3 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
4 @c 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
6 @include configdoc.texi
7 @c (configdoc.texi is generated by the Makefile)
12 @macro gcctabopt{body}
18 @c Configure for the generation of man pages
55 * Ld: (ld). The GNU linker.
61 This file documents the @sc{gnu} linker LD version @value{VERSION}.
63 Copyright (C) 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000,
64 2001, 2002, 2003, 2004 Free Software Foundation, Inc.
68 Permission is granted to copy, distribute and/or modify this document
69 under the terms of the GNU Free Documentation License, Version 1.1
70 or any later version published by the Free Software Foundation;
71 with no Invariant Sections, with no Front-Cover Texts, and with no
72 Back-Cover Texts. A copy of the license is included in the
73 section entitled ``GNU Free Documentation License''.
75 Permission is granted to process this file through Tex and print the
76 results, provided the printed document carries copying permission
77 notice identical to this one except for the removal of this paragraph
78 (this paragraph not being relevant to the printed manual).
84 @setchapternewpage odd
85 @settitle Using LD, the GNU linker
88 @subtitle The GNU linker
90 @subtitle @code{ld} version 2
91 @subtitle Version @value{VERSION}
92 @author Steve Chamberlain
93 @author Ian Lance Taylor
98 \hfill Red Hat Inc\par
99 \hfill nickc\@credhat.com, doc\@redhat.com\par
100 \hfill {\it Using LD, the GNU linker}\par
101 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
103 \global\parindent=0pt % Steve likes it this way.
106 @vskip 0pt plus 1filll
107 @c man begin COPYRIGHT
108 Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000, 2001,
109 2002, 2003, 2004 Free Software Foundation, Inc.
111 Permission is granted to copy, distribute and/or modify this document
112 under the terms of the GNU Free Documentation License, Version 1.1
113 or any later version published by the Free Software Foundation;
114 with no Invariant Sections, with no Front-Cover Texts, and with no
115 Back-Cover Texts. A copy of the license is included in the
116 section entitled ``GNU Free Documentation License''.
121 @c FIXME: Talk about importance of *order* of args, cmds to linker!
126 This file documents the @sc{gnu} linker ld version @value{VERSION}.
128 This document is distributed under the terms of the GNU Free
129 Documentation License. A copy of the license is included in the
130 section entitled ``GNU Free Documentation License''.
133 * Overview:: Overview
134 * Invocation:: Invocation
135 * Scripts:: Linker Scripts
137 * Machine Dependent:: Machine Dependent Features
141 * H8/300:: ld and the H8/300
144 * Renesas:: ld and other Renesas micros
147 * i960:: ld and the Intel 960 family
150 * ARM:: ld and the ARM family
153 * HPPA ELF32:: ld and HPPA 32-bit ELF
156 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
159 * TI COFF:: ld and the TI COFF
162 * Win32:: ld and WIN32 (cygwin/mingw)
165 * Xtensa:: ld and Xtensa Processors
168 @ifclear SingleFormat
171 @c Following blank line required for remaining bug in makeinfo conds/menus
173 * Reporting Bugs:: Reporting Bugs
174 * MRI:: MRI Compatible Script Files
175 * GNU Free Documentation License:: GNU Free Documentation License
183 @cindex @sc{gnu} linker
184 @cindex what is this?
187 @c man begin SYNOPSIS
188 ld [@b{options}] @var{objfile} @dots{}
192 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
193 the Info entries for @file{binutils} and
198 @c man begin DESCRIPTION
200 @command{ld} combines a number of object and archive files, relocates
201 their data and ties up symbol references. Usually the last step in
202 compiling a program is to run @command{ld}.
204 @command{ld} accepts Linker Command Language files written in
205 a superset of AT&T's Link Editor Command Language syntax,
206 to provide explicit and total control over the linking process.
210 This man page does not describe the command language; see the
211 @command{ld} entry in @code{info}, or the manual
212 ld: the GNU linker, for full details on the command language and
213 on other aspects of the GNU linker.
216 @ifclear SingleFormat
217 This version of @command{ld} uses the general purpose BFD libraries
218 to operate on object files. This allows @command{ld} to read, combine, and
219 write object files in many different formats---for example, COFF or
220 @code{a.out}. Different formats may be linked together to produce any
221 available kind of object file. @xref{BFD}, for more information.
224 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
225 linkers in providing diagnostic information. Many linkers abandon
226 execution immediately upon encountering an error; whenever possible,
227 @command{ld} continues executing, allowing you to identify other errors
228 (or, in some cases, to get an output file in spite of the error).
235 @c man begin DESCRIPTION
237 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
238 and to be as compatible as possible with other linkers. As a result,
239 you have many choices to control its behavior.
245 * Options:: Command Line Options
246 * Environment:: Environment Variables
250 @section Command Line Options
258 The linker supports a plethora of command-line options, but in actual
259 practice few of them are used in any particular context.
260 @cindex standard Unix system
261 For instance, a frequent use of @command{ld} is to link standard Unix
262 object files on a standard, supported Unix system. On such a system, to
263 link a file @code{hello.o}:
266 ld -o @var{output} /lib/crt0.o hello.o -lc
269 This tells @command{ld} to produce a file called @var{output} as the
270 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
271 the library @code{libc.a}, which will come from the standard search
272 directories. (See the discussion of the @samp{-l} option below.)
274 Some of the command-line options to @command{ld} may be specified at any
275 point in the command line. However, options which refer to files, such
276 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
277 which the option appears in the command line, relative to the object
278 files and other file options. Repeating non-file options with a
279 different argument will either have no further effect, or override prior
280 occurrences (those further to the left on the command line) of that
281 option. Options which may be meaningfully specified more than once are
282 noted in the descriptions below.
285 Non-option arguments are object files or archives which are to be linked
286 together. They may follow, precede, or be mixed in with command-line
287 options, except that an object file argument may not be placed between
288 an option and its argument.
290 Usually the linker is invoked with at least one object file, but you can
291 specify other forms of binary input files using @samp{-l}, @samp{-R},
292 and the script command language. If @emph{no} binary input files at all
293 are specified, the linker does not produce any output, and issues the
294 message @samp{No input files}.
296 If the linker cannot recognize the format of an object file, it will
297 assume that it is a linker script. A script specified in this way
298 augments the main linker script used for the link (either the default
299 linker script or the one specified by using @samp{-T}). This feature
300 permits the linker to link against a file which appears to be an object
301 or an archive, but actually merely defines some symbol values, or uses
302 @code{INPUT} or @code{GROUP} to load other objects. Note that
303 specifying a script in this way merely augments the main linker script;
304 use the @samp{-T} option to replace the default linker script entirely.
307 For options whose names are a single letter,
308 option arguments must either follow the option letter without intervening
309 whitespace, or be given as separate arguments immediately following the
310 option that requires them.
312 For options whose names are multiple letters, either one dash or two can
313 precede the option name; for example, @samp{-trace-symbol} and
314 @samp{--trace-symbol} are equivalent. Note---there is one exception to
315 this rule. Multiple letter options that start with a lower case 'o' can
316 only be preceeded by two dashes. This is to reduce confusion with the
317 @samp{-o} option. So for example @samp{-omagic} sets the output file
318 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
321 Arguments to multiple-letter options must either be separated from the
322 option name by an equals sign, or be given as separate arguments
323 immediately following the option that requires them. For example,
324 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
325 Unique abbreviations of the names of multiple-letter options are
328 Note---if the linker is being invoked indirectly, via a compiler driver
329 (e.g. @samp{gcc}) then all the linker command line options should be
330 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
331 compiler driver) like this:
334 gcc -Wl,--startgroup foo.o bar.o -Wl,--endgroup
337 This is important, because otherwise the compiler driver program may
338 silently drop the linker options, resulting in a bad link.
340 Here is a table of the generic command line switches accepted by the GNU
344 @kindex -a@var{keyword}
345 @item -a@var{keyword}
346 This option is supported for HP/UX compatibility. The @var{keyword}
347 argument must be one of the strings @samp{archive}, @samp{shared}, or
348 @samp{default}. @samp{-aarchive} is functionally equivalent to
349 @samp{-Bstatic}, and the other two keywords are functionally equivalent
350 to @samp{-Bdynamic}. This option may be used any number of times.
353 @cindex architectures
355 @item -A@var{architecture}
356 @kindex --architecture=@var{arch}
357 @itemx --architecture=@var{architecture}
358 In the current release of @command{ld}, this option is useful only for the
359 Intel 960 family of architectures. In that @command{ld} configuration, the
360 @var{architecture} argument identifies the particular architecture in
361 the 960 family, enabling some safeguards and modifying the
362 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
363 family}, for details.
365 Future releases of @command{ld} may support similar functionality for
366 other architecture families.
369 @ifclear SingleFormat
370 @cindex binary input format
371 @kindex -b @var{format}
372 @kindex --format=@var{format}
375 @item -b @var{input-format}
376 @itemx --format=@var{input-format}
377 @command{ld} may be configured to support more than one kind of object
378 file. If your @command{ld} is configured this way, you can use the
379 @samp{-b} option to specify the binary format for input object files
380 that follow this option on the command line. Even when @command{ld} is
381 configured to support alternative object formats, you don't usually need
382 to specify this, as @command{ld} should be configured to expect as a
383 default input format the most usual format on each machine.
384 @var{input-format} is a text string, the name of a particular format
385 supported by the BFD libraries. (You can list the available binary
386 formats with @samp{objdump -i}.)
389 You may want to use this option if you are linking files with an unusual
390 binary format. You can also use @samp{-b} to switch formats explicitly (when
391 linking object files of different formats), by including
392 @samp{-b @var{input-format}} before each group of object files in a
395 The default format is taken from the environment variable
400 You can also define the input format from a script, using the command
403 see @ref{Format Commands}.
407 @kindex -c @var{MRI-cmdfile}
408 @kindex --mri-script=@var{MRI-cmdfile}
409 @cindex compatibility, MRI
410 @item -c @var{MRI-commandfile}
411 @itemx --mri-script=@var{MRI-commandfile}
412 For compatibility with linkers produced by MRI, @command{ld} accepts script
413 files written in an alternate, restricted command language, described in
415 @ref{MRI,,MRI Compatible Script Files}.
418 the MRI Compatible Script Files section of GNU ld documentation.
420 Introduce MRI script files with
421 the option @samp{-c}; use the @samp{-T} option to run linker
422 scripts written in the general-purpose @command{ld} scripting language.
423 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
424 specified by any @samp{-L} options.
426 @cindex common allocation
433 These three options are equivalent; multiple forms are supported for
434 compatibility with other linkers. They assign space to common symbols
435 even if a relocatable output file is specified (with @samp{-r}). The
436 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
437 @xref{Miscellaneous Commands}.
439 @cindex entry point, from command line
440 @kindex -e @var{entry}
441 @kindex --entry=@var{entry}
443 @itemx --entry=@var{entry}
444 Use @var{entry} as the explicit symbol for beginning execution of your
445 program, rather than the default entry point. If there is no symbol
446 named @var{entry}, the linker will try to parse @var{entry} as a number,
447 and use that as the entry address (the number will be interpreted in
448 base 10; you may use a leading @samp{0x} for base 16, or a leading
449 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
450 and other ways of specifying the entry point.
452 @kindex --exclude-libs
453 @item --exclude-libs @var{lib},@var{lib},...
454 Specifies a list of archive libraries from which symbols should not be automatically
455 exported. The library names may be delimited by commas or colons. Specifying
456 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
457 automatic export. This option is available only for the i386 PE targeted
458 port of the linker and for ELF targeted ports. For i386 PE, symbols
459 explicitly listed in a .def file are still exported, regardless of this
460 option. For ELF targeted ports, symbols affected by this option will
461 be treated as hidden.
463 @cindex dynamic symbol table
465 @kindex --export-dynamic
467 @itemx --export-dynamic
468 When creating a dynamically linked executable, add all symbols to the
469 dynamic symbol table. The dynamic symbol table is the set of symbols
470 which are visible from dynamic objects at run time.
472 If you do not use this option, the dynamic symbol table will normally
473 contain only those symbols which are referenced by some dynamic object
474 mentioned in the link.
476 If you use @code{dlopen} to load a dynamic object which needs to refer
477 back to the symbols defined by the program, rather than some other
478 dynamic object, then you will probably need to use this option when
479 linking the program itself.
481 You can also use the version script to control what symbols should
482 be added to the dynamic symbol table if the output format supports it.
483 See the description of @samp{--version-script} in @ref{VERSION}.
485 @ifclear SingleFormat
486 @cindex big-endian objects
490 Link big-endian objects. This affects the default output format.
492 @cindex little-endian objects
495 Link little-endian objects. This affects the default output format.
501 @itemx --auxiliary @var{name}
502 When creating an ELF shared object, set the internal DT_AUXILIARY field
503 to the specified name. This tells the dynamic linker that the symbol
504 table of the shared object should be used as an auxiliary filter on the
505 symbol table of the shared object @var{name}.
507 If you later link a program against this filter object, then, when you
508 run the program, the dynamic linker will see the DT_AUXILIARY field. If
509 the dynamic linker resolves any symbols from the filter object, it will
510 first check whether there is a definition in the shared object
511 @var{name}. If there is one, it will be used instead of the definition
512 in the filter object. The shared object @var{name} need not exist.
513 Thus the shared object @var{name} may be used to provide an alternative
514 implementation of certain functions, perhaps for debugging or for
515 machine specific performance.
517 This option may be specified more than once. The DT_AUXILIARY entries
518 will be created in the order in which they appear on the command line.
523 @itemx --filter @var{name}
524 When creating an ELF shared object, set the internal DT_FILTER field to
525 the specified name. This tells the dynamic linker that the symbol table
526 of the shared object which is being created should be used as a filter
527 on the symbol table of the shared object @var{name}.
529 If you later link a program against this filter object, then, when you
530 run the program, the dynamic linker will see the DT_FILTER field. The
531 dynamic linker will resolve symbols according to the symbol table of the
532 filter object as usual, but it will actually link to the definitions
533 found in the shared object @var{name}. Thus the filter object can be
534 used to select a subset of the symbols provided by the object
537 Some older linkers used the @option{-F} option throughout a compilation
538 toolchain for specifying object-file format for both input and output
540 @ifclear SingleFormat
541 The @sc{gnu} linker uses other mechanisms for this purpose: the
542 @option{-b}, @option{--format}, @option{--oformat} options, the
543 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
544 environment variable.
546 The @sc{gnu} linker will ignore the @option{-F} option when not
547 creating an ELF shared object.
549 @cindex finalization function
551 @item -fini @var{name}
552 When creating an ELF executable or shared object, call NAME when the
553 executable or shared object is unloaded, by setting DT_FINI to the
554 address of the function. By default, the linker uses @code{_fini} as
555 the function to call.
559 Ignored. Provided for compatibility with other tools.
565 @itemx --gpsize=@var{value}
566 Set the maximum size of objects to be optimized using the GP register to
567 @var{size}. This is only meaningful for object file formats such as
568 MIPS ECOFF which supports putting large and small objects into different
569 sections. This is ignored for other object file formats.
571 @cindex runtime library name
573 @kindex -soname=@var{name}
575 @itemx -soname=@var{name}
576 When creating an ELF shared object, set the internal DT_SONAME field to
577 the specified name. When an executable is linked with a shared object
578 which has a DT_SONAME field, then when the executable is run the dynamic
579 linker will attempt to load the shared object specified by the DT_SONAME
580 field rather than the using the file name given to the linker.
583 @cindex incremental link
585 Perform an incremental link (same as option @samp{-r}).
587 @cindex initialization function
589 @item -init @var{name}
590 When creating an ELF executable or shared object, call NAME when the
591 executable or shared object is loaded, by setting DT_INIT to the address
592 of the function. By default, the linker uses @code{_init} as the
595 @cindex archive files, from cmd line
596 @kindex -l@var{archive}
597 @kindex --library=@var{archive}
598 @item -l@var{archive}
599 @itemx --library=@var{archive}
600 Add archive file @var{archive} to the list of files to link. This
601 option may be used any number of times. @command{ld} will search its
602 path-list for occurrences of @code{lib@var{archive}.a} for every
603 @var{archive} specified.
605 On systems which support shared libraries, @command{ld} may also search for
606 libraries with extensions other than @code{.a}. Specifically, on ELF
607 and SunOS systems, @command{ld} will search a directory for a library with
608 an extension of @code{.so} before searching for one with an extension of
609 @code{.a}. By convention, a @code{.so} extension indicates a shared
612 The linker will search an archive only once, at the location where it is
613 specified on the command line. If the archive defines a symbol which
614 was undefined in some object which appeared before the archive on the
615 command line, the linker will include the appropriate file(s) from the
616 archive. However, an undefined symbol in an object appearing later on
617 the command line will not cause the linker to search the archive again.
619 See the @option{-(} option for a way to force the linker to search
620 archives multiple times.
622 You may list the same archive multiple times on the command line.
625 This type of archive searching is standard for Unix linkers. However,
626 if you are using @command{ld} on AIX, note that it is different from the
627 behaviour of the AIX linker.
630 @cindex search directory, from cmd line
632 @kindex --library-path=@var{dir}
633 @item -L@var{searchdir}
634 @itemx --library-path=@var{searchdir}
635 Add path @var{searchdir} to the list of paths that @command{ld} will search
636 for archive libraries and @command{ld} control scripts. You may use this
637 option any number of times. The directories are searched in the order
638 in which they are specified on the command line. Directories specified
639 on the command line are searched before the default directories. All
640 @option{-L} options apply to all @option{-l} options, regardless of the
641 order in which the options appear.
643 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
644 by the @dfn{sysroot prefix}, a path specified when the linker is configured.
647 The default set of paths searched (without being specified with
648 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
649 some cases also on how it was configured. @xref{Environment}.
652 The paths can also be specified in a link script with the
653 @code{SEARCH_DIR} command. Directories specified this way are searched
654 at the point in which the linker script appears in the command line.
657 @kindex -m @var{emulation}
658 @item -m@var{emulation}
659 Emulate the @var{emulation} linker. You can list the available
660 emulations with the @samp{--verbose} or @samp{-V} options.
662 If the @samp{-m} option is not used, the emulation is taken from the
663 @code{LDEMULATION} environment variable, if that is defined.
665 Otherwise, the default emulation depends upon how the linker was
673 Print a link map to the standard output. A link map provides
674 information about the link, including the following:
678 Where object files and symbols are mapped into memory.
680 How common symbols are allocated.
682 All archive members included in the link, with a mention of the symbol
683 which caused the archive member to be brought in.
687 @cindex read-only text
692 Turn off page alignment of sections, and mark the output as
693 @code{NMAGIC} if possible.
697 @cindex read/write from cmd line
701 Set the text and data sections to be readable and writable. Also, do
702 not page-align the data segment, and disable linking against shared
703 libraries. If the output format supports Unix style magic numbers,
704 mark the output as @code{OMAGIC}. Note: Although a writable text section
705 is allowed for PE-COFF targets, it does not conform to the format
706 specification published by Microsoft.
711 This option negates most of the effects of the @option{-N} option. It
712 sets the text section to be read-only, and forces the data segment to
713 be page-aligned. Note - this option does not enable linking against
714 shared libraries. Use @option{-Bdynamic} for this.
716 @kindex -o @var{output}
717 @kindex --output=@var{output}
718 @cindex naming the output file
719 @item -o @var{output}
720 @itemx --output=@var{output}
721 Use @var{output} as the name for the program produced by @command{ld}; if this
722 option is not specified, the name @file{a.out} is used by default. The
723 script command @code{OUTPUT} can also specify the output file name.
725 @kindex -O @var{level}
726 @cindex generating optimized output
728 If @var{level} is a numeric values greater than zero @command{ld} optimizes
729 the output. This might take significantly longer and therefore probably
730 should only be enabled for the final binary.
733 @kindex --emit-relocs
734 @cindex retain relocations in final executable
737 Leave relocation sections and contents in fully linked exececutables.
738 Post link analysis and optimization tools may need this information in
739 order to perform correct modifications of executables. This results
740 in larger executables.
742 This option is currently only supported on ELF platforms.
745 @cindex relocatable output
747 @kindex --relocatable
750 Generate relocatable output---i.e., generate an output file that can in
751 turn serve as input to @command{ld}. This is often called @dfn{partial
752 linking}. As a side effect, in environments that support standard Unix
753 magic numbers, this option also sets the output file's magic number to
755 @c ; see @option{-N}.
756 If this option is not specified, an absolute file is produced. When
757 linking C++ programs, this option @emph{will not} resolve references to
758 constructors; to do that, use @samp{-Ur}.
760 When an input file does not have the same format as the output file,
761 partial linking is only supported if that input file does not contain any
762 relocations. Different output formats can have further restrictions; for
763 example some @code{a.out}-based formats do not support partial linking
764 with input files in other formats at all.
766 This option does the same thing as @samp{-i}.
768 @kindex -R @var{file}
769 @kindex --just-symbols=@var{file}
770 @cindex symbol-only input
771 @item -R @var{filename}
772 @itemx --just-symbols=@var{filename}
773 Read symbol names and their addresses from @var{filename}, but do not
774 relocate it or include it in the output. This allows your output file
775 to refer symbolically to absolute locations of memory defined in other
776 programs. You may use this option more than once.
778 For compatibility with other ELF linkers, if the @option{-R} option is
779 followed by a directory name, rather than a file name, it is treated as
780 the @option{-rpath} option.
784 @cindex strip all symbols
787 Omit all symbol information from the output file.
790 @kindex --strip-debug
791 @cindex strip debugger symbols
794 Omit debugger symbol information (but not all symbols) from the output file.
798 @cindex input files, displaying
801 Print the names of the input files as @command{ld} processes them.
803 @kindex -T @var{script}
804 @kindex --script=@var{script}
806 @item -T @var{scriptfile}
807 @itemx --script=@var{scriptfile}
808 Use @var{scriptfile} as the linker script. This script replaces
809 @command{ld}'s default linker script (rather than adding to it), so
810 @var{commandfile} must specify everything necessary to describe the
811 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
812 the current directory, @code{ld} looks for it in the directories
813 specified by any preceding @samp{-L} options. Multiple @samp{-T}
816 @kindex -u @var{symbol}
817 @kindex --undefined=@var{symbol}
818 @cindex undefined symbol
819 @item -u @var{symbol}
820 @itemx --undefined=@var{symbol}
821 Force @var{symbol} to be entered in the output file as an undefined
822 symbol. Doing this may, for example, trigger linking of additional
823 modules from standard libraries. @samp{-u} may be repeated with
824 different option arguments to enter additional undefined symbols. This
825 option is equivalent to the @code{EXTERN} linker script command.
830 For anything other than C++ programs, this option is equivalent to
831 @samp{-r}: it generates relocatable output---i.e., an output file that can in
832 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
833 @emph{does} resolve references to constructors, unlike @samp{-r}.
834 It does not work to use @samp{-Ur} on files that were themselves linked
835 with @samp{-Ur}; once the constructor table has been built, it cannot
836 be added to. Use @samp{-Ur} only for the last partial link, and
837 @samp{-r} for the others.
