var{org}
598 @kindex -Ttext @var{org}
599 @cindex segment origins, cmd line
600 Use @var{org} as the starting address for---respectively---the
601 @code{bss}, @code{data}, or the @code{text} segment of the output file.
602 @var{org} must be a single hexadecimal integer;
603 for compatibility with other linkers, you may omit the leading
604 @samp{0x} usually associated with hexadecimal values.
606 @item -T @var{commandfile}
607 @itemx -T@var{commandfile}
608 @kindex -T @var{script}
610 Read link commands from the file @var{commandfile}. These commands
611 replace @code{ld}'s default link script (rather than adding
612 to it), so @var{commandfile} must specify everything necessary to describe
613 the target format. @xref{Commands}. If @var{commandfile} does not
614 exist, @code{ld} looks for it in the directories specified by any
615 preceding @samp{-L} options. Multiple @samp{-T} options accumulate.
619 @cindex input files, displaying
621 Print the names of the input files as @code{ld} processes them.
623 @kindex -traditional-format
624 @cindex traditional format
625 @item -traditional-format
626 For some targets, the output of @code{ld} is different in some ways from
627 the output of some existing linker. This switch requests @code{ld} to
628 use the traditional format instead.
631 For example, on SunOS, @code{ld} combines duplicate entries in the
632 symbol string table. This can reduce the size of an output file with
633 full debugging information by over 30 percent. Unfortunately, the SunOS
634 @code{dbx} program can not read the resulting program (@code{gdb} has no
635 trouble). The @samp{-traditional-format} switch tells @code{ld} to not
636 combine duplicate entries.
638 @item -u @var{symbol}
639 @kindex -u @var{symbol}
640 @cindex undefined symbol
641 Force @var{symbol} to be entered in the output file as an undefined symbol.
642 Doing this may, for example, trigger linking of additional modules from
643 standard libraries. @samp{-u} may be repeated with different option
644 arguments to enter additional undefined symbols.
645 @c Nice idea, but no such command: This option is equivalent
646 @c to the @code{EXTERN} linker command.
651 For anything other than C++ programs, this option is equivalent to
652 @samp{-r}: it generates relocatable output---i.e., an output file that can in
653 turn serve as input to @code{ld}. When linking C++ programs, @samp{-Ur}
654 @emph{does} resolve references to constructors, unlike @samp{-r}.
655 It does not work to use @samp{-Ur} on files that were themselves linked
656 with @samp{-Ur}; once the constructor table has been built, it cannot
657 be added to. Use @samp{-Ur} only for the last partial link, and
658 @samp{-r} for the others.
663 Display the version number for @code{ld} and list the linker emulations
664 supported. Display which input files can and cannot be opened.
669 Display the version number for @code{ld}.
673 Display the version number for @code{ld} and exit.
677 @cindex warnings, on combining symbols
678 @cindex combining symbols, warnings on
679 Warn when a common symbol is combined with another common symbol or with
680 a symbol definition. Unix linkers allow this somewhat sloppy practice,
681 but linkers on some other operating systems do not. This option allows
682 you to find potential problems from combining global symbols.
683 Unfortunately, some C libraries use this practice, so you may get some
684 warnings about symbols in the libraries as well as in your programs.
686 There are three kinds of global symbols, illustrated here by C examples:
690 A definition, which goes in the initialized data section of the output
694 An undefined reference, which does not allocate space.
695 There must be either a definition or a common symbol for the
699 A common symbol. If there are only (one or more) common symbols for a
700 variable, it goes in the uninitialized data area of the output file.
701 The linker merges multiple common symbols for the same variable into a
702 single symbol. If they are of different sizes, it picks the largest
703 size. The linker turns a common symbol into a declaration, if there is
704 a definition of the same variable.
707 The @samp{-warn-common} option can produce five kinds of warnings. Each
708 warning consists of a pair of lines: the first describes the symbol just
709 encountered, and the second describes the previous symbol encountered
710 with the same name. One or both of the two symbols will be a common
715 Turning a common symbol into a reference, because there is already a
716 definition for the symbol.
718 @var{file}(@var{section}): warning: common of `@var{symbol}'
719 overridden by definition
720 @var{file}(@var{section}): warning: defined here
724 Turning a common symbol into a reference, because a later definition for
725 the symbol is encountered. This is the same as the previous case,
726 except that the symbols are encountered in a different order.
728 @var{file}(@var{section}): warning: definition of `@var{symbol}'
730 @var{file}(@var{section}): warning: common is here
734 Merging a common symbol with a previous same-sized common symbol.
736 @var{file}(@var{section}): warning: multiple common
738 @var{file}(@var{section}): warning: previous common is here
742 Merging a common symbol with a previous larger common symbol.
744 @var{file}(@var{section}): warning: common of `@var{symbol}'
745 overridden by larger common
746 @var{file}(@var{section}): warning: larger common is here
750 Merging a common symbol with a previous smaller common symbol. This is
751 the same as the previous case, except that the symbols are
752 encountered in a different order.
754 @var{file}(@var{section}): warning: common of `@var{symbol}'
755 overriding smaller common
756 @var{file}(@var{section}): warning: smaller common is here
761 @cindex local symbols, deleting
762 @cindex L, deleting symbols beginning
764 If @samp{-s} or @samp{-S} is also specified, delete only local symbols
765 beginning with @samp{L}.
768 @cindex deleting local symbols
770 If @samp{-s} or @samp{-S} is also specified, delete all local symbols,
771 not just those beginning with @samp{L}.
773 @item -y @var{symbol}
774 @kindex -y @var{symbol}
775 @cindex symbol tracing
776 Print the name of each linked file in which @var{symbol} appears. This
777 option may be given any number of times. On many systems it is necessary
778 to prepend an underscore.
780 This option is useful when you have an undefined symbol in your link but
781 don't know where the reference is coming from.
783 @item -( @var{archives} -)
784 @itemx --start-group @var{archives} --end-group
786 @cindex groups of archives
787 The @var{archives} should be a list of archive files. They may be
788 either explicit file names, or @samp{-l} options.
790 The specified archives are searched repeatedly until no new undefined
791 references are created. Normally, an archive is searched only once in
792 the order that it is specified on the command line. If a symbol in that
793 archive is needed to resolve an undefined symbol referred to by an
794 object in an archive that appears later on the command line, the linker
795 would not be able to resolve that reference. By grouping the archives,
796 they all be searched repeatedly until all possible references are
799 Using this option has a significant performance cost. It is best to use
800 it only when there are unavoidable circular references between two or
806 @section Environment Variables
808 You can change the behavior of @code{ld} with the environment
809 variable @code{GNUTARGET}.
812 @cindex default input format
813 @code{GNUTARGET} determines the input-file object format if you don't
814 use @samp{-b} (or its synonym @samp{-format}). Its value should be one
815 of the BFD names for an input format (@pxref{BFD}). If there is no
816 @code{GNUTARGET} in the environment, @code{ld} uses the natural format
817 of the target. If @code{GNUTARGET} is set to @code{default} then BFD attempts to discover the
818 input format by examining binary input files; this method often
819 succeeds, but there are potential ambiguities, since there is no method
820 of ensuring that the magic number used to specify object-file formats is
821 unique. However, the configuration procedure for BFD on each system
822 places the conventional format for that system first in the search-list,
823 so ambiguities are resolved in favor of convention.
827 @chapter Command Language
829 @cindex command files
830 The command language provides explicit control over the link process,
831 allowing complete specification of the mapping between the linker's
832 input files and its output. It controls:
841 addresses of sections
843 placement of common blocks
846 You may supply a command file (also known as a link script) to the
847 linker either explicitly through the @samp{-T} option, or implicitly as
848 an ordinary file. If the linker opens a file which it cannot recognize
849 as a supported object or archive format, it reports an error.
852 * Scripts:: Linker Scripts
853 * Expressions:: Expressions
854 * MEMORY:: MEMORY Command
855 * SECTIONS:: SECTIONS Command
856 * Entry Point:: The Entry Point
857 * Option Commands:: Option Commands
861 @section Linker Scripts
862 The @code{ld} command language is a collection of statements; some are
863 simple keywords setting a particular option, some are used to select and
864 group input files or name output files; and two statement
865 types have a fundamental and pervasive impact on the linking process.
867 @cindex fundamental script commands
868 @cindex commands, fundamental
869 @cindex output file layout
870 @cindex layout of output file
871 The most fundamental command of the @code{ld} command language is the
872 @code{SECTIONS} command (@pxref{SECTIONS}). Every meaningful command
873 script must have a @code{SECTIONS} command: it specifies a
874 ``picture'' of the output file's layout, in varying degrees of detail.
875 No other command is required in all cases.
877 The @code{MEMORY} command complements @code{SECTIONS} by describing the
878 available memory in the target architecture. This command is optional;
879 if you don't use a @code{MEMORY} command, @code{ld} assumes sufficient
880 memory is available in a contiguous block for all output.
884 You may include comments in linker scripts just as in C: delimited
885 by @samp{/*} and @samp{*/}. As in C, comments are syntactically
886 equivalent to whitespace.
890 @cindex expression syntax
892 Many useful commands involve arithmetic expressions. The syntax for
893 expressions in the command language is identical to that of C
894 expressions, with the following features:
897 All expressions evaluated as integers and
898 are of ``long'' or ``unsigned long'' type.
900 All constants are integers.
902 All of the C arithmetic operators are provided.
904 You may reference, define, and create global variables.
906 You may call special purpose built-in functions.
910 * Integers:: Integers
911 * Symbols:: Symbol Names
912 * Location Counter:: The Location Counter
913 * Operators:: Operators
914 * Evaluation:: Evaluation
915 * Assignment:: Assignment: Defining Symbols
916 * Arithmetic Functions:: Built-In Functions
921 @cindex integer notation
922 @cindex octal integers
923 An octal integer is @samp{0} followed by zero or more of the octal
924 digits (@samp{01234567}).
929 @cindex decimal integers
930 A decimal integer starts with a non-zero digit followed by zero or
931 more digits (@samp{0123456789}).
936 @cindex hexadecimal integers
938 A hexadecimal integer is @samp{0x} or @samp{0X} followed by one or
939 more hexadecimal digits chosen from @samp{0123456789abcdefABCDEF}.
944 @cindex negative integers
945 To write a negative integer, use
946 the prefix operator @samp{-}; @pxref{Operators}.
951 @cindex scaled integers
952 @cindex K and M integer suffixes
953 @cindex M and K integer suffixes
954 @cindex suffixes for integers
955 @cindex integer suffixes
956 Additionally the suffixes @code{K} and @code{M} may be used to scale a
960 @c END TEXI2ROFF-KILL
961 @code{1024} or @code{1024*1024}
965 ${\rm 1024}$ or ${\rm 1024}^2$
967 @c END TEXI2ROFF-KILL
968 respectively. For example, the following all refer to the same quantity:
977 @subsection Symbol Names
980 @cindex quoted symbol names
982 Unless quoted, symbol names start with a letter, underscore, or point
983 and may include any letters, underscores, digits, points,
984 and hyphens. Unquoted symbol names must not conflict with any
985 keywords. You can specify a symbol which contains odd characters or has
986 the same name as a keyword, by surrounding the symbol name in double quotes:
989 "with a space" = "also with a space" + 10;
992 Since symbols can contain many non-alphabetic characters, it is safest
993 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
994 whereas @samp{A - B} is an expression involving subtraction.
996 @node Location Counter
997 @subsection The Location Counter
1000 @cindex location counter
1001 @cindex current output location
1002 The special linker variable @dfn{dot} @samp{.} always contains the
1003 current output location counter. Since the @code{.} always refers to
1004 a location in an output section, it must always appear in an
1005 expression within a @code{SECTIONS} command. The @code{.} symbol
1006 may appear anywhere that an ordinary symbol is allowed in an
1007 expression, but its assignments have a side effect. Assigning a value
1008 to the @code{.} symbol will cause the location counter to be moved.
1010 This may be used to create holes in the output section. The location
1011 counter may never be moved backwards.
1026 In the previous example, @code{file1} is located at the beginning of the
1027 output section, then there is a 1000 byte gap. Then @code{file2}
1028 appears, also with a 1000 byte gap following before @code{file3} is
1029 loaded. The notation @samp{= 0x1234} specifies what data to write in
1030 the gaps (@pxref{Section Options}).
1038 @subsection Operators
1039 @cindex Operators for arithmetic
1040 @cindex arithmetic operators
1041 @cindex precedence in expressions
1042 The linker recognizes the standard C set of arithmetic operators, with
1043 the standard bindings and precedence levels:
1046 @c END TEXI2ROFF-KILL
1048 precedence associativity Operators Notes
1054 5 left == != > < <= >=
1060 11 right &= += -= *= /= (2)
1064 (1) Prefix operators
1065 (2) @xref{Assignment}
1069 \vskip \baselineskip
1070 %"lispnarrowing" is the extra indent used generally for @example
1071 \hskip\lispnarrowing\vbox{\offinterlineskip
1074 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
1075 height2pt&\omit&&\omit&&\omit&\cr
1076 &Precedence&& Associativity &&{\rm Operators}&\cr
1077 height2pt&\omit&&\omit&&\omit&\cr
1079 height2pt&\omit&&\omit&&\omit&\cr
1081 % '176 is tilde, '~' in tt font
1082 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
1083 &2&&left&&* / \%&\cr
1086 &5&&left&&== != > < <= >=&\cr
1089 &8&&left&&{\&\&}&\cr
1092 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
1094 height2pt&\omit&&\omit&&\omit&\cr}
1099 @obeylines@parskip=0pt@parindent=0pt
1100 @dag@quad Prefix operators.
1101 @ddag@quad @xref{Assignment}.