839 @kindex --unique[=@var{SECTION}]
840 @item --unique[=@var{SECTION}]
841 Creates a separate output section for every input section matching
842 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
843 missing, for every orphan input section. An orphan section is one not
844 specifically mentioned in a linker script. You may use this option
845 multiple times on the command line; It prevents the normal merging of
846 input sections with the same name, overriding output section assignments
856 Display the version number for @command{ld}. The @option{-V} option also
857 lists the supported emulations.
860 @kindex --discard-all
861 @cindex deleting local symbols
864 Delete all local symbols.
867 @kindex --discard-locals
868 @cindex local symbols, deleting
869 @cindex L, deleting symbols beginning
871 @itemx --discard-locals
872 Delete all temporary local symbols. For most targets, this is all local
873 symbols whose names begin with @samp{L}.
875 @kindex -y @var{symbol}
876 @kindex --trace-symbol=@var{symbol}
877 @cindex symbol tracing
878 @item -y @var{symbol}
879 @itemx --trace-symbol=@var{symbol}
880 Print the name of each linked file in which @var{symbol} appears. This
881 option may be given any number of times. On many systems it is necessary
882 to prepend an underscore.
884 This option is useful when you have an undefined symbol in your link but
885 don't know where the reference is coming from.
887 @kindex -Y @var{path}
889 Add @var{path} to the default library search path. This option exists
890 for Solaris compatibility.
892 @kindex -z @var{keyword}
893 @item -z @var{keyword}
894 The recognized keywords are:
898 Combines multiple reloc sections and sorts them to make dynamic symbol
899 lookup caching possible.
902 Disallows undefined symbols in object files. Undefined symbols in
903 shared libraries are still allowed.
906 This option is only meaningful when building a shared object.
907 It marks the object so that its runtime initialization will occur
908 before the runtime initialization of any other objects brought into
909 the process at the same time. Similarly the runtime finalization of
910 the object will occur after the runtime finalization of any other
914 Marks the object that its symbol table interposes before all symbols
915 but the primary executable.
918 Marks the object that its filters be processed immediately at
922 Allows multiple definitions.
925 Disables multiple reloc sections combining.
928 Disables production of copy relocs.
931 Marks the object that the search for dependencies of this object will
932 ignore any default library search paths.
935 Marks the object shouldn't be unloaded at runtime.
938 Marks the object not available to @code{dlopen}.
941 Marks the object can not be dumped by @code{dldump}.
944 When generating an executable or shared library, mark it to tell the
945 dynamic linker to resolve all symbols when the program is started, or
946 when the shared library is linked to using dlopen, instead of
947 deferring function call resolution to the point when the function is
951 Marks the object may contain $ORIGIN.
955 Other keywords are ignored for Solaris compatibility.
958 @cindex groups of archives
959 @item -( @var{archives} -)
960 @itemx --start-group @var{archives} --end-group
961 The @var{archives} should be a list of archive files. They may be
962 either explicit file names, or @samp{-l} options.
964 The specified archives are searched repeatedly until no new undefined
965 references are created. Normally, an archive is searched only once in
966 the order that it is specified on the command line. If a symbol in that
967 archive is needed to resolve an undefined symbol referred to by an
968 object in an archive that appears later on the command line, the linker
969 would not be able to resolve that reference. By grouping the archives,
970 they all be searched repeatedly until all possible references are
973 Using this option has a significant performance cost. It is best to use
974 it only when there are unavoidable circular references between two or
977 @kindex --accept-unknown-input-arch
978 @kindex --no-accept-unknown-input-arch
979 @item --accept-unknown-input-arch
980 @itemx --no-accept-unknown-input-arch
981 Tells the linker to accept input files whose architecture cannot be
982 recognised. The assumption is that the user knows what they are doing
983 and deliberately wants to link in these unknown input files. This was
984 the default behaviour of the linker, before release 2.14. The default
985 behaviour from release 2.14 onwards is to reject such input files, and
986 so the @samp{--accept-unknown-input-arch} option has been added to
987 restore the old behaviour.
990 @kindex --no-as-needed
992 @itemx --no-as-needed
993 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
994 on the command line after the @option{--as-needed} option. Normally,
995 the linker will add a DT_NEEDED tag for each dynamic library mentioned
996 on the command line, regardless of whether the library is actually
997 needed. @option{--as-needed} causes DT_NEEDED tags to only be emitted
998 for libraries that satisfy some reference from regular objects.
999 @option{--no-as-needed} restores the default behaviour.
1001 @kindex --add-needed
1002 @kindex --no-add-needed
1004 @itemx --no-add-needed
1005 This option affects the treatment of dynamic libraries from ELF
1006 DT_NEEDED tags in dynamic libraries mentioned on the command line after
1007 the @option{--no-add-needed} option. Normally, the linker will add
1008 a DT_NEEDED tag for each dynamic library from DT_NEEDED tags.
1009 @option{--no-add-needed} causes DT_NEEDED tags will never be emitted
1010 for those libraries from DT_NEEDED tags. @option{--add-needed} restores
1011 the default behaviour.
1013 @kindex -assert @var{keyword}
1014 @item -assert @var{keyword}
1015 This option is ignored for SunOS compatibility.
1019 @kindex -call_shared
1023 Link against dynamic libraries. This is only meaningful on platforms
1024 for which shared libraries are supported. This option is normally the
1025 default on such platforms. The different variants of this option are
1026 for compatibility with various systems. You may use this option
1027 multiple times on the command line: it affects library searching for
1028 @option{-l} options which follow it.
1032 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1033 section. This causes the runtime linker to handle lookups in this
1034 object and its dependencies to be performed only inside the group.
1035 @option{--unresolved-symbols=report-all} is implied. This option is
1036 only meaningful on ELF platforms which support shared libraries.
1046 Do not link against shared libraries. This is only meaningful on
1047 platforms for which shared libraries are supported. The different
1048 variants of this option are for compatibility with various systems. You
1049 may use this option multiple times on the command line: it affects
1050 library searching for @option{-l} options which follow it. This
1051 option also implies @option{--unresolved-symbols=report-all}.
1055 When creating a shared library, bind references to global symbols to the
1056 definition within the shared library, if any. Normally, it is possible
1057 for a program linked against a shared library to override the definition
1058 within the shared library. This option is only meaningful on ELF
1059 platforms which support shared libraries.
1061 @kindex --check-sections
1062 @kindex --no-check-sections
1063 @item --check-sections
1064 @itemx --no-check-sections
1065 Asks the linker @emph{not} to check section addresses after they have
1066 been assigned to see if there any overlaps. Normally the linker will
1067 perform this check, and if it finds any overlaps it will produce
1068 suitable error messages. The linker does know about, and does make
1069 allowances for sections in overlays. The default behaviour can be
1070 restored by using the command line switch @option{--check-sections}.
1072 @cindex cross reference table
1075 Output a cross reference table. If a linker map file is being
1076 generated, the cross reference table is printed to the map file.
1077 Otherwise, it is printed on the standard output.
1079 The format of the table is intentionally simple, so that it may be
1080 easily processed by a script if necessary. The symbols are printed out,
1081 sorted by name. For each symbol, a list of file names is given. If the
1082 symbol is defined, the first file listed is the location of the
1083 definition. The remaining files contain references to the symbol.
1085 @cindex common allocation
1086 @kindex --no-define-common
1087 @item --no-define-common
1088 This option inhibits the assignment of addresses to common symbols.
1089 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1090 @xref{Miscellaneous Commands}.
1092 The @samp{--no-define-common} option allows decoupling
1093 the decision to assign addresses to Common symbols from the choice
1094 of the output file type; otherwise a non-Relocatable output type
1095 forces assigning addresses to Common symbols.
1096 Using @samp{--no-define-common} allows Common symbols that are referenced
1097 from a shared library to be assigned addresses only in the main program.
1098 This eliminates the unused duplicate space in the shared library,
1099 and also prevents any possible confusion over resolving to the wrong
1100 duplicate when there are many dynamic modules with specialized search
1101 paths for runtime symbol resolution.
1103 @cindex symbols, from command line
1104 @kindex --defsym @var{symbol}=@var{exp}
1105 @item --defsym @var{symbol}=@var{expression}
1106 Create a global symbol in the output file, containing the absolute
1107 address given by @var{expression}. You may use this option as many
1108 times as necessary to define multiple symbols in the command line. A
1109 limited form of arithmetic is supported for the @var{expression} in this
1110 context: you may give a hexadecimal constant or the name of an existing
1111 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1112 constants or symbols. If you need more elaborate expressions, consider
1113 using the linker command language from a script (@pxref{Assignments,,
1114 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1115 space between @var{symbol}, the equals sign (``@key{=}''), and
1118 @cindex demangling, from command line
1119 @kindex --demangle[=@var{style}]
1120 @kindex --no-demangle
1121 @item --demangle[=@var{style}]
1122 @itemx --no-demangle
1123 These options control whether to demangle symbol names in error messages
1124 and other output. When the linker is told to demangle, it tries to
1125 present symbol names in a readable fashion: it strips leading
1126 underscores if they are used by the object file format, and converts C++
1127 mangled symbol names into user readable names. Different compilers have
1128 different mangling styles. The optional demangling style argument can be used
1129 to choose an appropriate demangling style for your compiler. The linker will
1130 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1131 is set. These options may be used to override the default.
1133 @cindex dynamic linker, from command line
1134 @kindex -I@var{file}
1135 @kindex --dynamic-linker @var{file}
1136 @item --dynamic-linker @var{file}
1137 Set the name of the dynamic linker. This is only meaningful when
1138 generating dynamically linked ELF executables. The default dynamic
1139 linker is normally correct; don't use this unless you know what you are
1143 @kindex --fatal-warnings
1144 @item --fatal-warnings
1145 Treat all warnings as errors.
1147 @kindex --force-exe-suffix
1148 @item --force-exe-suffix
1149 Make sure that an output file has a .exe suffix.
1151 If a successfully built fully linked output file does not have a
1152 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1153 the output file to one of the same name with a @code{.exe} suffix. This
1154 option is useful when using unmodified Unix makefiles on a Microsoft
1155 Windows host, since some versions of Windows won't run an image unless
1156 it ends in a @code{.exe} suffix.
1158 @kindex --gc-sections
1159 @kindex --no-gc-sections
1160 @cindex garbage collection
1161 @item --no-gc-sections
1162 @itemx --gc-sections
1163 Enable garbage collection of unused input sections. It is ignored on
1164 targets that do not support this option. This option is not compatible
1165 with @samp{-r}. The default behaviour (of not performing this garbage
1166 collection) can be restored by specifying @samp{--no-gc-sections} on
1173 Print a summary of the command-line options on the standard output and exit.
1175 @kindex --target-help
1177 Print a summary of all target specific options on the standard output and exit.
1180 @item -Map @var{mapfile}
1181 Print a link map to the file @var{mapfile}. See the description of the
1182 @option{-M} option, above.
1184 @cindex memory usage
1185 @kindex --no-keep-memory
1186 @item --no-keep-memory
1187 @command{ld} normally optimizes for speed over memory usage by caching the
1188 symbol tables of input files in memory. This option tells @command{ld} to
1189 instead optimize for memory usage, by rereading the symbol tables as
1190 necessary. This may be required if @command{ld} runs out of memory space
1191 while linking a large executable.
1193 @kindex --no-undefined
1195 @item --no-undefined
1197 Report unresolved symbol references from regular object files. This
1198 is done even if the linker is creating a non-symbolic shared library.
1199 The switch @option{--[no-]allow-shlib-undefined} controls the
1200 behaviour for reporting unresolved references found in shared
1201 libraries being linked in.
1203 @kindex --allow-multiple-definition
1205 @item --allow-multiple-definition
1207 Normally when a symbol is defined multiple times, the linker will
1208 report a fatal error. These options allow multiple definitions and the
1209 first definition will be used.
1211 @kindex --allow-shlib-undefined
1212 @kindex --no-allow-shlib-undefined
1213 @item --allow-shlib-undefined
1214 @itemx --no-allow-shlib-undefined
1215 Allows (the default) or disallows undefined symbols in shared libraries.
1216 This switch is similar to @option{--no-undefined} except that it
1217 determines the behaviour when the undefined symbols are in a
1218 shared library rather than a regular object file. It does not affect
1219 how undefined symbols in regular object files are handled.
1221 The reason that @option{--allow-shlib-undefined} is the default is that
1222 the shared library being specified at link time may not be the same as
1223 the one that is available at load time, so the symbols might actually be
1224 resolvable at load time. Plus there are some systems, (eg BeOS) where
1225 undefined symbols in shared libraries is normal. (The kernel patches
1226 them at load time to select which function is most appropriate
1227 for the current architecture. This is used for example to dynamically
1228 select an appropriate memset function). Apparently it is also normal
1229 for HPPA shared libraries to have undefined symbols.
1231 @kindex --no-undefined-version
1232 @item --no-undefined-version
1233 Normally when a symbol has an undefined version, the linker will ignore
1234 it. This option disallows symbols with undefined version and a fatal error
1235 will be issued instead.
1237 @kindex --default-symver
1238 @item --default-symver
1239 Create and use a default symbol version (the soname) for unversioned
1242 @kindex --default-imported-symver
1243 @item --default-imported-symver
1244 Create and use a default symbol version (the soname) for unversioned
1247 @kindex --no-warn-mismatch
1248 @item --no-warn-mismatch
1249 Normally @command{ld} will give an error if you try to link together input
1250 files that are mismatched for some reason, perhaps because they have
1251 been compiled for different processors or for different endiannesses.
1252 This option tells @command{ld} that it should silently permit such possible
1253 errors. This option should only be used with care, in cases when you
1254 have taken some special action that ensures that the linker errors are
1257 @kindex --no-whole-archive
1258 @item --no-whole-archive
1259 Turn off the effect of the @option{--whole-archive} option for subsequent
1262 @cindex output file after errors
1263 @kindex --noinhibit-exec
1264 @item --noinhibit-exec
1265 Retain the executable output file whenever it is still usable.
1266 Normally, the linker will not produce an output file if it encounters
1267 errors during the link process; it exits without writing an output file
1268 when it issues any error whatsoever.
1272 Only search library directories explicitly specified on the
1273 command line. Library directories specified in linker scripts
1274 (including linker scripts specified on the command line) are ignored.
1276 @ifclear SingleFormat
1278 @item --oformat @var{output-format}
1279 @command{ld} may be configured to support more than one kind of object
1280 file. If your @command{ld} is configured this way, you can use the
1281 @samp{--oformat} option to specify the binary format for the output
1282 object file. Even when @command{ld} is configured to support alternative
1283 object formats, you don't usually need to specify this, as @command{ld}
1284 should be configured to produce as a default output format the most
1285 usual format on each machine. @var{output-format} is a text string, the
1286 name of a particular format supported by the BFD libraries. (You can
1287 list the available binary formats with @samp{objdump -i}.) The script
1288 command @code{OUTPUT_FORMAT} can also specify the output format, but
1289 this option overrides it. @xref{BFD}.
1293 @kindex --pic-executable
1295 @itemx --pic-executable
1296 @cindex position independent executables
1297 Create a position independent executable. This is currently only supported on
1298 ELF platforms. Position independent executables are similar to shared
1299 libraries in that they are relocated by the dynamic linker to the virtual
1300 address the OS chooses for them (which can vary between invocations). Like
1301 normal dynamically linked executables they can be executed and symbols
1302 defined in the executable cannot be overridden by shared libraries.
1306 This option is ignored for Linux compatibility.
1310 This option is ignored for SVR4 compatibility.
1313 @cindex synthesizing linker
1314 @cindex relaxing addressing modes
1316 An option with machine dependent effects.
1318 This option is only supported on a few targets.
1321 @xref{H8/300,,@command{ld} and the H8/300}.
1324 @xref{i960,, @command{ld} and the Intel 960 family}.
1327 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1330 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1333 On some platforms, the @samp{--relax} option performs global
1334 optimizations that become possible when the linker resolves addressing
1335 in the program, such as relaxing address modes and synthesizing new
1336 instructions in the output object file.
1338 On some platforms these link time global optimizations may make symbolic
1339 debugging of the resulting executable impossible.
1342 the case for the Matsushita MN10200 and MN10300 family of processors.
1346 On platforms where this is not supported, @samp{--relax} is accepted,
1350 @cindex retaining specified symbols
1351 @cindex stripping all but some symbols
1352 @cindex symbols, retaining selectively
1353 @item --retain-symbols-file @var{filename}
1354 Retain @emph{only} the symbols listed in the file @var{filename},
1355 discarding all others. @var{filename} is simply a flat file, with one
1356 symbol name per line. This option is especially useful in environments
1360 where a large global symbol table is accumulated gradually, to conserve
1363 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1364 or symbols needed for relocations.
1366 You may only specify @samp{--retain-symbols-file} once in the command
1367 line. It overrides @samp{-s} and @samp{-S}.
1370 @item -rpath @var{dir}
1371 @cindex runtime library search path
1373 Add a directory to the runtime library search path. This is used when
1374 linking an ELF executable with shared objects. All @option{-rpath}
1375 arguments are concatenated and passed to the runtime linker, which uses
1376 them to locate shared objects at runtime. The @option{-rpath} option is
1377 also used when locating shared objects which are needed by shared
1378 objects explicitly included in the link; see the description of the
1379 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1380 ELF executable, the contents of the environment variable
1381 @code{LD_RUN_PATH} will be used if it is defined.
1383 The @option{-rpath} option may also be used on SunOS. By default, on
1384 SunOS, the linker will form a runtime search patch out of all the
1385 @option{-L} options it is given. If a @option{-rpath} option is used, the
1386 runtime search path will be formed exclusively using the @option{-rpath}
1387 options, ignoring the @option{-L} options. This can be useful when using
1388 gcc, which adds many @option{-L} options which may be on NFS mounted
1391 For compatibility with other ELF linkers, if the @option{-R} option is
1392 followed by a directory name, rather than a file name, it is treated as
1393 the @option{-rpath} option.
1397 @cindex link-time runtime library search path
1399 @item -rpath-link @var{DIR}
1400 When using ELF or SunOS, one shared library may require another. This
1401 happens when an @code{ld -shared} link includes a shared library as one
1404 When the linker encounters such a dependency when doing a non-shared,
1405 non-relocatable link, it will automatically try to locate the required
1406 shared library and include it in the link, if it is not included
1407 explicitly. In such a case, the @option{-rpath-link} option
1408 specifies the first set of directories to search. The
1409 @option{-rpath-link} option may specify a sequence of directory names
1410 either by specifying a list of names separated by colons, or by
1411 appearing multiple times.
1413 This option should be used with caution as it overrides the search path
1414 that may have been hard compiled into a shared library. In such a case it
1415 is possible to use unintentionally a different search path than the
1416 runtime linker would do.
1418 The linker uses the following search paths to locate required shared
1422 Any directories specified by @option{-rpath-link} options.
1424 Any directories specified by @option{-rpath} options. The difference
1425 between @option{-rpath} and @option{-rpath-link} is that directories
1426 specified by @option{-rpath} options are included in the executable and
1427 used at runtime, whereas the @option{-rpath-link} option is only effective
1428 at link time. It is for the native linker only.
1430 On an ELF system, if the @option{-rpath} and @code{rpath-link} options
1431 were not used, search the contents of the environment variable
1432 @code{LD_RUN_PATH}. It is for the native linker only.
1434 On SunOS, if the @option{-rpath} option was not used, search any
1435 directories specified using @option{-L} options.
1437 For a native linker, the contents of the environment variable
1438 @code{LD_LIBRARY_PATH}.
1440 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1441 @code{DT_RPATH} of a shared library are searched for shared
1442 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1443 @code{DT_RUNPATH} entries exist.
1445 The default directories, normally @file{/lib} and @file{/usr/lib}.
1447 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1448 exists, the list of directories found in that file.
1451 If the required shared library is not found, the linker will issue a
1452 warning and continue with the link.
1459 @cindex shared libraries
1460 Create a shared library. This is currently only supported on ELF, XCOFF
1461 and SunOS platforms. On SunOS, the linker will automatically create a
1462 shared library if the @option{-e} option is not used and there are
1463 undefined symbols in the link.
1466 @kindex --sort-common
1467 This option tells @command{ld} to sort the common symbols by size when it
1468 places them in the appropriate output sections. First come all the one
1469 byte symbols, then all the two byte, then all the four byte, and then
1470 everything else. This is to prevent gaps between symbols due to
1471 alignment constraints.
1473 @kindex --sort-section name
1474 @item --sort-section name
1475 This option will apply @code{SORT_BY_NAME} to all wildcard section
1476 patterns in the linker script.
1478 @kindex --sort-section alignment
1479 @item --sort-section alignment
1480 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1481 patterns in the linker script.
1483 @kindex --split-by-file
1484 @item --split-by-file [@var{size}]
1485 Similar to @option{--split-by-reloc} but creates a new output section for
1486 each input file when @var{size} is reached. @var{size} defaults to a
1487 size of 1 if not given.
1489 @kindex --split-by-reloc
1490 @item --split-by-reloc [@var{count}]
1491 Tries to creates extra sections in the output file so that no single
1492 output section in the file contains more than @var{count} relocations.
1493 This is useful when generating huge relocatable files for downloading into
1494 certain real time kernels with the COFF object file format; since COFF
1495 cannot represent more than 65535 relocations in a single section. Note
1496 that this will fail to work with object file formats which do not
1497 support arbitrary sections. The linker will not split up individual
1498 input sections for redistribution, so if a single input section contains
1499 more than @var{count} relocations one output section will contain that
1500 many relocations. @var{count} defaults to a value of 32768.
1504 Compute and display statistics about the operation of the linker, such
1505 as execution time and memory usage.
1508 @item --sysroot=@var{directory}
1509 Use @var{directory} as the location of the sysroot, overriding the
1510 configure-time default. This option is only supported by linkers
1511 that were configured using @option{--with-sysroot}.
1513 @kindex --traditional-format
1514 @cindex traditional format
1515 @item --traditional-format
1516 For some targets, the output of @command{ld} is different in some ways from
1517 the output of some existing linker. This switch requests @command{ld} to
1518 use the traditional format instead.
1521 For example, on SunOS, @command{ld} combines duplicate entries in the
1522 symbol string table. This can reduce the size of an output file with
1523 full debugging information by over 30 percent. Unfortunately, the SunOS
1524 @code{dbx} program can not read the resulting program (@code{gdb} has no
1525 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1526 combine duplicate entries.
1528 @kindex --section-start @var{sectionname}=@var{org}
1529 @item --section-start @var{sectionname}=@var{org}
1530 Locate a section in the output file at the absolute
1531 address given by @var{org}. You may use this option as many
1532 times as necessary to locate multiple sections in the command
1534 @var{org} must be a single hexadecimal integer;
1535 for compatibility with other linkers, you may omit the leading
1536 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1537 should be no white space between @var{sectionname}, the equals
1538 sign (``@key{=}''), and @var{org}.