1104 @c END TEXI2ROFF-KILL
1107 @subsection Evaluation
1109 @cindex lazy evaluation
1110 @cindex expression evaluation order
1111 The linker uses ``lazy evaluation'' for expressions; it only calculates
1112 an expression when absolutely necessary. The linker needs the value of
1113 the start address, and the lengths of memory regions, in order to do any
1114 linking at all; these values are computed as soon as possible when the
1115 linker reads in the command file. However, other values (such as symbol
1116 values) are not known or needed until after storage allocation. Such
1117 values are evaluated later, when other information (such as the sizes of
1118 output sections) is available for use in the symbol assignment
1122 @subsection Assignment: Defining Symbols
1123 @cindex assignment in scripts
1124 @cindex symbol definition, scripts
1125 @cindex variables, defining
1126 You may create global symbols, and assign values (addresses) to global
1127 symbols, using any of the C assignment operators:
1130 @item @var{symbol} = @var{expression} ;
1131 @itemx @var{symbol} &= @var{expression} ;
1132 @itemx @var{symbol} += @var{expression} ;
1133 @itemx @var{symbol} -= @var{expression} ;
1134 @itemx @var{symbol} *= @var{expression} ;
1135 @itemx @var{symbol} /= @var{expression} ;
1138 Two things distinguish assignment from other operators in @code{ld}
1142 Assignment may only be used at the root of an expression;
1143 @samp{a=b+3;} is allowed, but @samp{a+b=3;} is an error.
1148 You must place a trailing semicolon (``@key{;}'') at the end of an
1149 assignment statement.
1152 Assignment statements may appear:
1155 as commands in their own right in an @code{ld} script; or
1157 as independent statements within a @code{SECTIONS} command; or
1159 as part of the contents of a section definition in a
1160 @code{SECTIONS} command.
1163 The first two cases are equivalent in effect---both define a symbol with
1164 an absolute address. The last case defines a symbol whose address is
1165 relative to a particular section (@pxref{SECTIONS}).
1167 @cindex absolute and relocatable symbols
1168 @cindex relocatable and absolute symbols
1169 @cindex symbols, relocatable and absolute
1170 When a linker expression is evaluated and assigned to a variable, it is
1171 given either an absolute or a relocatable type. An absolute expression
1172 type is one in which the symbol contains the value that it will have in
1173 the output file; a relocatable expression type is one in which the
1174 value is expressed as a fixed offset from the base of a section.
1176 The type of the expression is controlled by its position in the script
1177 file. A symbol assigned within a section definition is created relative
1178 to the base of the section; a symbol assigned in any other place is
1179 created as an absolute symbol. Since a symbol created within a
1180 section definition is relative to the base of the section, it
1181 will remain relocatable if relocatable output is requested. A symbol
1182 may be created with an absolute value even when assigned to within a
1183 section definition by using the absolute assignment function
1184 @code{ABSOLUTE}. For example, to create an absolute symbol whose address
1185 is the last byte of an output section named @code{.data}:
1191 _edata = ABSOLUTE(.) ;
1196 The linker tries to put off the evaluation of an assignment until all
1197 the terms in the source expression are known (@pxref{Evaluation}). For
1198 instance, the sizes of sections cannot be known until after allocation,
1199 so assignments dependent upon these are not performed until after
1200 allocation. Some expressions, such as those depending upon the location
1201 counter @dfn{dot}, @samp{.} must be evaluated during allocation. If the
1202 result of an expression is required, but the value is not available,
1203 then an error results. For example, a script like the following
1206 text 9+this_isnt_constant :
1211 @kindex Non constant expression
1213 will cause the error message ``@code{Non constant expression for initial
1216 @node Arithmetic Functions
1217 @subsection Arithmetic Functions
1218 @cindex functions in expression language
1219 The command language includes a number of built-in
1220 functions for use in link script expressions.
1222 @item ABSOLUTE(@var{exp})
1223 @kindex ABSOLUTE(@var{exp})
1224 @cindex expression, absolute
1225 Return the absolute (non-relocatable, as opposed to non-negative) value
1226 of the expression @var{exp}. Primarily useful to assign an absolute
1227 value to a symbol within a section definition, where symbol values are
1228 normally section-relative.
1230 @item ADDR(@var{section})
1231 @kindex ADDR(@var{section})
1232 @cindex section address
1233 Return the absolute address of the named @var{section}. Your script must
1234 previously have defined the location of that section. In the following
1235 example, @code{symbol_1} and @code{symbol_2} are assigned identical
1241 start_of_output_1 = ABSOLUTE(.);
1246 symbol_1 = ADDR(.output1);
1247 symbol_2 = start_of_output_1;
1252 @item ALIGN(@var{exp})
1253 @kindex ALIGN(@var{exp})
1254 @cindex rounding up location counter
1255 Return the result of the current location counter (@code{.}) aligned to
1256 the next @var{exp} boundary. @var{exp} must be an expression whose
1257 value is a power of two. This is equivalent to
1259 (. + @var{exp} - 1) & ~(@var{exp} - 1)
1262 @code{ALIGN} doesn't change the value of the location counter---it just
1263 does arithmetic on it. As an example, to align the output @code{.data}
1264 section to the next @code{0x2000} byte boundary after the preceding
1265 section and to set a variable within the section to the next
1266 @code{0x8000} boundary after the input sections:
1269 .data ALIGN(0x2000): @{
1271 variable = ALIGN(0x8000);
1276 The first use of @code{ALIGN} in this example specifies the location of
1277 a section because it is used as the optional @var{start} attribute of a
1278 section definition (@pxref{Section Options}). The second use simply
1279 defines the value of a variable.
1281 The built-in @code{NEXT} is closely related to @code{ALIGN}.
1283 @item DEFINED(@var{symbol})
1284 @kindex DEFINED(@var{symbol})
1285 @cindex symbol defaults
1286 Return 1 if @var{symbol} is in the linker global symbol table and is
1287 defined, otherwise return 0. You can use this function to provide default
1288 values for symbols. For example, the following command-file fragment shows how
1289 to set a global symbol @code{begin} to the first location in the
1290 @code{.text} section---but if a symbol called @code{begin} already
1291 existed, its value is preserved:
1296 begin = DEFINED(begin) ? begin : . ;
1302 @item NEXT(@var{exp})
1303 @kindex NEXT(@var{exp})
1304 @cindex unallocated address, next
1305 Return the next unallocated address that is a multiple of @var{exp}.
1306 This function is closely related to @code{ALIGN(@var{exp})}; unless you
1307 use the @code{MEMORY} command to define discontinuous memory for the
1308 output file, the two functions are equivalent.
1310 @item SIZEOF(@var{section})
1311 @kindex SIZEOF(@var{section})
1312 @cindex section size
1313 Return the size in bytes of the named @var{section}, if that section has
1314 been allocated. In the following example, @code{symbol_1} and
1315 @code{symbol_2} are assigned identical values:
1316 @c What does it return if the section hasn't been allocated? 0?