1540 @kindex -Tbss @var{org}
1541 @kindex -Tdata @var{org}
1542 @kindex -Ttext @var{org}
1543 @cindex segment origins, cmd line
1544 @item -Tbss @var{org}
1545 @itemx -Tdata @var{org}
1546 @itemx -Ttext @var{org}
1547 Same as --section-start, with @code{.bss}, @code{.data} or
1548 @code{.text} as the @var{sectionname}.
1550 @kindex --unresolved-symbols
1551 @item --unresolved-symbols=@var{method}
1552 Determine how to handle unresolved symbols. There are four possible
1553 values for @samp{method}:
1557 Do not report any unresolved symbols.
1560 Report all unresolved symbols. This is the default.
1562 @item ignore-in-object-files
1563 Report unresolved symbols that are contained in shared libraries, but
1564 ignore them if they come from regular object files.
1566 @item ignore-in-shared-libs
1567 Report unresolved symbols that come from regular object files, but
1568 ignore them if they come from shared libraries. This can be useful
1569 when creating a dynamic binary and it is known that all the shared
1570 libraries that it should be referencing are included on the linker's
1574 The behaviour for shared libraries on their own can also be controlled
1575 by the @option{--[no-]allow-shlib-undefined} option.
1577 Normally the linker will generate an error message for each reported
1578 unresolved symbol but the option @option{--warn-unresolved-symbols}
1579 can change this to a warning.
1585 Display the version number for @command{ld} and list the linker emulations
1586 supported. Display which input files can and cannot be opened. Display
1587 the linker script being used by the linker.
1589 @kindex --version-script=@var{version-scriptfile}
1590 @cindex version script, symbol versions
1591 @itemx --version-script=@var{version-scriptfile}
1592 Specify the name of a version script to the linker. This is typically
1593 used when creating shared libraries to specify additional information
1594 about the version hierarchy for the library being created. This option
1595 is only meaningful on ELF platforms which support shared libraries.
1598 @kindex --warn-common
1599 @cindex warnings, on combining symbols
1600 @cindex combining symbols, warnings on
1602 Warn when a common symbol is combined with another common symbol or with
1603 a symbol definition. Unix linkers allow this somewhat sloppy practise,
1604 but linkers on some other operating systems do not. This option allows
1605 you to find potential problems from combining global symbols.
1606 Unfortunately, some C libraries use this practise, so you may get some
1607 warnings about symbols in the libraries as well as in your programs.
1609 There are three kinds of global symbols, illustrated here by C examples:
1613 A definition, which goes in the initialized data section of the output
1617 An undefined reference, which does not allocate space.
1618 There must be either a definition or a common symbol for the
1622 A common symbol. If there are only (one or more) common symbols for a
1623 variable, it goes in the uninitialized data area of the output file.
1624 The linker merges multiple common symbols for the same variable into a
1625 single symbol. If they are of different sizes, it picks the largest
1626 size. The linker turns a common symbol into a declaration, if there is
1627 a definition of the same variable.
1630 The @samp{--warn-common} option can produce five kinds of warnings.
1631 Each warning consists of a pair of lines: the first describes the symbol
1632 just encountered, and the second describes the previous symbol
1633 encountered with the same name. One or both of the two symbols will be
1638 Turning a common symbol into a reference, because there is already a
1639 definition for the symbol.
1641 @var{file}(@var{section}): warning: common of `@var{symbol}'
1642 overridden by definition
1643 @var{file}(@var{section}): warning: defined here
1647 Turning a common symbol into a reference, because a later definition for
1648 the symbol is encountered. This is the same as the previous case,
1649 except that the symbols are encountered in a different order.
1651 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1653 @var{file}(@var{section}): warning: common is here
1657 Merging a common symbol with a previous same-sized common symbol.
1659 @var{file}(@var{section}): warning: multiple common
1661 @var{file}(@var{section}): warning: previous common is here
1665 Merging a common symbol with a previous larger common symbol.
1667 @var{file}(@var{section}): warning: common of `@var{symbol}'
1668 overridden by larger common
1669 @var{file}(@var{section}): warning: larger common is here
1673 Merging a common symbol with a previous smaller common symbol. This is
1674 the same as the previous case, except that the symbols are
1675 encountered in a different order.
1677 @var{file}(@var{section}): warning: common of `@var{symbol}'
1678 overriding smaller common
1679 @var{file}(@var{section}): warning: smaller common is here
1683 @kindex --warn-constructors
1684 @item --warn-constructors
1685 Warn if any global constructors are used. This is only useful for a few
1686 object file formats. For formats like COFF or ELF, the linker can not
1687 detect the use of global constructors.
1689 @kindex --warn-multiple-gp
1690 @item --warn-multiple-gp
1691 Warn if multiple global pointer values are required in the output file.
1692 This is only meaningful for certain processors, such as the Alpha.
1693 Specifically, some processors put large-valued constants in a special
1694 section. A special register (the global pointer) points into the middle
1695 of this section, so that constants can be loaded efficiently via a
1696 base-register relative addressing mode. Since the offset in
1697 base-register relative mode is fixed and relatively small (e.g., 16
1698 bits), this limits the maximum size of the constant pool. Thus, in
1699 large programs, it is often necessary to use multiple global pointer
1700 values in order to be able to address all possible constants. This
1701 option causes a warning to be issued whenever this case occurs.
1704 @cindex warnings, on undefined symbols
1705 @cindex undefined symbols, warnings on
1707 Only warn once for each undefined symbol, rather than once per module
1710 @kindex --warn-section-align
1711 @cindex warnings, on section alignment
1712 @cindex section alignment, warnings on
1713 @item --warn-section-align
1714 Warn if the address of an output section is changed because of
1715 alignment. Typically, the alignment will be set by an input section.
1716 The address will only be changed if it not explicitly specified; that
1717 is, if the @code{SECTIONS} command does not specify a start address for
1718 the section (@pxref{SECTIONS}).
1720 @kindex --warn-shared-textrel
1721 @item --warn-shared-textrel
1722 Warn if the linker adds a DT_TEXTREL to a shared object.
1724 @kindex --warn-unresolved-symbols
1725 @item --warn-unresolved-symbols
1726 If the linker is going to report an unresolved symbol (see the option
1727 @option{--unresolved-symbols}) it will normally generate an error.
1728 This option makes it generate a warning instead.
1730 @kindex --error-unresolved-symbols
1731 @item --error-unresolved-symbols
1732 This restores the linker's default behaviour of generating errors when
1733 it is reporting unresolved symbols.
1735 @kindex --whole-archive
1736 @cindex including an entire archive
1737 @item --whole-archive
1738 For each archive mentioned on the command line after the
1739 @option{--whole-archive} option, include every object file in the archive
1740 in the link, rather than searching the archive for the required object
1741 files. This is normally used to turn an archive file into a shared
1742 library, forcing every object to be included in the resulting shared
1743 library. This option may be used more than once.
1745 Two notes when using this option from gcc: First, gcc doesn't know
1746 about this option, so you have to use @option{-Wl,-whole-archive}.
1747 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
1748 list of archives, because gcc will add its own list of archives to
1749 your link and you may not want this flag to affect those as well.
1752 @item --wrap @var{symbol}
1753 Use a wrapper function for @var{symbol}. Any undefined reference to
1754 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
1755 undefined reference to @code{__real_@var{symbol}} will be resolved to
1758 This can be used to provide a wrapper for a system function. The
1759 wrapper function should be called @code{__wrap_@var{symbol}}. If it
1760 wishes to call the system function, it should call
1761 @code{__real_@var{symbol}}.
1763 Here is a trivial example:
1767 __wrap_malloc (size_t c)
1769 printf ("malloc called with %zu\n", c);
1770 return __real_malloc (c);
1774 If you link other code with this file using @option{--wrap malloc}, then
1775 all calls to @code{malloc} will call the function @code{__wrap_malloc}
1776 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
1777 call the real @code{malloc} function.
1779 You may wish to provide a @code{__real_malloc} function as well, so that
1780 links without the @option{--wrap} option will succeed. If you do this,
1781 you should not put the definition of @code{__real_malloc} in the same
1782 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
1783 call before the linker has a chance to wrap it to @code{malloc}.
1785 @kindex --enable-new-dtags
1786 @kindex --disable-new-dtags
1787 @item --enable-new-dtags
1788 @itemx --disable-new-dtags
1789 This linker can create the new dynamic tags in ELF. But the older ELF
1790 systems may not understand them. If you specify
1791 @option{--enable-new-dtags}, the dynamic tags will be created as needed.
1792 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
1793 created. By default, the new dynamic tags are not created. Note that
1794 those options are only available for ELF systems.
1796 @kindex --hash-size=@var{number}
1797 Set the default size of the linker's hash tables to a prime number
1798 close to @var{number}. Increasing this value can reduce the length of
1799 time it takes the linker to perform its tasks, at the expense of
1800 increasing the linker's memory requirements. Similarly reducing this
1801 value can reduce the memory requirements at the expense of speed.
1803 @kindex --reduce-memory-overheads
1804 @item --reduce-memory-overheads
1805 This option reduces memory requirements at ld runtime, at the expense of
1806 linking speed. This was introduced to to select the old O(n^2) algorithm
1807 for link map file generation, rather than the new O(n) algorithm which uses
1808 about 40% more memory for symbol storage.
1810 Another affect of the switch is to set the default hash table size to
1811 1021, which again saves memory at the cost of lengthening the linker's
1812 run time. This is not done however if the @option{--hash-size} switch
1815 The @option{--reduce-memory-overheads} switch may be also be used to
1816 enable other tradeoffs in future versions of the linker.
1822 @subsection Options Specific to i386 PE Targets
1824 @c man begin OPTIONS
1826 The i386 PE linker supports the @option{-shared} option, which causes
1827 the output to be a dynamically linked library (DLL) instead of a
1828 normal executable. You should name the output @code{*.dll} when you
1829 use this option. In addition, the linker fully supports the standard
1830 @code{*.def} files, which may be specified on the linker command line
1831 like an object file (in fact, it should precede archives it exports
1832 symbols from, to ensure that they get linked in, just like a normal
1835 In addition to the options common to all targets, the i386 PE linker
1836 support additional command line options that are specific to the i386
1837 PE target. Options that take values may be separated from their
1838 values by either a space or an equals sign.
1842 @kindex --add-stdcall-alias
1843 @item --add-stdcall-alias
1844 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
1845 as-is and also with the suffix stripped.
1846 [This option is specific to the i386 PE targeted port of the linker]
1849 @item --base-file @var{file}
1850 Use @var{file} as the name of a file in which to save the base
1851 addresses of all the relocations needed for generating DLLs with
1853 [This is an i386 PE specific option]
1857 Create a DLL instead of a regular executable. You may also use
1858 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
1860 [This option is specific to the i386 PE targeted port of the linker]
1862 @kindex --enable-stdcall-fixup
1863 @kindex --disable-stdcall-fixup
1864 @item --enable-stdcall-fixup
1865 @itemx --disable-stdcall-fixup
1866 If the link finds a symbol that it cannot resolve, it will attempt to
1867 do ``fuzzy linking'' by looking for another defined symbol that differs
1868 only in the format of the symbol name (cdecl vs stdcall) and will
1869 resolve that symbol by linking to the match. For example, the
1870 undefined symbol @code{_foo} might be linked to the function
1871 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
1872 to the function @code{_bar}. When the linker does this, it prints a
1873 warning, since it normally should have failed to link, but sometimes
1874 import libraries generated from third-party dlls may need this feature
1875 to be usable. If you specify @option{--enable-stdcall-fixup}, this
1876 feature is fully enabled and warnings are not printed. If you specify
1877 @option{--disable-stdcall-fixup}, this feature is disabled and such
1878 mismatches are considered to be errors.
1879 [This option is specific to the i386 PE targeted port of the linker]
1881 @cindex DLLs, creating
1882 @kindex --export-all-symbols
1883 @item --export-all-symbols
1884 If given, all global symbols in the objects used to build a DLL will
1885 be exported by the DLL. Note that this is the default if there
1886 otherwise wouldn't be any exported symbols. When symbols are
1887 explicitly exported via DEF files or implicitly exported via function
1888 attributes, the default is to not export anything else unless this
1889 option is given. Note that the symbols @code{DllMain@@12},
1890 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
1891 @code{impure_ptr} will not be automatically
1892 exported. Also, symbols imported from other DLLs will not be
1893 re-exported, nor will symbols specifying the DLL's internal layout
1894 such as those beginning with @code{_head_} or ending with
1895 @code{_iname}. In addition, no symbols from @code{libgcc},
1896 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
1897 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
1898 not be exported, to help with C++ DLLs. Finally, there is an
1899 extensive list of cygwin-private symbols that are not exported
1900 (obviously, this applies on when building DLLs for cygwin targets).
1901 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
1902 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
1903 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
1904 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
1905 @code{cygwin_premain3}, and @code{environ}.
1906 [This option is specific to the i386 PE targeted port of the linker]
1908 @kindex --exclude-symbols
1909 @item --exclude-symbols @var{symbol},@var{symbol},...
1910 Specifies a list of symbols which should not be automatically
1911 exported. The symbol names may be delimited by commas or colons.
1912 [This option is specific to the i386 PE targeted port of the linker]
1914 @kindex --file-alignment
1915 @item --file-alignment
1916 Specify the file alignment. Sections in the file will always begin at
1917 file offsets which are multiples of this number. This defaults to
1919 [This option is specific to the i386 PE targeted port of the linker]
1923 @item --heap @var{reserve}
1924 @itemx --heap @var{reserve},@var{commit}
1925 Specify the amount of memory to reserve (and optionally commit) to be
1926 used as heap for this program. The default is 1Mb reserved, 4K
1928 [This option is specific to the i386 PE targeted port of the linker]
1931 @kindex --image-base
1932 @item --image-base @var{value}
1933 Use @var{value} as the base address of your program or dll. This is
1934 the lowest memory location that will be used when your program or dll
1935 is loaded. To reduce the need to relocate and improve performance of
1936 your dlls, each should have a unique base address and not overlap any
1937 other dlls. The default is 0x400000 for executables, and 0x10000000
1939 [This option is specific to the i386 PE targeted port of the linker]
1943 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
1944 symbols before they are exported.
1945 [This option is specific to the i386 PE targeted port of the linker]
1947 @kindex --large-address-aware
1948 @item --large-address-aware
1949 If given, the appropriate bit in the ``Charateristics'' field of the COFF
1950 header is set to indicate that this executable supports virtual addresses
1951 greater than 2 gigabytes. This should be used in conjuction with the /3GB
1952 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
1953 section of the BOOT.INI. Otherwise, this bit has no effect.
1954 [This option is specific to PE targeted ports of the linker]
1956 @kindex --major-image-version
1957 @item --major-image-version @var{value}
1958 Sets the major number of the ``image version''. Defaults to 1.
1959 [This option is specific to the i386 PE targeted port of the linker]
1961 @kindex --major-os-version
1962 @item --major-os-version @var{value}
1963 Sets the major number of the ``os version''. Defaults to 4.
1964 [This option is specific to the i386 PE targeted port of the linker]
1966 @kindex --major-subsystem-version
1967 @item --major-subsystem-version @var{value}
1968 Sets the major number of the ``subsystem version''. Defaults to 4.
1969 [This option is specific to the i386 PE targeted port of the linker]
1971 @kindex --minor-image-version
1972 @item --minor-image-version @var{value}
1973 Sets the minor number of the ``image version''. Defaults to 0.
1974 [This option is specific to the i386 PE targeted port of the linker]
1976 @kindex --minor-os-version
1977 @item --minor-os-version @var{value}
1978 Sets the minor number of the ``os version''. Defaults to 0.
1979 [This option is specific to the i386 PE targeted port of the linker]
1981 @kindex --minor-subsystem-version
1982 @item --minor-subsystem-version @var{value}
1983 Sets the minor number of the ``subsystem version''. Defaults to 0.
1984 [This option is specific to the i386 PE targeted port of the linker]
1986 @cindex DEF files, creating
1987 @cindex DLLs, creating
1988 @kindex --output-def
1989 @item --output-def @var{file}
1990 The linker will create the file @var{file} which will contain a DEF
1991 file corresponding to the DLL the linker is generating. This DEF file
1992 (which should be called @code{*.def}) may be used to create an import
1993 library with @code{dlltool} or may be used as a reference to
1994 automatically or implicitly exported symbols.
1995 [This option is specific to the i386 PE targeted port of the linker]
1997 @cindex DLLs, creating
1998 @kindex --out-implib
1999 @item --out-implib @var{file}
2000 The linker will create the file @var{file} which will contain an
2001 import lib corresponding to the DLL the linker is generating. This
2002 import lib (which should be called @code{*.dll.a} or @code{*.a}
2003 may be used to link clients against the generated DLL; this behaviour
2004 makes it possible to skip a separate @code{dlltool} import library
2006 [This option is specific to the i386 PE targeted port of the linker]
2008 @kindex --enable-auto-image-base
2009 @item --enable-auto-image-base
2010 Automatically choose the image base for DLLs, unless one is specified
2011 using the @code{--image-base} argument. By using a hash generated
2012 from the dllname to create unique image bases for each DLL, in-memory
2013 collisions and relocations which can delay program execution are
2015 [This option is specific to the i386 PE targeted port of the linker]
2017 @kindex --disable-auto-image-base
2018 @item --disable-auto-image-base
2019 Do not automatically generate a unique image base. If there is no
2020 user-specified image base (@code{--image-base}) then use the platform
2022 [This option is specific to the i386 PE targeted port of the linker]
2024 @cindex DLLs, linking to
2025 @kindex --dll-search-prefix
2026 @item --dll-search-prefix @var{string}
2027 When linking dynamically to a dll without an import library,
2028 search for @code{<string><basename>.dll} in preference to
2029 @code{lib<basename>.dll}. This behaviour allows easy distinction
2030 between DLLs built for the various "subplatforms": native, cygwin,
2031 uwin, pw, etc. For instance, cygwin DLLs typically use
2032 @code{--dll-search-prefix=cyg}.
2033 [This option is specific to the i386 PE targeted port of the linker]
2035 @kindex --enable-auto-import
2036 @item --enable-auto-import
2037 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2038 DATA imports from DLLs, and create the necessary thunking symbols when
2039 building the import libraries with those DATA exports. Note: Use of the
2040 'auto-import' extension will cause the text section of the image file
2041 to be made writable. This does not conform to the PE-COFF format
2042 specification published by Microsoft.
2044 Using 'auto-import' generally will 'just work' -- but sometimes you may
2047 "variable '<var>' can't be auto-imported. Please read the
2048 documentation for ld's @code{--enable-auto-import} for details."
2050 This message occurs when some (sub)expression accesses an address
2051 ultimately given by the sum of two constants (Win32 import tables only
2052 allow one). Instances where this may occur include accesses to member
2053 fields of struct variables imported from a DLL, as well as using a
2054 constant index into an array variable imported from a DLL. Any
2055 multiword variable (arrays, structs, long long, etc) may trigger
2056 this error condition. However, regardless of the exact data type
2057 of the offending exported variable, ld will always detect it, issue
2058 the warning, and exit.
2060 There are several ways to address this difficulty, regardless of the
2061 data type of the exported variable:
2063 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2064 of adjusting references in your client code for runtime environment, so
2065 this method works only when runtime environment supports this feature.
2067 A second solution is to force one of the 'constants' to be a variable --
2068 that is, unknown and un-optimizable at compile time. For arrays,
2069 there are two possibilities: a) make the indexee (the array's address)
2070 a variable, or b) make the 'constant' index a variable. Thus:
2073 extern type extern_array[];
2075 @{ volatile type *t=extern_array; t[1] @}
2081 extern type extern_array[];
2083 @{ volatile int t=1; extern_array[t] @}
2086 For structs (and most other multiword data types) the only option
2087 is to make the struct itself (or the long long, or the ...) variable:
2090 extern struct s extern_struct;
2091 extern_struct.field -->
2092 @{ volatile struct s *t=&extern_struct; t->field @}
2098 extern long long extern_ll;
2100 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2103 A third method of dealing with this difficulty is to abandon
2104 'auto-import' for the offending symbol and mark it with
2105 @code{__declspec(dllimport)}. However, in practise that
2106 requires using compile-time #defines to indicate whether you are
2107 building a DLL, building client code that will link to the DLL, or
2108 merely building/linking to a static library. In making the choice
2109 between the various methods of resolving the 'direct address with
2110 constant offset' problem, you should consider typical real-world usage:
2118 void main(int argc, char **argv)@{
2119 printf("%d\n",arr[1]);
2129 void main(int argc, char **argv)@{
2130 /* This workaround is for win32 and cygwin; do not "optimize" */
2131 volatile int *parr = arr;
2132 printf("%d\n",parr[1]);
2139 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2140 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2141 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2142 #define FOO_IMPORT __declspec(dllimport)
2146 extern FOO_IMPORT int arr[];
2149 void main(int argc, char **argv)@{
2150 printf("%d\n",arr[1]);
2154 A fourth way to avoid this problem is to re-code your
2155 library to use a functional interface rather than a data interface
2156 for the offending variables (e.g. set_foo() and get_foo() accessor
2158 [This option is specific to the i386 PE targeted port of the linker]
2160 @kindex --disable-auto-import
2161 @item --disable-auto-import
2162 Do not attempt to do sophisticated linking of @code{_symbol} to
2163 @code{__imp__symbol} for DATA imports from DLLs.
2164 [This option is specific to the i386 PE targeted port of the linker]
2166 @kindex --enable-runtime-pseudo-reloc
2167 @item --enable-runtime-pseudo-reloc
2168 If your code contains expressions described in --enable-auto-import section,
2169 that is, DATA imports from DLL with non-zero offset, this switch will create
2170 a vector of 'runtime pseudo relocations' which can be used by runtime
2171 environment to adjust references to such data in your client code.
2172 [This option is specific to the i386 PE targeted port of the linker]
2174 @kindex --disable-runtime-pseudo-reloc
2175 @item --disable-runtime-pseudo-reloc
2176 Do not create pseudo relocations for non-zero offset DATA imports from
2177 DLLs. This is the default.
2178 [This option is specific to the i386 PE targeted port of the linker]
2180 @kindex --enable-extra-pe-debug
2181 @item --enable-extra-pe-debug
2182 Show additional debug info related to auto-import symbol thunking.
2183 [This option is specific to the i386 PE targeted port of the linker]
2185 @kindex --section-alignment
2186 @item --section-alignment
2187 Sets the section alignment. Sections in memory will always begin at
2188 addresses which are a multiple of this number. Defaults to 0x1000.
2189 [This option is specific to the i386 PE targeted port of the linker]
2193 @item --stack @var{reserve}
2194 @itemx --stack @var{reserve},@var{commit}
2195 Specify the amount of memory to reserve (and optionally commit) to be
2196 used as stack for this program. The default is 2Mb reserved, 4K
2198 [This option is specific to the i386 PE targeted port of the linker]
2201 @item --subsystem @var{which}
2202 @itemx --subsystem @var{which}:@var{major}
2203 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2204 Specifies the subsystem under which your program will execute. The
2205 legal values for @var{which} are @code{native}, @code{windows},
2206 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2207 the subsystem version also. Numeric values are also accepted for
2209 [This option is specific to the i386 PE targeted port of the linker]
2216 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2218 @c man begin OPTIONS
2220 The 68HC11 and 68HC12 linkers support specific options to control the
2221 memory bank switching mapping and trampoline code generation.