1324 symbol_1 = .end - .start ;
1325 symbol_2 = SIZEOF(.output);
1330 @item SIZEOF_HEADERS
1331 @kindex SIZEOF_HEADERS
1333 @itemx sizeof_headers
1334 @kindex sizeof_headers
1335 Return the size in bytes of the output file's headers. You can use this number
1336 as the start address of the first section, if you choose, to facilitate
1342 @section Memory Layout
1344 @cindex regions of memory
1345 @cindex discontinuous memory
1346 @cindex allocating memory
1347 The linker's default configuration permits allocation of all available memory.
1348 You can override this configuration by using the @code{MEMORY} command. The
1349 @code{MEMORY} command describes the location and size of blocks of
1350 memory in the target. By using it carefully, you can describe which
1351 memory regions may be used by the linker, and which memory regions it
1352 must avoid. The linker does not shuffle sections to fit into the
1353 available regions, but does move the requested sections into the correct
1354 regions and issue errors when the regions become too full.
1356 A command file may contain at most one use of the @code{MEMORY}
1357 command; however, you can define as many blocks of memory within it as
1358 you wish. The syntax is:
1363 @var{name} (@var{attr}) : ORIGIN = @var{origin}, LENGTH = @var{len}
1368 @cindex naming memory regions
1370 is a name used internally by the linker to refer to the region. Any
1371 symbol name may be used. The region names are stored in a separate
1372 name space, and will not conflict with symbols, file names or section
1373 names. Use distinct names to specify multiple regions.
1375 @cindex memory region attributes
1377 is an optional list of attributes, permitted for compatibility with the
1378 AT&T linker but not used by @code{ld} beyond checking that the
1379 attribute list is valid. Valid attribute lists must be made up of the
1380 characters ``@code{LIRWX}''. If you omit the attribute list, you may
1381 omit the parentheses around it as well.
1387 is the start address of the region in physical memory. It is
1388 an expression that must evaluate to a constant before
1389 memory allocation is performed. The keyword @code{ORIGIN} may be
1390 abbreviated to @code{org} or @code{o} (but not, for example, @samp{ORG}).
1396 is the size in bytes of the region (an expression).
1397 The keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
1400 For example, to specify that memory has two regions available for
1401 allocation---one starting at 0 for 256 kilobytes, and the other
1402 starting at @code{0x40000000} for four megabytes:
1407 rom : ORIGIN = 0, LENGTH = 256K
1408 ram : org = 0x40000000, l = 4M
1412 Once you have defined a region of memory named @var{mem}, you can direct
1413 specific output sections there by using a command ending in
1414 @samp{>@var{mem}} within the @code{SECTIONS} command (@pxref{Section
1415 Options}). If the combined output sections directed to a region are too
1416 big for the region, the linker will issue an error message.
1419 @section Specifying Output Sections
1422 The @code{SECTIONS} command controls exactly where input sections are
1423 placed into output sections, their order in the output file, and to
1424 which output sections they are allocated.
1426 You may use at most one @code{SECTIONS} command in a script file,
1427 but you can have as many statements within it as you wish. Statements
1428 within the @code{SECTIONS} command can do one of three things:
1432 define the entry point;
1435 assign a value to a symbol;
1438 describe the placement of a named output section, and which input
1439 sections go into it.
1442 You can also use the first two operations---defining the entry point and
1443 defining symbols---outside the @code{SECTIONS} command: @pxref{Entry
1444 Point}, and @pxref{Assignment}. They are permitted here as well for
1445 your convenience in reading the script, so that symbols and the entry
1446 point can be defined at meaningful points in your output-file layout.
1448 If you do not use a @code{SECTIONS} command, the linker places each input
1449 section into an identically named output section in the order that the
1450 sections are first encountered in the input files. If all input sections
1451 are present in the first file, for example, the order of sections in the
1452 output file will match the order in the first input file.
1455 * Section Definition:: Section Definitions
1456 * Section Placement:: Section Placement
1457 * Section Data Expressions:: Section Data Expressions
1458 * Section Options:: Optional Section Attributes
1461 @node Section Definition
1462 @subsection Section Definitions
1463 @cindex section definition
1464 The most frequently used statement in the @code{SECTIONS} command is
1465 the @dfn{section definition}, which specifies the
1466 properties of an output section: its location, alignment, contents,
1467 fill pattern, and target memory region. Most of
1468 these specifications are optional; the simplest form of a section
1477 @cindex naming output sections
1479 @var{secname} is the name of the output section, and @var{contents} a
1480 specification of what goes there---for example, a list of input files or
1481 sections of input files (@pxref{Section Placement}). As you might
1482 assume, the whitespace shown is optional. You do need the colon
1483 @samp{:} and the braces @samp{@{@}}, however.
1485 @var{secname} must meet the constraints of your output format. In
1486 formats which only support a limited number of sections, such as
1487 @code{a.out}, the name must be one of the names supported by the format
1488 (@code{a.out}, for example, allows only @code{.text}, @code{.data} or
1489 @code{.bss}). If the output format supports any number of sections, but
1490 with numbers and not names (as is the case for Oasys), the name should be
1491 supplied as a quoted numeric string. A section name may consist of any
1492 sequence of characters, but any name which does not conform to the standard
1493 @code{ld} symbol name syntax must be quoted.
1494 @xref{Symbols, , Symbol Names}.
1496 The linker will not create output sections which do not have any
1497 contents. This is for convenience when referring to input sections that
1498 may or may not exist. For example,
1502 will only create a @samp{.foo} section in the output file if there is a
1503 @samp{.foo} section in at least one input file.
1505 @node Section Placement
1506 @subsection Section Placement
1508 @cindex contents of a section
1509 In a section definition, you can specify the contents of an output
1510 section by listing particular input files, by listing particular
1511 input-file sections, or by a combination of the two. You can also place
1512 arbitrary data in the section, and define symbols relative to the
1513 beginning of the section.
1515 The @var{contents} of a section definition may include any of the
1516 following kinds of statement. You can include as many of these as you
1517 like in a single section definition, separated from one another by
1521 @kindex @var{filename}
1522 @cindex input files, section defn
1523 @cindex files, including in output sections
1524 @item @var{filename}
1525 You may simply name a particular input file to be placed in the current
1526 output section; @emph{all} sections from that file are placed in the
1527 current section definition. If the file name has already been mentioned
1528 in another section definition, with an explicit section name list, then
1529 only those sections which have not yet been allocated are used.
1531 To specify a list of particular files by name:
1533 .data : @{ afile.o bfile.o cfile.o @}
1536 The example also illustrates that multiple statements can be included in
1537 the contents of a section definition, since each file name is a separate
1540 @kindex @var{filename}(@var{section})
1541 @cindex files and sections, section defn
1542 @item @var{filename}( @var{section} )
1543 @itemx @var{filename}( @var{section}, @var{section}, @dots{} )
1544 @itemx @var{filename}( @var{section} @var{section} @dots{} )
1545 You can name one or more sections from your input files, for
1546 insertion in the current output section. If you wish to specify a list
1547 of input-file sections inside the parentheses, you may separate the
1548 section names by either commas or whitespace.
1550 @cindex input sections to output section
1551 @kindex *(@var{section})
1552 @item * (@var{section})
1553 @itemx * (@var{section}, @var{section}, @dots{})
1554 @itemx * (@var{section} @var{section} @dots{})
1555 Instead of explicitly naming particular input files in a link control
1556 script, you can refer to @emph{all} files from the @code{ld} command
1557 line: use @samp{*} instead of a particular file name before the
1558 parenthesized input-file section list.
1560 If you have already explicitly included some files by name, @samp{*}
1561 refers to all @emph{remaining} files---those whose places in the output
1562 file have not yet been defined.
1564 For example, to copy sections @code{1} through @code{4} from an Oasys file
1565 into the @code{.text} section of an @code{a.out} file, and sections @code{13}
1566 and @code{14} into the @code{.data} section:
1579 @cindex @code{[@var{section}@dots{}]}, not supported
1580 @samp{[ @var{section} @dots{} ]} used to be accepted as an alternate way
1581 to specify named sections from all unallocated input files. Because
1582 some operating systems (VMS) allow brackets in file names, that notation
1583 is no longer supported.
1585 @cindex uninitialized data
1586 @cindex commons in output
1588 @item @var{filename}@code{( COMMON )}
1590 Specify where in your output file to place uninitialized data
1591 with this notation. @code{*(COMMON)} by itself refers to all
1592 uninitialized data from all input files (so far as it is not yet
1593 allocated); @var{filename}@code{(COMMON)} refers to uninitialized data
1594 from a particular file. Both are special cases of the general
1595 mechanisms for specifying where to place input-file sections:
1596 @code{ld} permits you to refer to uninitialized data as if it
1597 were in an input-file section named @code{COMMON}, regardless of the
1598 input file's format.
1601 For example, the following command script arranges the output file into
1602 three consecutive sections, named @code{.text}, @code{.data}, and
1603 @code{.bss}, taking the input for each from the correspondingly named
1604 sections of all the input files:
1608 .text : @{ *(.text) @}
1609 .data : @{ *(.data) @}
1610 .bss : @{ *(.bss) *(COMMON) @}
1614 The following example reads all of the sections from file @code{all.o}
1615 and places them at the start of output section @code{outputa} which
1616 starts at location @code{0x10000}. All of section @code{.input1} from
1617 file @code{foo.o} follows immediately, in the same output section. All
1618 of section @code{.input2} from @code{foo.o} goes into output section
1619 @code{outputb}, followed by section @code{.input1} from @code{foo1.o}.
1620 All of the remaining @code{.input1} and @code{.input2} sections from any
1621 files are written to output section @code{outputc}.
1643 @node Section Data Expressions
1644 @subsection Section Data Expressions
1646 @cindex expressions in a section
1647 The foregoing statements arrange, in your output file, data originating
1648 from your input files. You can also place data directly in an output
1649 section from the link command script. Most of these additional
1650 statements involve expressions; @pxref{Expressions}. Although these
1651 statements are shown separately here for ease of presentation, no such
1652 segregation is needed within a section definition in the @code{SECTIONS}
1653 command; you can intermix them freely with any of the statements we've
1657 @cindex input filename symbols
1658 @cindex filename symbols
1659 @kindex CREATE_OBJECT_SYMBOLS
1660 @item CREATE_OBJECT_SYMBOLS
1661 Create a symbol for each input file
1662 in the current section, set to the address of the first byte of
1663 data written from that input file. For instance, with @code{a.out}
1664 files it is conventional to have a symbol for each input file. You can
1665 accomplish this by defining the output @code{.text} section as follows:
1670 CREATE_OBJECT_SYMBOLS
1672 _etext = ALIGN(0x2000);
1678 If @code{sample.ld} is a file containing this script, and @code{a.o},
1679 @code{b.o}, @code{c.o}, and @code{d.o} are four input files with
1680 contents like the following---
1690 @samp{ld -M -T sample.ld a.o b.o c.o d.o} would create a map like this,
1691 containing symbols matching the object file names:
1693 00000000 A __DYNAMIC
1696 00002020 T _afunction
1699 00002038 T _bfunction
1702 00002050 T _cfunction
1705 00002068 T _dfunction
1715 @kindex @var{symbol} = @var{expression} ;
1716 @kindex @var{symbol} @var{f}= @var{expression} ;
1717 @item @var{symbol} = @var{expression} ;
1718 @itemx @var{symbol} @var{f}= @var{expression} ;
1719 @var{symbol} is any symbol name (@pxref{Symbols}). ``@var{f}=''
1720 refers to any of the operators @code{&= += -= *= /=} which combine
1721 arithmetic and assignment.
1723 @cindex assignment, in section defn
1724 When you assign a value to a symbol within a particular section
1725 definition, the value is relative to the beginning of the section
1726 (@pxref{Assignment}). If you write
1732 .data : @{ @dots{} rel = 14 ; @dots{} @}
1733 abs2 = 14 + ADDR(.data);
1738 @c FIXME: Try above example!
1740 @code{abs} and @code{rel} do not have the same value; @code{rel} has the
1741 same value as @code{abs2}.
1743 @kindex BYTE(@var{expression})
1744 @kindex SHORT(@var{expression})
1745 @kindex LONG(@var{expression})
1746 @kindex QUAD(@var{expression})
1747 @cindex direct output
1748 @item BYTE(@var{expression})
1749 @itemx SHORT(@var{expression})
1750 @itemx LONG(@var{expression})
1751 @itemx QUAD(@var{expression})
1752 By including one of these four statements in a section definition, you
1753 can explicitly place one, two, four, or eight bytes (respectively) at
1754 the current address of that section. @code{QUAD} is only supported when
1755 using a 64 bit host or target.
1757 @ifclear SingleFormat
1758 Multiple-byte quantities are represented in whatever byte order is
1759 appropriate for the output file format (@pxref{BFD}).
1762 @item FILL(@var{expression})
1763 @kindex FILL(@var{expression})
1764 @cindex holes, filling
1765 @cindex unspecified memory
1766 Specify the ``fill pattern'' for the current section. Any otherwise
1767 unspecified regions of memory within the section (for example, regions
1768 you skip over by assigning a new value to the location counter @samp{.})
1769 are filled with the two least significant bytes from the
1770 @var{expression} argument. A @code{FILL} statement covers memory
1771 locations @emph{after} the point it occurs in the section definition; by
1772 including more than one @code{FILL} statement, you can have different
1773 fill patterns in different parts of an output section.