2225 @kindex --no-trampoline
2226 @item --no-trampoline
2227 This option disables the generation of trampoline. By default a trampoline
2228 is generated for each far function which is called using a @code{jsr}
2229 instruction (this happens when a pointer to a far function is taken).
2231 @kindex --bank-window
2232 @item --bank-window @var{name}
2233 This option indicates to the linker the name of the memory region in
2234 the @samp{MEMORY} specification that describes the memory bank window.
2235 The definition of such region is then used by the linker to compute
2236 paging and addresses within the memory window.
2245 @section Environment Variables
2247 @c man begin ENVIRONMENT
2249 You can change the behaviour of @command{ld} with the environment variables
2250 @ifclear SingleFormat
2253 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2255 @ifclear SingleFormat
2257 @cindex default input format
2258 @code{GNUTARGET} determines the input-file object format if you don't
2259 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2260 of the BFD names for an input format (@pxref{BFD}). If there is no
2261 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2262 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2263 attempts to discover the input format by examining binary input files;
2264 this method often succeeds, but there are potential ambiguities, since
2265 there is no method of ensuring that the magic number used to specify
2266 object-file formats is unique. However, the configuration procedure for
2267 BFD on each system places the conventional format for that system first
2268 in the search-list, so ambiguities are resolved in favor of convention.
2272 @cindex default emulation
2273 @cindex emulation, default
2274 @code{LDEMULATION} determines the default emulation if you don't use the
2275 @samp{-m} option. The emulation can affect various aspects of linker
2276 behaviour, particularly the default linker script. You can list the
2277 available emulations with the @samp{--verbose} or @samp{-V} options. If
2278 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2279 variable is not defined, the default emulation depends upon how the
2280 linker was configured.
2282 @kindex COLLECT_NO_DEMANGLE
2283 @cindex demangling, default
2284 Normally, the linker will default to demangling symbols. However, if
2285 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2286 default to not demangling symbols. This environment variable is used in
2287 a similar fashion by the @code{gcc} linker wrapper program. The default
2288 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2295 @chapter Linker Scripts
2298 @cindex linker scripts
2299 @cindex command files
2300 Every link is controlled by a @dfn{linker script}. This script is
2301 written in the linker command language.
2303 The main purpose of the linker script is to describe how the sections in
2304 the input files should be mapped into the output file, and to control
2305 the memory layout of the output file. Most linker scripts do nothing
2306 more than this. However, when necessary, the linker script can also
2307 direct the linker to perform many other operations, using the commands
2310 The linker always uses a linker script. If you do not supply one
2311 yourself, the linker will use a default script that is compiled into the
2312 linker executable. You can use the @samp{--verbose} command line option
2313 to display the default linker script. Certain command line options,
2314 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2316 You may supply your own linker script by using the @samp{-T} command
2317 line option. When you do this, your linker script will replace the
2318 default linker script.
2320 You may also use linker scripts implicitly by naming them as input files
2321 to the linker, as though they were files to be linked. @xref{Implicit
2325 * Basic Script Concepts:: Basic Linker Script Concepts
2326 * Script Format:: Linker Script Format
2327 * Simple Example:: Simple Linker Script Example
2328 * Simple Commands:: Simple Linker Script Commands
2329 * Assignments:: Assigning Values to Symbols
2330 * SECTIONS:: SECTIONS Command
2331 * MEMORY:: MEMORY Command
2332 * PHDRS:: PHDRS Command
2333 * VERSION:: VERSION Command
2334 * Expressions:: Expressions in Linker Scripts
2335 * Implicit Linker Scripts:: Implicit Linker Scripts
2338 @node Basic Script Concepts
2339 @section Basic Linker Script Concepts
2340 @cindex linker script concepts
2341 We need to define some basic concepts and vocabulary in order to
2342 describe the linker script language.
2344 The linker combines input files into a single output file. The output
2345 file and each input file are in a special data format known as an
2346 @dfn{object file format}. Each file is called an @dfn{object file}.
2347 The output file is often called an @dfn{executable}, but for our
2348 purposes we will also call it an object file. Each object file has,
2349 among other things, a list of @dfn{sections}. We sometimes refer to a
2350 section in an input file as an @dfn{input section}; similarly, a section
2351 in the output file is an @dfn{output section}.
2353 Each section in an object file has a name and a size. Most sections
2354 also have an associated block of data, known as the @dfn{section
2355 contents}. A section may be marked as @dfn{loadable}, which mean that
2356 the contents should be loaded into memory when the output file is run.
2357 A section with no contents may be @dfn{allocatable}, which means that an
2358 area in memory should be set aside, but nothing in particular should be
2359 loaded there (in some cases this memory must be zeroed out). A section
2360 which is neither loadable nor allocatable typically contains some sort
2361 of debugging information.
2363 Every loadable or allocatable output section has two addresses. The
2364 first is the @dfn{VMA}, or virtual memory address. This is the address
2365 the section will have when the output file is run. The second is the
2366 @dfn{LMA}, or load memory address. This is the address at which the
2367 section will be loaded. In most cases the two addresses will be the
2368 same. An example of when they might be different is when a data section
2369 is loaded into ROM, and then copied into RAM when the program starts up
2370 (this technique is often used to initialize global variables in a ROM
2371 based system). In this case the ROM address would be the LMA, and the
2372 RAM address would be the VMA.
2374 You can see the sections in an object file by using the @code{objdump}
2375 program with the @samp{-h} option.
2377 Every object file also has a list of @dfn{symbols}, known as the
2378 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2379 has a name, and each defined symbol has an address, among other
2380 information. If you compile a C or C++ program into an object file, you
2381 will get a defined symbol for every defined function and global or
2382 static variable. Every undefined function or global variable which is
2383 referenced in the input file will become an undefined symbol.
2385 You can see the symbols in an object file by using the @code{nm}
2386 program, or by using the @code{objdump} program with the @samp{-t}
2390 @section Linker Script Format
2391 @cindex linker script format
2392 Linker scripts are text files.
2394 You write a linker script as a series of commands. Each command is
2395 either a keyword, possibly followed by arguments, or an assignment to a
2396 symbol. You may separate commands using semicolons. Whitespace is
2399 Strings such as file or format names can normally be entered directly.
2400 If the file name contains a character such as a comma which would
2401 otherwise serve to separate file names, you may put the file name in
2402 double quotes. There is no way to use a double quote character in a
2405 You may include comments in linker scripts just as in C, delimited by
2406 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2409 @node Simple Example
2410 @section Simple Linker Script Example
2411 @cindex linker script example
2412 @cindex example of linker script
2413 Many linker scripts are fairly simple.
2415 The simplest possible linker script has just one command:
2416 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2417 memory layout of the output file.
2419 The @samp{SECTIONS} command is a powerful command. Here we will
2420 describe a simple use of it. Let's assume your program consists only of
2421 code, initialized data, and uninitialized data. These will be in the
2422 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2423 Let's assume further that these are the only sections which appear in
2426 For this example, let's say that the code should be loaded at address
2427 0x10000, and that the data should start at address 0x8000000. Here is a
2428 linker script which will do that:
2433 .text : @{ *(.text) @}
2435 .data : @{ *(.data) @}
2436 .bss : @{ *(.bss) @}
2440 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2441 followed by a series of symbol assignments and output section
2442 descriptions enclosed in curly braces.
2444 The first line inside the @samp{SECTIONS} command of the above example
2445 sets the value of the special symbol @samp{.}, which is the location
2446 counter. If you do not specify the address of an output section in some
2447 other way (other ways are described later), the address is set from the
2448 current value of the location counter. The location counter is then
2449 incremented by the size of the output section. At the start of the
2450 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2452 The second line defines an output section, @samp{.text}. The colon is
2453 required syntax which may be ignored for now. Within the curly braces
2454 after the output section name, you list the names of the input sections
2455 which should be placed into this output section. The @samp{*} is a
2456 wildcard which matches any file name. The expression @samp{*(.text)}
2457 means all @samp{.text} input sections in all input files.
2459 Since the location counter is @samp{0x10000} when the output section
2460 @samp{.text} is defined, the linker will set the address of the
2461 @samp{.text} section in the output file to be @samp{0x10000}.
2463 The remaining lines define the @samp{.data} and @samp{.bss} sections in
2464 the output file. The linker will place the @samp{.data} output section
2465 at address @samp{0x8000000}. After the linker places the @samp{.data}
2466 output section, the value of the location counter will be
2467 @samp{0x8000000} plus the size of the @samp{.data} output section. The
2468 effect is that the linker will place the @samp{.bss} output section
2469 immediately after the @samp{.data} output section in memory.
2471 The linker will ensure that each output section has the required
2472 alignment, by increasing the location counter if necessary. In this
2473 example, the specified addresses for the @samp{.text} and @samp{.data}
2474 sections will probably satisfy any alignment constraints, but the linker
2475 may have to create a small gap between the @samp{.data} and @samp{.bss}
2478 That's it! That's a simple and complete linker script.
2480 @node Simple Commands
2481 @section Simple Linker Script Commands
2482 @cindex linker script simple commands
2483 In this section we describe the simple linker script commands.
2486 * Entry Point:: Setting the entry point
2487 * File Commands:: Commands dealing with files
2488 @ifclear SingleFormat
2489 * Format Commands:: Commands dealing with object file formats
2492 * Miscellaneous Commands:: Other linker script commands
2496 @subsection Setting the Entry Point
2497 @kindex ENTRY(@var{symbol})
2498 @cindex start of execution
2499 @cindex first instruction
2501 The first instruction to execute in a program is called the @dfn{entry
2502 point}. You can use the @code{ENTRY} linker script command to set the
2503 entry point. The argument is a symbol name:
2508 There are several ways to set the entry point. The linker will set the
2509 entry point by trying each of the following methods in order, and
2510 stopping when one of them succeeds:
2513 the @samp{-e} @var{entry} command-line option;
2515 the @code{ENTRY(@var{symbol})} command in a linker script;
2517 the value of the symbol @code{start}, if defined;
2519 the address of the first byte of the @samp{.text} section, if present;
2521 The address @code{0}.
2525 @subsection Commands Dealing with Files
2526 @cindex linker script file commands
2527 Several linker script commands deal with files.
2530 @item INCLUDE @var{filename}
2531 @kindex INCLUDE @var{filename}
2532 @cindex including a linker script
2533 Include the linker script @var{filename} at this point. The file will
2534 be searched for in the current directory, and in any directory specified
2535 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
2538 @item INPUT(@var{file}, @var{file}, @dots{})
2539 @itemx INPUT(@var{file} @var{file} @dots{})
2540 @kindex INPUT(@var{files})
2541 @cindex input files in linker scripts
2542 @cindex input object files in linker scripts
2543 @cindex linker script input object files
2544 The @code{INPUT} command directs the linker to include the named files
2545 in the link, as though they were named on the command line.
2547 For example, if you always want to include @file{subr.o} any time you do
2548 a link, but you can't be bothered to put it on every link command line,
2549 then you can put @samp{INPUT (subr.o)} in your linker script.
2551 In fact, if you like, you can list all of your input files in the linker
2552 script, and then invoke the linker with nothing but a @samp{-T} option.
2554 In case a @dfn{sysroot prefix} is configured, and the filename starts
2555 with the @samp{/} character, and the script being processed was
2556 located inside the @dfn{sysroot prefix}, the filename will be looked
2557 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
2558 open the file in the current directory. If it is not found, the
2559 linker will search through the archive library search path. See the
2560 description of @samp{-L} in @ref{Options,,Command Line Options}.
2562 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
2563 name to @code{lib@var{file}.a}, as with the command line argument
2566 When you use the @code{INPUT} command in an implicit linker script, the
2567 files will be included in the link at the point at which the linker
2568 script file is included. This can affect archive searching.
2570 @item GROUP(@var{file}, @var{file}, @dots{})
2571 @itemx GROUP(@var{file} @var{file} @dots{})
2572 @kindex GROUP(@var{files})
2573 @cindex grouping input files
2574 The @code{GROUP} command is like @code{INPUT}, except that the named
2575 files should all be archives, and they are searched repeatedly until no
2576 new undefined references are created. See the description of @samp{-(}
2577 in @ref{Options,,Command Line Options}.
2579 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
2580 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
2581 @kindex AS_NEEDED(@var{files})
2582 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
2583 commands, among other filenames. The files listed will be handled
2584 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
2585 with the exception of ELF shared libraries, that will be added only
2586 when they are actually needed. This construct essentially enables
2587 @option{--as-needed} option for all the files listed inside of it
2588 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
2591 @item OUTPUT(@var{filename})
2592 @kindex OUTPUT(@var{filename})
2593 @cindex output file name in linker scripot
2594 The @code{OUTPUT} command names the output file. Using
2595 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
2596 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
2597 Line Options}). If both are used, the command line option takes
2600 You can use the @code{OUTPUT} command to define a default name for the
2601 output file other than the usual default of @file{a.out}.
2603 @item SEARCH_DIR(@var{path})
2604 @kindex SEARCH_DIR(@var{path})
2605 @cindex library search path in linker script
2606 @cindex archive search path in linker script
2607 @cindex search path in linker script
2608 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
2609 @command{ld} looks for archive libraries. Using
2610 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
2611 on the command line (@pxref{Options,,Command Line Options}). If both
2612 are used, then the linker will search both paths. Paths specified using
2613 the command line option are searched first.
2615 @item STARTUP(@var{filename})
2616 @kindex STARTUP(@var{filename})
2617 @cindex first input file
2618 The @code{STARTUP} command is just like the @code{INPUT} command, except
2619 that @var{filename} will become the first input file to be linked, as
2620 though it were specified first on the command line. This may be useful
2621 when using a system in which the entry point is always the start of the
2625 @ifclear SingleFormat
2626 @node Format Commands
2627 @subsection Commands Dealing with Object File Formats
2628 A couple of linker script commands deal with object file formats.
2631 @item OUTPUT_FORMAT(@var{bfdname})
2632 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
2633 @kindex OUTPUT_FORMAT(@var{bfdname})
2634 @cindex output file format in linker script
2635 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
2636 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
2637 exactly like using @samp{--oformat @var{bfdname}} on the command line
2638 (@pxref{Options,,Command Line Options}). If both are used, the command
2639 line option takes precedence.
2641 You can use @code{OUTPUT_FORMAT} with three arguments to use different
2642 formats based on the @samp{-EB} and @samp{-EL} command line options.
2643 This permits the linker script to set the output format based on the
2646 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
2647 will be the first argument, @var{default}. If @samp{-EB} is used, the
2648 output format will be the second argument, @var{big}. If @samp{-EL} is
2649 used, the output format will be the third argument, @var{little}.
2651 For example, the default linker script for the MIPS ELF target uses this
2654 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
2656 This says that the default format for the output file is
2657 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
2658 option, the output file will be created in the @samp{elf32-littlemips}
2661 @item TARGET(@var{bfdname})
2662 @kindex TARGET(@var{bfdname})
2663 @cindex input file format in linker script
2664 The @code{TARGET} command names the BFD format to use when reading input
2665 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
2666 This command is like using @samp{-b @var{bfdname}} on the command line
2667 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
2668 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
2669 command is also used to set the format for the output file. @xref{BFD}.
2673 @node Miscellaneous Commands
2674 @subsection Other Linker Script Commands
2675 There are a few other linker scripts commands.
2678 @item ASSERT(@var{exp}, @var{message})
2680 @cindex assertion in linker script
2681 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
2682 with an error code, and print @var{message}.
2684 @item EXTERN(@var{symbol} @var{symbol} @dots{})
2686 @cindex undefined symbol in linker script
2687 Force @var{symbol} to be entered in the output file as an undefined
2688 symbol. Doing this may, for example, trigger linking of additional
2689 modules from standard libraries. You may list several @var{symbol}s for
2690 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
2691 command has the same effect as the @samp{-u} command-line option.
2693 @item FORCE_COMMON_ALLOCATION
2694 @kindex FORCE_COMMON_ALLOCATION
2695 @cindex common allocation in linker script
2696 This command has the same effect as the @samp{-d} command-line option:
2697 to make @command{ld} assign space to common symbols even if a relocatable
2698 output file is specified (@samp{-r}).
2700 @item INHIBIT_COMMON_ALLOCATION
2701 @kindex INHIBIT_COMMON_ALLOCATION
2702 @cindex common allocation in linker script
2703 This command has the same effect as the @samp{--no-define-common}
2704 command-line option: to make @code{ld} omit the assignment of addresses
2705 to common symbols even for a non-relocatable output file.
2707 @item NOCROSSREFS(@var{section} @var{section} @dots{})
2708 @kindex NOCROSSREFS(@var{sections})
2709 @cindex cross references
2710 This command may be used to tell @command{ld} to issue an error about any
2711 references among certain output sections.
2713 In certain types of programs, particularly on embedded systems when
2714 using overlays, when one section is loaded into memory, another section
2715 will not be. Any direct references between the two sections would be
2716 errors. For example, it would be an error if code in one section called
2717 a function defined in the other section.
2719 The @code{NOCROSSREFS} command takes a list of output section names. If
2720 @command{ld} detects any cross references between the sections, it reports
2721 an error and returns a non-zero exit status. Note that the
2722 @code{NOCROSSREFS} command uses output section names, not input section
2725 @ifclear SingleFormat
2726 @item OUTPUT_ARCH(@var{bfdarch})
2727 @kindex OUTPUT_ARCH(@var{bfdarch})
2728 @cindex machine architecture
2729 @cindex architecture
2730 Specify a particular output machine architecture. The argument is one
2731 of the names used by the BFD library (@pxref{BFD}). You can see the
2732 architecture of an object file by using the @code{objdump} program with
2733 the @samp{-f} option.
2738 @section Assigning Values to Symbols
2739 @cindex assignment in scripts
2740 @cindex symbol definition, scripts
2741 @cindex variables, defining
2742 You may assign a value to a symbol in a linker script. This will define
2743 the symbol as a global symbol.
2746 * Simple Assignments:: Simple Assignments
2750 @node Simple Assignments
2751 @subsection Simple Assignments
2753 You may assign to a symbol using any of the C assignment operators:
2756 @item @var{symbol} = @var{expression} ;
2757 @itemx @var{symbol} += @var{expression} ;
2758 @itemx @var{symbol} -= @var{expression} ;
2759 @itemx @var{symbol} *= @var{expression} ;
2760 @itemx @var{symbol} /= @var{expression} ;
2761 @itemx @var{symbol} <<= @var{expression} ;
2762 @itemx @var{symbol} >>= @var{expression} ;
2763 @itemx @var{symbol} &= @var{expression} ;
2764 @itemx @var{symbol} |= @var{expression} ;
2767 The first case will define @var{symbol} to the value of
2768 @var{expression}. In the other cases, @var{symbol} must already be
2769 defined, and the value will be adjusted accordingly.
2771 The special symbol name @samp{.} indicates the location counter. You
2772 may only use this within a @code{SECTIONS} command.
2774 The semicolon after @var{expression} is required.
2776 Expressions are defined below; see @ref{Expressions}.
2778 You may write symbol assignments as commands in their own right, or as
2779 statements within a @code{SECTIONS} command, or as part of an output
2780 section description in a @code{SECTIONS} command.
2782 The section of the symbol will be set from the section of the
2783 expression; for more information, see @ref{Expression Section}.
2785 Here is an example showing the three different places that symbol
2786 assignments may be used:
2797 _bdata = (. + 3) & ~ 3;
2798 .data : @{ *(.data) @}
2802 In this example, the symbol @samp{floating_point} will be defined as
2803 zero. The symbol @samp{_etext} will be defined as the address following
2804 the last @samp{.text} input section. The symbol @samp{_bdata} will be
2805 defined as the address following the @samp{.text} output section aligned
2806 upward to a 4 byte boundary.
2811 In some cases, it is desirable for a linker script to define a symbol
2812 only if it is referenced and is not defined by any object included in
2813 the link. For example, traditional linkers defined the symbol
2814 @samp{etext}. However, ANSI C requires that the user be able to use
2815 @samp{etext} as a function name without encountering an error. The
2816 @code{PROVIDE} keyword may be used to define a symbol, such as
2817 @samp{etext}, only if it is referenced but not defined. The syntax is
2818 @code{PROVIDE(@var{symbol} = @var{expression})}.
2820 Here is an example of using @code{PROVIDE} to define @samp{etext}:
2833 In this example, if the program defines @samp{_etext} (with a leading
2834 underscore), the linker will give a multiple definition error. If, on
2835 the other hand, the program defines @samp{etext} (with no leading
2836 underscore), the linker will silently use the definition in the program.
2837 If the program references @samp{etext} but does not define it, the
2838 linker will use the definition in the linker script.
2841 @section SECTIONS Command
2843 The @code{SECTIONS} command tells the linker how to map input sections
2844 into output sections, and how to place the output sections in memory.
2846 The format of the @code{SECTIONS} command is:
2850 @var{sections-command}
2851 @var{sections-command}
2856 Each @var{sections-command} may of be one of the following:
2860 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
2862 a symbol assignment (@pxref{Assignments})
2864 an output section description
2866 an overlay description
2869 The @code{ENTRY} command and symbol assignments are permitted inside the
2870 @code{SECTIONS} command for convenience in using the location counter in
2871 those commands. This can also make the linker script easier to
2872 understand because you can use those commands at meaningful points in
2873 the layout of the output file.
2875 Output section descriptions and overlay descriptions are described
2878 If you do not use a @code{SECTIONS} command in your linker script, the
2879 linker will place each input section into an identically named output
2880 section in the order that the sections are first encountered in the
2881 input files. If all input sections are present in the first file, for
2882 example, the order of sections in the output file will match the order
2883 in the first input file. The first section will be at address zero.
2886 * Output Section Description:: Output section description
2887 * Output Section Name:: Output section name
2888 * Output Section Address:: Output section address
2889 * Input Section:: Input section description
2890 * Output Section Data:: Output section data
2891 * Output Section Keywords:: Output section keywords
2892 * Output Section Discarding:: Output section discarding
2893 * Output Section Attributes:: Output section attributes
2894 * Overlay Description:: Overlay description
2897 @node Output Section Description
2898 @subsection Output Section Description
2899 The full description of an output section looks like this:
2902 @var{section} [@var{address}] [(@var{type})] :
2903 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
2905 @var{output-section-command}
2906 @var{output-section-command}
2908 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
2912 Most output sections do not use most of the optional section attributes.
2914 The whitespace around @var{section} is required, so that the section
2915 name is unambiguous. The colon and the curly braces are also required.