1776 @node Section Options
1777 @subsection Optional Section Attributes
1778 @cindex section defn, full syntax
1779 Here is the full syntax of a section definition, including all the
1785 @var{secname} @var{start} BLOCK(@var{align}) (NOLOAD) : AT ( @var{ldadr} )
1786 @{ @var{contents} @} >@var{region} =@var{fill}
1791 @var{secname} and @var{contents} are required. @xref{Section
1792 Definition}, and @pxref{Section Placement} for details on
1793 @var{contents}. The remaining elements---@var{start},
1794 @code{BLOCK(@var{align)}}, @code{(NOLOAD)}, @code{AT ( @var{ldadr} )},
1795 @code{>@var{region}}, and @code{=@var{fill}}---are all optional.
1798 @cindex start address, section
1799 @cindex section start
1800 @cindex section address
1802 You can force the output section to be loaded at a specified address by
1803 specifying @var{start} immediately following the section name.
1804 @var{start} can be represented as any expression. The following
1805 example generates section @var{output} at location
1811 output 0x40000000: @{
1818 @kindex BLOCK(@var{align})
1819 @cindex section alignment
1820 @cindex aligning sections
1821 @item BLOCK(@var{align})
1822 You can include @code{BLOCK()} specification to advance
1823 the location counter @code{.} prior to the beginning of the section, so
1824 that the section will begin at the specified alignment. @var{align} is
1828 @cindex prevent unnecessary loading
1829 @cindex loading, preventing
1831 Use @samp{(NOLOAD)} to prevent a section from being loaded into memory
1832 each time it is accessed. For example, in the script sample below, the
1833 @code{ROM} segment is addressed at memory location @samp{0} and does not
1834 need to be loaded into each object file:
1838 ROM 0 (NOLOAD) : @{ @dots{} @}
1843 @kindex AT ( @var{ldadr} )
1844 @cindex specify load address
1845 @cindex load address, specifying
1846 @item AT ( @var{ldadr} )
1847 The expression @var{ldadr} that follows the @code{AT} keyword specifies
1848 the load address of the section. The default (if you do not use the
1849 @code{AT} keyword) is to make the load address the same as the
1850 relocation address. This feature is designed to make it easy to build a
1851 ROM image. For example, this @code{SECTIONS} definition creates two
1852 output sections: one called @samp{.text}, which starts at @code{0x1000},
1853 and one called @samp{.mdata}, which is loaded at the end of the
1854 @samp{.text} section even though its relocation address is
1855 @code{0x2000}. The symbol @code{_data} is defined with the value
1861 .text 0x1000 : @{ *(.text) _etext = . ; @}
1863 AT ( ADDR(.text) + SIZEOF ( .text ) )
1864 @{ _data = . ; *(.data); _edata = . ; @}
1866 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
1870 The run-time initialization code (for C programs, usually @code{crt0})
1871 for use with a ROM generated this way has to include something like
1872 the following, to copy the initialized data from the ROM image to its runtime
1879 /* ROM has data at end of text; copy it. */
1880 while (dst < _edata) @{
1885 for (dst = _bstart; dst< _bend; dst++)
1889 @kindex >@var{region}
1890 @cindex section, assigning to memory region
1891 @cindex memory regions and sections
1893 Assign this section to a previously defined region of memory.
1897 @cindex section fill pattern
1898 @cindex fill pattern, entire section
1900 Including @code{=@var{fill}} in a section definition specifies the
1901 initial fill value for that section. You may use any expression to
1902 specify @var{fill}. Any unallocated holes in the current output section
1903 when written to the output file will be filled with the two least
1904 significant bytes of the value, repeated as necessary. You can also
1905 change the fill value with a @code{FILL} statement in the @var{contents}
1906 of a section definition.
1911 @section The Entry Point
1912 @kindex ENTRY(@var{symbol})
1913 @cindex start of execution
1914 @cindex first instruction
1915 The linker command language includes a command specifically for
1916 defining the first executable instruction in an output file (its
1917 @dfn{entry point}). Its argument is a symbol name:
1922 Like symbol assignments, the @code{ENTRY} command may be placed either
1923 as an independent command in the command file, or among the section
1924 definitions within the @code{SECTIONS} command---whatever makes the most
1925 sense for your layout.
1927 @cindex entry point, defaults
1928 @code{ENTRY} is only one of several ways of choosing the entry point.
1929 You may indicate it in any of the following ways (shown in descending
1930 order of priority: methods higher in the list override methods lower down).
1933 the @samp{-e} @var{entry} command-line option;
1935 the @code{ENTRY(@var{symbol})} command in a linker control script;
1937 the value of the symbol @code{start}, if present;
1939 the value of the symbol @code{_main}, if present;
1941 the address of the first byte of the @code{.text} section, if present;
1943 The address @code{0}.
1946 For example, you can use these rules to generate an entry point with an
1947 assignment statement: if no symbol @code{start} is defined within your
1948 input files, you can simply define it, assigning it an appropriate
1956 The example shows an absolute address, but you can use any expression.
1957 For example, if your input object files use some other symbol-name
1958 convention for the entry point, you can just assign the value of
1959 whatever symbol contains the start address to @code{start}:
1962 start = other_symbol ;
1965 @node Option Commands
1966 @section Option Commands
1967 The command language includes a number of other commands that you can
1968 use for specialized purposes. They are similar in purpose to
1969 command-line options.
1972 @kindex CONSTRUCTORS
1973 @cindex C++ constructors, arranging in link
1974 @cindex constructors, arranging in link
1976 This command ties up C++ style constructor and destructor records. The
1977 details of the constructor representation vary from one object format to
1978 another, but usually lists of constructors and destructors appear as
1979 special sections. The @code{CONSTRUCTORS} command specifies where the
1980 linker is to place the data from these sections, relative to the rest of
1981 the linked output. Constructor data is marked by the symbol
1982 @w{@code{__CTOR_LIST__}} at the start, and @w{@code{__CTOR_LIST_END}} at
1983 the end; destructor data is bracketed similarly, between
1984 @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_LIST_END}}. (The compiler
1985 must arrange to actually run this code; GNU C++ calls constructors from
1986 a subroutine @code{__main}, which it inserts automatically into the
1987 startup code for @code{main}, and destructors from @code{_exit}.)
1994 These keywords were used in some older linkers to request a particular
1995 math subroutine library. @code{ld} doesn't use the keywords, assuming
1996 instead that any necessary subroutines are in libraries specified using
1997 the general mechanisms for linking to archives; but to permit the use of
1998 scripts that were written for the older linkers, the keywords
1999 @code{FLOAT} and @code{NOFLOAT} are accepted and ignored.
2001 @kindex FORCE_COMMON_ALLOCATION
2002 @cindex common allocation
2003 @item FORCE_COMMON_ALLOCATION
2004 This command has the same effect as the @samp{-d} command-line option:
2005 to make @code{ld} assign space to common symbols even if a relocatable
2006 output file is specified (@samp{-r}).
2008 @kindex INPUT ( @var{files} )
2009 @cindex binary input files
2010 @item INPUT ( @var{file}, @var{file}, @dots{} )
2011 @itemx INPUT ( @var{file} @var{file} @dots{} )
2012 Use this command to include binary input files in the link, without
2013 including them in a particular section definition.
2014 Specify the full name for each @var{file}, including @samp{.a} if
2017 @code{ld} searches for each @var{file} through the archive-library
2018 search path, just as for files you specify on the command line.
2019 See the description of @samp{-L} in @ref{Options,,Command Line
2023 @item MAP ( @var{name} )
2024 @kindex MAP ( @var{name} )
2025 @c MAP(...) appears to look for an F in the arg, ignoring all other
2026 @c chars; if it finds one, it sets "map_option_f" to true. But nothing
2027 @c checks map_option_f. Apparently a stub for the future...
2030 @item OUTPUT ( @var{filename} )
2031 @kindex OUTPUT ( @var{filename} )
2032 @cindex naming the output file
2033 Use this command to name the link output file @var{filename}. The
2034 effect of @code{OUTPUT(@var{filename})} is identical to the effect of
2035 @w{@samp{-o @var{filename}}}, which overrides it. You can use this
2036 command to supply a default output-file name other than @code{a.out}.
2038 @ifclear SingleFormat
2039 @item OUTPUT_ARCH ( @var{bfdname} )
2040 @kindex OUTPUT_ARCH ( @var{bfdname} )
2041 @cindex machine architecture, output
2042 Specify a particular output machine architecture, with one of the names
2043 used by the BFD back-end routines (@pxref{BFD}). This command is often
2044 unnecessary; the architecture is most often set implicitly by either the
2045 system BFD configuration or as a side effect of the @code{OUTPUT_FORMAT}
2048 @item OUTPUT_FORMAT ( @var{bfdname} )
2049 @kindex OUTPUT_FORMAT ( @var{bfdname} )
2050 @cindex format, output file
2051 When @code{ld} is configured to support multiple object code formats,
2052 you can use this command to specify a particular output format.
2053 @var{bfdname} is one of the names used by the BFD back-end routines
2054 (@pxref{BFD}). The effect is identical to the effect of the
2055 @samp{-oformat} command-line option. This selection affects only
2056 the output file; the related command @code{TARGET} affects primarily
2060 @item SEARCH_DIR ( @var{path} )
2061 @kindex SEARCH_DIR ( @var{path} )
2062 @cindex path for libraries
2063 @cindex search path, libraries
2064 Add @var{path} to the list of paths where @code{ld} looks for
2065 archive libraries. @code{SEARCH_DIR(@var{path})} has the same
2066 effect as @samp{-L@var{path}} on the command line.
2068 @item STARTUP ( @var{filename} )
2069 @kindex STARTUP ( @var{filename} )
2070 @cindex first input file
2071 Ensure that @var{filename} is the first input file used in the link
2074 @ifclear SingleFormat
2075 @item TARGET ( @var{format} )
2076 @cindex input file format
2077 @kindex TARGET ( @var{format} )
2078 When @code{ld} is configured to support multiple object code formats,
2079 you can use this command to change the input-file object code format
2080 (like the command-line option @samp{-b} or its synonym @samp{-format}).
2081 The argument @var{format} is one of the strings used by BFD to name
2082 binary formats. If @code{TARGET} is specified but @code{OUTPUT_FORMAT}
2083 is not, the last @code{TARGET} argument is also used as the default
2084 format for the @code{ld} output file. @xref{BFD}.
2087 If you don't use the @code{TARGET} command, @code{ld} uses the value of
2088 the environment variable @code{GNUTARGET}, if available, to select the
2089 output file format. If that variable is also absent, @code{ld} uses
2090 the default format configured for your machine in the BFD libraries.
2095 @node Machine Dependent
2096 @chapter Machine Dependent Features
2098 @cindex machine dependencies
2099 @code{ld} has additional features on some platforms; the following
2100 sections describe them. Machines where @code{ld} has no additional
2101 functionality are not listed.
2104 * H8/300:: @code{ld} and the H8/300
2105 * i960:: @code{ld} and the Intel 960 family
2109 @c FIXME! This could use @raisesections/@lowersections, but there seems to be a conflict
2110 @c between those and node-defaulting.
2116 @section @code{ld} and the H8/300
2118 @cindex H8/300 support
2119 For the H8/300, @code{ld} can perform these global optimizations when
2120 you specify the @samp{-relax} command-line option.
2123 @item relaxing address modes
2124 @cindex relaxing on H8/300
2125 @code{ld} finds all @code{jsr} and @code{jmp} instructions whose
2126 targets are within eight bits, and turns them into eight-bit
2127 program-counter relative @code{bsr} and @code{bra} instructions,
2130 @item synthesizing instructions
2131 @cindex synthesizing on H8/300
2132 @c FIXME: specifically mov.b, or any mov instructions really?
2133 @code{ld} finds all @code{mov.b} instructions which use the
2134 sixteen-bit absolute address form, but refer to the top
2135 page of memory, and changes them to use the eight-bit address form.
2136 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
2137 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
2138 top page of memory).
2147 @c This stuff is pointless to say unless you're especially concerned
2148 @c with Hitachi chips; don't enable it for generic case, please.
2150 @chapter @code{ld} and other Hitachi chips
2152 @code{ld} also supports the H8/300H, the H8/500, and the Hitachi SH. No
2153 special features, commands, or command-line options are required for
2163 @section @code{ld} and the Intel 960 family
2165 @cindex i960 support
2167 You can use the @samp{-A@var{architecture}} command line option to
2168 specify one of the two-letter names identifying members of the 960
2169 family; the option specifies the desired output target, and warns of any
2170 incompatible instructions in the input files. It also modifies the
2171 linker's search strategy for archive libraries, to support the use of
2172 libraries specific to each particular architecture, by including in the
2173 search loop names suffixed with the string identifying the architecture.
2175 For example, if your @code{ld} command line included @w{@samp{-ACA}} as
2176 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
2177 paths, and in any paths you specify with @samp{-L}) for a library with
2188 The first two possibilities would be considered in any event; the last
2189 two are due to the use of @w{@samp{-ACA}}.
2191 You can meaningfully use @samp{-A} more than once on a command line, since
2192 the 960 architecture family allows combination of target architectures; each
2193 use will add another pair of name variants to search for when @w{@samp{-l}}
2194 specifies a library.
2196 @cindex @code{-relax} on i960
2197 @cindex relaxing on i960
2198 @code{ld} supports the @samp{-relax} option for the i960 family. If you
2199 specify @samp{-relax}, @code{ld} finds all @code{balx} and @code{calx}
2200 instructions whose targets are within 24 bits, and turns them into
2201 24-bit program-counter relative @code{bal} and @code{cal}
2202 instructions, respectively. @code{ld} also turns @code{cal}
2203 instructions into @code{bal} instructions when it determines that the
2204 target subroutine is a leaf routine (that is, the target subroutine does
2205 not itself call any subroutines).