2916 The line breaks and other white space are optional.
2918 Each @var{output-section-command} may be one of the following:
2922 a symbol assignment (@pxref{Assignments})
2924 an input section description (@pxref{Input Section})
2926 data values to include directly (@pxref{Output Section Data})
2928 a special output section keyword (@pxref{Output Section Keywords})
2931 @node Output Section Name
2932 @subsection Output Section Name
2933 @cindex name, section
2934 @cindex section name
2935 The name of the output section is @var{section}. @var{section} must
2936 meet the constraints of your output format. In formats which only
2937 support a limited number of sections, such as @code{a.out}, the name
2938 must be one of the names supported by the format (@code{a.out}, for
2939 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
2940 output format supports any number of sections, but with numbers and not
2941 names (as is the case for Oasys), the name should be supplied as a
2942 quoted numeric string. A section name may consist of any sequence of
2943 characters, but a name which contains any unusual characters such as
2944 commas must be quoted.
2946 The output section name @samp{/DISCARD/} is special; @ref{Output Section
2949 @node Output Section Address
2950 @subsection Output Section Address
2951 @cindex address, section
2952 @cindex section address
2953 The @var{address} is an expression for the VMA (the virtual memory
2954 address) of the output section. If you do not provide @var{address},
2955 the linker will set it based on @var{region} if present, or otherwise
2956 based on the current value of the location counter.
2958 If you provide @var{address}, the address of the output section will be
2959 set to precisely that. If you provide neither @var{address} nor
2960 @var{region}, then the address of the output section will be set to the
2961 current value of the location counter aligned to the alignment
2962 requirements of the output section. The alignment requirement of the
2963 output section is the strictest alignment of any input section contained
2964 within the output section.
2968 .text . : @{ *(.text) @}
2973 .text : @{ *(.text) @}
2976 are subtly different. The first will set the address of the
2977 @samp{.text} output section to the current value of the location
2978 counter. The second will set it to the current value of the location
2979 counter aligned to the strictest alignment of a @samp{.text} input
2982 The @var{address} may be an arbitrary expression; @ref{Expressions}.
2983 For example, if you want to align the section on a 0x10 byte boundary,
2984 so that the lowest four bits of the section address are zero, you could
2985 do something like this:
2987 .text ALIGN(0x10) : @{ *(.text) @}
2990 This works because @code{ALIGN} returns the current location counter
2991 aligned upward to the specified value.
2993 Specifying @var{address} for a section will change the value of the
2997 @subsection Input Section Description
2998 @cindex input sections
2999 @cindex mapping input sections to output sections
3000 The most common output section command is an input section description.
3002 The input section description is the most basic linker script operation.
3003 You use output sections to tell the linker how to lay out your program
3004 in memory. You use input section descriptions to tell the linker how to
3005 map the input files into your memory layout.
3008 * Input Section Basics:: Input section basics
3009 * Input Section Wildcards:: Input section wildcard patterns
3010 * Input Section Common:: Input section for common symbols
3011 * Input Section Keep:: Input section and garbage collection
3012 * Input Section Example:: Input section example
3015 @node Input Section Basics
3016 @subsubsection Input Section Basics
3017 @cindex input section basics
3018 An input section description consists of a file name optionally followed
3019 by a list of section names in parentheses.
3021 The file name and the section name may be wildcard patterns, which we
3022 describe further below (@pxref{Input Section Wildcards}).
3024 The most common input section description is to include all input
3025 sections with a particular name in the output section. For example, to
3026 include all input @samp{.text} sections, you would write:
3031 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
3032 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
3033 match all files except the ones specified in the EXCLUDE_FILE list. For
3036 (*(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors))
3038 will cause all .ctors sections from all files except @file{crtend.o} and
3039 @file{otherfile.o} to be included.
3041 There are two ways to include more than one section:
3047 The difference between these is the order in which the @samp{.text} and
3048 @samp{.rdata} input sections will appear in the output section. In the
3049 first example, they will be intermingled, appearing in the same order as
3050 they are found in the linker input. In the second example, all
3051 @samp{.text} input sections will appear first, followed by all
3052 @samp{.rdata} input sections.
3054 You can specify a file name to include sections from a particular file.
3055 You would do this if one or more of your files contain special data that
3056 needs to be at a particular location in memory. For example:
3061 If you use a file name without a list of sections, then all sections in
3062 the input file will be included in the output section. This is not
3063 commonly done, but it may by useful on occasion. For example:
3068 When you use a file name which does not contain any wild card
3069 characters, the linker will first see if you also specified the file
3070 name on the linker command line or in an @code{INPUT} command. If you
3071 did not, the linker will attempt to open the file as an input file, as
3072 though it appeared on the command line. Note that this differs from an
3073 @code{INPUT} command, because the linker will not search for the file in
3074 the archive search path.
3076 @node Input Section Wildcards
3077 @subsubsection Input Section Wildcard Patterns
3078 @cindex input section wildcards
3079 @cindex wildcard file name patterns
3080 @cindex file name wildcard patterns
3081 @cindex section name wildcard patterns
3082 In an input section description, either the file name or the section
3083 name or both may be wildcard patterns.
3085 The file name of @samp{*} seen in many examples is a simple wildcard
3086 pattern for the file name.
3088 The wildcard patterns are like those used by the Unix shell.
3092 matches any number of characters
3094 matches any single character
3096 matches a single instance of any of the @var{chars}; the @samp{-}
3097 character may be used to specify a range of characters, as in
3098 @samp{[a-z]} to match any lower case letter
3100 quotes the following character
3103 When a file name is matched with a wildcard, the wildcard characters
3104 will not match a @samp{/} character (used to separate directory names on
3105 Unix). A pattern consisting of a single @samp{*} character is an
3106 exception; it will always match any file name, whether it contains a
3107 @samp{/} or not. In a section name, the wildcard characters will match
3108 a @samp{/} character.
3110 File name wildcard patterns only match files which are explicitly
3111 specified on the command line or in an @code{INPUT} command. The linker
3112 does not search directories to expand wildcards.
3114 If a file name matches more than one wildcard pattern, or if a file name
3115 appears explicitly and is also matched by a wildcard pattern, the linker
3116 will use the first match in the linker script. For example, this
3117 sequence of input section descriptions is probably in error, because the
3118 @file{data.o} rule will not be used:
3120 .data : @{ *(.data) @}
3121 .data1 : @{ data.o(.data) @}
3124 @cindex SORT_BY_NAME
3125 Normally, the linker will place files and sections matched by wildcards
3126 in the order in which they are seen during the link. You can change
3127 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
3128 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
3129 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
3130 into ascending order by name before placing them in the output file.
3132 @cindex SORT_BY_ALIGNMENT
3133 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
3134 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
3135 ascending order by alignment before placing them in the output file.
3138 @code{SORT} is an alias for @code{SORT_BY_NAME}.
3140 When there are nested section sorting commands in linker script, there
3141 can be at most 1 level of nesting for section sorting commands.
3145 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
3146 It will sort the input sections by name first, then by alignment if 2
3147 sections have the same name.
3149 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
3150 It will sort the input sections by alignment first, then by name if 2
3151 sections have the same alignment.
3153 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
3154 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
3156 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
3157 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
3159 All other nested section sorting commands are invalid.
3162 When both command line section sorting option and linker script
3163 section sorting command are used, section sorting command always
3164 takes precedence over the command line option.
3166 If the section sorting command in linker script isn't nested, the
3167 command line option will make the section sorting command to be
3168 treated as nested sorting command.
3172 @code{SORT_BY_NAME} (wildcard section pattern ) with
3173 @option{--sort-sections alignment} is equivalent to
3174 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
3176 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
3177 @option{--sort-section name} is equivalent to
3178 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
3181 If the section sorting command in linker script is nested, the
3182 command line option will be ignored.
3184 If you ever get confused about where input sections are going, use the
3185 @samp{-M} linker option to generate a map file. The map file shows
3186 precisely how input sections are mapped to output sections.
3188 This example shows how wildcard patterns might be used to partition
3189 files. This linker script directs the linker to place all @samp{.text}
3190 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
3191 The linker will place the @samp{.data} section from all files beginning
3192 with an upper case character in @samp{.DATA}; for all other files, the
3193 linker will place the @samp{.data} section in @samp{.data}.
3197 .text : @{ *(.text) @}
3198 .DATA : @{ [A-Z]*(.data) @}
3199 .data : @{ *(.data) @}
3200 .bss : @{ *(.bss) @}
3205 @node Input Section Common
3206 @subsubsection Input Section for Common Symbols
3207 @cindex common symbol placement
3208 @cindex uninitialized data placement
3209 A special notation is needed for common symbols, because in many object
3210 file formats common symbols do not have a particular input section. The
3211 linker treats common symbols as though they are in an input section
3212 named @samp{COMMON}.
3214 You may use file names with the @samp{COMMON} section just as with any
3215 other input sections. You can use this to place common symbols from a
3216 particular input file in one section while common symbols from other
3217 input files are placed in another section.
3219 In most cases, common symbols in input files will be placed in the
3220 @samp{.bss} section in the output file. For example:
3222 .bss @{ *(.bss) *(COMMON) @}
3225 @cindex scommon section
3226 @cindex small common symbols
3227 Some object file formats have more than one type of common symbol. For
3228 example, the MIPS ELF object file format distinguishes standard common
3229 symbols and small common symbols. In this case, the linker will use a
3230 different special section name for other types of common symbols. In
3231 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
3232 symbols and @samp{.scommon} for small common symbols. This permits you
3233 to map the different types of common symbols into memory at different
3237 You will sometimes see @samp{[COMMON]} in old linker scripts. This
3238 notation is now considered obsolete. It is equivalent to
3241 @node Input Section Keep
3242 @subsubsection Input Section and Garbage Collection
3244 @cindex garbage collection
3245 When link-time garbage collection is in use (@samp{--gc-sections}),
3246 it is often useful to mark sections that should not be eliminated.
3247 This is accomplished by surrounding an input section's wildcard entry
3248 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
3249 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
3251 @node Input Section Example
3252 @subsubsection Input Section Example
3253 The following example is a complete linker script. It tells the linker
3254 to read all of the sections from file @file{all.o} and place them at the
3255 start of output section @samp{outputa} which starts at location
3256 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
3257 follows immediately, in the same output section. All of section
3258 @samp{.input2} from @file{foo.o} goes into output section
3259 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
3260 All of the remaining @samp{.input1} and @samp{.input2} sections from any
3261 files are written to output section @samp{outputc}.
3289 @node Output Section Data
3290 @subsection Output Section Data
3292 @cindex section data
3293 @cindex output section data
3294 @kindex BYTE(@var{expression})
3295 @kindex SHORT(@var{expression})
3296 @kindex LONG(@var{expression})
3297 @kindex QUAD(@var{expression})
3298 @kindex SQUAD(@var{expression})
3299 You can include explicit bytes of data in an output section by using
3300 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
3301 an output section command. Each keyword is followed by an expression in
3302 parentheses providing the value to store (@pxref{Expressions}). The
3303 value of the expression is stored at the current value of the location
3306 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
3307 store one, two, four, and eight bytes (respectively). After storing the
3308 bytes, the location counter is incremented by the number of bytes
3311 For example, this will store the byte 1 followed by the four byte value
3312 of the symbol @samp{addr}:
3318 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
3319 same; they both store an 8 byte, or 64 bit, value. When both host and
3320 target are 32 bits, an expression is computed as 32 bits. In this case
3321 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
3322 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
3324 If the object file format of the output file has an explicit endianness,
3325 which is the normal case, the value will be stored in that endianness.
3326 When the object file format does not have an explicit endianness, as is
3327 true of, for example, S-records, the value will be stored in the
3328 endianness of the first input object file.
3330 Note---these commands only work inside a section description and not
3331 between them, so the following will produce an error from the linker:
3333 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
3335 whereas this will work:
3337 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
3340 @kindex FILL(@var{expression})
3341 @cindex holes, filling
3342 @cindex unspecified memory
3343 You may use the @code{FILL} command to set the fill pattern for the
3344 current section. It is followed by an expression in parentheses. Any
3345 otherwise unspecified regions of memory within the section (for example,
3346 gaps left due to the required alignment of input sections) are filled
3347 with the value of the expression, repeated as
3348 necessary. A @code{FILL} statement covers memory locations after the
3349 point at which it occurs in the section definition; by including more
3350 than one @code{FILL} statement, you can have different fill patterns in
3351 different parts of an output section.
3353 This example shows how to fill unspecified regions of memory with the
3359 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
3360 section attribute, but it only affects the
3361 part of the section following the @code{FILL} command, rather than the
3362 entire section. If both are used, the @code{FILL} command takes
3363 precedence. @xref{Output Section Fill}, for details on the fill
3366 @node Output Section Keywords
3367 @subsection Output Section Keywords
3368 There are a couple of keywords which can appear as output section
3372 @kindex CREATE_OBJECT_SYMBOLS
3373 @cindex input filename symbols
3374 @cindex filename symbols
3375 @item CREATE_OBJECT_SYMBOLS
3376 The command tells the linker to create a symbol for each input file.
3377 The name of each symbol will be the name of the corresponding input
3378 file. The section of each symbol will be the output section in which
3379 the @code{CREATE_OBJECT_SYMBOLS} command appears.
3381 This is conventional for the a.out object file format. It is not
3382 normally used for any other object file format.
3384 @kindex CONSTRUCTORS
3385 @cindex C++ constructors, arranging in link
3386 @cindex constructors, arranging in link
3388 When linking using the a.out object file format, the linker uses an
3389 unusual set construct to support C++ global constructors and
3390 destructors. When linking object file formats which do not support
3391 arbitrary sections, such as ECOFF and XCOFF, the linker will
3392 automatically recognize C++ global constructors and destructors by name.
3393 For these object file formats, the @code{CONSTRUCTORS} command tells the
3394 linker to place constructor information in the output section where the
3395 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
3396 ignored for other object file formats.
3398 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
3399 constructors, and the symbol @w{@code{__DTOR_LIST}} marks the end. The
3400 first word in the list is the number of entries, followed by the address
3401 of each constructor or destructor, followed by a zero word. The
3402 compiler must arrange to actually run the code. For these object file
3403 formats @sc{gnu} C++ normally calls constructors from a subroutine
3404 @code{__main}; a call to @code{__main} is automatically inserted into
3405 the startup code for @code{main}. @sc{gnu} C++ normally runs
3406 destructors either by using @code{atexit}, or directly from the function
3409 For object file formats such as @code{COFF} or @code{ELF} which support
3410 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
3411 addresses of global constructors and destructors into the @code{.ctors}
3412 and @code{.dtors} sections. Placing the following sequence into your
3413 linker script will build the sort of table which the @sc{gnu} C++
3414 runtime code expects to see.
3418 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
3423 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
3429 If you are using the @sc{gnu} C++ support for initialization priority,
3430 which provides some control over the order in which global constructors
3431 are run, you must sort the constructors at link time to ensure that they
3432 are executed in the correct order. When using the @code{CONSTRUCTORS}
3433 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
3434 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
3435 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
3438 Normally the compiler and linker will handle these issues automatically,
3439 and you will not need to concern yourself with them. However, you may
3440 need to consider this if you are using C++ and writing your own linker
3445 @node Output Section Discarding
3446 @subsection Output Section Discarding
3447 @cindex discarding sections
3448 @cindex sections, discarding
3449 @cindex removing sections
3450 The linker will not create output section which do not have any
3451 contents. This is for convenience when referring to input sections that
3452 may or may not be present in any of the input files. For example:
3457 will only create a @samp{.foo} section in the output file if there is a
3458 @samp{.foo} section in at least one input file.
3460 If you use anything other than an input section description as an output
3461 section command, such as a symbol assignment, then the output section
3462 will always be created, even if there are no matching input sections.
3465 The special output section name @samp{/DISCARD/} may be used to discard
3466 input sections. Any input sections which are assigned to an output
3467 section named @samp{/DISCARD/} are not included in the output file.
3469 @node Output Section Attributes
3470 @subsection Output Section Attributes
3471 @cindex output section attributes
3472 We showed above that the full description of an output section looked
3476 @var{section} [@var{address}] [(@var{type})] :
3477 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
3479 @var{output-section-command}
3480 @var{output-section-command}
3482 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3485 We've already described @var{section}, @var{address}, and
3486 @var{output-section-command}. In this section we will describe the
3487 remaining section attributes.
3490 * Output Section Type:: Output section type
3491 * Output Section LMA:: Output section LMA
3492 * Forced Input Alignment:: Forced Input Alignment
3493 * Output Section Region:: Output section region
3494 * Output Section Phdr:: Output section phdr
3495 * Output Section Fill:: Output section fill
3498 @node Output Section Type
3499 @subsubsection Output Section Type
3500 Each output section may have a type. The type is a keyword in
3501 parentheses. The following types are defined:
3505 The section should be marked as not loadable, so that it will not be
3506 loaded into memory when the program is run.
3511 These type names are supported for backward compatibility, and are
3512 rarely used. They all have the same effect: the section should be
3513 marked as not allocatable, so that no memory is allocated for the
3514 section when the program is run.
3518 @cindex prevent unnecessary loading
3519 @cindex loading, preventing
3520 The linker normally sets the attributes of an output section based on
3521 the input sections which map into it. You can override this by using
3522 the section type. For example, in the script sample below, the
3523 @samp{ROM} section is addressed at memory location @samp{0} and does not
3524 need to be loaded when the program is run. The contents of the
3525 @samp{ROM} section will appear in the linker output file as usual.
3529 ROM 0 (NOLOAD) : @{ @dots{} @}
3535 @node Output Section LMA
3536 @subsubsection Output Section LMA
3537 @kindex AT>@var{lma_region}
3538 @kindex AT(@var{lma})
3539 @cindex load address
3540 @cindex section load address
3541 Every section has a virtual address (VMA) and a load address (LMA); see
3542 @ref{Basic Script Concepts}. The address expression which may appear in
3543 an output section description sets the VMA (@pxref{Output Section
3546 The linker will normally set the LMA equal to the VMA. You can change
3547 that by using the @code{AT} keyword. The expression @var{lma} that
3548 follows the @code{AT} keyword specifies the load address of the
3551 Alternatively, with @samp{AT>@var{lma_region}} expression, you may
3552 specify a memory region for the section's load address. @xref{MEMORY}.
3553 Note that if the section has not had a VMA assigned to it then the
3554 linker will use the @var{lma_region} as the VMA region as well.
3555 @xref{Output Section Region}.
3557 @cindex ROM initialized data
3558 @cindex initialized data in ROM
3559 This feature is designed to make it easy to build a ROM image. For
3560 example, the following linker script creates three output sections: one
3561 called @samp{.text}, which starts at @code{0x1000}, one called
3562 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
3563 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
3564 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
3565 defined with the value @code{0x2000}, which shows that the location
3566 counter holds the VMA value, not the LMA value.
3572 .text 0x1000 : @{ *(.text) _etext = . ; @}
3574 AT ( ADDR (.text) + SIZEOF (.text) )
3575 @{ _data = . ; *(.data); _edata = . ; @}
3577 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
3582 The run-time initialization code for use with a program generated with
3583 this linker script would include something like the following, to copy
3584 the initialized data from the ROM image to its runtime address. Notice
3585 how this code takes advantage of the symbols defined by the linker
3590 extern char _etext, _data, _edata, _bstart, _bend;
3591 char *src = &_etext;
3594 /* ROM has data at end of text; copy it. */
3595 while (dst < &_edata) @{
3600 for (dst = &_bstart; dst< &_bend; dst++)
3605 @node Forced Input Alignment
3606 @subsubsection Forced Input Alignment
3607 @kindex SUBALIGN(@var{subsection_align})
3608 @cindex forcing input section alignment
3609 @cindex input section alignment
3610 You can force input section alignment within an output section by using
3611 SUBALIGN. The value specified overrides any alignment given by input
3612 sections, whether larger or smaller.
3614 @node Output Section Region
3615 @subsubsection Output Section Region
3616 @kindex >@var{region}
3617 @cindex section, assigning to memory region
3618 @cindex memory regions and sections
3619 You can assign a section to a previously defined region of memory by
3620 using @samp{>@var{region}}. @xref{MEMORY}.
3622 Here is a simple example:
3625 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
3626 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
3630 @node Output Section Phdr
3631 @subsubsection Output Section Phdr
3633 @cindex section, assigning to program header
3634 @cindex program headers and sections
3635 You can assign a section to a previously defined program segment by
3636 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
3637 one or more segments, then all subsequent allocated sections will be
3638 assigned to those segments as well, unless they use an explicitly
3639 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
3640 linker to not put the section in any segment at all.
3642 Here is a simple example:
3645 PHDRS @{ text PT_LOAD ; @}
3646 SECTIONS @{ .text : @{ *(.text) @} :text @}
3650 @node Output Section Fill
3651 @subsubsection Output Section Fill
3652 @kindex =@var{fillexp}
3653 @cindex section fill pattern
3654 @cindex fill pattern, entire section
3655 You can set the fill pattern for an entire section by using
3656 @samp{=@var{fillexp}}. @var{fillexp} is an expression
3657 (@pxref{Expressions}). Any otherwise unspecified regions of memory
3658 within the output section (for example, gaps left due to the required
3659 alignment of input sections) will be filled with the value, repeated as
3660 necessary. If the fill expression is a simple hex number, ie. a string
3661 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
3662 an arbitrarily long sequence of hex digits can be used to specify the
3663 fill pattern; Leading zeros become part of the pattern too. For all
3664 other cases, including extra parentheses or a unary @code{+}, the fill
3665 pattern is the four least significant bytes of the value of the
3666 expression. In all cases, the number is big-endian.
3668 You can also change the fill value with a @code{FILL} command in the
3669 output section commands; (@pxref{Output Section Data}).
3671 Here is a simple example:
3674 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
3678 @node Overlay Description
3679 @subsection Overlay Description
3682 An overlay description provides an easy way to describe sections which
3683 are to be loaded as part of a single memory image but are to be run at
3684 the same memory address. At run time, some sort of overlay manager will
3685 copy the overlaid sections in and out of the runtime memory address as
3686 required, perhaps by simply manipulating addressing bits. This approach
3687 can be useful, for example, when a certain region of memory is faster
3690 Overlays are described using the @code{OVERLAY} command. The
3691 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
3692 output section description. The full syntax of the @code{OVERLAY}
3693 command is as follows:
3696 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
3700 @var{output-section-command}
3701 @var{output-section-command}
3703 @} [:@var{phdr}@dots{}] [=@var{fill}]
3706 @var{output-section-command}
3707 @var{output-section-command}
3709 @} [:@var{phdr}@dots{}] [=@var{fill}]
3711 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
3715 Everything is optional except @code{OVERLAY} (a keyword), and each
3716 section must have a name (@var{secname1} and @var{secname2} above). The
3717 section definitions within the @code{OVERLAY} construct are identical to
3718 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
3719 except that no addresses and no memory regions may be defined for
3720 sections within an @code{OVERLAY}.
3722 The sections are all defined with the same starting address. The load
3723 addresses of the sections are arranged such that they are consecutive in
3724 memory starting at the load address used for the @code{OVERLAY} as a
3725 whole (as with normal section definitions, the load address is optional,
3726 and defaults to the start address; the start address is also optional,
3727 and defaults to the current value of the location counter).
3729 If the @code{NOCROSSREFS} keyword is used, and there any references
3730 among the sections, the linker will report an error. Since the sections
3731 all run at the same address, it normally does not make sense for one
3732 section to refer directly to another. @xref{Miscellaneous Commands,
3735 For each section within the @code{OVERLAY}, the linker automatically
3736 defines two symbols. The symbol @code{__load_start_@var{secname}} is
3737 defined as the starting load address of the section. The symbol
3738 @code{__load_stop_@var{secname}} is defined as the final load address of
3739 the section. Any characters within @var{secname} which are not legal
3740 within C identifiers are removed. C (or assembler) code may use these
3741 symbols to move the overlaid sections around as necessary.
3743 At the end of the overlay, the value of the location counter is set to
3744 the start address of the overlay plus the size of the largest section.
3746 Here is an example. Remember that this would appear inside a
3747 @code{SECTIONS} construct.
3750 OVERLAY 0x1000 : AT (0x4000)
3752 .text0 @{ o1/*.o(.text) @}
3753 .text1 @{ o2/*.o(.text) @}
3758 This will define both @samp{.text0} and @samp{.text1} to start at
3759 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
3760 @samp{.text1} will be loaded immediately after @samp{.text0}. The
3761 following symbols will be defined: @code{__load_start_text0},
3762 @code{__load_stop_text0}, @code{__load_start_text1},
3763 @code{__load_stop_text1}.
3765 C code to copy overlay @code{.text1} into the overlay area might look
3770 extern char __load_start_text1, __load_stop_text1;
3771 memcpy ((char *) 0x1000, &__load_start_text1,
3772 &__load_stop_text1 - &__load_start_text1);
3776 Note that the @code{OVERLAY} command is just syntactic sugar, since
3777 everything it does can be done using the more basic commands. The above
3778 example could have been written identically as follows.
3782 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
3783 __load_start_text0 = LOADADDR (.text0);
3784 __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0);
3785 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
3786 __load_start_text1 = LOADADDR (.text1);
3787 __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1);
3788 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
3793 @section MEMORY Command
3795 @cindex memory regions
3796 @cindex regions of memory
3797 @cindex allocating memory
3798 @cindex discontinuous memory
3799 The linker's default configuration permits allocation of all available
3800 memory. You can override this by using the @code{MEMORY} command.
3802 The @code{MEMORY} command describes the location and size of blocks of
3803 memory in the target. You can use it to describe which memory regions
3804 may be used by the linker, and which memory regions it must avoid. You
3805 can then assign sections to particular memory regions. The linker will
3806 set section addresses based on the memory regions, and will warn about
3807 regions that become too full. The linker will not shuffle sections
3808 around to fit into the available regions.
3810 A linker script may contain at most one use of the @code{MEMORY}
3811 command. However, you can define as many blocks of memory within it as
3812 you wish. The syntax is:
3817 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
3823 The @var{name} is a name used in the linker script to refer to the
3824 region. The region name has no meaning outside of the linker script.
3825 Region names are stored in a separate name space, and will not conflict
3826 with symbol names, file names, or section names. Each memory region
3827 must have a distinct name.
3829 @cindex memory region attributes
3830 The @var{attr} string is an optional list of attributes that specify
3831 whether to use a particular memory region for an input section which is
3832 not explicitly mapped in the linker script. As described in
3833 @ref{SECTIONS}, if you do not specify an output section for some input
3834 section, the linker will create an output section with the same name as
3835 the input section. If you define region attributes, the linker will use
3836 them to select the memory region for the output section that it creates.
3838 The @var{attr} string must consist only of the following characters:
3853 Invert the sense of any of the preceding attributes
3856 If a unmapped section matches any of the listed attributes other than
3857 @samp{!}, it will be placed in the memory region. The @samp{!}
3858 attribute reverses this test, so that an unmapped section will be placed
3859 in the memory region only if it does not match any of the listed
3865 The @var{origin} is an numerical expression for the start address of
3866 the memory region. The expression must evaluate to a constant and it
3867 cannot involve any symbols. The keyword @code{ORIGIN} may be
3868 abbreviated to @code{org} or @code{o} (but not, for example,
3874 The @var{len} is an expression for the size in bytes of the memory
3875 region. As with the @var{origin} expression, the expression must
3876 be numerical only and must evaluate to a constant. The keyword
3877 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
3879 In the following example, we specify that there are two memory regions
3880 available for allocation: one starting at @samp{0} for 256 kilobytes,
3881 and the other starting at @samp{0x40000000} for four megabytes. The
3882 linker will place into the @samp{rom} memory region every section which
3883 is not explicitly mapped into a memory region, and is either read-only
3884 or executable. The linker will place other sections which are not
3885 explicitly mapped into a memory region into the @samp{ram} memory
3892 rom (rx) : ORIGIN = 0, LENGTH = 256K
3893 ram (!rx) : org = 0x40000000, l = 4M
3898 Once you define a memory region, you can direct the linker to place
3899 specific output sections into that memory region by using the
3900 @samp{>@var{region}} output section attribute. For example, if you have
3901 a memory region named @samp{mem}, you would use @samp{>mem} in the
3902 output section definition. @xref{Output Section Region}. If no address
3903 was specified for the output section, the linker will set the address to
3904 the next available address within the memory region. If the combined
3905 output sections directed to a memory region are too large for the
3906 region, the linker will issue an error message.
3908 It is possible to access the origin and length of a memory in an
3909 expression via the @code{ORIGIN(@var{memory})} and
3910 @code{LENGTH(@var{memory})} functions:
3914 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
3919 @section PHDRS Command
3921 @cindex program headers
3922 @cindex ELF program headers
3923 @cindex program segments
3924 @cindex segments, ELF
3925 The ELF object file format uses @dfn{program headers}, also knows as
3926 @dfn{segments}. The program headers describe how the program should be
3927 loaded into memory. You can print them out by using the @code{objdump}
3928 program with the @samp{-p} option.
3930 When you run an ELF program on a native ELF system, the system loader
3931 reads the program headers in order to figure out how to load the
3932 program. This will only work if the program headers are set correctly.
3933 This manual does not describe the details of how the system loader
3934 interprets program headers; for more information, see the ELF ABI.
3936 The linker will create reasonable program headers by default. However,
3937 in some cases, you may need to specify the program headers more
3938 precisely. You may use the @code{PHDRS} command for this purpose. When
3939 the linker sees the @code{PHDRS} command in the linker script, it will
3940 not create any program headers other than the ones specified.
3942 The linker only pays attention to the @code{PHDRS} command when
3943 generating an ELF output file. In other cases, the linker will simply
3944 ignore @code{PHDRS}.
3946 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
3947 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
3953 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
3954 [ FLAGS ( @var{flags} ) ] ;
3959 The @var{name} is used only for reference in the @code{SECTIONS} command
3960 of the linker script. It is not put into the output file. Program
3961 header names are stored in a separate name space, and will not conflict
3962 with symbol names, file names, or section names. Each program header
3963 must have a distinct name.
3965 Certain program header types describe segments of memory which the
3966 system loader will load from the file. In the linker script, you
3967 specify the contents of these segments by placing allocatable output
3968 sections in the segments. You use the @samp{:@var{phdr}} output section
3969 attribute to place a section in a particular segment. @xref{Output
3972 It is normal to put certain sections in more than one segment. This
3973 merely implies that one segment of memory contains another. You may
3974 repeat @samp{:@var{phdr}}, using it once for each segment which should
3975 contain the section.
3977 If you place a section in one or more segments using @samp{:@var{phdr}},
3978 then the linker will place all subsequent allocatable sections which do
3979 not specify @samp{:@var{phdr}} in the same segments. This is for
3980 convenience, since generally a whole set of contiguous sections will be
3981 placed in a single segment. You can use @code{:NONE} to override the
3982 default segment and tell the linker to not put the section in any
3987 You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after
3988 the program header type to further describe the contents of the segment.
3989 The @code{FILEHDR} keyword means that the segment should include the ELF
3990 file header. The @code{PHDRS} keyword means that the segment should
3991 include the ELF program headers themselves.
3993 The @var{type} may be one of the following. The numbers indicate the
3994 value of the keyword.
3997 @item @code{PT_NULL} (0)
3998 Indicates an unused program header.
4000 @item @code{PT_LOAD} (1)
4001 Indicates that this program header describes a segment to be loaded from
4004 @item @code{PT_DYNAMIC} (2)
4005 Indicates a segment where dynamic linking information can be found.
4007 @item @code{PT_INTERP} (3)
4008 Indicates a segment where the name of the program interpreter may be
4011 @item @code{PT_NOTE} (4)
4012 Indicates a segment holding note information.
4014 @item @code{PT_SHLIB} (5)
4015 A reserved program header type, defined but not specified by the ELF
4018 @item @code{PT_PHDR} (6)
4019 Indicates a segment where the program headers may be found.
4021 @item @var{expression}
4022 An expression giving the numeric type of the program header. This may
4023 be used for types not defined above.
4026 You can specify that a segment should be loaded at a particular address
4027 in memory by using an @code{AT} expression. This is identical to the
4028 @code{AT} command used as an output section attribute (@pxref{Output
4029 Section LMA}). The @code{AT} command for a program header overrides the
4030 output section attribute.
4032 The linker will normally set the segment flags based on the sections
4033 which comprise the segment. You may use the @code{FLAGS} keyword to
4034 explicitly specify the segment flags. The value of @var{flags} must be
4035 an integer. It is used to set the @code{p_flags} field of the program
4038 Here is an example of @code{PHDRS}. This shows a typical set of program
4039 headers used on a native ELF system.
4045 headers PT_PHDR PHDRS ;
4047 text PT_LOAD FILEHDR PHDRS ;
4049 dynamic PT_DYNAMIC ;
4055 .interp : @{ *(.interp) @} :text :interp
4056 .text : @{ *(.text) @} :text
4057 .rodata : @{ *(.rodata) @} /* defaults to :text */
4059 . = . + 0x1000; /* move to a new page in memory */
4060 .data : @{ *(.data) @} :data
4061 .dynamic : @{ *(.dynamic) @} :data :dynamic
4068 @section VERSION Command
4069 @kindex VERSION @{script text@}
4070 @cindex symbol versions
4071 @cindex version script
4072 @cindex versions of symbols
4073 The linker supports symbol versions when using ELF. Symbol versions are
4074 only useful when using shared libraries. The dynamic linker can use
4075 symbol versions to select a specific version of a function when it runs
4076 a program that may have been linked against an earlier version of the
4079 You can include a version script directly in the main linker script, or
4080 you can supply the version script as an implicit linker script. You can
4081 also use the @samp{--version-script} linker option.
4083 The syntax of the @code{VERSION} command is simply
4085 VERSION @{ version-script-commands @}
4088 The format of the version script commands is identical to that used by
4089 Sun's linker in Solaris 2.5. The version script defines a tree of
4090 version nodes. You specify the node names and interdependencies in the
4091 version script. You can specify which symbols are bound to which
4092 version nodes, and you can reduce a specified set of symbols to local
4093 scope so that they are not globally visible outside of the shared
4096 The easiest way to demonstrate the version script language is with a few
4118 This example version script defines three version nodes. The first
4119 version node defined is @samp{VERS_1.1}; it has no other dependencies.
4120 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
4121 a number of symbols to local scope so that they are not visible outside
4122 of the shared library; this is done using wildcard patterns, so that any
4123 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
4124 is matched. The wildcard patterns available are the same as those used
4125 in the shell when matching filenames (also known as ``globbing'').
4127 Next, the version script defines node @samp{VERS_1.2}. This node
4128 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
4129 to the version node @samp{VERS_1.2}.
4131 Finally, the version script defines node @samp{VERS_2.0}. This node
4132 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
4133 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
4135 When the linker finds a symbol defined in a library which is not
4136 specifically bound to a version node, it will effectively bind it to an
4137 unspecified base version of the library. You can bind all otherwise
4138 unspecified symbols to a given version node by using @samp{global: *;}
4139 somewhere in the version script.
4141 The names of the version nodes have no specific meaning other than what
4142 they might suggest to the person reading them. The @samp{2.0} version
4143 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
4144 However, this would be a confusing way to write a version script.
4146 Node name can be omited, provided it is the only version node
4147 in the version script. Such version script doesn't assign any versions to
4148 symbols, only selects which symbols will be globally visible out and which
4152 @{ global: foo; bar; local: *; @};
4155 When you link an application against a shared library that has versioned
4156 symbols, the application itself knows which version of each symbol it
4157 requires, and it also knows which version nodes it needs from each
4158 shared library it is linked against. Thus at runtime, the dynamic
4159 loader can make a quick check to make sure that the libraries you have
4160 linked against do in fact supply all of the version nodes that the
4161 application will need to resolve all of the dynamic symbols. In this
4162 way it is possible for the dynamic linker to know with certainty that
4163 all external symbols that it needs will be resolvable without having to
4164 search for each symbol reference.
4166 The symbol versioning is in effect a much more sophisticated way of
4167 doing minor version checking that SunOS does. The fundamental problem
4168 that is being addressed here is that typically references to external
4169 functions are bound on an as-needed basis, and are not all bound when
4170 the application starts up. If a shared library is out of date, a
4171 required interface may be missing; when the application tries to use
4172 that interface, it may suddenly and unexpectedly fail. With symbol
4173 versioning, the user will get a warning when they start their program if
4174 the libraries being used with the application are too old.
4176 There are several GNU extensions to Sun's versioning approach. The
4177 first of these is the ability to bind a symbol to a version node in the
4178 source file where the symbol is defined instead of in the versioning
4179 script. This was done mainly to reduce the burden on the library
4180 maintainer. You can do this by putting something like:
4182 __asm__(".symver original_foo,foo@@VERS_1.1");
4185 in the C source file. This renames the function @samp{original_foo} to
4186 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
4187 The @samp{local:} directive can be used to prevent the symbol
4188 @samp{original_foo} from being exported. A @samp{.symver} directive
4189 takes precedence over a version script.
4191 The second GNU extension is to allow multiple versions of the same
4192 function to appear in a given shared library. In this way you can make
4193 an incompatible change to an interface without increasing the major
4194 version number of the shared library, while still allowing applications
4195 linked against the old interface to continue to function.
4197 To do this, you must use multiple @samp{.symver} directives in the
4198 source file. Here is an example:
4201 __asm__(".symver original_foo,foo@@");
4202 __asm__(".symver old_foo,foo@@VERS_1.1");
4203 __asm__(".symver old_foo1,foo@@VERS_1.2");
4204 __asm__(".symver new_foo,foo@@@@VERS_2.0");
4207 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
4208 unspecified base version of the symbol. The source file that contains this
4209 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
4210 @samp{old_foo1}, and @samp{new_foo}.
4212 When you have multiple definitions of a given symbol, there needs to be
4213 some way to specify a default version to which external references to
4214 this symbol will be bound. You can do this with the
4215 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
4216 declare one version of a symbol as the default in this manner; otherwise
4217 you would effectively have multiple definitions of the same symbol.
4219 If you wish to bind a reference to a specific version of the symbol
4220 within the shared library, you can use the aliases of convenience
4221 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
4222 specifically bind to an external version of the function in question.
4224 You can also specify the language in the version script:
4227 VERSION extern "lang" @{ version-script-commands @}
4230 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
4231 The linker will iterate over the list of symbols at the link time and
4232 demangle them according to @samp{lang} before matching them to the
4233 patterns specified in @samp{version-script-commands}.
4236 @section Expressions in Linker Scripts
4239 The syntax for expressions in the linker script language is identical to
4240 that of C expressions. All expressions are evaluated as integers. All
4241 expressions are evaluated in the same size, which is 32 bits if both the
4242 host and target are 32 bits, and is otherwise 64 bits.
4244 You can use and set symbol values in expressions.
4246 The linker defines several special purpose builtin functions for use in
4250 * Constants:: Constants
4251 * Symbols:: Symbol Names
4252 * Location Counter:: The Location Counter
4253 * Operators:: Operators
4254 * Evaluation:: Evaluation
4255 * Expression Section:: The Section of an Expression
4256 * Builtin Functions:: Builtin Functions
4260 @subsection Constants
4261 @cindex integer notation
4262 @cindex constants in linker scripts
4263 All constants are integers.
4265 As in C, the linker considers an integer beginning with @samp{0} to be
4266 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
4267 hexadecimal. The linker considers other integers to be decimal.
4269 @cindex scaled integers
4270 @cindex K and M integer suffixes
4271 @cindex M and K integer suffixes
4272 @cindex suffixes for integers
4273 @cindex integer suffixes
4274 In addition, you can use the suffixes @code{K} and @code{M} to scale a
4278 @c END TEXI2ROFF-KILL
4279 @code{1024} or @code{1024*1024}
4283 ${\rm 1024}$ or ${\rm 1024}^2$
4285 @c END TEXI2ROFF-KILL
4286 respectively. For example, the following all refer to the same quantity:
4294 @subsection Symbol Names
4295 @cindex symbol names
4297 @cindex quoted symbol names
4299 Unless quoted, symbol names start with a letter, underscore, or period
4300 and may include letters, digits, underscores, periods, and hyphens.
4301 Unquoted symbol names must not conflict with any keywords. You can
4302 specify a symbol which contains odd characters or has the same name as a
4303 keyword by surrounding the symbol name in double quotes:
4306 "with a space" = "also with a space" + 10;
4309 Since symbols can contain many non-alphabetic characters, it is safest
4310 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
4311 whereas @samp{A - B} is an expression involving subtraction.
4313 @node Location Counter
4314 @subsection The Location Counter
4317 @cindex location counter
4318 @cindex current output location
4319 The special linker variable @dfn{dot} @samp{.} always contains the
4320 current output location counter. Since the @code{.} always refers to a
4321 location in an output section, it may only appear in an expression
4322 within a @code{SECTIONS} command. The @code{.} symbol may appear
4323 anywhere that an ordinary symbol is allowed in an expression.
4326 Assigning a value to @code{.} will cause the location counter to be
4327 moved. This may be used to create holes in the output section. The
4328 location counter may never be moved backwards.
4344 In the previous example, the @samp{.text} section from @file{file1} is
4345 located at the beginning of the output section @samp{output}. It is
4346 followed by a 1000 byte gap. Then the @samp{.text} section from
4347 @file{file2} appears, also with a 1000 byte gap following before the
4348 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
4349 specifies what data to write in the gaps (@pxref{Output Section Fill}).
4351 @cindex dot inside sections
4352 Note: @code{.} actually refers to the byte offset from the start of the
4353 current containing object. Normally this is the @code{SECTIONS}
4354 statement, whose start address is 0, hence @code{.} can be used as an
4355 absolute address. If @code{.} is used inside a section description
4356 however, it refers to the byte offset from the start of that section,
4357 not an absolute address. Thus in a script like this:
4375 The @samp{.text} section will be assigned a starting address of 0x100
4376 and a size of exactly 0x200 bytes, even if there is not enough data in
4377 the @samp{.text} input sections to fill this area. (If there is too
4378 much data, an error will be produced because this would be an attempt to
4379 move @code{.} backwards). The @samp{.data} section will start at 0x500
4380 and it will have an extra 0x600 bytes worth of space after the end of
4381 the values from the @samp{.data} input sections and before the end of
4382 the @samp{.data} output section itself.
4386 @subsection Operators
4387 @cindex operators for arithmetic
4388 @cindex arithmetic operators
4389 @cindex precedence in expressions
4390 The linker recognizes the standard C set of arithmetic operators, with
4391 the standard bindings and precedence levels:
4394 @c END TEXI2ROFF-KILL
4396 precedence associativity Operators Notes
4402 5 left == != > < <= >=
4408 11 right &= += -= *= /= (2)
4412 (1) Prefix operators
4413 (2) @xref{Assignments}.
4417 \vskip \baselineskip
4418 %"lispnarrowing" is the extra indent used generally for smallexample
4419 \hskip\lispnarrowing\vbox{\offinterlineskip
4422 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
4423 height2pt&\omit&&\omit&&\omit&\cr
4424 &Precedence&& Associativity &&{\rm Operators}&\cr
4425 height2pt&\omit&&\omit&&\omit&\cr
4427 height2pt&\omit&&\omit&&\omit&\cr
4429 % '176 is tilde, '~' in tt font
4430 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
4431 &2&&left&&* / \%&\cr
4434 &5&&left&&== != > < <= >=&\cr
4437 &8&&left&&{\&\&}&\cr
4440 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
4442 height2pt&\omit&&\omit&&\omit&\cr}
4447 @obeylines@parskip=0pt@parindent=0pt
4448 @dag@quad Prefix operators.
4449 @ddag@quad @xref{Assignments}.
4452 @c END TEXI2ROFF-KILL
4455 @subsection Evaluation
4456 @cindex lazy evaluation
4457 @cindex expression evaluation order
4458 The linker evaluates expressions lazily. It only computes the value of
4459 an expression when absolutely necessary.
4461 The linker needs some information, such as the value of the start
4462 address of the first section, and the origins and lengths of memory
4463 regions, in order to do any linking at all. These values are computed
4464 as soon as possible when the linker reads in the linker script.
4466 However, other values (such as symbol values) are not known or needed
4467 until after storage allocation. Such values are evaluated later, when
4468 other information (such as the sizes of output sections) is available
4469 for use in the symbol assignment expression.
4471 The sizes of sections cannot be known until after allocation, so
4472 assignments dependent upon these are not performed until after
4475 Some expressions, such as those depending upon the location counter
4476 @samp{.}, must be evaluated during section allocation.
4478 If the result of an expression is required, but the value is not
4479 available, then an error results. For example, a script like the
4485 .text 9+this_isnt_constant :
4491 will cause the error message @samp{non constant expression for initial
4494 @node Expression Section
4495 @subsection The Section of an Expression
4496 @cindex expression sections
4497 @cindex absolute expressions
4498 @cindex relative expressions
4499 @cindex absolute and relocatable symbols
4500 @cindex relocatable and absolute symbols
4501 @cindex symbols, relocatable and absolute
4502 When the linker evaluates an expression, the result is either absolute
4503 or relative to some section. A relative expression is expressed as a
4504 fixed offset from the base of a section.
4506 The position of the expression within the linker script determines
4507 whether it is absolute or relative. An expression which appears within
4508 an output section definition is relative to the base of the output
4509 section. An expression which appears elsewhere will be absolute.
4511 A symbol set to a relative expression will be relocatable if you request
4512 relocatable output using the @samp{-r} option. That means that a
4513 further link operation may change the value of the symbol. The symbol's
4514 section will be the section of the relative expression.
4516 A symbol set to an absolute expression will retain the same value
4517 through any further link operation. The symbol will be absolute, and
4518 will not have any particular associated section.
4520 You can use the builtin function @code{ABSOLUTE} to force an expression
4521 to be absolute when it would otherwise be relative. For example, to
4522 create an absolute symbol set to the address of the end of the output
4523 section @samp{.data}:
4527 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
4531 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
4532 @samp{.data} section.