2212 @ifclear SingleFormat
2217 @cindex object file management
2218 @cindex object formats available
2220 The linker accesses object and archive files using the BFD libraries.
2221 These libraries allow the linker to use the same routines to operate on
2222 object files whatever the object file format. A different object file
2223 format can be supported simply by creating a new BFD back end and adding
2224 it to the library. To conserve runtime memory, however, the linker and
2225 associated tools are usually configured to support only a subset of the
2226 object file formats available. You can use @code{objdump -i}
2227 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
2228 list all the formats available for your configuration.
2230 @cindex BFD requirements
2231 @cindex requirements for BFD
2232 As with most implementations, BFD is a compromise between
2233 several conflicting requirements. The major factor influencing
2234 BFD design was efficiency: any time used converting between
2235 formats is time which would not have been spent had BFD not
2236 been involved. This is partly offset by abstraction payback; since
2237 BFD simplifies applications and back ends, more time and care
2238 may be spent optimizing algorithms for a greater speed.
2240 One minor artifact of the BFD solution which you should bear in
2241 mind is the potential for information loss. There are two places where
2242 useful information can be lost using the BFD mechanism: during
2243 conversion and during output. @xref{BFD information loss}.
2246 * BFD outline:: How it works: an outline of BFD
2250 @section How it works: an outline of BFD
2251 @cindex opening object files
2252 @include bfdsumm.texi
2256 @appendix MRI Compatible Script Files
2257 @cindex MRI compatibility
2258 To aid users making the transition to @sc{gnu} @code{ld} from the MRI
2259 linker, @code{ld} can use MRI compatible linker scripts as an
2260 alternative to the more general-purpose linker scripting language
2261 described in @ref{Commands,,Command Language}. MRI compatible linker
2262 scripts have a much simpler command set than the scripting language
2263 otherwise used with @code{ld}. @sc{gnu} @code{ld} supports the most
2264 commonly used MRI linker commands; these commands are described here.
2266 In general, MRI scripts aren't of much use with the @code{a.out} object
2267 file format, since it only has three sections and MRI scripts lack some
2268 features to make use of them.
2270 You can specify a file containing an MRI-compatible script using the
2271 @samp{-c} command-line option.
2273 Each command in an MRI-compatible script occupies its own line; each
2274 command line starts with the keyword that identifies the command (though
2275 blank lines are also allowed for punctuation). If a line of an
2276 MRI-compatible script begins with an unrecognized keyword, @code{ld}
2277 issues a warning message, but continues processing the script.
2279 Lines beginning with @samp{*} are comments.
2281 You can write these commands using all upper-case letters, or all
2282 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
2283 The following list shows only the upper-case form of each command.
2286 @cindex @code{ABSOLUTE} (MRI)
2287 @item ABSOLUTE @var{secname}
2288 @item ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
2289 Normally, @code{ld} includes in the output file all sections from all
2290 the input files. However, in an MRI-compatible script, you can use the
2291 @code{ABSOLUTE} command to restrict the sections that will be present in
2292 your output program. If the @code{ABSOLUTE} command is used at all in a
2293 script, then only the sections named explicitly in @code{ABSOLUTE}
2294 commands will appear in the linker output. You can still use other
2295 input sections (whatever you select on the command line, or using
2296 @code{LOAD}) to resolve addresses in the output file.
2298 @cindex @code{ALIAS} (MRI)
2299 @item ALIAS @var{out-secname}, @var{in-secname}
2300 Use this command to place the data from input section @var{in-secname}
2301 in a section called @var{out-secname} in the linker output file.
2303 @var{in-secname} may be an integer.
2305 @cindex @code{BASE} (MRI)
2306 @item BASE @var{expression}
2307 Use the value of @var{expression} as the lowest address (other than
2308 absolute addresses) in the output file.
2310 @cindex @code{CHIP} (MRI)
2311 @item CHIP @var{expression}
2312 @itemx CHIP @var{expression}, @var{expression}
2313 This command does nothing; it is accepted only for compatibility.
2315 @cindex @code{END} (MRI)
2317 This command does nothing whatever; it's only accepted for compatibility.
2319 @cindex @code{FORMAT} (MRI)
2320 @item FORMAT @var{output-format}
2321 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
2322 language, but restricted to one of these output formats:
2326 S-records, if @var{output-format} is @samp{S}
2329 IEEE, if @var{output-format} is @samp{IEEE}
2332 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
2336 @cindex @code{LIST} (MRI)
2337 @item LIST @var{anything}@dots{}
2338 Print (to the standard output file) a link map, as produced by the
2339 @code{ld} command-line option @samp{-M}.
2341 The keyword @code{LIST} may be followed by anything on the
2342 same line, with no change in its effect.
2344 @cindex @code{LOAD} (MRI)
2345 @item LOAD @var{filename}
2346 @item LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
2347 Include one or more object file @var{filename} in the link; this has the
2348 same effect as specifying @var{filename} directly on the @code{ld}
2351 @cindex @code{NAME} (MRI)
2352 @item NAME @var{output-name}
2353 @var{output-name} is the name for the program produced by @code{ld}; the
2354 MRI-compatible command @code{NAME} is equivalent to the command-line
2355 option @samp{-o} or the general script language command @code{OUTPUT}.
2357 @cindex @code{ORDER} (MRI)
2358 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
2359 @itemx ORDER @var{secname} @var{secname} @var{secname}
2360 Normally, @code{ld} orders the sections in its output file in the
2361 order in which they first appear in the input files. In an MRI-compatible
2362 script, you can override this ordering with the @code{ORDER} command. The
2363 sections you list with @code{ORDER} will appear first in your output
2364 file, in the order specified.
2366 @cindex @code{PUBLIC} (MRI)
2367 @item PUBLIC @var{name}=@var{expression}
2368 @itemx PUBLIC @var{name},@var{expression}
2369 @itemx PUBLIC @var{name} @var{expression}
2370 Supply a value (@var{expression}) for external symbol
2371 @var{name} used in the linker input files.
2373 @cindex @code{SECT} (MRI)
2374 @item SECT @var{secname}, @var{expression}
2375 @itemx SECT @var{secname}=@var{expression}
2376 @itemx SECT @var{secname} @var{expression}
2377 You can use any of these three forms of the @code{SECT} command to
2378 specify the start address (@var{expression}) for section @var{secname}.
2379 If you have more than one @code{SECT} statement for the same
2380 @var{secname}, only the @emph{first} sets the start address.
2390 % I think something like @colophon should be in texinfo. In the
2392 \long\def\colophon{\hbox to0pt{}\vfill
2393 \centerline{The body of this manual is set in}
2394 \centerline{\fontname\tenrm,}
2395 \centerline{with headings in {\bf\fontname\tenbf}}
2396 \centerline{and examples in {\tt\fontname\tentt}.}
2397 \centerline{{\it\fontname\tenit\/} and}
2398 \centerline{{\sl\fontname\tensl\/}}
2399 \centerline{are used for emphasis.}\vfill}
2401 % Blame: pesch@cygnus.com, 28mar91.