4534 @node Builtin Functions
4535 @subsection Builtin Functions
4536 @cindex functions in expressions
4537 The linker script language includes a number of builtin functions for
4538 use in linker script expressions.
4541 @item ABSOLUTE(@var{exp})
4542 @kindex ABSOLUTE(@var{exp})
4543 @cindex expression, absolute
4544 Return the absolute (non-relocatable, as opposed to non-negative) value
4545 of the expression @var{exp}. Primarily useful to assign an absolute
4546 value to a symbol within a section definition, where symbol values are
4547 normally section relative. @xref{Expression Section}.
4549 @item ADDR(@var{section})
4550 @kindex ADDR(@var{section})
4551 @cindex section address in expression
4552 Return the absolute address (the VMA) of the named @var{section}. Your
4553 script must previously have defined the location of that section. In
4554 the following example, @code{symbol_1} and @code{symbol_2} are assigned
4561 start_of_output_1 = ABSOLUTE(.);
4566 symbol_1 = ADDR(.output1);
4567 symbol_2 = start_of_output_1;
4573 @item ALIGN(@var{align})
4574 @itemx ALIGN(@var{exp},@var{align})
4575 @kindex ALIGN(@var{align})
4576 @kindex ALIGN(@var{exp},@var{align})
4577 @cindex round up location counter
4578 @cindex align location counter
4579 @cindex round up expression
4580 @cindex align expression
4581 Return the location counter (@code{.}) or arbitrary expression aligned
4582 to the next @var{align} boundary. The single operand @code{ALIGN}
4583 doesn't change the value of the location counter---it just does
4584 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
4585 expression to be aligned upwards (@code{ALIGN(@var{align})} is
4586 equivalent to @code{ALIGN(., @var{align})}).
4588 Here is an example which aligns the output @code{.data} section to the
4589 next @code{0x2000} byte boundary after the preceding section and sets a
4590 variable within the section to the next @code{0x8000} boundary after the
4595 .data ALIGN(0x2000): @{
4597 variable = ALIGN(0x8000);
4603 The first use of @code{ALIGN} in this example specifies the location of
4604 a section because it is used as the optional @var{address} attribute of
4605 a section definition (@pxref{Output Section Address}). The second use
4606 of @code{ALIGN} is used to defines the value of a symbol.
4608 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
4610 @item BLOCK(@var{exp})
4611 @kindex BLOCK(@var{exp})
4612 This is a synonym for @code{ALIGN}, for compatibility with older linker
4613 scripts. It is most often seen when setting the address of an output
4616 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4617 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4618 This is equivalent to either
4620 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
4624 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
4627 depending on whether the latter uses fewer @var{commonpagesize} sized pages
4628 for the data segment (area between the result of this expression and
4629 @code{DATA_SEGMENT_END}) than the former or not.
4630 If the latter form is used, it means @var{commonpagesize} bytes of runtime
4631 memory will be saved at the expense of up to @var{commonpagesize} wasted
4632 bytes in the on-disk file.
4634 This expression can only be used directly in @code{SECTIONS} commands, not in
4635 any output section descriptions and only once in the linker script.
4636 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
4637 be the system page size the object wants to be optimized for (while still
4638 working on system page sizes up to @var{maxpagesize}).
4643 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
4646 @item DATA_SEGMENT_END(@var{exp})
4647 @kindex DATA_SEGMENT_END(@var{exp})
4648 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
4649 evaluation purposes.
4652 . = DATA_SEGMENT_END(.);
4655 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
4656 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
4657 This defines the end of the @code{PT_GNU_RELRO} segment when
4658 @samp{-z relro} option is used. Second argument is returned.
4659 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
4660 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
4661 @var{exp} + @var{offset} is aligned to the most commonly used page
4662 boundary for particular target. If present in the linker script,
4663 it must always come in between @code{DATA_SEGMENT_ALIGN} and
4664 @code{DATA_SEGMENT_END}.
4667 . = DATA_SEGMENT_RELRO_END(24, .);
4670 @item DEFINED(@var{symbol})
4671 @kindex DEFINED(@var{symbol})
4672 @cindex symbol defaults
4673 Return 1 if @var{symbol} is in the linker global symbol table and is
4674 defined before the statement using DEFINED in the script, otherwise
4675 return 0. You can use this function to provide
4676 default values for symbols. For example, the following script fragment
4677 shows how to set a global symbol @samp{begin} to the first location in
4678 the @samp{.text} section---but if a symbol called @samp{begin} already
4679 existed, its value is preserved:
4685 begin = DEFINED(begin) ? begin : . ;
4693 @item LENGTH(@var{memory})
4694 @kindex LENGTH(@var{memory})
4695 Return the length of the memory region named @var{memory}.
4697 @item LOADADDR(@var{section})
4698 @kindex LOADADDR(@var{section})
4699 @cindex section load address in expression
4700 Return the absolute LMA of the named @var{section}. This is normally
4701 the same as @code{ADDR}, but it may be different if the @code{AT}
4702 attribute is used in the output section definition (@pxref{Output
4706 @item MAX(@var{exp1}, @var{exp2})
4707 Returns the maximum of @var{exp1} and @var{exp2}.
4710 @item MIN(@var{exp1}, @var{exp2})
4711 Returns the minimum of @var{exp1} and @var{exp2}.
4713 @item NEXT(@var{exp})
4714 @kindex NEXT(@var{exp})
4715 @cindex unallocated address, next
4716 Return the next unallocated address that is a multiple of @var{exp}.
4717 This function is closely related to @code{ALIGN(@var{exp})}; unless you
4718 use the @code{MEMORY} command to define discontinuous memory for the
4719 output file, the two functions are equivalent.
4721 @item ORIGIN(@var{memory})
4722 @kindex ORIGIN(@var{memory})
4723 Return the origin of the memory region named @var{memory}.
4725 @item SEGMENT_START(@var{segment}, @var{default})
4726 @kindex SEGMENT_START(@var{segment}, @var{default})
4727 Return the base address of the named @var{segment}. If an explicit
4728 value has been given for this segment (with a command-line @samp{-T}
4729 option) that value will be returned; otherwise the value will be
4730 @var{default}. At present, the @samp{-T} command-line option can only
4731 be used to set the base address for the ``text'', ``data'', and
4732 ``bss'' sections, but you use @code{SEGMENT_START} with any segment
4735 @item SIZEOF(@var{section})
4736 @kindex SIZEOF(@var{section})
4737 @cindex section size
4738 Return the size in bytes of the named @var{section}, if that section has
4739 been allocated. If the section has not been allocated when this is
4740 evaluated, the linker will report an error. In the following example,
4741 @code{symbol_1} and @code{symbol_2} are assigned identical values:
4750 symbol_1 = .end - .start ;
4751 symbol_2 = SIZEOF(.output);
4756 @item SIZEOF_HEADERS
4757 @itemx sizeof_headers
4758 @kindex SIZEOF_HEADERS
4760 Return the size in bytes of the output file's headers. This is
4761 information which appears at the start of the output file. You can use
4762 this number when setting the start address of the first section, if you
4763 choose, to facilitate paging.
4765 @cindex not enough room for program headers
4766 @cindex program headers, not enough room
4767 When producing an ELF output file, if the linker script uses the
4768 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
4769 number of program headers before it has determined all the section
4770 addresses and sizes. If the linker later discovers that it needs
4771 additional program headers, it will report an error @samp{not enough
4772 room for program headers}. To avoid this error, you must avoid using
4773 the @code{SIZEOF_HEADERS} function, or you must rework your linker
4774 script to avoid forcing the linker to use additional program headers, or
4775 you must define the program headers yourself using the @code{PHDRS}
4776 command (@pxref{PHDRS}).
4779 @node Implicit Linker Scripts
4780 @section Implicit Linker Scripts
4781 @cindex implicit linker scripts
4782 If you specify a linker input file which the linker can not recognize as
4783 an object file or an archive file, it will try to read the file as a
4784 linker script. If the file can not be parsed as a linker script, the
4785 linker will report an error.
4787 An implicit linker script will not replace the default linker script.
4789 Typically an implicit linker script would contain only symbol
4790 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
4793 Any input files read because of an implicit linker script will be read
4794 at the position in the command line where the implicit linker script was
4795 read. This can affect archive searching.
4798 @node Machine Dependent
4799 @chapter Machine Dependent Features
4801 @cindex machine dependencies
4802 @command{ld} has additional features on some platforms; the following
4803 sections describe them. Machines where @command{ld} has no additional
4804 functionality are not listed.
4808 * H8/300:: @command{ld} and the H8/300
4811 * i960:: @command{ld} and the Intel 960 family
4814 * ARM:: @command{ld} and the ARM family
4817 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
4820 * MMIX:: @command{ld} and MMIX
4823 * MSP430:: @command{ld} and MSP430
4826 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
4829 * TI COFF:: @command{ld} and TI COFF
4832 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
4835 * Xtensa:: @command{ld} and Xtensa Processors
4846 @section @command{ld} and the H8/300
4848 @cindex H8/300 support
4849 For the H8/300, @command{ld} can perform these global optimizations when
4850 you specify the @samp{--relax} command-line option.
4853 @cindex relaxing on H8/300
4854 @item relaxing address modes
4855 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
4856 targets are within eight bits, and turns them into eight-bit
4857 program-counter relative @code{bsr} and @code{bra} instructions,
4860 @cindex synthesizing on H8/300
4861 @item synthesizing instructions
4862 @c FIXME: specifically mov.b, or any mov instructions really?
4863 @command{ld} finds all @code{mov.b} instructions which use the
4864 sixteen-bit absolute address form, but refer to the top
4865 page of memory, and changes them to use the eight-bit address form.
4866 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
4867 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
4868 top page of memory).
4870 @item bit manipulation instructions
4871 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
4872 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
4873 which use 32 bit and 16 bit absolute address form, but refer to the top
4874 page of memory, and changes them to use the 8 bit address form.
4875 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
4876 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
4877 the top page of memory).
4879 @item system control instructions
4880 @command{ld} finds all @code{ldc.w, stc.w} instrcutions which use the
4881 32 bit absolute address form, but refer to the top page of memory, and
4882 changes them to use 16 bit address form.
4883 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
4884 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
4885 the top page of memory).
4895 @c This stuff is pointless to say unless you're especially concerned
4896 @c with Renesas chips; don't enable it for generic case, please.
4898 @chapter @command{ld} and Other Renesas Chips
4900 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
4901 H8/500, and SH chips. No special features, commands, or command-line
4902 options are required for these chips.
4912 @section @command{ld} and the Intel 960 Family
4914 @cindex i960 support
4916 You can use the @samp{-A@var{architecture}} command line option to
4917 specify one of the two-letter names identifying members of the 960
4918 family; the option specifies the desired output target, and warns of any
4919 incompatible instructions in the input files. It also modifies the
4920 linker's search strategy for archive libraries, to support the use of
4921 libraries specific to each particular architecture, by including in the
4922 search loop names suffixed with the string identifying the architecture.
4924 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
4925 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
4926 paths, and in any paths you specify with @samp{-L}) for a library with
4939 The first two possibilities would be considered in any event; the last
4940 two are due to the use of @w{@samp{-ACA}}.
4942 You can meaningfully use @samp{-A} more than once on a command line, since
4943 the 960 architecture family allows combination of target architectures; each
4944 use will add another pair of name variants to search for when @w{@samp{-l}}
4945 specifies a library.
4947 @cindex @option{--relax} on i960
4948 @cindex relaxing on i960
4949 @command{ld} supports the @samp{--relax} option for the i960 family. If
4950 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
4951 @code{calx} instructions whose targets are within 24 bits, and turns
4952 them into 24-bit program-counter relative @code{bal} and @code{cal}
4953 instructions, respectively. @command{ld} also turns @code{cal}
4954 instructions into @code{bal} instructions when it determines that the
4955 target subroutine is a leaf routine (that is, the target subroutine does
4956 not itself call any subroutines).
4973 @node M68HC11/68HC12
4974 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
4976 @cindex M68HC11 and 68HC12 support
4978 @subsection Linker Relaxation
4980 For the Motorola 68HC11, @command{ld} can perform these global
4981 optimizations when you specify the @samp{--relax} command-line option.
4984 @cindex relaxing on M68HC11
4985 @item relaxing address modes
4986 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
4987 targets are within eight bits, and turns them into eight-bit
4988 program-counter relative @code{bsr} and @code{bra} instructions,
4991 @command{ld} also looks at all 16-bit extended addressing modes and
4992 transforms them in a direct addressing mode when the address is in
4993 page 0 (between 0 and 0x0ff).
4995 @item relaxing gcc instruction group
4996 When @command{gcc} is called with @option{-mrelax}, it can emit group
4997 of instructions that the linker can optimize to use a 68HC11 direct
4998 addressing mode. These instructions consists of @code{bclr} or
4999 @code{bset} instructions.
5003 @subsection Trampoline Generation
5005 @cindex trampoline generation on M68HC11
5006 @cindex trampoline generation on M68HC12
5007 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
5008 call a far function using a normal @code{jsr} instruction. The linker
5009 will also change the relocation to some far function to use the
5010 trampoline address instead of the function address. This is typically the
5011 case when a pointer to a function is taken. The pointer will in fact
5012 point to the function trampoline.
5020 @section @command{ld} and the ARM family
5022 @cindex ARM interworking support
5023 @kindex --support-old-code
5024 For the ARM, @command{ld} will generate code stubs to allow functions calls
5025 betweem ARM and Thumb code. These stubs only work with code that has
5026 been compiled and assembled with the @samp{-mthumb-interwork} command
5027 line option. If it is necessary to link with old ARM object files or
5028 libraries, which have not been compiled with the -mthumb-interwork
5029 option then the @samp{--support-old-code} command line switch should be
5030 given to the linker. This will make it generate larger stub functions
5031 which will work with non-interworking aware ARM code. Note, however,
5032 the linker does not support generating stubs for function calls to
5033 non-interworking aware Thumb code.
5035 @cindex thumb entry point
5036 @cindex entry point, thumb
5037 @kindex --thumb-entry=@var{entry}
5038 The @samp{--thumb-entry} switch is a duplicate of the generic
5039 @samp{--entry} switch, in that it sets the program's starting address.
5040 But it also sets the bottom bit of the address, so that it can be
5041 branched to using a BX instruction, and the program will start
5042 executing in Thumb mode straight away.
5046 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
5047 executables. This option is only valid when linking big-endian objects.
5048 The resulting image will contain big-endian data and little-endian code.
5051 @kindex --target1-rel
5052 @kindex --target1-abs
5053 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
5054 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
5055 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
5056 and @samp{--target1-abs} switches override the default.
5059 @kindex --target2=@var{type}
5060 The @samp{--target2=type} switch overrides the default definition of the
5061 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
5062 meanings, and target defaults are as follows:
5065 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
5067 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
5069 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
5083 @section @command{ld} and HPPA 32-bit ELF Support
5084 @cindex HPPA multiple sub-space stubs
5085 @kindex --multi-subspace
5086 When generating a shared library, @command{ld} will by default generate
5087 import stubs suitable for use with a single sub-space application.
5088 The @samp{--multi-subspace} switch causes @command{ld} to generate export
5089 stubs, and different (larger) import stubs suitable for use with
5090 multiple sub-spaces.
5092 @cindex HPPA stub grouping
5093 @kindex --stub-group-size=@var{N}
5094 Long branch stubs and import/export stubs are placed by @command{ld} in
5095 stub sections located between groups of input sections.
5096 @samp{--stub-group-size} specifies the maximum size of a group of input
5097 sections handled by one stub section. Since branch offsets are signed,
5098 a stub section may serve two groups of input sections, one group before
5099 the stub section, and one group after it. However, when using
5100 conditional branches that require stubs, it may be better (for branch
5101 prediction) that stub sections only serve one group of input sections.
5102 A negative value for @samp{N} chooses this scheme, ensuring that
5103 branches to stubs always use a negative offset. Two special values of
5104 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
5105 @command{ld} to automatically size input section groups for the branch types
5106 detected, with the same behaviour regarding stub placement as other
5107 positive or negative values of @samp{N} respectively.
5109 Note that @samp{--stub-group-size} does not split input sections. A
5110 single input section larger than the group size specified will of course
5111 create a larger group (of one section). If input sections are too
5112 large, it may not be possible for a branch to reach its stub.
5125 @section @code{ld} and MMIX
5126 For MMIX, there is a choice of generating @code{ELF} object files or
5127 @code{mmo} object files when linking. The simulator @code{mmix}
5128 understands the @code{mmo} format. The binutils @code{objcopy} utility
5129 can translate between the two formats.
5131 There is one special section, the @samp{.MMIX.reg_contents} section.
5132 Contents in this section is assumed to correspond to that of global
5133 registers, and symbols referring to it are translated to special symbols,
5134 equal to registers. In a final link, the start address of the
5135 @samp{.MMIX.reg_contents} section corresponds to the first allocated
5136 global register multiplied by 8. Register @code{$255} is not included in
5137 this section; it is always set to the program entry, which is at the
5138 symbol @code{Main} for @code{mmo} files.
5140 Symbols with the prefix @code{__.MMIX.start.}, for example
5141 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special;
5142 there must be only one each, even if they are local. The default linker
5143 script uses these to set the default start address of a section.
5145 Initial and trailing multiples of zero-valued 32-bit words in a section,
5146 are left out from an mmo file.
5159 @section @code{ld} and MSP430
5160 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
5161 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
5162 just pass @samp{-m help} option to the linker).
5164 @cindex MSP430 extra sections
5165 The linker will recognize some extra sections which are MSP430 specific:
5168 @item @samp{.vectors}
5169 Defines a portion of ROM where interrupt vectors located.
5171 @item @samp{.bootloader}
5172 Defines the bootloader portion of the ROM (if applicable). Any code
5173 in this section will be uploaded to the MPU.
5175 @item @samp{.infomem}
5176 Defines an information memory section (if applicable). Any code in
5177 this section will be uploaded to the MPU.
5179 @item @samp{.infomemnobits}
5180 This is the same as the @samp{.infomem} section except that any code
5181 in this section will not be uploaded to the MPU.
5183 @item @samp{.noinit}
5184 Denotes a portion of RAM located above @samp{.bss} section.
5186 The last two sections are used by gcc.
5200 @section @command{ld}'s Support for Various TI COFF Versions
5201 @cindex TI COFF versions
5202 @kindex --format=@var{version}
5203 The @samp{--format} switch allows selection of one of the various
5204 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
5205 also supported. The TI COFF versions also vary in header byte-order
5206 format; @command{ld} will read any version or byte order, but the output
5207 header format depends on the default specified by the specific target.
5220 @section @command{ld} and WIN32 (cygwin/mingw)
5222 This section describes some of the win32 specific @command{ld} issues.
5223 See @ref{Options,,Command Line Options} for detailed decription of the
5224 command line options mentioned here.
5227 @cindex import libraries
5228 @item import libraries
5229 The standard Windows linker creates and uses so-called import
5230 libraries, which contains information for linking to dll's. They are
5231 regular static archives and are handled as any other static
5232 archive. The cygwin and mingw ports of @command{ld} have specific
5233 support for creating such libraries provided with the
5234 @samp{--out-implib} command line option.
5236 @item exporting DLL symbols
5237 @cindex exporting DLL symbols
5238 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
5241 @item using auto-export functionality
5242 @cindex using auto-export functionality
5243 By default @command{ld} exports symbols with the auto-export functionality,
5244 which is controlled by the following command line options:
5247 @item --export-all-symbols [This is the default]
5248 @item --exclude-symbols
5249 @item --exclude-libs
5252 If, however, @samp{--export-all-symbols} is not given explicitly on the
5253 command line, then the default auto-export behavior will be @emph{disabled}
5254 if either of the following are true:
5257 @item A DEF file is used.
5258 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
5261 @item using a DEF file
5262 @cindex using a DEF file
5263 Another way of exporting symbols is using a DEF file. A DEF file is
5264 an ASCII file containing definitions of symbols which should be
5265 exported when a dll is created. Usually it is named @samp{<dll
5266 name>.def} and is added as any other object file to the linker's
5267 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
5270 gcc -o <output> <objectfiles> <dll name>.def
5273 Using a DEF file turns off the normal auto-export behavior, unless the
5274 @samp{--export-all-symbols} option is also used.
5276 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
5279 LIBRARY "xyz.dll" BASE=0x10000000
5287 This example defines a base address and three symbols. The third
5288 symbol is an alias for the second. For the complete format
5289 specification see ld/deffilep.y in the binutils sources.
5291 @cindex creating a DEF file
5292 While linking a shared dll, @command{ld} is able to create a DEF file
5293 with the @samp{--output-def <file>} command line option.
5295 @item Using decorations
5296 @cindex Using decorations
5297 Another way of marking symbols for export is to modify the source code
5298 itself, so that when building the DLL each symbol to be exported is
5302 __declspec(dllexport) int a_variable
5303 __declspec(dllexport) void a_function(int with_args)
5306 All such symbols will be exported from the DLL. If, however,
5307 any of the object files in the DLL contain symbols decorated in
5308 this way, then the normal auto-export behavior is disabled, unless
5309 the @samp{--export-all-symbols} option is also used.
5311 Note that object files that wish to access these symbols must @emph{not}
5312 decorate them with dllexport. Instead, they should use dllimport,
5316 __declspec(dllimport) int a_variable
5317 __declspec(dllimport) void a_function(int with_args)
5320 This complicates the structure of library header files, because
5321 when included by the library itself the header must declare the
5322 variables and functions as dllexport, but when included by client
5323 code the header must declare them as dllimport. There are a number
5324 of idioms that are typically used to do this; often client code can
5325 omit the __declspec() declaration completely. See
5326 @samp{--enable-auto-import} and @samp{automatic data imports} for more
5330 @cindex automatic data imports
5331 @item automatic data imports
5332 The standard Windows dll format supports data imports from dlls only
5333 by adding special decorations (dllimport/dllexport), which let the
5334 compiler produce specific assembler instructions to deal with this
5335 issue. This increases the effort necessary to port existing Un*x
5336 code to these platforms, especially for large
5337 c++ libraries and applications. The auto-import feature, which was
5338 initially provided by Paul Sokolovsky, allows one to omit the
5339 decorations to archieve a behavior that conforms to that on POSIX/Un*x
5340 platforms. This feature is enabled with the @samp{--enable-auto-import}
5341 command-line option, although it is enabled by default on cygwin/mingw.
5342 The @samp{--enable-auto-import} option itself now serves mainly to
5343 suppress any warnings that are ordinarily emitted when linked objects
5344 trigger the feature's use.
5346 auto-import of variables does not always work flawlessly without
5347 additional assistance. Sometimes, you will see this message
5349 "variable '<var>' can't be auto-imported. Please read the
5350 documentation for ld's @code{--enable-auto-import} for details."
5352 The @samp{--enable-auto-import} documentation explains why this error
5353 occurs, and several methods that can be used to overcome this difficulty.
5354 One of these methods is the @emph{runtime pseudo-relocs} feature, described
5357 @cindex runtime pseudo-relocation
5358 For complex variables imported from DLLs (such as structs or classes),
5359 object files typically contain a base address for the variable and an
5360 offset (@emph{addend}) within the variable--to specify a particular
5361 field or public member, for instance. Unfortunately, the runtime loader used
5362 in win32 environments is incapable of fixing these references at runtime
5363 without the additional information supplied by dllimport/dllexport decorations.
5364 The standard auto-import feature described above is unable to resolve these
5367 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
5368 be resolved without error, while leaving the task of adjusting the references
5369 themselves (with their non-zero addends) to specialized code provided by the
5370 runtime environment. Recent versions of the cygwin and mingw environments and
5371 compilers provide this runtime support; older versions do not. However, the
5372 support is only necessary on the developer's platform; the compiled result will
5373 run without error on an older system.
5375 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
5378 @cindex direct linking to a dll
5379 @item direct linking to a dll
5380 The cygwin/mingw ports of @command{ld} support the direct linking,
5381 including data symbols, to a dll without the usage of any import
5382 libraries. This is much faster and uses much less memory than does the
5383 traditional import library method, expecially when linking large
5384 libraries or applications. When @command{ld} creates an import lib, each
5385 function or variable exported from the dll is stored in its own bfd, even
5386 though a single bfd could contain many exports. The overhead involved in
5387 storing, loading, and processing so many bfd's is quite large, and explains the
5388 tremendous time, memory, and storage needed to link against particularly
5389 large or complex libraries when using import libs.
5391 Linking directly to a dll uses no extra command-line switches other than
5392 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
5393 of names to match each library. All that is needed from the developer's
5394 perspective is an understanding of this search, in order to force ld to
5395 select the dll instead of an import library.
5398 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
5399 to find, in the first directory of its search path,
5410 before moving on to the next directory in the search path.
5412 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
5413 where @samp{<prefix>} is set by the @command{ld} option
5414 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
5415 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
5418 Other win32-based unix environments, such as mingw or pw32, may use other
5419 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
5420 was originally intended to help avoid name conflicts among dll's built for the
5421 various win32/un*x environments, so that (for example) two versions of a zlib dll
5422 could coexist on the same machine.
5424 The generic cygwin/mingw path layout uses a @samp{bin} directory for
5425 applications and dll's and a @samp{lib} directory for the import
5426 libraries (using cygwin nomenclature):
5432 libxxx.dll.a (in case of dll's)
5433 libxxx.a (in case of static archive)
5436 Linking directly to a dll without using the import library can be
5439 1. Use the dll directly by adding the @samp{bin} path to the link line
5441 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
5444 However, as the dll's often have version numbers appended to their names
5445 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
5446 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
5447 not versioned, and do not have this difficulty.
5449 2. Create a symbolic link from the dll to a file in the @samp{lib}
5450 directory according to the above mentioned search pattern. This
5451 should be used to avoid unwanted changes in the tools needed for
5455 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
5458 Then you can link without any make environment changes.
5461 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
5464 This technique also avoids the version number problems, because the following is
5471 libxxx.dll.a -> ../bin/cygxxx-5.dll
5474 Linking directly to a dll without using an import lib will work
5475 even when auto-import features are exercised, and even when
5476 @samp{--enable-runtime-pseudo-relocs} is used.
5478 Given the improvements in speed and memory usage, one might justifiably
5479 wonder why import libraries are used at all. There are two reasons:
5481 1. Until recently, the link-directly-to-dll functionality did @emph{not}
5482 work with auto-imported data.
5484 2. Sometimes it is necessary to include pure static objects within the
5485 import library (which otherwise contains only bfd's for indirection
5486 symbols that point to the exports of a dll). Again, the import lib
5487 for the cygwin kernel makes use of this ability, and it is not
5488 possible to do this without an import lib.
5490 So, import libs are not going away. But the ability to replace
5491 true import libs with a simple symbolic link to (or a copy of)
5492 a dll, in most cases, is a useful addition to the suite of tools
5493 binutils makes available to the win32 developer. Given the
5494 massive improvements in memory requirements during linking, storage
5495 requirements, and linking speed, we expect that many developers
5496 will soon begin to use this feature whenever possible.
5498 @item symbol aliasing
5500 @item adding additional names
5501 Sometimes, it is useful to export symbols with additional names.
5502 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
5503 exported as @samp{_foo} by using special directives in the DEF file
5504 when creating the dll. This will affect also the optional created
5505 import library. Consider the following DEF file:
5508 LIBRARY "xyz.dll" BASE=0x61000000
5515 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
5517 Another method for creating a symbol alias is to create it in the
5518 source code using the "weak" attribute:
5521 void foo () @{ /* Do something. */; @}
5522 void _foo () __attribute__ ((weak, alias ("foo")));
5525 See the gcc manual for more information about attributes and weak
5528 @item renaming symbols
5529 Sometimes it is useful to rename exports. For instance, the cygwin
5530 kernel does this regularly. A symbol @samp{_foo} can be exported as
5531 @samp{foo} but not as @samp{_foo} by using special directives in the
5532 DEF file. (This will also affect the import library, if it is
5533 created). In the following example:
5536 LIBRARY "xyz.dll" BASE=0x61000000
5542 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
5546 Note: using a DEF file disables the default auto-export behavior,
5547 unless the @samp{--export-all-symbols} command line option is used.
5548 If, however, you are trying to rename symbols, then you should list
5549 @emph{all} desired exports in the DEF file, including the symbols
5550 that are not being renamed, and do @emph{not} use the
5551 @samp{--export-all-symbols} option. If you list only the
5552 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
5553 to handle the other symbols, then the both the new names @emph{and}
5554 the original names for the renamed symbols will be exported.
5555 In effect, you'd be aliasing those symbols, not renaming them,
5556 which is probably not what you wanted.
5558 @cindex weak externals
5559 @item weak externals
5560 The Windows object format, PE, specifies a form of weak symbols called
5561 weak externals. When a weak symbol is linked and the symbol is not
5562 defined, the weak symbol becomes an alias for some other symbol. There
5563 are three variants of weak externals:
5565 @item Definition is searched for in objects and libraries, historically
5566 called lazy externals.
5567 @item Definition is searched for only in other objects, not in libraries.
5568 This form is not presently implemented.
5569 @item No search; the symbol is an alias. This form is not presently
5572 As a GNU extension, weak symbols that do not specify an alternate symbol
5573 are supported. If the symbol is undefined when linking, the symbol
5574 uses a default value.
5588 @section @code{ld} and Xtensa Processors
5590 @cindex Xtensa processors
5591 The default @command{ld} behavior for Xtensa processors is to interpret
5592 @code{SECTIONS} commands so that lists of explicitly named sections in a
5593 specification with a wildcard file will be interleaved when necessary to
5594 keep literal pools within the range of PC-relative load offsets. For
5595 example, with the command:
5607 @command{ld} may interleave some of the @code{.literal}
5608 and @code{.text} sections from different object files to ensure that the
5609 literal pools are within the range of PC-relative load offsets. A valid
5610 interleaving might place the @code{.literal} sections from an initial
5611 group of files followed by the @code{.text} sections of that group of
5612 files. Then, the @code{.literal} sections from the rest of the files
5613 and the @code{.text} sections from the rest of the files would follow.
5615 @cindex @option{--relax} on Xtensa
5616 @cindex relaxing on Xtensa
5617 Relaxation is enabled by default for the Xtensa version of @command{ld} and
5618 provides two important link-time optimizations. The first optimization
5619 is to combine identical literal values to reduce code size. A redundant
5620 literal will be removed and all the @code{L32R} instructions that use it
5621 will be changed to reference an identical literal, as long as the
5622 location of the replacement literal is within the offset range of all
5623 the @code{L32R} instructions. The second optimization is to remove
5624 unnecessary overhead from assembler-generated ``longcall'' sequences of
5625 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
5626 range of direct @code{CALL@var{n}} instructions.
5628 For each of these cases where an indirect call sequence can be optimized
5629 to a direct call, the linker will change the @code{CALLX@var{n}}
5630 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
5631 instruction, and remove the literal referenced by the @code{L32R}
5632 instruction if it is not used for anything else. Removing the
5633 @code{L32R} instruction always reduces code size but can potentially
5634 hurt performance by changing the alignment of subsequent branch targets.
5635 By default, the linker will always preserve alignments, either by
5636 switching some instructions between 24-bit encodings and the equivalent
5637 density instructions or by inserting a no-op in place of the @code{L32R}
5638 instruction that was removed. If code size is more important than
5639 performance, the @option{--size-opt} option can be used to prevent the
5640 linker from widening density instructions or inserting no-ops, except in
5641 a few cases where no-ops are required for correctness.
5643 The following Xtensa-specific command-line options can be used to
5646 @cindex Xtensa options
5650 Since the Xtensa version of @code{ld} enables the @option{--relax} option
5651 by default, the @option{--no-relax} option is provided to disable
5655 When optimizing indirect calls to direct calls, optimize for code size
5656 more than performance. With this option, the linker will not insert
5657 no-ops or widen density instructions to preserve branch target
5658 alignment. There may still be some cases where no-ops are required to
5659 preserve the correctness of the code.
5667 @ifclear SingleFormat
5672 @cindex object file management
5673 @cindex object formats available
5675 The linker accesses object and archive files using the BFD libraries.
5676 These libraries allow the linker to use the same routines to operate on
5677 object files whatever the object file format. A different object file
5678 format can be supported simply by creating a new BFD back end and adding
5679 it to the library. To conserve runtime memory, however, the linker and
5680 associated tools are usually configured to support only a subset of the
5681 object file formats available. You can use @code{objdump -i}
5682 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
5683 list all the formats available for your configuration.
5685 @cindex BFD requirements
5686 @cindex requirements for BFD
5687 As with most implementations, BFD is a compromise between
5688 several conflicting requirements. The major factor influencing
5689 BFD design was efficiency: any time used converting between
5690 formats is time which would not have been spent had BFD not
5691 been involved. This is partly offset by abstraction payback; since
5692 BFD simplifies applications and back ends, more time and care
5693 may be spent optimizing algorithms for a greater speed.
5695 One minor artifact of the BFD solution which you should bear in
5696 mind is the potential for information loss. There are two places where
5697 useful information can be lost using the BFD mechanism: during
5698 conversion and during output. @xref{BFD information loss}.
5701 * BFD outline:: How it works: an outline of BFD
5705 @section How It Works: An Outline of BFD
5706 @cindex opening object files
5707 @include bfdsumm.texi
5710 @node Reporting Bugs
5711 @chapter Reporting Bugs
5712 @cindex bugs in @command{ld}
5713 @cindex reporting bugs in @command{ld}
5715 Your bug reports play an essential role in making @command{ld} reliable.
5717 Reporting a bug may help you by bringing a solution to your problem, or
5718 it may not. But in any case the principal function of a bug report is
5719 to help the entire community by making the next version of @command{ld}
5720 work better. Bug reports are your contribution to the maintenance of
5723 In order for a bug report to serve its purpose, you must include the
5724 information that enables us to fix the bug.
5727 * Bug Criteria:: Have you found a bug?
5728 * Bug Reporting:: How to report bugs
5732 @section Have You Found a Bug?
5733 @cindex bug criteria
5735 If you are not sure whether you have found a bug, here are some guidelines:
5738 @cindex fatal signal
5739 @cindex linker crash
5740 @cindex crash of linker
5742 If the linker gets a fatal signal, for any input whatever, that is a
5743 @command{ld} bug. Reliable linkers never crash.
5745 @cindex error on valid input
5747 If @command{ld} produces an error message for valid input, that is a bug.
5749 @cindex invalid input
5751 If @command{ld} does not produce an error message for invalid input, that
5752 may be a bug. In the general case, the linker can not verify that
5753 object files are correct.
5756 If you are an experienced user of linkers, your suggestions for
5757 improvement of @command{ld} are welcome in any case.
5761 @section How to Report Bugs
5763 @cindex @command{ld} bugs, reporting
5765 A number of companies and individuals offer support for @sc{gnu}
5766 products. If you obtained @command{ld} from a support organization, we
5767 recommend you contact that organization first.
5769 You can find contact information for many support companies and
5770 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
5773 Otherwise, send bug reports for @command{ld} to
5774 @samp{bug-binutils@@gnu.org}.
5776 The fundamental principle of reporting bugs usefully is this:
5777 @strong{report all the facts}. If you are not sure whether to state a
5778 fact or leave it out, state it!
5780 Often people omit facts because they think they know what causes the
5781 problem and assume that some details do not matter. Thus, you might
5782 assume that the name of a symbol you use in an example does not
5783 matter. Well, probably it does not, but one cannot be sure. Perhaps
5784 the bug is a stray memory reference which happens to fetch from the
5785 location where that name is stored in memory; perhaps, if the name
5786 were different, the contents of that location would fool the linker
5787 into doing the right thing despite the bug. Play it safe and give a
5788 specific, complete example. That is the easiest thing for you to do,
5789 and the most helpful.
5791 Keep in mind that the purpose of a bug report is to enable us to fix
5792 the bug if it is new to us. Therefore, always write your bug reports
5793 on the assumption that the bug has not been reported previously.
5795 Sometimes people give a few sketchy facts and ask, ``Does this ring a
5796 bell?'' This cannot help us fix a bug, so it is basically useless. We
5797 respond by asking for enough details to enable us to investigate.
5798 You might as well expedite matters by sending them to begin with.
5800 To enable us to fix the bug, you should include all these things:
5804 The version of @command{ld}. @command{ld} announces it if you start it with
5805 the @samp{--version} argument.
5807 Without this, we will not know whether there is any point in looking for
5808 the bug in the current version of @command{ld}.
5811 Any patches you may have applied to the @command{ld} source, including any
5812 patches made to the @code{BFD} library.
5815 The type of machine you are using, and the operating system name and
5819 What compiler (and its version) was used to compile @command{ld}---e.g.
5823 The command arguments you gave the linker to link your example and
5824 observe the bug. To guarantee you will not omit something important,
5825 list them all. A copy of the Makefile (or the output from make) is
5828 If we were to try to guess the arguments, we would probably guess wrong
5829 and then we might not encounter the bug.
5832 A complete input file, or set of input files, that will reproduce the
5833 bug. It is generally most helpful to send the actual object files
5834 provided that they are reasonably small. Say no more than 10K. For
5835 bigger files you can either make them available by FTP or HTTP or else
5836 state that you are willing to send the object file(s) to whomever
5837 requests them. (Note - your email will be going to a mailing list, so
5838 we do not want to clog it up with large attachments). But small
5839 attachments are best.
5841 If the source files were assembled using @code{gas} or compiled using
5842 @code{gcc}, then it may be OK to send the source files rather than the
5843 object files. In this case, be sure to say exactly what version of
5844 @code{gas} or @code{gcc} was used to produce the object files. Also say
5845 how @code{gas} or @code{gcc} were configured.
5848 A description of what behavior you observe that you believe is
5849 incorrect. For example, ``It gets a fatal signal.''
5851 Of course, if the bug is that @command{ld} gets a fatal signal, then we
5852 will certainly notice it. But if the bug is incorrect output, we might
5853 not notice unless it is glaringly wrong. You might as well not give us
5854 a chance to make a mistake.
5856 Even if the problem you experience is a fatal signal, you should still
5857 say so explicitly. Suppose something strange is going on, such as, your
5858 copy of @command{ld} is out of synch, or you have encountered a bug in the
5859 C library on your system. (This has happened!) Your copy might crash
5860 and ours would not. If you told us to expect a crash, then when ours
5861 fails to crash, we would know that the bug was not happening for us. If
5862 you had not told us to expect a crash, then we would not be able to draw
5863 any conclusion from our observations.
5866 If you wish to suggest changes to the @command{ld} source, send us context
5867 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
5868 @samp{-p} option. Always send diffs from the old file to the new file.
5869 If you even discuss something in the @command{ld} source, refer to it by
5870 context, not by line number.
5872 The line numbers in our development sources will not match those in your
5873 sources. Your line numbers would convey no useful information to us.
5876 Here are some things that are not necessary:
5880 A description of the envelope of the bug.
5882 Often people who encounter a bug spend a lot of time investigating
5883 which changes to the input file will make the bug go away and which
5884 changes will not affect it.
5886 This is often time consuming and not very useful, because the way we
5887 will find the bug is by running a single example under the debugger
5888 with breakpoints, not by pure deduction from a series of examples.
5889 We recommend that you save your time for something else.
5891 Of course, if you can find a simpler example to report @emph{instead}
5892 of the original one, that is a convenience for us. Errors in the
5893 output will be easier to spot, running under the debugger will take
5894 less time, and so on.
5896 However, simplification is not vital; if you do not want to do this,
5897 report the bug anyway and send us the entire test case you used.
5900 A patch for the bug.
5902 A patch for the bug does help us if it is a good one. But do not omit
5903 the necessary information, such as the test case, on the assumption that
5904 a patch is all we need. We might see problems with your patch and decide
5905 to fix the problem another way, or we might not understand it at all.
5907 Sometimes with a program as complicated as @command{ld} it is very hard to
5908 construct an example that will make the program follow a certain path
5909 through the code. If you do not send us the example, we will not be
5910 able to construct one, so we will not be able to verify that the bug is
5913 And if we cannot understand what bug you are trying to fix, or why your
5914 patch should be an improvement, we will not install it. A test case will
5915 help us to understand.
5918 A guess about what the bug is or what it depends on.
5920 Such guesses are usually wrong. Even we cannot guess right about such
5921 things without first using the debugger to find the facts.
5925 @appendix MRI Compatible Script Files
5926 @cindex MRI compatibility
5927 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
5928 linker, @command{ld} can use MRI compatible linker scripts as an
5929 alternative to the more general-purpose linker scripting language
5930 described in @ref{Scripts}. MRI compatible linker scripts have a much
5931 simpler command set than the scripting language otherwise used with
5932 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
5933 linker commands; these commands are described here.
5935 In general, MRI scripts aren't of much use with the @code{a.out} object
5936 file format, since it only has three sections and MRI scripts lack some
5937 features to make use of them.
5939 You can specify a file containing an MRI-compatible script using the
5940 @samp{-c} command-line option.
5942 Each command in an MRI-compatible script occupies its own line; each
5943 command line starts with the keyword that identifies the command (though
5944 blank lines are also allowed for punctuation). If a line of an
5945 MRI-compatible script begins with an unrecognized keyword, @command{ld}
5946 issues a warning message, but continues processing the script.
5948 Lines beginning with @samp{*} are comments.
5950 You can write these commands using all upper-case letters, or all
5951 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
5952 The following list shows only the upper-case form of each command.
5955 @cindex @code{ABSOLUTE} (MRI)
5956 @item ABSOLUTE @var{secname}
5957 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
5958 Normally, @command{ld} includes in the output file all sections from all
5959 the input files. However, in an MRI-compatible script, you can use the
5960 @code{ABSOLUTE} command to restrict the sections that will be present in
5961 your output program. If the @code{ABSOLUTE} command is used at all in a
5962 script, then only the sections named explicitly in @code{ABSOLUTE}
5963 commands will appear in the linker output. You can still use other
5964 input sections (whatever you select on the command line, or using
5965 @code{LOAD}) to resolve addresses in the output file.
5967 @cindex @code{ALIAS} (MRI)
5968 @item ALIAS @var{out-secname}, @var{in-secname}
5969 Use this command to place the data from input section @var{in-secname}
5970 in a section called @var{out-secname} in the linker output file.
5972 @var{in-secname} may be an integer.
5974 @cindex @code{ALIGN} (MRI)
5975 @item ALIGN @var{secname} = @var{expression}
5976 Align the section called @var{secname} to @var{expression}. The
5977 @var{expression} should be a power of two.
5979 @cindex @code{BASE} (MRI)
5980 @item BASE @var{expression}
5981 Use the value of @var{expression} as the lowest address (other than
5982 absolute addresses) in the output file.
5984 @cindex @code{CHIP} (MRI)
5985 @item CHIP @var{expression}
5986 @itemx CHIP @var{expression}, @var{expression}
5987 This command does nothing; it is accepted only for compatibility.
5989 @cindex @code{END} (MRI)
5991 This command does nothing whatever; it's only accepted for compatibility.
5993 @cindex @code{FORMAT} (MRI)
5994 @item FORMAT @var{output-format}
5995 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
5996 language, but restricted to one of these output formats:
6000 S-records, if @var{output-format} is @samp{S}
6003 IEEE, if @var{output-format} is @samp{IEEE}
6006 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
6010 @cindex @code{LIST} (MRI)
6011 @item LIST @var{anything}@dots{}
6012 Print (to the standard output file) a link map, as produced by the
6013 @command{ld} command-line option @samp{-M}.
6015 The keyword @code{LIST} may be followed by anything on the
6016 same line, with no change in its effect.
6018 @cindex @code{LOAD} (MRI)
6019 @item LOAD @var{filename}
6020 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
6021 Include one or more object file @var{filename} in the link; this has the
6022 same effect as specifying @var{filename} directly on the @command{ld}
6025 @cindex @code{NAME} (MRI)
6026 @item NAME @var{output-name}
6027 @var{output-name} is the name for the program produced by @command{ld}; the
6028 MRI-compatible command @code{NAME} is equivalent to the command-line
6029 option @samp{-o} or the general script language command @code{OUTPUT}.
6031 @cindex @code{ORDER} (MRI)
6032 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
6033 @itemx ORDER @var{secname} @var{secname} @var{secname}
6034 Normally, @command{ld} orders the sections in its output file in the
6035 order in which they first appear in the input files. In an MRI-compatible
6036 script, you can override this ordering with the @code{ORDER} command. The
6037 sections you list with @code{ORDER} will appear first in your output
6038 file, in the order specified.
6040 @cindex @code{PUBLIC} (MRI)
6041 @item PUBLIC @var{name}=@var{expression}
6042 @itemx PUBLIC @var{name},@var{expression}
6043 @itemx PUBLIC @var{name} @var{expression}
6044 Supply a value (@var{expression}) for external symbol
6045 @var{name} used in the linker input files.
6047 @cindex @code{SECT} (MRI)
6048 @item SECT @var{secname}, @var{expression}
6049 @itemx SECT @var{secname}=@var{expression}
6050 @itemx SECT @var{secname} @var{expression}
6051 You can use any of these three forms of the @code{SECT} command to
6052 specify the start address (@var{expression}) for section @var{secname}.
6053 If you have more than one @code{SECT} statement for the same
6054 @var{secname}, only the @emph{first} sets the start address.
6065 % I think something like @colophon should be in texinfo. In the
6067 \long\def\colophon{\hbox to0pt{}\vfill
6068 \centerline{The body of this manual is set in}
6069 \centerline{\fontname\tenrm,}
6070 \centerline{with headings in {\bf\fontname\tenbf}}
6071 \centerline{and examples in {\tt\fontname\tentt}.}
6072 \centerline{{\it\fontname\tenit\/} and}
6073 \centerline{{\sl\fontname\tensl\/}}
6074 \centerline{are used for emphasis.}\vfill}
6076 % Blame: doc@cygnus.com, 28mar91.